Towards a trust-based heuristic for securing the internet of vehicles

<p dir="ltr">Over the past few decades or so, considerable technological advancements in wireless communication and sensor networking technologies have fostered the paradigm of the Internet of Vehicles (IoV) as an emerging reality, wherein vehicles not only communicate amongst one another but also with the supporting roadside infrastructure, vulnerable pedestrians, and backbone network in a bid to disseminate both safety-critical and non-safety messages. However, as a distributed network that is highly dynamic in nature, IoV networks offer a wide attack surface. Whilst the cryptographic-based solutions are capable of handling external attacks, the internal attacks are ones that pose a considerable challenge. Trust mechanisms are, therefore, highly indispensable for detecting and mitigating internal attacks in IoV networks. Trust, in the context of an IoV network, manifests the confidence of a vehicle (trustor) over the other (trustee) for performing a particular task. Accordingly, vehicles possessing a higher trust score gets more privileges in an IoV network, thereby making trust a lucrative target for attackers. Consequently, trust-based attacks, including but not limited to, good-mouth attacks, bad-mouth attacks, on-off attacks, and opportunistic service attacks, attempt to mislead trust-based mechanisms to misplace trust and hence jeopardize the IoV network for malicious gains. The impact of a trust-based attack can be substantial since it could put the lives of drivers and pedestrians at risk. Moreover, trust-based attacks are typically sophisticated and well-disguised, thereby making them quite difficult to detect and mitigate.</p><p dir="ltr">This research thesis, therefore, offers a comprehensive survey pertinent to IoV-based trust management by delineating the (a) state-of-the-art trust-based mechanisms, i.e., non learningbased (conventional) and learning-based (traditional machine learning, reinforcement learning, and deep learning) ones, (b) trust-based attacks in an IoV network, and (c) trust factors (attributes), datasets, and simulators employed by the state-of-the-art trust-based mechanisms. One of the key contributions of this research thesis is envisaging of a trust heuristic that is capable of detecting and mitigating the trust-based attacks, i.e., self-promotion attacks, on-off attacks, and opportunistic service attacks, in an IoV network. The said heuristic takes into consideration direct trust that employs subjective logic theory and the stability of a vehicle’s behavior over the time in an IoV network, indirect trust along with the credibility of the recommending vehicles, and a dynamically weighted role-based trust. Extensive simulations have been carried out and a critical analysis pertinent to the same has been delineated to demonstrate the efficacy of the envisaged trust-based heuristic.</p>