Research On Cooperative Tracking Method For Distributed Multisensor Network

With the rapid development of computer,sensor and communication equipment,a large number of multi-sensor networks for complex applications are emerging.The information fusion of multiple sensors is an effective way to improve the target detection performance.However,the limited sensor capability and the uncertain target environment make it impossible for the sensor network to track the same target with all sensors.Thus,how to distribute the limited sensor resources to collaborative track targets has become a key problem of target collaborative tracking.As a number of sensors and targets increase,the complexity of target collaborative tracking is increase.Distributed processing can avoid the single sensor failure on the whole network performance and reduce the communication requirements in the large-scale sensor network limited by the capacity.In this study,we focus on the cooperative target tracking problem in distributed sensor networks from the point of view of distributed fusion and sensor management.The main contents are as follows:1.A distributed cooperative asynchronous sensor network cooperative tracking algorithm is proposed to solve the problem of large-scale sensor network with limited capacity constraints of the fusion center in the large-scale asynchronous sensor networks.Firstly,the distributed deployed sensors are clustering into multiple subnets.The sensor sets are selected to cooperative track the target in each sub-network at the next time based on the maximum information gain criteria.Then the asynchronous sequential fusion method is carried out for selected sensors in each subnet.Moreover multi-hop information transfer between sensors is use to determine the optimal transmission path and data fusion from the perspective of reducing energy consumption.Finally simulation results show that the proposed algorithm is effective.2.A multi-sensor multi-target cooperative tracking algorithm based on PCRLB criteria is proposed considering the time-varying measurement variance.Firstly,the time-varying measurement variance of each sensor is given and its effect on the tracking accuracy is analyzed.The PCRLB criteria are used to allocate the sensor set to track the target at the next time considering the PCRLB criteria are not associated with the filter algorithm.In addition,the converted coordinates Kalman filtering algorithm is developed to cooperative tracking by coordinate transformation of the measurement variances.Finally the simulation results verify the efficiency of thealgorithm.3.A decentralized sensor network cooperative tracking algorithm based on PCRLB is proposed to deal with the limitation of communication between sensors in large-scale sensor networks.The decentralized PCRLB index is calculated for the structure of distributed sensor network with tree topological structure.The sensor selection is based on the decentralized PCRLB criteria,and the local fusion center is estimated by parallel filtering.Simulation results show that the proposed decentralized multi-sensor cooperative tracking algorithm is more effective compared with the centralized fusion algorithm.