Multisensor tracking of a maneuvering target using multiple maneuver model PDA

Abstract

This paper presents a new method for multisensor tracking of a maneuvering target in dense environment. This is the extension and improvement of the conventional probabilistic data association (PDA). In addition to radar measurement, the use of accurate angle data from an infrared (IR) sensor is proposed to improve tracking performance, We propose IR-integrated multiple maneuver model PDA (IM3PDA) filter which combines IR and radar sensor data. The maneuver acceleration is selected from a time invariant set of discrete values and follows a Markov process. In this method, the maneuver of a target increases the prediction covariance as compared with that which is obtained by the standard Kalman filter equations, and so, validation gate size varies automatically with the maneuver of the target. The performance of this method is evaluated in terms of tracking success rates and position estimation accuracy by computer simulation.