Research On Dynamic Landing System Of Vehicle-mounted Quadrotor

With the rapid development of quadrotor UAV application technology,the quadrotor UAV autonomous tracking and landing technology for mobile platforms integrates target identification,tracking and landing decision and flight control,which has broad application prospects and important research significance.Through the preliminary study,we found that there are still some problems with this technology,firstly,the target recognition by visual method may have the problem of target failure due to partial obscuration and the problem of poor accuracy when using visual algorithm to solve the position,secondly,it is difficult to carry out stable tracking,and finally,there is the problem that the flight control deviation of the UAV in the process of autonomous tracking and landing leads to the accuracy and The problem of low success rate.In view of the above problems,the main research work of this paper includes the following aspects.First,a distributed embedded visual marker for guiding UAVs to land autonomously is designed based on Ar Uco markers,which consists of multiple small Ar Uco markers at the periphery and large Ar Uco markers in the form of nesting in the middle,and UAVs can be relatively positioned by visual recognition of them at different heights and complex environment backgrounds.Based on this,the relative position acquisition method of multiple marker position fusion is proposed for the problem of low accuracy of traditional position solution method,which improves the position solution accuracy and also solves the problem that position solution cannot be performed when the visual marker is partially obscured.Secondly,an active vision algorithm for estimating the motion speed of a moving target is designed to obtain the speed of the target by performing displacement differential between the position of the UAV before and after a period of interval and the relative position of the UAV and the target.After analyzing the shortcomings of the direct landing strategy used in previous UAV autonomous landing,a phased dynamic landing strategy is proposed to improve the success rate of UAV autonomous landing.Based on the above research basis,a string-level PID flight control algorithm with automatically adjustable control parameters is designed,which calculates parameter correction values based on the current position of the UAV and the position deviation between the UAV and the target to adjust the control parameters and improve the accuracy of the UAV autonomous landing.Finally,based on the quadrotor UAV experimental platform mounted on the unmanned vehicle,the above visual recognition and position solving algorithms and the autonomous flight control algorithm are experimentally verified.Under the experimental conditions of initial flight altitude of 6～8m and visual markers of 0.6m side length,the UAV was tested for autonomous tracking and landing relative to dynamic targets,and the UAV landing was achieved and the landing error could be kept within ±0.3m.The experimental results show that the system combined with the above methods can meet the needs of UAV dynamic autonomous landing.