A historical information feedback multiple-target tracker

The traditional PHD tracker (PHDT) generates the newborn PHD from the prior knowledge, thus can not be applied when the prior newborn target knowledge is unavailable. To this end, we propose a historical information feedback multiple-target tracker (HIFMTT) as an improvement to the traditional PHDT. The HIFMTT generates the newborn PHD through processing the historical observation and estimating results with a feedback structure, and thus can overcome the above mentioned difficulty. For the real application, we also construct the PF-HIFMTT algorithm by embedding the particle filter into the HIFMTT framework. The simulation results demonstrate the effectiveness of the PF-HIFMTT algorithm in two different scenarios.