Search

High definition (HD) map is an indispensable part of autonomous driving, and its development will greatly affect the level of autonomous driving. The existing HD map models are diverse, lacking the systems thinking of “human-vehicle-road-map”, and the vehicle-oriented consideration is not thoughtful enough. Based on the consideration of more explicit layers, more comprehensive elements and more prominent user characteristics, we attempt to refine and enrich the data logic structure of intelligent high precision map, and elaborate the contents and functions of static data layer, road real-time information layer, vehicle dynamic information layer and user model layer in detail. We point out three key technologies, the collaborative updating mechanism of “cloud-edge-vehicle”, the intersection safety information framework, and the navigation decision-making driven by traffic event. It describes how these three key technologies support the global and local path navigation decisions and the control of autonomous vehicle. Decision supports of HD map for autonomous driving are explained from four aspects, perception, positioning, decision-making, and control. [ABSTRACT FROM AUTHOR]

Objectives: The “non-visual” and machine-oriented characteristics of high definition map distinguish them from traditional human-oriented spatiotemporal products. Correspondingly, the transmission model describing the relationships between map subjects, objects, and their products also faces significant changes. Existing high definition map information transmission models have reconstructed these relationships, including the addition of user-specific information and its transmission. However, there are still shortcomings in the use of human-oriented map language instead of machine-oriented language as the information transmission tool. To address this, we combine the transmission characteristics of map information in autonomous driving and construct a machine-oriented cognitive high definition map information transmission model. Methods: We propose three extensions to the existing map information transmission model: Substituting GIS language for map language, integrating user-specific information into the user layer of high definition map, and expanding action guidance to action practice. Results: The research results show that the constructed machine-oriented cognitive high definition map information transmission model has extended the subject of map information cognition from humans to machines, adapting to the full artificial intelligence characteristics of high definition map in the transmission process of perception, localization, planning, and control. Conclusions: The proposed model contributes to accurately grasping the essence and content structure of high definition map, enhancing cognitive performance. Additionally, it plays an important guiding role in driving services, including content selection, expression methods, system functional framework design, and improving transmission efficiency. [ABSTRACT FROM AUTHOR]

Objectives: High definition map (HDM) is the “sensor” for automated driving (AD), which integrates the real-time data collected by various sensors and the prior data collected in the early stage, and serves the application of AD, including providing the necessary prior data and the corresponding over-the-horizon perception, calibration positioning, dynamic planning and decision-making and control. However, there still exists a gap between the supply of HDM data and the demand for AD knowledge, including difficulties in data retrieval due to the large volume of data and challenges in semantic understanding and intelligent decision-making due to weak data correlation. Therefore, how to balance the data supply of HDM and the knowledge demand of AD is the main goal of this paper. Methods: Knowledge graph (KG) is a representation of knowledge in a graph structure to describe entities and their relationships, involving entity extraction and relationship extraction, so that it can make the AD with interpretable, understandable and inferential. This means that KG can serve as an alternative and explicit simulation of the human mind and map cognition that driverless vehicles are missing. Therefore, we introduce KG on the basis of HDM data, and propose a framework of HDM-KG-driving task, so as to support the transformation of HDM data to AD knowledge. The construction of HDM-KG adopts the top-down method, that is, the pattern layer is first followed by the data layer. As the conceptual hierarchy of KG, pattern layer defines the concept, attribute and relation of map ontology, and it can explicitly describe the indirect information and implicit correlation information of map domain from the perspective of traffic. The data layer is an instantiation of the pattern layer, a way to populate the pattern layer through instance matching. Results: To verify the validity of the proposed method, a virtual simulation dataset based on OpenStreetMap data is constructed and converted into two formats. The first is the OpenDRIVE format, which is used to build the KG of static data, including the static knowledge of roads, lanes, intersections, road signs and road markings. The other is the data format of CARLA AD-simulator, which is used for AD simulation and the construction of dynamic real-time KG, including the dynamic knowledge of self-driving cars and vehicles in front of them, pedestrians and other traffic participants. Through the application of the resulting HDM-KG in semantic query, knowledge reasoning and local decision planning, the results show that the HDM-KG explicitly expresses the semantic relationship, enhances the application of HDM in AD, and can further improve AD by connecting additional knowledge such as safety requirements, traffic rules or scene background. Conclusions: HDM-KG can structurally and explicitly express the traffic semantics contained in map objects and the semantic relations between objects, traffic rule constraints, traffic scenes and other knowledge, and realize the three-layer evolution of data-information-knowledge of data sources-HDM-KG, providing support for semantic relation query, logical reasoning and motion planning required by driving tasks. In addition, the introduction of KG eliminates the need for driving tasks to perceive and predict objects and semantic relationships from data collected by sensors, which are explicitly defined by HDM-KG, reducing computational latency and improving safety and reliability. The top-down method of constructing KG will cost more labor and time, and the subsequent research will carry out the bottom-up method to automatically fill the KG with multi-source data. [ABSTRACT FROM AUTHOR]

Objectives: High definition(HD)map plays an important role in autonomous driving. Most existing research focuses on high-resolution mapping in terms of road geometries as well as the updating of static road information in HD map. However, there is still a lack of research on dynamic information involved in road traffic networks. We focus on enriching the content of HD map in terms of the dynamic information including the dynamic information, it's meta data structure and inforamation interaction. Methods: We propose a comprehensive set of content that HD map need to cover for the dynamic information based on existing work on HD map data organization and data standards. Based on this, combined with cloud ends of HD map systems and vehicle ends of autonomous driving vehicles as two information interaction terminals, a method for information interaction between cloud ends and vehicle ends under different combination modes (vehicle-cloud, vehicle-vehicle, and cloud-vehicle modes) used in HD map information system is proposed to facilitate a timely capture of dynamic information in road environments. Results: (1) The vehicle-cloud information interaction mode is suitable for self-driving cars. The vehicle collects vehicle dynamic information in real time and uploads it to the cloud of the HD map information system after preprocessing, so as to realize the sharing of perception information of different vehicles in the road environment on the cloud. (2)The vehicle-to-vehicle information interaction mode is used between different self-driving cars. This mode is mainly aimed at obtaining partially vehicle dynamic information directly through vehicleto-vehicle interaction during driving. (3)The cloud-vehicle information interaction mode is suitable for vehicles connected to the HD map information system platform. The cloud stores dynamic information, and the vehicle requests the demanding information according to its own needs. Therefore, The HD map information system continuously performs information interaction in three interaction modes, and updates information simultaneously, maintaining the freshness of dynamic information and enhancing the robustness of HD map. Conclusions: However, due to the complexity and variability of the road environment, the content of dynamic information interaction we propose can only cover what happens during vehicle driving to a certain extent. For the road information that appears with a small probability, we do not conduct in-depth research, which is also the direction we need to discuss and study in the future. [ABSTRACT FROM AUTHOR]