Multi Target Tracking Using Zero-Backscan Multiple Hypothesis Tracking
Data association or correlation in Multi Target Tracking (MTT) is generally achieved by using n-backscan Multiple Hypothesis Tracking (MHT) or Joint Probabilistic Data Association (JPDA). These methods are computationally heavy due to deferred decision in MHT and batch processing in JPDA leading to delay in correlation. Hence, a new correlation algorithm is needed to reduce the time taken in correlation. This paper proposes a new correlation algorithm, namely, Zero-Backscan Multiple Hypothesis Tracking (ZBMHT), and the execution time of ZBMHT is less than n-backscan MHT due to sequential processing of hypotheses. It is observed that the proposed correlation logic takes lesser time than the existing n-backscan MHT and JPDA. The results were analyzed for increased number of targets.