TDOA-FDOA based Multiple Target Detection and Tracking in the Presence of Measurement Errors and Biases

Details

11:30 - 12:45 | Wed 6 Jul | Salisbury B | S13.6

Session: Localization and tracking (indoor and outdoor)

Abstract

This paper considers joint detection and tracking of multiple targets using a sequence of time difference of arrival (TDOA) and frequency difference of arrival (FDOA) measurements obtained from a pair of receivers. The number of targets, as well as the position and velocity of each target are assumed to be unknown and time-varying. A particle filtering based data association (PFDA) method is employed to detect and track multiple targets. The birth, survival and death models together with the motion model are defined to describe the target dynamics, and different hypotheses on the measurement-target associations are generated and evaluated by using a particle filtering (PF). Since errors and biases can be present in the measurements due to sensor synchronization errors and non line of sight wave propagation, a bank of random walk modes are included to model the biases and errors. The likelihood of the particles are thus computed by combining all these modes so that the biased measurement can also be associated with the target track. Empirical experiments are conducted to demonstrate the performance of our proposed approach.