Blockchain Based Security and Trust Mechanisms for Vehicular Ad hoc Networks

Alharthi, Abdullah and Ni, Qiang and Jiang, Richard (2023) Blockchain Based Security and Trust Mechanisms for Vehicular Ad hoc Networks. PhD thesis, Lancaster University.

Text (2023AbdullahPhD)
2023AbdullahPhD.pdf - Published Version
Download (2MB)

Abstract

In the near future, intelligent vehicles (IV) will be part of the Internet of Things (IoT) and will offer valuable services and opportunities that could revolutionize human life in smart cities. The Vehicular Ad-hoc Network (VANET) is the core structure of intelligent vehicles. It ensures the accuracy and security of communication in vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) modes to enhance road safety and decrease traffic congestion. However, VANET is subject to security vulnerabilities such as denial-of-service (DoS), replay attacks and Sybil attacks that may undermine the security and privacy of the network. Such issues may lead to the transmission of incorrect information from a malicious node to other nodes in the network. Therefore, a biometrics blockchain (BBC) framework to secure data sharing among vehicles in VANET and to retain statuary data in a conventional and trusted system is designed. In the proposed framework, the biometric information is used to keep a record of the genuine identity of the message sender, thus preserving privacy and provide conditional anonymity. The suggested BBC approach provides security and trust among vehicles in VANET, as well as the capability to track identities as needed. To show the feasibility of the suggested framework utilizing the urban mobility model, simulations in OMNeT++, veins, and SUMO were performed. The framework's performance is assessed in respect to packet delivery rate, packet loss rate, and computational cost. The results demonstrate that our unique model outperforms previous techniques. Vehicles, on the other hand, find it challenging to assess the authenticity of received messages in non-trusted environment. The primary challenges in VANET are trust, data accuracy, and dependability of data broadcast over the communication channel. To protect against these threats, the majority of researchers have proposed cryptography-based solutions to verify the sender's legitimacy but are incapable of preventing the broadcast of false or malicious messages from a legal sender. Therefore, in this thesis, we propose a formal technique to compute and classify trust for vehicles in networks. Vehicles can evaluate the received messages, calculate the vehicle’s reputation, and check the message correctness based on numerous factors. The latest reputation of the vehicle will be stored on the blockchain. A machine learning approach is employed to classify the trust. Comprehensive tests are carried out using the dataset in order to measure the efficiency of the proposed ensemble-based learning and feature selection based on random forest. The outcomes of the experimentations reveal that the suggested ensemble learning approach with attribute selection produces an accuracy of 99.98%, which is higher than the baseline models studied in this thesis.