Research On Autonomous Flight Control Of UAV Based On Multidimensional Data Fusion

Because of its flexibility,convenience and vertical takeoff and landing,rotor UAV is loved by researchers and UAV enthusiasts.Benefiting from the development of intelligent chip technology,UAV can get rid of the control of ground control station to a certain extent and realize autonomous flight with the calculation of airborne computer.However,the complex flight environment puts forward new requirements for the autonomous control of UAV,which requires UAV to be able to independently complete real-time dynamic route planning,timely avoid static/dynamic obstacles,complete autonomous takeoff and landing and other actions in the process of mission execution,and should continue the challenge of accurate positioning in case of individual sensor failure.Therefore,making the UAV have the perception ability of multi-dimensional data and perfect autonomous obstacle avoidance ability is the key to realize the autonomous flight of UAV.The autonomous flight control of UAV in complex flight environment is studied.The specific work is as follows: The position and attitude data fusion technology of multi-sensor is studied,and a fusion technology is adopted to solve the problem that a single sensor affects the positioning and attitude estimation of UAV due to noise and failure.This thesis studies the technology of converting the point cloud map constructed by lidar SLAM into European signed distance field map,and puts forward the loop detection method of laser and image fusion,which ensures the global consistency of laser slam and prepares for the path planning algorithm.This thesis studies the dynamic re planning method based on trajectory optimization,constructs a high-order B-spline curve based on re planning,and dynamically re-plans the path combined with the flight constraint equation.The motion model of dynamic obstacle is studied,and a UAV dynamic obstacle avoidance algorithm based on time redistribution strategy is proposed to solve the real-time performance of dynamic re planning path algorithm.The autonomous flight control model of UAV is studied.It is proposed to use PID control model and LQR control rate to track the trajectory of UAV,so as to realize the autonomous flight of UAV.Based on the robot operating system,the experiment carries out the modeling and Simulation of UAV autonomous flight on the gazebo platform,and visualizes the experimental results with Rviz tool.The performance of data fusion algorithm and dynamic programming algorithm in the system and the reliability of UAV autonomous flight model are verified through experimental data analysis.