Vehicle specifications.

Electric Power Steering (EPS) systems enhance driving comfort and safety. However, their performance often degrades under varying operating conditions due to external disturbances and modeling uncertainties. Traditional control methods, which typically rely on fixed parameters or neglect disturbance dynamics, struggle to maintain robustness and adaptability across diverse scenarios. This article presents an improved control strategy integrating Active Disturbance Rejection Control (ADRC) with advanced soft computing techniques to address these challenges. The proposed method introduces two key innovations: optimizing the tracking differentiator’s speed factor using a genetic algorithm and dynamically tuning state feedback control parameters through a fuzzy inference system. This hybrid approach enhances the disturbance rejection capability of ADRC and significantly improves system adaptability and tracking accuracy. Simulation results validate the effectiveness of the proposed controller, demonstrating low tracking errors (1.875% at low speed and 1.373% at high speed) and disturbance estimation accuracy exceeding 90%. Compared to conventional controllers, the proposed method exhibits superior robustness, reduced steady-state error, and improved performance across a wide range of operating conditions. These results confirm the potential of integrating ADRC with intelligent optimization techniques for advanced control in automotive mechatronic systems.

电动助力转向系统（Electric Power Steering, EPS）可有效提升驾驶舒适性与行车安全性。然而在多变的运行工况下，受外部扰动与建模不确定性的影响，其性能往往会出现衰减。传统控制方法通常依赖固定参数或忽略扰动动态特性，难以在多样场景中维持鲁棒性与适应性。本文提出一种改进控制策略，将自抗扰控制（Active Disturbance Rejection Control, ADRC）与先进软计算技术相结合，以应对上述挑战。该方法包含两项核心创新：一是通过遗传算法优化跟踪微分器的速度因子，二是借助模糊推理系统对状态反馈控制参数进行动态整定。这种混合策略增强了ADRC的扰动抑制能力，显著提升了系统的适应性与跟踪精度。仿真结果验证了所提控制器的有效性：低速工况下跟踪误差为1.875%，高速工况下为1.373%，扰动估计精度超过90%。与传统控制器相比，所提方法展现出更优的鲁棒性、更低的稳态误差，且在宽范围运行工况下性能均得到改善。上述结果证实，将ADRC与智能优化技术相结合，可用于汽车机电一体化系统的先进控制，具备良好的应用潜力。