Research On Autonomous Landing Technology Of UAV For Mobile Platform

With the development of UAV systems,the requirements for the coordination and autonomy of UAVs in performing tasks are constantly increasing.The autonomous landing of UAVs is one of the difficult problems.The core difficulty lies in positioning and on-board computing.and control etc.The research of this thesis is oriented to the autonomous landing system of the UAV under the mobile platform.By studying the visual positioning algorithm of the UAV,the trajectory prediction algorithm of the ground moving target and the flight control algorithm,a UAV autonomous landing system on the mobile platform is designed and realized.system,and the average landing accuracy does not exceed 0.3m.In view of the above problems,the main research work of this thesis is as follows:1.Design and implement a UAV visual positioning algorithm.Design a new type of marker composed of multiple markers to avoid the loss of the target due to the change of the camera’s field of view during the landing process;according to the pinhole camera model and the coordinate system transformation matrix,solve the P4 P problem,and obtain the relationship between the drone and the marker.The pose relationship of,the weighted fusion method is used to fuse the pose information calculated by multiple markers,so that the accuracy of pose estimation is improved by 2.6%;the unscented Kalman filter algorithm is used to estimate the speed and acceleration of the target movement,According to the second-order model car motion model,predict the motion trajectory of the target in a limited time in the future.2.Based on the model predictive controller,design the position controller during the landing process of the UAV,take the future trajectory prediction of the moving target as the reference information,realize the trajectory tracking control,and carry out the corresponding simulation to verify the performance and state quantity of the controller Whether it exceeds the physical limitations of the UAV;the calculation complexity of the model predictive control solution is too large to be applied in embedded devices,and the interior point method is used to optimize and solve it to speed up the solution speed;The maximum number of iterations reduces the time of iterative calculation and further speeds up the calculation speed,reducing the calculation time by an order of magnitude and ensuring real-time performance.3.Design and implement the UAV autonomous landing system,clarify the requirements of each hardware module,determine the hardware selection,and build a rotary-wing UAV experimental platform;complete the visual positioning module,moving target trajectory prediction module,position control module and The software implementation of the landing logic control module,and the use of the ROS system to connect various modules in series,combined with the on-board computer,sensor and flight controller,realize the control of the UAV by the above algorithm;based on the Gazebo software,the simulation experiment of the above system is carried out to verify Landing accuracy in a simulated environment;use the UAV platform to conduct flight tests to study the landing accuracy of the autonomous landing system in the actual environment.