Research Of Robot Servo System Based On Visual Tracking

At present, with further research of the robot technology, more and more industrial robots are used in manufacturing industry. Traditional industrial robots are used in “teach and playback” way, repeating preset actions in a fixed environment. they can not be able to work regularly when the operating environment or the job object changes. It is a challenging subject that how to improve the ability and intelligent level of industrial robots to adapt to environmental change. It has certain scientific significance and market prospects. Robot visual servo is a technology which uses the camera to obtain the variable quantity of the visual information as input to control the robot joint motion. And it can improve industrial robot’s ability of perceiving the surrounding environment change and broaden its application range. Combined with common positioning and tracking engineering problems, the robot visual servo system was discussed. In this thesis, its main work was as follows..First, three axis robot arm motion platform was designed and implemented. Based on the analysis and research of the controller, servo system and mechanical structure, the working platform of the manipulator arm with three-dimensional space movement ability was designed. And the acceleration and deceleration control and interpolation algorithms of the control system were derived and analyzed. The whole debugging proved that the platform could work normally, and completed the task of linear interpolation and arc interpolation.Second, machine vision subsystem was designed. Through selection and analysis of light source, camera and lens, the hardware platform of the machine vision was designed. In the aspect of software, the image processing algorithms such as image filtering, threshold segmentation, morphological processing and image moment calculation were studied. The effective image features for visual servo system were extracted at last.Third, the visual servo control algorithm without calibration was designed. A positioning algorithm based on Jacobi matrix was designed to locate the stationary object with a single eye Eye-in-Hand visual servo structure. In view of the problem of moving object tracking, a nonlinear relation model was established from image feature space to joint motion space, and a tracking algorithm based on the BP neural network was designed. Finally, the robot visual servo system was improved combining with the software development technology, and the feasibility of positioning and tracking algorithms were verified.To sum up, this thesis completed the development of visual servo system based on the related theoretical research and engineering practice, it realized the function of positioning and tracking, and laid the foundation for the further research. The anticipated goal of the subject research has been achieved.