The intelligent flexible manufacturing system with the characteristics of many varieties, small batch, customized production represent the future development direction of modern manufacturing industries. And the industrial robot, which is a typical highly flexible motion platform, will play a role of “catalyst†in the upgrading process of the traditional manufacturing to the advanced intelligent manufacturing. However, due to the structural characteristic and the working environment, the industrial robot has a poor absolute accuracy and low long-term stability, which has limited its application in the new product lines. The existing robot compensation approach remains in the exploration of perfect kinematic model or effective estimation algorithm, which has not considered the kinematic characteristic of the industrial robot or enhanced the control accuracy or autonomous performance of the industrial robot in the industrial field.This paper aims at the problem of positioning error compensation and accuracy maintenance of the industrial robot working in the manufacturing field, and has presented a feasible, reliable and innovative graded compensation technique and an in- line self- maintenance approach. Based on the exploration of the pose error distribution rules and the mechanism of the error sources, we have developed the graded error compensation method in the levels of joint error modeling, geometrical error modeling and establishing a precise grid array in the robot joint space, the graded error compensation method could improve the absolute accuracy of the industrial robot to the level of repeatability. The basic principle, main problem and key techniques of the graded error compensation method is detailly elaborated and the main content of this paper is arranged as follows:1. Based on the two assessment indicators of the industrial robot(repeatability and accuracy), the distribution rules of the robot pose error and the mechanism of the error sources are analyzed in detail, the main error resources that influent the robot pose accuracy are divided into the following three classes: joint error, geometrical error and non-geometrical error, and a graded error compensation mechanism is established based on this.2. With respect to the joint error, an error compensation approach is proposed based on the deep analysis of the joint error rules, a compliance error compensation model as well as a transmission error compensation model of parallel mechanism in the robot are established.3. With respect to the geometrical error, an error compensation method based on the absolute positioning error is proposed, and the optimal calibration pose selection strategy is researched in order to improve the calibration efficiency. Also, a robot self calibration method based on the fixed-point constraint is researched, which has been developed and extended to be an inline thermal error compensation method for the industrial robot working in the industrial field.4. The non- geometrical error calibration method based on the grid division in the robot joint space is researched based on the structural characteristics and motion features. In order to balance the effect and efficiency of the non-geometrical error calibration method, the optimal grid division method is also researched based on distribution rules of the positioning error in the robot joint space.5. The application of the graded error compensation and accuracy maintenance method for robot off- line programming and high-precision flexible measurement of the sheet metal parts are researched. A high-precision off- line programming method is proposed based on “fault-target-pose iterative algorithmâ€. And for the high-precision high-efficiency measurement of the large sheet metal parts in the automobile plants, a flexible visual measurement platform is constructed by combining the industrial robot with the visual sensor. |