On /spl alpha/-/spl beta/ target tracking: the probability of target escape

Abstract

The paper considers the discrete time Kalman and H/sub /spl infin// approach to the two-state /spl alpha/-/spl beta/ target tracking problem. A closed form steady-state solution for the /spl alpha/-/spl beta/ parameter selection for H/sub /spl infin// tracker, which is similar to the tracking index process for the MSE (Kalman) criterion, is presented. The H/sub /spl infin// and Kalman processes model the radar/target system with the objective of keeping the target within the radar beamwidth. The two trackers are compared considering two cases that differ in input maneuver disturbances, random maneuver and constant acceleration. An example illustrates the performance of the /spl alpha/-/spl beta/ tracker with respect to keeping the target within the beamwidth in terms of probability of escape.