Research On Visual Servoing Control Method Of Service Robot

At present,the main application areas of robots are limited to industrial production,but people are beginning to hope that robots can serve daily life.In family life,robots must be able to sense changes in the external environment and make a series of adjustments.Vision sensor has the advantages of large amount of information,non-contact and so on.It has become a hot topic in the robot research to introduce vision sensor into robot control system.Among them,kinematics analysis of robotic arm,extraction of image feature points and methods of visual servoing control are the basis for the study of robotic visual servoing control.Therefore,it is of great practical significance to research these several aspects.Taking the service robot as the research object,the paper researched several questions including the inverse kinematics model of the 7-DOF redundant robotic arm,the extraction efficiency of image feature point and the control strategy of visual servoing and so on,which provided an effective control method for visual servoing motion of the service robot and the method was verified by experiments.Firstly,a neural network model was established to solve the inverse kinematics problem of a 7-DOF redundant robotic arm.The effects of the learning rate,the selection of excitation function,the number of neural networks,and the number of neurons in the hidden layer to the fitting precision of the final model were analyzed in detail.Based on the analysis results,the neural network model was optimized.Secondly,different algorithms were applied to extract and match the feature points of the image and the ORB(Oriented FAST and Rotated Brief)algorithm was selected in terms of matching accuracy and computational efficiency.The extraction method of feature point in the algorithm was simplified to improve the computational efficiency of the algorithm and solve the real-time problems of the robot visual servoing control.At the same time,the matching feature points in the algorithm are screened to eliminate the mismatched points and improve the positioning accuracy of the target.Then the strategy of combing global camera and local camera was adopted to achieve accurate positioning of the target object.For the image-based visual servoing control method adopted by the partial camera,a speed controller was designed and a visual servoing control system was established.In order to verify the stability of the system,a visual positioning tracking simulation experiment was designed to verify it.Finally,relevant task experiments were designed for the research content of this paper,including the visual follow-up experiment of dynamic objects and the experiment of grabbing the cups and pouring water with both arms.The feasibility and practicability of the proposed method were verified by the experimental results and data analysis.