智能驾驶道路高精地图基础数据规范研究

Research on basic data specification of high-precision road map for intelligent driving

针对目前部分智能驾驶语义信息存在短缺,高精地图检索效率不足的问题,该文结合浙江省内外6千多公里的道路高精地图数据实际生产项目,提出了一种相对统一协调的、相对基础的、相对共性的智能网联汽车道路高精地图基础数据规范和数据模型。该基础数据模型将高精地图基础数据划分为8大图层,包含17种要素大类,162种要素小类,并在基础数据模型方面针对要素属性进行特别设计,除定义143项基本属性外,新增了64项关联属性,形成15类关联规则,最大化地覆盖道路要素全集。通过车商、图商在封闭道路与开放道路的实地测试应用,充分验证了基础地理数据规范与基础数据模型的可行性、有效性、灵活性,能够在一定程度上辅助智能驾驶车辆对复杂道路环境的认知,支撑智能驾驶应用,实现车路的有效协同。

This paper combines the actual production project of road high-precision map data of more than 6,000 kilometers inside and outside Zhejiang Province, and puts forward a relatively unified and coordinated, relatively basic and relatively common base data specification and data model for road high-precision maps of intelligent networked vehicles in response to the shortage of semantic information of some intelligent driving and the insufficient retrieval efficiency of high-precision maps at present. The base data model divides the base data of the high-precision map into 8 major layers, containing 17 major classes of elements and 162 subclasses of elements. In terms of the base data model, it is specially designed for the attributes of the elements, and in addition to defining 143 basic attributes, 64 new associated attributes to form 15 types of association rules, maximizing the coverage of the full set of road elements. Through the field test and application of car dealers and map dealers on closed roads and open roads, the feasibility, validity and flexibility of the basic geographic data specification and basic data model have been fully verified, which can, to a certain extent, assist the cognition of intelligent driving vehicles on the complex road environment, support the application of intelligent driving, and realize the effective collaboration between vehicles and roads.