Use of Extended Kalman Filter in position sensor fault detection for stepper motors

Stepper motors are often used as actuators in harsh environments such as nuclear power plants or particle accelerators due to their high reliability and relatively precise open loop positioning. Their significant disadvantage, however, is their jerky motion profile, caused by the stepping action and introducing vibrations to the actuated system. These vibrations can reduce the lifetime of the mechanics when compared to the smooth motion profile that can be achieved using closed-loop position control. In order to have the reliability of open-loop control, but the smoother motion of closed-loop control, real-time position sensor fault detection can allow sensors to be used in normal operation and, if a fault is detected, switch to open-loop. In this paper, sensor fault detection for stepper motors using the Extended Kalman Filter is studied. A novel method to detect a sensor fault is proposed and compared to a standard one.