Huang Yingbo, Wu Jiande, Na Jing, Han Shichang, Gao Guanbin
IEEE Trans Cybern. 2022 Aug;52(8):8504-8514. doi: 10.1109/TCYB.2021.3063225. Epub 2022 Jul 19.
This article proposes a novel control method for vehicle active suspension systems in the presence of time-varying input delay and unknown nonlinearities. An unknown system dynamics estimator (USDE), which employs first-order low-pass filter operations and has only one tuning parameter, is constructed to deal with unknown nonlinearities. With this USDE, the widely used function approximators (e.g., neural networks and fuzzy-logic systems) are not needed, and the intermediate variables and observer used in the traditional estimators are not required. This estimator has a reduced computational burden, trivial parameter tuning and guaranteed convergence. Moreover, a predictor-based compensation strategy is developed to handle the time-varying input delay. Finally, we combine the suggested USDE and predictor to design a feedback controller to attenuate the vibrations of vehicle body and retain the required suspension performances. Theoretical analysis is carried out via the Lyapunov-Krasovkii functional to prove the stability of the closed-loop system. Simulation results based on professional vehicle simulation software Carsim are provided to show the efficiency of the proposed control scheme.
本文针对存在时变输入延迟和未知非线性的车辆主动悬架系统提出了一种新颖的控制方法。构建了一种未知系统动力学估计器（USDE），它采用一阶低通滤波器运算且仅有一个调谐参数，用于处理未知非线性。有了这个USDE，就不需要广泛使用的函数逼近器（如神经网络和模糊逻辑系统），也不需要传统估计器中使用的中间变量和观测器。该估计器具有降低的计算负担、简单的参数调整和保证的收敛性。此外，还开发了一种基于预测器的补偿策略来处理时变输入延迟。最后，我们将所提出的USDE和预测器相结合，设计了一个反馈控制器，以衰减车身振动并保持所需的悬架性能。通过Lyapunov-Krasovkii泛函进行理论分析，以证明闭环系统的稳定性。提供了基于专业车辆仿真软件Carsim的仿真结果，以展示所提出控制方案的有效性。
