Extended target tracking algorithm based on improved density peak clustering

This paper proposes an improved density peak clustering (IDPC) measurement set partition algorithm to solve the problem of measurement set partition in the extended target Gaussian mixture probability hypothesis density (ET-GM-PHD) filter. Firstly, the IDPC algorithm removes the low local density clutter-generated measurements to obtain the most likely target-generated measurements. Secondly, cluster the remaining measurements is able to obtain spatially close clusters and cluster centers. Finally, according to the projection of the mean vectors of the predicted Gaussian component with higher weight on each cluster, an accurately measurement set partition is obtained. The simulation results show that, compared with the existing measurement set partition methods, this proposed algorithm can greatly reduce the calculation time while maintaining tracking accuracy.