Abstract:
In order to compare these algorithms on track fusion accurately, it is necessary to construct a metric of high reparability. This paper improved the metric, global optimal sub pattern assignment (GOSPA) distance measure of track fusion, and got a new metric, HGOSPA, by introducing the concept of uncertainty into the computation of track distance error in GOSPA distance. By contrast experiment, it was shown that the metric HGOSPA can sensitively reflect the pros and cons of track fusion algorithms, due to considering the uncertainty. Furthermore, the experimental results also validated that the evaluation results of HGOSPA were consistent with the true situation. That is,HGOSPA can attain the correct evaluation of different track fusion algorithms even if without the true value.