Joint detection, tracking and classification of multiple extended targets using PHD filter

To account for joint detection, tracking, and classification (JDTC) of multiple extended targets and to estimate the close extended targets, a recursive algorithm based on the extended target probability hypothesis density (ET-PHD) filter was proposed with the particle implementation. The feature measurements were incorporated into the filter. In the prediction stage, the particles were propagated according to their classes. A joint update was implemented by updating the particle weights in the update stage. Then the particles were classified by their class. The particles with the same class label and their corresponding weights represented the estimated PHD distribution of the corresponding class. The algorithm has a flexible modularized structure with the computational order O(mn). A simulation example involving the tracking of two closely spaced parallel moving targets and two crossing moving targets from different classes indicates that the algorithm is capable of estimating the classes, numbers and states of extended targets simultaneously. Moreover, the average OSPA distance reduced by more than 50% compared with the traditional algorithm. © 2015, Editorial Board of Journal of Shanghai Jiao Tong University. All right reserved.