SSP'05 -- Proceedings

	IEEE/SP 13th workshop on Statistical Signal Processing July, 17-20, 2005 - Bordeaux - France

Information regarding the paper

Title

Sensor network particle filters: motes as particles

Author(s)

Mark Coates	McGill University
Garrick Ing	McGill University

Get the paper in PDF format

To obtain Acrobat Reader (version 5 minimum required) necessary to his read.

Abstract

We describe an algorithm for tracking an object using particle filtering in a sensor network comprised of smart dust-type motes. We investigate the situation where the motes are equipped with binary proximity sensors, low-power lasers and optical receivers for communication with nearby motes, and corner-cube arrays for communication with a central transceiver. The particle filter we describe is largely decentralized; a central transceiver performs no processing beyond a summation and weighted average. Individual motes act as the particles in that they represent candidate positions of the object. Propagation of the particle filter is performed through activation of appropriate neighbouring nodes with weighted messages. We provide simulation results of tracking a maneuvering object, comparing performance with a centralized particle filter.