Improving Track Continuity Using Track Segment Association

In this paper we present a new track segment association technique to improve track continuity in large-scale target tracking problems where track breakages are common. A representative Airborne Early Warning (AEW) system scenario, which is a challenging environment due to highly maneuvering targets, close target formations, large measurement errors, long sampling intervals, and low detection probabilities, provides the motivation for the new technique. In order to combine broken track segments and improve track continuity, a new track segment association algorithm using a discrete optimization approach is presented in this paper. Forward prediction and backward retrodiction give the estimates (at a common time) of the pairs of track segments that are candidates for possible matching. The representative AEW scenario considered consists of 120 targets over 120 scans with straight, zigzag, oval, and jinking motions. Simulation results on this scenario show significant improvement of the mean track life and position, speed and course RMS errors with track segment association.