Research On Relative Navigation Technology Of UAV Based On Multi-source Information Fusion

Intelligent swarming is an important development direction in the field of UAV.Relative navigation and positioning,as the core technology,is the key for UAV to complete its mission.This includes two meanings.One is to obtain the relative navigation information between UAVs to ensure the consistency and safety of autonomous formation flight.The other is that multiple UAVs can complement each other by configuring different sensor loads to locate the target,which can improve the mission capability and overall efficiency of UAVs.Aiming at these two problems in multi-aircraft cooperative relative navigation,this paper mainly studies the relative navigation problem based on multi-source information fusion on the basis of comprehensive analysis of the requirements of relative navigation system,aiming at improving the comprehensive performance of UAV relative navigation system.Firstly,the SINS/DGPS integrated relative navigation is studied,and an inertial/satellite high precision relative navigation algorithm is proposed.The relative state recurrence model of inertial navigation system is studied,and the relative attitude quaternion and relative velocity error equation are deduced.Then,for the satellite navigation system,the GPS model based on mobile reference station and the method of ambiguity resolution are studied.Based on the fully parallel and distributed system framework,the fusion structure of two-stage filter is adopted.This method can effectively improve the accuracy of inertial/satellite integrated relative navigation system.Then,in order to further improve the reliability of UAV relative navigation,angle and range information is introduced to assist SINS/DGPS combination.The relative positioning model of passive/active navigation system is studied,and a fusion algorithm based on double-layer nested structure is proposed,the simulation analysis is carried out.Following,aiming at the relative positioning between UAV and non-cooperative targets,this paper studies and analyses the idea of Levenberg-Marquardt optimization and improves the iterative Kalman filtering algorithm,and proposes a LM-IEKF algorithm.Considering the problem of multiplicative noise caused by signal correlation characteristics,a Modified LM-IEKF method based on adaptive correction of measurement noise is further proposed,which improves the accuracy performance of the system.In order to improve the adaptive ability of the relative navigation system in the case of GPS rejection,the all-source relative navigation algorithm based on multi-sensor cooperation is studied.An adaptive reconfigurable filtering algorithm based on centralized fusion architecture is proposed to meet the requirements of compatibility,flexibility and fault tolerance of all-source relative navigation system.At the same time,a multi-model filtering algorithm based on fully parallel and distributed fusion architecture is proposed,which realizes the flexible switching between multiple models of relative navigation system.The method of residual detection can effectively isolate the failure sensors or measurement outliers and ensure the quality of navigation information.Finally,a semi-physical experiment platform and a comprehensive simulation software are designed for the relative positioning of SINS/DGPS integrated and non-cooperative targets respectively.The algorithm proposed in this paper is validated and evaluated effectively,which provides an important reference for the application of multi-source information fusion technology in the field of UAV relative navigation.