Research On External Force Estimation And Anti-disturbance Control Of Unmanned Aerial Vehicle Using Vision

In recent years,quadrotor UAV(Unmanned Aerial Vehicle)have been rapidly applied to aerial operations,logistics and transportation,photography,agriculture and other fields.Due to the diverse and unstructured application scenarios,the external forces to which they are subjected during operation have become more complex.However,tasks such as antidisturbance control and collision detection require the use of these external forces.Therefore,it is important to study the design of estimators for immunity control and collision detection for external forces applied to UAV.However,problems such as response lag,more dependencies and limited applications pose many challenges for external force estimation.Computer vision is increasingly used in the localization,navigation and obstacle avoidance of UAV.Therefore,this paper focuses on the method of estimating the external forces of UAV by visually acquiring the poses and applies it to the anti-disturbance control of UAV trajectory tracking,with the following main research elements:(1)A quadrotor UAV system model is studied.The kinematic and dynamical models of the quadrotor UAV are derived by rigid-body transformation and Newton-Euler method,based on which generalized momentum is introduced into the quadrotor UAV.(2)An external force estimator with applied vision is proposed.First,to address the problems that the current external force estimation method of visual inertia does not consider the external moment and the inability of IMU(Inertial Measurement Unit)and vision to measure the angular acceleration of attitude,the Fast Super-Twisting Momentum Observer(FSTMO)external moment estimation method without relying on acceleration is proposed in this work.Then,a filtering method is used to fuse vision and IMU to obtain velocity information.Finally,comparison experiments with commonly used algorithms are conducted in MATLAB/Simulink and ROS(Robot Operating System)respectively,and the results show that the performance of the proposed algorithm for external moment estimation is significantly improved.(3)UAV immunity controller with the introduction of external force compensation is designed.Firstly,a hierarchical control framework for UAV trajectory tracking is designed to realize UAV trajectory tracking without external forces.Secondly,the attitude controller is improved by introducing a CSTA(Conditioned Super-Twisting Algorithm)in the attitude loop to resist input saturation,and adding a linear term and an external moment estimation term to the CSTA to obtain a LCCSTA(Linear Compensated CSTA)with compensation.then,an external force estimation term is introduced in the position controller.Finally,a comparative experiment of anti-disturbance control was conducted in MATLAB/Simulink,and the experimental results showed that LCCSTA has stronger anti-disturbance performance against external forces compared with PID and CSTA.(4)The proposed algorithm was verified in a real system.First,the UAV experimental platform was built indoors.Then,the required non-standard hardware was developed and debugged and the experimental equipment was calibrated.Finally,the external force estimation and immunity control experiments were designed and completed in the experimental platform.