Research On Decentralized Kalman Filter Cluster Cooperative Navigation Algorithm

Multi body systems are widely used in many fields.Due to the fact that navigation capability is a fundamental capability of moving bodies,and the fact that multi body collaborative navigation has many advantages that a single moving body cannot navigate alone,multi body collaborative navigation is gradually becoming a hot research topic.Collaborative navigation technology integrates its own navigation information with observed relative measurement information and navigation information of other moving objects,corrects errors in its own navigation information,improves the accuracy and reliability of collaborative navigation systems,and thereby improves the accuracy and robustness of collaborative navigation systems,effectively solving the problems of poor positioning accuracy and weak anti-interference ability of a single moving object in complex environments.The paper analyzed the navigation principle of collaborative navigation system,established a mathematical model of cluster collaborative navigation system,designed a decentralized Kalman filter collaborative navigation algorithm,simulated and analyzed two different relative observation information algorithms based on relative distance and relative distance heading angle,developed a distributed collaborative navigation system simulation platform,and verified the positioning accuracy and correctness of the collaborative navigation algorithm through simulation.The main research content is as follows:Firstly,the principle of cooperative navigation technology,cooperative navigation mode and cooperative navigation data fusion structure are studied,and the moment parameter expression and information parameter expression of Gaussian filter and Gaussian filter are derived.A mathematical model of the cluster collaborative navigation system was established,and the motion and observation equations of the robot cluster system were derived based on the robot as the research object.Secondly,a decentralized Kalman filter collaborative navigation algorithm was designed.Adopting a decentralized data fusion structure,a model based on relative distance information and a model based on relative distance heading angle information were established.Derive the relative distance information model and the state equation and measurement equation based on the relative distance heading angle information model,determine the collaborative information of nodes,and conduct simulation analysis of the algorithm.Once again,analyze the observability and collaborative positioning capabilities of two types of collaborative navigation systems that observe information.Analyzed the observability of collaborative navigation systems based on relative distance and relative distance heading angle as observation information,analyzed the impact of different relative measurement information and positioning capabilities on the positioning accuracy of collaborative navigation systems,and conducted simulation verification.Finally,complete the development of a distributed collaborative navigation simulation platform and validate the collaborative navigation algorithm through simulation.We have designed a collaborative navigation simulation platform scheme,developed the hardware and software of the platform,and verified the collaborative navigation algorithm through simulation.