Active/Passive Radar Information Fusion And Its Application To Maneuvering Target Tracking

Information fusion technology is rapidly developed and get more and more interesting in recent years. Information fusion can be widely used in variant fields that require accurate and reliable information and status assessment according to the data from multi-sensor or multi-source. Active/passive radar target-tracking system is a typical two-sensor system in which different type measurements can be obtained. In active/passive radar systems, target motion states estimated must be fused together for further usage.This thesis is mainly focused in maneuvering target information fusion in active/passive radar tracking systems. First, maneuvering target models and information fusion theory are discussed. For tracking high maneuvering targets, a new Kalman filter algorithm based on fuzzy-logic theory is presented, in which filter eovariance matrix is adjusted adaptive to filter innovation in a fuzzy manner. Simulation results prove this filter scheme has certain advantage to Kalman filter, and is especially more reliable and accurate in high maneuvering target-tracking conditions. Then different target data fusion scheme are studied. Two target models (i.e. constant velocity model and constant acceleration model) are considered in study. Kalman filter (including extended Kalman filter) and eovariance fuzzy-logic based on Kalman filter are used in target state estimation. The accuracy is compared for equal-weighted fusion and covarianse-weighted fusion scheme. The results show that covariance-weighted fusion scheme is more reliable for high-g target tracking.