RAO-BLACKWELLIZED VARIABLE RATE PARTICLE FILTERING FOR HANDSETTRACKING IN COMMUNICATION AND SENSOR NETWORKS

摘要

Knowledge about the position of a user s mobile handset device constitutes a valuable information for emerging location-based services and applications. While satellite-based navigation systems are the chosen technology for outdoor, rural and semi-rural environments, location in urban canyon or indoor navigation is still unsolved. In this paper, we exploit the fact that handsets are usually surrounded by different kinds of communication and sensor networks, which can be used to enhance the coverage, accuracy and robustness of satellite-based systems. After outlining models for motion, measurements and positioning, we depict a common scenario and propose a methodology for handset tracking. We show the suitability of Particle Filtering to the problem at hand by the appliance of a modified version of a recently proposed algorithm, the Variable Rate Particle Filtering (VRPF). The novelty relies on the inclusion of a Rao-Blackwellization procedure that significantly reduces the computational load. Details about the implementation and some significant numerical results of computer simulations are also provided.