Ground target tracking with experimental data of the PAMIR system

The ultimate goal of a tracking algorithm is the generation of high-quality tracks in terms of track precision and track stability. But several challenges are generally present which tend to strongly deteriorate the performance of standard tracking filters. Based on airborne radar measurements, the most significant challenges in ground target tracking are high false alarm backgrounds, closely-spaced targets, sequences of missed detections due to terrain or technical obscuration and the generally complex motion of ground targets. To counterbalance these challenges, Doppler blind zone and road information are in particular appropriate for which simulation studies indicate a significant improvement in tracking performance. But so far, the performance of extended tracking filters has been analyzed only rarely based on realistic data which generally comprise additional unexpected effects and a higher complexity. Therefore, this work evaluates an MHT algorithm augmented by blind zone and road network information based on experimental data to assess the tracking performance under realistic conditions. The data set was recorded by the PAMIR sensor of Fraunhofer FHR during a flight campaign conducted in 2009.