Adaptive Speed Control of Electric Vehicles Based on Multi-Agent Fuzzy Q-Learning

Electric vehicles (EVs) are expected to account for a significant percentage of the vehicle market in the coming years. The speed regulator is the inseparable part of EVs to solve challenges such as managing battery energy or adapting to road conditions. Since the operation of the EV system is time-variant due to parametric variations and road situations, the regulation of such systems is a complex task that cannot be effectively accomplished by the conventional deterministic control methodologies. To address the control challenges corresponding to the EV, the ultra-local model (ULM) with extended state observer (ESO) is developed for the speed tracking problem of the system. The Fuzzy Q-learning multi-age system (MAS) is also adopted to adaptively regulate the gains of the ULM controller in an online manner. The robustness of the established controller has been assessed by experimental data based on the new European driving cycle (NEDC) and altering some critical parameters of the EV test system. Moreover, the real-time verifications for the tracking problem of NEDC are also conducted by hardware-In-the Loop (HiL) simulation in the EV system.