您的位置:首页  > 论文页面

基于微雷达检测器的行人检测及轨迹重构方法研究

发表时间:2016-10-31  浏览量:1824  下载量:713
全部作者: 倪颖,李旭红
作者单位: 同济大学道路与交通工程教育部重点实验室
摘 要: 微雷达检测器是一种可直接埋设于路面中进行主动行人检测的新型检测器,可弥补其他检测器受光线和天气影响及检测精度低的缺点。本研究利用美国加州丹维尔市微雷达检测器试点路口的数据,开发一种行人轨迹重构的分析方法,即分析不同对象—行人/自行车—之间的轨迹特征;通过“标准轨迹平移”法估计全部对象的过街轨迹,挖掘轨迹特征。在时空轨迹基础上,结合行人/自行车过街的速度特征,实现过街对象的分类统计。最后,利用VISSIM仿真平台,分析行人流量对检测算法的影响趋势。研究表明,随着人行横道处行人流量的增加,检测精度下降幅度在10%以内,总体检测精度在75%以上。
关 键 词: 交通运输系统工程;行人检测;微雷达检测器;轨迹重构;仿真评价
Title: Micro-radar-based pedestrian detection and trajectory reconstruction
Author: NI Ying, LI Xuhong
Organization: Key Laboratory of Road and Traffic Engineering of Ministry of Education, Tongji University
Abstract: Micro-radar detector is a kind of new paved-in detector that can detect pedestrians at crosswalks, which compensates the disadvantages of other detectors, for example, video-based detectors are easily handicapped due to light and weather conditions and low accuracy of microwave detectors. Based on data collected at a demonstration crosswalk with micro-radar detectors installed in Danville, CA, a methodology that reconstruct pedestrian/bicyclist trajectories is proposed. On the basis of information collected by micro-radar detection, we could gain space-time scatter diagram. Given by different characteristics between pedestrians and bicyclists, standard trajectory curves can be generated, by comparing the shifted standard trajectories and data collected, trajectories can be reconstructed. Further, analysis of the impact of pedestrian flow on the detection algorithm by VISSIM is demonstrated. The research shows that the detection precision is reduced in the range of less than 10%, with the increase of crossing pedestrian flow, the overall accuracy is above 75%.
Key words: transportation systems engineering; pedestrian detection; micro-radar detector; trajectory reconstruction; simulation
发表期数: 2016年10月第20期
引用格式: 倪颖,李旭红. 基于微雷达检测器的行人检测及轨迹重构方法研究[J]. 中国科技论文在线精品论文,2016,9(20):2099-2105.
 
2 评论数 0
暂无评论
友情链接