Research On Navigation Algorithm Of UAV Based On Multi-sensor Fusion

In recent years,with the advantages of small size,strong mobility and low cost,UAV has been widely used in the fields of enemy investigation,terrain survey,security monitoring and disaster rescue.In order to adapt to such a complex application scenario,UAV can not reliably complete the task by relying on a single sensor.How to make the UAV with multiple sensors to achieve reliable positioning and navigation in the unknown indoor and outdoor environment has become an urgent problem to be solved.The research results in the thesis are of great reference value in engineering application.The thesis focuses on the positioning and navigation algorithm of UAV in indoor structured environment and outdoor scene that GPS is degraded or denied,and based on the secondary development of open source flight control pixhawk,a set of small four-rotor UAV verification platform is built.The core sensors include lidar,single point lidar,GPS module(with magnetometer)and IMU.For indoor,the lidar SLAM(Simultaneous localization and mapping)algorithm which belong to graph-based optimization is used to complete the synchronous positioning and map construction of the UAV in the XY plane,and threedimensional positioning can be achieved,after the single point laser measured the height of the UAV in the Z axis.A correction scheme combined with IMU is proposed to solve the threeaxis positioning distortion caused by UAV tilting.For outdoors,the traditional multi-rotor UAV is mostly based on GPS for positioning,which can not accurately estimate the position of the UAV in the environment where there is no GPS signal or only degraded GPS signal.Therefore,the thesis uses GPS module and lidar for integrated navigation and data fusion based on extended Kalman filter.In order to improve the accuracy of fusion positioning,a method to evaluate the quality of GPS data is designed to eliminate the GPS data with large errors.Secondly,the navigation system architecture of the thesis is established based on ROS Navigation,and according to the actual application scenario of UAV,an improved A* algorithm is proposed for global path planning.Finally,flight tests are carried out in indoor rooms,indoor corridors and outdoor near buildings.The experimental results show that the navigation algorithm designed in the thesis can provide accurate indoor positioning information for the UAV,eliminate the cumulative error,and can correct the positioning distortion caused by UAV tilting very well,and in the outdoor environment where GPS is degraded,it can provide accurate and continuous positioning information for the UAV and reliably complete the navigation mission.The global path generated by the improved A* algorithm is shorter and smoother,which is beneficial to the control and smooth flight of UAV.