Real-time adaptive fuzzy density clustering for multi-target data association

Abstract

The problem of data association for tracking multiple targets based on using the ship-borne radar is addressed in this study. A robust fuzzy density clustering algorithm is proposed, that contains three steps. At first, a customized form of adaptive density clustering is used to determine valid measurements for each target’s state. In the second step, the degree of fuzzy membership for each valid measurement is determined based on the maximum entropy approach. At the final step, the measurements with a maximum degree of membership are used for updating the position of the targets. The proposed approach does not require gating techniques and led to the reduction of steps in comparison with other data association methods. In addition, the effect of ship movement in the performance of the tracking filter, based on the adaptive extended Kalman filter (AEKF) was studied. The efficiency and effectiveness of the proposed algorithm are compared with the nearest neighbor (NN) with Mahalanobis distance and Fuzzy nearest neighbor (FNN) methods. The results demonstrate the main advantages of the proposed algorithm, including its simplicity and suitability for real-time target tracking in cluttered environments.