基于低精度GPS智能车视觉导航方法.pdfVIP

45 2 上 海 交 通 大 学 学 报 Vol. 45 No. 2 2011 2 JOU RNAL OF SH ANGH AI JIAOT ONG UNIVERSITY Feb. 2011 : 2011) 02016 05+ 017 GPS a a b a 梁 昆, 杨 明, 王春香 , 王 冰 ( a. ; b. , 200240) : 针对室外环境提出了一种基于低精度GPS 视觉导航方法. 采用多传感器融合 方法解 决道路和路口 导航问题. 对于路口环境, 通过对道路标志( 如人行横道线) 检测, 实现路口 初 定位; 基于全局地图信息, 采用扩展卡尔曼滤波融合GPS 和里程计信息, 进行路口导航. 真实环境 下 实验结果表明该方法, 具有良好 有效性和实时性. : 智能车; 视觉导航; 道路检测; 扩展卡尔曼滤波; 路口定位 : TP 242. 6 : A Low Accuracy GPS and Vision Based Met hod for Intelligent Vehicle Navigation a a b a LIA N GK un , YAN G Ming , WAN G Chunxiang , WAN G Bing ( a. Department of Automation; b. Research Institute of Robotics, Shanghai Jiaotong U niversity , Shanghai 200240, China) Abstract: T his paper proposed a visual navigation method for outdoor environment based on low accuracy GPS. Outdoor environment consists of intersections and roads. The navigation methods for the neatly structured environment are not suitable for the environment w ith intersections. Thus, this paper applies a multisensor fusion method to navigation in the road intersections and the common roads situation. For the intersection issues, preliminary location is conducted by intersection detection ( such as zebra cross ing) . T hen, based on w orld map, an extended Kalman filter( EKF) is used for fusing the GPS and odome ter information to position in the intersections. T he experimental results in the real situation prove that the method is efficient and can w ell satisfy the realtime request. Key words: intelligent vehicle; visual navig