Addressing the greedy-target problem in multiple-hypothesis tracking

Track-oriented multiple-hypothesis tracking (MHT) is arguably the most powerful current methodology in multi-sensor multi-target tracking. In this paper, we provide an accessible introduction to this methodology using a similar formalism to the one introduced by Kurien in the 1980 s. Once the formalism is in place, we show immediately that the target-death problem inherent in the probability hypothesis density (PHD) filter is not present in the MHT. On the other hand, we show that the MHT suffers from a problem of its own, that we term the “greedy-target problem”. We explain how this problem arises and identify a simple mechanism to mitigate its effect.