基于视觉、轮速和单轴陀螺仪的清扫车定位

Localization Method of Sweeper Vehicle Based on Vision, Wheel Speed and Single-Axis Gyroscope

针对清扫车在特定工况行驶的特点,利用车辆上现有的传感器(包括摄像头、单轴陀螺仪、车速信息)进行定位.所提出的系统包括地图模块、信号源模块及融合模块.地图模块提供车道线的局部地图信息.信号源模块中的视觉模块提供车辆相对车道线的航向和距车道线的距离信息,单轴陀螺仪提供车辆的横摆角速度信息,车辆信息模块提供车速信息.融合模块融合以上信息进行定位,在长时间无卫星信号的情况下利用视觉信息和车道线的地图融合进行位置的修正可以得到分米级的定位精度,加上航向信息修正后定位精度能够得到进一步的提升.

A novel sweeper vehicle localization method was proposed using sensors available on current sweeper vehicle including camera,single-axis gyroscope and velocity of vehicle.The proposed system includes three parts:map module,signal sources module and fusion module.Local map module provides the lane map information for fusion module.Vision module of signal sources module can afford distance and heading information of vehicle relative to detected lane;singleaxis gyroscope provides yaw rate of vehicle and vehicle speed can be obtained from vehicle module.Fusion module fuses these information for localization.Research results show that the fusion of vision and lane line information to correct position can achieve decimeterlevel accuracy for a long time without satellite signal.The positioning accuracy can be further improved with the correction of the heading information from lane line.