Residual-Based Fault Detection of Abnormal Joint Running State of Industrial Collaborative Robot

As one of the most critical electromechanical devices in intelligent manufacturing scenarios, it is particularly important to study the state monitoring methods of industrial robots. The most used method currently is to train diagnostic models based on machine learning algorithms, but this approach has the problem of poor universality. Therefore, a residual-based fault detection method of abnormal running state of the UR5 robot based on the mathematical model was proposed in this paper. Firstly, research is conducted on the mathematical model of the UR5 collaborative robot to obtain signals such as position and current that can reflect the operating status. Then combined with the monitoring of actual values during operation. The residual values between the target and actual signal could be calculated to achieve the monitoring of abnormal joint conditions. The result shows that the residual of the current signal between normal and abnormal state shows an increase trend as the abnormal force magnitude increases across various trajectories. The case study demonstrates a robust methodology for accurately diagnosing the abnormal operational condition of an industrial robot.