The Research Of Monocular Vision Servo Modular Manipulator Control System

Manipulator is a leading edge subject for worldwide research of MEMS(Micro Electro Mechanical Systems). Modular manipulator can fulfill the requirement of easy modification, construction and expansibility; Not only routine works are required, human vision, thoughts, study and autonomous ability are specially needed in respects of national defense, industry and medical treatment. Based on the reasons above, this paper works on a deep theoretical and experimental study of modular manipulator control system based on monocular vision servo system, and thus established a modular manipulator control system based on monocular vision servo system with location variable. This paper proceeds as follows:Firstly, this paper analyzes the method to construct the matrix and the coordinate system of dimensional position and orientation, and it uses D-H method to build up a transformation matrix in order to achieve the solution of the kinematics function. In addition, it unties the inverse kinematics function using algebraic method, and conducts experiments on both kinematics function and inverse kinematics function. Statistics indicate that the error chiefly comes from the machine design and solving trigonometric functions.Secondly, this paper designs a digital image processing algorithm combining adaptive median filter and adaptive wiener filter in order to eliminate the noise of the image. The position and orientation of the target object in relation to the manipulator coordinate system is identified by using statistics algorithm. The experiments indicate that the designed algorithm of the filter has largely promote the SNR(Signal to Noise Ratio) of the original digital image. Clustering analysis in statistics is used to obtain the image features of the central pixel coordinates of the image, half axial length on horizontal axis of the vertical axis, image area and perimeter.Thirdly, this paper analyzes several pattern of manipulator visual servo. It designs the Fuzzy PID motion controller for every joint and analyzes the stability of the controller focusing on the motion planning. It designs the linear interpolation trajectory planning algorithm for the terminal of the manipulator and the fourth order three uniform b-spline interpolation path planning algorithm, and it analyzes the reliability of the algorithm via simulation. It also conducts the comprehensive motion planning and experimental analysis of the terminal of the manipulator and the modular joints. The experiments have proved the feasibility of the trajectory planning algorithm of the designed terminal and the modular joints of the manipulator.