Variable-Numbered Multi-Target Tracking Algorithm Based On Universal Track Initiation Model With Adaptive Threshold

This paper firstly proposed a universal track initiation method with adaptive threshold. Using this method, performance under heavy clutter tracking environment is improved. Besides, joint probabilistic data association filter is the common data association method, which only deal with the fixed-numbered multi-target tracking problem. Introducing the data pre-association process, an improved JPDAF algorithm is proposed to accomplish variable-numbered multi-target tracking, which is validated by simulation results.