Research on Direct Yaw Moment Control Based on Neural Sliding Mode Control for Four-Wheel Actuated Electric Vehicles

Lateral instability issues such as sharp turns or emergency obstacle avoidance in extreme driving scenarios were solved using direct yaw moment control (DYC). The DYC method was structured hierarchically and included a combination of sliding mode control (SMC) and radial basis function neural network (RBFNN) to calculate the additional required yaw moment for the upper-level control. SMC effectively managed uncertainties and ensured vehicle stability, while RBFNN approximated the entire uncertainties and vehicle dynamic model. The limitations of SMC were addressed in the chattering behavior of control signals and a system model in control design. This model-free control approach for the upper level showed high control performance without relying on a system model, which effectively minimized chattering phenomena in control signals. The lower-level control was designed to distribute torque to each wheel based on the calculated yaw moment. To distribute torque reasonably, a strategy was developed considering the load ratio in the vertical direction of each wheel. The stability of the proposed method was confirmed using Lyapunov stable theory. A simulation was conducted in MATLAB/Simulink, demonstrating that the proposed controller effectively ensured driving stability by closely tracking the desired yaw rate. The controller improved rapid stabilization and convergence and ensured high precision by reducing chattering and robustness in dynamic conditions. The proposed controller functioned effectively without a detailed vehicle dynamics model owing to the implementation of RBFNN-based approximation.