Some years ago we were pondering about what a tool could look like to assess road networks’ readiness for automated driving. Remember this blog?: “What could an assessment tool for Connected and Automated Vehicle (CAV) readiness for roads look like?” This year, we were challenged to rethink our views and see what could be a meaningful, internationally acceptable classification of Smart Roads. Basically meaning: what key functionalities and characteristics should roads have to accommodate self-driving cars and how can this be assessed? And which roads around the globe possess these favourable features? In an inspiring partnership with some of the leading international experts on the topic, we put our views together. We came across very interesting research done so far, of which this blogpost provides a brief state of the art.

Automated driving today

Vehicles being manufactured and released on the road today are featuring functions that fall in the Society of Automotive Engineers (SAE) automation levels 2 and 2+, labelled “partial driving automation”, where the driver retains the responsibility for the Object and Event Detection and Response (OEDR). Japan is the first country to authorize Level 3 vehicle, where the system is responsible for the OEDR. In other regions, plans to introduce conditional automated vehicles from 2020 have been postponed due to legislation.

The introduction of higher levels of automation, characterised by more automated functions, less driver involvement and more Transfer of Control (ToC) instances, are hampered by the complexity of the traffic scenarios they would need to be able to handle, as well as the driving speed they have to safely operate at. The recent adoption of the regulation for Automated Lane Keeping Systems (ALKS) up to 60 km/h by the UN Economic Commission for Europe (UNECE) is the first step towards enabling the deployment of such systems in certain traffic environments. Wider deployment of SAE levels 3 and 4 automation depend on developments in international legislation (in particular regarding operating speeds) and is expected to be rolled out in the coming decade (2020-2030). For a long time however, we will see mixed fleets of various levels of automation on our roads.

Initiatives and future needs

The physical and digital road infrastructure have a fundamental role to play in the deployment of CAVs. Governments are starting preparations for this transition, with first policy initiatives being launched. For example, in Europe, the Road Infrastructure Safety Management (RISM) Directive has been revised, requiring a network-wide road assessment, systematic and pro-active risk-mapping and definition of general performance requirements for road markings and road signs. EU Members States are required to transpose the update in national regulations by the end of 2021.

Similarly, road infrastructure standards might need upgrading and stronger harmonisation. In some environments, roads might need to be adapted, for example to extend the number of safe harbours to accommodate for ToC issues and associated Minimum Risk Manoeuvres (MRM). The perception of automated vehicles of their surroundings is limited by the range and capability of on-board sensors. Digitalisation is necessary to support CAVs in understanding their environment and connect to traffic management procedures. Static digital map data and dynamic traffic information enable a better anticipation of the road ahead. For higher automation levels, connectivity with the road infrastructure and other vehicles is an essential aspect. The requirements for both physical and digital infrastructure related to Connected and Automated Mobility (CAM) impose new constraints and imply new responsibilities for road authorities and operators. In particular in complex urban and rural environments, the safe operation of automated vehicles will require up-to-date High-Definition maps, reliable digitised traffic signs and rules and real time information.

Classification efforts

Several classification schemes are in use or being proposed to describe the capabilities and operation conditions of CAVs. Examples include:

• The above-mentioned SAE levels of automation;
• The Operational Design Domains (ODDs), describing the specific conditions in which automated driving systems are designed to properly operate. Specific ODD-attributes related to physical and digital infrastructure have been identified by Kulmala et al (2018) [1]. While the ODD is a widely accepted concept, there is no harmonised definition yet.
• The Infrastructure Support Levels for Automated Driving (ISAD) concept, describing the readiness of the road network to provide the required infrastructure support for automated functions. The ISAD concept was developed in the framework of the EU-funded INFRAMIX project [2].
• A Connected Roadway Classification System (CRCS) has been developed in the framework of the National Cooperative Highway Research Program (NCHRP) in the United States, identifying three approaches to classify infrastructure improvements to support CAVs [3].
• A Smart Road Infrastructure Standardization Initiative led by the National Highway Research Institute in China, involving Huawai, Baidu and Alibaba among other stakeholders with standardisation plans through SAC TC204 [4].

The next steps

The recommendation starting to emerge from relevant initiatives is that roads should be classified according to both their ISAD and ODD characteristics. However, further development depends on standardisation; several standards related to physical and digital infrastructure for CAVs [5] are still under development. A number of discussion fora have been recently initiated on definitions and classification of infrastructure and related data elements for Automated Driving in the different regions and at international level. Examples include: the CCAM Platform and the Network of National ITS Associations’ concertation on physical and digital infrastructure in Europe, OECD/ITF Working Group on Preparing Transport Infrastructure for Automated Vehicles and the Trilateral EU-US-Japan ART Physical Digital Infrastructure sub-Working Group.

So far, research and categorisation efforts have mainly focused on the requirements for the ODD, often based on SAE automation levels (see e.g. Amditis et al, 2019). The most relevant and advanced initiatives that provide promising building blocks for initial indicators are the ISAD concept introduced by the EU-funded INFRAMIX project and the CRCS methodology. Harmonisation efforts are only now starting to get underway.

Want to know more?

Are you working for a road agency and would you be interested to assess how CAV-proof your network is? Contact us to start a conversation about mapping its readiness. We can develop a tailored methodology and indicators, based on available data (or including new data to be collected), and support you with the collection, interpretation and presentation of results. You can reach us with an email to [email protected].

Acknowledgements: The authors would like to thank Stephen T’Siobbel, Paul Carlson and Cristian Gonzalez for their contributions.