基于RBFNN观测器的车辆非线性悬架预测控制方法  

Predictive Control Method of Vehicle Nonlinear Suspension Based on RBFNN Observer

作　　者：江深[1] 张海兰 JIANG Shen;ZHANG Hai-lan(Guangdong University Of Science&Technology School of Mechanical Electrical Engineering,Dongguan Guangdong 523000,China;School of Mechanical Electronic and Information Engineering,China University of Mining and Technology(Beijing),Beijing 100083,China)

机构地区：[1]广东科技学院机电工程学院,广东东莞523000 [2]中国矿业大学(北京)机电与信息工程学院,北京100083

出　　处：《计算机仿真》2022年第1期102-105,123,共5页Computer Simulation

摘　　要：传统车辆非线性悬架预测控制方法存在车身垂直加速度和能量消耗较大、悬架动扰度和车轮动载荷较高问题,悬架控制效果不理想。于是提出基于RBFNN观测器的车辆非线性悬架预测控制方法。构建四分之一非线性悬架模型,通过改进粒子群算法中的数据聚类和参数辨识,构建线性非段仿射(PWA)模型。结合RBFNN观测器和多模型控制理论研究PWA模型的滚动时域优化控制问题,获取最优控制信号,实现车辆非线性悬架预测控制。仿真结果表明,所提方法能够有效降低车身垂直加速度和能量消耗,悬架动扰度和车轮动载荷均较低。Traditionally, vehicle nonlinear suspension predictive control methods have defects, such as fast body vertical acceleration, high energy consumption, large suspension dynamic disturbance, wheel dynamic load and poor suspension control effect. Therefore, a vehicle nonlinear suspension predictive control method based on RBFNN observer is put forward in this paper. Firstly, a quarter nonlinear suspension model was built. Secondly, according to the results of data clustering and parameter identification in the improved particle swarm optimization algorithm, the linear non-segment affine(PWA) model was also constructed. Then, based on RBFNN observer and multi-model control theory, the rolling time-domain optimal control problem of PWA model was systematically investigated to obtain the optimal control signal. Finally, the vehicle nonlinear suspension predictive control was achieved. The simulation results show that this method not only has low vertical acceleration and energy consumption, but also reduces suspension dynamic disturbance and wheel dynamic load.

关 键 词：观测器 车辆 非线性悬架 预测控制 改进粒子群算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]