视觉导航智能车辆的路径跟踪预瞄控制  

Preview Control for Road Following of Vision Guided Intelligent Vehicle

作　　者：陈无畏[1] 李进[1] 王檀彬[1] 李碧春[2] 

机构地区：[1]合肥工业大学机械与汽车工程学院,合肥230009 [2]中国科学院合肥智能机械研究所,合肥230031

出　　处：《机械工程学报》2008年第10期277-282,共6页Journal of Mechanical Engineering

基　　金：安徽省自然科学基金(050420302);清华大学汽车安全与节能国家重点实验室开放基金(KF2007-06)资助项目。

摘　　要：视觉导航智能车辆通过机器视觉获取导航路径信息并进行自动跟踪。常规的反馈控制仅利用当前的路径信息,在复杂条件或突发状况下难以有效工作。通过对在线获取的具体路径信息进行预瞄,可以对未来的路径信息加以利用,从而提高智能车辆行驶的安全性和有效性。为此,在获取的路径信息中确定当前路径和未来路径分别用于反馈和预瞄,并设计预瞄加反馈的控制器。该控制器能够自动根据预瞄路径的弯曲程度调整智能车辆的预瞄距离和行驶速度,从而提高跟踪精度和行驶稳定性。仿真和试验结果表明,该控制器与一般反馈式控制器相比跟踪误差较小,行驶更稳定。The vision-guided intelligent vehicle acquires the road information by machine vision, and follows the road automatically. Ordinary feedback control only uses the present road information, so the intelligent vehicle can hardly work efficiently in complicated or emergent conditions. By previewing the on-line acquired road information, future navigation information can be utilized in order to improve safety and efficiency. So road information is classified into present information and future information, and they can be used in feedback and preview respectively. A preview-feedback controller is designed. The controller can automatically adjust the preview distance and the vehicle speed according to the previewed road, so tracking accuracy and driving stability can be improved. The simulation and experiments show that this controller has better accuracy and stability than the ordinary feedback controller.

关 键 词：视觉导航 智能车辆 路径跟踪 反馈控制 预瞄控制 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]