images CHAPTER 14

Multitarget Tracking and Classification in Collaborative Sensor Networks via Sequential Monte Carlo Methods

Tom Vercauteren1 and Xiaodong Wang2

14.1 INTRODUCTION

The convergence of recent developments in microelectromechanical systems (MEMS), microprocessors, and ad hoc networking protocols have enabled low-power and low-cost sensor nodes to collaborate and achieve large tasks [1]. Individually, each node owns limited sensing, communicating and computing capabilities, but, when a large number of them are used in conjunction, it is possible to achieve a reliable and robust network. These devices collect measurements from the physical environment, communicate with each other, and carry out computations in order to transmit only the required data to the end user. The network is then able to perform event detection, event identification, and location sensing in a field under observation [2]. Typical applications of such sensor networks are environmental monitoring, military surveillance, and space exploration, to name a few.

This chapter focuses on the problem of jointly tracking and classifying several targets evolving within densely scattered sensor nodes. On the one hand, multiple target tracking tackles the issue of sequentially estimating the state of a possible varying number of objects; and on the other hand, classification deals with the identification of those objects ...

Get Handbook on Array Processing and Sensor Networks now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.