Abstract: The uncertain maneuver of a target is analysed and divided into four modes: non-maneuver, critical maneuver, weak maneuver, and strong maneuver. Based on that, a new adaptive estimator of target state-the interacting dual adaptive models estimator is set up. With the combination of a 2nd-order adaptive model of maneuver detection and a 3rd-order adaptive model of maneuver level and a way of the interacting adaptive combination, it reaches "all-round" adaptive ability of tracking the various motions of a target. The new estimator is used to track the five elementary motions of a target in Monto-Carlo simulation. The simulation shows that the estimator has the expected performance.

Key words: maneuver adaptive estimator, target tracking