摘要
为解决智能车辆在城市环境中的定位问题,使用激光雷达检测和提取圆柱形路标的中心位置,并与已知地图进行数据关联;以CyberC3智能车辆为移动平台,建立了车辆运动模型和传感器测量模型,使用扩展卡尔曼滤波算法对激光雷达和编码器数据进行融合,计算车辆的全局位姿.实验结果证明该方法可以获得较高的定位精度,解决了车辆自主导航的关键问题,并可推广应用到基于自然特征的全局定位.
Aiming at estimating location of intelligent vehicle in an urban environment, a laser scanner was used to detect and extract the position of the center of each cylindrical landmark. The result of feature extraction is associated with known map. Based on a mobile platform named CyberC3 intelligent vehicle, models of vehicle motion and sensor measurement were built accordingly. An extended Kalman filter algo- rithm completes the fusion of data from laser scanner and encoders so as to calculate global pose of the ve- hicle. The experimental results show high accuracy of pose estimation and the key problem in autonomous navigation of vehicle is resolved. The application of this method can be extended to natural feature based global localization.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2007年第6期894-898,共5页
Journal of Shanghai Jiaotong University
基金
欧盟亚洲信息技术与通信计划(CN/Asia-IT&C/002(88667))
欧盟第六框架CyberCars-2(EC-FP6-IST-028062)资助项目
关键词
智能车辆
定位
激光雷达
路标
扩展卡尔曼滤波器
intelligent vehicle
localization
laser scanner
landmark
extended Kalman filter （EKF）
