Studying drift maneuvers helps improve vehicle handling and broaden the dynamic performance envelope of autonomous vehicles. This paper presents a hierarchical drift control scheme for distributed drive electric vehicles, based on terminal sliding mode control (TSMC). A saturation velocity planner is developed to transform pose errors into bounded reference states, thereby improving stability in drifting trajectory tracking. Subsequently, a TSMC is designed to track these reference states, leveraging its finite-time convergence and robustness. To further improve robustness and adaptability, an radial basis function neural network (RBFNN) approximator is incorporated into the framework for real-time compensation of modeling inaccuracies and unmodeled dynamics. In addition, an optimization-based inverse vehicle dynamics model is employed to map the desired state derivatives to vehicle control inputs. Lastly, co-simulation experiments in Simulink and CarSim demonstrate that the proposed method outperforms baseline control schemes in trajectory tracking and drifting stability.
Sio et al. (Tue,) studied this question.