A Novel Fault-Tolerant Control Strategy for Multimotor Servo System Based on its Redundant Mechanical Structure

Due to the redundant design in mechanical structure, multimotor servo systems have inherent advantages in dealing with motor faults. Many existing researchers utilize the estimated effectiveness factor to design the partial loss-of-effectiveness (LOE) control signals. However, the status of the faulty motor may deteriorate at any time and exhibit strong unpredictability. Consequently, working with a partial LOE motor may potentially reduce system stability and reliability. In this article, we design a fault tolerant control strategy for a representative four-motor servo system with unexpected motor fault. The faulty motor will be immediately turned off after being detected, and the other remaining motors will maintain satisfactory tracking and synchronization performance using the proposed control method. The mechanical and mathematical descriptions of the asymmetric faulty servo system with unknown inertia and unbalanced torque are first introduced. An improved command filtered backstepping control scheme is proposed to achieve high precision fault-tolerant control. The radial basis neural network is utilized to estimate the nonlinearities, and a novel finite-time command filter and error compensation subsystem without signum functions are designed decreasing the excitation of high frequency components. Experiments are designed to analyze the fault-tolerant control performance of the four-motor servo system.