Track Association And Attribute-Track Fusion In Data Fusion System

In this paper, the methods of track association and attribute fusion in data fusion system are studied. An improved Nearest Neighbor arithmetic is presented, which adds the direction of velocity of target based on the general Nearest Neighbor (NN) arithmetic. This arithmetic that not only uses the velocity, distance as the correlation gate, and also introduces the angle difference of velocity between the measurement point and the track as other basis of correlation, could solve the problem of false association when targets are dense and tracks are crossing. To improve the real-time performance of tracks association in the circumstances of large quantities of targets, an Approximate Multi-Sensor Joint Probabilistic Data Association (AMSJPDA) method is discussed. AMSJPDA redefines the conception of cluster in the MSJPDA method and introduce the conception of approximate cluster, which makes the number of targets in the cluster fewer. So the efficiency of calculation is improved highly.The accuracy of conventional radar and ESM association algorithm that use azimuth as the basis of correlation is low. To solve this problem, using the high precision of attribute parameters from ESM sensor, Multiple Factor Synthetic Correlation (MFSC) arithmetic is developed. MFSC combines recognition reasoning with attribute association effectively by making good use the historical data. And the rate of correct correlation is improved. At the same time, with cross bearing method which use sensors of different places locates the target the problem of correlation of new start -up radiation source is resolved.In addition, this paper present a performance testing system of target tracing based on the target tracing requirements of the testing program, and makes a comprehensive testing, takes statistics in detail and analyzes deeply.