Suboptimal Data Association Technique for Multiple-Target Tracking in Dense Clutter Environment

Abstract

In multiple target tracking (MTT) systems that track targets with less-than-unity probability of detection in the presence of false alarms (FA), data association is very important. Data association is responsible for deciding which of the received multiple measurements should update which track. Some data association techniques use a unique pairing to update a track; i.e. at most one observation is used to update a track. An alternative approach is to use all of the validated measurements with different weights (probabilities), known as probabilistic data association (PDA). Due to the increase in the FA rate or low probability of target detection, most of the data association algorithms begin to fail. In this paper, we introduce a new suboptimal PDA technique for MTT in dense clutter environment. The proposed technique is based on merging the probabilistic nearest-neighbor filter (PNNF) with the PDA algorithm. The main idea is based on high-weighting the measurements that has minimum statistical distance from the predicted position of the target. The state updating equation in Kalman filter uses the combined innovation as in Joint Probabilistic Data Association method which is defined as the weighted sum of the residuals associated with many observations. Due to its simplicity in calculations and robustness, this technique can be used for real-time applications even though in dense clutter environments. We applied the proposed algorithm in tracking multiple targets in presence of various clutter densities. Results showed better performance when compared to Nearest-Neighbor and All-Neighbors approaches in different clutter densities and noise measurements.