Documentation
Content
School of Electrical, Electronic & Computer Engineering
EERI 429
Active RFID Feasibility study
S.S. Cox 20285132
Study leader:
Prof. J.E.W. Holm
OCTOBER 2010
Faculty of
Engineering
School for Electronic and Computer Engineering
Acknowledgements
This final year project would not be complete without acknowledging all who contributed. Firstly I would like to thank God for granting me the ability to complete this project; To my fiancé Jorene Bosman, thank you for your patience and understanding; To my parents Stephen and Marina Cox, thank you for sculpting my consciousness in such a unique way and for all your sacrifices; To my study leader Prof. J.E.W. Holm, for the inspiration and creative thinking; To Prof. L. Liebenberg, for inspiration and leadership; To HVAC International, for providing me with the project; To Becker Electronics, for lending me the hardware to complete this project. To Arno de Coning, Andreas Alberts, Andre Botha, Henri Marais and Walter Booysen.
School of Electrical, Electronic and Computer Engineering
ii
School for Electronic and Computer Engineering
Executive summary
Tracking of personnel in mines is critical to the safety of miners, technicians, managers and contractors. Currently, a system is used that makes use of contactless access control cards. However, personnel cannot be tracked once they have gone underground. In order to alleviate this problem, an active RFID solution is proposed. Since the actual operational performance of such a system is not known, it is required for this design project to simulate a mining environment with different scenarios. Such a simulation will allow the system designer to make critical design decisions and plan the layout of a system. A simulation package to this effect does not exist and is a critical need. Therefore, the aim of this project is to create a simulation package that is based on actual hardware, an actual environment, and realistic scenarios in a mining environment. A study of wave propagation in mines was done, available (commercial) hardware was analysed, and a software simulation package was created that allows a designer to plan a tracking system.
School of Electrical, Electronic and Computer Engineering
iii
School for Electronic and Computer Engineering
Declaration
I, Stephen Cox, declare that dissertation is a presentation of my own original work, conducted under the supervision of Johann Holm. Whenever contributions of others are involved, every effort is made to indicate this clearly, with due reference to the literature. No part of this work has been submitted in the past, or is being submitted, for a degree or examination at any other university or course.
Signed on this ___ day of _____________ 2010, in Potchefstroom.
_________________________________ S.S. Cox
School of Electrical, Electronic and Computer Engineering
iv
School for Electronic and Computer Engineering
Table of Contents
CHAPTER 1 - INTRODUCTION ................................................................................................................................................. 1 1.1 1.2 1.3 1.4 1.5 BACKGROUND ............................................................................................................................................................... 2 PROBLEM STATEMENT .................................................................................................................................................. 2
METHODOLOGY ................................................................................................................................................................. 5
PROJECT MANAGEMENT ..................................................................................................................................................... 9 DOCUMENT LAYOUT......................................................................................................................................................... 13
CHAPTER 2 - LITERATURE STUDY .......................................................................................................................................... 14 2.1 2.2 2.3 2.4 2.5 LITERATURE STUDY ........................................................................................................................................................... 15
TECHNOLOGY STUDY ......................................................................................................................................................... 20
EXISTING SYSTEMS ........................................................................................................................................................... 25 SHORTFALL OF THE EXISTING SYSTEMS AND TECHNOLOGIES ........................................................................................... 26 DESIGN PHILOSOPHY ........................................................................................................................................................ 26
CHAPTER 3 – PRELIMINARY DESIGN ..................................................................................................................................... 28 3.1 3.2
FUNCTIONAL ARCHITECTURE ............................................................................................................................................... 29
INTERFACES .................................................................................................................................................................... 30
CHAPTER 4 – DETAIL DESIGN ................................................................................................................................................ 31 4.1 4.2 4.3 4.4 SYSTEM DESIGN .............................................................................................................................................................. 32 MINE SIMULATOR ............................................................................................................................................................ 41 READER SIMULATOR ......................................................................................................................................................... 48 SERVER DISPLAY .............................................................................................................................................................. 53
CHAPTER 5 – VALIDATION .................................................................................................................................................... 58 5.1 5.2 5.3 5.4 5.5 5.6 OVERVIEW ..................................................................................................................................................................... 59 ENVIRONMENT ................................................................................................................................................................ 59 HARDWARE USED ............................................................................................................................................................ 62 TESTS ............................................................................................................................................................................ 62 RESULTS......................................................................................................................................................................... 63 DISCUSSION .................................................................................................................................................................... 66
CHAPTER 6 – IMPLEMENTATION AND EVALUATION ............................................................................................................ 67 6.1 6.2 6.3 IMPLEMENTATION ............................................................................................................................................................ 68 EVALUATION ................................................................................................................................................................... 69 SYSTEM TEST ................................................................................................................................................................... 69
CHAPTER 7 - CONCLUSION AND RECOMMENDATIONS ........................................................................................ 73 7.1 7.2 PROJECT HISTORY - REVISITED ............................................................................................................................................. 74 CONCLUSION .................................................................................................................................................................. 74
School of Electrical, Electronic and Computer Engineering
v
School for Electronic and Computer Engineering 7.3 7.4 RECOMMENDATIONS ........................................................................................................................................................ 75 ECSA OUTCOMES ............................................................................................................................................................ 75
APPENDIX A – TURN IT IN REPORT ....................................................................................................................................... 80
School of Electrical, Electronic and Computer Engineering
vi
School for Electronic and Computer Engineering
List of Figures
Figure 1 : System layout .............................................................................................................. 3 Figure 2 : System design concept that will be simulated .............................................................. 5 Figure 3 : Work breakdown structure ........................................................................................... 6 Figure 4 : Project schedule .......................................................................................................... 7 Figure 5 : Data Model Design ...................................................................................................... 7 Figure 6 : RFID data logger Design ............................................................................................. 8 Figure 7 : Server Software Design ............................................................................................... 8 Figure 8 : Tracking System Design .............................................................................................. 9 Figure 9 : Active RFID relevant to EM spectrum ........................................................................ 16 Figure 10 : Plane wave incident onto a plane boundary ............................................................ 17 Figure 11 : Finding the point of reflection using Fermat‟s principle ............................................ 17 Figure 12 : Refraction ................................................................................................................ 18 Figure 13 : Wave scattering ....................................................................................................... 18 Figure 14 : 2 Path model ............................................................................................................ 19 Figure 15 : Savi ST602 [12] ....................................................................................................... 20 Figure 16 : TCAP 302 ................................................................................................................ 21 Figure 17 : TCAP 302 fitted on a standard mining cap lamp battery .......................................... 21 Figure 18 : Wavetrend TG100-A [5] ........................................................................................... 21 Figure 19 : Ekahau T301A [6] .................................................................................................... 22 Figure 20 : Savi SR-650 [7] ........................................................................................................ 22 Figure 21 : UATR 400 [8] ........................................................................................................... 23 Figure 22 : Wavetrend RX201 [9]............................................................................................... 23 Figure 23 : Wavetrend RX900 [9] ............................................................................................... 24 Figure 24 : Handheld reader [12] ............................................................................................... 25 Figure 25 : Impact infrastructure ................................................................................................ 25 Figure 26 : Functional system overview ..................................................................................... 29 Figure 27 : Detail system overview ............................................................................................ 32 Figure 28 : MineSimulatorDrawing screenshot .......................................................................... 32 Figure 29 : MineSimulatorDrawing's route planner .................................................................... 33
School of Electrical, Electronic and Computer Engineering
vii
School for Electronic and Computer Engineering
Figure 30 : Reader propagation model ...................................................................................... 34 Figure 31 : MineSimulator version 2 .......................................................................................... 35 Figure 32 : Reader Simulator Screenshot .................................................................................. 36 Figure 33 : Data logger script ..................................................................................................... 37 Figure 34 : Server Control software flow .................................................................................... 39 Figure 35 : ServerDisplay version 3 ........................................................................................... 41 Figure 36 : Movement calculation .............................................................................................. 44 Figure 37 : Final reader propagation model ............................................................................... 44 Figure 38 : Simplification of the functional flow of MineSimulator .............................................. 47 Figure 39 : Visual identification indication .................................................................................. 48 Figure 40 : SET tag data ............................................................................................................ 51 Figure 41 : GET tag data............................................................................................................ 51 Figure 42 : Reader Simulator flow .............................................................................................. 52 Figure 43 : ServerDisplay version 1 ........................................................................................... 53 Figure 44 : ServerDisplay........................................................................................................... 54 Figure 45 : Locating algorithm .................................................................................................... 55 Figure 46 : ServerDisplay version 3 ........................................................................................... 56 Figure 47 : Test layout ............................................................................................................... 59 Figure 48 : Section 1 .................................................................................................................. 60 Figure 49 : Section 1 (Left) and section 4 (Right) ....................................................................... 60 Figure 50 : Section 3 .................................................................................................................. 61 Figure 51 : Section 2 (Right) and section 4 (Left) ....................................................................... 61 Figure 52 : RFID Tag placement ................................................................................................ 61 Figure 53 : Backer Electronics HTRD100ZA .............................................................................. 62 Figure 54 : Becker Electronics TPER200WW ............................................................................ 62
School of Electrical, Electronic and Computer Engineering
viii
School for Electronic and Computer Engineering
List of Tables
Table 1 : Preliminary budget ...................................................................................................... 10 Table 2 : Risk table .................................................................................................................... 11 Table 3 : Route calculation list details ........................................................................................ 34 Table 4 : Data logger database structure ................................................................................... 38 Table 5 : Server database structure ........................................................................................... 39
School of Electrical, Electronic and Computer Engineering
ix
School for Electronic and Computer Engineering
List of Abbreviations
RFID ISO IEC GPS SABS NWU ID PC ECSA ELO EMS SCADA API EMPA PHP SQL RTLS POI Radio frequency identification International Organisation for Standardization International Electrotechnical Commission Global Positioning System South-Africa Bureau of Standards North-West University Identification Personal Computer Engineering Council of South-Africa Exit Level Outcome Energy Management System Supervisory Control and Data Acquisition Application Programming Interface Energy Management Personnel Awareness PHP: Hypertext Preprocessor Structured Query Language Real-time locating systems Point of interest
School of Electrical, Electronic and Computer Engineering
x
Chapter 1 - Introduction
CHAPTER 1 - INTRODUCTION
The aim of this chapter is to provide an overview of the design project. This includes making a formal problem statement that forms the basis and motivation of the project. Furthermore, the chapter contains objectives and the scope of the design project. To describe the project a work breakdown structure, design methodology, quality assurance plan, preliminary budget, risk analysis, and ECSA outcome assessment are given.
School of Electrical, Electronic and Computer Engineering
1
Chapter 1 - Introduction 1.1 BACKGROUND In South Africa the mining sector is a major component of the economy. One of the most challenging aspects of mines is safety. One element of safety in mines is the location of people and assets. GPS tracking on open ground is the most common tracking system today. Since GPS uses satellites for tracking it will not be applicable to underground applications. The most commonly used tracking in applications without GPS is localized identification-based tracking. The most commonly used identification tracking is the well-known barcode system. For tracking applications in mines, the barcode system will not be suitable due to physical limitations. The recent advancement and increasing research and development on RFID showed that RFID can be used as an effective tracking mechanism. RFID tracking can be used to increase safety, offer more security, and increase energy efficiency in a mine. In this project, only safety will be focused on although the results of the research will also be used in energy management systems (to know which areas in a mine may be shut down to save energy, it is necessary to know if miners are present in those areas). This project is innovative in the sense that there will be a unique simulation of an active RFID real-time tracking system. This simulation will be used as a layout tool for optimizing a tracking system‟s design. This system will generate data that is otherwise very expensive to acquire from an actual, physical system. 1.2 PROBLEM STATEMENT
Verify the feasibility of an active RFID tracking system for use in the underground mining sector by using software simulation. 1.2.1 Project Scope Project requirements were received from Hades Technologies (Pty) Ltd (the client). Areas to be researched: Existing technologies and systems; Tag and reader functioning; Software simulation procedures; Adjustable parameters; Database software; RF propagation effects;
School of Electrical, Electronic and Computer Engineering
2
Chapter 1 - Introduction RFID RTLS system functioning.
Outcomes to be met when the project is finished: Implementation and demonstration of the system simulation; System adjustability and customisability; Database support and integration support with other software; Software that handle events; Reader propagation model for validation.
The following figure is the author‟s interpretation of the provided specifications and shows the system to be simulated.
Figure 1 : System layout
School of Electrical, Electronic and Computer Engineering
3
Chapter 1 - Introduction 1.2.2 Project Objectives 1.2.2.1 Primary Objectives
Investigate the feasibility of an underground tracking system using active RFID tracking system simulation. 1.2.2.2 Secondary Objectives
Extend the simulator design to verify safety and energy efficiency applications. 1.2.2.3 Critical Success Factors
In order to deliver a successful system, the following must be in place: An RFID tracking system simulation to identify the last known position of personnel; Software that reads events and displays them; Validation of the RFID reader propagation model. 1.2.2.4 Deliverables
The following deliverables shall be delivered per milestone: Milestone 1 – Project proposal documentation; Milestone 2 - Introduction and Literature Study of final documentation; Milestone 3 – Presentation of proposal and conceptual design; Milestone 4 – Chapters 1 to 3 of final documentation; Milestone 5 – Demonstrate project; Milestone 6 – Fully demonstrate project; Milestone 7 – Hand in concept of final documentation; Milestone 8 – Final report. At milestone 5 there will be the following: RFID RTLS concept; Sub system simulation software; System integration concept.
School of Electrical, Electronic and Computer Engineering
4
Chapter 1 - Introduction 1.3 METHODOLOGY 1.3.1 Proposed design methodology
Figure 2 : System design concept that will be simulated
School of Electrical, Electronic and Computer Engineering
5
Chapter 1 - Introduction Systems engineering will be used throughout the project [1]. The following is the design process that will be followed: Definition of need; Need analysis; Preliminary design; Detail design and development. The following is a breakdown of the documentation process: Project proposal; Introduction and literature study; Preliminary design; Final report. 1.3.2 Work Breakdown Structure
NWU Project 2010
Research
Design
Develop
Current RFID Technologies
Tag and Reader functioning
Data model
RFID Tracking System Simulator
Software simulation procedures Radio Frequency Propagation effects
Database Software
Simulator
RFID RTLS System functioning
Figure 3 : Work breakdown structure
School of Electrical, Electronic and Computer Engineering
6
Chapter 1 - Introduction 1.3.3 Schedule Milestone 1 (05-Feb) Milestone 2 (19-Mrt) Milestone 3 (17-Apr) Milestone 4 (01-June) Research (02-June) Milestone 5 (16-Jul) System adjustments (17-Jul) Milestone 6&7 (13-Oct) Milestone 8 (19-Oct)
•Project proposal •Introduction and literature study •Conceptual design •Chapters 1,2 and 3 •System components, software and communication infrastructure •Concept demonstration •Optimisation •Demonstration and concept of final documentation •Final documentation
Figure 4 : Project schedule
1.3.4 Design Proposal The following figures are a graphical illustration of the proposed designs.
Data model
Database software
RFID data protocols
Interface
RFID data logger
Server
Figure 5 : Data Model Design
School of Electrical, Electronic and Computer Engineering
7
Chapter 1 - Introduction
RFID data logger
Hardware
Software
Enclosure
Computer
Network
Data Model
Data acquisition software
Figure 6 : RFID data logger Design
Server
Software
Hardware
Data Model
GUI
Data acquisition software
Computer
Network
Integration with EMS
Tracking results
Figure 7 : Server Software Design
School of Electrical, Electronic and Computer Engineering
8
Chapter 1 - Introduction
RFID Tracking System Hardware Required RFID data logger Single-board Computer
Software
Server
RFID
Database
Computer
Tag
Reader
Data logger
Simulator
Figure 8 : Tracking System Design 1.4 PROJECT MANAGEMENT 1.4.1 Available resources The following people are available to contact in regards of my project: Prof. J.E.W. Holm (Mining technology); Prof. Hoffman (RFID technologies); Mnr. T. Tromp (Becker Electronics); Mr. P. Van Huyssteen (Mining network technology); Mr. H. Marais (Software programming); Mr. A. Alberts (Database design). Available Software SuperNec (Antenna simulation); Matlab (Simulations using large volume mathematics); Microsoft Visual Studio (GUI design); MySQL (Database design). PHP (Data logger software) Available Finances NWU (Maximum R2500);
School of Electrical, Electronic and Computer Engineering
9
Chapter 1 - Introduction Hades Technologies (Pty) Ltd. 1.4.2 Preliminary Budget The following is an approximation of the development costs to verify the simulation by the author: RFID Tag R 330.00 RFID Reader R 13,000.00 RFID data logger Computer R 3,000.00 UPS R 500.00 Casing R 1,000.00 Server (Industrial standard) Computer R 25,000.00 UPS R 5,000.00 Casing R 10,000.00 Table 1 : Preliminary budget
School of Electrical, Electronic and Computer Engineering
10
Chapter 1 - Introduction 1.4.3 Associated risks Calculating a risk factor involves determining a fail probability and a consequence factor. [2] The associated risk is the product of the fail probability and the consequence factor. Both factors are between 0% and 100%, where 0% is no failure and 100% is total failure.
Time [A] System Integration Communication protocol Data model Functional unit component research Cost analysis Application layer Implementation 70% 65% 80% 60% 30% 30% 60% Resources [B] 50% 60% 50% 50% 40% 60% 40% Cost [C] 30% 40% 0% 0% 0% 0% 10% Fail Probability [D=0.4xA+0.35xB+0.25xC] 53% 57% 50% 42% 26% 33% 41% Consequence [E] 100% 70% 75% 90% 100% 75% 60% Risk Factor [F=DxE] 53% 40% 37% 37% 26% 25% 24% Global risk factor 77% 58% 54% 54% 38% 36% 35%
Table 2 : Risk table
Probability of failure is calculated by looking at previous data. This previous data can also be personal experience. Probability of failure is obtained by looking at time-, resourceand cost failures. Each of these factors has a selected weight determined by judging the project circumstances; Time failure is when an item causes delays in the design process or when insufficient time had been allowed within which to complete the project; Resource failure is when people cannot act as resources or when components or equipment fail, or when it is difficult to understand a resource; Cost failure is when an item‟s price or cost fluctuates over time, or an item causes additional expenses (such as services cost); The consequence factor is obtained by determining the impact of the item‟s failure on the project as a whole [2]; The global risk factor is a normalised value when the average of the risk factors is set to 50%. An item that has a risk factor above 50% is a risk above normal and more attention should be paid to that item.
School of Electrical, Electronic and Computer Engineering
11
Chapter 1 - Introduction 1.4.4 Quality assurance The following will determine the quality of the project: RFID software components with minimal user operation; System scalability; Software customisability; Software simulation results; Reader propagation model accuracy; System integration. 1.4.5 ECSA exit level outcomes All relevant information on ECSA ELO‟s was gathered from [3]. 1.4.5.1 Problem solving
In this project there exists a need to be solved. 1.4.5.2 Application of scientific and engineering knowledge
Working with a radio frequency application there will be mathematics and physics involved. 1.4.5.3 Engineering Design
There will be a system design involved in this project. 1.4.5.4 Investigations, experiments and data analysis
For the designs in this project there will be an investigation, experiment and data analysis. 1.4.5.5 Engineering methods, skills and tools
In this project there will be an investigation in mining communication infrastructure. 1.4.5.6 Professional and technical communication
In this project there will be communication with mining professionals and project managers. 1.4.5.7 Impact of engineering activity
This project will increase the safety in mines. 1.4.5.8 Individual, team and multidisciplinary working
In this project there will be interaction with project manager, mining professionals and technicians. The author of this document will be responsible for this project.
School of Electrical, Electronic and Computer Engineering
12
Chapter 1 - Introduction 1.4.5.9 Independent learning ability
The author of this document will work strategically, complete tasks effectively and will deliver work on time. 1.4.5.10 Engineering Professionalism Being professional is a moral decision and will be done throughout the whole project. 1.5 DOCUMENT LAYOUT Documentation started with an introduction that disused that problem and its background. After that a more detailed background was discussed in the literature study. Following the detail background was a preliminary design, which purposed an possible design to solve the problem. This design was discussed in detail after the purposing design was evaluated. After the detail design validation was done to strengthen the design. Then implementation and evaluation was discussed and finally the conclusion was done.
This concludes chapter 1, where a formal problem statement was made and objectives of the project and the scope of the project were defined. A work breakdown structure, design methodology, quality assurance plan, preliminary budget, risk analysis and ECSA outcome assessment were given.
School of Electrical, Electronic and Computer Engineering
13
CHAPTER 2 - LITERATURE STUDY
The aim of this chapter is to give a theoretical background on components of the RFID tracking system. It starts with an overview of existing systems and technologies and moves to design philosophy, and finally discusses the theoretical background of sub components.
School of Electrical, Electronic and Computer Engineering
14
2.1 LITERATURE STUDY 2.1.1 Real time locating systems (RTLS) RTLS‟s are used to identify and track the location of objects. Locating is achieved by attaching tags/badges to an object and then identifying these tags/badges by means of wireless signals at specific known locations the readers are placed. These readers will relay information to a central server. RTLS is a system that collects location information passively. [4] The RTLS standard excludes passive RFID indexing as well as beacon systems. Locating is achieved by the RTLS standard. This involves single/multiple read transactions with the readers, locating algorithm, relay of information and distance measuring between readers. RFID tracking is in a sense just like RTLS except that is does not determine the exact location of the object being tracked. 2.1.2 Electromagnetic (EM) spectrum The EM spectrum is the collection of all EM waves; categorized by frequencies. The behaviour of the waves depends on their frequency. Relevant equations are wavelength ( energy ( E
c f ) and
fh ); where c is the speed of light in a vacuum and h is Planck„s constant. [5]
School of Electrical, Electronic and Computer Engineering
15
Very Low Frequencies (3-30 kHz) Low Frequencies (30-300 kHz) Medium Frequencies (300-3000 kHz) High Frequencies (3-30 MHz) Very High Frequencies (30-300 MHz) Ultra High Frequencies (300-3000 MHz) Super Hign Frequencies (3-30 GHz) Extremenly high Frequencies (30-300 GHz) Terahertz (300+ GHz) Active RFID
Radio
Microwave
Infrared
Electromagnetic spectrum
Visible
Ultraviolet
X-rays
Gamma rays
Figure 9 : Active RFID relevant to EM spectrum 2.1.3 Radio waves The EM spectrum of importance in this document is the radio frequency spectrum. Governments regulate the use of this spectrum through frequency allocation. Collective oscillation of charge carriers in material causes radio waves. Radio waves are utilised by antennas. Antenna size is an important factor; it must be sized according to the principle of resonance. This will determine the bandwidth and optimal frequency on the antenna [6]. Radio waves can modulate information by means of varying the amplitude, frequency and phase of the wave within a defined frequency range. When EM radiation acts upon a conductor it couples to the conductor. The EM radiation travels along the conductor and induces an electric current on the surface by exiting electrons in the conductor; this effect is called the skin effect. [7] These waves are affected by reflection, refraction, diffraction, absorption, polarization and scattering.
School of Electrical, Electronic and Computer Engineering
16
2.1.3.1
Propagation
Some of the most interesting aspects of waves are propagation effects. This is important due to energy losses and time delays in transmission of waves. In underground mining the propagation effects include reflection, refraction, scattering and multipath fading. The effects will be discussed in the following paragraphs. Reflection First we look at lossless scenarios. Depending on the angle of incidence the incoming wave energy is redirected to transmission-, reflection- and absorbed energy. This can be proved by using Snell‟s laws. Figure 10 shows a wave reflecting and transmitting some of its energy. In the figure
r i
; angle of reflection is equal to the angle of incidence. [8]
Figure 10 : Plane wave incident onto a plane boundary Furthermore, if the wave is reflected from a surface it will most likely choose a path as defined in Fermat‟s principle. Thus it is highly likely that a wave will follow the minimum time to the destination. This can be seen in Figure 11.
Figure 11 : Finding the point of reflection using Fermat’s principle 17
School of Electrical, Electronic and Computer Engineering
Refraction Refraction is the phenomenon when it seems as if a wave bends. This effect can normally be seen when light waves move trough different mediums or when observing something that is underwater. The effect is due to difference in refraction indices. This effect can be described using Snell‟s law which states the following:
sin sin
1 2
n1 , where n is the respective refraction n2
index. The effect can be seen in the following figure. [8] [9]
Figure 12 : Refraction Scattering Scattering is an effect when the surface is not smooth and the energy of the wave is not effectively directed in the path of the wave. The energy is radiated in other directions. This can be seen in the following figure [8] – this causes energy loss.
Figure 13 : Wave scattering Multipath fading In non-perfect media there is attenuation which is an indication of loss at a specific distance. When one combines the above mentioned effects and look again at the 2 path model, the multipath fading concept makes sense.
School of Electrical, Electronic and Computer Engineering
18
Figure 14 : 2 Path model Referring to Figure 14 there are two rays. The one is the direct path which a transmitted wave will follow and the other is a reflected wave. Bringing into account attenuation, refraction speed loss and scattering energy loss, it can be said that the reflected wave will arrive with a delay and decreased energy at the receiver. [9] Multipath fading contains most of the effect discussed and it is the effect that is expected to be seen in underground radio frequency communication. 2.1.4 Radio frequency identification RFID uses radio waves for communication. Various protocols have been defined and standardised. There exist three types of RFID technologies,.namely passive, semi-passive and active RFID. RFID is a recently rediscovered technology (it came into being due to development in radio technology during the world wars). This technology existed in the 1980‟s b ut was too expensive to implement. Due to decreasing electrical component costs this technology has emerged as a viable technology. The main advantage that active RFID has over other identification technologies is its range. Barcodes work exceptionally well but lack range; they need to be close to the reader. RFID has endless potential due to its wireless operation. Passive RFID tags use the energy from a magnetic field and do not contain a battery. It then uses the same energy to return communication. Semi-passive RFID tags use a battery but not when returning communication, when communicating back with a reader the tag relies on the magnetic energy from the reader. Active RFID tags use a battery to power itself and for communications. Active RFID tags can use higher frequencies due to the use of a battery [10]. 2.1.5 SCADA This system is usually used where an industrial process requires data acquisition and control. In simple terms this is the supervisory system to control a process, thus having a central server and clients. This system is ideal to control the RFID tracking system. Care should be taken in designing the hardware and software architecture.
School of Electrical, Electronic and Computer Engineering
19
Functions of the SCADA system include; trending, logging, report generation and automation. The benefits of using this type of system design is that is will be dependable and data can be gathered at a single point [11].
2.2 TECHNOLOGY STUDY 2.2.1 RFID tags A RFID tag consists of an antenna and a processing unit with storage. The different types of tags will be discussed later in section 2.1.4. This device has different protocols for receiving / sending data to readers, but usually has a unique ID within the system. This device acts as an identifying medium for assets. Thus one can use this device for locating assets within a tracking system. The following sections present different commercially available tags. 2.2.1.1 Savi Technologies
Savi Technology Inc makes an active RFID tag named the ST-602 asset tag [12]. This tag can send data at 433 MHz and receive data at 123 kHz. It is low-power device and has a transmission range of 122 meters [12]. The tag has dimensions 6.2cm x 4.3cm x 1.2cm. The tag can be seen in Figure 15.
Figure 15 : Savi ST602 [12] Savi Technology Inc also makes an active RFID sensor tag. This tag has an on-board temperature and humidity sensor with adjustable alarms. Savi Technology Inc operates mainly in tracking containers. 2.2.1.2 Becker Electronics
Becker Electronics is part of Becker Mining who makes an active RFID tag named the TCAP 302. This is a beacon tag and has the unique feature that it fits inside a miner‟s cap lamp battery. This tag sends and receives data at 433 MHz. The tag can be seen in Figure 16 and Figure 17.
School of Electrical, Electronic and Computer Engineering
20
Figure 16 : TCAP 302
Figure 17 : TCAP 302 fitted on a standard mining cap lamp battery
2.2.1.3 Wavetrend Holdings
Wavetrend Holdings Limited makes an active RFID beacon tag named the TG100-A. This tag sends data at 433 MHz and transmits its ID at a predefined interval. This is a low-power device and has a long battery life. This tag features an anti-tampering function and has the following dimensions; 6.4cm x 3cm x 0.9cm [13]. The tag can be seen in Figure 18.
Figure 18 : Wavetrend TG100-A [13] 2.2.1.4 Ekahau
Ekahau makes an active RFID tag named the T301A [14]. This tag transmits and receives data at 2.4 GHz. This tag can easily be integrated into an existing Wi-Fi network that uses the 802.11a/b/n standards. This tag is a low-power device, has excellent security, supports Ethernet protocols and has motion detection [14]. The tag has dimensions 4.5cm x 5.5cm x 1.9cm. The tag can be seen in Figure 19.
School of Electrical, Electronic and Computer Engineering
21
Figure 19 : Ekahau T301A [14] The Becker Electronics tags are suitable for mining and were selected due to their form factor and their robustness. 2.2.2 RFID readers An RFID reader is a device that can interpret data received from RFID tags and can relay this data on a network interface. Most active RFID readers do not send a signal and listen for tag transmissions on a specific frequency from RFID tags. These readers normally support more than one RFID protocol or standard. 2.2.2.1 Savi Technologies
Savi Technology Inc makes an active RFID reader named SR-650 [15]. This reader supports any tag using the ISO 18000-7 standard, has an omni-directional range of 122 meters, receives data on 433 MHz and supports Ethernet/RS485 [15]. The reader has dimensions 30cm x 14cm. The reader can be seen in Figure 20.
Figure 20 : Savi SR-650 [15] 2.2.2.2 Becker Electronics
Becker Electronics make an RFID reader named UATR 400 [16]. This reader only supports the Backer family tags. This reads tags at 433MHz and can transmit the data to RS232/RS485/RS422. What makes this reader unique is that it complies with IP68 standard and it was designed for South African mines. The reader has dimensions 22cm x 12cm x 9.5cm. The reader can be seen in Figure 21.
School of Electrical, Electronic and Computer Engineering
22
Figure 21 : UATR 400 [16]
2.2.2.3
Wavetrend Holdings
Wavetrend Holdings Limited makes an RFID reader named RX201 [17]. This reader supports only the Wavetrend family tags. This reader receives data at 433 MHz, interfaces with RS232/RS485 and supports an external antenna [17]. The reader has dimensions 8.4cm x 4cm x 1.9 cm. The reader can be seen in Figure 22.
Figure 22 : Wavetrend RX201 [17] Wavetrend Holdings Limited also makes an RFID reader named RX900. This reader supports only the Wavetrend family tags. This reader receives data at 433 MHz, interfaces with Ethernet and supports an external antenna. The reader has dimensions 13cm x 14cm x 2.8cm. The reader can be seen in Figure 23.
School of Electrical, Electronic and Computer Engineering
23
Figure 23 : Wavetrend RX900 [17] Once more, the Becker Electronics readers were used due to their robustness and effectiveness in the mining environment. 2.2.3 RFID software RFID software is required to translate the data - collected by the readers - into a user readable format. Each company designing RFID tags and readers makes its own software according to its own needs and preferences. Some of the companies also make an API to let programmers interface with their software. Without the API there is no access to the data from the readers due to companies‟ privacy policies.
School of Electrical, Electronic and Computer Engineering
24
2.3 EXISTING SYSTEMS 2.3.1 Locatx (RTLS) LOCATX is a company who designed an RTLS using RFID technology at UHF and VHF frequencies [18]. The RTLS system operates at dual frequencies (13.56MHz and 433 MHz). The RFID tag incorporates a photo of the miner and the handheld reader is also used for security; the reader is shown in Figure 24. Furthermore the system has an alarm equal to a panic button and access violation notification. [18]
Figure 24 : Handheld reader [18] 2.3.2 Mine site technologies (Impact) Mine site technologies is an Australian based company specialising in mining communication. This company developed a personnel and asset tracking system called Impact. The infrastructure of Impact can be seen in Figure 25. This system has been fully developed, tested and implemented; according to the datasheet. [19]
Figure 25 : Impact infrastructure 2.3.3 Other RTLS’s Most of the RFID RTLS systems available have the same structure. This structure entails tags, readers and a central server with appropriate software. More advanced systems like Wavetrend‟s Total Enterprise Asset Visibility has a unique approach that puts it above the rest, due to their research focussing on RTLS‟s.
School of Electrical, Electronic and Computer Engineering
25
RTLS has a different connotation when used in logistics above ground than when used underground. Underground systems do not require excessive performance in terms of location accuracy since the people and assets are constrained to move along specific routes. Therefore, simplistic tagging is the preferred technology for underground – such as the Becker Electronics system. 2.4 SHORTFALL OF THE EXISTING SYSTEMS AND TECHNOLOGIES Currently, existing systems are expensive and thus not directly applicable to use in South African underground mines. As a result, the systems must be implemented by distributors, resulting in extra costs and time delays. Current technologies also use Wi-Fi frequencies for communication. At these high frequencies the effects of radio waves get accentuated and more attention must be given to correct this; thus becoming more expensive. The inter-networking protocols and standards are difficult to implement for Wi-Fi systems underground (due to physical and environmental challenges). There exist systems using a carrier frequency of 433 MHz, but these systems are not designed for underground use. Becker mining is the only company that designs RFID hardware for underground usage as far as could be ascertained. The tag, reader and single-board computer technologies available today are exactly what are needed to create an RTLS system for underground use. Some of these systems lack verification for use in underground conditions. Most of these systems cannot function when a central server is down; this however can be fixed with a data logger near the RFID readers. The simulation of all the components of an RTLS is possible, but care must be taken not to deviate from the required functionality. 2.5 DESIGN PHILOSOPHY The systems engineering life cycle is useful when designing a system that will be implemented. This life cycle contains preliminary design, detail design, production, utilisation and phase-out. For the purposes of this project the preliminary- and detail design will be done [1]. 2.5.1 Preliminary design In the preliminary design, subsystems are designed to adhere to specification and interfaces between these subsystems are defined. After completion of this design, a development specification is defined. Important steps in the preliminary design are the functional analysis, trade-off studies and development specification.
School of Electrical, Electronic and Computer Engineering
26
2.5.2 Detail design From the detail design emanate the final specifications of the design. These specifications may produce a change in the development specification. Important steps in the detail design are the final specifications and the development of a prototype.
This concludes chapter 2. The aim of this chapter was to give a theoretical background on components of the RFID tracking system. The chapter started with a discussion of the theoretical background of sub components, then moved to an investigation into existing systems and finally defined the design philosophy.
School of Electrical, Electronic and Computer Engineering
27
CHAPTER 3 – PRELIMINARY DESIGN
The aim of this chapter is to give physical definition to the theories explained so far. Here the functional architecture, process flow and the data model are presented.
School of Electrical, Electronic and Computer Engineering
28
3.1 FUNCTIONAL ARCHITECTURE The functionality of the system to be developed is shown below. This is a basic overview of the system and will be used to reference and describe the functional units of the system.
Figure 26 : Functional system overview 3.1.1 Input Each system requires input to start the processing. Even default parameters are inputs. The system in Figure 26 has mine layout and system parameters as inputs. 3.1.1.1 Mine layout
The mine layout consists of a grid. Each item in the grid must correspond to an object. The object will range from walls to readers. A grid is use to simplify the process of designing a layout. 3.1.1.2 System parameters
The system parameters consist of the asset movement characteristics, reader propagation model, RFID hardware specifications, and communication parameters. 3.1.2 Simulation Some hardware components of RFID RTLS will need to be simulated. These components are a RFID reader and the asset movement. 3.1.2.1 Asset movement
The movement of assets and people inside a mine needs to be simulated. The system input in 3.1.1 will be used to generate movement of assets and people. This movement must be accurate and representative of actual movement. The tag/reader interface will be simulated with a reader propagation model. 3.1.2.2 RFID Reader
The basic functioning of a RFID reader is a device that receives data from RFID tags and stores it in a buffer for collection from a data logger or server PC. This can also be simulated by using
School of Electrical, Electronic and Computer Engineering
29
software. The reader simulator will consist of threading, buffers and communication components. 3.1.3 Data collection Data collection has 2 phases. Each phase has its own functionality and operates independently of the other. 3.1.3.1 Data logger
The data logger hardware is a scaled down PC that runs a simple PHP script. The operation of the data logger is to create a temporary data buffer in case the server goes down and the RFID reader‟s buffer overflows. Thus the data logger must be able to collect data from the reader simulator and store it in a buffer. The server can then collect the data from this buffer. 3.1.3.2 Server Control software
The server needs to collect data from the readers/data loggers. The server needs software to collect the data and store it in the correct format in a central database. 3.1.4 Data interpretation The data that was collect by 0 needs to be interpreted. Server display software will need to get a copy of the server‟s data and run an algorithm through it to locate assets. 3.2 INTERFACES 3.2.1 IF1 The RFID tag has a wireless interface with the RFID tag. This interface is done using radio frequency communication. 3.2.2 IF2 The interface between the RFID reader and the logger can be RS485/RS422/RS232. The most common interface is RS485. RS485 can easily be used by the data logger computer, it functions the same as a serial port. 3.2.3 IF3 The interface between the data logger and the server control is a MySQL data connection. 3.2.4 IF4 The interface between the server control and the server display is a MySQL data connection. This concludes chapter 3. The aim of this chapter was to give a physical definition to the theories explained so far. Here the functional architecture, process flow and the data model have been presented.
School of Electrical, Electronic and Computer Engineering
30
CHAPTER 4 – DETAIL DESIGN
The chapter provides the design in more detail. The chapter starts with a system overview and then discusses the sub-system components in detail.
School of Electrical, Electronic and Computer Engineering
31
4.1 SYSTEM DESIGN
Figure 27 : Detail system overview 4.1.1 Input 4.1.1.1 Mine Drawing
The program MineSimulatorDrawing is used to create a mine layout. MineSimulatorDrawing uses the same drawing functions as MineSimulator and the user is able to edit the layout. The layout is then saved and can be used by MineSimulator.
Figure 28 : MineSimulatorDrawing screenshot In Figure 28 it can be seen that MineSimulatorDrawing was implemented using a drawing grid. In the grid blocks can be drawn to indicate an object type. The blocks can be categorised as wall, walking space, point of interest and a reader. An important feature of 32
School of Electrical, Electronic and Computer Engineering
MineSimulatorDrawing is the route planner. MineSimulator will simulate mobile assets and people walking around the mine by using the pre-calculated routes. These routes are saved in a file that the MineSimulator can read.
Figure 29 : MineSimulatorDrawing's route planner The route planner can be seen in Figure 29. A route has 4 components; number, level, point and coordinates: 1. (R) Route numbers start at 1 and count numerically onwards. The user must input the route number manually. Each possible route to the end of each tunnel must have a unique number; 2. (L) Route levels start at 0 and count numerically onwards. The level is increased when the increase button is pressed. Route levels are to identify the progress of a route; 3. (P) Route points are used to identify the POI‟s of a route at a certain level. Each level must have 5 points; 4. (X#Y#) Route coordinates is used to link a point to the layout grid. An X- and Y-point is defined.
In Figure 29 there is a list called “Current POI‟s”. These entries follow a structure which can be seen in Table 3. .
School of Electrical, Electronic and Computer Engineering
33
Table 3 : Route calculation list details
Entry item R L P X#Y# Example 1 0 0 X13Y19 Explanation Route number Route level Route point at that level X- and Y-point in layout grid
The combination of the layout and route design lets the user prepare a mine model for the simulator. 4.1.1.2 Asset movement parameters
Important parameters for asset movement calculation are their average walking speed. The average walking speed of a human being is between 4.5 km/h to 5.2 km/h. In a mine the environmental conditions worsen the ability to maintain that speed. So it is been decided that the speed to work with is 3.5 km/h to 4 km/h. The RFID tag used to identify the asset also has some adjustable parameters. The Becker mining tag has a transmission interval of 2 to 3 seconds. 4.1.1.3 Reader propagation model
For validation, testing was done in an active underground mine to determine the RFID reader‟s propagation model. The details of this research are discussed in chapter 5. This model is reduced to a piecewise linear model to simplify the programming and implementation. The implemented model can be seen in Figure 30.
100
90 Percentage Packet loss 80 70 60 50 40 30 20 10 0 0 5 10 15 20 25 Distance in meters
Figure 30 : Reader propagation model
School of Electrical, Electronic and Computer Engineering
34
4.1.2 Simulation 4.1.2.1 MineSimulator
MineSimulator is a program that simulates the movement of people and assets that host RFID tags. Movement near RFID readers results in a reader propagation model evaluation. When a tag is within reach of a reader, the RFID tag data is sent to the reader simulator. This is the main function of the MineSimulator. Figure 31 is a screenshot of the program in action. The white dots in the figure are the assets moving in the mine and the large circles are the reader detection areas based on the propagation models. The mine layout can also be seen in this figure. MineSimulator will be discussed in 4.2.
Figure 31 : MineSimulator version 2 4.1.2.2 Statistics
Read accuracy statistics are generated by MineSimulator. These statistics identify the blind spots, packet errors and collisions for each reader and each tag. The relevant terms are as follows: Blind spot – If an asset is within the reader‟s range and there is no propagation circle covering it, then the asset fails to be identified since the tag could not be read;
School of Electrical, Electronic and Computer Engineering
35
Packet error – If an asset is within the reader‟s range and is located within a propagation circle with, for example, 50% packet error rate, then the asset has a probability of 50% to been successfully identified. A packet error results in a transmission failure; Collision – If an asset is within the reader‟s range and transmits in the same interval as any other asset, then the asset fails to be identified; Unique reader seen – If an asset enters the reader‟s range for the first time; Unique reader read – If an asset has entered a reader‟s range and was correctly identified; Safety rating – . This calculates a safety rating of a specific
layout, which is the number of unique times a tag was read by a reader ( read) divided by the number of times that tag was inside a reader‟s detection area (seen). 4.1.2.3 Reader Simulator
The functioning of the reader was discussed in 0. The reader was simulated by using threading, TCP sockets and buffers. The functioning of the reader simulator will be discussed in chapter 4.3. The reader can simulate a given number of RFID readers each in their separate thread.
Figure 32 : Reader Simulator Screenshot 4.1.3 Data collection Data collection starts by creating a temporary buffer for data storage, hence the use of a data logger in the simulator. The data logger retrieves data from the RFID readers and stores them in a MySQL database. The data is the collected from the data logger by the server PC. The software that will be responsible for this is Server Control.
School of Electrical, Electronic and Computer Engineering
36
The database structures are unique for the server and the data logger, but there is a shared structure. These data structures will be discussed in chapter 4.1.3.2 and chapter 4.1.3.4. 4.1.3.1 Data logger
The data logger was not simulated, but created as a working application that can be used in an RFID RTLS. The data logger software was written in PHP using a combination of object orientated and procedural functions. The basic flow of operation can be seen in Figure 33. The data logger software is a PHP script that can be executed in a Linux environment. Most singleboard computers can run a basic Linux distribution. Linux is a more stable and resource efficient operating system. There must be a MySQL database available locally with the data structure of the data logger database which will be discussed in chapter 4.1.3.2.
Figure 33 : Data logger script The data logger firsts checks to see if the server is online. If the server is online, the script will check if the data logger ID has been updated recently. If the server is offline, the script ends. In future version it will be able to continue. This feature is incomplete due to time constrains, but all the functions needed are already in place. The check for the ID update is necessary and needs to identify the data being logged. It is saved to a file to be used in offline mode. The ID update process involves a connection to the server database. The data logger ID can be found by using the data logger‟s IP-address and hostname.
School of Electrical, Electronic and Computer Engineering
37
When the above checks are done, the script continues and retrieves a list of all the readers. The readers are identified by IP-address and port number. The script goes through each entry and creates a TCP socket connection to the reader. When the connection is made, the data is retrieved using the protocol discussed in 0 The data that is retrieved is then stored in the local MySQL database. <screenshot> There are minor features not shown on the diagram. The script also creates a lock file. This file is created when the script is started and deleted when the script exits. This is used to prevent the script from being executed multiple times, only one copy of the script should run simultaneously. The script also has the feature to do DNS lookups. This is useful when operation is a network that has dynamic IP-addresses. Using DNS lookups, it can be prevented that the data is retrieved from the wrong addresses. 4.1.3.2 Data logger database
The data logger database is used as a temporary buffer for the data. The database structure can be seen in Table 4. Table 4 : Data logger database structure Table Data Field d_id d_tagid d_readerid Explanation Unique data entry ID. Used as a data index. Unique tag ID. Used to identify assets. Unique reader ID Used to identify reader; determine location. Unique data logger ID. Used to identify data logger; relation to readers. Date and time the data was acquired. Unique ID to identify the reader. Reader hostname used in DNS lookups. Reader IP-address. Reader location in mine grid. Data logger responsible for the data buffer of the reader. 38
d_dataloggerid
d_datetime
Reader
r_id r_devname r_ip r_location r_dataloggerid
School of Electrical, Electronic and Computer Engineering
r_port
Reader communication port.
4.1.3.3
Server Control
Server Control software is responsible for data collection from the data loggers. This software is not simulated and can be used in a RFID RTLS together with the data logger software. The operational flow of the Server Control can be seen in Figure 34. <sreenshot> Version 1.0 was the first functioning application and version 1.1 added new data structures and corrected threading errors.
Figure 34 : Server Control software flow 4.1.3.4 Server database
Table 5 : Server database structure Table Data Field d_id d_tagid Explanation Unique data entry ID. Used as a data index. Unique tag ID. Used to identify assets. 39
School of Electrical, Electronic and Computer Engineering
d_readerid
d_dataloggerid
d_datetime_seen
d_datetime_store
Unique reader ID Used to identify reader; determine location. Unique data logger ID. Used to identify data logger; relation to readers. Date and time the data was acquired by the data logger or reader. Date and time the data was acquired by the server control. The delay between the date seen and date stored determines data accuracy. Unique ID to identify the reader. Reader hostname used in DNS lookups. Reader IP-address. Reader location in mine grid. Data logger responsible for the data buffer of the reader. Reader communication port. Indicates the route level in a hierarchy format. Unique data logger ID to identify the data logger. Data logger hostname used in DNS lookups. Data logger IP-address. Real location of the data logger to indicate the data logger‟s position relative to the mine‟s levels.
Reader
r_id r_devname r_ip r_location r_dataloggerid r_port r_route_level
Datalogger
d_id d_pcname d_ipaddr d_location
4.1.4 Data interpretation Data interpretation is done by the Server Display. Server Display uses buffers and data from the server‟s MySQL database to interpret the data. T his will be discussed in 4.4. The data is interpreted so that the user can understand the data generated by the system. MySQL‟s .Net connector is used in this module [20].
School of Electrical, Electronic and Computer Engineering
40
Figure 35 : ServerDisplay version 3 4.2 MINE SIMULATOR 4.2.1 Versions 4.2.1.1 Version 1.0
This was the first functioning application. An asset could move around the mine with the first version of the route algorithm and was using the first version of the reader‟s propagation model . The route algorithm was based on a pre-recorded route being saved in a file. The route was loaded into the memory and started at the first entry and moved through the entries until it reached the end. When the end was reach the asset has moved through the mine. The reader‟s propagation model was based on an on/off method. If the asset was within the reader‟s range when the asset‟s tag transmitted its data, the asset was identified. 4.2.1.2 Version 1.1
This version introduced a few new features namely statistical collection, simulation pausing and version 2 of the reader propagation model. The propagation model in this version was based on linear blind spots. There were circles that represented the detection areas. When the asset was in the range of the reader, the asset‟s tag transmitted its data and if the asset was not in the
School of Electrical, Electronic and Computer Engineering
41
detection circle; the asset was in the blind spot. In reality the blind spot was not linear, but for practical purposes it was sufficient to simplify the model to reduce programming complexity. <reader propagation screenshot> The statistical collection is to review simulation results and will be discussed in 0. The thread required to save the data was a challenge, due to object synchronization. The movement calculation could be using an object when the logging thread tried to use it. This will be discussed in 0. When the user paused the simulation, the people and assets stopped moving in the mine. This would allow the user to review their positions. 4.2.1.3 Version 1.2
This version was the first version to function in the RFID RTLS. This version introduced socket communication, adjustable simulation parameters, tag collision detection, version 2 of the route algorithm and a more efficient threading usage. The socket communication was introduced to notify the reader simulator of successful asset identification. When an asset was successfully identified by evaluating the reader propagation model, the reader simulator was notified by using the protocol in 0. The adjustable parameters were an important part of creating an adjustable simulation. These parameters are discussed in 0. Tag collision is a real problem RF-devices faces. Modelling this into the software was an important part of successfully creating a simulator. Version 2 of the route algorithm was a new design and replaced version 1. Version 2 will be discussed in 4.2.2.1. Efficient threading usage will be discussed in 0. 4.2.1.4 Version 2
This is the final version and introduced version 3 of the reader propagation model. The new version of the propagation model was obtained from a validation in an active mine and will be discussed in detail in chapter 5. The new model also used packet loss as an indication of the reader circle‟s probability to detect an asset. The packet loss count was added to the statistical collection. Version 3 of the reader propagation model will be discussed in 4.2.2.3.
School of Electrical, Electronic and Computer Engineering
42
4.2.2 Features 4.2.2.1 Route algorithm First, pre-
Moving assets around in a mine requires an algorithm to calculate the route.
calculated routes were used, which resulted in extra time spent on calculating the routes and programming difficulties in determining the unique reader reads/seen statistic. The new algorithm involved recording POI‟s to determine the level of a specific route. Specific terminology involved in the route calculation is discussed in 4.1.1.1. Before a person or asset enters a mine, a route number is pseudo-randomly chosen. Each route number has a route file linked to it. The route file contains the route levels linked to a mine grid point. Thus after a route has been chosen, the movement calculation can identify a mine grid point to navigate an asset to. When the movement calculation has navigated the asset to a next level POI, the following POI is chosen using the current route information. If the end of route has been reached, the route is reversed; forcing the person or asset to go back out of the mine along the same track. 4.2.2.2 Asset movement calculation
In the previous paragraph the movement calculation‟s navigation input was discussed, namely the routing algorithm. The movement calculation routine knows the current position of the asset, the next POI position and the movement speed. The movement speed is defined in terms of a minimum and a maximum speed. The following will description explains the movement vectors as can be seen in Figure 36. The current position of the asset is (X1,Y1). The destination, the next POI, a person or asset wants to move to is (X2,Y2). P1 is the angle between the asset and its next POI, thus .
P1 has a pseudo-random angle adjustment of ±15º, resulting in the new angle P2, thus with . K is the magnitude of the assets movement. K is
pseudo-randomly generated using the minimum and maximum movement speed. Finally the new position (X3,Y3) is calculated. Thus . and
School of Electrical, Electronic and Computer Engineering
43
Figure 36 : Movement calculation 4.2.2.3 Reader propagation model
The final propagation model is based on linear packet loss. The model can be seen in Figure 37. The different colours indicate a packet error rate. The colour was calculated using red, green and blue (RGB). The formulas is: and ,
. This will result in a red colour for a packet loss of 90% and a green colour for a
packet loss of 0%. The specifics of the packet error will be discussed in chapter 5.
Figure 37 : Final reader propagation model The distance from the reader is calculated by getting the distance from each reader on the mine layout and determining if the distance is within the largest reader circle. The circles in the model indicate an identifiable area inside which a tag can be read. When the asset falls inside these circles, the position of the asset will be evaluated to determine if the asset is identified. Evaluation is done by detecting the packet error rate, checking the blind
School of Electrical, Electronic and Computer Engineering
44
spots and collision detection. The packet error rate is loaded from file. The blind spots are calculated by gaps in the identifiable circles. The collision detection is a longer process. Tag collision detection involved capturing al the tag pings at a calculation interval. When all the assets‟ movements were calculated and their tag pings recorded, it was verified for tag collisions. A tag collision will occur when two or more tags tried to communicate with s reader inside the same time interval. Thus, when two or more tags transmit inside the same time interval, they will have a collision and none of them will be successfully identified. The tag collision count is a key indicator of tag read failure. 4.2.2.4 Statistical collection
The statistical collection is used to review the simulation results. Each asset and reader hold a summary of the identification statistics. After a simulation, the user can see which reader has been the least efficient and can review each asset‟s movement and efficiency data. Data collection required a thread due to the file I/O latencies caused by multiple file writes per asset movement calculation. At first, the application was very slow and the latencies made the simulation inaccurate due to assets skipping important areas in the reader‟s propagation model. Efficiency data included: the amount of transmissions, unique readers met, unique readers seen, tag collision, blind spots and packet errors. 4.2.2.5 Data logging
Data logging in any software application is a handy debugging tool as it can also be used to find bugs in software. In most of the software developed for this project, the logging was done by using buffers and a logging thread. The thread operates continuously throughout program execution. The thread writes the buffer to a file and then clears the buffer. After clearing the buffer, the thread goes in a sleep state for 100ms and repeats the operation. 4.2.2.6 Adjustable simulation parameters
The adjustable parameters allow the user to change simulation parameters. The parameters that can be changed are movement speed, grid dimensions, tag transmission interval, save file location and communication details. Changing these settings will cause the simulation to restart with the new settings. Changing simulation parameters creates an adjustable simulation process. The parameters were saved using Microsoft.Net. 4.2.2.7 Efficient memory and processing power usage [21] [22]
Efficient processing power and memory usage become a key factor when a significant amount of calculation has to be done. Reallocation of memory and poor usage of processing power
School of Electrical, Electronic and Computer Engineering
45
slows the simulator down. Memory reallocation was reduced by creating more global objects and reusing them instead of creating new objects all the time. To use the processing power more efficiently, a lot of time was spent to analyse and optimize program flow. Another bottleneck was threading synchronisation. When threads use the same object and have different execution speeds, they will try to access an object simultaneously. This was corrected by using a synchronization lock. The object at hand was securely put on hold for exclusive use by the thread and released upon completion. This created an insignificant latency. 4.2.2.8 Reader simulator communication handler
The reader simulator‟s communications protocol is discussed in 0. The communication is based on TCP sockets. The communication handler is a thread that operates independently of the other tasks. The thread starts and checks if there is any data in the communications buffer. If there is not data, the thread goes in a sleep state for 100ms. If there is data, the handler creates a socket communication channel to the reader simulator and sends the data to the reader. After the data is sent the communication channel is closed and the handler continues its operation as normal.
School of Electrical, Electronic and Computer Engineering
46
4.2.3 Functional flow
Figure 38 : Simplification of the functional flow of MineSimulator 4.2.3.1 Setup
MineSimulator starts by creating a MineDrawing object. This is the class created to handle the simulation. A class was used to simplify the process of creating an API from this simulator. When the MineDrawing object is created, it initiates a process to start the simulation setup. Firstly, it loads the settings from the saved files and creates the necessary objects in the memory. Thereafter it creates an image of the mine layout without people or assets. This image will be used throughout the simulation process as the people‟s or assets‟ movement and reader‟s changes will be drawn on this canvas.
School of Electrical, Electronic and Computer Engineering
47
4.2.3.2
Threading
After setup, three threads are started, namely logging, moving, and communication threads. These threads are processed continuously until the MineDrawing object is destroyed. The logging thread was discussed in 0. The moving thread cycles through the currently active people or assets and evaluates the reader propagation model as discussed in 4.2.2.3. When the model has been parsed and the successful identifications (tag reads) have been transmitted, a new image is created by simply drawing the result onto the mine image created in setup. If the person or asset has been successfully identified, the appropriate reader circle is coloured in yellow to visually illustrate a successful read. This can be seen in Figure 39.
Figure 39 : Visual identification indication The communication thread is discussed in 0. 4.2.3.3 Exiting
When the application exits the statistics are written to file and the threads are stopped. When stopping the treads the application must wait for a thread to close. Threads will close when a termination flag is sent to them. Data can be lost by forcing a thread to close, but this is the last resort when the application closes. 4.3 READER SIMULATOR 4.3.1 Versions 4.3.1.1 Version 1
This first version of the application was created to be a standalone application in the sense that one ReaderSimulator will run on PC. This application was also created when there was a total different system design, before the design of MineSimulator. The first system design had a reader manager, which was used to manually move assets from reader to reader. When the data logger connected to the ReaderSimulator, there was a TCP socket connection made from
School of Electrical, Electronic and Computer Engineering
48
the ReaderSimulator to the reader manager. The reader manager would then report the assets at that specific reader. The connection to the reader manager was terminated and the ReaderSimulator reported to the data logger of the assets near itself. So the reader manager was used as a tool to test the data logger and server control software. 4.3.1.2 Version 2
The next version of ReaderSimulator was created to eliminate the need for more hardware. Thus threading was used in the application to function as multiple readers on one PC. This proved to be very useful when demonstrating a mine layout with limited hardware. This version did not used the threading model very well and resulted in an application not in control of the threads that are running. So the application started and worked without any user control. When the application was exited the threads was forced to close. 4.3.1.3 Version 2.1
This version was the first version to eliminate the reader manager. The new version made use of an internal buffer to store data received by using the ReaderSimulator communication protocol. This protocol will be discussed in 0. This version also made use of a class to manage threading. This will be discussed in 4.3.1.5. 4.3.1.4 Version 2.2
Version 2.1 had one major disadvantage. Every transaction of all the readers was displayed in one GUI component; the implication of this is that due to threading synchronization it slowed down the application. The queue of synchronization request increased with time and after a few minutes the application slowed down the whole system. This caused a huge latency in real time and data interpretation - the latency was about a 5 minutes. This was reduced by only allowing the reader threads to report when data was received and not reporting of every transaction. This reduced the system latency down to 30 seconds. 4.3.1.5 Threading
A class is used to manage the threading and reporting of the readers. An array of this class is created to initialize more than one reader and controlling it at the same time. This proved a success. The class could be used multiple times and still be controlled from the parent application. The class uses a buffer to store data (discussed in 4.3.1.6) and has a communication protocol that will be discussed in 0. The advantage of using the class in an array was that when the application exited, the threads could be stopped safely.
School of Electrical, Electronic and Computer Engineering
49
4.3.1.6
Buffer
The buffer was created to store data received from MineSimulator. The buffer was created using a simple string array structure. This buffer is easily expandable. The buffer does not use threading synchronization due to the fact that the thread is the only one able to use the buffer. The buffer is never erased, so it can cause a memory leak when used for an extended time period. Instead of clearing the buffer, each data entry has an active flag. When the flag is set, the data is still perceived as tracking data at that reader. If the flag is not set, the data is ignored. The memory leak was not fixed due to time constrains. 4.3.1.7 Information display
Information display, in an application using multiple threads, is a problem. The balance between information data and application speed must be struck. The final version displays debugging data which shows that an application has started and that the reader is online. The only information displayed from the reader thread is when the reader receives data. Previously incoming connection details, protocol details and data transactions were displayed. Everything display in the GUI is also written to file using the same architecture as the MineSimulator‟s data logging thread. 4.3.1.8 ReaderSimulator communication protocol
This protocol was created to simplify and standardize the data interactions of the software modules in the system. The protocol is visualized in Figure 40and Figure 41. The fist command received when handling incoming connections determines the flow of data. If the SET command is received the data will flow into the ReaderSimulator‟s buffer. When the GET command is received the data will flow outwards to the application that made the connection. The SET command will result in the ReaderSimulator‟s communication thread waiting for incoming data and storing it in the buffer. The GET command will evaluate the ReaderSimulator‟s communication thread‟s buffer for active data entries. These entries will be combined and sent back to the application that made the connection. The reader simulator receives data inputs/requests on a port saved in a database. The ports start at 6000. The functional flow of the reader simulator is simplified in Figure 42. The ports can be calculated by reading the mine layout save file and incrementing the port number by one. The port numbers is saved in a database for use by the data logger and the server control/display.
School of Electrical, Electronic and Computer Engineering
50
Figure 40 : SET tag data
Figure 41 : GET tag data
School of Electrical, Electronic and Computer Engineering
51
4.3.2 Functional flow
Figure 42 : Reader Simulator flow
4.3.2.1
Setup
ReaderSimulator starts by reading configuration data from a file. This configuration contains the number of readers and determines the type of information to log. 4.3.2.2 Threading
After setup, all the threads can be started. These threads were discussed in 4.3.1.5. The operation of the thread is simplified in Figure 42. 4.3.2.3 Exiting
When the application is closed, each reader thread is safely closed as well.
School of Electrical, Electronic and Computer Engineering
52
4.4 SERVER DISPLAY 4.4.1 Versions 4.4.1.1 Version 1
The first reader manager was used as discussed in 4.3.1.1. The basic working of this model was to translate data received from the data logger. This is the main function of the ServerDisplay, but in this version the display was not visually attractive. This can be seen in Figure 43.
Figure 43 : ServerDisplay version 1 The application would gather data from the MySQL database and get the unique data entries. These entries would then be written into pre-defined spaces. This restricted the application‟s scalability. In this version of ServerDisplay version 1 of the data gathering SQL statement was used. This version seems to work in small amount of data, but was later discovered not to be reliable. 4.4.1.2 Version 2
This version eliminated the reader manager. The data display was done by executing version 2 of the gathering SQL statement of the server. The data returned was the unique location of every asset in terms of the data logger and reader placement. This was a more logical display than visually informative display. 4.4.1.3 Version 3
The final ServerDisplay provides a visual simulation of the mine to display the location of assets. The solution is to display the same mine layout as the MineSimulator. Now the locations
School of Electrical, Electronic and Computer Engineering
53
of the assets are known in terms of readers since the locations of the readers are known. Thus, one can simply display the current assets at a reader in order to determine their location. A display and locating model was developed and is discussed in 4.4.2.1. 4.4.2 Features 4.4.2.1 Display method
The display method was based on the same method used for drawing the MineSimulator‟s display, but without people or assets moving around the mine. This model replaced the reader‟s propagation circles with the amount of assets located near the reader or moving around the reader. The display received the data from the locating algorithm. The display can be seen in Figure 44.
Figure 44 : ServerDisplay In the Figure 44 it can be seen that two types of reader information are displayed. The first is the forward/backward display. This is readers that have other readers after them in a logical sense of flow. Using sector based location the movement can only be indicated when an asset moves from one reader to another. The second information type is the people or asset count. This is when there is no reader after the last reader in a route (at the end of a route). When this happens, one cannot determine if an asset moving past a reader is moving forward of backward in the mine. 4.4.2.2 Locating algorithm
The locating algorithm determines where an asset is currently placed and the asset‟s direction. Thus the locating algorithm receives the last known location of each asset continuously and calculates the movement dynamics.
School of Electrical, Electronic and Computer Engineering
54
This algorithm uses a buffer for every reader. If data is received, it is defined accordingly as forward / backward / count. The algorithm can be seen in Figure 45. The algorithm works by adding the assets to the reader‟s buffer and removing them when they move to a different reader. This creates a history of the asset‟s movement through the mine. This history is used to determine the movement direction.
Figure 45 : Locating algorithm
School of Electrical, Electronic and Computer Engineering
55
4.4.3 Functional flow The following section provides the functional flow of the simulator application.
Figure 46 : ServerDisplay version 3 4.4.3.1 Setup
The application starts and loads the required settings for creating an image. Thereafter, the background image is created that will be used to draw the movement information through the simulation process. Part of the setup is to determine all readers‟ configuration information that handles the display method and location algorithms. 4.4.3.2 Data flow
The data flow in the functional flow is indicated in Figure 46. This is to show how the object handling the display can retrieve data and continuously retrieve updated data. Data is sent to configure the readers at first. Thereafter the data is updated by a timer on the parent application. 4.4.3.3 Updating timer
The timer on the parent application is use to query the database with the SQL statement in Error! Reference source not found.. After the data has been received, it is sent to the object andling the display drawing. The timer executes every 100ms (chosen from empirical testing).
School of Electrical, Electronic and Computer Engineering
56
4.4.3.4
Image creation
The image creation works the same as the image creation of MineSimulator. The extra information that is added to the image was discussed in 4.4.2.1. This concludes chapter 4. The aim of this chapter was to give a system overview and discuss the sub-system components in detail.
School of Electrical, Electronic and Computer Engineering
57
CHAPTER 5 – VALIDATION
This chapter provides validation data for the design used in the reader propagation model. An overview of the validation process will be given and the validation will be briefly discussed. The model was used to determine the propagation characteristics of RF inside the haulage of an underground mine.
School of Electrical, Electronic and Computer Engineering
58
5.1 OVERVIEW The reader propagation model used in MineSimulator was validated in an active underground mine. The details of the mine cannot be discussed due to mutual non-disclosure agreement. The validation was done 3km underground. Test was done to determine a propagation model for an active RFID reader. Test generated data was compiled into one piecewise linear model - this model was used in MineSimulator. 5.2 ENVIRONMENT The humidity underground was high and there was water present near the testing area. There were also pipes transporting water and air. The layout of the test is shown in Figure 47. The width of the haulage was approximate 4 meters with a height of 5 meters. In section 4 there was an electric motor located at about 17 meters. There were no other wireless devices present near the testing area.
Section1
Motor
Section4 Section3 Reader
Section2
Figure 47 : Test layout
School of Electrical, Electronic and Computer Engineering
59
Figure 48 : Section 1
Figure 49 : Section 1 (Left) and section 4 (Right)
School of Electrical, Electronic and Computer Engineering
60
Figure 50 : Section 3
Figure 51 : Section 2 (Right) and section 4 (Left)
Figure 52 : RFID Tag placement
School of Electrical, Electronic and Computer Engineering
61
5.3 HARDWARE USED The hardware used for testing was sponsored by Becker Electronics. A portable active RFID reader and active RFID tags were used. The reader that was used is a HTRD100ZA as can be seen in Figure 53. This reader was designed to read any Backer family RFID tag. This reader displays the RSSI reading it has calculated when a transmission from a RFID tag had been successful.
Figure 53 : Backer Electronics HTRD100ZA The RFID tags used are TPER200WW as can be seen in Figure 54. This tag pseudo-randomly transmits its identification number in a window of 2 to 3 seconds.
Figure 54 : Becker Electronics TPER200WW 5.4 TESTS In Figure 47 the tests were carried out by measuring RSSI values in sections 1, 2, 3 and 4. In the middle of the haulage the measurements were spaced at 0.5 meters and all remaining measurements were made in 1 meter steps. The RSSI values were measured and the packet success rate was indicated by listening how fast the packets were received. A value of 5 was chosen for an excellent packet success rate,
School of Electrical, Electronic and Computer Engineering
62
while a value of 1 was chosen for a weak packet success rate – that is, a high packet error rate resulted in a value of 1 and a low packet error rate in a value of 5. The tag placement and environment details can be seen in 5.2. The tests took about 3 hours and the data was recorded manually. Three tests were done in each section. Reception along the centre and sides of the haulage was characterized. The measurement results are shown in the following section. 5.5 RESULTS 5.5.1 Section 1
60.00 50.00 40.00 RSSI 6 5 4
30.00
20.00 10.00 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
3
2 1 0
Distance (Meters)
Center Center Packet Success Left Left Packet Success Right Right Packet Success
School of Electrical, Electronic and Computer Engineering
63
5.5.2 Section 2
60.00 50.00 40.00 RSSI 30.00 20.00 10.00 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Distance (Meters) Center Center Packet Success Left Left Packet Success Right Right Packet Success 6 5 4 3 2 1 0
5.5.3 Section 3
60.00 6
50.00
5
40.00
RSSI
4
30.00
3
20.00
2
10.00
1
0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 Distance (Meters) Center Center Packet Success Left Left Packet Success Right Right Packet Success
0
School of Electrical, Electronic and Computer Engineering
64
5.5.4 Section 4
60.00 6
50.00
5
40.00 RSSI
4
30.00
3
20.00
2
10.00
1
0.00 0 2 4 6 Distance (Meters) Center Center Packet Success Left Left Packet Success Right Right Packet Success 8 10 12
0
5.5.5 Global average
50.00 45.00 40.00 5.00 6.00
35.00
30.00 RSSI 25.00 20.00 15.00 10.00 5.00 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Distance (Meters) RSSI Packet success Packet success used in MineSimulator
4.00
3.00
2.00
1.00
0.00
School of Electrical, Electronic and Computer Engineering
65
5.6 DISCUSSION The characterization tests provided a successful model that could be used in MineSimulator. The RF effects discussed in 0 produced the desired effect that summarized propagation effects such as multipath fading. This can be witnessed in 5.5.2 at the edge of reception where significant fading is evident - refer to the variation in signal strength at the fringe. This was due to the multipath fading from reflections against the haulage walls. To improve reception in the haulage, one may do the following: Spatial diversity – use more than one antenna; Polarisation – use cross-polarised or circularly polarised antennas; Frequency – use more than one frequency.
This concludes chapter 5. The aim of this chapter was to validate the reader propagation model from measurements taken in an underground mine. The model appears to be piecewise linear and was used accordingly in all simulations.
School of Electrical, Electronic and Computer Engineering
66
CHAPTER 6 – IMPLEMENTATION AND EVALUATION
This chapter presents the implementation layout (architecture) of the physical design from chapter 4 and addresses the evaluation of the test layout.
School of Electrical, Electronic and Computer Engineering
67
6.1 IMPLEMENTATION
6.1.1 PC1 PC1 was used to run two simulation programs from an IP-address of 192.168.0.3. There was a two-screen setup with each simulation program running on each screen. This resulted in an aesthetically acceptable appearance. 6.1.2 PC2 PC2 was used to run three layers of computing. The first layer was the PC‟s main operating system. Linux was used in this demonstration. The Linux operating system was used to create the environment for the data logger. Thus there was a MySQL [23] service running with the data logger database setup and a PHP [24] execution environment. The second and third layers were created using virtual computing environments. This was done by using VirtualBox [25]. The one virtual PC ran the ServerDisplay application and the other ran the ServerControl application. 6.1.3 PC3 PC3 was just to display the results of the PHP script that was executed on PC2‟s first layer. Putty [26] was used to accomplish this.
School of Electrical, Electronic and Computer Engineering
68
6.2 EVALUATION 6.2.1 MineSimulator tests The statistical collection of MineSimulator was to test and ensure functionality of the simulator. MineSimulator can function without any other application so the testing there had been a straight forward run-in test. The simulator was run for one day without any interruption. The tests revealed that MineSimulator was stable terms of its functional capability. 6.2.2 ReaderSimulator tests ReaderSimulator was a more testable application in the sense that most of the functionality was done in the background. An application was created to perform a fuzzy fire test on ReaderSimulator. This was done by sending 100 data items at once to the ReaderSimulator. This tested the stability of the ReaderSimulator when handling a large volume of data. 6.2.3 Data collection tester ServerDisplay, ServerControl and the data logger script could be tested as whole. This was done by using the test application created for the ReaderSimulator and recording the data flow performance through the data collection process. 6.3 SYSTEM TEST The system could only be tested as whole when the setup in 6.1 was done. The test required all modeles to run and an active simulation in MineSimulator. The verification of the system to be functioning correctly was done by visually inspecting ServerDisplay and comparing it to MineSimulator‟s assets movement. 6.3.1 Overview MineSimulator was run to analyse the movement of 52 assets. The locations of each asset were successfully traced and the statistics is show in 6.3.2 and 6.3.3. 6.3.2 Asset statistics
Person 0 1 2 3 4 5 6 7 8 9 Tag Tag Tag Tag Read Miss Collision Blindspot 79 36 3 0 72 37 6 0 77 39 4 0 78 44 6 0 81 51 5 0 113 60 2 0 77 43 4 0 121 64 3 0 79 43 5 0 80 47 7 0 Tag Readers Readers PacketError Met Read 33 4 4 31 4 4 35 4 4 38 4 4 46 4 4 58 6 6 39 4 4 61 6 6 38 4 4 40 4 4
School of Electrical, Electronic and Computer Engineering
69
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
114 74 74 107 119 76 108 79 77 123 106 121 73 118 76 110 120 71 75 107 81 119 110 78 72 76 111 122 79 109 118 72 77 79 73 78 107 80 116 107 77 110 70
60 36 42 58 54 45 51 42 42 68 51 64 42 70 41 57 67 33 47 64 47 73 61 42 43 49 63 64 46 52 68 32 44 41 39 47 55 48 65 65 46 59 37
5 6 5 3 4 5 6 5 7 8 4 7 5 3 8 8 5 7 5 2 9 6 5 3 5 12 3 4 5 6 11 6 12 6 3 8 4 7 5 12 9 6 3
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
55 30 37 55 50 40 45 37 35 60 47 57 37 67 33 49 62 26 42 62 38 67 56 39 38 37 60 60 41 46 57 26 32 35 36 39 51 41 60 53 37 53 34
6 4 4 6 6 4 6 4 4 6 6 6 4 6 4 6 6 4 4 6 4 6 6 4 4 4 6 6 4 6 6 4 4 4 4 4 6 4 6 6 4 6 4
6 4 4 6 6 4 6 4 4 6 6 6 4 6 4 6 6 4 4 6 4 6 6 4 4 4 6 6 4 6 6 4 4 4 4 4 6 4 6 6 4 6 4
School of Electrical, Electronic and Computer Engineering
70
6.3.3 Reader Statistics
Reader Tag Tag Tag Blind Tag Packet Tag Unique Unique transmissions Collisions Spots Errors Reads transmissions Reads in radius in radius (11:21) 2067 181 0 997 889 106 106 (51:21) 886 33 0 459 394 46 46 (83:40) 421 6 0 218 197 20 20 (46:4) 662 51 0 336 275 34 34 (21:41) 453 18 0 199 236 26 26 (57:38) 417 14 0 172 231 26 26
6.3.4 Information logging Only a part of the log is shown.
Date 08:36:55:744 08:36:55:754 08:36:55:944 08:36:56:174 08:36:56:184 08:36:56:214 08:37:08:392 08:37:08:592 08:37:08:792 08:37:08:993 08:37:09:193 08:37:09:393 08:37:09:594 08:37:09:794 08:37:09:994 08:37:10:194 08:37:10:405 08:37:10:645 08:37:10:845 08:37:11:046 08:37:11:296 08:37:11:496 08:37:11:717 08:37:11:937 08:37:18:116 Type Info Info Info Info Info Info Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel First Time Reader 08:37:18:116 RandomCheck2 08:37:18:116 RandomCheck1 08:37:18:116 PacketError Info MineDrawing Object created Starting Initialization Thread Loading done Creating GridImage done Starting Caculation Thread Stoping Initialization Thread Person 0 added. Route=2 Person 1 added. Route=1 Person 2 added. Route=2 Person 3 added. Route=1 Person 4 added. Route=2 Person 5 added. Route=3 Person 6 added. Route=1 Person 7 added. Route=4 Person 8 added. Route=1 Person 9 added. Route=2 Person 10 added. Route=3 Person 11 added. Route=1 Person 12 added. Route=2 Person 13 added. Route=3 Person 14 added. Route=4 Person 15 added. Route=1 Person 16 added. Route=3 Person 17 added. Route=2 Person 0 is the first time at reader(11:21) with distance 24.17183852705 m Rate=80 Packet Error With Person 0 @Reader(11:21) with distance 72.5155155811499
School of Electrical, Electronic and Computer Engineering
71
08:37:18:386 First Time Reader 08:37:18:386 RandomCheck2 08:37:18:386 RandomCheck1 08:37:18:386 PacketError 08:37:18:927 First Time Reader 08:37:18:927 RandomCheck2
Person 1 is the first time at reader(11:21) with distance 24.3732549232874 m Rate=80 Packet Error With Person 1 @Reader(11:21) with distance 73.1197647698623 Person 3 is the first time at reader(11:21) with distance 24.2131736411768 m Rate=80
6.3.5 Test results The test shows a safety rating of 100% and this is confirmed by the reader and asset statistics. The reader at with location (11:21) in 6.3.3 is the reader at the entrance and needs to identify the larges amount of data. This concludes chapter 6. The aim of this chapter was to discuss the implementation and evaluation of the design in chapter 4.
School of Electrical, Electronic and Computer Engineering
72
CHAPTER 7 - CONCLUSION AND RECOMMENDATIONS
In this chapter we provide an overview of the complete project and reflect on the success of the design. We make recommendations for use in similar projects (and future iterations of this project) and discuss the five ECSA Exit Level Outcomes that were addressed in the completion of this project.
School of Electrical, Electronic and Computer Engineering
73
7.1 PROJECT HISTORY - REVISITED A feasibility study of an RFID RTLS was required. After researching current technologies and systems, the need for a simulator design became evident. After designing the simulator, it was integrated into a simulation that could be verified by doing software functional testing with validation of the RF propagation model. All of the software modules were created using VisualBasic.Net [27] [28] – except one module that was created with a PHP script. These modules were all planned, designed and tested by the author. 7.2 CONCLUSION The systems engineering approach to software simulation is ideal for software system development. An active radio frequency identification real-time locating system was successfully planned, designed and verified in a simulated environment. Part of the simulation was also validated by a propagation model testing in an active underground mine. It is possible to use an application to determine the optimal design for an RFID RTLS. Using this application not can it only save costs but also maximize safety. Designs will not always be a success from the start; this is why the design is an iterative process. Analysis is done, the design is conceptualised and goes through a detail design process. Thereafter can the design be tested and, if needed, be adjusted. Software design is a speciality upon which needs to be focussed. Simply rushing through the correct design process will not be successful. The design process needs to be understood first and must be followed correctly. The engineering curriculum creates the knowledge to successfully complete this project. Software design is not limited to software itself and involved mathematics and the logical thought process learned in engineering. The simulator can be used to optimize and reduce cost when designing a RFID RTLS. This is very useful in a sense that a RTLS can be planned to achieve a certain resolution of location data; thus the balance between costs and detail information can be found while still insuring the safety of such a system. The statistical collection feature of MineSimulator proved to generate valuable data in planning a RTLS. Critical readers that have a high amount of traffic must be account for by insuring that the antenna quality of those readers is reliable. The validation in an active mine was necessary to create a simulator that is close to reality and to identify the problems of using an RFID RTLS.
School of Electrical, Electronic and Computer Engineering
74
7.3 RECOMMENDATIONS The systems engineering process needs to be understood before it is used – is it not as simplistic as it initially appears. Future developers will have to understand the process before attempting to adapt the simulator in this design project. The simulation revealed that the most critical reader is the reader at the entrance. Therefore, this readers needs to be reliable. General recommendations regarding to the software will be to used object orientation from the start, plan an application instead of improvising along the way, use backups and have a professional programming style. This was all done in this design project by experiencing the limitations of ad hoc development. Future recommendation for MineSimulator is to add more moveable objects that have an influence on the reader propagation model, such as locomotives and trains. More research can thus be done on the reader propagation model and scenarios - such as trapped miners - could be integrated into the simulation. 7.4 ECSA OUTCOMES 7.4.1 Problem solving In engineering, problem solving is done on a daily basis. This project started with a problem statement and was then analysed and solved. 7.4.2 Engineering Design Engineering design is a very unique philosophy that is tried and tested. This project implemented engineering design into a software application. 7.4.3 Investigations, experiments and data analysis Science requires investigation, experiment and analysis to prove a theory - engineering is applied science. This project had a problem statement which was investigated, analysed, implemented and validated. 7.4.4 Professional and technical communication Engineering communication is an unavoidable function and must be done to insure that the design meets al the client‟s needs. This project involved negotiating with hardware manufactures and mine management and writing this report.
School of Electrical, Electronic and Computer Engineering
75
7.4.5 Independent learning ability Engineering is obsolete when no one learns from the process. Much has been learnt from this project (particularly about engineering method and programming) and will most certainly be used in future designs.
School of Electrical, Electronic and Computer Engineering
76
References
[1] B.S. Blanchard and W.J. Fabrycky, Systems engineering and analysis, 4th ed.: Peason Education, 2006. [2] H Steyn et al., Project Management - A Multi-Disciplinary Approach, 2nd ed. Pretoria, South Africa: FPM Publishing, 2009. [3] Engineering Council of South Africa. (2004, July) Whole Qualification Standard for Bachelor of Science in Engineering (BSc(Eng))/Bachelors of Engineering (BEng): NQF Level 7. http://www.ecsa.co.za/index.asp?x=documents&y=guidelines. [4] Ajay Malik, RTLS For Dummies.: Wiley, 2009. [5] David M Pozar, Microwave Engineering, 3rd ed.: Wiley, 2005. [6] Charles L Hutchinson and Larry D Wolfgang, The ARRL handbook for radio amateurs, 68th ed.: American Radio Relay League, 1990. [7] R.D. Straw, The ARRL antenna book, 21st ed.: American Radio Relay League, 2003. [8] S.R. SAUNDERS and A ARAGO N-ZAVALA, ANTENNAS AND PROPAGATION FOR WIRELESS COMMUNICATION SYSTEMS, 2nd ed. West Sussex, England: John Wiley & Sons, 2007. [9] I.F. Akyildiz, Z Sun, and M.C. Vuran, "Signal propagation techniques for wireless underground," Physical Communication, vol. 2, pp. 167-183, 2009. [10 D Dobkin, The RF in RFID.: Elsevier, 2008. ] [11 Salter W. Daneels A. (1999, December) What is SCADA? PDF. ] [12 Savi Technologies. (2010, March) Savi RFID Hardware. [Online].
] http://www.savi.com/media-form.php?d=ST-602_DS_09-09a.pdf&n=Savi%20ST602%20Asset%20Tag [13 Wavetrend. (2010, March) Wavetrend Products Tags. [Online].
] http://www.wavetrend.net/UserFiles/Files/TG100-A-PI-02-LT.pdf [14 Ekahau. (2010, March) Ekahau RTLS. [Online].
] http://www.ekahau.com/images/stories/documents/ekahau_t301a_datasheet.pdf [15 Savi Technologies. (2010, March) Savi Active RFID Readers. [Online]. 77
School of Electrical, Electronic and Computer Engineering
] http://www.savi.com/media-form.php?d=SR-650_DS_09-09a.pdf&n=Savi%20SR650%20Fixed%20Reader [16 Becker Electronics. (2010, September) Active Tag Reader Datasheet UATR400ZAA. Email. ] [17 Wavetrend. (2010, March) Wavetrend Products Readers. [Online].
] http://www.wavetrend.net/UserFiles/Files/L-RX201-04-PI.pdf [18 Cybertrack. (2010, March) RFID Insitude S.A. [Online].
] http://www.rfidtec.co.za/backend/media/11132007111709/Locatxrealtimelocatingsystems.p df [19 Mine Site Technologies. (2010, March) Mine Site Technologies. [Online].
] http://www.minesite.com.au/ [20 (2010, October) MySQL.Net Connector. [Online].
] http://dev.mysql.com/downloads/connector/net/ [21 Roberto M. Amadioa and Silvano Dal Ziliob, "Resource control for synchronous cooperative ] threads," Theoretical Computer Science, vol. 358, no. 2, pp. 229-254, August 2006. [22 Gyu Sang Choi and Chita R. Das, "A Superscalar software architecture model for Multi] Core Processors," The Journal of Systems and Software, vol. 83, no. 10, pp. 1823-1837, October 2010. [23 (2010, October) MySQL database software. [Online]. http://www.mysql.com/ ] [24 (2010, October) PHP scripting language. [Online]. http://www.php.net/ ] [25 (2010, October) VirtualBox. [Online]. http://www.virtualbox.org/ ] [26 (2010, October) putty ssh client. [Online].
] http://www.chiark.greenend.org.uk/~sgtatham/putty/ [27 (2010, October) Microsoft.Net. [Online]. http://www.microsoft.com/net/ ] [28 (2010, October) Microsoft VisualBasic. [Online]. http://msdn.microsoft.com/en-
] us/vbasic/default.aspx
School of Electrical, Electronic and Computer Engineering
78
[29 Ieiworld.
(2010,
March)
Ieiworld
-
WAFER-945GSE.
[Online].
] http://www.ieiworld.com/files/product/08274643094896902127/catalog/en-US/WAFER945GSE_20090715.pdf [30 Beckhoff. (2010, March) Beckhoff Industrial PC. [Online].
] http://download.beckhoff.com/download/Document/Catalog/Main_Catalog/english/separatepages/Industrial_PC/C6915.pdf [31 E Manson et al., "Radio wave propagation in arch-shaped tunnels: Measurements," ] Comptes Rendus Physique, vol. IV, no. 11, pp. 44-53, January 2010. [32 S.R. Saunders and A Arafo N-zavala, Antennas and propagation for wireless ] communication systems, 2nd ed. West Sussex, England: John Wiley & Sons, 2007.
School of Electrical, Electronic and Computer Engineering
79
APPENDIX A – TURN IT IN REPORT
School of Electrical, Electronic and Computer Engineering
80
School of Electrical, Electronic and Computer Engineering
81
EERI 429
Active RFID Feasibility study
S.S. Cox 20285132
Study leader:
Prof. J.E.W. Holm
OCTOBER 2010
Faculty of
Engineering
School for Electronic and Computer Engineering
Acknowledgements
This final year project would not be complete without acknowledging all who contributed. Firstly I would like to thank God for granting me the ability to complete this project; To my fiancé Jorene Bosman, thank you for your patience and understanding; To my parents Stephen and Marina Cox, thank you for sculpting my consciousness in such a unique way and for all your sacrifices; To my study leader Prof. J.E.W. Holm, for the inspiration and creative thinking; To Prof. L. Liebenberg, for inspiration and leadership; To HVAC International, for providing me with the project; To Becker Electronics, for lending me the hardware to complete this project. To Arno de Coning, Andreas Alberts, Andre Botha, Henri Marais and Walter Booysen.
School of Electrical, Electronic and Computer Engineering
ii
School for Electronic and Computer Engineering
Executive summary
Tracking of personnel in mines is critical to the safety of miners, technicians, managers and contractors. Currently, a system is used that makes use of contactless access control cards. However, personnel cannot be tracked once they have gone underground. In order to alleviate this problem, an active RFID solution is proposed. Since the actual operational performance of such a system is not known, it is required for this design project to simulate a mining environment with different scenarios. Such a simulation will allow the system designer to make critical design decisions and plan the layout of a system. A simulation package to this effect does not exist and is a critical need. Therefore, the aim of this project is to create a simulation package that is based on actual hardware, an actual environment, and realistic scenarios in a mining environment. A study of wave propagation in mines was done, available (commercial) hardware was analysed, and a software simulation package was created that allows a designer to plan a tracking system.
School of Electrical, Electronic and Computer Engineering
iii
School for Electronic and Computer Engineering
Declaration
I, Stephen Cox, declare that dissertation is a presentation of my own original work, conducted under the supervision of Johann Holm. Whenever contributions of others are involved, every effort is made to indicate this clearly, with due reference to the literature. No part of this work has been submitted in the past, or is being submitted, for a degree or examination at any other university or course.
Signed on this ___ day of _____________ 2010, in Potchefstroom.
_________________________________ S.S. Cox
School of Electrical, Electronic and Computer Engineering
iv
School for Electronic and Computer Engineering
Table of Contents
CHAPTER 1 - INTRODUCTION ................................................................................................................................................. 1 1.1 1.2 1.3 1.4 1.5 BACKGROUND ............................................................................................................................................................... 2 PROBLEM STATEMENT .................................................................................................................................................. 2
METHODOLOGY ................................................................................................................................................................. 5
PROJECT MANAGEMENT ..................................................................................................................................................... 9 DOCUMENT LAYOUT......................................................................................................................................................... 13
CHAPTER 2 - LITERATURE STUDY .......................................................................................................................................... 14 2.1 2.2 2.3 2.4 2.5 LITERATURE STUDY ........................................................................................................................................................... 15
TECHNOLOGY STUDY ......................................................................................................................................................... 20
EXISTING SYSTEMS ........................................................................................................................................................... 25 SHORTFALL OF THE EXISTING SYSTEMS AND TECHNOLOGIES ........................................................................................... 26 DESIGN PHILOSOPHY ........................................................................................................................................................ 26
CHAPTER 3 – PRELIMINARY DESIGN ..................................................................................................................................... 28 3.1 3.2
FUNCTIONAL ARCHITECTURE ............................................................................................................................................... 29
INTERFACES .................................................................................................................................................................... 30
CHAPTER 4 – DETAIL DESIGN ................................................................................................................................................ 31 4.1 4.2 4.3 4.4 SYSTEM DESIGN .............................................................................................................................................................. 32 MINE SIMULATOR ............................................................................................................................................................ 41 READER SIMULATOR ......................................................................................................................................................... 48 SERVER DISPLAY .............................................................................................................................................................. 53
CHAPTER 5 – VALIDATION .................................................................................................................................................... 58 5.1 5.2 5.3 5.4 5.5 5.6 OVERVIEW ..................................................................................................................................................................... 59 ENVIRONMENT ................................................................................................................................................................ 59 HARDWARE USED ............................................................................................................................................................ 62 TESTS ............................................................................................................................................................................ 62 RESULTS......................................................................................................................................................................... 63 DISCUSSION .................................................................................................................................................................... 66
CHAPTER 6 – IMPLEMENTATION AND EVALUATION ............................................................................................................ 67 6.1 6.2 6.3 IMPLEMENTATION ............................................................................................................................................................ 68 EVALUATION ................................................................................................................................................................... 69 SYSTEM TEST ................................................................................................................................................................... 69
CHAPTER 7 - CONCLUSION AND RECOMMENDATIONS ........................................................................................ 73 7.1 7.2 PROJECT HISTORY - REVISITED ............................................................................................................................................. 74 CONCLUSION .................................................................................................................................................................. 74
School of Electrical, Electronic and Computer Engineering
v
School for Electronic and Computer Engineering 7.3 7.4 RECOMMENDATIONS ........................................................................................................................................................ 75 ECSA OUTCOMES ............................................................................................................................................................ 75
APPENDIX A – TURN IT IN REPORT ....................................................................................................................................... 80
School of Electrical, Electronic and Computer Engineering
vi
School for Electronic and Computer Engineering
List of Figures
Figure 1 : System layout .............................................................................................................. 3 Figure 2 : System design concept that will be simulated .............................................................. 5 Figure 3 : Work breakdown structure ........................................................................................... 6 Figure 4 : Project schedule .......................................................................................................... 7 Figure 5 : Data Model Design ...................................................................................................... 7 Figure 6 : RFID data logger Design ............................................................................................. 8 Figure 7 : Server Software Design ............................................................................................... 8 Figure 8 : Tracking System Design .............................................................................................. 9 Figure 9 : Active RFID relevant to EM spectrum ........................................................................ 16 Figure 10 : Plane wave incident onto a plane boundary ............................................................ 17 Figure 11 : Finding the point of reflection using Fermat‟s principle ............................................ 17 Figure 12 : Refraction ................................................................................................................ 18 Figure 13 : Wave scattering ....................................................................................................... 18 Figure 14 : 2 Path model ............................................................................................................ 19 Figure 15 : Savi ST602 [12] ....................................................................................................... 20 Figure 16 : TCAP 302 ................................................................................................................ 21 Figure 17 : TCAP 302 fitted on a standard mining cap lamp battery .......................................... 21 Figure 18 : Wavetrend TG100-A [5] ........................................................................................... 21 Figure 19 : Ekahau T301A [6] .................................................................................................... 22 Figure 20 : Savi SR-650 [7] ........................................................................................................ 22 Figure 21 : UATR 400 [8] ........................................................................................................... 23 Figure 22 : Wavetrend RX201 [9]............................................................................................... 23 Figure 23 : Wavetrend RX900 [9] ............................................................................................... 24 Figure 24 : Handheld reader [12] ............................................................................................... 25 Figure 25 : Impact infrastructure ................................................................................................ 25 Figure 26 : Functional system overview ..................................................................................... 29 Figure 27 : Detail system overview ............................................................................................ 32 Figure 28 : MineSimulatorDrawing screenshot .......................................................................... 32 Figure 29 : MineSimulatorDrawing's route planner .................................................................... 33
School of Electrical, Electronic and Computer Engineering
vii
School for Electronic and Computer Engineering
Figure 30 : Reader propagation model ...................................................................................... 34 Figure 31 : MineSimulator version 2 .......................................................................................... 35 Figure 32 : Reader Simulator Screenshot .................................................................................. 36 Figure 33 : Data logger script ..................................................................................................... 37 Figure 34 : Server Control software flow .................................................................................... 39 Figure 35 : ServerDisplay version 3 ........................................................................................... 41 Figure 36 : Movement calculation .............................................................................................. 44 Figure 37 : Final reader propagation model ............................................................................... 44 Figure 38 : Simplification of the functional flow of MineSimulator .............................................. 47 Figure 39 : Visual identification indication .................................................................................. 48 Figure 40 : SET tag data ............................................................................................................ 51 Figure 41 : GET tag data............................................................................................................ 51 Figure 42 : Reader Simulator flow .............................................................................................. 52 Figure 43 : ServerDisplay version 1 ........................................................................................... 53 Figure 44 : ServerDisplay........................................................................................................... 54 Figure 45 : Locating algorithm .................................................................................................... 55 Figure 46 : ServerDisplay version 3 ........................................................................................... 56 Figure 47 : Test layout ............................................................................................................... 59 Figure 48 : Section 1 .................................................................................................................. 60 Figure 49 : Section 1 (Left) and section 4 (Right) ....................................................................... 60 Figure 50 : Section 3 .................................................................................................................. 61 Figure 51 : Section 2 (Right) and section 4 (Left) ....................................................................... 61 Figure 52 : RFID Tag placement ................................................................................................ 61 Figure 53 : Backer Electronics HTRD100ZA .............................................................................. 62 Figure 54 : Becker Electronics TPER200WW ............................................................................ 62
School of Electrical, Electronic and Computer Engineering
viii
School for Electronic and Computer Engineering
List of Tables
Table 1 : Preliminary budget ...................................................................................................... 10 Table 2 : Risk table .................................................................................................................... 11 Table 3 : Route calculation list details ........................................................................................ 34 Table 4 : Data logger database structure ................................................................................... 38 Table 5 : Server database structure ........................................................................................... 39
School of Electrical, Electronic and Computer Engineering
ix
School for Electronic and Computer Engineering
List of Abbreviations
RFID ISO IEC GPS SABS NWU ID PC ECSA ELO EMS SCADA API EMPA PHP SQL RTLS POI Radio frequency identification International Organisation for Standardization International Electrotechnical Commission Global Positioning System South-Africa Bureau of Standards North-West University Identification Personal Computer Engineering Council of South-Africa Exit Level Outcome Energy Management System Supervisory Control and Data Acquisition Application Programming Interface Energy Management Personnel Awareness PHP: Hypertext Preprocessor Structured Query Language Real-time locating systems Point of interest
School of Electrical, Electronic and Computer Engineering
x
Chapter 1 - Introduction
CHAPTER 1 - INTRODUCTION
The aim of this chapter is to provide an overview of the design project. This includes making a formal problem statement that forms the basis and motivation of the project. Furthermore, the chapter contains objectives and the scope of the design project. To describe the project a work breakdown structure, design methodology, quality assurance plan, preliminary budget, risk analysis, and ECSA outcome assessment are given.
School of Electrical, Electronic and Computer Engineering
1
Chapter 1 - Introduction 1.1 BACKGROUND In South Africa the mining sector is a major component of the economy. One of the most challenging aspects of mines is safety. One element of safety in mines is the location of people and assets. GPS tracking on open ground is the most common tracking system today. Since GPS uses satellites for tracking it will not be applicable to underground applications. The most commonly used tracking in applications without GPS is localized identification-based tracking. The most commonly used identification tracking is the well-known barcode system. For tracking applications in mines, the barcode system will not be suitable due to physical limitations. The recent advancement and increasing research and development on RFID showed that RFID can be used as an effective tracking mechanism. RFID tracking can be used to increase safety, offer more security, and increase energy efficiency in a mine. In this project, only safety will be focused on although the results of the research will also be used in energy management systems (to know which areas in a mine may be shut down to save energy, it is necessary to know if miners are present in those areas). This project is innovative in the sense that there will be a unique simulation of an active RFID real-time tracking system. This simulation will be used as a layout tool for optimizing a tracking system‟s design. This system will generate data that is otherwise very expensive to acquire from an actual, physical system. 1.2 PROBLEM STATEMENT
Verify the feasibility of an active RFID tracking system for use in the underground mining sector by using software simulation. 1.2.1 Project Scope Project requirements were received from Hades Technologies (Pty) Ltd (the client). Areas to be researched: Existing technologies and systems; Tag and reader functioning; Software simulation procedures; Adjustable parameters; Database software; RF propagation effects;
School of Electrical, Electronic and Computer Engineering
2
Chapter 1 - Introduction RFID RTLS system functioning.
Outcomes to be met when the project is finished: Implementation and demonstration of the system simulation; System adjustability and customisability; Database support and integration support with other software; Software that handle events; Reader propagation model for validation.
The following figure is the author‟s interpretation of the provided specifications and shows the system to be simulated.
Figure 1 : System layout
School of Electrical, Electronic and Computer Engineering
3
Chapter 1 - Introduction 1.2.2 Project Objectives 1.2.2.1 Primary Objectives
Investigate the feasibility of an underground tracking system using active RFID tracking system simulation. 1.2.2.2 Secondary Objectives
Extend the simulator design to verify safety and energy efficiency applications. 1.2.2.3 Critical Success Factors
In order to deliver a successful system, the following must be in place: An RFID tracking system simulation to identify the last known position of personnel; Software that reads events and displays them; Validation of the RFID reader propagation model. 1.2.2.4 Deliverables
The following deliverables shall be delivered per milestone: Milestone 1 – Project proposal documentation; Milestone 2 - Introduction and Literature Study of final documentation; Milestone 3 – Presentation of proposal and conceptual design; Milestone 4 – Chapters 1 to 3 of final documentation; Milestone 5 – Demonstrate project; Milestone 6 – Fully demonstrate project; Milestone 7 – Hand in concept of final documentation; Milestone 8 – Final report. At milestone 5 there will be the following: RFID RTLS concept; Sub system simulation software; System integration concept.
School of Electrical, Electronic and Computer Engineering
4
Chapter 1 - Introduction 1.3 METHODOLOGY 1.3.1 Proposed design methodology
Figure 2 : System design concept that will be simulated
School of Electrical, Electronic and Computer Engineering
5
Chapter 1 - Introduction Systems engineering will be used throughout the project [1]. The following is the design process that will be followed: Definition of need; Need analysis; Preliminary design; Detail design and development. The following is a breakdown of the documentation process: Project proposal; Introduction and literature study; Preliminary design; Final report. 1.3.2 Work Breakdown Structure
NWU Project 2010
Research
Design
Develop
Current RFID Technologies
Tag and Reader functioning
Data model
RFID Tracking System Simulator
Software simulation procedures Radio Frequency Propagation effects
Database Software
Simulator
RFID RTLS System functioning
Figure 3 : Work breakdown structure
School of Electrical, Electronic and Computer Engineering
6
Chapter 1 - Introduction 1.3.3 Schedule Milestone 1 (05-Feb) Milestone 2 (19-Mrt) Milestone 3 (17-Apr) Milestone 4 (01-June) Research (02-June) Milestone 5 (16-Jul) System adjustments (17-Jul) Milestone 6&7 (13-Oct) Milestone 8 (19-Oct)
•Project proposal •Introduction and literature study •Conceptual design •Chapters 1,2 and 3 •System components, software and communication infrastructure •Concept demonstration •Optimisation •Demonstration and concept of final documentation •Final documentation
Figure 4 : Project schedule
1.3.4 Design Proposal The following figures are a graphical illustration of the proposed designs.
Data model
Database software
RFID data protocols
Interface
RFID data logger
Server
Figure 5 : Data Model Design
School of Electrical, Electronic and Computer Engineering
7
Chapter 1 - Introduction
RFID data logger
Hardware
Software
Enclosure
Computer
Network
Data Model
Data acquisition software
Figure 6 : RFID data logger Design
Server
Software
Hardware
Data Model
GUI
Data acquisition software
Computer
Network
Integration with EMS
Tracking results
Figure 7 : Server Software Design
School of Electrical, Electronic and Computer Engineering
8
Chapter 1 - Introduction
RFID Tracking System Hardware Required RFID data logger Single-board Computer
Software
Server
RFID
Database
Computer
Tag
Reader
Data logger
Simulator
Figure 8 : Tracking System Design 1.4 PROJECT MANAGEMENT 1.4.1 Available resources The following people are available to contact in regards of my project: Prof. J.E.W. Holm (Mining technology); Prof. Hoffman (RFID technologies); Mnr. T. Tromp (Becker Electronics); Mr. P. Van Huyssteen (Mining network technology); Mr. H. Marais (Software programming); Mr. A. Alberts (Database design). Available Software SuperNec (Antenna simulation); Matlab (Simulations using large volume mathematics); Microsoft Visual Studio (GUI design); MySQL (Database design). PHP (Data logger software) Available Finances NWU (Maximum R2500);
School of Electrical, Electronic and Computer Engineering
9
Chapter 1 - Introduction Hades Technologies (Pty) Ltd. 1.4.2 Preliminary Budget The following is an approximation of the development costs to verify the simulation by the author: RFID Tag R 330.00 RFID Reader R 13,000.00 RFID data logger Computer R 3,000.00 UPS R 500.00 Casing R 1,000.00 Server (Industrial standard) Computer R 25,000.00 UPS R 5,000.00 Casing R 10,000.00 Table 1 : Preliminary budget
School of Electrical, Electronic and Computer Engineering
10
Chapter 1 - Introduction 1.4.3 Associated risks Calculating a risk factor involves determining a fail probability and a consequence factor. [2] The associated risk is the product of the fail probability and the consequence factor. Both factors are between 0% and 100%, where 0% is no failure and 100% is total failure.
Time [A] System Integration Communication protocol Data model Functional unit component research Cost analysis Application layer Implementation 70% 65% 80% 60% 30% 30% 60% Resources [B] 50% 60% 50% 50% 40% 60% 40% Cost [C] 30% 40% 0% 0% 0% 0% 10% Fail Probability [D=0.4xA+0.35xB+0.25xC] 53% 57% 50% 42% 26% 33% 41% Consequence [E] 100% 70% 75% 90% 100% 75% 60% Risk Factor [F=DxE] 53% 40% 37% 37% 26% 25% 24% Global risk factor 77% 58% 54% 54% 38% 36% 35%
Table 2 : Risk table
Probability of failure is calculated by looking at previous data. This previous data can also be personal experience. Probability of failure is obtained by looking at time-, resourceand cost failures. Each of these factors has a selected weight determined by judging the project circumstances; Time failure is when an item causes delays in the design process or when insufficient time had been allowed within which to complete the project; Resource failure is when people cannot act as resources or when components or equipment fail, or when it is difficult to understand a resource; Cost failure is when an item‟s price or cost fluctuates over time, or an item causes additional expenses (such as services cost); The consequence factor is obtained by determining the impact of the item‟s failure on the project as a whole [2]; The global risk factor is a normalised value when the average of the risk factors is set to 50%. An item that has a risk factor above 50% is a risk above normal and more attention should be paid to that item.
School of Electrical, Electronic and Computer Engineering
11
Chapter 1 - Introduction 1.4.4 Quality assurance The following will determine the quality of the project: RFID software components with minimal user operation; System scalability; Software customisability; Software simulation results; Reader propagation model accuracy; System integration. 1.4.5 ECSA exit level outcomes All relevant information on ECSA ELO‟s was gathered from [3]. 1.4.5.1 Problem solving
In this project there exists a need to be solved. 1.4.5.2 Application of scientific and engineering knowledge
Working with a radio frequency application there will be mathematics and physics involved. 1.4.5.3 Engineering Design
There will be a system design involved in this project. 1.4.5.4 Investigations, experiments and data analysis
For the designs in this project there will be an investigation, experiment and data analysis. 1.4.5.5 Engineering methods, skills and tools
In this project there will be an investigation in mining communication infrastructure. 1.4.5.6 Professional and technical communication
In this project there will be communication with mining professionals and project managers. 1.4.5.7 Impact of engineering activity
This project will increase the safety in mines. 1.4.5.8 Individual, team and multidisciplinary working
In this project there will be interaction with project manager, mining professionals and technicians. The author of this document will be responsible for this project.
School of Electrical, Electronic and Computer Engineering
12
Chapter 1 - Introduction 1.4.5.9 Independent learning ability
The author of this document will work strategically, complete tasks effectively and will deliver work on time. 1.4.5.10 Engineering Professionalism Being professional is a moral decision and will be done throughout the whole project. 1.5 DOCUMENT LAYOUT Documentation started with an introduction that disused that problem and its background. After that a more detailed background was discussed in the literature study. Following the detail background was a preliminary design, which purposed an possible design to solve the problem. This design was discussed in detail after the purposing design was evaluated. After the detail design validation was done to strengthen the design. Then implementation and evaluation was discussed and finally the conclusion was done.
This concludes chapter 1, where a formal problem statement was made and objectives of the project and the scope of the project were defined. A work breakdown structure, design methodology, quality assurance plan, preliminary budget, risk analysis and ECSA outcome assessment were given.
School of Electrical, Electronic and Computer Engineering
13
CHAPTER 2 - LITERATURE STUDY
The aim of this chapter is to give a theoretical background on components of the RFID tracking system. It starts with an overview of existing systems and technologies and moves to design philosophy, and finally discusses the theoretical background of sub components.
School of Electrical, Electronic and Computer Engineering
14
2.1 LITERATURE STUDY 2.1.1 Real time locating systems (RTLS) RTLS‟s are used to identify and track the location of objects. Locating is achieved by attaching tags/badges to an object and then identifying these tags/badges by means of wireless signals at specific known locations the readers are placed. These readers will relay information to a central server. RTLS is a system that collects location information passively. [4] The RTLS standard excludes passive RFID indexing as well as beacon systems. Locating is achieved by the RTLS standard. This involves single/multiple read transactions with the readers, locating algorithm, relay of information and distance measuring between readers. RFID tracking is in a sense just like RTLS except that is does not determine the exact location of the object being tracked. 2.1.2 Electromagnetic (EM) spectrum The EM spectrum is the collection of all EM waves; categorized by frequencies. The behaviour of the waves depends on their frequency. Relevant equations are wavelength ( energy ( E
c f ) and
fh ); where c is the speed of light in a vacuum and h is Planck„s constant. [5]
School of Electrical, Electronic and Computer Engineering
15
Very Low Frequencies (3-30 kHz) Low Frequencies (30-300 kHz) Medium Frequencies (300-3000 kHz) High Frequencies (3-30 MHz) Very High Frequencies (30-300 MHz) Ultra High Frequencies (300-3000 MHz) Super Hign Frequencies (3-30 GHz) Extremenly high Frequencies (30-300 GHz) Terahertz (300+ GHz) Active RFID
Radio
Microwave
Infrared
Electromagnetic spectrum
Visible
Ultraviolet
X-rays
Gamma rays
Figure 9 : Active RFID relevant to EM spectrum 2.1.3 Radio waves The EM spectrum of importance in this document is the radio frequency spectrum. Governments regulate the use of this spectrum through frequency allocation. Collective oscillation of charge carriers in material causes radio waves. Radio waves are utilised by antennas. Antenna size is an important factor; it must be sized according to the principle of resonance. This will determine the bandwidth and optimal frequency on the antenna [6]. Radio waves can modulate information by means of varying the amplitude, frequency and phase of the wave within a defined frequency range. When EM radiation acts upon a conductor it couples to the conductor. The EM radiation travels along the conductor and induces an electric current on the surface by exiting electrons in the conductor; this effect is called the skin effect. [7] These waves are affected by reflection, refraction, diffraction, absorption, polarization and scattering.
School of Electrical, Electronic and Computer Engineering
16
2.1.3.1
Propagation
Some of the most interesting aspects of waves are propagation effects. This is important due to energy losses and time delays in transmission of waves. In underground mining the propagation effects include reflection, refraction, scattering and multipath fading. The effects will be discussed in the following paragraphs. Reflection First we look at lossless scenarios. Depending on the angle of incidence the incoming wave energy is redirected to transmission-, reflection- and absorbed energy. This can be proved by using Snell‟s laws. Figure 10 shows a wave reflecting and transmitting some of its energy. In the figure
r i
; angle of reflection is equal to the angle of incidence. [8]
Figure 10 : Plane wave incident onto a plane boundary Furthermore, if the wave is reflected from a surface it will most likely choose a path as defined in Fermat‟s principle. Thus it is highly likely that a wave will follow the minimum time to the destination. This can be seen in Figure 11.
Figure 11 : Finding the point of reflection using Fermat’s principle 17
School of Electrical, Electronic and Computer Engineering
Refraction Refraction is the phenomenon when it seems as if a wave bends. This effect can normally be seen when light waves move trough different mediums or when observing something that is underwater. The effect is due to difference in refraction indices. This effect can be described using Snell‟s law which states the following:
sin sin
1 2
n1 , where n is the respective refraction n2
index. The effect can be seen in the following figure. [8] [9]
Figure 12 : Refraction Scattering Scattering is an effect when the surface is not smooth and the energy of the wave is not effectively directed in the path of the wave. The energy is radiated in other directions. This can be seen in the following figure [8] – this causes energy loss.
Figure 13 : Wave scattering Multipath fading In non-perfect media there is attenuation which is an indication of loss at a specific distance. When one combines the above mentioned effects and look again at the 2 path model, the multipath fading concept makes sense.
School of Electrical, Electronic and Computer Engineering
18
Figure 14 : 2 Path model Referring to Figure 14 there are two rays. The one is the direct path which a transmitted wave will follow and the other is a reflected wave. Bringing into account attenuation, refraction speed loss and scattering energy loss, it can be said that the reflected wave will arrive with a delay and decreased energy at the receiver. [9] Multipath fading contains most of the effect discussed and it is the effect that is expected to be seen in underground radio frequency communication. 2.1.4 Radio frequency identification RFID uses radio waves for communication. Various protocols have been defined and standardised. There exist three types of RFID technologies,.namely passive, semi-passive and active RFID. RFID is a recently rediscovered technology (it came into being due to development in radio technology during the world wars). This technology existed in the 1980‟s b ut was too expensive to implement. Due to decreasing electrical component costs this technology has emerged as a viable technology. The main advantage that active RFID has over other identification technologies is its range. Barcodes work exceptionally well but lack range; they need to be close to the reader. RFID has endless potential due to its wireless operation. Passive RFID tags use the energy from a magnetic field and do not contain a battery. It then uses the same energy to return communication. Semi-passive RFID tags use a battery but not when returning communication, when communicating back with a reader the tag relies on the magnetic energy from the reader. Active RFID tags use a battery to power itself and for communications. Active RFID tags can use higher frequencies due to the use of a battery [10]. 2.1.5 SCADA This system is usually used where an industrial process requires data acquisition and control. In simple terms this is the supervisory system to control a process, thus having a central server and clients. This system is ideal to control the RFID tracking system. Care should be taken in designing the hardware and software architecture.
School of Electrical, Electronic and Computer Engineering
19
Functions of the SCADA system include; trending, logging, report generation and automation. The benefits of using this type of system design is that is will be dependable and data can be gathered at a single point [11].
2.2 TECHNOLOGY STUDY 2.2.1 RFID tags A RFID tag consists of an antenna and a processing unit with storage. The different types of tags will be discussed later in section 2.1.4. This device has different protocols for receiving / sending data to readers, but usually has a unique ID within the system. This device acts as an identifying medium for assets. Thus one can use this device for locating assets within a tracking system. The following sections present different commercially available tags. 2.2.1.1 Savi Technologies
Savi Technology Inc makes an active RFID tag named the ST-602 asset tag [12]. This tag can send data at 433 MHz and receive data at 123 kHz. It is low-power device and has a transmission range of 122 meters [12]. The tag has dimensions 6.2cm x 4.3cm x 1.2cm. The tag can be seen in Figure 15.
Figure 15 : Savi ST602 [12] Savi Technology Inc also makes an active RFID sensor tag. This tag has an on-board temperature and humidity sensor with adjustable alarms. Savi Technology Inc operates mainly in tracking containers. 2.2.1.2 Becker Electronics
Becker Electronics is part of Becker Mining who makes an active RFID tag named the TCAP 302. This is a beacon tag and has the unique feature that it fits inside a miner‟s cap lamp battery. This tag sends and receives data at 433 MHz. The tag can be seen in Figure 16 and Figure 17.
School of Electrical, Electronic and Computer Engineering
20
Figure 16 : TCAP 302
Figure 17 : TCAP 302 fitted on a standard mining cap lamp battery
2.2.1.3 Wavetrend Holdings
Wavetrend Holdings Limited makes an active RFID beacon tag named the TG100-A. This tag sends data at 433 MHz and transmits its ID at a predefined interval. This is a low-power device and has a long battery life. This tag features an anti-tampering function and has the following dimensions; 6.4cm x 3cm x 0.9cm [13]. The tag can be seen in Figure 18.
Figure 18 : Wavetrend TG100-A [13] 2.2.1.4 Ekahau
Ekahau makes an active RFID tag named the T301A [14]. This tag transmits and receives data at 2.4 GHz. This tag can easily be integrated into an existing Wi-Fi network that uses the 802.11a/b/n standards. This tag is a low-power device, has excellent security, supports Ethernet protocols and has motion detection [14]. The tag has dimensions 4.5cm x 5.5cm x 1.9cm. The tag can be seen in Figure 19.
School of Electrical, Electronic and Computer Engineering
21
Figure 19 : Ekahau T301A [14] The Becker Electronics tags are suitable for mining and were selected due to their form factor and their robustness. 2.2.2 RFID readers An RFID reader is a device that can interpret data received from RFID tags and can relay this data on a network interface. Most active RFID readers do not send a signal and listen for tag transmissions on a specific frequency from RFID tags. These readers normally support more than one RFID protocol or standard. 2.2.2.1 Savi Technologies
Savi Technology Inc makes an active RFID reader named SR-650 [15]. This reader supports any tag using the ISO 18000-7 standard, has an omni-directional range of 122 meters, receives data on 433 MHz and supports Ethernet/RS485 [15]. The reader has dimensions 30cm x 14cm. The reader can be seen in Figure 20.
Figure 20 : Savi SR-650 [15] 2.2.2.2 Becker Electronics
Becker Electronics make an RFID reader named UATR 400 [16]. This reader only supports the Backer family tags. This reads tags at 433MHz and can transmit the data to RS232/RS485/RS422. What makes this reader unique is that it complies with IP68 standard and it was designed for South African mines. The reader has dimensions 22cm x 12cm x 9.5cm. The reader can be seen in Figure 21.
School of Electrical, Electronic and Computer Engineering
22
Figure 21 : UATR 400 [16]
2.2.2.3
Wavetrend Holdings
Wavetrend Holdings Limited makes an RFID reader named RX201 [17]. This reader supports only the Wavetrend family tags. This reader receives data at 433 MHz, interfaces with RS232/RS485 and supports an external antenna [17]. The reader has dimensions 8.4cm x 4cm x 1.9 cm. The reader can be seen in Figure 22.
Figure 22 : Wavetrend RX201 [17] Wavetrend Holdings Limited also makes an RFID reader named RX900. This reader supports only the Wavetrend family tags. This reader receives data at 433 MHz, interfaces with Ethernet and supports an external antenna. The reader has dimensions 13cm x 14cm x 2.8cm. The reader can be seen in Figure 23.
School of Electrical, Electronic and Computer Engineering
23
Figure 23 : Wavetrend RX900 [17] Once more, the Becker Electronics readers were used due to their robustness and effectiveness in the mining environment. 2.2.3 RFID software RFID software is required to translate the data - collected by the readers - into a user readable format. Each company designing RFID tags and readers makes its own software according to its own needs and preferences. Some of the companies also make an API to let programmers interface with their software. Without the API there is no access to the data from the readers due to companies‟ privacy policies.
School of Electrical, Electronic and Computer Engineering
24
2.3 EXISTING SYSTEMS 2.3.1 Locatx (RTLS) LOCATX is a company who designed an RTLS using RFID technology at UHF and VHF frequencies [18]. The RTLS system operates at dual frequencies (13.56MHz and 433 MHz). The RFID tag incorporates a photo of the miner and the handheld reader is also used for security; the reader is shown in Figure 24. Furthermore the system has an alarm equal to a panic button and access violation notification. [18]
Figure 24 : Handheld reader [18] 2.3.2 Mine site technologies (Impact) Mine site technologies is an Australian based company specialising in mining communication. This company developed a personnel and asset tracking system called Impact. The infrastructure of Impact can be seen in Figure 25. This system has been fully developed, tested and implemented; according to the datasheet. [19]
Figure 25 : Impact infrastructure 2.3.3 Other RTLS’s Most of the RFID RTLS systems available have the same structure. This structure entails tags, readers and a central server with appropriate software. More advanced systems like Wavetrend‟s Total Enterprise Asset Visibility has a unique approach that puts it above the rest, due to their research focussing on RTLS‟s.
School of Electrical, Electronic and Computer Engineering
25
RTLS has a different connotation when used in logistics above ground than when used underground. Underground systems do not require excessive performance in terms of location accuracy since the people and assets are constrained to move along specific routes. Therefore, simplistic tagging is the preferred technology for underground – such as the Becker Electronics system. 2.4 SHORTFALL OF THE EXISTING SYSTEMS AND TECHNOLOGIES Currently, existing systems are expensive and thus not directly applicable to use in South African underground mines. As a result, the systems must be implemented by distributors, resulting in extra costs and time delays. Current technologies also use Wi-Fi frequencies for communication. At these high frequencies the effects of radio waves get accentuated and more attention must be given to correct this; thus becoming more expensive. The inter-networking protocols and standards are difficult to implement for Wi-Fi systems underground (due to physical and environmental challenges). There exist systems using a carrier frequency of 433 MHz, but these systems are not designed for underground use. Becker mining is the only company that designs RFID hardware for underground usage as far as could be ascertained. The tag, reader and single-board computer technologies available today are exactly what are needed to create an RTLS system for underground use. Some of these systems lack verification for use in underground conditions. Most of these systems cannot function when a central server is down; this however can be fixed with a data logger near the RFID readers. The simulation of all the components of an RTLS is possible, but care must be taken not to deviate from the required functionality. 2.5 DESIGN PHILOSOPHY The systems engineering life cycle is useful when designing a system that will be implemented. This life cycle contains preliminary design, detail design, production, utilisation and phase-out. For the purposes of this project the preliminary- and detail design will be done [1]. 2.5.1 Preliminary design In the preliminary design, subsystems are designed to adhere to specification and interfaces between these subsystems are defined. After completion of this design, a development specification is defined. Important steps in the preliminary design are the functional analysis, trade-off studies and development specification.
School of Electrical, Electronic and Computer Engineering
26
2.5.2 Detail design From the detail design emanate the final specifications of the design. These specifications may produce a change in the development specification. Important steps in the detail design are the final specifications and the development of a prototype.
This concludes chapter 2. The aim of this chapter was to give a theoretical background on components of the RFID tracking system. The chapter started with a discussion of the theoretical background of sub components, then moved to an investigation into existing systems and finally defined the design philosophy.
School of Electrical, Electronic and Computer Engineering
27
CHAPTER 3 – PRELIMINARY DESIGN
The aim of this chapter is to give physical definition to the theories explained so far. Here the functional architecture, process flow and the data model are presented.
School of Electrical, Electronic and Computer Engineering
28
3.1 FUNCTIONAL ARCHITECTURE The functionality of the system to be developed is shown below. This is a basic overview of the system and will be used to reference and describe the functional units of the system.
Figure 26 : Functional system overview 3.1.1 Input Each system requires input to start the processing. Even default parameters are inputs. The system in Figure 26 has mine layout and system parameters as inputs. 3.1.1.1 Mine layout
The mine layout consists of a grid. Each item in the grid must correspond to an object. The object will range from walls to readers. A grid is use to simplify the process of designing a layout. 3.1.1.2 System parameters
The system parameters consist of the asset movement characteristics, reader propagation model, RFID hardware specifications, and communication parameters. 3.1.2 Simulation Some hardware components of RFID RTLS will need to be simulated. These components are a RFID reader and the asset movement. 3.1.2.1 Asset movement
The movement of assets and people inside a mine needs to be simulated. The system input in 3.1.1 will be used to generate movement of assets and people. This movement must be accurate and representative of actual movement. The tag/reader interface will be simulated with a reader propagation model. 3.1.2.2 RFID Reader
The basic functioning of a RFID reader is a device that receives data from RFID tags and stores it in a buffer for collection from a data logger or server PC. This can also be simulated by using
School of Electrical, Electronic and Computer Engineering
29
software. The reader simulator will consist of threading, buffers and communication components. 3.1.3 Data collection Data collection has 2 phases. Each phase has its own functionality and operates independently of the other. 3.1.3.1 Data logger
The data logger hardware is a scaled down PC that runs a simple PHP script. The operation of the data logger is to create a temporary data buffer in case the server goes down and the RFID reader‟s buffer overflows. Thus the data logger must be able to collect data from the reader simulator and store it in a buffer. The server can then collect the data from this buffer. 3.1.3.2 Server Control software
The server needs to collect data from the readers/data loggers. The server needs software to collect the data and store it in the correct format in a central database. 3.1.4 Data interpretation The data that was collect by 0 needs to be interpreted. Server display software will need to get a copy of the server‟s data and run an algorithm through it to locate assets. 3.2 INTERFACES 3.2.1 IF1 The RFID tag has a wireless interface with the RFID tag. This interface is done using radio frequency communication. 3.2.2 IF2 The interface between the RFID reader and the logger can be RS485/RS422/RS232. The most common interface is RS485. RS485 can easily be used by the data logger computer, it functions the same as a serial port. 3.2.3 IF3 The interface between the data logger and the server control is a MySQL data connection. 3.2.4 IF4 The interface between the server control and the server display is a MySQL data connection. This concludes chapter 3. The aim of this chapter was to give a physical definition to the theories explained so far. Here the functional architecture, process flow and the data model have been presented.
School of Electrical, Electronic and Computer Engineering
30
CHAPTER 4 – DETAIL DESIGN
The chapter provides the design in more detail. The chapter starts with a system overview and then discusses the sub-system components in detail.
School of Electrical, Electronic and Computer Engineering
31
4.1 SYSTEM DESIGN
Figure 27 : Detail system overview 4.1.1 Input 4.1.1.1 Mine Drawing
The program MineSimulatorDrawing is used to create a mine layout. MineSimulatorDrawing uses the same drawing functions as MineSimulator and the user is able to edit the layout. The layout is then saved and can be used by MineSimulator.
Figure 28 : MineSimulatorDrawing screenshot In Figure 28 it can be seen that MineSimulatorDrawing was implemented using a drawing grid. In the grid blocks can be drawn to indicate an object type. The blocks can be categorised as wall, walking space, point of interest and a reader. An important feature of 32
School of Electrical, Electronic and Computer Engineering
MineSimulatorDrawing is the route planner. MineSimulator will simulate mobile assets and people walking around the mine by using the pre-calculated routes. These routes are saved in a file that the MineSimulator can read.
Figure 29 : MineSimulatorDrawing's route planner The route planner can be seen in Figure 29. A route has 4 components; number, level, point and coordinates: 1. (R) Route numbers start at 1 and count numerically onwards. The user must input the route number manually. Each possible route to the end of each tunnel must have a unique number; 2. (L) Route levels start at 0 and count numerically onwards. The level is increased when the increase button is pressed. Route levels are to identify the progress of a route; 3. (P) Route points are used to identify the POI‟s of a route at a certain level. Each level must have 5 points; 4. (X#Y#) Route coordinates is used to link a point to the layout grid. An X- and Y-point is defined.
In Figure 29 there is a list called “Current POI‟s”. These entries follow a structure which can be seen in Table 3. .
School of Electrical, Electronic and Computer Engineering
33
Table 3 : Route calculation list details
Entry item R L P X#Y# Example 1 0 0 X13Y19 Explanation Route number Route level Route point at that level X- and Y-point in layout grid
The combination of the layout and route design lets the user prepare a mine model for the simulator. 4.1.1.2 Asset movement parameters
Important parameters for asset movement calculation are their average walking speed. The average walking speed of a human being is between 4.5 km/h to 5.2 km/h. In a mine the environmental conditions worsen the ability to maintain that speed. So it is been decided that the speed to work with is 3.5 km/h to 4 km/h. The RFID tag used to identify the asset also has some adjustable parameters. The Becker mining tag has a transmission interval of 2 to 3 seconds. 4.1.1.3 Reader propagation model
For validation, testing was done in an active underground mine to determine the RFID reader‟s propagation model. The details of this research are discussed in chapter 5. This model is reduced to a piecewise linear model to simplify the programming and implementation. The implemented model can be seen in Figure 30.
100
90 Percentage Packet loss 80 70 60 50 40 30 20 10 0 0 5 10 15 20 25 Distance in meters
Figure 30 : Reader propagation model
School of Electrical, Electronic and Computer Engineering
34
4.1.2 Simulation 4.1.2.1 MineSimulator
MineSimulator is a program that simulates the movement of people and assets that host RFID tags. Movement near RFID readers results in a reader propagation model evaluation. When a tag is within reach of a reader, the RFID tag data is sent to the reader simulator. This is the main function of the MineSimulator. Figure 31 is a screenshot of the program in action. The white dots in the figure are the assets moving in the mine and the large circles are the reader detection areas based on the propagation models. The mine layout can also be seen in this figure. MineSimulator will be discussed in 4.2.
Figure 31 : MineSimulator version 2 4.1.2.2 Statistics
Read accuracy statistics are generated by MineSimulator. These statistics identify the blind spots, packet errors and collisions for each reader and each tag. The relevant terms are as follows: Blind spot – If an asset is within the reader‟s range and there is no propagation circle covering it, then the asset fails to be identified since the tag could not be read;
School of Electrical, Electronic and Computer Engineering
35
Packet error – If an asset is within the reader‟s range and is located within a propagation circle with, for example, 50% packet error rate, then the asset has a probability of 50% to been successfully identified. A packet error results in a transmission failure; Collision – If an asset is within the reader‟s range and transmits in the same interval as any other asset, then the asset fails to be identified; Unique reader seen – If an asset enters the reader‟s range for the first time; Unique reader read – If an asset has entered a reader‟s range and was correctly identified; Safety rating – . This calculates a safety rating of a specific
layout, which is the number of unique times a tag was read by a reader ( read) divided by the number of times that tag was inside a reader‟s detection area (seen). 4.1.2.3 Reader Simulator
The functioning of the reader was discussed in 0. The reader was simulated by using threading, TCP sockets and buffers. The functioning of the reader simulator will be discussed in chapter 4.3. The reader can simulate a given number of RFID readers each in their separate thread.
Figure 32 : Reader Simulator Screenshot 4.1.3 Data collection Data collection starts by creating a temporary buffer for data storage, hence the use of a data logger in the simulator. The data logger retrieves data from the RFID readers and stores them in a MySQL database. The data is the collected from the data logger by the server PC. The software that will be responsible for this is Server Control.
School of Electrical, Electronic and Computer Engineering
36
The database structures are unique for the server and the data logger, but there is a shared structure. These data structures will be discussed in chapter 4.1.3.2 and chapter 4.1.3.4. 4.1.3.1 Data logger
The data logger was not simulated, but created as a working application that can be used in an RFID RTLS. The data logger software was written in PHP using a combination of object orientated and procedural functions. The basic flow of operation can be seen in Figure 33. The data logger software is a PHP script that can be executed in a Linux environment. Most singleboard computers can run a basic Linux distribution. Linux is a more stable and resource efficient operating system. There must be a MySQL database available locally with the data structure of the data logger database which will be discussed in chapter 4.1.3.2.
Figure 33 : Data logger script The data logger firsts checks to see if the server is online. If the server is online, the script will check if the data logger ID has been updated recently. If the server is offline, the script ends. In future version it will be able to continue. This feature is incomplete due to time constrains, but all the functions needed are already in place. The check for the ID update is necessary and needs to identify the data being logged. It is saved to a file to be used in offline mode. The ID update process involves a connection to the server database. The data logger ID can be found by using the data logger‟s IP-address and hostname.
School of Electrical, Electronic and Computer Engineering
37
When the above checks are done, the script continues and retrieves a list of all the readers. The readers are identified by IP-address and port number. The script goes through each entry and creates a TCP socket connection to the reader. When the connection is made, the data is retrieved using the protocol discussed in 0 The data that is retrieved is then stored in the local MySQL database. <screenshot> There are minor features not shown on the diagram. The script also creates a lock file. This file is created when the script is started and deleted when the script exits. This is used to prevent the script from being executed multiple times, only one copy of the script should run simultaneously. The script also has the feature to do DNS lookups. This is useful when operation is a network that has dynamic IP-addresses. Using DNS lookups, it can be prevented that the data is retrieved from the wrong addresses. 4.1.3.2 Data logger database
The data logger database is used as a temporary buffer for the data. The database structure can be seen in Table 4. Table 4 : Data logger database structure Table Data Field d_id d_tagid d_readerid Explanation Unique data entry ID. Used as a data index. Unique tag ID. Used to identify assets. Unique reader ID Used to identify reader; determine location. Unique data logger ID. Used to identify data logger; relation to readers. Date and time the data was acquired. Unique ID to identify the reader. Reader hostname used in DNS lookups. Reader IP-address. Reader location in mine grid. Data logger responsible for the data buffer of the reader. 38
d_dataloggerid
d_datetime
Reader
r_id r_devname r_ip r_location r_dataloggerid
School of Electrical, Electronic and Computer Engineering
r_port
Reader communication port.
4.1.3.3
Server Control
Server Control software is responsible for data collection from the data loggers. This software is not simulated and can be used in a RFID RTLS together with the data logger software. The operational flow of the Server Control can be seen in Figure 34. <sreenshot> Version 1.0 was the first functioning application and version 1.1 added new data structures and corrected threading errors.
Figure 34 : Server Control software flow 4.1.3.4 Server database
Table 5 : Server database structure Table Data Field d_id d_tagid Explanation Unique data entry ID. Used as a data index. Unique tag ID. Used to identify assets. 39
School of Electrical, Electronic and Computer Engineering
d_readerid
d_dataloggerid
d_datetime_seen
d_datetime_store
Unique reader ID Used to identify reader; determine location. Unique data logger ID. Used to identify data logger; relation to readers. Date and time the data was acquired by the data logger or reader. Date and time the data was acquired by the server control. The delay between the date seen and date stored determines data accuracy. Unique ID to identify the reader. Reader hostname used in DNS lookups. Reader IP-address. Reader location in mine grid. Data logger responsible for the data buffer of the reader. Reader communication port. Indicates the route level in a hierarchy format. Unique data logger ID to identify the data logger. Data logger hostname used in DNS lookups. Data logger IP-address. Real location of the data logger to indicate the data logger‟s position relative to the mine‟s levels.
Reader
r_id r_devname r_ip r_location r_dataloggerid r_port r_route_level
Datalogger
d_id d_pcname d_ipaddr d_location
4.1.4 Data interpretation Data interpretation is done by the Server Display. Server Display uses buffers and data from the server‟s MySQL database to interpret the data. T his will be discussed in 4.4. The data is interpreted so that the user can understand the data generated by the system. MySQL‟s .Net connector is used in this module [20].
School of Electrical, Electronic and Computer Engineering
40
Figure 35 : ServerDisplay version 3 4.2 MINE SIMULATOR 4.2.1 Versions 4.2.1.1 Version 1.0
This was the first functioning application. An asset could move around the mine with the first version of the route algorithm and was using the first version of the reader‟s propagation model . The route algorithm was based on a pre-recorded route being saved in a file. The route was loaded into the memory and started at the first entry and moved through the entries until it reached the end. When the end was reach the asset has moved through the mine. The reader‟s propagation model was based on an on/off method. If the asset was within the reader‟s range when the asset‟s tag transmitted its data, the asset was identified. 4.2.1.2 Version 1.1
This version introduced a few new features namely statistical collection, simulation pausing and version 2 of the reader propagation model. The propagation model in this version was based on linear blind spots. There were circles that represented the detection areas. When the asset was in the range of the reader, the asset‟s tag transmitted its data and if the asset was not in the
School of Electrical, Electronic and Computer Engineering
41
detection circle; the asset was in the blind spot. In reality the blind spot was not linear, but for practical purposes it was sufficient to simplify the model to reduce programming complexity. <reader propagation screenshot> The statistical collection is to review simulation results and will be discussed in 0. The thread required to save the data was a challenge, due to object synchronization. The movement calculation could be using an object when the logging thread tried to use it. This will be discussed in 0. When the user paused the simulation, the people and assets stopped moving in the mine. This would allow the user to review their positions. 4.2.1.3 Version 1.2
This version was the first version to function in the RFID RTLS. This version introduced socket communication, adjustable simulation parameters, tag collision detection, version 2 of the route algorithm and a more efficient threading usage. The socket communication was introduced to notify the reader simulator of successful asset identification. When an asset was successfully identified by evaluating the reader propagation model, the reader simulator was notified by using the protocol in 0. The adjustable parameters were an important part of creating an adjustable simulation. These parameters are discussed in 0. Tag collision is a real problem RF-devices faces. Modelling this into the software was an important part of successfully creating a simulator. Version 2 of the route algorithm was a new design and replaced version 1. Version 2 will be discussed in 4.2.2.1. Efficient threading usage will be discussed in 0. 4.2.1.4 Version 2
This is the final version and introduced version 3 of the reader propagation model. The new version of the propagation model was obtained from a validation in an active mine and will be discussed in detail in chapter 5. The new model also used packet loss as an indication of the reader circle‟s probability to detect an asset. The packet loss count was added to the statistical collection. Version 3 of the reader propagation model will be discussed in 4.2.2.3.
School of Electrical, Electronic and Computer Engineering
42
4.2.2 Features 4.2.2.1 Route algorithm First, pre-
Moving assets around in a mine requires an algorithm to calculate the route.
calculated routes were used, which resulted in extra time spent on calculating the routes and programming difficulties in determining the unique reader reads/seen statistic. The new algorithm involved recording POI‟s to determine the level of a specific route. Specific terminology involved in the route calculation is discussed in 4.1.1.1. Before a person or asset enters a mine, a route number is pseudo-randomly chosen. Each route number has a route file linked to it. The route file contains the route levels linked to a mine grid point. Thus after a route has been chosen, the movement calculation can identify a mine grid point to navigate an asset to. When the movement calculation has navigated the asset to a next level POI, the following POI is chosen using the current route information. If the end of route has been reached, the route is reversed; forcing the person or asset to go back out of the mine along the same track. 4.2.2.2 Asset movement calculation
In the previous paragraph the movement calculation‟s navigation input was discussed, namely the routing algorithm. The movement calculation routine knows the current position of the asset, the next POI position and the movement speed. The movement speed is defined in terms of a minimum and a maximum speed. The following will description explains the movement vectors as can be seen in Figure 36. The current position of the asset is (X1,Y1). The destination, the next POI, a person or asset wants to move to is (X2,Y2). P1 is the angle between the asset and its next POI, thus .
P1 has a pseudo-random angle adjustment of ±15º, resulting in the new angle P2, thus with . K is the magnitude of the assets movement. K is
pseudo-randomly generated using the minimum and maximum movement speed. Finally the new position (X3,Y3) is calculated. Thus . and
School of Electrical, Electronic and Computer Engineering
43
Figure 36 : Movement calculation 4.2.2.3 Reader propagation model
The final propagation model is based on linear packet loss. The model can be seen in Figure 37. The different colours indicate a packet error rate. The colour was calculated using red, green and blue (RGB). The formulas is: and ,
. This will result in a red colour for a packet loss of 90% and a green colour for a
packet loss of 0%. The specifics of the packet error will be discussed in chapter 5.
Figure 37 : Final reader propagation model The distance from the reader is calculated by getting the distance from each reader on the mine layout and determining if the distance is within the largest reader circle. The circles in the model indicate an identifiable area inside which a tag can be read. When the asset falls inside these circles, the position of the asset will be evaluated to determine if the asset is identified. Evaluation is done by detecting the packet error rate, checking the blind
School of Electrical, Electronic and Computer Engineering
44
spots and collision detection. The packet error rate is loaded from file. The blind spots are calculated by gaps in the identifiable circles. The collision detection is a longer process. Tag collision detection involved capturing al the tag pings at a calculation interval. When all the assets‟ movements were calculated and their tag pings recorded, it was verified for tag collisions. A tag collision will occur when two or more tags tried to communicate with s reader inside the same time interval. Thus, when two or more tags transmit inside the same time interval, they will have a collision and none of them will be successfully identified. The tag collision count is a key indicator of tag read failure. 4.2.2.4 Statistical collection
The statistical collection is used to review the simulation results. Each asset and reader hold a summary of the identification statistics. After a simulation, the user can see which reader has been the least efficient and can review each asset‟s movement and efficiency data. Data collection required a thread due to the file I/O latencies caused by multiple file writes per asset movement calculation. At first, the application was very slow and the latencies made the simulation inaccurate due to assets skipping important areas in the reader‟s propagation model. Efficiency data included: the amount of transmissions, unique readers met, unique readers seen, tag collision, blind spots and packet errors. 4.2.2.5 Data logging
Data logging in any software application is a handy debugging tool as it can also be used to find bugs in software. In most of the software developed for this project, the logging was done by using buffers and a logging thread. The thread operates continuously throughout program execution. The thread writes the buffer to a file and then clears the buffer. After clearing the buffer, the thread goes in a sleep state for 100ms and repeats the operation. 4.2.2.6 Adjustable simulation parameters
The adjustable parameters allow the user to change simulation parameters. The parameters that can be changed are movement speed, grid dimensions, tag transmission interval, save file location and communication details. Changing these settings will cause the simulation to restart with the new settings. Changing simulation parameters creates an adjustable simulation process. The parameters were saved using Microsoft.Net. 4.2.2.7 Efficient memory and processing power usage [21] [22]
Efficient processing power and memory usage become a key factor when a significant amount of calculation has to be done. Reallocation of memory and poor usage of processing power
School of Electrical, Electronic and Computer Engineering
45
slows the simulator down. Memory reallocation was reduced by creating more global objects and reusing them instead of creating new objects all the time. To use the processing power more efficiently, a lot of time was spent to analyse and optimize program flow. Another bottleneck was threading synchronisation. When threads use the same object and have different execution speeds, they will try to access an object simultaneously. This was corrected by using a synchronization lock. The object at hand was securely put on hold for exclusive use by the thread and released upon completion. This created an insignificant latency. 4.2.2.8 Reader simulator communication handler
The reader simulator‟s communications protocol is discussed in 0. The communication is based on TCP sockets. The communication handler is a thread that operates independently of the other tasks. The thread starts and checks if there is any data in the communications buffer. If there is not data, the thread goes in a sleep state for 100ms. If there is data, the handler creates a socket communication channel to the reader simulator and sends the data to the reader. After the data is sent the communication channel is closed and the handler continues its operation as normal.
School of Electrical, Electronic and Computer Engineering
46
4.2.3 Functional flow
Figure 38 : Simplification of the functional flow of MineSimulator 4.2.3.1 Setup
MineSimulator starts by creating a MineDrawing object. This is the class created to handle the simulation. A class was used to simplify the process of creating an API from this simulator. When the MineDrawing object is created, it initiates a process to start the simulation setup. Firstly, it loads the settings from the saved files and creates the necessary objects in the memory. Thereafter it creates an image of the mine layout without people or assets. This image will be used throughout the simulation process as the people‟s or assets‟ movement and reader‟s changes will be drawn on this canvas.
School of Electrical, Electronic and Computer Engineering
47
4.2.3.2
Threading
After setup, three threads are started, namely logging, moving, and communication threads. These threads are processed continuously until the MineDrawing object is destroyed. The logging thread was discussed in 0. The moving thread cycles through the currently active people or assets and evaluates the reader propagation model as discussed in 4.2.2.3. When the model has been parsed and the successful identifications (tag reads) have been transmitted, a new image is created by simply drawing the result onto the mine image created in setup. If the person or asset has been successfully identified, the appropriate reader circle is coloured in yellow to visually illustrate a successful read. This can be seen in Figure 39.
Figure 39 : Visual identification indication The communication thread is discussed in 0. 4.2.3.3 Exiting
When the application exits the statistics are written to file and the threads are stopped. When stopping the treads the application must wait for a thread to close. Threads will close when a termination flag is sent to them. Data can be lost by forcing a thread to close, but this is the last resort when the application closes. 4.3 READER SIMULATOR 4.3.1 Versions 4.3.1.1 Version 1
This first version of the application was created to be a standalone application in the sense that one ReaderSimulator will run on PC. This application was also created when there was a total different system design, before the design of MineSimulator. The first system design had a reader manager, which was used to manually move assets from reader to reader. When the data logger connected to the ReaderSimulator, there was a TCP socket connection made from
School of Electrical, Electronic and Computer Engineering
48
the ReaderSimulator to the reader manager. The reader manager would then report the assets at that specific reader. The connection to the reader manager was terminated and the ReaderSimulator reported to the data logger of the assets near itself. So the reader manager was used as a tool to test the data logger and server control software. 4.3.1.2 Version 2
The next version of ReaderSimulator was created to eliminate the need for more hardware. Thus threading was used in the application to function as multiple readers on one PC. This proved to be very useful when demonstrating a mine layout with limited hardware. This version did not used the threading model very well and resulted in an application not in control of the threads that are running. So the application started and worked without any user control. When the application was exited the threads was forced to close. 4.3.1.3 Version 2.1
This version was the first version to eliminate the reader manager. The new version made use of an internal buffer to store data received by using the ReaderSimulator communication protocol. This protocol will be discussed in 0. This version also made use of a class to manage threading. This will be discussed in 4.3.1.5. 4.3.1.4 Version 2.2
Version 2.1 had one major disadvantage. Every transaction of all the readers was displayed in one GUI component; the implication of this is that due to threading synchronization it slowed down the application. The queue of synchronization request increased with time and after a few minutes the application slowed down the whole system. This caused a huge latency in real time and data interpretation - the latency was about a 5 minutes. This was reduced by only allowing the reader threads to report when data was received and not reporting of every transaction. This reduced the system latency down to 30 seconds. 4.3.1.5 Threading
A class is used to manage the threading and reporting of the readers. An array of this class is created to initialize more than one reader and controlling it at the same time. This proved a success. The class could be used multiple times and still be controlled from the parent application. The class uses a buffer to store data (discussed in 4.3.1.6) and has a communication protocol that will be discussed in 0. The advantage of using the class in an array was that when the application exited, the threads could be stopped safely.
School of Electrical, Electronic and Computer Engineering
49
4.3.1.6
Buffer
The buffer was created to store data received from MineSimulator. The buffer was created using a simple string array structure. This buffer is easily expandable. The buffer does not use threading synchronization due to the fact that the thread is the only one able to use the buffer. The buffer is never erased, so it can cause a memory leak when used for an extended time period. Instead of clearing the buffer, each data entry has an active flag. When the flag is set, the data is still perceived as tracking data at that reader. If the flag is not set, the data is ignored. The memory leak was not fixed due to time constrains. 4.3.1.7 Information display
Information display, in an application using multiple threads, is a problem. The balance between information data and application speed must be struck. The final version displays debugging data which shows that an application has started and that the reader is online. The only information displayed from the reader thread is when the reader receives data. Previously incoming connection details, protocol details and data transactions were displayed. Everything display in the GUI is also written to file using the same architecture as the MineSimulator‟s data logging thread. 4.3.1.8 ReaderSimulator communication protocol
This protocol was created to simplify and standardize the data interactions of the software modules in the system. The protocol is visualized in Figure 40and Figure 41. The fist command received when handling incoming connections determines the flow of data. If the SET command is received the data will flow into the ReaderSimulator‟s buffer. When the GET command is received the data will flow outwards to the application that made the connection. The SET command will result in the ReaderSimulator‟s communication thread waiting for incoming data and storing it in the buffer. The GET command will evaluate the ReaderSimulator‟s communication thread‟s buffer for active data entries. These entries will be combined and sent back to the application that made the connection. The reader simulator receives data inputs/requests on a port saved in a database. The ports start at 6000. The functional flow of the reader simulator is simplified in Figure 42. The ports can be calculated by reading the mine layout save file and incrementing the port number by one. The port numbers is saved in a database for use by the data logger and the server control/display.
School of Electrical, Electronic and Computer Engineering
50
Figure 40 : SET tag data
Figure 41 : GET tag data
School of Electrical, Electronic and Computer Engineering
51
4.3.2 Functional flow
Figure 42 : Reader Simulator flow
4.3.2.1
Setup
ReaderSimulator starts by reading configuration data from a file. This configuration contains the number of readers and determines the type of information to log. 4.3.2.2 Threading
After setup, all the threads can be started. These threads were discussed in 4.3.1.5. The operation of the thread is simplified in Figure 42. 4.3.2.3 Exiting
When the application is closed, each reader thread is safely closed as well.
School of Electrical, Electronic and Computer Engineering
52
4.4 SERVER DISPLAY 4.4.1 Versions 4.4.1.1 Version 1
The first reader manager was used as discussed in 4.3.1.1. The basic working of this model was to translate data received from the data logger. This is the main function of the ServerDisplay, but in this version the display was not visually attractive. This can be seen in Figure 43.
Figure 43 : ServerDisplay version 1 The application would gather data from the MySQL database and get the unique data entries. These entries would then be written into pre-defined spaces. This restricted the application‟s scalability. In this version of ServerDisplay version 1 of the data gathering SQL statement was used. This version seems to work in small amount of data, but was later discovered not to be reliable. 4.4.1.2 Version 2
This version eliminated the reader manager. The data display was done by executing version 2 of the gathering SQL statement of the server. The data returned was the unique location of every asset in terms of the data logger and reader placement. This was a more logical display than visually informative display. 4.4.1.3 Version 3
The final ServerDisplay provides a visual simulation of the mine to display the location of assets. The solution is to display the same mine layout as the MineSimulator. Now the locations
School of Electrical, Electronic and Computer Engineering
53
of the assets are known in terms of readers since the locations of the readers are known. Thus, one can simply display the current assets at a reader in order to determine their location. A display and locating model was developed and is discussed in 4.4.2.1. 4.4.2 Features 4.4.2.1 Display method
The display method was based on the same method used for drawing the MineSimulator‟s display, but without people or assets moving around the mine. This model replaced the reader‟s propagation circles with the amount of assets located near the reader or moving around the reader. The display received the data from the locating algorithm. The display can be seen in Figure 44.
Figure 44 : ServerDisplay In the Figure 44 it can be seen that two types of reader information are displayed. The first is the forward/backward display. This is readers that have other readers after them in a logical sense of flow. Using sector based location the movement can only be indicated when an asset moves from one reader to another. The second information type is the people or asset count. This is when there is no reader after the last reader in a route (at the end of a route). When this happens, one cannot determine if an asset moving past a reader is moving forward of backward in the mine. 4.4.2.2 Locating algorithm
The locating algorithm determines where an asset is currently placed and the asset‟s direction. Thus the locating algorithm receives the last known location of each asset continuously and calculates the movement dynamics.
School of Electrical, Electronic and Computer Engineering
54
This algorithm uses a buffer for every reader. If data is received, it is defined accordingly as forward / backward / count. The algorithm can be seen in Figure 45. The algorithm works by adding the assets to the reader‟s buffer and removing them when they move to a different reader. This creates a history of the asset‟s movement through the mine. This history is used to determine the movement direction.
Figure 45 : Locating algorithm
School of Electrical, Electronic and Computer Engineering
55
4.4.3 Functional flow The following section provides the functional flow of the simulator application.
Figure 46 : ServerDisplay version 3 4.4.3.1 Setup
The application starts and loads the required settings for creating an image. Thereafter, the background image is created that will be used to draw the movement information through the simulation process. Part of the setup is to determine all readers‟ configuration information that handles the display method and location algorithms. 4.4.3.2 Data flow
The data flow in the functional flow is indicated in Figure 46. This is to show how the object handling the display can retrieve data and continuously retrieve updated data. Data is sent to configure the readers at first. Thereafter the data is updated by a timer on the parent application. 4.4.3.3 Updating timer
The timer on the parent application is use to query the database with the SQL statement in Error! Reference source not found.. After the data has been received, it is sent to the object andling the display drawing. The timer executes every 100ms (chosen from empirical testing).
School of Electrical, Electronic and Computer Engineering
56
4.4.3.4
Image creation
The image creation works the same as the image creation of MineSimulator. The extra information that is added to the image was discussed in 4.4.2.1. This concludes chapter 4. The aim of this chapter was to give a system overview and discuss the sub-system components in detail.
School of Electrical, Electronic and Computer Engineering
57
CHAPTER 5 – VALIDATION
This chapter provides validation data for the design used in the reader propagation model. An overview of the validation process will be given and the validation will be briefly discussed. The model was used to determine the propagation characteristics of RF inside the haulage of an underground mine.
School of Electrical, Electronic and Computer Engineering
58
5.1 OVERVIEW The reader propagation model used in MineSimulator was validated in an active underground mine. The details of the mine cannot be discussed due to mutual non-disclosure agreement. The validation was done 3km underground. Test was done to determine a propagation model for an active RFID reader. Test generated data was compiled into one piecewise linear model - this model was used in MineSimulator. 5.2 ENVIRONMENT The humidity underground was high and there was water present near the testing area. There were also pipes transporting water and air. The layout of the test is shown in Figure 47. The width of the haulage was approximate 4 meters with a height of 5 meters. In section 4 there was an electric motor located at about 17 meters. There were no other wireless devices present near the testing area.
Section1
Motor
Section4 Section3 Reader
Section2
Figure 47 : Test layout
School of Electrical, Electronic and Computer Engineering
59
Figure 48 : Section 1
Figure 49 : Section 1 (Left) and section 4 (Right)
School of Electrical, Electronic and Computer Engineering
60
Figure 50 : Section 3
Figure 51 : Section 2 (Right) and section 4 (Left)
Figure 52 : RFID Tag placement
School of Electrical, Electronic and Computer Engineering
61
5.3 HARDWARE USED The hardware used for testing was sponsored by Becker Electronics. A portable active RFID reader and active RFID tags were used. The reader that was used is a HTRD100ZA as can be seen in Figure 53. This reader was designed to read any Backer family RFID tag. This reader displays the RSSI reading it has calculated when a transmission from a RFID tag had been successful.
Figure 53 : Backer Electronics HTRD100ZA The RFID tags used are TPER200WW as can be seen in Figure 54. This tag pseudo-randomly transmits its identification number in a window of 2 to 3 seconds.
Figure 54 : Becker Electronics TPER200WW 5.4 TESTS In Figure 47 the tests were carried out by measuring RSSI values in sections 1, 2, 3 and 4. In the middle of the haulage the measurements were spaced at 0.5 meters and all remaining measurements were made in 1 meter steps. The RSSI values were measured and the packet success rate was indicated by listening how fast the packets were received. A value of 5 was chosen for an excellent packet success rate,
School of Electrical, Electronic and Computer Engineering
62
while a value of 1 was chosen for a weak packet success rate – that is, a high packet error rate resulted in a value of 1 and a low packet error rate in a value of 5. The tag placement and environment details can be seen in 5.2. The tests took about 3 hours and the data was recorded manually. Three tests were done in each section. Reception along the centre and sides of the haulage was characterized. The measurement results are shown in the following section. 5.5 RESULTS 5.5.1 Section 1
60.00 50.00 40.00 RSSI 6 5 4
30.00
20.00 10.00 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
3
2 1 0
Distance (Meters)
Center Center Packet Success Left Left Packet Success Right Right Packet Success
School of Electrical, Electronic and Computer Engineering
63
5.5.2 Section 2
60.00 50.00 40.00 RSSI 30.00 20.00 10.00 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Distance (Meters) Center Center Packet Success Left Left Packet Success Right Right Packet Success 6 5 4 3 2 1 0
5.5.3 Section 3
60.00 6
50.00
5
40.00
RSSI
4
30.00
3
20.00
2
10.00
1
0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 Distance (Meters) Center Center Packet Success Left Left Packet Success Right Right Packet Success
0
School of Electrical, Electronic and Computer Engineering
64
5.5.4 Section 4
60.00 6
50.00
5
40.00 RSSI
4
30.00
3
20.00
2
10.00
1
0.00 0 2 4 6 Distance (Meters) Center Center Packet Success Left Left Packet Success Right Right Packet Success 8 10 12
0
5.5.5 Global average
50.00 45.00 40.00 5.00 6.00
35.00
30.00 RSSI 25.00 20.00 15.00 10.00 5.00 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Distance (Meters) RSSI Packet success Packet success used in MineSimulator
4.00
3.00
2.00
1.00
0.00
School of Electrical, Electronic and Computer Engineering
65
5.6 DISCUSSION The characterization tests provided a successful model that could be used in MineSimulator. The RF effects discussed in 0 produced the desired effect that summarized propagation effects such as multipath fading. This can be witnessed in 5.5.2 at the edge of reception where significant fading is evident - refer to the variation in signal strength at the fringe. This was due to the multipath fading from reflections against the haulage walls. To improve reception in the haulage, one may do the following: Spatial diversity – use more than one antenna; Polarisation – use cross-polarised or circularly polarised antennas; Frequency – use more than one frequency.
This concludes chapter 5. The aim of this chapter was to validate the reader propagation model from measurements taken in an underground mine. The model appears to be piecewise linear and was used accordingly in all simulations.
School of Electrical, Electronic and Computer Engineering
66
CHAPTER 6 – IMPLEMENTATION AND EVALUATION
This chapter presents the implementation layout (architecture) of the physical design from chapter 4 and addresses the evaluation of the test layout.
School of Electrical, Electronic and Computer Engineering
67
6.1 IMPLEMENTATION
6.1.1 PC1 PC1 was used to run two simulation programs from an IP-address of 192.168.0.3. There was a two-screen setup with each simulation program running on each screen. This resulted in an aesthetically acceptable appearance. 6.1.2 PC2 PC2 was used to run three layers of computing. The first layer was the PC‟s main operating system. Linux was used in this demonstration. The Linux operating system was used to create the environment for the data logger. Thus there was a MySQL [23] service running with the data logger database setup and a PHP [24] execution environment. The second and third layers were created using virtual computing environments. This was done by using VirtualBox [25]. The one virtual PC ran the ServerDisplay application and the other ran the ServerControl application. 6.1.3 PC3 PC3 was just to display the results of the PHP script that was executed on PC2‟s first layer. Putty [26] was used to accomplish this.
School of Electrical, Electronic and Computer Engineering
68
6.2 EVALUATION 6.2.1 MineSimulator tests The statistical collection of MineSimulator was to test and ensure functionality of the simulator. MineSimulator can function without any other application so the testing there had been a straight forward run-in test. The simulator was run for one day without any interruption. The tests revealed that MineSimulator was stable terms of its functional capability. 6.2.2 ReaderSimulator tests ReaderSimulator was a more testable application in the sense that most of the functionality was done in the background. An application was created to perform a fuzzy fire test on ReaderSimulator. This was done by sending 100 data items at once to the ReaderSimulator. This tested the stability of the ReaderSimulator when handling a large volume of data. 6.2.3 Data collection tester ServerDisplay, ServerControl and the data logger script could be tested as whole. This was done by using the test application created for the ReaderSimulator and recording the data flow performance through the data collection process. 6.3 SYSTEM TEST The system could only be tested as whole when the setup in 6.1 was done. The test required all modeles to run and an active simulation in MineSimulator. The verification of the system to be functioning correctly was done by visually inspecting ServerDisplay and comparing it to MineSimulator‟s assets movement. 6.3.1 Overview MineSimulator was run to analyse the movement of 52 assets. The locations of each asset were successfully traced and the statistics is show in 6.3.2 and 6.3.3. 6.3.2 Asset statistics
Person 0 1 2 3 4 5 6 7 8 9 Tag Tag Tag Tag Read Miss Collision Blindspot 79 36 3 0 72 37 6 0 77 39 4 0 78 44 6 0 81 51 5 0 113 60 2 0 77 43 4 0 121 64 3 0 79 43 5 0 80 47 7 0 Tag Readers Readers PacketError Met Read 33 4 4 31 4 4 35 4 4 38 4 4 46 4 4 58 6 6 39 4 4 61 6 6 38 4 4 40 4 4
School of Electrical, Electronic and Computer Engineering
69
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
114 74 74 107 119 76 108 79 77 123 106 121 73 118 76 110 120 71 75 107 81 119 110 78 72 76 111 122 79 109 118 72 77 79 73 78 107 80 116 107 77 110 70
60 36 42 58 54 45 51 42 42 68 51 64 42 70 41 57 67 33 47 64 47 73 61 42 43 49 63 64 46 52 68 32 44 41 39 47 55 48 65 65 46 59 37
5 6 5 3 4 5 6 5 7 8 4 7 5 3 8 8 5 7 5 2 9 6 5 3 5 12 3 4 5 6 11 6 12 6 3 8 4 7 5 12 9 6 3
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
55 30 37 55 50 40 45 37 35 60 47 57 37 67 33 49 62 26 42 62 38 67 56 39 38 37 60 60 41 46 57 26 32 35 36 39 51 41 60 53 37 53 34
6 4 4 6 6 4 6 4 4 6 6 6 4 6 4 6 6 4 4 6 4 6 6 4 4 4 6 6 4 6 6 4 4 4 4 4 6 4 6 6 4 6 4
6 4 4 6 6 4 6 4 4 6 6 6 4 6 4 6 6 4 4 6 4 6 6 4 4 4 6 6 4 6 6 4 4 4 4 4 6 4 6 6 4 6 4
School of Electrical, Electronic and Computer Engineering
70
6.3.3 Reader Statistics
Reader Tag Tag Tag Blind Tag Packet Tag Unique Unique transmissions Collisions Spots Errors Reads transmissions Reads in radius in radius (11:21) 2067 181 0 997 889 106 106 (51:21) 886 33 0 459 394 46 46 (83:40) 421 6 0 218 197 20 20 (46:4) 662 51 0 336 275 34 34 (21:41) 453 18 0 199 236 26 26 (57:38) 417 14 0 172 231 26 26
6.3.4 Information logging Only a part of the log is shown.
Date 08:36:55:744 08:36:55:754 08:36:55:944 08:36:56:174 08:36:56:184 08:36:56:214 08:37:08:392 08:37:08:592 08:37:08:792 08:37:08:993 08:37:09:193 08:37:09:393 08:37:09:594 08:37:09:794 08:37:09:994 08:37:10:194 08:37:10:405 08:37:10:645 08:37:10:845 08:37:11:046 08:37:11:296 08:37:11:496 08:37:11:717 08:37:11:937 08:37:18:116 Type Info Info Info Info Info Info Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel Personnel First Time Reader 08:37:18:116 RandomCheck2 08:37:18:116 RandomCheck1 08:37:18:116 PacketError Info MineDrawing Object created Starting Initialization Thread Loading done Creating GridImage done Starting Caculation Thread Stoping Initialization Thread Person 0 added. Route=2 Person 1 added. Route=1 Person 2 added. Route=2 Person 3 added. Route=1 Person 4 added. Route=2 Person 5 added. Route=3 Person 6 added. Route=1 Person 7 added. Route=4 Person 8 added. Route=1 Person 9 added. Route=2 Person 10 added. Route=3 Person 11 added. Route=1 Person 12 added. Route=2 Person 13 added. Route=3 Person 14 added. Route=4 Person 15 added. Route=1 Person 16 added. Route=3 Person 17 added. Route=2 Person 0 is the first time at reader(11:21) with distance 24.17183852705 m Rate=80 Packet Error With Person 0 @Reader(11:21) with distance 72.5155155811499
School of Electrical, Electronic and Computer Engineering
71
08:37:18:386 First Time Reader 08:37:18:386 RandomCheck2 08:37:18:386 RandomCheck1 08:37:18:386 PacketError 08:37:18:927 First Time Reader 08:37:18:927 RandomCheck2
Person 1 is the first time at reader(11:21) with distance 24.3732549232874 m Rate=80 Packet Error With Person 1 @Reader(11:21) with distance 73.1197647698623 Person 3 is the first time at reader(11:21) with distance 24.2131736411768 m Rate=80
6.3.5 Test results The test shows a safety rating of 100% and this is confirmed by the reader and asset statistics. The reader at with location (11:21) in 6.3.3 is the reader at the entrance and needs to identify the larges amount of data. This concludes chapter 6. The aim of this chapter was to discuss the implementation and evaluation of the design in chapter 4.
School of Electrical, Electronic and Computer Engineering
72
CHAPTER 7 - CONCLUSION AND RECOMMENDATIONS
In this chapter we provide an overview of the complete project and reflect on the success of the design. We make recommendations for use in similar projects (and future iterations of this project) and discuss the five ECSA Exit Level Outcomes that were addressed in the completion of this project.
School of Electrical, Electronic and Computer Engineering
73
7.1 PROJECT HISTORY - REVISITED A feasibility study of an RFID RTLS was required. After researching current technologies and systems, the need for a simulator design became evident. After designing the simulator, it was integrated into a simulation that could be verified by doing software functional testing with validation of the RF propagation model. All of the software modules were created using VisualBasic.Net [27] [28] – except one module that was created with a PHP script. These modules were all planned, designed and tested by the author. 7.2 CONCLUSION The systems engineering approach to software simulation is ideal for software system development. An active radio frequency identification real-time locating system was successfully planned, designed and verified in a simulated environment. Part of the simulation was also validated by a propagation model testing in an active underground mine. It is possible to use an application to determine the optimal design for an RFID RTLS. Using this application not can it only save costs but also maximize safety. Designs will not always be a success from the start; this is why the design is an iterative process. Analysis is done, the design is conceptualised and goes through a detail design process. Thereafter can the design be tested and, if needed, be adjusted. Software design is a speciality upon which needs to be focussed. Simply rushing through the correct design process will not be successful. The design process needs to be understood first and must be followed correctly. The engineering curriculum creates the knowledge to successfully complete this project. Software design is not limited to software itself and involved mathematics and the logical thought process learned in engineering. The simulator can be used to optimize and reduce cost when designing a RFID RTLS. This is very useful in a sense that a RTLS can be planned to achieve a certain resolution of location data; thus the balance between costs and detail information can be found while still insuring the safety of such a system. The statistical collection feature of MineSimulator proved to generate valuable data in planning a RTLS. Critical readers that have a high amount of traffic must be account for by insuring that the antenna quality of those readers is reliable. The validation in an active mine was necessary to create a simulator that is close to reality and to identify the problems of using an RFID RTLS.
School of Electrical, Electronic and Computer Engineering
74
7.3 RECOMMENDATIONS The systems engineering process needs to be understood before it is used – is it not as simplistic as it initially appears. Future developers will have to understand the process before attempting to adapt the simulator in this design project. The simulation revealed that the most critical reader is the reader at the entrance. Therefore, this readers needs to be reliable. General recommendations regarding to the software will be to used object orientation from the start, plan an application instead of improvising along the way, use backups and have a professional programming style. This was all done in this design project by experiencing the limitations of ad hoc development. Future recommendation for MineSimulator is to add more moveable objects that have an influence on the reader propagation model, such as locomotives and trains. More research can thus be done on the reader propagation model and scenarios - such as trapped miners - could be integrated into the simulation. 7.4 ECSA OUTCOMES 7.4.1 Problem solving In engineering, problem solving is done on a daily basis. This project started with a problem statement and was then analysed and solved. 7.4.2 Engineering Design Engineering design is a very unique philosophy that is tried and tested. This project implemented engineering design into a software application. 7.4.3 Investigations, experiments and data analysis Science requires investigation, experiment and analysis to prove a theory - engineering is applied science. This project had a problem statement which was investigated, analysed, implemented and validated. 7.4.4 Professional and technical communication Engineering communication is an unavoidable function and must be done to insure that the design meets al the client‟s needs. This project involved negotiating with hardware manufactures and mine management and writing this report.
School of Electrical, Electronic and Computer Engineering
75
7.4.5 Independent learning ability Engineering is obsolete when no one learns from the process. Much has been learnt from this project (particularly about engineering method and programming) and will most certainly be used in future designs.
School of Electrical, Electronic and Computer Engineering
76
References
[1] B.S. Blanchard and W.J. Fabrycky, Systems engineering and analysis, 4th ed.: Peason Education, 2006. [2] H Steyn et al., Project Management - A Multi-Disciplinary Approach, 2nd ed. Pretoria, South Africa: FPM Publishing, 2009. [3] Engineering Council of South Africa. (2004, July) Whole Qualification Standard for Bachelor of Science in Engineering (BSc(Eng))/Bachelors of Engineering (BEng): NQF Level 7. http://www.ecsa.co.za/index.asp?x=documents&y=guidelines. [4] Ajay Malik, RTLS For Dummies.: Wiley, 2009. [5] David M Pozar, Microwave Engineering, 3rd ed.: Wiley, 2005. [6] Charles L Hutchinson and Larry D Wolfgang, The ARRL handbook for radio amateurs, 68th ed.: American Radio Relay League, 1990. [7] R.D. Straw, The ARRL antenna book, 21st ed.: American Radio Relay League, 2003. [8] S.R. SAUNDERS and A ARAGO N-ZAVALA, ANTENNAS AND PROPAGATION FOR WIRELESS COMMUNICATION SYSTEMS, 2nd ed. West Sussex, England: John Wiley & Sons, 2007. [9] I.F. Akyildiz, Z Sun, and M.C. Vuran, "Signal propagation techniques for wireless underground," Physical Communication, vol. 2, pp. 167-183, 2009. [10 D Dobkin, The RF in RFID.: Elsevier, 2008. ] [11 Salter W. Daneels A. (1999, December) What is SCADA? PDF. ] [12 Savi Technologies. (2010, March) Savi RFID Hardware. [Online].
] http://www.savi.com/media-form.php?d=ST-602_DS_09-09a.pdf&n=Savi%20ST602%20Asset%20Tag [13 Wavetrend. (2010, March) Wavetrend Products Tags. [Online].
] http://www.wavetrend.net/UserFiles/Files/TG100-A-PI-02-LT.pdf [14 Ekahau. (2010, March) Ekahau RTLS. [Online].
] http://www.ekahau.com/images/stories/documents/ekahau_t301a_datasheet.pdf [15 Savi Technologies. (2010, March) Savi Active RFID Readers. [Online]. 77
School of Electrical, Electronic and Computer Engineering
] http://www.savi.com/media-form.php?d=SR-650_DS_09-09a.pdf&n=Savi%20SR650%20Fixed%20Reader [16 Becker Electronics. (2010, September) Active Tag Reader Datasheet UATR400ZAA. Email. ] [17 Wavetrend. (2010, March) Wavetrend Products Readers. [Online].
] http://www.wavetrend.net/UserFiles/Files/L-RX201-04-PI.pdf [18 Cybertrack. (2010, March) RFID Insitude S.A. [Online].
] http://www.rfidtec.co.za/backend/media/11132007111709/Locatxrealtimelocatingsystems.p df [19 Mine Site Technologies. (2010, March) Mine Site Technologies. [Online].
] http://www.minesite.com.au/ [20 (2010, October) MySQL.Net Connector. [Online].
] http://dev.mysql.com/downloads/connector/net/ [21 Roberto M. Amadioa and Silvano Dal Ziliob, "Resource control for synchronous cooperative ] threads," Theoretical Computer Science, vol. 358, no. 2, pp. 229-254, August 2006. [22 Gyu Sang Choi and Chita R. Das, "A Superscalar software architecture model for Multi] Core Processors," The Journal of Systems and Software, vol. 83, no. 10, pp. 1823-1837, October 2010. [23 (2010, October) MySQL database software. [Online]. http://www.mysql.com/ ] [24 (2010, October) PHP scripting language. [Online]. http://www.php.net/ ] [25 (2010, October) VirtualBox. [Online]. http://www.virtualbox.org/ ] [26 (2010, October) putty ssh client. [Online].
] http://www.chiark.greenend.org.uk/~sgtatham/putty/ [27 (2010, October) Microsoft.Net. [Online]. http://www.microsoft.com/net/ ] [28 (2010, October) Microsoft VisualBasic. [Online]. http://msdn.microsoft.com/en-
] us/vbasic/default.aspx
School of Electrical, Electronic and Computer Engineering
78
[29 Ieiworld.
(2010,
March)
Ieiworld
-
WAFER-945GSE.
[Online].
] http://www.ieiworld.com/files/product/08274643094896902127/catalog/en-US/WAFER945GSE_20090715.pdf [30 Beckhoff. (2010, March) Beckhoff Industrial PC. [Online].
] http://download.beckhoff.com/download/Document/Catalog/Main_Catalog/english/separatepages/Industrial_PC/C6915.pdf [31 E Manson et al., "Radio wave propagation in arch-shaped tunnels: Measurements," ] Comptes Rendus Physique, vol. IV, no. 11, pp. 44-53, January 2010. [32 S.R. Saunders and A Arafo N-zavala, Antennas and propagation for wireless ] communication systems, 2nd ed. West Sussex, England: John Wiley & Sons, 2007.
School of Electrical, Electronic and Computer Engineering
79
APPENDIX A – TURN IT IN REPORT
School of Electrical, Electronic and Computer Engineering
80
School of Electrical, Electronic and Computer Engineering
81
