Fault Detection and Correction in Omni Bundle Robot Using EKF

Abstract

In this paper, the entire focus is on fault detection and correction using extended Kalman filter (EKF) on industrially used omni bundle robot. Here, the authors considered a problem where a robot does monotonous job and repetitively traverse the same trajectory every time. Due to wear and tear or any fluctuation in input signals, the robot may not track the desired trajectory. Faults may arise at various levels in robots, for example, in actuator, encoder, and dynamical parameters. Actuator fault detection using encoder sensor will be the main focus in this research paper. The dynamics of robotic arm link is modeled using a DC motor attached with the link, and the encoder attached with the motor will provide position feedback. The fault is detected using EKF state estimation, and the fault is corrected using a PID controller feedback into the motor input torque.