The Airborne Multi-sensor Target Recognition Based On Data Fusion

Fighters are as one of the important forces in the battlefield, playing an important role in the getting air control. A large number of precision sensors of the aircraft make itself become an important integrated sensor, forming a multi-sensor complex system. Fighters both fusions the airborne multi-sensor information, but also integrate information transmits from other platforms over the battlefield, so airborne multi-sensor data fusion application is the inevitable trend of development of advanced fighters.In this paper, airborne multi-sensor data fusion system has been discussed mainly.The major works are as follows:(1) Analysis of airborne multi-sensor data fusion system platform, the main research questions, data fusion needs and objectives. For the particularity of the fighter avionics, establish the multi-sensor data fusion system architecture and model of airborne avionics system based on the evidence theory.(2) Discuss of the evidence theory combination rule problems--how to use the combination rules when the evidences are at a high degree of conflict. Study of Dempster combination rule and its improvement methods. The classic evidence conflict coefficient can not effectively measure the extent of the conflict; we present a correlation coefficient method. Combined the correlation coefficient, the classic conflict coefficient and the pignistic reliability distance to determine the degree of conflict between the evidence, based on ordered weighted averaging operator (OWA), proposed a comprehensive measure of conflict between the evidence parameters. With the new coefficient of the evidence presents a conflict of evidence theory to improve the traditional method. The new method is effective handling of Dempster combination rule does not apply when the evidences are at a high degree of conflict.(3) Based on aircraft platform, research to combine belief function of multi-sensor's evidences by DS evidence combining rule. Present the methods of get BPAs and ways to deal with conflicts of evidences. And the BPA-based fusion decision rule is adopted. Therefore, an approach of fusing uncertain information from different sources is proposed. Finally, a simulation scenario is presented to prove the validity of this data fusion method.