Particle Filters for Rss-Based Localization in Wireless Sensor Networks: An Experimental Study

This paper focuses on the development of a radio localization technique for a wireless sensor network infrastructure where a large number of simple power-aware nodes are spread in indoor environments. Fixed and moving nodes exchange radio messages but can only measure mutual power figures such as the received signal strength (RSS) indicator. Local maximum likelihood estimation from propagation models suffers from false alarm problems due to incorrect position information, complex indoor propagation effects and simple hardware radio architectures. Here, we propose a Bayesian approach to estimate and track the position of a moving node from power maps obtained through field measurements. To lower the computational power required by grid-based algorithms, we exploit particle filter techniques that implement an irregular sampling of the a-posteriori probability space. Finally, experimental results are presented and discussed.