Hidden Markov Model for target tracking with UWB radar systems

In this paper we demonstrate the application of Hidden Markov Models (HMM) for localization and tracking in ultra wide band (UWB) radar networks. To improve localization, a Voronoi region based approach is utilized to form a HMM for detection and tracking of mobile target. The observations used for the HMM localization are obtained from the power delay profile of the received signals. In UWB systems the use of entire power delay profiles instead of the total power only, allows to reach higher localization accuracy. This is due to the power delay profile being a measure of the power as well as the time of arrival. Simulation results suggest a performance gain of 7dB over the maximum likelihood estimation for localization in presence of path loss at intermediate values of signal to noise ratio (SNR).