Localization and Dynamic Tracking Using Wireless-Networked Sensors and Multi-Agent Technology: First Steps.
ZD Deng,WX Zhang
2002-01-01
IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences
Abstract:SUMMARY We describe in this paper our experience of develop-ing a large-scale, highly distributed multi-agent system using wireless-networked sensors. We provide solutions to the problems of localization(position estimation) and dynamic, real-time mobile object tracking, whichwe call PET problems for short, using wireless sensor networks. We pro-pose system architectures and a set of distributed algorithms for organiz-ing and scheduling cooperative computation in distributed environments,as well as distributed algorithms for localization and real-time object track-ing. Based on these distributed algorithms, we develop and implement ahardware system and software simulator for the PET problems. Finally, wepresent some experimental results on distance measurement accuracy usingradio signal strengths of the wireless sensors and discuss future work.key words: localization and mobile object tracking, distributed algorithms,multi-agent systems, wireless sensor network, and MEMS. 1. IntroductionIn recent years, the technology of micro-, electro-mechanical systems (MEMS) has made rapid advances.Various smart devices, such as sensors and actuators withsome information processing and communication capabili-ties embedded within, have been developed and deployed inmany real-world applications [9]–[11]. The application do-mains of such smart devices are numerous, including newgeneration pervasive computing systems, avionics and plantautomation, building and environmental monitoring, multi-hop routing discovery [6], augmented reality and virtue re-ality systems, active badge [13], and multiple robotics sys-tems. To use smart devices in these applications, it requiresto connect a large number of sensors and actuators, up tothousands, tens of thousands or even millions of units, andto integrate and embed sensing, communication, signal anddata processing, and control functionson individualdevices.The key to large-scale networked, embedded MEMS is themechanism with which individual units are programmed towork as a coherent piece toward achieving common goals.Due to the distributed and real-time nature of most applica-tions and the size of such a large system, information pro-cessing and decision making very often need to be carriedout on the units where actions take place.