A Smoothing Algorithm for Improved Tracking in Clutter and Multitarget Environment

The probabilistic data association filter (PDAF) has been proven to be a very good practical approximation to the otherwise computationally impractical, optimal or nearly optimal algorithms for the problem of tracking targets in cluttered or multi-target environments. In this paper, we develop a smoothing algorithm in the spirit of the PDAF, called the PDAS, in order to incorporate the advantages of smoothing techniques to the tracking problem. Various methods of using the smoothing techniques without undue or excessive increase in the computational method are briefly described.