A novel method of target tracking based on tracks splitting

Aiming at the problem of more than one measurement in a tracking gate, this paper introduces the concept of true trajectory probability based on the using of tracks splitting method, therefore the process of target tracking falls into two parts: states update and true trajectory probability update. True trajectory probability denotes the possibility of the trajectory being the track of the actual target. The updated values of the true trajectory probability depend on the prior information of measurements in current time. The method also obviates the need for track initiation and track termination. Proper assignment and dynamic adjusting of the probability gate make the real-time tracking more effective as well as reducing the consumption of the system resources.