MAP-PF/LRDT Multitarget Tracking for Next Generation Over the Horizon Radar - MP 88-10

Detection, localization, and tracking in over the horizon radar (OTHR) systems is a challenging problem due to the existence of multiple ray mode propagation paths between the transmitter, target, and receiver. Traditional OTHR multi-target tracking techniques use three isolated processes: slant azimuth/range/Doppler estimation (detection), followed by association of detections to targets and mode paths, followed by fusion of associated slant detections into a single track in Earth coordinates for each target. We propose to use the maximum a posteriori penalty function (MAP-PF)/ Likelihood Ratio Detection and Tracking (LRDT) approach to formulate the detection and tracking problem as a MAP estimation problem operating directly on the radar array data. The result is a two-step estimation process similar to traditional methods, except the processes are coupled via the penalty function and the data association step is eliminated. In the azimuth/range/Doppler estimation process, the penalty function guides the estimator to look for detections at slant coordinates of possible mode paths for the current target state. In the track estimation process, the penalty function adaptively determines the influence of mode path estimates on the final track estimates. The algorithm will be implemented in Matlab and tested on Relocatable OTHR (ROTHR) data. BENEFIT: MAP-PF/LRDT has the potential to improve target detection, localization, and tracking accuracy while reducing computational complexity in next generation OTHR systems.