Target tracking in uncertain multipath environments using Viterbi data association

This paper addresses the problem of accurate geolocation of targets or emitters in an uncertain multipath propagation channel. We propose a new approach which exploits differences between target and multipath dynamics to jointly estimate target position and channel parameters. The concept is derived here for an uncertain indoor propagation channel where multipath specular reflections from walls can be adequately captured using ray-trace models. A multi-mode Viterbi data association algorithm (MM-VDA) is presented which determines the unknown correspondences between measurements and multipath propagation modes in the presence of false alarms. The method is evaluated based on simulations where a moving radio frequency (RF) emitter is tracked in an uncertain multipath indoor channel even when the direct-path return is occasionally not present.