High Maneuvering Multiple-Underwater Robot Tracking with Optimal Two-Stage Kalman Filter and Competitive Hopfield Neural Network Based Data Fusion

Abstract

Tracking high maneuvering underwater targets with radar is very complex, because the acceleration is not directly measurable. This may cause the tracking errors which will diverge the estimation when maneuver happened. In this article Optimal Two-Stage Kalman Filter (OTSKF) used for tracking purpose, which can estimate the acceleration in maneuvering targets. Under this method, Acceleration of the targets has been identified and it is estimated at the moment. Also an effective data fusion method plays critical role in order to track the objects accurately. We use Competitive Hopfield Neural Network (CHNN) to associate obtained data from radar measurements with existing target. In this paper we investigate new method, which combine these two algorithms for tracking highly maneuvering targets. As simulation results show, to solve multiple-underwater vehicles tracking problem, investigated method have good ability and suitable performance in high maneuvering targets.