汽车半主动座椅悬架自适应模糊神经滑模控制  

Adaptive neural-fuzzy sliding mode control for semi-active seat suspensions

作　　者：贾继良 赵清海[1] 杨景周 陈满 JIA Jiliang;ZHAO Qinghai;YANG Jingzhou;CHEN Man(College of Mechanical and Electronic Engineering,Qingdao University,Qingdao 266071)

机构地区：[1]青岛大学机电工程学院,山东青岛266071

出　　处：《机械设计》2024年第4期28-35,共8页Journal of Machine Design

基　　金：国家自然科学基金资助项目(51875396)。

摘　　要：针对含有人体模型的5自由度座椅悬架系统,设计一种基于自适应模糊神经网络(Adaptive Neuro-Fuzzy Inference System,ANFIS)的滑模控制器(ANFIS-SMC)。首先,设计一种时变滑模面,通过调整其斜率使系统状态点快速到达滑模面,从而提升系统控制速度;其次,通过ANFIS对切换增益在线调节,用切换项消除干扰,实现控制器在复杂多源干扰下精确控制;最后,仿真验证采用双曲正切函数代替切换项中的符号函数,使得输出更连续,有效降低抖振。仿真结果表明:该控制器能有效提高系统的跟踪性能和响应速度,对不确定性干扰具有较好的鲁棒性,带有该控制器的座椅悬架乘坐舒适性明显改善。A sliding mode controller(ANFIS-SMC)based on Adaptive Neuro-Fuzzy Inference System(ANFIS)was designed for a 5-DOF seat suspension system with a human body model.Firstly,a time-varying sliding mode surface is designed.By adjusting its slope,the system state points can reach the sliding mode surface quickly to improve the system’s control speed;Secondly,ANFIS is used to adjust the switching gain online,and the switching item is used to eliminate the interference to realize the precise control of the controller under the complex multi-source interference;Using the hyperbolic tangent function replace the sign function in the switching term,making the output more continuous and effectively reducing chattering.The simulation results show that the controller can effectively improve the system’s tracking performance and response speed and is robust to uncertain interference.The seat suspension ride comfort with the controller is improved.

关 键 词：半主动座椅悬架 滑模控制 自适应模糊神经网络 鲁棒性 

分 类 号：U463.8[机械工程—车辆工程]