2. Kolloquium Straßenbau - September 2021 305 Consequences of connected and automated driving to physical and digital high-level road infrastructure Sandra Ulrich ARND IDC GmbH & Co KG, Vienna, Austria Risto Kulmala Traficon Ltd, Espoo, Finland Abstract The development of automated driving functions in vehicles and road network operation automation is progressing steadily. As amendments to infrastructure are costly and timely there is a strong need for research for them to be planned accordingly. The CEDR-funded project MANTRA has investigated the consequences of automated vehicles on physical and digital road infrastructures as well as their planning, construction, maintenance and operation. Concrete consequences of selected automated vehicle functions as well as requirements resulting from their operational design domain (ODD) definition to infrastructure up until the year 2040 have been assessed. The need to define and provide the required Operational Design Domain (ODD) to enable highly automated driving were identified as most pressing for the road authorities and operators. This paper summarizes the results and provides infrastructure recommendations to support highly automated driving covering the areas of traffic management, road maintenance, crisis management, traffic information services, road planning and building, road works planning, physical infrastructure, enforcement, ITS systems, digital infrastructure, road user charging and new core businesses. 1. Introduction 1.1 Objective Automation has been advancing very quickly during the past years in all domains of the society. In transport, the development started in the late 1980s but has advanced extremely rapidly during the past 7-8 years, mainly due to technology giants like Google stepping into the business earlier dominated by automobile manufacturers only. Research and technology development activities around road vehicle automation have very much focused on the basic enablers of automated driving, such as sensing, positioning, vehicle control and artificial intelligence. So far, road authorities and operators have not been heavily engaged in the development of automated driving automation until recently. CEDR provided its position paper on road vehicle automation in 2016, and has initiated a working group on CAD, while many NRAs have developed specific strategies and action plans for CAD. There are two main reasons for this. First, CAD will have a major impact on the NRAs’ policy goals related to safety, efficiency and environment as well as their core business of network operation. Second, NRAs have a major influence in determining where higher level automated driving can in fact take place. With this in mind, the Conference of European Directors of Roads CEDR launched its Automation research call in late 2017. MANTRA “Making full use of Automation for National Transport and Road Authorities - NRA Core Business” has responded to the questions posed as CEDR Call’s Topic A: How will automation change the core business of NRA’s, by answering the following questions: • What are the influences of automation on the core business in relation to road safety, traffic efficiency, the environment, customer service, maintenance and construction processes? • How will the current core business on operations & services, planning & building and information and communication technology (ICT) change in the future? 1.2 Methodology MANTRA work started with the analysis of vehicle penetrations and Operational Design Domain (ODD) coverage of NRA-relevant automation functions up to 2040 (Aigner et al. 2019). The next MANTRA deliverable (van der Tuin et al. 2020) on the impacts of connected and automated driving (CAD) studied the impacts related 306 2. Kolloquium Straßenbau - September 2021 Consequences of connected and automated driving to physical and digital high-level road infrastructure to the role and policy targets of NRAs. In parallel the impacts and the resulting consequences and therefore necessary changes to infrastructure were assessed (Ulrich et al. 2020). The overall methodology followed the process as shown in Fig. 1. Starting with the current status quo on European highways, inputs from several sources formed the basis of the assessment. The initial starting point was a set of candidate automation functions all based on the latest definitions in ER- TRAC (2019). Through scientific analysis of deployment and the collective selection with CEDR of most significant automated functions and use cases during workshops, the following functions shown in Fig. 2 have been chosen to be further studied. Fig. 1: Overall Methodology to find required changes to infrastructure Fig. 2: Selected use cases including level according to SAE J3016 2. Kolloquium Straßenbau - September 2021 307 Consequences of connected and automated driving to physical and digital high-level road infrastructure Fig. 3: Assessment of the infrastructure impact from three different angles (Ulrich et al. 2020) The infrastructure consequence analysis was thereby tackled from three directions in order to structurally cover the crucial ones, referred to as impact categories, as shown in Fig. 3. The assessment of these three impact categories included literature analysis, expertise of the consortium as well as structured interviews with selected experts, workshops with CEDR CAD WG and a workshop with experts from road authorities, operators, automotive, civil design and construction companies, telecommunications industry and research/ academia stakeholders. The consequences related to infrastructures of the road operators and authorities are described below for the different activity areas of the road operators, focusing on technical rather than legal aspects. 2. Results - Impact on Infrastructure 2.1 Traffic management The concept of cooperative traffic management needs to be fully developed and implemented building on the work carried out among other e.g. in the TM2.0 (2018), SOCRATES 2.0 (2018), and C-ITS Platform (EC 2017). Traffic management will become an integral part of overall mobility management. In an ecosystem enhanced by significant decarbonisation and privacy priorities together with high degrees of digitalisation, traffic management is anticipated to most probably by 2040 become closely integrated with fleet management, at least with regard to ODD management also with e.g. minimum risk manoeuvres (Ulrich et al. 2020). Hence, we need to establish real-time two-way connectivity between traffic management and vehicles. This needs to be done directly and/ or via OEM or service provider clouds. Furthermore, the connectivity should be used to share safety and traffic management related data. The latter will also include traffic rules and regulations as well as ODD-related data such as for example geofences due to or affecting ODD, or incidents, events or conditions affecting the ODD. Specific access points to digital traffic rules and regulations (e.g. a Trusted Digital Regulations Access Point) and ODDs need likely to be set up to facilitate the cooperative traffic management in practice, with high level data security. The traffic management systems have to be digitized, and the traffic circulation and traffic management plans upgraded to include mobility management and also ODD management aspects. Tools such as geofencing need to be adapted for deployment. Quite likely, the contents of these plans need to be evolving during the whole transition period from fully human-operated to a situation, where close to 100% of the vehicles are highly automated. The digital traffic management systems will provide real-time information to HD maps and the local dynamic maps in the vehicles via the access points or also directly in specific cases such as e.g. road work zones. Traffic management for events and incidents including shortand long-term road works should be enhanced and harmonised to maximise efficiency. (Ulrich et al. 2020) Standards need to be developed for the exchange of digital traffic rules, traffic management plans, and ODD management related data as well as the related access points, including the data security solutions. Further standards or similar are needed for the harmonised traffic management and marking of road work zones and incident sites. (Ulrich et al. 2020) 2.2 Road maintenance In road operation, maintenance and traffic management, automation can certainly contribute to increase safety of operational workers as well as road users, improve traffic flow and optimize operational cost but only in combination with connectivity. Thereby, we need integrated connectivity of operational vehicles and road maintenance work zones with a traffic management centre equipped to inform vehicles in real time about such works. The recommendation is quite similar as with traffic management. 308 2. Kolloquium Straßenbau - September 2021 Consequences of connected and automated driving to physical and digital high-level road infrastructure Road inspections, minor repairs, winter maintenance, incident management, and other traditional works will also be necessary in the future. Nowadays they are carried out by operational workers who are always at risk due to high-speed traffic right next to them. Supporting them in the most critical operational tasks, like work zone protection on fast lane and winter maintenance with automated driverless vehicles will take away main safety hazards. This requires the further development of the technological readiness of the systems and the related legal framework. The digital infrastructure enabling the positioning of the vehicles and according standardized, connected communication with the traffic management centre are key for the safe implementation. Road maintenance can also benefit from new condition data sources made possible through additional vehicle sensors and V2I communication. Vehicles providing road condition data to the TMC promise major improvements for predictive maintenance. This should cover cracks, rutting and skid resistance data from vehicle sensors. Overall the digital part of an operations management centre and the traffic management centre will need to merge and have integrated communication standards rather sooner than later. (Ulrich et al. 2020) 2.3 Crisis management This field is potentially fuelled by anticipated increases in severe weather conditions in Europe, as well as by increased expectations into adequate management and mitigation activities. Higher degrees of dependability on communication infrastructure add to the criticality. Crisis management is closely linked to traffic management. This again is important in both directions: informing road users quickly as well as using digital infrastructure of sensors, cameras and vehicles to make the traffic management centres aware of new incidents as quickly as possible for fast reaction times. The increasing eCall information needs to be provided directly to the responsible traffic management centre to accelerate the crisis management. 2.4 Traffic information services The role of traffic information is changing. From a policy relying on providing information on traffic conditions and problems to the driver with the final decision-making by the driver we are going for a policy, where the traffic managers make the overall decisions on behalf of the individual drivers and automated vehicles. This is especially true in large cities and busy peri-urban road networks prone to incidents with considerable consequences to travellers and hauliers. The role is also changing due to information’s increasing importance to the transport system, because what was only desirable for human drivers, is essential for highly automated vehicles (Sweatman 2019). Highly automated vehicles need to be aware of everything happening on the route ahead, also beyond their own sensors. Hence, the quality of traffic information needs to improve from the levels of today. Due to the fact that the automated vehicles with their advanced sensors will be part of the solution themselves, the quality of the traffic information will gradually improve with increased fleet penetration of connectivity and high-level automation. The prerequisite for the improvement is that the stakeholders involved - drivers and OEMs governing the data created by their vehicles, service providers and road operators governing the data from their customers and own monitoring stations - are willing to share their data. This could follow from the Data for Road Safety initiative of the European Data Task Force having a 12-month trial of the concept of sharing vehicle originated road safety related data among the stakeholders involving member states, OEMs and service providers (DTF 2019). To ensure the quality of traffic information, stakeholders need to use appropriate quality assurance methods and processes. While this is a standard practice for commercial stakeholders, many road authorities and operators do not have such quality assurance in place. In the future, the road users (drivers, automated vehicles, vulnerable road users) will receive information increasingly via their onboard devices. These can be devices embedded in the vehicle by the OEMs or aftermarket or nomadic devices attached to the dashboard of the vehicle. Unfortunately, today the OEMs, service providers and app developers use a large variety of pictograms and message content in presenting the information to the user of the device. Often the contents and pictogram differ considerably from that shown by the road operator (Haspel 2019). For the sake of safety, it would be good to harmonise at least the pictograms used by the different stakeholders, but preferably the whole message content (Kamalski and Rytkönen 2015). Automated driving systems would also benefit from a harmonised, consistent use of the pictograms (Ulrich et al. 2020). 2.5 Road planning and building The planning of new roads obviously needs to consider and make provisions for mixed traffic and connected automated vehicles. These new roads however will only be a very minor part of the network used by automated vehicles. Therefore, it is important to define standards for rehabilitation and extensions of existing roads considering the necessary equipment. This way the road networks will be upgraded step by step as part of the continual maintenance program. Infrastructure support levels for automated driving (ISAD) as developed in the project Inframix (Carreras et al. 2018) should be further defined to provide very clear guidelines for necessary digital and physical infrastructure a like. New road planning in the future needs to involve the assessment of the new sections and dependent on their importance and segment a categorization in ISAD levels as well as the operational 2. Kolloquium Straßenbau - September 2021 309 Consequences of connected and automated driving to physical and digital high-level road infrastructure Design Domain (ODD) requirements of the highly automated vehicles. The ODD requirements should be built into the design guidelines for new roads planning and for rehabilitations of existing roads. New road construction makes the integration of digital infrastructure, partly included also in ISAD levels, much easier compared to upgrades during rehabilitations of existing roads. Design guidelines considering all this will need to be developed for planning of new roads as well as for upgrades of existing ones. Some countries already started to develop such guidelines for infrastructure (e.g. U.S. DOT 2018 and Zencic 2019) but also admit that it is an ongoing approach also facing the challenges of limited, concrete exchange with automated vehicle developers in terms of ODDs. One element of new road planning and construction is the application of the BIM (building information modelling) methodology to ensure the parallel development of a so called digital twin of the new road that includes all necessary design, material and operational data for each asset. This will also provide the basis for road operators’ information exchange and provisions for HD maps. (Ulrich et al. 2020) 2.6 Road works planning Planned road works as part of routine maintenance works, rehabilitation or even new roads are not only core business of road operators but also heavily affect traffic flow and road safety requiring close cooperation with traffic management. Starting with a network analysis to avoid conflicting road work zones in close vicinity, the exact location, planned layout, duration and any other relevant technical information needs to be exchanged with the traffic management centres following a standardized process. During the planning of road work zones - in particular in safety critical areas - cooperative connected safety trailers and temporary sensors should be considered to enable continuous live communication with the TMC. Thus, any changes to the road work zone layout, position of or incidents around the road work zone are communicated directly to the TMC and further on to the road users. Road works planning of the future therefore goes beyond picking the optimal time slots and planning the local traffic management layout. The standardized information exchange with defined communication protocols has to be compulsory. Guidelines for necessary sensors in road work zones need to be developed. (Ulrich et al. 2020) 2.7 Physical infrastructure Road operators are partly able to influence whether or not specific automated driving use cases (such as e.g. truck platooning or highway autopilot) are going to be allowed on their networks and which adaptions are necessary. Physical infrastructure adaptions are very costly, need to be planned far ahead and are also heavily regulated in each country with technical standards. Amendments therefore need to be well thought through. The elements most affected are either the road guidance systems (signs, markings, etc.) which are crucial for the ODD of the use cases or the more extensive elements related to the road geometry and structural adaptations. If road operators want to enable the potentially positive effects of automated driving in terms of safety, traffic flow and such they are advised to make according provisions so their infrastructure supports the ODD. Most required infrastructure support will be on the digital part, and physical infrastructure amendments should be very carefully selected. It is necessary to try to limit the dependence on physical infrastructure because of the cost (Vreeswijk 2019). The tricky aspect for decisions is the constant evolution of the ODDs- This evolution is driven by customer demand, and enabled by the improvement of vehicle sensors - for instance, sensors being able to deal with different kinds of weather conditions - and vehicle software - for instance, AI being able to deal with safe manoeuvring of the vehicle also in interaction with vulnerable road users in complicated urban environments. The technological development in the areas of sensors and software is currently very fast, and also hard to predict with any certainty. The overarching recommendation to NRAs is however to analyse their networks and prioritize where deployment of CAD use cases is most suitable and sensible. The likely actions deals with pavement design and maintenance standards review and adaption, pavement enforcement on truck platooning routes, additional emergency bays, wide shoulders and safe harbours, passenger pick-up/ drop-off points, amendments to general road design, changes in ramps and junctions, road markings, road signs’ machine readability and digital twins, and amendments to road furniture (landmarks, gantries, gates). More details are given by Ulrich et al (2020). 2.8 Enforcement The whole area of enforcement will be heavily affected by digitization and connectivity in close relation with changes in traffic management, bringing opportunities of improved cross-border and cross-entity cooperation. The enforcement of allowed weights (and dimensions) will become more critical with the potential of automated high capacity goods vehicles and truck platoons increasing loads on pavement and bridges. The integration of weigh-in-motion (WIM) systems in the pavements and bridges with legally accurate measurements will allow for continuous measurements with less necessary infrastructural and personnel resources that are now required in designated weight control parking areas. Dimensions can be checked already now visually through toll cameras but legally those are not accurate enough as are the WIM systems. The V2I information exchange with connected traffic management enables 310 2. Kolloquium Straßenbau - September 2021 Consequences of connected and automated driving to physical and digital high-level road infrastructure direct enforcement through the necessity of data provision from vehicles on their speed, weight, environmental category, etc. While desirable for road operators and police, the issue is very sensitive in terms of privacy, data security and also market competitiveness. (Ulrich et al. 2020) 2.9 ITS systems This means the traditional ITS systems utilised by road authorities and operators, primarily systems deployed on the roads and in traffic management centres. The information and guidance currently provided via variable or static message signs can be replaced with data provided via cooperative ITS or other messages provided to the on-board systems in the vehicles. During the transition period, which can last to 2040 or even beyond, the human-operated, unconnected vehicles are also on the roads, and their drivers have to be considered. This means that at least all regulatory signs need to be maintained, while considerable number of human-operated unconnected vehicles use the roads. (RWS 2018) The informative and route guidance signs, however, can gradually be abandoned. Likely this can be dealt with by not renewing the signs, when they had reached the end of their life cycle. In addition to variable and static message signs, the road operators have equipped their roads with roadside stations often in connection with monitoring systems (loop, radar and other traffic detectors, road weather sensors, cameras, etc.). The increasing penetration of connected vehicles will improve the possibilities of utilising the data from the connected vehicles and there-by obtaining monitoring data from the whole network instead of the cross-sections equipped with fixed monitoring systems. Despite this, the road operators should still maintain and install fixed monitoring stations. First, the fixed stations are needed for the use of forecasting and nowcasting the conditions on the road network. Second, it is not wise for the road operators to rely solely on other stakeholders to provide the data needed by road operators in their core business. The future needs here are difficult to predict, and to mitigate impacts from this uncertainty, the road stations should not be rigid single-purpose components but should be adapted flexibly to meet the changing needs of the road operators. Hence, the traditional roadside stations at the end of their lifecycle should be replaced by flexible roadside stations that respond to current as well as future needs (RWS 2018). 2.10 Digital infrastructure As parts of the digital infrastructure have also been discussed above, this part contains high-definition maps, satellite positioning, communication infrastructure, and fleet supervision centres. The consequences for HD maps have been described in detail by the DIRIZON project (Malone, et al. 2019). The road operators are expected to provide data for the HD maps to road map and service providers directly or via national access points. The profiles, formats, structures and procedures needed to handle data streams are to be specified and tested in agreement with other stakeholders, and especially the HD map providers. The road network data will need to be digitized including any landmarks supporting accurate vehicle positioning. This will be carried out by HD map providers, but also road authorities and road operators may want to have it done for themselves as HD maps of the roads and their (sub-)structures can be regarded as a key asset of the road operators with regard to their core business. By 2040, the feedback loops for maintaining data quality will be established, the digital traffic rules included, the HD maps localization quality reached, most of the physical and digital infrastructure elements digitised and available to HD maps, and HD digital map will achieve the data quality levels required for the decision-making process in an automated vehicle (Malone et al. 2019) Specific attention needs to be given to including ODD attribute related data in the HD digital maps especially for physical infrastructure attributes, which may not be provided by the road operators throughout the road network due to their high costs. Examples of such are, for instance, wide shoulders, safe harbours and game fences. The availability and location of such attributes is essential for the highly automated vehicles in order to determine the existence of their ODD. Highly automated vehicles utilise several independent positioning methods such as satellite positioning and inertial positioning, mobile phone network positioning as well as car sensors and HD map positioning (Koskinen et al. 2018). Satellite positioning is the basic positioning solution, and it has been shown to reach the desired 5 cm accuracy when supported by RTK (Real Time Kinetics) land stations. Such or similar stations should be provided especially in challenging environments such as northern latitudes and mountainous areas. Communication is developing fast and will likely do so during the next decades as well. The basic communication types will most likely still be vehicle to vehicle short range, vehicle to infrastructure short range, and vehicle to infrastructure medium/ long range. The last mentioned will likely be provided via cellular networks, but the short range V2I communications will need communication beacons beside or over the road, connected to different servers (road operators, vehicle manufacturers, service providers, fleet managers, etc.) via trunk communications such as fibre optic cabling. Road authorities and operators benefiting from the connectivity can invest in the trunk communication and road side communication station investments in cases, where such investments are not made by other stakeholders due to their customer needs. Remote operation centres to monitor and supervise fleets of automated vehicles are needed by several use cases of highly automated driving, if not all of them. As the 2. Kolloquium Straßenbau - September 2021 311 Consequences of connected and automated driving to physical and digital high-level road infrastructure fleets will mostly belong to other stakeholders, such centres will be the responsibility of these other stakeholders. Some national road authorities and many road operators deal with the operational maintenance and winter maintenance of their road networks. Thereby, those road authorities and operations need to set up their fleet supervision centres. 2.11 Road user charging In satellite-based systems there exist only virtual toll plazas, if any. Consequently, properly equipped automated vehicles can behave as traditional vehicles in these systems. Modern DSRC-based tolling systems are based on the “multi-lane free-flow” principle. In these systems, properly equipped automated vehicles can also behave as traditional ones. Hence, changes in the physical infrastructure are only required on roads with traditional toll plazas. At the toll plaza area approaches, gates and exits, standardised markings should be used to indicate the routes and lanes to be used by highly automated vehicles. Automatic payment lanes need to be included in the toll plaza setup. Concerning the physical infrastructure, the road charge information needs to be a part of the dynamic layer of the HD map, for instance in the rules and regulations part. 2.12 New core businesses Connected and highly automated driving has enormous change potential. Significant cornerstones in the road transport system will be blending with a broader IOT and AI ecosystems. This also means potentially new core businesses for road operators. In general, new core businesses will not so much relate to infrastructure provision but rather shift the focus to an even more service provider oriented business model for road operators. Some candidates for such services are listed below: - Elements in a broader mobility-as-a-service ecosystem, where travel time is used for productive or recreational purposes by the travellers - Integrating (and potentially mitigating) a potentially increasing number of services and non-traditional vehicle concepts and services - Mitigating issues of a highly fragmented communication network reality in Europe (e.g. mitigate end of network / end of high quality communication infra-structure impacts, including expectation management) - Validating quality of service in communication infrastructure and map infrastructure - Facilitating, in a freight automation context, entirely new forms of vehicles in terms of length and behaviour - also taken up proactively to mitigate risks of alternative service providers - More dynamic parking management 3. Conclusions This paper compiles recommendations for NRAs on the expected impact of highly automated driving and the respective necessary changes to physical and digital infrastructure that can support cooperative, connected and automated road traffic. The work was carried out as part of the CEDR project MANTRA and detailed in a separate deliverable (Ulrich et al. 2020). Introducing connected and automated mobility on public roads is expected to effectively address several traffic safety, efficiency and environmental problems. While the expected impacts to infrastructure are manifold those resulting from the need to provide the required Operational Design Domain (ODD) for highly automated vehicles were identified as most pressing. The identified impacts obviously only reflect the recommended actions in order to enlarge the automated vehicle’s ODD coverage as far as economically feasible. There are some inherent difficulties in supporting the ODDs as they depend on the capabilities of the sensors and software including AI of the automated vehicles, and these capabilities are improving quite quickly with the evolution of related technologies. The results reflect the knowledge of the likely function and ODDs of highly automated vehicles at the time of writing at the end of 2019. It is likely that technology, market and policy developments will change the importance, benefits and costs of the individual changes in the physical and digital road infrastructure considerably until 2040. References [1] Aigner, W., Kulmala, R., Ulrich, S. (2019): Vehicle fleet penetrations and ODD coverage of NRA-relevant automation functions up to 2040. MANTRA: Making full use of Automation for National Transport and Road Authorities - NRA Core Business, Deliverable 2.1 [2] Carreras, A., Daura, X., Erhart, J., Ruehrup, S. (2018): Road infrastructure support levels for automated driving. 25th ITS World Congress, Copenhagen, Denmark, 2018. [3] DTF (2019): Data for Road Safety. European Data Task Force. https: / / www.dataforroadsafety.eu/ . [4] EC (2017): C-ITS Platform Final report Phase II. Cooperative Intelligent Transport Systems Towards Cooperative, Connected and Automated Mobility. [5] Haspel, U. (2019): C2SBA und C2NBA. Aktueller Stand und Planungen. Bayerische Staatsbauverwaltung, Zentralstelle Verkehrsmanagement. [6] Kamalski, T., Rytkönen, M. (2015): iMobility Forum SafeAPP WG. Presentation at the iMobility Forum Steering Group Meeting. ERTICO, 19/ 11/ 2015. [7] Koskinen, J., Kuusniemi, H., Hyyppä, J., Thombre, S., Kirkko-Jaakkola, M. (2018): Positioning, location data and GNSS as solution for autonomous