Elsevier

Computer Networks

Volume 172, 8 May 2020, 107150
Computer Networks

Collision avoidance in 5G using MEC and NFV: The vulnerable road user safety use case

https://doi.org/10.1016/j.comnet.2020.107150Get rights and content

Abstract

Automotive is considered one of the driving use cases for the 5th Generation (5G) systems, which currently formulates numerous scenarios and Key Performance Indicators (KPIs), via advanced Vehicle-to-everything (V2X) services and applications. Minimum end-to-end delay, as well as advanced contextual awareness requirements, pose novel architectural and functional challenges. This paper exploits two key enablers, namely Multiple Access/Mobile Edge Computing (MEC) and Network Function Virtualization (NFV), and acts in a two-fold manner: Firstly, it proposes a hybrid architecture for 5G systems, which exploits the afore-mentioned technologies, and performs computing resources’ selection among MEC and/or centralized, cloud-based resources (as VNFs), towards efficient service orchestration. The second contribution of this paper is a novel V2X service and algorithm, namely VRU-safe, that operates on top of the proposed architecture. VRU-Safe is an efficient, lightweight, low time complexity scheme, capable of identifying and predicting potential imminent road hazards between moving vehicles and Vulnerable Road Users (VRUs). The performance and viability of the proposed solutions are evaluated in a real-world 5G testbed in Europe.

Introduction

Fifth Generation (5G) networks promise highly flexible and programmable end-to-end communication aiming to further enhance several network performance aspects. Increased performance in terms of throughput, latency and reliability is expected in order to meet challenging requirements from diverse services and user demands [1]. Among 5G use cases, the automotive vertical domain establishes an undoubted key driver for 5G systems. Connected and fully automated vehicles will ultimately lead to safer transportation via ultra-low delay network performance with the support of the roadside infrastructure, aiming at radically reducing fatalities on the road, while at the same time achieving lower environmental impact [2].

Very recently, the 5G Infrastructure Public Private Partnership (5G-PPP) released an initial strategic agenda for Connected and Automated Mobility (CAM) in Europe [3]. In this agenda, key elements are highlighted in order to progress and stimulate investments into a pan-European network of 5G corridors, as a first strategic step towards large scale deployment and other high value services related to connected vehicles. Vehicle-to-everything (V2X) communication enablers have been introduced by 3GPP in Long Term Evolution (LTE) since Release 14 [4], via the Cellular V2X concept (i.e., C-V2X), - as well as its successor set of features, i.e., eV2X -, in several 5G use cases [5], [6], as well as by the European Telecommunications Standards Institute (ETSI) [7]. 3GPP aims to further enhance several aspects of V2X communications in Release 16 [8]. In particular, collision avoidance, VRU safety and hazardous situation detection is identified as one of the key use cases according to 5G Automotive Association (5GAA) [9]. Additionally, the requirements for different V2X use cases has been identified in [10], in relation to the end-to-end latency, reliability, data rate per vehicle, communication range, etc. In the same work, the specific requirements for the VRU, Vehicle-to-Pedestrian use case (V2P) have been identified as well, highlighting a maximum end-to-end latency of 100 ms, reliability higher than 95%, and data rate ranging between 5 and 10 kb/s. Also, in [11], the authors provide a comprehensive analysis on the V2X requirements according to ETSI and 3GPP, and focus on a thorough discussion on the 5G requirements, particularly for the Radio Access.

Towards this direction, 5G networks call for the support of multiple network slices on a common and programmable infrastructure, i.e. multiple logical networks with different configurations, optimized for specific traffic service types. Recent research efforts target to the design of 5G network slice(s) customized for vehicle-to-everything services, which involve vehicles exchanging data with each other, with the infrastructure and any communicating entity for improved transport fluidity, safety, and comfort on the road [12]. The potential of the network slicing concept for V2X has been also initially investigated by 3GPP [13] and also unveiled in [14], [15].

Although automotive-driven services involve different traffic types, each with diverse requirements, -primarily in terms of end-to-end latency, reliability and bandwidth-, these requirements could be clustered in two major kinds of applications: (a) large scale applications, that involve non delay-sensitive communications between road users and application servers located remotely, such as fleet management or traffic monitoring; (b) time critical applications, that deal with short-lived information (hence time-critical), such as vehicles’ dynamic parameters and sensor data, relevant in proximity of the area where this data been generated. To this end, its treatment in proximity to the users is also required. Anticipated Collision Avoidance is a typical use case requiring this kind of time-critical information processing.

Virtualized cloud resources offer robust performance, via powerful processing and computing capabilities. On the other hand, MEC capabilities are introduced towards the enablement of placing storage and computation resources at the network edge, in the proximity of the radio access network. By processing data locally and accelerating data streams through various techniques, MEC reduces both the end-to-end latency, as well as the traffic overhead towards the core network. A crucial trade-off, thus, results between the two architectural choices, with MEC-based processing, on the one hand, minimizing end-to-end communication delay; on the other hand, minimizing the processing/computing delays via robust and powerful virtualized cloud resources, having however to face higher transmission end-to-end delays between the RAN and the cloud.

The remainder of this article is organized as follows. Section II presents the state of the art. In Section III, we present in detail the proposed architecture and the VRU-Safe service. Section IV provides the performance evaluation-related details, the experimental set-up and discusses the obtained results. The article concludes in the final section, with a summary recapping the main findings and discussing the next steps of this work.

Section snippets

Related work

Recently, MEC-based solutions empowered the automotive sector facilitating the direct exchange of vehicle related information between (mobile) nodes via the underlying communication network, highlighting novel challenging use cases and applications [16], [17], [18], [19]. Resource-heavy applications will leverage from edge computing characteristics, by minimizing computational cost, latency and supporting data processing strategies [20]. Besides the end-to-end delay reduction, MEC-based

The proposed architecture and collision avoidance service

The proposed service, namely VRU-Safe, aims at timely predicting and avoiding anticipated collision events between VRUs and vehicles in their vicinity; at the same time, it builds on top of a novel architecture towards 5G, capable of dynamically steering the computing requests towards either MEC or centralized cloud computing resources (operating as VNFs), towards efficient V2X service operation and delay minimization for critical scenarios.

The end-to-end design, deployment and orchestration of

Evaluation

This section provides the evaluation outcomes of the proposed framework; in the first part we present the configuration of the collision identification algorithm, via evaluating the F1-score, Recall, Precision and Accuracy metrics for different values of the tolerance window, collision threshold and immediate threshold parameters; in the second part of the evaluation, we perform a real-world experiment in different mobility and intersection locations with real, human-controlled vehicles, in

Conclusion and future directions

A hybrid architecture for V2X services was presented, which exploits MEC and cloud-based resources in a coordinated manner towards optimizing the end-to-end performance of the system. On top of the proposed architecture a novel V2X service - namely VRU-safe - was presented and thoroughly evaluated in a real network wireless automotive testbed, and in a diverse set of mobility scenarios. All in all, it is shown that VRU-safe service is able to provide accurate and timely predictions of potential

Declaration of Competing Interest

None.

Acknowledgments

The research leading to these results has been performed under the H2020 project 5GinFIRE [48].

Dr. Sokratis N. Barmpounakis is a Postdoctoral Research Associate in SCAN group, Dept. of Informatics and Telecommunications, in the National and Kapodistrian University of Athens (NKUA). He holds a Ph.D. in “Context-based Resource Management and Slicing for SDN-enabled 5G Smart, Connected Environments”, since May 2018. Dr. Barmpounakis obtained his Engineering Diploma in Electrical and Computer Engineering, from the National Technical University of Athens (NTUA) in 2010. After completing his

References (48)

  • J. Wang et al.

    A survey of vehicle to everything (V2X) testing

    Sensors

    (2019)
  • N. Lubbe et al.

    Pedestrian crossing situations: Quantification of comfort boundaries to guide intervention timing

    Accident Anal. Prevent.

    (2014)
  • M. Simsek et al.

    On the Flexibility and Autonomy of 5G Wireless Networks

    (2017)
  • 5G Automotive Association (5GAA), Toward fully connected vehicles: Edge computing for advanced automotive...
  • 5G Infrastructure Public Private Partnership (5G-PPP), 5G Strategic deployment agenda for connected and automated...
  • J. Lee et al.

    LTE-Advanced in 3GPP Rel -13/14: An evolution toward 5G

    (2016)
  • S. Chen et al.

    Vehicle-to-Everything (v2x) Services Supported by LTE-Based Systems and 5G

    (2017)
  • 5G communication automotive research and innovation (5G Car) H2020 project, https://5gcar.eu/[Accessed November...
  • 5G communication automotive research and innovation (5G Croco) H2020 project, https://5gcroco.eu/[Accessed November...
  • Study on MEC Support for V2X Use Cases

    (Sept. 2018)
  • Toward Fully Connected Vehicles: Edge Computing White Paper

    (2017)
  • M. Boban et al.

    Connected roads of the future: Use cases, requirements, and design considerations for vehicle-To-everything communications

    IEEE Veh. Technol. Mag.

    (2018)
  • B.M. Masini et al.

    Radio access for future 5G vehicular networks

  • Toward fully connected vehicles: edge computing for advanced automotive communications, White Paper. Available...
  • 3GPP TR 23.786 v0.8.0. Technical specification group services and system aspects; Study on architecture enhancements...
  • C. Campolo et al.

    5G network slicing for vehicle-to everything services

    IEEE Wireless Commun.

    (2017)
  • S. Chen et al.

    Vehicle-to-Everything (v2x) Services Supported by LTE-Based Systems and 5G

    (2017)
  • F. Giust et al.

    Multi-Access Edge Computing: The Driver Behind the Wheel of 5G-Connected Cars

    (2018)
  • A Platform for Computing at the Mobile Edge: Joint Solution with HPE, Saguna, and AWS, Amazon White Paper, Available...
  • D. Grewe et al.

    Information-centric mobile edge computing for connected vehicle environments: challenges and research directions

  • P. Corcoran et al.

    Mobile-Edge Computing and the Internet of Things for Consumers: Extending Cloud Computing and Services to the Edge of the Network

    (2016)
  • K. Zhang et al.

    Mobile-Edge Computing for Vehicular Networks: A Promising Network Paradigm with Predictive Off-Loading

    (2017)
  • Emara, M., Filippou, C. M., Dario, S. (2017). MEC-aware Cell Association for 5G Heterogeneous...
  • A. Aissioui et al.

    On Enabling 5G Automotive Systems Using Follow Me Edge-Cloud Concept

    (2018)
  • Cited by (24)

    • Autonomous vehicles in 5G and beyond: A survey

      2023, Vehicular Communications
      Citation Excerpt :

      The authors mainly focus on service migration for V2X applications to reduce the duration of non-availability of services provided by the virtualized migrating instance (service downtime). Sokratis Barmpounakis et al. [79] proposed an architecture for 5G systems that uses MEC and NFV to select the computing resources for V2X applications. A novel algorithm, VRU-Safe, is proposed by the authors that operates on top of the proposed architecture to identify and predict road hazards like vehicle collisions.

    • Driving under influence: Robust controller migration for MEC-enabled platooning

      2022, Computer Communications
      Citation Excerpt :

      Intelligent Transportation Systems (ITS) of the future aim to increase road throughput, improve safety and reduce emissions from vehicles [2,3].

    • A survey on road safety and traffic efficiency vehicular applications based on C-V2X technologies

      2022, Vehicular Communications
      Citation Excerpt :

      In these works the focus was on latency, and the results highlight the importance of processing the CAM information to predict collisions close to the edge of the network: perhaps using MEC in the RSU, if D2D technologies are used [127]; or using MEC in base stations if Uu interfaces are used [121] [128]. The evaluation in [121] and [127] was done with real equipment in a lab environment, and system level simulations were carried out in [128]. A similar application for V2I event-driven messages is infrastructure-based collision warning, which is intended to prevent collisions between vehicles (for example, in intersections).

    View all citing articles on Scopus

    Dr. Sokratis N. Barmpounakis is a Postdoctoral Research Associate in SCAN group, Dept. of Informatics and Telecommunications, in the National and Kapodistrian University of Athens (NKUA). He holds a Ph.D. in “Context-based Resource Management and Slicing for SDN-enabled 5G Smart, Connected Environments”, since May 2018. Dr. Barmpounakis obtained his Engineering Diploma in Electrical and Computer Engineering, from the National Technical University of Athens (NTUA) in 2010. After completing his studies, he moved to Geneva, Switzerland. At University of Geneva, he worked for 2 years as a Researcher/Developer in various projects, Swiss and European, as a member of the QoL group. Since March 2013, he is a researcher in the SCAN group, actively involved in numerous European Projects and industrial contracts, having also undertaken managerial roles, coordinating the University of Athens team. Dr. Barmpounakis is a co- inventor of 2 patents, approved and filed in the World Intellectual Property organisation. His main fields of interest are: 5G, LTE, Software Defined Networking, Context-Aware Resource and Mobility Management and Collaborative Data Sharing mechanisms. He is also a member of the Technical Chamber of Greece.

    Mr. George Tsiatsios obtained his four-year Bachelor of Science (B.Sc.) from National and Kapodistrian University of Athens (NKUA), in the Department of Informatics and Telecommunications, with a specialization in Telecommunications and Networks. He is currently a student at NCSR Demokritos, in the field of Data Science and works as a Researcher at the Software Centric & Autonomic Networking (SCAN) Lab of UoA. His main fields of interest are AI, Data Analytics, Software-Defined Networking (SDN), Software-Defined Wireless Local Area Networking (SDWLAN), Future networks and 5G.

    Mr. Michael Papadakis graduated from the Department of Informatics and Telecommunications, NKUA, and is currently completing his M.Sc. at the University of Piraeus, Dept. of Digital Systems, Object Oriented Architecture and Services. He is collaborator with SCAN group, University of Athens and with TERRA SPATIUM, Geo-information and Space Products & Services S.A. His field of studies include Cloud Computing, SDN, SDR, 5G communications, as well as conceptual models, system analysis, decentralized architectures and smart maritime applications. He is currently researching on computer vision and deep neural networks on real time communications.

    Mr. Evangelos Mitsianis holds a Computer Engineer degree from University of Thessaly and a Msc. degree in Machine Learning from National and Kapodistrian University of Athens. Currently, he is Ph.D. candidate of National and Kapodistrian University of Athens in Informatics and Telecommunications department. He serves as research associate in Self-evolving Cognitive and Autonomic Networking Lab (Scan-Lab). Mr. Mitsianis main research interests are focus in designing reliable, fast and power-aware artificial intelligence and bigdata applications, using cutting edge technologies. Specifically, he is interested in algorithms and computational methods used in Data Science, Machine learning and Deep Learning for Image Analysis, Audio-Speech recognition, and Video scene recognition. Also, he focus on applied artificial intelligence for next generation telecommunication networks.

    Mr. Nikolaos Koursioumpas holds a M.Sc. degree with a specialization in Computer Networking at the Department of Informatics and Telecommunications, National and Kapodistrian University of Athens. He obtained his B.Sc from the same University with a specialization in Networks and Telecommunications. He is currently working as a research associate in Big Data Engineering in SCAN lab.

    Assoc. Prof. Nancy Alonistioti is an Associate Professor in the Department of Informatics Telecommunications (Dept. of Informatics and Telecommunications, NKUA). She has over 20 years of experience in numerous national and European projects, including project/ technical management experience. She is currently leading the SCAN group activities in the Dept. of Informatics in the National, NKUA. She has served as member of the Future Internet Assembly Steering Committee. She has also undertaken management activities as a PMT member and WP Leader of the FP6 IST- E2R and FP7 E3 projects and the PC of FP7 ICT Self-NET and FP7 ICT CONSERN projects. She is member of the ETSI Experts group and the Greek standardization group ELOT (5G, smart citiy autonomic communications). She has over 100 publications in the area of mobile networks, Future Internet/NGI, cognitive management, autonomic communications and reconfigurable mobile systems. She is co-author of 4 WO Patents and has more than 2500 citation in her work.

    View full text