An improved particle filter based on 4D discontinuous measurements

During the target detection of the airborne pulse Doppler (PD) radar, a 4D measurement comprised of range, azimuth, elevation and radial velocity can be obtained to track the target exactly. But the factors including the Doppler blind zone (DBZ) would result in discontinuous measurements which breaks the continuity of the target track. To be consistent with the practice, we constitute the 3D relative motion model between the airborne radar and the target based on moving platform, and deduce the character of the relative radial velocity in DBZ. According to this, an improved particle filter with relative maneuver detection is proposed. This algorithm is able to adjust the moving direction of the particles according to the type of maneuver, and maintain the tracking by predicting the possible region where the target reappear from the DBZ. The simulation results indicate the validity and feasibility of this method.