5G networks are perfect technology to serve as communications channel for vehicles trying to communicate with their environment (V2X – Vehicle to everything). There is a need to transfer information between RSUs (Road-Side Units) to vehicle OBUs (On-Board Units). Such information exchange can benefit drivers, pedestrians and traffic control. It can improve traffic safety and emergency services efficiency. Modern and future vehicles will also need support for advanced features like autonomous and remote driving as well as infotainment and augmented reality.

Figure 1: Future traffic management will be based on telecommunications between vehicles and road-side units. This can be very beneficial in emergency situations inside tunnels (photo: Peter Zidar)

Such innovative use of mobile networks introduces new challenges that are poorly addressed by previous generations of mobile networks. That is one of the reasons for upgrading these networks to 5G. The emerging 5G is improving data transfer rates, latency, QoS, security, reliability and service management. Therefore, 5G is in unique position to provide new and improved road safety features, provided by infrastructure, device manufacturers and third party application developers. In particular, 5G slicing can provide a network dedicated resources for V2X communications. These can also be used for road safety and emergency services.

Mobile network operators are in good position to host cloud based platform for independent application developers focusing on traffic safety and other automotive applications. The goal of European project 5G-IANA is to provide a prototype platform Automotive Open Experimental Platform (AOEP). The issues discovered during testing of this experimental platform will provide important data and guidance for future open automotive platforms operated by network operators.

Specific use cases tested in this experimental infrastructure will give Telekom Slovenije better understanding of requirements such applications have. Use Case 7 will investigate support for emergency vehicles approaching location of traffic accident inside tunnel located on country border. The planned solution will involve transfer of video and other infrastructure data to incoming emergency vehicle. Received information can be vital for better and faster rescue efforts and giving victims efficient help.

Author: Peter Zidar

Work on Local Dynamic Maps (LDM) implementation is still in its early stages, as the LDM standards only define how information shall be structured in databases, while the mechanism to fuse or link information across different layers is left undefined. A working LDM component, as a real-time database inside the vehicle, is an attractive solution to multi-ADAS systems, which may feed a real-time LDM database that serves as a central point of information inside the vehicle, exposing fused and structured information to other components (e.g., decision-making systems). In 5G-IANA we will be implementing a real-time LDM component, as a database reachable by vehicles through V2X communications and deployed in road-side units (RSU) or as on-board units (OBU), making use of the three pillars that guide a successful fusion strategy: utilisation of standards (with conversions between domains), middlewares to unify multiple ADAS sources, and linkage of data via semantic concepts.

Advanced Driver Assistance Functions (ADAS), such as Forward Collision Warning (FCW), Automated Braking (AB), Lane Departure Warning (LDW) or Blind Spot Detection (BSD) are already operational in our cars. Their utilisation in a vehicle has demonstrated exceptional performance to increase comfort and safety, with reductions of crashes between 27% to 50%. The research and innovation in ADAS and in Automated Driving (AD) functions is continuously growing and the market expects many new functions in the context of SAE-L3 vehicles. Particularly relevant will be the integration and coordinated co-existence of multiple functions, from multiple cars, potentially from different vendors, and with possibly overlapping capabilities, input needs and power consumption requirements.

The integration of an ecosystem of ADAS has become a major challenge, because their simple accumulation in a vehicle can produce undesired system conflicts and also decrease user acceptance because of the increased complexity of the vehicle. Car manufacturers are, therefore, keen to research on integrated solutions, which coordinate functions to operate harmonized, balance power and processing consumption, and overall improving driving safety.

One of the main challenges is the ability to fuse data from heterogeneous sources from diverse domains that have emerged and evolved independently from each other during the last decade: perception (e.g., sensing devices such as cameras or LIDARs with detection capabilities), communication (e.g., V2X systems, with standardized messages [1]), or digital maps (e.g., standard-definition or high-definition road topologies). The issue is that these domains have establish data formats, standards, and conventions for domain-specific use cases which suddenly clash into the common need to fuse multi-ADAS information real-time in a vehicle to provide the next step of autonomous driving.

Figure 1: iLDM implementation using Neo4j and RTMaps [3].

Provided each domain is governed by its own inertia, huge alignment efforts are required to interoperate perception, communication and digital map systems. Current approaches focus on creating inter-domain standards (e.g., ISO LDM, ASAM simulation branch), utilise multi-sensor application middlewares (e.g., RTMaps, ROS, ADTF) and apply semantic alignment of concepts [2].

One significant example which have attracted the attention of the automotive industry is the Local Dynamic Map (LDM) concept. Initially defined as an structure of road information categorized in layers (from static to dynamic information), methods to implement the structure itself have started to appear as a response to the publication of the LDM-related ETSI/ISO standards. In [3] an interoperable LDM (iLDM) implementation was presented, including a data model aligned with the recently published ASAM OpenLABEL standard [4], to leverage the interconnection of sensor data recording vehicle set-ups with ground truth generation.

Authors: Marcos Nieto, Mikel García, Itziar Urbieta (Vicomtech)

References

[1] ETSI EN 302 895 V1.1.1. Intelligent Transport Systems (ITS); Vehicular Communications; Basic Set of Applications; Local Dynamic Map (LDM) Basic Set of Applications-ETSI EN 302 895. 2014. Available online: https://www.etsi.org/deliver/etsi_en/302800_302899/302895/01.01.01_60/en_302895v010101p.pdf (accessed on 16 June 2021).

[2] Urbieta, I.; Nieto, M.; García, M.; Otaegui, O. Design and Implementation of an Ontology for Semantic Labeling and Testing: Automotive Global Ontology (AGO). Appl. Sci. 2021, 11, 7782.

[3] García, M.; Urbieta, I.;Nieto, M.; González de Mendibil, J.;Otaegui, O. iLDM: An Interoperable Graph-Based Local Dynamic Map.Vehicles 2022, 4, 42–59. https://doi.org/10.3390/vehicles4010003.

[4] ASAM. ASAM OpenLABEL V1.0.0. Available online: https://www.asam.net/project-detail/asam-OpenLABEL-v100/ (accessed on 6 July 2021).

The ongoing deployment of 5G mobile networks can enable the development of several new CCAM applications thanks to the increase of 5G network performance with respect to the previous mobile network generations. The CCAM application can exploit the large bandwidth and the reduced latency to provide innovative services related to safety and infotainment to a huge number of vehicles. Moreover, safety-related CCAM applications can leverage the softwarization of the 5G network to provide services with a higher level of reliability and with Service Level Agreements (SLAs) that can enable the real implementation of services that today are only feasible as demonstrator or Proof of Concept.

The spread of CCAM applications can be fostered by the availability of flexible communication boards. Third-party developers (e.g., SMEs) can take advantage of realistic testing environments to conceive and to develop new services devoted to the automotive sector.

One of the main objectives of 5G-IANA is to provide to third-party developers an open experimentation platform that includes the availability of testing 5G communication resources. This platform comprises boards devoted to automotive communications (V2X): On-Board Unit (OBU) for enabling communication from vehicles, and Road-Side Unit (RSU) for providing communication from the roadside. The purpose is to make available 5G-enabled V2X communication boards that can provide an adequate testing environment to CCAM developers.

The flexibility of the V2X communication boards is mainly ensured by the virtualization approach that is pursued in 5G-IANA. The V2X OBU and RSU will support virtualization features to easily support the onboarding of third-party applications. Furthermore, OBU and RSU will integrate management and orchestration features making the onboarding of applications an easy and automatic process.

Developers will be further facilitated by the availability of ready to use software modules that are offered on the V2X OBU and RSU. These components are related to the provision of communication capabilities or of basic functionalities that can be exploited by several applications. The developers can avoid starting the development from scratch by leveraging these already implemented components. This approach will let the developers to focus on the implementation of new complex and advanced services. They can exploit easy to use APIs to deal with the already available software modules reducing the development time and effort.

One difficult aspect in the development of CCAM application is indeed the complexity related to the automotive specific context. The main role of the provided basic components on the V2X OBU and RSU is to hide the complexity of the automotive world and let the developers focusing on the high-level behaviour of their applications. In this way, developers don’t have to be experts on specific aspects such as the format and encoding of C-ITS messages or about the management of position and time. Moreover, the CCAM applications can exploit a reliable and secure channel since message integrity and authentication is managed by the V2X OBU and RSU implementing the secure communication compliant to the relevant ETSI standards. The V2X OBU and RSU will provide 5G Uu communication interface, and they will also offer short-reach communication capability thanks to ETSI ITS-G5 and C-V2X modems.

Authors: Daniele Brevi, Edoardo Bonetto

The shift towards automation in the automotive sector is giving rise to a plethora of cooperative distributed applications characterized by Quality-of-Service (QoS) constraints on the underlying 5G communication and computing infrastructure. In turn, this has fostered efforts to estimate QoS conditions and pro-actively adjust the network and/or service configuration, especially in cases of expected QoS degradation. For instance, the 5G-Automotive Association (5G-AA) has already proposed a framework for mobile networks to deliver In-advance QoS Notifications (IQN) to applications, in the so-called context of Predictive QoS[1].

The estimation of upcoming QoS conditions heavily builds on the use of historical data so as to gain insights of future system behavior. In the case of Artificial Intelligence and Machine Learning (AI/ML) this corresponds to the training of a ML model with historical data, so as to later use it for inference, based on the current conditions. Training is typically characterized by the need for large volumes of data (and corresponding 5G bandwidth), the correspondingly high computation load and the typically loose latency requirements. In the context of Predictive QoS, multiple data sources can be envisioned, including the 5G mobile network, having the vehicle itself as the natural focal point of past QoS experience and contextual information.

Applying the well-established centralized ML practices, where data are collected to a central location (e.g., a data lake) for training purposes, followed by the dispatching of the trained ML model, immediately reveals a particularly challenging environment: (i) training data can be of high volume consuming non-negligible network resources for collection, and/or (ii) subject to privacy concerns such as vehicle trajectory, while (iii) the dynamics of mobility call for a continuous learning process able to adapt to evolving and short-term conditions. The advent of Distributed Machine Learning (DML), including Federated Learning (FL)[2] promises to address some of these challenges by realizing multi-node ML systems that bring the (un-)trained model to the data (and not the other way around) and hence (re-)train, collect and aggregate model instances in repetitive (a)synchronous steps (Figure 1).

Figure 1: The DML-FL Framework: 1) A Model Aggregator within the central cloud server selects a subset of clients and dispatches the current global model to them, 2) The clients perform local model training, 3) The clients upload the local models back to the central server, and 4) The central server aggregates the local models.

However, adopting DML/FL in the context of 5G-enabled services/applications presents significant practical challenges when it comes to the overall Management and Orchestration (MANO) processes. Realizing a DML/FL scheme requires the inclusion of vehicle On-Board Units (OBUs) and Road-Side Units (RSUs) within the broader operational scope of MANO processes, which comes with challenges related to the integration of the virtualization and programmability capabilities on the corresponding devices, as well as the integration within the MANO fabric in the presence of intermittent connectivity/availability. Then, the distributed character of the training process poses the requirement for advanced DML/FL MANO primitives, reducing the complexity of processes such as client selection, overlay topology formation and placement e.g., hierarchical FL, model aggregation and/or data transfer (where applicable), from the application-level implementation, down to the simple consumption of corresponding interfaces and/or the definition of corresponding policies.

Focusing on the case of AI/ML-enabled Predictive QoS, 5G-IANA identifies and addresses these challenges with the purpose of providing generic MANO primitives and NetApp VNF support for the realization of DML/FL services/applications. Our work will build on active/passive network monitoring data produced in NOKIA 5G testbed in Ulm, Germany, which consists of 5 sites-with 3 radio cells each. The monitoring data will be used to feed a DML/FL-enabled Predictive QoS service, with the purpose of eventually delivering IQNs for consumption by other services. The spatio-temporal dimensioning of the overall service will be carefully assessed, also feeding to the corresponding service Life-cycle Management (LCM) operations e.g., selection of ML model and/or model aggregation server corresponding to spatio-temporal QoS maps of the region of interest (Figure 2).

Figure 2: DML/FL-enabled Predictive QoS with geo-fencing: Single-model (instance) approach, adopting a global ML

model for all areas (top), versus multi-model (instance) approach, adopting multiple ML-model(s) instances per area (bottom).

Authors: Konstantinos Katsaros, Nehal Baganal-Krishna, Amr Rizk, Eirini Liotou, George Drainakis, Markus Wimmer, Steffen Schulz.

[1] 5GAA Automotive Association, “Making 5G Proactive and Predictive for the Automotive Industry,” White Paper, Dec 2019.

[2] H. B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. Y. Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Proc. 20th Int. Conf. Artif. Intell. Statist., 2017, pp. 1273–1282

ICCS, our project coordinator, introduces the project and provides a general introduction to its technical objectives, methodology and main expected outcomes

5G-based Automotive-related services (i.e., Connected and Automated Mobility services) are a broad range of digital services in and around vehicles including both safety-related and other commercial services provided, enabled, or supported by 5G networks. The rollout of 5G is expected to become a “game changer”. The prospect that 5G will be a unified multi-service platform, serving not only the traditional mobile broadband market but also enabling digital transformation in a number of vertical industries, is expected to result in the creation of unprecedented opportunities for innovation and economic growth.

In view of this opportunity, 5G-IANA aims at providing an open 5G experimentation platform, on top of which third party experimenters, i.e., Small and Medium Enterprises (SMEs) in the Automotive-related 5G-PPP vertical will have the opportunity to develop, deploy and test their services. An Automotive Open Experimental Platform (AOEP) will be specified, as the whole set of hardware and software resources that provides the computational and communication/transport infrastructure as well as the management and orchestration components, coupled with an enhanced NetApp Toolkit tailored to the Automotive sector. 5G-IANA will expose to experimenters secured and standardized Application Programming Interfaces (APIs) for facilitating all the different steps towards the production stage of a new service.

5G-IANA will target different virtualization technologies integrating different Management and Orchestration (MANO) frameworks for enabling the deployment of the end-to-end network services across different domains (vehicles, road infrastructure, Multi-access Edge Computing (MEC) nodes and cloud resources). 5G-IANA NetApp toolkit will be linked with a new Automotive Virtual Network Functions (VNFs) Repository including an extended list of ready to use open accessible Automotive-related VNFs and NetApp templates, that will form a repository for SMEs to use and develop new applications.

Furthermore, 5G-IANA will develop a Distributed Artificial Intelligence / Machine Learning (AI/ML) (DML) framework, that will provide functionalities for simplified management and orchestration of collections of AI/ML service components and will allow ML-based applications to penetrate the Automotive world, due to its inherent privacy preserving nature.

5G-IANA will be demonstrated through seven Automotive-related use cases in two 5G Stand Alone (SA) testbeds, which are:

1. Remote driving
2. Manoeuvres coordination for autonomous driving (vehicle-side and road-side manoeuvre coordination)
3. Infotainment
4. Augmented reality (AR) content delivery for vehicular networks
5. Parking circulation & high-risk driving hotspot detection
6. Network status monitoring
7. Situational awareness in cross border road tunnel accidents.

Moving beyond technological challenges, and exploiting input from the demonstration activities, 5G-IANA will perform a multi-stakeholder cost-benefit analysis that will identify and validate market conditions for innovative, yet sustainable business models supporting a long-term roadmap towards the pan-European deployment of 5G as key advanced Automotive services enabler.

INSTITUTE OF COMMUNICATION AND COMPUTER SYSTEMS (ICCS)