Two-tier target tracking framework in distributed sensor networks

Document Type

Conference Proceeding

Publication Date

1-1-2014

Abstract

Copyright © 2014 Inderscience Enterprises Ltd. A distributive two-tier tracking framework is developed to track mobile targets in sensor networks. The algorithm utilises the efficiency of binary sensor reading to ensure fast isolation of the target, and then it uses a small cluster of sensors around the target to compute accurate target location. The framework has two tiers, in tier-one, binary sensor reading and Viterbi algorithm are used, it is a fast and distributed way of roughly locate the target. In tier-two, a detailed localisation using maximum likelihood estimation is done within a small cluster of sensors to compute the accurate location of the target. By dividing the framework into two tiers, scaling the size of the network only affects the computation cost in tier one, where binary data are used. The computation cost in tier 2 remains unchanged. The framework is first implemented in a chained-form network with sensors deployed in a regular pattern, it is then extended to other regularly deployed network in full 2D space. The two-tier tracking framework is implemented using ultrasonic sensor network in a real-world hallway-room environment in our lab.

Publication Title

International Journal of Sensor Networks

Link to Full Text

Share

COinS