Multi-target angle tracking via antenna array

Abstract

A new algorithm for tracking the angles of arrival of multiple moving targets using sensor array outputs is proposed. In the standard approaches to multi-target angle tracking, estimates of the target angle trajectories are obtained from noisy measurements of the target positions provided by one or more sensors. A key problem in multi-target tracking is the measurement/target data association problem; that is, the problem of determining which measurement in the measurement set corresponds to which targets. To overcome this problem, we apply the marked subspace technique, directly to the ESPRIT algorithm without arising the data association problem. By using single spatial rotation invariance and single temporal rotation invariance algorithms, we can individually estimate the directions-of-arrival /spl theta//sub i/(t)'s and the temporal frequencies /spl omega//sub i/'s of the targets. Applying the marked subspace technique to put different markers on the signal subspaces, we can easily obtain the parameter set {/spl theta//sub i/(t), /spl omega//sub i/} which are auto-paired and belongs to the same ith target. Hence, if the temporal frequencies of the received signals are distinct and kept constant over the entire tracking time, we can recognize which /spl theta//sub i/(t)'s belong to which targets by recognizing temporal frequencies /spl omega//sub i/'s at each time t=T, 2T, 3T, ..., where T is a measurement interval. Therefore, multi-target angle tracking can be performed without the data association problems. Simulation results show the effectiveness of the proposed method in comparison to the recursively updated method.