Abstract
To enhance the safety and electrification of intelligent transportation, this article develops an innovative dual three-phase permanent magnet synchronous motor steer-by-wire (SbW) system. The architecture uses triple-redundant orthogonal decoupling to convert the dual three-phase motor into three two-phase redundant motors, driving the steering system through a common rigid shaft output torque. We propose a dual-layer control strategy: the outer layer uses an angle tracking controller based on non-singular fast terminal sliding mode control and hyper basis function neural network adaptive interval observer, which more accurately estimates disturbance boundaries. The inner layer is a torque synchronization controller based on actor–critic (AC) learning with adaptive mean deviation coupling, overcoming limitations of fixed proportional–integral (PI) gains and suppressing torque asynchrony from parameter mismatch, interference, and faults. AC learning reduces storage and computational complexity. Experimental results show that this strategy achieves fast torque synchronization under interference, mismatch, and fault conditions, improving angle tracking performance and system robustness.
Funding Information
  • National Natural Science Foundation of China (52372390)
  • National Natural Science Foundation of China (52405112)
  • Anhui New Energy Vehicle Integration Technology Industry Innovation Research Institute (202305C12020003)

This publication has 32 references indexed in Scilit: