新型视觉区域智能车辆导航控制器设计  

Controller Design for CyberCar Based on Vision Navigation

作　　者：王荣本[1] 张荣辉[1] 储江伟[1] 金立生[1] 游峰[1] 

机构地区：[1]吉林大学交通学院,吉林长春130025

出　　处：《公路交通科技》2006年第10期103-107,共5页Journal of Highway and Transportation Research and Development

基　　金：中国博士后科学基金资助项目(2004036397);吉林大学青年教师基金资助项目

摘　　要：简要介绍了基于机器视觉导航区域智能车辆(CyberCar)的导航原理和组成。首先采用逆M序列作为辨识输入信号和最小二乘算法得到车辆转向系统的系统辨识特征方程,结合预瞄运动学模型和车辆二自由度转向动力学模型,从而建立车辆基于视觉预瞄的转向动力学控制数学模型,根据线性二次型最优控制理论得到状态线性反馈的最优控制规律。通过仿真分析和试验,验证了最优控制器在CyberCar户外路径跟踪过程中平稳、可靠。The guiding principle and composition of CyberCar based on machine vision was introduced. Applying IM sequence signals as input response signals and least squares method to establish the dynamic equation for CyberCar steering system by system identification experiments combined with the preview kinematics model and two-degree steering dynamic model of vehicle, therefore the steering control mathematics model based on preview kinematics for CyberCar was established. The optimal control of a linear state-variable feedback was obtained through the linear quadratic form optimal control theory. It was proved by the simulation analysis and outdoor CyberCar experiments that the optimal controller could trace the path steadily and reliably.

关 键 词：系统辨识 最优控制器 智能车辆 CyberCar 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]