The optimization of window length in target maneuver detection

Abstract

Decision-based approach of target tracking is an important technical field of maneuvering target tracking. In decision-based techniques, detection of maneuver onset is the key factor determining tracking performance. In order to minimize the detection delay of maneuver onset, the window length of maneuver detection was optimized under the Neyman-Pearson criterion. The optimum window length was obtained under various maneuver magnitude. Existence of the optimum window length was verified by computer simulations. The effects of processing interval, probability of false alarm of maneuver detection and maneuver magnitude to the optimum window length were given through computer simulation. In applications, one can consult the simulation results and use priori knowledge such as maneuver magnitude to choose sliding window length so as to get better tracking performance. (4 pages)