Autonomous Quadrotor Landing Control System Design

Rotorcraft has been widely studied in various fields,including the autonomous landing technology of rotorcraft.In order to realize the accurate and stable landing of rotorcraft on the unknown platform and improve its autonomous decision-making ability,as well as to improve China’s maritime military combat capability,the rotorcraft can land on the ship with high precision and low cost,Therefore,how to design an autonomous landing control system has become a hot topic.First of all,this paper studies the X-type quadrotor autonomous landing control system,including the control system designed by traditional control method and the quadrotor autonomous landing control system based on Deep Deterministic Policy Gradient(DDPG).According to the X-type structure of the quadrotor,the mathematical model of the quadrotor is established,including the definition of the coordinate system,the dynamic model,the kinematic model and the relationship model between torque and speed,and the system simulation experiment is carried out.The correctness of the quadrotor model is tested by pitching,rolling,yawing and hovering actions,which lays the foundation for the design of the controller.Secondly,the landing control of quadrotor is realized by using control method.The basic concept of traditional control algorithm,the design method of controller,the simulation process of building model,and the parameter tuning process based on time domain method are introduced respectively.Finally,the wind disturbance model in the natural environment is added to control the quadrotor to land on the platform in the presence of wind disturbance,which is in line with the environment of the landing platform of the marine UAV.Finally,based on the model free deep reinforcement learning theory,an autonomous landing controller for X-type quadrotor is designed.It includes the introduction of reinforcement learning related theoretical basis and Deep Deterministic Policy Gradient algorithm,the theoretical basis model and application process of DDPG algorithm.The simulation experiment uses Tensorflow framework based on Python language to build quadrotor model and algorithm framework.Through network training,the final quadrotor autonomous landing system is obtained,The Deep Deterministic Policy Gradient is successfully applied to the quadrotor model,so that the UAV can land on the platform stably at a low speed with high accuracy.