智能驾驶场景中高精地图动静态数据关联方法
High Definition Map Dynamic and Static Data Association Method for Intelligent Driving Scenarios
王舒曼 1应申 2蒋跃文 1张闯 1李霖 1刘经南3
作者信息
- 1. 武汉大学资源与环境科学学院,湖北 武汉,430079
- 2. 武汉大学资源与环境科学学院,湖北 武汉,430079;武汉大学地理信息系统教育部重点实验室,湖北 武汉,430079;武汉大学自然资源部数字制图与国土信息应用重点实验室,湖北 武汉,430079;地球空间信息技术协同创新中心,湖北 武汉,430079
- 3. 武汉大学卫星导航定位技术研究中心,湖北 武汉,430079;湖北珞珈实验室,湖北 武汉,430079
- 折叠
摘要
在智能驾驶场景中,对高精地图的实时动态信息和静态数据的协同调用可以支持智能驾驶系统准确地重构道路行驶场景,并针对复杂的道路环境和突发事件做出安全高效的决策.因此,动态数据与静态数据之间的关联和实时重构是实现智能驾驶车辆路径规划及决策的关键技术.针对目前高精地图模型中动静态数据关联原则共识不够造成耦合实时性弱的问题,提出并论述了动静态数据的关联原则,依此原则,且基于关系数据理论和高精地图的属性属地区块更新机制,进一步提出了高精地图动静态数据的关联和重构的强、弱两种关联方法.通过自动驾驶宏观、中观、微观3种场景决策支持应用分析了动静态数据关联方法对智能驾驶的影响,说明高精地图动静态数据关联关系的建立能够支持智能驾驶系统的规划决策和控制,为车辆实现安全、高效、舒适的智能驾驶奠定了基础.
Abstract
Objectives:In the intelligent driving scene,the collaborative calls of real-time dynamic informa-tion and static data of high definition maps can support the intelligent driving system to accurately recon-struct the road driving scenes,and make safe and efficient decisions for complex road environment and emergencies.Therefore,the correlation between dynamic data and static data is a key technique for achieving vehicle's intelligent path planning and decision-making.Methods:To solve the problem of weak real-time coupling of dynamic and static data in current high definition map model,we propose an association method of dynamic and static data within high definition map based on the association principles of dynamic and static data,and this association method depends and can be triggered by update mechanism upon custom attri-butes and specific geographic locations.Results:We propose a high definition map dynamic and static data correlation method and analyze the impacts on intelligent driving through different macro-,meso-and mi-croscales driving levels to verify the method's validity.Conclusions:The establishment of the dynamic and static data association relationship of the high definition map not only supports the planning,decision-making and control of the auto-driving system,but also lays the foundation for the realization of safe,efficient and comfortable intelligent driving of vehicles.
关键词
高精地图动静态数据关联/数据存储模型/智能驾驶决策支持Key words
high definition map dynamic and static data association/data storage model/intelligent driving decision support引用本文复制引用
基金项目
国家重点研发计划(2021YFB2501101)
出版年
2024
NETL
NSTL
维普
万方数据