Accurate Murty's algorithm for multitarget top hypothesis extraction

In most hypothesis-oriented Multiple Hypothesis Tracking (MHT) implementations, the target-to-measurement data association is typically solved by using the Murty's algorithm. However, the Murty's algorithm has no control over the diversity of target-to-measurement associations - often the top associations vary only slightly. In addition, in practical tracking solutions, tracks are often grouped as tentative or continued. It was observed with real data sets that in the associations, the top hypotheses consist of mostly similar associations with the same confirmed tracks along with some permutations of new measurements. The result is that a fixed set of confirmed tracks dominate diversity of the association tree. To overcome this problem, a modified Murty's algorithm, which can achieve any user defined (or adaptable) diversity of track-to-measurement association of different types of tracks, is proposed in this paper. Numerical examples are provided to demonstrate the improved efficiency in hypotheses generation by the proposed method.