| The premise of the quadrotor UAVs to achieve autonomous flight is able to better perceive its position and speed information.Real-time acquisition of motion status information during flight is a necessary condition for achieving stable flight and complex mission.The global positioning system(GPS)is currently the mainstream tool for the positioning of quadrotor UAVs.However,when the signal strength is unstable or even disappears due to environmental influences,the acquisition of the position information of the quadrotor UAVs becomes a top priority.The quadrotor UAVs has the characteristics of high-order nonlinearity,parameter uncertainty,strong coupling and under-actuation.The research of flight control algorithm still faces great challenges.The main task of this paper is to design a control algorithm based on visual information to solve the problem of controlling the posture of the quadrotor UAVs under the interference of GPS signals.Based on extensive reading of literature and previous research work,this paper proposes some innovative or improved control algorithms for the following problems:First,based on the assumption that the depth value of the camera is known and the target appears in the camera field of view,a vision-based backstepping control algorithm is proposed.The algorithm proposes a virtual image plane method,which solves the problem of condition numbers in the image Jacobian matrix in the mainstream method(spherical perspective model).Then,based on the assumption that the target appears in the camera field of view,a vision-based adaptive control algorithm on the basis of the previous chapter is proposed.The algorithm designs an adaptive law for the problem that the camera depth value is unknown.The attitude angle controller is designed by the backstepping method.The method can also be extended to the cases where other parameters are unknown or disturbed in the quadrotor model,which is beneficial to improve the robustness of the system.Finally,consdering the problem that the target might be out of the camera field of view in a practical flight,a vision-based input saturation control method is proposed.By a nested saturation,the algorithm first designs the control law of the height control system limits the thrust to a predefined compact set,and the positive thrust is guaranteed by setting a positive lower bound to the compact set.Then the control law of the lateral subsystem is designed to restrict roll and pitch to a compact set.This avoids large roll and pitch when the image feature error is large,and thus this helps to keep the visual target in the camera field of view.The feasibility of the proposed algorithm is verified by Matlab/Simulink simulation,and the simulation results are analyzed in detail.The simulation results show that the control methods of the vision-based quadrotor UAVs not only effectively address the problems,but also ensure the control performance,which provides an important theoretical basis for the broad application of quadrotor UAVs in real life. |
