Study Of Target Tracking Techniques In Multi-sensor Data Fusion System

Multi-sensor data fusion is a comprehensive discipline for multi-source information processing in which multi-target tracking is the most popular research field. How to improve the precision of target tracking. in multi-sensor data fusion system is mainly studied in this paper.In this paper, an introduction is first given to the general principle of multi-sensor data fusion and target tracking, and then the method of tracking multiple maneuvering targets is studied in single radar environment, finally, the data preparation for sensor network and track association and fusion in sensor network system are introduced and discussed.Due to the complexity of some algorithms for tracking multiple maneuvering targets in clutter, a new method is presented based on adaptive a-βfilter and improved FDA(Fuzzy Data Association), it uses the maneuvering index to adjust the size of the gate adaptively. This method has low computational load, and is easy to realize in the TWS(Track While Scan) systems.Due to all kinds of fuzziness in multi-sensor multi-target track fusion, a new track fusion method is presented based on improved FCM(Fuzzy C-Means) algorithm, in this method, track association and fusion are combined, not only the range information but also the azimuth information are considered to reduce the fuzziness and improve the precision of track fusion. Monte Carlo simulation results show its good fusion performance.