汽车工程 ›› 2024, Vol. 46 ›› Issue (10): 1744-1754.doi: 10.19562/j.chinasae.qcgc.2024.10.003

• • 上一篇    下一篇

基于扰动观测器的主动悬架切换控制算法研究

陈潇凯1(),陈丰1,刘向2,刘宏宇1,王笑宇1   

  1. 1.北京理工大学机械与车辆学院,北京 100081
    2.南阳淅减汽车减振器有限公司,南阳 473000
  • 收稿日期:2024-03-22 修回日期:2024-05-09 出版日期:2024-10-25 发布日期:2024-10-21
  • 通讯作者: 陈潇凯 E-mail:chenxiaokai@263.net
  • 基金资助:
    河南省重大科技专项项目(231100240300);国家自然科学基金区域联合基金重点项目(U22A2069)

Research on DOB-Based Switching Control Algorithm for Active Suspension System

Xiaokai Chen1(),Feng Chen1,Xiang Liu2,Hongyu Liu1,Xiaoyu Wang1   

  1. 1.School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081
    2.Nanyang CIJAN Automobile Shock Absorber Co. , Ltd. , Nanyang 473000
  • Received:2024-03-22 Revised:2024-05-09 Online:2024-10-25 Published:2024-10-21
  • Contact: Xiaokai Chen E-mail:chenxiaokai@263.net

摘要:

悬架控制需要实现乘坐舒适性和操纵稳定性之间的良好折中,此外还需要考虑系统的不确定性,是一项复杂的任务。本文以兼顾悬架的动力学性能指标、算法鲁棒性与成本因素为出发点,提出了一种基于扰动观测器的次优-非奇异终端滑模切换控制算法(DOB-SNTSM)。首先,以簧载质量加速度信息为输入,通过卡尔曼滤波器设计,实现了悬架动挠度和簧载质量速度的有效估计。然后,针对悬架系统中的不确定项估计,设计了一种扰动观测器,并将扰动估计值作为前馈补偿。接下来,以滑模面函数为依据,提出了一种次优-非奇异终端滑模切换控制算法,并与扰动观测器的前馈补偿相结合,共同构成一种新的主动悬架控制策略。最后,分别进行了凸包路面和平稳随机路面下的仿真和台架试验验证,结果表明,扰动观测器的引入能显著提升悬架的乘坐舒适性指标,相比经典的天棚控制、理想状态LQR方法、不带有扰动观测器的SNTSM算法,新算法不仅很好地实现各项悬架性能指标的均衡,而且能够仅利用簧载质量加速度信息就可以达到接近理想状态LQR的控制效果,同时,控制器切换方案可以显著提升算法鲁棒性。

关键词: 主动悬架, 扰动观测器, 切换控制, 次优控制, 非奇异终端滑模控制, 卡尔曼滤波

Abstract:

Suspension control requires good balance between ride comfort and driving stability, while considering system uncertainties, which is a complex task. In this paper, a disturbance observer-based suboptimal-nonsingular terminal sliding mode switching control algorithm (DOB-SNTSM) is proposed, with considerations of suspension dynamic performance indicators, algorithm robustness, and cost factors. Firstly, using spring mass acceleration information as input and by Kalman filter design, effective estimation of suspension deflection and spring mass velocity is achieved. Subsequently, a disturbance observer is devised to estimate uncertainties within the suspension system, with the disturbance estimation serving as feedforward compensation. Next, based on the sliding mode surface function, a suboptimal-nonsingular terminal sliding mode switching control algorithm is proposed, integrating with the feedforward compensation from the disturbance observer to formulate a novel active suspension control strategy. Finally, simulation and bench tests are conducted on both convex road surfaces and smooth random road surfaces. The results show that the introduction of disturbance observers can significantly improve the ride comfort index of the suspension. Compared to the SNTSM algorithm with the classical sky-hook control, the ideal state LQR method and without disturbance observer, the new algorithm not only effectively balances various suspension performance indicators but also achieves control effect close to the ideal state LQR using solely spring mass acceleration information. Additionally, the controller switching scheme significantly enhances algorithm robustness.

Key words: active suspension, disturbance observer, switching control, suboptimal control, nonsingular terminal sliding mode control, Kalman filter