Target Track Initiation Comparison and Optimization∗

This paper proposes a procedure for comparison of target tracking filters in cluttered environment, using simulated experiments. Important simulation issues, including track generation and retention across simulation runs and true track definition are discussed. False track discrimination is identified as an important target tracking filter performance measure. This measure has two components; confirmed true track statistics and confirmed false track statistics. In order to meaningfully compare false track discrimination capabilities of target tracking filters, we suggest optimizing true track statistics, while using false track statistics as the optimization constrains. The optimization procedure is described, and the importance of correct parameter selection is il-