节点文献

基于ADAMS和Simulink的神经网络控制4WS汽车联合仿真研究

The Co-Simulation Research of 4WS Vehicle with Neural Network Controlled Based on ADAMS and Simulink

【作者】 岳本龙

【导师】 张天侠; 周淑文;

【作者基本信息】 东北大学 , 车辆工程, 2008, 硕士

【摘要】 汽车的四轮转向(Four-Wheel-Steering,4WS)技术作为一项先进的汽车主动底盘控制技术,是车辆主动底盘技术的重要分支,能有效的改善汽车高速行驶时的操控稳定性能,减小汽车质心侧偏角,减小车辆横摆率和横向加速度之间的相差,低速行驶时减小了车辆转弯半径,提高了车辆操纵灵活性。本文介绍了四轮转向技术研究的目的和意义,描述了四轮转向技术的基本原理及优点,分析了四轮转向控制方法的研究现状和发展趋势。从仿真技术发展的角度,分析了4WS研究存在的问题以及运用虚拟仿真技术的现实意义。为建立四轮转向汽车虚拟样机模型,研究了多体系动力学软件MSC.ADAMS的动力学理论和计算方法,并详细的介绍了多体动力学仿真软件中的一些重要模块。在轿车模块ADAMS/Car的模板系统中,通过修改一个转向系统的转向齿条,建立了后转向系统;修改车身系统使之与后转向系统连接,并与其他子系统一起建立了四轮转向整车模型。详细介绍了神经网络以及BP神经网络的有关知识,在Simulink中建立了四轮转向神经网络控制模型,取得需要的数据,确定了BP神经网络的算法以及相关参数,并训练网络模型,实现了四轮转向汽车的神经网络控制以及ADAMS与MATLAB的联合仿真,探索汽车在高速转向情况下横摆角速度和质心侧偏角等表征汽车行驶稳定性的曲线参数,肯定了四轮转向汽车神经网络控制的稳定性及可行性,提高了汽车行驶的安全性。

【Abstract】 As an advanced vehicle active chassis control technology, Four-Wheel-Steering (4WS) is an important branch of active chassis control technologies. It can effectively improve the stability of the vehicle at a high speed, reduce the body side slip angle, body yaw rate and so on. It can also reduce the turning radius and makes you drive easily when the car is running at a law speed.This paper introduces the purpose and significance of four-wheel steering technology research, describes the basic principle and advantage of 4WS technology, analysises the research status and develop trend of 4WS control method. From the simulation point of view, it analysises some existed problems on the 4WS research, as well as the practical significance of using virtual simulation technology. In order to create 4WS model, the dynamics theory and calculate method of multi-system dynamics software MSC.ADAMS are studied. It also introduces some important modules of this software. In the system of ADAMS/Car template, I created a rear steering system by modifying the steering rack of a steering system. I modified the body system so as to connect to the rear steering system. At last I created the 4WS full vehicle model using these two systems and some other necessary systems. Details of neural network and BP neural network are mentioned in the paper. I created a 4WS vehicle model with neural network controlled in Simulink, got necessary data, established the algorithm and parameters of BP neural network, and then trained the network model. At last, I made it successful to control the 4WS vehicle with neural network. The co-simulation based on ADAMS and Matlab is executed successfully. By comparing the curves parameters of yaw rate、side slip angle and some other parameters react to the stability of vehicle at a high speed, I make sure that the 4WS vehicle with neural network controlled has high stability. It makes drive more safely.

  • 【网络出版投稿人】 东北大学
  • 【网络出版年期】2012年 03期
  • 【分类号】TP183;U463.6
  • 【被引频次】7
  • 【下载频次】528
节点文献中: