Discrete Prescribed Performance Adaptive Dynamic Programming for Energy-Saving Torque Vectoring Control of FWID-EVs Under Extreme Conditions

This work is dedicated to addressing the optimal torque vectoring control (TVC) issue for the path following and energy saving of the four-wheel-independent-drive electric vehicle (FWID-EV) under extreme conditions. First, an improved cubature Kalman filter (CKF) is designed to observe the vehicle sideslip angle. In particular, a fractional-order difference (FOD) operator is developed to tackle the mismatches of the vehicle model, and an adjustable coefficient is utilized to improve the accuracy of the observer. Second, a discrete prescribed performance adaptive dynamic programming (DPP-ADP) algorithm is proposed to guarantee the transient performance and achieve the path following and energy saving of the FWID-EV by optimizing the torques of the four in-wheel motors. Third, a single parameter-guided learning (SPGL) strategy is presented to enhance the convergence of the DPP-ADP and ensure that the tracking errors are restricted in the prescribed performance bounds. The software-in-the-loop (SIL) and the hardware-in-the-loop (HIL) tests are performed under emergency maneuvers to examine the effectiveness and practicability of the proposed method, respectively.