Adaptive control of dual-motor autonomous steering system for intelligent vehicles via Bi-LSTM and fuzzy methods

Control Engineering Practice(2023)

引用 8|浏览22
In order to achieve better trajectory tracking of intelligent vehicle on different road surfaces, which include underling the weather of rain and snow, an adaptive tire cornering stiffness strategy (ACS) and a trajectory tracking autonomous steering control strategy were proposed in this paper. Firstly, the tire–road friction coefficient estimator was designed to obtain the tire–road friction coefficient based on Bi-LSTM neural network, then combining the tire cornering stiffness coefficient by fuzzy controller to obtain the online tire cornering stiffness. Secondly, considering the uncertainty of tire cornering stiffness, a trajectory tracking upper-level controller based on model predictive control (MPC) algorithm was proposed to calculate the optimal front wheel angle. Thirdly, in order to better adapt to various road conditions, a dynamic equivalent stiffness model is proposed to replace the self-aligning torque model during vehicle steering, and a lower-level controller for pinion gear target position tracking was designed based on the sliding mode control (SMC) algorithm. In addition, this paper also designed a dual-motor control strategy, which cooperates with the pinion gear target position tracking lower-level controller to quickly and accurately achieve the target front wheel angle. Finally, the simulation and HiL test results show that the proposed control algorithm greatly improves the trajectory tracking accuracy and stability of the intelligent vehicle on roads with different tire–road friction coefficients.
Tire–road friction coefficient,Adaptive tire cornering stiffness strategy (ACS),Bi-LSTM neural network,Equivalent stiffness model,Autonomous steering system,Hardware-in-Loop (HiL)
AI 理解论文