智轨电车多源环境感知系统

Multi-source Environment Perception System for Autonomous-rail Rapid Tram

智轨电车的多源环境感知系统是车载平台与运行环境的交互纽带,其包括基于激光雷达的障碍物感知子系统、基于毫米波雷达的前侧向障碍物感知子系统以及基于多摄像头的360度环视子系统。基于激光雷达的障碍物感知子系统采用地面分割算法、点云聚类算法和数据关联算法,实现了对运行前方障碍物的检测和跟踪;基于毫米波雷达的前侧向障碍物感知子系统采用目标检测算法和跟踪算法,实现侧向以及前向障碍物探测;基于分布式鱼眼摄像头的360度环视子系统运用图像拼接算法,实现智轨电车周围障碍物感知和预警。实车试验结果表明,该环境感知系统可有效地提高智轨电车的运行安全系数,为车辆智能驾驶提供了全面的环境信息。

The multi-source environment perception system of the autonomous-rail rapid tram is an interactive link between the vehicle platform and the operating environment,which includes the obstacle perception subsystem based on LiDAR,the front-side obstacle perception subsystem based on millimeter wave radar and the 360-degree detection subsystem based on multi-camera.The ground segmentation algorithm,point cloud clustering algorithm and data association algorithm are used to detect and track the obstacles in front of the autonomous-rail rapid tram in the obstacle perception subsystem based on the LiDAR.The target detection algorithm and tracking algorithm are used to sense the side and forward obstacles of the autonomous-rail rapid tram in the obstacle perception subsystem based on the millimeter wave radar.The image mosaic algorithm is used for perception and early warning of obstacles around the autonomous-rail rapid tram in the 360 degree detection subsystem based on distributed fish eye camera.The results of the autonomous-rail rapid tram test show that the environment perception system proposed in this paper can effectively improve the operation safety factor of the autonomousrail rapid tram and provide comprehensive environmental information for intelligent driving of vehicles.