An operational calibration approach of industrial robots through a motion capture system and an artificial neural network ELM

In the industry, robots’ absolute accuracy is critical. This study provides a new calibration method to increase the robot’s absolute accuracy. This method relies on a motion capture device as a measurement tool that can capture the robot’s continuous motion state in any conceivable domain. Geometric parameters are determined by conducting overall measurements and assessing each joint individually, and the Extreme Learning Machine (ELM) neural network is in charge of compensating for non-geometric error sources that are difficult or impossible to model correctly or completely. The combination of model-based parameter identification and ELM neural network–based compensation approaches is an effective solution for the correction of all robot error sources. To verify the method’s effectiveness and correctness, a six-revolute joint robot GSK RB03 is used across simulation and experiment. After calibration, the robot’s position accuracy ranges from 7.440 to 0.159 mm, and its orientation accuracy ranges from 3.073 to 0.077 degrees. The method’s practical applicability and correctness are determined.