Power and robustness of a track-loss detector based on Kolmogorov-Smirnov tests

It is important in the practical application of data association algorithms to target tracking in cluttered environments to be able to effectively and efficiently detect track-loss in the absence of truth data. We recently developed a track-loss detector using Kolmogorov-Smirnov tests to determine the regime of filter operation in the absence of truth data for data association algorithms where estimates of the innovations covariance are available. Among advantages to this method are that confidence intervals are associated with the regime tests, and that the number of measurement samples required can be determined adaptively. In this paper, results are developed for the exact and asymptotic power of the test, as well as a measure of the robustness of the method to errors in parameters such as clutter density. Simulation results are provided