Investigation of position accuracy of industrial robots and online methods for accuracy improvement in machining processes

Industrial 6DOF robots have become a tool for various machining processes due to its universality - an ability to perform any type of movement in space - and low price compared to standard 5-Axis milling machines. However, main limited factors stems from the fact that the robot accuracy with serial kinematics is much lower in comparison with CNC machines. Using the robots for machining is therefore limited to applications with lower geometrical accuracy and machining non-hard materials. In this contribution, an experimental investigation of sources of errors, having the impact on the product accuracy and surface quality, is presented. Based on this investigation focused on KUKA robots, an online method for compensating the main source of errors - backlash errors resulting from the drive reversion - is proposed. The main emphasis is put on online (real time) error compensation based on KUKA RSI (Robot Sensor Interface) technology package and sensing the actual robot positions. Experimental results given at the end of this contribution summarizes the potential for further accuracy improvements and using the robots for machining processes maintaining the accuracy in acceptable limits.