Visual Servoing of Flying Robot Based on Fuzzy Adaptive Linear Active Disturbance Rejection Control

摘要

Visual servoing is the most common control scheme for military and civil robots. For the flying robot, a nonlinear and strongly coupled system, the traditional control methods have the problems of poor adaptive ability and sensitivity to the disturbance in the complex environment. This brief deals with the disadvantages of slow response speed and weak anti-interference ability of flying robot visual servoing by using a fuzzy adaptive linear active disturbance rejection control (FLADRC) scheme. To improve the response speed and robustness of the system, the effect of the controller parameters on the performance is analyzed and quantified into fuzzy rules, and the fuzzy adaptive method is used to adjust the parameters of the linear extended state observer in real-time. Furthermore, a high-speed and high-precision visual positioning algorithm based on YOLOv4 and efficient large-scale stereo matching (ELAS) is proposed. Finally, the performances of FLADRC, LADRC, PID, and Fuzzy-PID are compared by simulation experiments, and the effectiveness of the proposed scheme is proved.