The Research Of Visual Servo System On The Orthogonal Coordinate Robot Based On The Internet

Today, with the development of the society, the demand for the robot is more and more. The research of the advanced intelligentized robot is popular issue. Visual control is the importance method to intelligentize the robot, which realizes by the independent analysis of the environment information obtained by the robotic vision. Neural network can simulate the complicated function, and acts as the core of the intelligentized robotic control. It is the importance direction to make the Internet into the robotic technology. After the research of the interrelated theory, this paper first brings forward the systemic configuration of neural network based visual servo system on the orthogonal coordinate robot based on the Internet.The robotic visual control involves multi-subject, such as image processing, kinematics, dynamics, control theory and computer program etc. This paper realizes position closed loop control on the orthogonal-coordinate robot by the image information in the Internet, which is obtained by visual sensor, namely visual servo. The neural network is the core of the movement controller. The paper lucubrates the theory and realized methods of the visual servo control on the orthogonal coordinate robot. Firstly the paper summarizes the research status of the robot visual servo, the neural network and the remote supervisor. Secondly, The paper theoretically brings forward the systemic configuration of neural network based visual servo system on the orthogonal coordinate robot based on the Internet. The neuron network controller is designed instead of normal robot movement controller, because that the robot control is the nonlinear and strong coupling character. Next, the image processor, the movement controller and the remote supervisory controller are designed. The paper lucubrates the theory of the digital image manipulation. By the conclusion obtained by comparing the various methods, the optimal image manipulation method is found. The neural network program is compiled by VC. In the Internet the harmonic control between the image processor and the movement controller is realized by the user-defined protocols. After the character analysis of the remote supervisor data included in the movement parameters and the locale image data, the data transfer is completed by the multithreading technology in the Internet. Finally, an experimental system is built and used to11evaluate the performance of the robotic visual servo system. These results for position control experiments indicate that the system performance is good. It is practicable in the package, assembly and transit domain.