Efficient multiple hypotheses tracking scheme using adaptive number of ‘K’ best hypotheses for target tracking in clutter

In a cluttered target tracking environment multiple hypotheses tracking (MHT) based algorithms improve the data association by considering a batch of measurements. To reduce the computational complexity generated by the exponential growth of hypotheses, the number of hypotheses is limited to k in the k best MHT algorithm. In conventional k best MHT algorithm the value of k is fixed. This paper proposes a method to keep the value of k adaptive depending on the scenario complexity and also depending on the likelihood of the valid hypotheses. The Monte Carlo simulation results carried out in this paper justifies the advantage of the proposed method compared to the fixed k-best MHT algorithm.