In the research project 'LaneS', a methodology was developed to reliably derive lane-specific road network graphs from standard GNSS-data on motorways. Based on that, 'LaneS - C-ITS' focuses on creating lane-specific road network graphs and traffic messages in complex, urban street networks for the purpose of C-ITS services and automated driving. Lane-specific traffic messages, which are georeferenced using highly precise map data, are relevant for automated driving, as they deliver a priori information on road and traffic conditions and thus facilitate the calculation of adequate driving strategies. Moreover, the predictive calculation of vehicle trajectories is supported, which helps to improve the calibration of passive safety systems of automated vehicles. Traditional map vendors in the navigation market (HERE, TomTom) have identified the demand for lane-specific road network graphs as important data source for automated driving and are currently conducting extensive efforts to develop respective data sources. However, the public sector, infrastructure operators, service providers or research facilities do not have comparable digital street maps. Thus, they are missing a map reference with the required lane-specific granularity for applications in automated driving, for example in order to apply traffic control strategies or communicate traffic messages with precise localization. Specifically, providers of C-ITS solutions (e.g. LOI-partner Siemens) are in high demand of lane-specific road network graphs for their future applications in automate driving in order to precisely reference cooperative messages.
The graph integration platform GIP does not fulfil requirements concerning lane-specific topological accuracy to date, while the GIP 2.0 standard already acknowledges the necessity of the lane-specific localization and referencing of ITS-measures. The identified gap between available digital road maps and requirements of maps for C-ITS services in automated driving is aimed to be closed in the project 'LaneS C-ITS'. The focus of the data-driven methodology to derive lane specific road network graphs lies on complex road situations in urban areas (e.g. intersections) as well as on dynamic changes of traffic infrastructure (e.g. temporary construction works), as according a priori data is most needed for the support of automated driving strategies in such situations. In this context, the following innovation areas are elaborated in the project:
- Automated, data-driven development of lane-specific road network graphs based on vehicle sensors, which are transmitted in standard C-ITS formats (CAM)
- Facilitating the widespread availability of required data
- Improvement of positional accuracy of input data for creating the lane-specific road network graph through fusion of vehicle sensor data and GNSS-data
- Mitigating GNSS inaccuracies
- Automated, data-driven construction of lane-specific road topologies in complex road situations (e.g. multilane intersections in dense urban road networks)
- Lane-specific road network graphs can be derived in urban areas, resulting in a area-wide applicability of the methodology
- Providing lane-specific road network graphs as Webservice for the usage as a highly precise referencing basis for C-ITS services and (highly) automated driving
- Development of a methodology to create lane-specific traffic messages in the context of traffic management/C-ITS and (highly) automated driving.
The developed prototypic service platform for creating and providing lane-specific road network graphs can be accessed and used by C-ITS-providers and public authorities as a Webservice. Especially the LOI-partner Siemens (former partner in the project 'LaneS') has identified great potential for improving their traffic-related applications through lane-specific services.