Research On Cooperative Tracking And Sensor Allocation Algorithm In Asynchronous Radar Networks

With the rapid development of sensor,computer,network and electronic equipment,a large number of multi-sensor networks for complex applications are emerging.The overall detection performance of the air defense system can be improved by cooperative detection of the radar networks.It is a key problem that allocating the limited sensor resources and collaborative tracking the targets in the radar networks.At present,Sensors in the radar networks have different sampling periods,inherent delays and communication delays.The sensor measurements and the fusion data are asynchronous.Therefore the thesis mainly studies cooperative tracking and sensor resource allocation methods in the asynchronous radar networking system.The main contributions are summarized as follows:(1)Some basic target tracking and fusion algorithm are presented.Moreover,a feedback sensor resource allocation framework based on the tracking performance is developed under the structure of multi-sensor cooperative tracking method.(2)In view of the asynchronous sampling measurements in the radar networks,a dynamic sensor resource allocation algorithm based on the asynchronous sequential fusion is proposed in the asynchronous radar networks.The algorithm estimates the target states with the sequential filtering fusion of asynchronous sampling measurements and selects the optimal sensor sets based on the differences between the expected covariance and the filtering covariance dynamically.Simulation analysis show that the proposed algorithm has lower computational load and higher target tracking accuracy compared with the dynamic sensor allocation algorithm based on pseudo-measurements.(3)In view of the asynchronous fusion of local fusion nodes in the distributed radar networks,a cooperative tracking algorithm based on distributed asynchronous track fusion is proposed.The algorithm uses the Least Squares(LS)algorithm to combine the information of the fusion nodes and obtain the virtual track.Then the fusion center obtains the global track by fusing the virtual track of each local fusion node.The algorithm dynamically allocates the sensor resource using the covariance control strategy and utilizes the track information to improve the trackingperformance.The proposed algorithm is compared with the centralized asynchronous track fusion algorithm.Simulation results show that the algorithm is better than the centralized asynchronous track fusion algorithm both in tracking accuracy and computational efficiency.(4)The Multiple target cooperative tracking method for the asynchronous radar network is studied.The multiple-sensor resource allocation model for tracking multi-target is developed based on the covariance control,which is solved by a heuristic algorithm.Then each target state is estimated by the real-time sequential filtering fusion of allocated asynchronous sensor measurements.The algorithm can operates in the different arrival time of each asynchronous fusion node,which decreases the computation burden and obtains the real-time tracking.Simulation results verify the effectiveness of the proposed algorithm.