Adaptive Detection Threshold Optimization for Multi-Target Tracking in Clutter.

摘要

The problem of selecting signal processing parameters, particularly detection thresholds, so as to optimize downstream tracking performance is examined further. Numerical simulations are used to establish the validity of certain approximations made previously. These simulations suggest that steady-state analysis is inadequate, and an adaptive threshold optimization scheme is proposed as an alternative. Finally, the original derivation of the Probabilistic Data Association Filter (PDAF), upon which the present work is founded, is augmented to account for finite gate size. (Author)