摘要
自动驾驶、车路协同等技术快速发展,其应用迫切需要高速公路信息感知、处理和发布更加准确、及时和有效,现有的高速公路机电系统技术架构和管理体制面临重大挑战。针对以上问题,提出了采用云边端协同控制的技术架构,实现跨区域、跨层级、跨业务的交通流协同控制,并搭建云控原型系统开展云边端协同能力验证。首先确定了系统时延、通信资源占用、计算资源占用等能力验证指标,提出了能力验证指标的具体测试方法。其次以高速公路车路协同典型应用场景交通事件预警和轨迹跟踪为例,搭建了云控原型系统,并将核心处理算法分别部署在云、边、端3个层级。最后基于云控原型系统,编制了测试方案,开展了时延、通信资源、计算资源等指标测试,形成测试结果。结果表明:在系统时延方面,端的时延约为250 ms,可实现百毫秒级的预警信息发布;将全部感知数据上传云平台占用大量的通信、存储和计算资源,需进行分级存储和计算。综上,云边端协同的技术架构能够适应自动驾驶和车路协同等技术发展,符合公路交通发展方向。
With the rapid development of autonomous driving,vehicle-infrastructure cooperation and other technologies,the application urgently needs more accurate,timely and effective perception,processing and release of expressway information.The existing technical architecture and management system of expressway electromechanical system are facing major challenges.In view of the above problems,the technical framework of cloud edge collaborative control is proposed to achieve trans-regional,trans-hierarchical and trans-business collaborative control of traffic flow,and a cloud control prototype system is built to verify the cloud edge collaboration capability.First,the system delay,communication resource occupation,computing resource occupation and other capability verification indicators are determined,and the specific testing methods of capability verification indicators are proposed.Second,taking traffic incident warning and track tracking in typical application scenarios of vehicle-road cooperation on expressways for examples,a cloud control prototype system is built,and the core processing algorithm is deployed at cloud,side and end levels respectively.Finally,based on the cloud control prototype system,a test scheme is prepared to test the indicators such as delay,communication resource and computing resource,and test result is formed.The result shows that(1)in terms of the system delay,the delay at the end is about 250 ms,which can realize the early-warning information release at the 100 ms level;(2)uploading all perceived data to the cloud platform takes up a lot of communication,storage and computing resources,and hierarchical storage and computing are required.To sum up,the technical architecture of cloud edge collaboration can adapt to the development of technologies such as automatic driving and vehicle-road collaboration,it conforms to the development direction of highway transport.
作者
卢立阳
朱丽丽
刘楠
刘博
LU Li-yang;ZHU Li-li;LIU Nan;LIU Bo(Research Institute of Highway,Ministry of Transport,Beijing 100088)
出处
《公路交通科技》
CAS
CSCD
北大核心
2022年第S01期154-160,共7页
Journal of Highway and Transportation Research and Development
关键词
智能交通
云边端协同
能力验证
系统时延
通信资源
计算资源
ITS
cloud side collaboration
capability verification
system delay
communication resource
computing resource
