Abstract: For probability hypothesis density (PHD) filter, that was not able to track birth targets of unknown position, PHD filter with adaptive target birth intensity was proposed. The track initiation algorithm was employed to detect positions of promising birth targets which were used to form the intensity function of birth targets, and an online estimation algorithm of spontaneous birth intensity was proposed. Adaptive target birth intensity was combined with the recursion of the PHD filter, and a solution to the PHD filter based on adaptive target birth intensity for linear Gaussian target dynamics was proposed. Simulation results demonstrate that the proposed tracker improves on effectively tracking birth targets of unknown positions in the scenario at any time.

Key words: multi-target tracking, probability hypothesis density, birth target intensity, online estimation, Gaussian mixture