中文摘要：

为了低成本、高效地实现自主车（ALV）可行域认知,设计了单线激光雷达倾斜扫描系统,避免水平放置造成的路面数据丢失;提出了一种基于距离与能量组合判据的可行域划分方法。筛选边缘跳变特征以及其他具有代表性的特征并将其表示出来,在求解道路特征点的同时,对环境反射能量进行特征提取,形成可行域的第二判据,实现高可靠性的边缘提取。通过实际的有阴影和障碍物的道路测试,结果表明：相比于昂贵的三维激光雷达的环境探测系统、单一判据的辨别系统,该系统可以很好地实现低成本、实时的ALV可行域认知,对道路内可行域的再划分更为准确,既减少了复杂场景的虚警、误报等不稳定现象,又实现了行道线、障碍物的可靠认知。

英文摘要：

In order to efficiently implement ALV feasible domain cognitive with low cost,tilt scanning system of single line lidar is designed to avoid data loss of roadbed caused by horizontal placement; a feasible region division method based on combination of distance and energy is proposed. Screening of edge jump features and other typical features and present,the characteristic points of the road are solved,at the same time,feature extraction on environment reflection energy is carried out and the second criteria of the feasible region are formed,achieve high reliability of edge extraction; through actual road test with shadow and obstacles,results show that compared with expensive 3D lidar environment detection system and single criterion distinguishing system,the system can achieve a good low cost and ALV feasible domain cognitive in real time,more accuraterely redivide feasible region within road,not only reduce false alarm in complex scenes,but also realize reliable cognition of lane,obstacles.