Elsevier

Information Fusion

Volume 50, October 2019, Pages 212-220
Information Fusion

Full Length Article
An efficient key distribution system for data fusion in V2X heterogeneous networks

https://doi.org/10.1016/j.inffus.2019.02.002Get rights and content

Highlights

  • A four level data fusion trust model in V2X heterogeneous networks.

  • An efficient location-based key distribution system to realize high trust level.

  • Deployment of GPGPU to accelerate security related calculation tasks.

Abstract

Data fusion in Vehicle-to-Everything (V2X) networks for different data types coming from different sources is the foundation for decision making in the smart vehicle driving systems. Different communication technologies have been combined to form a heterogeneous V2X network to support the data exchange. However, data fusion trust models are still designed for single use cases which cannot meet the general needs of Cooperative Intelligent Transport Systems (C-ITS). In this paper, we first define a data fusion trust architecture with different trust levels. Then, we propose an efficient and practical data fusion trust system for the multi-source and multi-formats of data exchange in the V2X heterogeneous networks. In particular, a location-based PKI system with acceleration brought by General Purpose Graphic Processing Unit (GPGPU) is presented for efficient key distribution with a high level of trust achieved. A performance evaluation is given to verify our data fusion trust system can meet the strict latency requirements in V2X networks.

Introduction

Data fusion in Cooperative Intelligent Transport Systems (C-ITS) is truly compelling to make transportation safer and more efficient in recent years. The data based applications deployed in C-ITS are promising to deeply change people’s driving experience by reducing traffic congestion, providing intelligent transportation algorithms, reducing significantly the number of traffic accidents, and finally realizing unmanned vehicles [1].

With the development of hardware and software, data can be collected and fused on different C-ITS components. We are in the big data era for transportation based communication systems. Smart modules, as shown in Fig. 1, such as vehicle cameras, sensors, and smart control systems are deployed on vehicles for data collection purpose. Thus, data collected from these modules both in physical environments and cyber systems is increasing rapidly in both volume and number of types [2]. On the other hand, improvement of data fusion and analytic algorithms especially artificial intelligent algorithms have become more and more efficient to meet the real-time data fusion needs in C-ITS networks.

Another important and fundamental technology for data fusion in C-ITS networks is the wireless communication between vehicles and other components in C-ITS networks. The two main categories of wireless communication technologies for V2X communication systems are IEEE 802.11p standard based [3] and cellular based [4]. For the IEEE 802.11p standard based communication technology, there are Dedicated Short-Range Communications (DSRC) standards in US [3] and Intelligent Transportation System (ITS)-G5 standards in Europe [3]. In fact, IEEE 802.11p can meet most V2X application requirement with the most stringent performance specifications. However, the cellular based communication technology such as LTE, 5G [4] can also be used as another option to build direct and stable network links such as for communication between vehicles and remote servers [4]. In Fig. 1, an example is given to illustrate the different communication technologies are deployed as a heterogeneous network for data exchange and data fusion based decision makings.

For data exchange in this heterogeneous communication network, concept of Vehicle-to-Everything (V2X) communication has been introduced and classified according to the communication parties [5]: Vehicle-to-Vehicle (V2V), Vehicle-to-Pedestrian (V2P), Vehicle-to-Infrastructure (V2I), and Vehicle-to-Network (V2N). Basically, this concept is used to sort the collected data in the application layer further for the data fusion based decision making systems on one vehicle. Among the four types of V2X communications, V2V and V2I are the most promising technologies for data exchange used for future C-ITS networks [5].

Data fusion and data analytics based on data collected from onboard sensors (cameras, sensors, radars, and control systems) and received from the heterogeneous V2X networks can derive high-level intelligent decision makings. In turn, these intelligent decisions made on each vehicle can then be sent to other components in C-ITS network for improving the traffic system. For drivers, cooperative data fusion can provide a better vision of an intersection and help detect other vehicles or pedestrians that the driver would overlook due to complex driving environments like obstacles, distraction, or bad weather. In the future, data fusion based applications on V2X heterogeneous networks will help to realize the unmanned vehicles on roads.

With the introduction of data fusion in V2X communication, data exchange between vehicles and other components in C-ITS networks is complicated as different data formats are transmitted through heterogeneous networks. Some of these data types are related to driving and safety applications, especially in an unmanned driving environment. Thus, one important issue is, based on one vehicle’s viewpoint, whether the data received from the V2X networks, can be trusted especially it is used for safety-related decision makings. For example, an attacker might compromise the transmitting system to intentionally send false or wrong data to compromise the whole C-ITS network. Also, data transmitted in the V2X network might be changed by an attacker to mislead vehicles which may lead to traffic accidents. Privacy may also be violated as eavesdropping and illegal tracking may happen as an attacker can receive the periodical broadcasting Cooperative Awareness Messages (CAM) [6] containing vehicles’ geographical position, speed, and some sensitive information.

Countermeasures like security and privacy protection schemes are introduced to build the trust system for data exchanged in V2X networks. Research works such as [7] have deployed a Public Key Infrastructure (PKI) to distribute temporary keys for providing security and privacy for data exchange in V2X networks. However, extra latency is introduced by security schemes such as key distribution and other security related calculations such as algorithms providing data source authentication, data content integrity, and data encryption. V2X is naturally not expected to have high latency toleration as safety-related applications require ultra-low latency (e.g. pre-crash sensing warning normally require a total latency of less than 50 ms [4]). Thus, building an efficient data fusion trust system for data exchanged in the V2X heterogeneous networks is the fundamental step for further data fusion in C-ITS.

In this article, our main contribution is to propose an efficient data fusion trust system including (1) a location-based key distribution scheme that can highly reduce latency due to current key distributions; (2) an acceleration computing solution by fully deploying the vehicles’ onboard GPUs for security related calculations.

We organize this paper following the order below. Section 2 gives an example of the specific question we are going to solve. Section 3 illustrates our proposed location-based key distribution as well as several details. Next, Section 4 shows the parallel computing model for accelerating security related calculations. Furthermore, Section 5 shows performance evaluation results for the proposed data fusion trust system. Section 6 presents a brief discussion and future works. Finally, Section 7 draws the conclusion of this study.

Section snippets

Problem definition

As mentioned in Section 1, from viewpoints of vehicles, data fusion in a V2X heterogeneous network must consider the issue that whether the data received through V2X networks can be trusted. Previous works [8] have demonstrated the related concepts of trust of data fusion in V2X networks. However, a general data fusion trust system architecture is still lacked. In this section, we first define four trust levels for the data fusion trust system which, to the best of our knowledge, is the first

A location-based key distribution system

In this section, we present a location-based key distribution system with pre-calculated and pre-distributed key pair pools for vehicles and RSUs. In this scheme, updating the temporary keys is not based on time period but based on the locations of the vehicles. For one time period (e.g. 24 h), temporary keys (ATs) are pre-distributed and downloaded to vehicles and RSUs. As long as one vehicle moves to one zone, the temporary keys will be updated according to its location. Thus, the updating

GPU accelerated security calculations

The latency in the V2X communication systems can also be increased as the security related calculations are introduced as shown in Section 2. In this section, we first list all security related calculations based on the level of trust we designed. Then a GPU based design will then be presented to give an evaluated performance compared with existing solutions.

Performance evaluation

In this section, the evaluation of performance is presented for our data fusion trust model. We simulate the heterogeneous network of V2X communications on NS-3 simulator [27] including short-range communication network between vehicles and RSUs and Internet protocols between RSUs and the key distribution center. Average latency due to key distribution is calculated for the location-based key distribution system with different prediction success ratio.

Discussions and future work

In this paper, we introduced the four level trust model for the data transmission on the V2X heterogeneous network. In fact, the key distribution process can also be seen as one of the data transmission process which must be deployed with the highest trust level. We hope for future work, a more general data transmission trust model for data fusion in V2X can be presented with including key distribution.

We claimed that a key distribution system such as a PKI is necessary for implementing the

Conclusions

In this article, we presented a data fusion trust system for V2X heterogeneous networks. Firstly a definition with four levels of data fusion trust is given according to V2X application needs. Then a location-based key distribution scheme is designed to efficiently distribute keys that build a foundation for all levels of data trust implementations for data fusion in V2X networks. Security related calculations are also considered and accelerated by GPGPU for data exchange including both secure

Acknowledgement

This work has been partially conducted in SCOOP, a franco-european pilot project for connected vehicles and roads under EC Grant No.2014-EU-TA-0669-S and INEA/CEF/TRAN/A2014/1042281.

This work is also supported by National Natural Science Foundation of China under Grant No.61873309, No.61572137, and No.61728202, and Shanghai 2018 Innovation Action Plan project under Grant No.18510760200-Research on smart city big data processing technology based on cloud-fog mixed mode.

References (29)

  • X. Tang et al.

    A hierarchical reliability-driven scheduling algorithm in grid systems

    Journal of Parallel and Distributed Computing

    (2012)
  • R. Glabb et al.

    Multi-mode operator for SHA-2 hash functions

    J. Syst. Archit.

    (2007)
  • C. Bettisworth et al.

    Status of the Dedicated Short-Range Communications Technology and Applications: Report to Congress

    Technical report

    (2015)
  • K. Abboud et al.

    Interworking of DSRC and cellular network technologies for V2X communications: a survey

    IEEE Trans. Veh. Technol.

    (2016)
  • A. Filippi et al.

    Ready to roll: why 802.11p beats LTE and 5G for V2X

    A White Paper by NXP Semiconductors

    (2016)
  • X. Wang et al.

    An overview of 3GPP cellular Vehicle-to-Everything standards

    GetMobile

    (2017)
  • ETSI

    302 637-2 v1. 3.1-Intelligent Transport Systems (ITS); Vehicular Communications; Basic Set of Applications; Part 2: Specification of Cooperative Awareness Basic Service

    (2014)
  • F. Haidar et al.

    On the performance evaluation of vehicular PKI protocol for V2X communications security

    IEEE 86th Vehicular Technology Conference (VTC-Fall)

    (2017)
  • M. Obst et al.

    Multi-sensor data fusion for checking plausibility of V2V communications by vision-based multiple-object tracking

    IEEE Vehicular Networking Conference (VNC)

    (2014)
  • P. Bellavista et al.

    V2X protocols for low-penetration-rate and cooperative traffic estimations

    IEEE 80th Vehicular Technology Conference (VTC Fall)

    (2014)
  • B. Glas et al.

    Prime field ECDSA signature processing for reconfigurable embedded systems

    International Journal of Reconfigurable Computing

    (2011)
  • M. Khalil-Hani et al.

    Hardware acceleration of openssl cryptographic functions for high-performance Internet security

    IEEE International Conference on Intelligent Systems, Modelling and Simulation (ISMS)

    (2010)
  • R.K. Dahal et al.

    Performance analysis of SHA-2 and SHA-3 finalists

    Int. J. Cryptogr.Inf. Secur.

    (2013)
  • J. Froehlich et al.

    Route Prediction from Trip Observations

    Technical Report, SAE Technical Paper

    (2008)
  • Cited by (57)

    • LPPTE: A lightweight privacy-preserving trust evaluation scheme for facilitating distributed data fusion in cooperative vehicular safety applications

      2021, Information Fusion
      Citation Excerpt :

      However, while the above schemes provide many brilliant ideas, there is no special consideration about potential attacks in cooperative vehicular safety applications. Nevertheless, due to the unique characteristics of vehicular networks, cooperative vehicular safety applications are vulnerable to multiple kinds of attacks which will greatly influence the accuracy of fusion results [5,11]. To improve the accuracy of fusion results, all the untrustworthy data should be filtered out before the distributed data fusion [5,12].

    • A trusted consensus fusion scheme for decentralized collaborated learning in massive IoT domain

      2021, Information Fusion
      Citation Excerpt :

      In the scene of massive IoT systems, there are many demands of data fusion and cooperation between devices[1,2].

    View all citing articles on Scopus
    View full text