Path planning for robot-assisted grinding processes

Path planning for a robot-assisted surface finishing system with an active torque controller is presented. We utilize a dexterous manipulator to attain the desired position and orientation in three-dimensional space during finishing processes. A single-axis active controller consists of a DC motor and a software observer is attached to the robot wrist and used to actuate a pneumatic hand-grinder. The torque observer is designed to sense the grinding contact force based on the driving current and output position of the motor. Zigzag and fractal paths on curved surfaces are designed for the grinding processes. In order to determine an ideal grinding condition, Taguchi's method for experimental design is utilized. We choose four grinding conditions, namely, path pattern, grinding contact pressure, tool diameter, and feed rate. Tendencies of these factors can be found from the experiments. In this research, the prototype of a robot-assisted finishing system is constructed and tested on a Tatung A330 robot. The experimental results show that the robot-assisted finishing system functions well under a variety of grinding conditions.