四轮转向车辆的径向基函数神经网络复合控制器设计

发布时间：2018-05-02 00:52

本文选题：车辆工程 + 四轮转向　；参考：《吉林大学学报(工学版)》2016年02期

【摘要】：建立了四轮转向车辆的八自由度动力学模型,并设计了四轮转向车辆的径向基函数神经网络控制器,该控制器采用直接闭环训练法以离线的方式进行训练。为了克服离线训练导致控制系统适应性差的缺点,本文同时设计了一个单神经元自适应PSD控制器,它与径向基函数神经网络控制器组成复合控制系统共同完成对四轮转向车辆的控制。结合建立的八自由度车辆动力学模型对设计的复合控制器进行仿真实验分析,结果表明,无论匀速工况还是变速工况,径向基函数神经网络复合控制器均可有效地改善四轮转向车辆的操作稳定性。
[Abstract]:The dynamic model of four-wheel steering vehicle with eight degrees of freedom is established, and the radial basis function neural network controller of four-wheel steering vehicle is designed. The controller is trained offline by direct closed-loop training method. In order to overcome the disadvantage that offline training leads to poor adaptability of the control system, a single neuron adaptive PSD controller is designed in this paper. The compound control system composed of radial basis function neural network controller and radial basis function neural network controller is used to control the four wheel steering vehicle. Combined with the established vehicle dynamics model of eight degrees of freedom, the simulation experiments of the designed composite controller are carried out. The results show that both uniform and variable speed conditions are obtained. The radial basis function neural network compound controller can effectively improve the operation stability of four wheel steering vehicle.
【作者单位】：吉林大学交通学院;
【基金】：吉林省科技厅国际合作项目(20130413056GH) 长春市科技局重大科技攻关计划项目(长科技合(2013021)号-13KG05) 清华大学汽车安全与节能国家重点实验室开放基金项目(KF14182) 吉林大学国家杰出青年基金后备人选培育计划项目
【分类号】：U469;U463.6

【相似文献】