Research On Bionic Navigation Algorithm For Autonomous Return Of Small UAV

This paper focuses on the research of visual / micro inertial integrated navigation algorithm under heading constraints,puts forward the online construction method of topology map for UAV autonomous return,designs the return node recursive algorithm based on navigation topology map,and takes Dajiang M300 RTK as the experimental platform.In the case of satellite denial/control link loss,The feasibility of the algorithm is verified by the autonomous return navigation and control experiment of UAV.1.The research on visual odometer / micro inertial integrated navigation algorithm under heading constraint is carried out.In the nonlinear optimization of visual/ inertial back-end,the heading angle residual is added to the objective function,and the output of polarized light / magnetic composite compass is introduced as heading observation,which improves the positioning and orientation accuracy of the integrated system.The on-board experiment is carried out to verify the feasibility of the algorithm.In order to improve the adaptability of visual odometer to the environment,a polarization visual image reflection transmission scene separation algorithm is proposed.It can improve the accuracy of target detection,image matching and visual navigation in the case of transmission / reflection scene superposition.2.An online construction method of topology map for UAV autonomous return is proposed.The measurement of space environment by topology map is realized through the identification,matching and positioning of scene nodes.A topology node design and node connected edge construction rule are given.For small UAV,a node image feature expression method suitable for its autonomous return is selected,Finally,the hardware in the loop simulation experiment is carried out to verify the feasibility of the online construction method of topology map and topology node3.A recursive algorithm of return node based on navigation topology is designed.The algorithm uses the navigation topology map information constructed in real time during the journey,stores the multi-source information of topology nodes,constructs the visual word bag model,realizes the node recurrence during the return journey,carries out loop detection and image pose PNP estimation,and corrects the navigation and control errors.In this paper,the algorithms of UAV’s homing path planning and multi position correction are designed based on different homing distance map,and the correctness of the algorithms of UAV’s homing path planning and multi position correction is verified based on different homing distance map,The hardware in the loop simulation experiment is carried out to smooth the trajectory and speed of the UAV in the return process.