Track-to-Track Association Based on Target Position Estimate and Type of Historical Information

Track-to-track association posses great significance in distributed sensor systems， and accurate association is beneficial for the sequent track-to-track fusion process. This paper proposes a new statistical method based on two kinds of historical information. On the one hand, the target position detected by the electro-optical sensor is transformed into Earth-Centered, Earth-Fixed, then Kalman filtering is used to form the target track. In each track, the average of state estimate and error covariance of historical sequence is calculated to obtain the target position statistic. On the other hand, the target type of historical information is counted to obtain the target type statistic. Finally, two kinds of statistics are tested by hypothesis, and the association results are comprehensively judged to complete the multi-track association of two sensors. The simulation results have verified the effectiveness and superiority of this method.