Research On Motion Control Of Robotic Arm Based On Vision Guidance

With the advancement of science and technology and the further upgrading of the industry,manipulator with visual perception have always been a research hotspot in the field of robotics.This paper focuses on the research of the vision-guided manipulator control system,using the image feature information collected by the camera to realize feedback control of the manipulator movement,focusing on the Jacobian matrix online estimation and joint control methods.The main work and innovations of the paper it can be summarized into the following parts:1.Systematically explained the development status of the Jacobian matrix online estimation and joint control method,discussed the advantages and disadvantages of various visual control methods,studied the pose expression and coordinate transformation method of the manipulator in space,here based on the establishment of the kinematics model of the manipulator,the forward and inverse kinematics analysis are carried out,and the accuracy of the model is verified through the simulation solving program.2.Aiming at the problem of image preprocessing in the vision system,this paper proposes an image enhancement algorithm based on an improved logarithmic image processing model in a low-light environment,and compares it with the single-scale and multi-scale Retinex image enhancement algorithms,and uses the SIFT algorithm for features identify the match.based on the principle of small hole imaging,a camera model is established,the relationship between camera motion and image feature changes is analyzed in detail,and the mapping relationship between image feature changes and changes in the joint angle of the manipulator is constructed.3.Take each element of the Jacobian matrix that symbolizes the mapping relationship between the target image feature and the manipulator joint space as the system state,and establish the state equation.aiming at the image noise in the online estimation of Jacobian matrix parameters of the Kalman filter algorithm,the noise will have a greater impact on the estimation accuracy,a multi-innovation Kalman filter algorithm is proposed.by expanding the single innovation to multiple innovations to increase the amount of innovation,the Kalman gain is expanded accordingly,and a more reasonable weighted multiple innovation Kalman filter algorithm is proposed based on this,which further improves the estimation accuracy of the Jacobian matrix.4.In view of the influence of modeling errors and external disturbances in joint control on control accuracy,and the chattering phenomenon caused by traditional sliding mode control,it is proposed to introduce feedback linearization methods and fuzzy control systems on the basis of traditional sliding mode control,and establish a new control law,using fuzzy control to adaptively adjust the switching part of the sliding mode,the stability analysis of Lyapunov shows that when there are system modeling errors and external disturbances,the designed control system is gradually stable.The proposed method effectively reduces the chattering phenomenon of the system,and the tracking performance is good.