Abstract: It is one of the frontier research directions of intelligent manufacturing to use robots as manufacturing equipment executor integrating intelligent vision sensors to realize small margin grinding/milling/cutting/drilling for large and complex parts. At present, the main problem confronting in application is that the machining error is hard to reduce. The study systematically investigates the geometric error modeling and accuracy control of visually-guided robotic machining, where the machining error reduction is regarded as the main objective and the robot/visual sensor/workpiece/tool are regarded as the main targets. The first I part introduces the research status of geometric error modeling and parameter identification, studies the spatial motion chain and metric index definition of robotic machining, and derives the static error (originating from workpiece/tool pose error) transmission model as well as dynamic error (originating from joint kinematic error and joint weak stiffness) transmission model. The second II part builds the objective function and calculation method of accurate identification for hand-eye/workpiece/tool pose parameters, and proposes a general pose optimization model of robotic machining (grinding/milling/cutting/drilling) for overall error control, where the joint kinematic error and weak stiffness deformation are both considered.

Key words: geometric error modeling, spatial motion chain, pose parameter identification, machining error transmission