Smoothing Data Association for Target Tracking in Clutter

Abstract

The proposed smoothing data association based on integrated probabilistic data association (IPDA) is applied for single target tracking in clutter. This algorithm operates two filters; one running forward in time (fIPDA) and other running backward in time (bIPDA). A new approach of fusing of each fIPDA prediction with bIPDA multi-track predictions with data association technique is derived to obtain smoothing predictions. The target state estimation is calculated by using smoothing predictions. Simulation is carried out to show improved false track discrimination performances over the existing algorithms.