神经网络控制在汽车主动悬架中的应用  

The Application of Neural Network Control on Automotive Active Suspensions

作　　者：王磊[1] 佘翊妮[1] 孟彬[1] 沈季胜[1] 

机构地区：[1]浙江工业大学车辆工程研究所,杭州310014

出　　处：《重庆工学院学报》2006年第5期15-18,50,共5页Journal of Chongqing Institute of Technology

摘　　要：建立了汽车悬架的两自由度1/4车数学模型,提出了一种基于RBF神经网络在线辨识的单神经元PID自适应控制方法,介绍了RBF网络在线辨识、单神经元PID控制器和控制算法,并将这种控制策略应用于汽车主动悬架1/4车模型.仿真是在Windows2000环境下用仿真软件Mat-lab6.1+Simulink进行的.仿真结果表明,这种控制策略能有效降低车身加速度的均方根值和悬架动挠度的均方根值,改善了乘坐的舒适性.Active control for vehicle suspension is one of the important research subjects in vehicle dynamics, automobile design and manufactures. This paper sets up a quarter car mathematical model of twoDOF, presents a single neural self-adaptive control method based on RBF network on-line recognition, and introduces RBF network online recognition, single neuron PID control and control algorithm. The control algorithm is applied to the active suspension by means of the control simulation software——Matlab6.1 ＋ Simulink for Windows2000. The results of simulations show that the control algorithm used can reduce the root mean square value of car body acceleration and the root mean square value of suspension dynamics deflection, so the riding quality can be greatly improved. The above research results will be valuable to the further study of suspension theory and suspension testing.

关 键 词：主动悬架 在线辩识 RBF网络 单神经元 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程] U463.33[自动化与计算机技术—控制科学与工程]