Variable Structure Control For Visual Servoing With Eye-in-Hand Robotic System

Robot visual servo control is a fundamental research subject with very important theoretical research sense and industry application prospect.Firstly a camera calibration technology based optimization RBF-DDA is presented. Optimized RBF-DDA can adaptive to determine hidden layer nodes and the central value of the Gaussian function .It overcomes the ills of excessive dependence on a priori knowledge of the original algorithm design parameters. This neural network camera calibration algorithm is simple and flexible, with high accuracy.Secondly, hand-eye robot visual servo system using points as image features was studied and realized image based positioning. There are two image-based servo control presented. The variable structure control method is applied for the design of the controller. It gives the camera translation and rotation speed and an adaptive update law is used to estimate the depth of selected features of the object. The system has made large-scale asymptotic stability. The other controller is designed without knowing the prior knowledge of the structure and dynamic calibration circumstances, Jacobian matrix using the recursive least squares estimation and a variable structure controller is designed from control view.Finally, a hand-eye visual servo system is set up based on MOTOMAN_UP6 robot, which constructed a hardware platform of the theoretical research and simulation experiment for robot real-time visual servo control. Based on this experimental device, a camera calibration technology based optimization RBF-DDA is held, the results show the accuracy and speed.At the same time, a software platform for the system was developed. An image-based closed-loop experiments are held the results verify the effectiveness of the control algorithm. Some problems that exist in the systemare analyzed. The experiment lays a foundation for the further development of the visual servo system.