Wireless Sensor Networks
Content
Anthony F. Hillen
Wireless Sensor Networks
Technology Overview The individual nodes that constitute a wireless sensor network are generally small in size and use power-efficient batteries to extend their operational longevity. Depending on its function, each node has a sensor board that facilitates the detection and measurement of heat, vibrations, air-pressure and magnetic fields (among other things). The “motes” developed at UC Berkeley are a typical example of such devices. Motes have a range of about 100 feet and feature a 7Mhz processor, 4Kb of RAM, 128Kb of programmable memory space, and utilize a ChipCon CC1000 radio for communication. Due to their deployment simplicity and low cost of about $200 per unit, motes can be distributed in spatially dense configurations within a given area. Motes make use of Tiny OS, an operating system designed from scratch to be as power-efficient as possible. Using less than half the capacity of an AA battery, Tiny OS can effectively run applications for months at a time. (Hellerstein, Hong, Madden, 2003) Motes within a given geographic location use networking software to selfassemble into ad-hoc networks, allowing data to be transferred to and from any node in its network, or if necessary, to a proxy (but unauthorized, non-peer/client) in close proximity (like a random cell-phone or laptop), thereby serving as a conduit to a wider network (like the internet). The nodes in wireless sensor networks can be employed to capture data about their geographic environment while seamlessly and instantly communicating that information with surrounding nodes, impervious to temporal or spatial limitation. Wireless sensor networks circumvent the hindrances of collecting information from
-1-
Anthony F. Hillen geographic locations otherwise inaccessible by human beings; from the nether ocean to enemy occupied territory. (Kumar, 2003) Sensor networks are amenable to both civilian and military deployment. In civilian scenarios, sensors used to monitor traffic, pollution, or infrastructure can be positioned by hand. In terms of the most basic military applications, such networks can be used to detect, classify, and track targets in a given territory (other applications will be discussed later). Civilian use of wireless sensor networks range from environmental purposes such as pollution and ecosystem analysis to law-enforcement activity like traffic monitoring and criminal surveillance. In their military context, discommodious or threatrich environments can be accurately and safely reconnoitered, determining sensor placement a priori is unnecessary as random and widespread sensor deployment can be achieved via aircraft. (Clouqueur, Veradej, Ramanathan, Saluja, 2003) Development Status Sensor technology has made substantial advancements thanks to innovative new research efforts. Some recent developments have been academic in nature, like tracking and monitoring animal migrations, bird habitats, or vineyards, while private-sector developments have included efficiency improvements like “condition-based” equipment maintenance. There are numerous examples of how wireless sensor networks are currently being used, for instance, biologists at UC Berkeley interested in studying how trees affect the temperature and humidity in their surrounding canopy use a network of trunk-attached motes to monitor the microclimates around the redwood trees in their botanical garden. (Hellerstein, Hong, Madden, 2003)
-2-
Anthony F. Hillen One of the most promising research endeavors currently underway is the development of a flexible and interference-resistant communication technology. Instead of being restricted to transmitting and receiving information on a pre-assigned block of spectrum, these radio devices would utilize “opportunistic spectrum access”. Such systems would facilitate faster and more efficient communication since static allotment would be complemented by instantaneous and opportunistic spectrum access. Sensor nodes utilizing such technology would access unused spectrum, detect, authorize and network surrounding nodes in a manner that reduces inter-node communication interference. (DARPA “neXt Generation” program) A second area of research worthy of mentioning employs “mobile swarm” sensor networks to facilitate asset management and multimedia streaming. Mobile swarms are clusters of sensor nodes located in close physical proximity to each other and possess similar mobility patterns. For example, a group of tanks or UAVs could constitute a swarm, presumably equipped with qualitatively superior sensors like hi-res cameras, and longer range radios with higher channel bandwidths than conventional motes. Sensor nodes attached to the swarm members can gather information about that individual member, like location or operating status, but it can also relay data captured by its “host” to other nodes in the swarm, other mobile swarms, or to a command center through a backbone network or satellite. (Gerla, Xu) There are three primary motivations supporting research and development in the field of wireless sensor networks: academic interest, corporate profit, civil value, and of course, military application. These strong and mutually supportive driving forces suggest a promising future for the technology. Although motes currently cost about $200 per unit,
-3-
Anthony F. Hillen prices have been dropping steadily and are expected to continue falling. Some projections suggest that the price will fall to around $10 a piece within the next few years and that the units themselves will shrink in size to about 2 cubic mm. It is safe to assume that the smaller and cheaper these sensors get, the more widely used they will become. Moore’s law indicates that in about ten years, devices as small as a mote will have processing and memory capabilities similar to a contemporary network server. (Hellerstein, Hong, Madden, 2003) Several challenges faced by sensor technology are worthy of closer scrutiny. Software development, for instance, has been particularly troublesome. This is primarily due to the sensor’s hardware limitations. Modern sensors like motes suffer from a dearth in processing speed, memory, radio bandwidth, and energy capacity. The problems with processing speed and memory are likely to be resolved in the near future. However, the shortage in bandwidth is due to insufficient energy, and because the energy density of commercial batteries has not changed much in the last ten years, it is unlikely that the challenges posed by battery capacity and radio bandwidth will be overcome anytime soon. Other problems involve developing a way of programming groups of sensors to undertake a variety of different tasks and creating reliable security protocols to ensure network integrity and guard against intrusion and denial-of-service threats. (Hellerstein, Hong, Madden, 2003) Although most challenges are developmental, the technology’s inherent potential to violate widely held standards of personal privacy implies that there are also social and legal obstacles to many of its civil applications. Critics are quick to point out the ways such technology can be misused, from tracking ones every movement to remotely
-4-
Anthony F. Hillen accessing personal information. It is not difficult to imagine scenarios whereby one need merely walk by the wrong person in a grocery store to share one’s home address, credit card number, or other identity information. These are valid concerns that warrant significant discussion and policy-formulation prior to any civil application. Security Implications In contrast to civilian applications, the military applications of wireless sensor networks must be fully understood, embraced, and implemented without delay. Until now, the United States’ military use of sensor technology has been limited to basic and relatively crude detectors that utilize sensor technology, the Unattended Ground Sensor (UGS) system and the AN/GSQ-187 Remote Battlefield Sensor System (REMBASS) are typical examples of such devices. Although technically “wireless” by definition (in the sense that they do not require external cables to function), these devices do not utilize the technology discussed in this report, nor do they form ad-hoc networks of any kind. These systems are capable of detecting vehicle and personnel activity, but would be incapable of providing potentially critical battlefield information in the form of real-time audio/video data. Wireless sensor networks can also provide a strategic advantage in urban and close-quarter combat situations. For instance, an orbiting UAV could automatically detect friendly forces in the area and transmit aerial reconnaissance data directly to a heads-up display build into the helmets of troops on the ground. If a ground unit required a topographical map of an area, it could transmit the request to a nearby tank or UAV which would then acquire the information from a satellite or databank at a command center.
-5-
Anthony F. Hillen Military integration of this technology will require simultaneous training and tactical adaptation. The ability of its operators to function effectively in an informationrich environment will ultimately depend on the quality of their training. US combat statistics from the first Gulf War indicate that an abundance of battlefield information actually degrades combat performance. Its recipients could not process all the information at every level of command so units in critical need of information had to sift through too much irrelevant data to locate the specific details they required. The confusion caused by information overload illustrates the importance of implementing training and tactical reform measures whenever new technology is introduced. (Davis, 2007) It should be noted that military use of wireless sensor networks need not be limited to information awareness purposes. While not the most creative of individuals, I can think of a few applications omitted from existing literature on the subject. First, integrating wireless sensor networking technology with anti-tank, anti-ship, or antipersonnel mines could facilitate a strategic self-repositioning function. Should an existing mine be detonated, the remaining mines/nodes in the network would detect the detonation’s location and adjust themselves accordingly, either filling in any gaps in the mine field or congregating in the area of activity. Another application might involve attaching wireless sensor nodes to handheld weapons. Potential benefits could include user-authentication, battlefield restriction (they become unusable when taken out of an AO), or perhaps “talking” with other weapons in the unit and automatically communicating the need for reinforcements or ammunition re-supply based on usage or environment data.
-6-
Anthony F. Hillen Regardless of application, wireless sensor networks portend significant defense and security implications, necessitating a response in the form of policy formulation. Any military seeking to become or remain a formidable force should carefully consider the strategic opportunities wireless sensor networking technology can provide. Given the cost of military conflict, in terms of both monetary expense and potential casualties, a prudent strategist must pay close attention to technologies that could result in an advantage of any sort. Military history suggests that success is not achieved by those who first acquire a new technology, but by those who accept it and learn to wield it effectively.
-7-
Anthony F. Hillen
References
T. Clouqueur, P. Veradej, P. Ramanathan, K. Saluja, “Sensor Deployment Strategy for Detection of Targets Traversing a Region,” Mobile Networks and Applications, 2003. D. Davis “Synthetic Battlespace Test-bed for the Analysis of New Intelligence Sensors, Platforms and Techniques: A National Intelligence Simulation Center,” University of Southern California, 2007. M. Gerla, K. Xu, “Multimedia Streaming in Large-Scale Sensor Networks with Mobile Swarms,” UCLA Computer Science Department. J. Hellerstein, W. Hong, S. Madden, “The Sensor Spectrum: Technology, Trends, and Requirements,” SIGMOD Record, December 2003. V. Kumar, “Sensor: The Atomic Computing Particle,” SIGMOD Record, December 2003. XG Working Group, “The XG Vision: Request for Comments,” DARPA, Version 2.0
-8-
Wireless Sensor Networks
Technology Overview The individual nodes that constitute a wireless sensor network are generally small in size and use power-efficient batteries to extend their operational longevity. Depending on its function, each node has a sensor board that facilitates the detection and measurement of heat, vibrations, air-pressure and magnetic fields (among other things). The “motes” developed at UC Berkeley are a typical example of such devices. Motes have a range of about 100 feet and feature a 7Mhz processor, 4Kb of RAM, 128Kb of programmable memory space, and utilize a ChipCon CC1000 radio for communication. Due to their deployment simplicity and low cost of about $200 per unit, motes can be distributed in spatially dense configurations within a given area. Motes make use of Tiny OS, an operating system designed from scratch to be as power-efficient as possible. Using less than half the capacity of an AA battery, Tiny OS can effectively run applications for months at a time. (Hellerstein, Hong, Madden, 2003) Motes within a given geographic location use networking software to selfassemble into ad-hoc networks, allowing data to be transferred to and from any node in its network, or if necessary, to a proxy (but unauthorized, non-peer/client) in close proximity (like a random cell-phone or laptop), thereby serving as a conduit to a wider network (like the internet). The nodes in wireless sensor networks can be employed to capture data about their geographic environment while seamlessly and instantly communicating that information with surrounding nodes, impervious to temporal or spatial limitation. Wireless sensor networks circumvent the hindrances of collecting information from
-1-
Anthony F. Hillen geographic locations otherwise inaccessible by human beings; from the nether ocean to enemy occupied territory. (Kumar, 2003) Sensor networks are amenable to both civilian and military deployment. In civilian scenarios, sensors used to monitor traffic, pollution, or infrastructure can be positioned by hand. In terms of the most basic military applications, such networks can be used to detect, classify, and track targets in a given territory (other applications will be discussed later). Civilian use of wireless sensor networks range from environmental purposes such as pollution and ecosystem analysis to law-enforcement activity like traffic monitoring and criminal surveillance. In their military context, discommodious or threatrich environments can be accurately and safely reconnoitered, determining sensor placement a priori is unnecessary as random and widespread sensor deployment can be achieved via aircraft. (Clouqueur, Veradej, Ramanathan, Saluja, 2003) Development Status Sensor technology has made substantial advancements thanks to innovative new research efforts. Some recent developments have been academic in nature, like tracking and monitoring animal migrations, bird habitats, or vineyards, while private-sector developments have included efficiency improvements like “condition-based” equipment maintenance. There are numerous examples of how wireless sensor networks are currently being used, for instance, biologists at UC Berkeley interested in studying how trees affect the temperature and humidity in their surrounding canopy use a network of trunk-attached motes to monitor the microclimates around the redwood trees in their botanical garden. (Hellerstein, Hong, Madden, 2003)
-2-
Anthony F. Hillen One of the most promising research endeavors currently underway is the development of a flexible and interference-resistant communication technology. Instead of being restricted to transmitting and receiving information on a pre-assigned block of spectrum, these radio devices would utilize “opportunistic spectrum access”. Such systems would facilitate faster and more efficient communication since static allotment would be complemented by instantaneous and opportunistic spectrum access. Sensor nodes utilizing such technology would access unused spectrum, detect, authorize and network surrounding nodes in a manner that reduces inter-node communication interference. (DARPA “neXt Generation” program) A second area of research worthy of mentioning employs “mobile swarm” sensor networks to facilitate asset management and multimedia streaming. Mobile swarms are clusters of sensor nodes located in close physical proximity to each other and possess similar mobility patterns. For example, a group of tanks or UAVs could constitute a swarm, presumably equipped with qualitatively superior sensors like hi-res cameras, and longer range radios with higher channel bandwidths than conventional motes. Sensor nodes attached to the swarm members can gather information about that individual member, like location or operating status, but it can also relay data captured by its “host” to other nodes in the swarm, other mobile swarms, or to a command center through a backbone network or satellite. (Gerla, Xu) There are three primary motivations supporting research and development in the field of wireless sensor networks: academic interest, corporate profit, civil value, and of course, military application. These strong and mutually supportive driving forces suggest a promising future for the technology. Although motes currently cost about $200 per unit,
-3-
Anthony F. Hillen prices have been dropping steadily and are expected to continue falling. Some projections suggest that the price will fall to around $10 a piece within the next few years and that the units themselves will shrink in size to about 2 cubic mm. It is safe to assume that the smaller and cheaper these sensors get, the more widely used they will become. Moore’s law indicates that in about ten years, devices as small as a mote will have processing and memory capabilities similar to a contemporary network server. (Hellerstein, Hong, Madden, 2003) Several challenges faced by sensor technology are worthy of closer scrutiny. Software development, for instance, has been particularly troublesome. This is primarily due to the sensor’s hardware limitations. Modern sensors like motes suffer from a dearth in processing speed, memory, radio bandwidth, and energy capacity. The problems with processing speed and memory are likely to be resolved in the near future. However, the shortage in bandwidth is due to insufficient energy, and because the energy density of commercial batteries has not changed much in the last ten years, it is unlikely that the challenges posed by battery capacity and radio bandwidth will be overcome anytime soon. Other problems involve developing a way of programming groups of sensors to undertake a variety of different tasks and creating reliable security protocols to ensure network integrity and guard against intrusion and denial-of-service threats. (Hellerstein, Hong, Madden, 2003) Although most challenges are developmental, the technology’s inherent potential to violate widely held standards of personal privacy implies that there are also social and legal obstacles to many of its civil applications. Critics are quick to point out the ways such technology can be misused, from tracking ones every movement to remotely
-4-
Anthony F. Hillen accessing personal information. It is not difficult to imagine scenarios whereby one need merely walk by the wrong person in a grocery store to share one’s home address, credit card number, or other identity information. These are valid concerns that warrant significant discussion and policy-formulation prior to any civil application. Security Implications In contrast to civilian applications, the military applications of wireless sensor networks must be fully understood, embraced, and implemented without delay. Until now, the United States’ military use of sensor technology has been limited to basic and relatively crude detectors that utilize sensor technology, the Unattended Ground Sensor (UGS) system and the AN/GSQ-187 Remote Battlefield Sensor System (REMBASS) are typical examples of such devices. Although technically “wireless” by definition (in the sense that they do not require external cables to function), these devices do not utilize the technology discussed in this report, nor do they form ad-hoc networks of any kind. These systems are capable of detecting vehicle and personnel activity, but would be incapable of providing potentially critical battlefield information in the form of real-time audio/video data. Wireless sensor networks can also provide a strategic advantage in urban and close-quarter combat situations. For instance, an orbiting UAV could automatically detect friendly forces in the area and transmit aerial reconnaissance data directly to a heads-up display build into the helmets of troops on the ground. If a ground unit required a topographical map of an area, it could transmit the request to a nearby tank or UAV which would then acquire the information from a satellite or databank at a command center.
-5-
Anthony F. Hillen Military integration of this technology will require simultaneous training and tactical adaptation. The ability of its operators to function effectively in an informationrich environment will ultimately depend on the quality of their training. US combat statistics from the first Gulf War indicate that an abundance of battlefield information actually degrades combat performance. Its recipients could not process all the information at every level of command so units in critical need of information had to sift through too much irrelevant data to locate the specific details they required. The confusion caused by information overload illustrates the importance of implementing training and tactical reform measures whenever new technology is introduced. (Davis, 2007) It should be noted that military use of wireless sensor networks need not be limited to information awareness purposes. While not the most creative of individuals, I can think of a few applications omitted from existing literature on the subject. First, integrating wireless sensor networking technology with anti-tank, anti-ship, or antipersonnel mines could facilitate a strategic self-repositioning function. Should an existing mine be detonated, the remaining mines/nodes in the network would detect the detonation’s location and adjust themselves accordingly, either filling in any gaps in the mine field or congregating in the area of activity. Another application might involve attaching wireless sensor nodes to handheld weapons. Potential benefits could include user-authentication, battlefield restriction (they become unusable when taken out of an AO), or perhaps “talking” with other weapons in the unit and automatically communicating the need for reinforcements or ammunition re-supply based on usage or environment data.
-6-
Anthony F. Hillen Regardless of application, wireless sensor networks portend significant defense and security implications, necessitating a response in the form of policy formulation. Any military seeking to become or remain a formidable force should carefully consider the strategic opportunities wireless sensor networking technology can provide. Given the cost of military conflict, in terms of both monetary expense and potential casualties, a prudent strategist must pay close attention to technologies that could result in an advantage of any sort. Military history suggests that success is not achieved by those who first acquire a new technology, but by those who accept it and learn to wield it effectively.
-7-
Anthony F. Hillen
References
T. Clouqueur, P. Veradej, P. Ramanathan, K. Saluja, “Sensor Deployment Strategy for Detection of Targets Traversing a Region,” Mobile Networks and Applications, 2003. D. Davis “Synthetic Battlespace Test-bed for the Analysis of New Intelligence Sensors, Platforms and Techniques: A National Intelligence Simulation Center,” University of Southern California, 2007. M. Gerla, K. Xu, “Multimedia Streaming in Large-Scale Sensor Networks with Mobile Swarms,” UCLA Computer Science Department. J. Hellerstein, W. Hong, S. Madden, “The Sensor Spectrum: Technology, Trends, and Requirements,” SIGMOD Record, December 2003. V. Kumar, “Sensor: The Atomic Computing Particle,” SIGMOD Record, December 2003. XG Working Group, “The XG Vision: Request for Comments,” DARPA, Version 2.0
-8-
