Dissertation > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Robotics > Robot

Variable Structure Control for Visual Servoing with Eye-in-Hand Robotic System

Author XuPeng
Tutor LiHuiGuang
School Yanshan University
Course Control Theory and Control Engineering
Keywords Visual servo RBF-DDA Variable Structure Control Depth estimation No calibration MOTOMAN robot
Type Master's thesis
Year 2007
Downloads 215
Quotes 2
Download Dissertation

The robot visual servo control is a basic research issue has important theoretical significance and broad prospects for industrial applications . This article first proposed an optimized RBF-DDA neural network - based camera calibration technology , optimized RBF-DDA neural network capable of adaptive determine the hidden layer nodes and the central value of the Gaussian function , to overcome the excessive dependence of the original algorithm posteriori knowledge of design parameters ills , this optimization neural network used in camera calibration in the algorithm is simple and flexible , with high precision . Secondly, eye-hand robot visual servo system using local feature , that feature points as the image feature information to achieve the image - based visual positioning . Designed two image - based control , a visual servo controller using variable structure control design , it gives the camera panning and rotating speed , and the use of adaptive update laws given depth trends , making the system has asymptotic stability within a wide range . Another case of no robot visual servo system parameters and dynamic structure of the system without prior knowledge calibration Jacobian matrix use of least squares estimation , and from the point of view of control , variable structure controller is designed , image characteristics of convergence to the desired value . Finally, based on the build MOTOMAN_UP6 robot visual servo control system hardware platform for theoretical research and simulation to create visual servo control system hardware environment . This experimental device based on RBF-DDA neural network optimized camera calibration , the experimental results demonstrate the accuracy and rapidity of the method . At the same time remote operation is developed on the basis of this hardware platform , the visual image-based closed-loop positioning experiments visual positioning , and analysis of the problems in the system , and to lay the foundation for the further development of the robot visual servo system .

Related Dissertations
More Dissertations