Research On Multi-node Collaborative Localization Method

Multi-node cooperative positioning can improve the accuracy and reliability of positioning by integrating the information of each node sensor and making use of the redundancy of information.It is widely used in military strike,emergency rescue,satellite networking and other aspects.It is of great significance to study efficient and convenient collaborative localization algorithm.Aiming at the problem of collaborative localization with short positioning time and no prior information of each node,this paper analyzes the centralized collaborative localization algorithm and points out its defects such as complicated information exchange process and large amount of information exchange.To solve the above problems,this paper proposes a distributed cooperative localization algorithm based on field theory,and proposes a weight calculation method by deducing the error variance expression.In this method,each node only broadcasts its own location information and makes its own positioning by receiving the broadcast information of other nodes.Each node can join and exit the collaborative network at any time without affecting the positioning work of other nodes.Simulation results show that the proposed method can avoid the defects of the centralized cooperative localization method and achieve the similar localization performance.Aiming at the problem of long positioning time and certain prior information of each node,this paper firstly improves the positioning accuracy by introducing extended kalman filter,and then studies the filtering method in the case of no noise prior information at all,and proposes the minimum posterior error filtering.Simulation results show that the performance of the two filtering algorithms is better than that of the non-filtering algorithm.Compared with the extended kalman filtering method,the least posterior error filtering method can work without any noise prior information,and can achieve better performance.