Research On Autonomous Landing System Of Rotor UAV Based On Visual Inspection

With the development of the drone industry,rotary-wing drones are widely used in various industries.Due to the limitation of battery energy,the UAV needs to land in the hangar at a fixed point in time to replace the battery when performing a flight mission.Since the traditional GPS-based UAV landing method cannot achieve the autonomous and precise landing of the UAV,the autonomous landing of the UAV based on visual positioning has been widely studied by scholars.This paper studies the autonomous landing of rotary-wing UAVs on the basis of visual positioning.The main work of the paper is as follows:(1)Combined with the research status of UAV visual positioning and autonomous landing at home and abroad,the overall scheme of the autonomous landing system of rotary-wing UAV based on visual inspection is proposed.(2)In order to facilitate the research on autonomous and precise landing of rotarywing drones,based on the DJI Mobile SDK,an Android software was developed to realize the detection and recognition of landing icons and the control of drone posture and attitude during the autonomous and precise landing of rotary-wing drones.(3)In order to reduce the difficulty of visual positioning during the landing of the drone,a simple landing icon consisting of a black square,a white circle and a black heading reference line was designed and detected and identified.According to the detection and recognition results,the position and attitude deviation of the rotor UAV during the landing process is solved.The running speed of the landing icon detection and recognition algorithm in Android software is less than 60 ms per frame.(4)In order to improve the landing accuracy of the UAV,a control algorithm of the horizontal position control system of the rotary-wing UAV is designed.The outer loop position controller is a fuzzy PID controller,and the inner loop speed controller is a PID controller.In order to facilitate the simulation of the control algorithm,a UAV simulation platform was built based on the DJI Assistant2 software,and the control algorithm was deployed to the Android mobile phone.The simulation of the control algorithm shows that,compared with the classical PID control algorithm,the control algorithm designed in this paper reduces the overshoot and adjustment time of the system,and has good anti-interference to the wind speed.(5)Several experiments have been carried out on the autonomous landing system of the rotor UAV based on visual inspection,and the landing error is within the range of 0.13 m,which meets the landing accuracy of the rotor UAV in the hangar.