Performance Analysis of MHT Algorithm in Cluttered Environment

Considering the high computational complexity of multiple hypothesis tracking (MHT), its performance in cluttered environment is analyzed. First, a practical implementation of MHT algorithm is designed. Then, the methods to reduce the computational complexity are presented. Finally, computation-time per one sample and reserved track hypotheses number are investigated in every sampling time when MHT use different delay depth by means of simulations. The results indicate that computation-time per one sample and reserved track hypotheses number increase quickly if delay depth increases. When the delay depth is not less than 3, the MHT algorithm can gain good tracking performance. The research results are helpful to design a real-time efficient MHT algorithm.