Abstract:

A multiple extended targets tracking algorithm based on Gaussian-mixture probability hypothesis density filter is proposed to solve the problems of track initiation and measurement partition of the extended target in the clutter environment. The proposed algorithm takes clustering tendency of measurement sets into account at the stage of track initiation. Then, the improved ordering points to identify the clustering structure (OPTICS) algorithm is used to extract the measurement cluster. Through the establishment of an augmented data set sequencing to represent the measurement set density structure, input parameters and initial point selection are not sensitive to the choice. Besides, the algorithm can filter the measurement from clutter. Simulation results show that the proposed algorithm has a better computational efficiency over the traditional algorithm at the stage of track initiation. In the clustering process, the proposed algorithm can be used to partition the measurement sets of different densities and shapes, adaptively determine the number of partitions and increase the computational efficiency.