Contribution of EU projects to CAD standardisation

Stakeholders of Connected and Autonomous Driving (CAD) are collaborating in Governmental Programmes, Industry associations as well as private initiatives, to identify and solve the challenges for an effective deployment Connected and Automated Driving. One of the main challenges is related to the standardisation of novel technologies that are integrated in the vehicle, as well as development, engineering and testing. The standard domains reported in the “Standardisation  Collection” webpage are reported in the following table.

Domain
Terms & Definitions
Management / Engineering Standards
AD/ADAS Functions
Testing, Verification and Validation
In-Vehicle Systems, Networks, Data and Interface Definition
Connectivity
Human Interaction
Artificial Intelligence
Safety
Privacy & Security
Map and Positioning

In this article, we give an overview how research and innovation (R&I) in Europe is addressing these domains. The following section presents a sample of European and National R&I projects. These projects were selected based on contributions to CAD standards and gaps/challenge identification. We tried to cover all the aforementioned standard domains.  The last chapter summarizes the analysis findings, which are currently being discussed within ARCADE. The final goal of ARCADE is to provide guidelines on future CAD standardization.  

Along the text to help the readers, we have highlight in:

  • italic the gap
  • in bold the recommendation and contribution

EU Projects Analysis

L3 Pilot

The L3Pilot project (2017-2021) is performing large scale Field Operational Tests and piloting AD with developed SAE Level 3 (L3) functions along different road networks. It plans to assess also Level 4 functions and connected automation. From the experimental data, L3Pilot aims at unified, de-facto standardised methods to ensure further development of AD applications (Code of Practice, CoP) and at a standardised Europe-wide piloting environment for AD. The L3 Pilot CoP draft document [1] mentions the compliance to existing standards but focusses on safety, security, related gaps, especially in the development and testing. For instance, it reports that ISO26262 standard does not address evaluation of HW tools, like measurement equipment, reference systems for data collection, and does not have a focus on a safe development. As example, the CoP refers to possible outcomes in the case the test failed (material damage or harm to people). Another gap in the standard is the training of safety drivers. For security, the CoP refers to ISO 21434 “Road vehicles – Cybersecurity engineering”, ISO standard under preparation. It also reports the need of OBDII standard evolution to integrate security requirements. Concerning Ethics standards, L3Pilot CoP reports that ethical standards do not need to be explicit standards but can also be implicit societal agreements, and can change over time. The document primarily refers to the Safety First milestone white paper [2] and its twelve principles for automated driving. On terminology,L3Pilot partnersreviewed and supported experts from the SAE international and the German Institute for Standardisation  (DIN) to develop a glossary of terms for Automated Driving, DIN SAE SPEC 91381 Terms and Definitions Related to Testing of Automated Vehicle Technologies.

Pegasus

The German Funded project PEGASUS (ended in 2019) developed tools and methods to ensure that automated driving is mature enough for market introduction. The goal was to achieve uniform technical standards for the verification of highly automated vehicle systems, and to answer key questions about safety and reliability. The Pegasus example is particularly relevant for the standardisation  of test scenarios for Automated Driving, and has proposed a method based on a 6 layer scenario [4].

HEADSTART

HEADSTART (2019-2021) aims to define testing and validation procedures of CAD functions including key technologies such as communications, cyber-security and positioning. The tests are both simulation and real-world fields to validate safety and security performance according to the key users’ needs. The project’s state of art deliverable D1.1 [5] highlights how GNSS security and integrity are still major challenges for CAD. Some standardisation activities have been started but the identification of unified procedure to validate the reliability of the GNSS based solutions is still under investigation. HEADSTART gives the current state of play of Map and positioning standards, referring to the ongoing CEN EN16803 standardisation activity, and mentioning relevant projects, such as SaPPART, ESCAPE and Galileo for Mobility.
Testing and validation by means of simulation allow to extensively test automated driving. Thus, combining of real trials with simulation is the only possible path for testing and validating highly automated driving functions. However, there is still way to go in terms of standardization, especially in the way simulation is compared with real world tests. Existing ASAM standards (mainly from Pegasus) are not sufficient. ASAM proposed to extend the standards to fulfill simulation domain requirements (including Vicomtech contribution). Standardization is still ongoing in spatial temporal events labeling; it is lacking in Operational Design Domain (ODD) definition (foreseen in Horizon Europe CL5-2021-D6 01-01) and in the comparison of simulation with real tests. There are gaps also in ontology definition, to have unified data meaning independently of the underlying IT platform (tackled in Horizon Europe CL5-2021-D6 01-02 [Ref. Horizon Europe Work Programme 2021-2022]). Another gap, in focus of HEADSTART project, is the inclusion of Key Enabling Technologies in simulators. These, as well as other technologies that are crucial for SAE L3, L4 and L5 but are currently not used for SAE L2, deserve special attention. For example in SAE L4 the system must perform well in navigation tasks. Reception of information from other vehicles, precise positioning and related data integrity are other key aspects for Highly Automated Driving (HAD) and have the following standard gaps: in the way these technologies are tested in real life and simulation and a common metrics for the ODD definition, depending con communication/positioning quality. HEADSTART is partially tackling this, but not in terms of standardization. Openscenario (ASAM) is used by the stakeholders, however, if such standards have to be used in validation of HAD, we may need ISO standards. Typically ISO or similar require 4 years for proposal. Considering e.g. HEADSTART end (2021) this would mean indicatively 2025 to fill the gaps.

AdaptIVe

In its final deliverable D1.0 [6] the AdaptIVe project (ended 2017) mentioned standardisation  gaps in its lessons learned. When they have investigated transitions from automated to manual driving, response time alone is not enough to provide information on how well drivers can handle a vehicle after re-taking control. Other metrics, e.g. steering and braking patterns, should be considered. In addition, physiological measurements are needed to understand drivers’ behaviour and their ability to re-take control. Standardisation is needed in this field. Another standardisation  gap found by AdaptIVE is connectivity in Valet Parking function.

CloudLSVA

CloudLSVA (ended 2018) integrates big data, video annotation and cloud based technologies for improved ADAS and Digital Mapping. The project [7] proposed a new format called Video Content Description (VCD). This format describes the scene not only enumerating the object, but also defining the relations and the spatial-temporal events among all elements. The project established an Open Group to trigger and support the development of standards for the video data set and video annotation. The VCD 4.3.1 is being designed to be compliant with OpenLABEL, which is a proposed standard of ASAM (Association for Standardisation of Automation and Measuring Systems) for Nov 2021. As the standard evolves into a standardisation project, VCD will evolve as well to become the first labeling toolset compliant with the standard.

BRAVE

The BRAVE project (final event in Feb 2021) focusses on AD needs and requirements from the viewpoint of the users, other road users affected and stakeholders. The Deliverable D4.5 [8] refers to the interaction between Autonomous Vehicle and Vulnerable Road user, referring to ISO/TR 23049 as the most central standard. The project gives recommendations for standardisation  of AV-pedestrian interaction signals. It lists the types of contents that could be reflected in external HMI of AVs, but also refers to the issues or conflicts that make it difficult to create a single, standardized concept for all AVs.

C-Roads Platform

The C-Roads Platform (2016-2023) is a joint initiative of European Member States and road operators, co-funded by CEF, for testing and implementing Cooperative Intelligent Transportation Systems (C-ITS)  based on Vehicle-Road Infrastructure communication, in light of cross-border harmonization and interoperability. C-ROADS, and in particular the Worlking Group “Technical Aspects” [9] refers to the set of standards enabling interoperable V2X communication (ETSI ITS G5 based on IEEE802.11p but also “hybrid communication” based on LTE). In particular, C-ROADS aims to harmonize how the standards are used in practice in EU member states. This process includes dedicated monitoring of standards, identification of relevant aspects and alignment with respective pilot requirements. Traffic management issues, such as usage and processing of data for traffic management, will be identified and integrated into the overall picture of influencing factors, which will also contain the link to urban environments. In order to ensure C-ITS service consistency for users, recommendations from the road operator’s point of view on the visualization or presentation of messages on HMIs are discussed.

ICT4CART

The main goal of ICT4CART (2018-2021) is to design, implement and test in real-life conditions a versatile ICT infrastructure [10] that will enable the transition towards higher levels of automation addressing connectivity, data management, cyber-security, data privacy and accurate localisation. ICT4CART builds on high-value use cases in urban and highway scenarios. In at least three use cases novel connectivity standards and AD functions are used. In the Smart Parking, the infrastructure-to-vehicle information about the available spot to the Autonomous Vehicle, uses extended messages with respect to the existing ETSI standard, namely Parking Availability Message (ETSI TR pending). In motorway lane merging, the road infrastructure identifies the incoming vehicles monitoring the highway lanes with cameras and collecting the information from standard Cooperative Awareness Messages (CAM) sent by the connected vehicles. The information is related to Collective Perception Messages (CPM), which is a standard message to share detections among vehicles, is in course of standardisation by ETSI in TS 103 324 (TR available as ETSI TR 103 562 V2.1.1 (2019-12) and has standardization gaps. The latter regard more generally the sharing of V2X environment data, key for autonomous driving: object trajectory prediction (also addressed in Maneuver Coordination Service), Redundancy Mitigation Techniques to limit the number of redundant messages, VRU nomenclature discrepancies between CPM and ETSI VRU related standards (ETSI TR 103 300-1 V2.2.1,     ETSI TS 103 300-2 V2.2.1,    ETSI TS 103 300-3 V2.1.2). Finally in ICT4CART new messages are addressed to dedicate lanes for Automated Vehicles at Toll stations, submitted to ETSI for TR and new use case of Wrong Way Driving for the update of ETSI TR102638. This again indicates the standardization gaps in V2X information sharing, whereby input is still being given from R&D projects to SDO’s.

ENSEMBLE

The ENSEMBLE project (2018-2021) is developing pre-standards for interoperability between trucks, platoons and logistics solution providers. A relevant contribution to ENSEMBLE is in the C-ITS standard of ETSI ITS G5, in particular the introduction of a platooning container in the Cooperative Awareness Message, with backward compatibility with existing standard [11]. The outcome of the ENSEMBLE project on platooning is essential for the platooning protocol in ETSI TC ITS. The platooning message will add to existing standard protocols (cooperative awareness message, CAM, and decentralized environmental notification message). It is not a broadcast protocol, only members of the platoon can communicate with each other in a secured and trusted way. Security is rooted in the European trusted domain based on public key infrastructure with the addition of encryption of platooning data. Messages are transmitted with a fixed rate of 20 MHz by all platoon members containing, for example, a high-resolution acceleration signal, which is crucial in an emergency braking situation. The next step for platooning ENSEMBLE in ETSI standardisation is to finalise a technical report identifying all details to be standardised. This forms the basis for the technical standard on platooning containing requirements that need to be fulfilled for creating an interoperable platooning system.

AUTOPILOT

AUTOPILOT (ended in Jan 2020) aimed at integrating Internet of Things (IoT) in Autonomous Driving. With IoT, objects are becoming “things” that can be addressed, recognized, localized and controlled via telecommunication. AUTOPILOT focused on the benefits of sharing data about road status, hazards, other vehicles, etc. for Highly Automated Driving. The interface for data sharing and object discovery is done through an IoT platform, using Machine to Machine (M2M) communication. In AUTOPILOT, data exchange via the IoT platform was applied Urban driving, Automated Valet Parking, Highway Pilot, Platooning and Real time car sharing. The project submitted 25 contributions to Standard Developing Organizations (SDOs), finding the existence of gaps in standardisation, in particular with respect to data models and the ETSI “oneM2M” IoT platform [12]. “oneM2M” (ETSI is one of the founders) is the global standards initiative covering requirements, architecture, Application Programming Interface (API) specifications, security solutions and interoperability for M2M and IoT technologies.

5G-PPP Projects

Regarding 5G connectivity [13] three pilot projects 5G Croco, 5G-CARMEN and 5G-Mobix, are going on in parallel and targeting 5G road testing for Connected and Highly Automated Driving. Their contributions to standard regard connectivity technologies, in particular 5G cellular technology and the related enabling components and technologies. Among the enabler we highlight Multi Access Edge computing (also referred to as “fog” computing in analogy to “cloud” computing), that considers computation, storage, and communication at the edge of the network (close to the cellular base stations) thus minimizing latency, which is key for Automated Driving.

Regarding standards contribution, 5G Croco project delivered research results to 3GPP Service and System Aspects (SA) SA2 and SA6, on the communication standards 3GPP Release 16 and Release 17 Study Items relating to communication Quality of Service, MEC, Mobile Network Operators interactions and tele-operated driving. 5G Croco also contributed to ETSI Technical Committee ITS, with intelligent transport system architecture extensions using Cellular infrastructure and Interoperability as well as C-ITS Security areas. Regarding pre-standardisation , 5G Croco has been contributing to several of Work Items of 5GAA, including tele-operating driving use case, quality of service and safety related features.

The 5G-Mobix project contributed to a white paper on IPv6-Based Vehicular Networking.

5G-CARMEN addresses 5G- service continuity when driving in AD mode from one country to another. Its contributions are thus related to the domain of 5G cross-border connectivity, specifically MEC connectivity (contribution to ETSI MEC and to the Internet Engineering Task Force – IETF), Network Functions Virtualization (to ETSI NVF) and network reselection improvement at the border (to 5GAA, in view of possible future standards).

RAINBOW

Referring to the aformenentioned edge/fog computing, the RAINBOW project (2020-2022) [14] designs and develops an open and trusted fog computing platform that facilitates the deployment and management of scalable, heterogeneous and secure IoT services and cross-cloud applications. The project plans to contribute to standardisation on fog computing (ETSI GS MEC-IEG 004).

5G-META

5G-META is a recently started project, relevant for CCAM on the topic of IoT data access standardisation. 5G-META is developing and integrating platform to ingest and deliver car data, and its objectives include CCAM use cases. The relevance for standards is in the data shared by OEM on the cloud/edge. This project is collaborating to ETSI-MEC standards, e.g. the interfaces for applications  to access vehicle data on MEC (ETSI MEC.030 V2X API), and to retrieve information nodes in a certain area (MEC Location API  ETSI GS MEC 013). Findings of 5G-META are expected to continue the work on IoT but also to contributed to standardization of Vehicle-to-Network (V2N) data sharing, addressed in AUTOPILOT (ETSI-M2M) and C-ROADS Task Force 4 (“Hybrid Communication”) but not standardized, yet. This is a remarkable gap, considering the increasing role of V2N in the future, also thanks to 5G.

SAFERtec

SAFERtec introduced a security assurance framework based on Common Criteria and composed of methods and tools for the cyber-security of vehicular V2X eco-systems [15]. SAFERtec has contributed to the development of the ETSI ITS TVRA document (TR 102 893 – TVRA) by introducing and strengthening privacy requirements (i.e. related to anonymity and unlikability) for the actors of the V2X communication. SAFERtec contributed to the document ETSI EN 302 890-2/ Facility Position & Time, related to the extension of security requirements. Furthermore, the project has carried-out experimental work to improve the adoption of plausibility checks as misbehavior detection, related to the ETSI TR 103 460 pre-standardisation study. Finally, SAFERtec collaborated with the Car2Car Consortium (C2C) in updating their Vehicle C-ITS System (VCS) protection profile (i.e. a set of criteria to evaluate the level of security of a specific technology or product).

CARAMEL

Another project expected to contribute to standardisation gaps in cybersecurity is the CARAMEL (2019-2022) project [16]. The Objective of the project is to address modern vehicle Cyber Security challenges proactively by employing Artificial Intelligence (AI) & Machine Learning (ML) techniques and to explore methods to mitigate the associated safety risks. The project foresees 3 key areas where Cyber Security innovations are required i) Autopilot, ii) Remote Pilot iii) V2X communications and iv) Electrical Charging stations, and so the use cases developed encompasses the technologies in these key areas. The projects intends to follow the layered approach with the notion of Defence in Depth. The first line of defence would be to extend and integrate advanced security technologies to secure the attack surface for i) Autonomous, ii) Connected, iii) Electro Mobility and iv) Remote Control vehicles. The second line of defence would be to design & develop tailored intrusion detection and prevention tools. And third line of defence would be to have a software solution for monitoring, detection and sending warnings/alarms about threats.
Also as part of V2X communication / Connected Vehicles, the project focuses on interoperability between radio technologies using forwarding policies executed in a MEC, which seeks to eliminate the communication gap between two (or more) different standards of radio communications being utilized by vehicles in a region to transmit V2X messages. This will enable ITS services to work with all the information available in the area and for vehicles to be able to receive all V2X messages without any restrictions of the radio technology being used.

Conclusion

The following table illustrates the impact of the projects on the standardisation  domains defined on the webpage “Standardisation  Collection” .

Project Terms & Definitions Management/ Engineering Standards AD/ADAS functions Testing, Verification & Validation In-Vehicle Systems, Networks, Data and Interface Definition Connectivity HMI Artificial Intelligence Safety Privacy & Security

Map and positioning

L3 Pilot  •    •          
Pegasus    •    •              
HEADSTART        •  •            
SaPPART                  
ESCAPE                  
Galileo for Mobility                  
AdaptIVe    •  •            
CloudLSVA          •          
BRAVE                     
C-Roads                    
ICT4CART                    
ENSEMBLE      •              
AUTOPILOT                    
5G Croco                    
5G-CARMEN                     
Mobix                    
RAINBOW                    
SAFERtec                  
CARAMEL                

Several projects have been found on Connectivity standards. R&I is also well diversified to cover all aspects of connected autonomous driving: data exchange, C-ITS technologies, 5G network connectivity, role of cloud/edge computing, internet-of-things, etc. The reason could be that there is more request for standardisation than other topics. Indeed, for an effective deployment of Cooperative Intelligent Transportation Systems, vehicles need to communicate to each other and with the infrastructure using standard protocols, a minimum common set of data, common interfaces and requirements. Many standardization aspects are still to be faced, for instance, the aspects regarding advanced use cases for Connected and Automated Driving, where V2X communication is used for actuation.

The topics of Map & position, Privacy & Security, Testing Verification and Validation are well represented by key projects in the field. In particular, testing of connected and automated driving vehicles is one of the most important R&I topics, and it is being tackled in real traffic conditions, test track and simulation. More projects on AD testing should start soon, and contribute to standards, too. The major topic of AD validation by combining real and simulated tests is tackled in the Headstart project, and although many steps forwards have been made, standardization in this field is still not complete.

Three of the sampled projects, L3Pilot, Pegasus, AdaptIVe target extended field operational tests on autonomous driving vehicles. They have addressed development guidelines, which are included in the domain Management/Engineering standards.

It was difficult to sample projects on AD/ADAS functionality standards, despite the several projects on ADAS/AD. For example, connectivity pilots are still working on the communication standards (what to exchange and how) rather than standardising the usage of connectivity for AD/ADAS functions (what to do with the exchanged data). The most advanced, function-based integration case is truck platooning in ENSEMBLE (note: the truck platooning challenge dates back in 2016, before the project itself).

Although many projects are addressing safety topics, in our analysis safety standards are covered only by L3Pilot. Almost all field operational tests have to follow safety procedures, therefore more contributions can be expected from ongoing projects. However, functional safety and safety of the intended functionality are key topics that should be further investigated, especially for connected vehicles.

Human Interaction standards have been tackled by two of the sampled projects, namely driver-interaction (AdaptiVe) and vulnerable-road users standards (BRAVE).

Two projects are dealing with in-vehicle systems, network and standard interface definition.

Two projects address Artificial Intelligence, Cloud LSVA for object annotation standard, CARAMEL for AI applied to cybersecurity.

References

  1. L3 Pilot Deliverable D2.2, Draft and results from pilot application of draft CoP (2020)
  2. Safety First for Automated Driving (SaFAD) white paper (2019)
  3. DIN SAE SPEC 91381 Terms and Definitions Related to Testing of Automated Vehicle Technologies (2019)
  4. Hendrik Weber, Julian Bock, Jens Klimke, Christian Roesener, Johannes Hiller, Robert Krajewski, Adrian Zlocki & Lutz Eckstein (2019) A framework for definition of logical scenarios for safety assurance of automated driving, Traffic Injury Prevention, 20:sup1, S65-S70.
  5. HEADSTART Deliverable D1.1 State of innovation of existing initiatives and gap analysis (2019); interview with O. Otaegui (Vicomtech)
  6. AdaptIVe Deliverable D1.0 Final project results (2017)
  7. https://cloud-lsva.eu/library/
  8. BRAVE Deliverable D4.5 Vehicle-VRU Interaction Concept Report (2020)
  9. https://www.c-roads.eu/platform/activities/wg-technical-aspects.html
  10. ICT4CART Deliverable D7.2: Adaptation of the reference architecture at the test sites (2020); interview with D. Brevi, E. Bonetto (Links Foundation)
  11. ENSEMBLE Deliverable D2.8 Platooning protocol definition and Communication strategy (2018)
  12. AUTOPILOT D5.8 Standards and conformance of IoT in AD
  13. 5G-PPP Projects Impact on SDO Technical Report, by 5G-IA and 5G-PPP (2020)
  14. https://rainbow-h2020.eu/concept-and-objectives/; interview with CARAMEL project team.

Feedback form

Have feedback on this section? Let us know!

Send feedback

Feedback

Please add your feedback in the field below.

Your feedback has been sent!
Thank you for your input.

An error occured...
Please try again later.