摘要
感知设备采集的交通数据是智能交通运输系统高效运行的关键,而交通数据的准确性和时效性与感知设备的空间密度和具体位置密切相关。在交通数据采集系统中,既要获得给定精度和完整度的交通数据,又要尽量减少在路网中的感知设备布置的数量。为保证多元感知设备在满足感知性能指标的同时,能服从布设成本约束的条件并合理地布设在高速公路上,提出一种基于DBSCAN聚类的高速公路路段划分方法和基于遗传算法的路测设备优化布设方法。其次,对基于DBSCAN聚类和遗传算法的多元感知设备布设方法进行验证,构建智慧高速感知设备的仿真平台,以杭甬高速G92和沪昆高速为案例,按照行业指南将路段划分为不同感知性能等级,并针对L1和L2级路段应用优化布设方法。结果表明:L2级路段采用雷视一体机进行优化布设,检测精度达到98.5%;L1级路段选择视频检测器进行优化布设,检测精度可达84.2%。各路段的行程时间估计精度均较高,验证了所提方法的有效性。
The traffic data collected by perception devices is crucial for the efficient operation of intelligent transportation systems,and the accuracy and timeliness of traffic data are closely related to the spatial density and specific location of perception devices.In the traffic data acquisition system,it is necessary to obtain traffic data with a given accuracy and completeness,while minimizing the number of sensing devices arranged in the road network.In order to ensure that multiple perception devices can comply with the constraints of deployment costs and be reasonably deployed on highways while meeting perception performance indicators,a DBSCAN clustering based method for highway section division and a genetic algorithm based method for optimizing the deployment of road testing devices are proposed.The deployment method of multi-dimensional perception devices based on DBSCAN clustering and genetic algorithm is verified,and a simulation platform for intelligent high-speed perception devices is constructed.By taking the Hangzhou Ningbo Expressway G92 and Shanghai Kunming Expressway as cases,road sections are divided into different perception performance levels according to industry guidelines,and the optimized deployment methods are applied for L1 and L2 level road sections.The results show that the L2 road section is optimized by means of Leishi all-in-one machine,and the detection accuracy can reach 98.5%.At the L1 road section,the video detector is selected for optimal layout,and the detection accuracy can reach 84.2%.The accuracy of travel time estimation of each road section is high,which can verify the effectiveness of the proposed method.
作者
荣文
王孜健
么新鹏
李一鸣
田彬
李林茜
RONG Wen;WANG Zijian;MO Xinpeng;LI Yiming;TIAN Bin;LI Linqian(Innovation Research Institute of High Speed Group Co.,Ltd.,Jinan 250014,China;School of Information Engineering,Chang’an University,Xi’an 710064,China)
出处
《现代电子技术》
北大核心
2024年第4期116-122,共7页
Modern Electronics Technique
关键词
智慧高速感知系统
路测设备布设
DBSCAN聚类
遗传算法
多元感知
多目标优化模型
intelligent high-speed perception system
layout of road testing equipment
DBSCAN clustering
genetic algorithm
multi perception
multi-objective optimization model
