Suboptimal JPDA for tracking in the presence of clutter and missed detections

A new heuristic for data association on multi-target tracking systems is presented. The algorithm is based on the existing JPDA (Joint Probabilistic Data Association) algorithm, more specifically, on the Suboptimal JPDA approximation. However, compared to Suboptimal JPDA, the presented method exhibits a significant improvement on tracking performance for scenarios in the presence of clutter and missed detections, at negligible increase of computation cost. In fact, it approaches the performance of the classic JPDA, requiring, however, drastically lower computational resources. It is also be shown that the proposed method is more robust with respect to trajectory crossings than the original Suboptimal JPDA.