Quickly and accurately landing on the ground recharging station or recovery device is the most critical phase for a quadrotor UAV to achieve air-ground robots cooperation.A vision-based modelling framework for autonomous landing of quadrotor UAV on stationary platform is proposed.Composite lable landing marker based on AprilTag is designed.The vision measurement and state information from the UAV flightautonomy system are fused by Kalman filter to improve the pose estimation precision of UAV.The desired acceleration of UAV is calculated by PID controller,then converted to attitude command of the UAV,making it fly to the desired position.Multiple simulation are carried on autonomous landing system of proposed UAV under dedicated simulator,and the simulation results show that the mean landing error of UAV is less than 5 cm,the effectiveness and accuracy of UAV autonomous landing system are verified.
维普
万方数据