Multisensor-multitarget tracking based on hierarchical association only using rangedoppler measurements

The azimuthal resolution of high frequency surface wave Radar (HFSWR) is usually coarse due to the difficulty to build a sufficiently huge antenna array. Effective solution to the problem of target locating and tracking with non-angular measurements is of significance. Once it is solved, the sensors, such as HFSWR, do not have to use antenna with large aperture size any more. Aiming at the existence of `Ghost point´ in the ranging cross locating system this paper presents a hierarchical association algorithm. Based on the Range-Doppler observed by a single sensor, the nearest neighbour joint probability data association algorithm (NNJPDA) is adopted firstly to establish local track data, then the Ghost points are eliminated according to the target speeds measured by the two sensors respectively. The 0-1 integer programming is used to eliminate Ghost points further and provide the final association results. Simulation results demonstrate the effectiveness of the proposed method.