Maneuver detection using the radar range rate measurement

Tracking maneuvering targets with radar is complicated because radar cannot measure target accelerations. We use the range rate measurement to calculate a new statistic that is a surrogate measurement of target acceleration under a coordinate turn model. Its distribution is found via simulation. A threshold test of the statistic turns out to be a reliable detector of a maneuver, and its receiver operating characteristic is found. A tracker that uses a threshold test of the new statistic of accelerations to detect maneuvers and set the state noise in a Kalman filter tracker is devleoped and compared to other, common maneuvering track filters. The new method outperforms a two mode interacting multiple model tracker and a noise switching tracker that switches based on the position measurement residuals