Securing the internet of vehicles – A trust management approach

Over the past few decades, the technological advancements in Vehicular Ad hoc Networks (VANETs) and the Internet of Things (IoT) have brought forth the promising paradigm of the Internet of Vehicles (IoV) which has attracted the attention of numerous researchers from both academia and industry. Today, this promising wireless communication technology plays an indispensable role as vehicles exchange low-latent safety critical messages with one another in a bid to make the road traffic more safer, efficient, and convenient. However, dissemination of malicious messages within the network not only significantly reduces the network performance but also becomes a source of threat for the passengers and the vulnerable pedestrians. Accordingly, a number of trust models have been recently proposed in the literature to ensure the identification and elimination of malicious vehicles from the network. These trust models primarily rely on the aggregation of different trust attributes, e.g., direct and indirect observations, and further evict malicious vehicles based on a particular threshold set on this composite trust value. Nevertheless, quantification of these trust attributes along with the weights associated with them and setting-up of the said threshold pose significant challenges especially owing to diverse influential factors in such a dynamic and distributed networking environment.

Accordingly, this thesis delineates on the convergence of the notion of trust with the IoV primarily in terms of its underlying rationale. It further sheds some light on the state of the art in the vehicular trust management, IoV architecture, and open challenges in the subject domain. Moreover, multiple unique trust management models have been developed with distinct features and objectives, including but not limited to, the quantification of influencing trust attributes, quantification of weights affiliated with these attributes, integration of context information, threshold definition, time-variant behavior analysis, and malevolent conduct detection by employing machine learning.

The first major contribution is conducting a comprehensive survey on the state of the art in the vehicular trust management focusing on the essential factors such as quantification of weights, quantification of threshold, misbehavior detection, attack resistance, etc. It further presents an overarching IoV architecture, constituents of the notion of trust, and attacks relating to the IoV in addition to open research challenges in the area of interest.

The second key contribution is proposing a novel trust management mechanism that utilizes context information in addition to employing relevant impacting quantities as weights to formulate trust evaluations. The primary emphasis is the quantification of weights associated with the contributing trust attributes and incorporating (a) attack resilience while constituting certain parameters (i.e., direct and indirect trust) and (b) an adaptive and flexible threshold to mitigate malevolent behavior.

The third significant contribution is analyzing the time-based behavior of the aggregated trust along with the contributing parameters to study the behavior of each vehicle and to identify suitable trust-based patterns for safety-critical and non-safety vehicular applications. Prior to the said analyses, a trust management model has been developed employing a different dataset (i.e., a real IoT dataset) to address the challenges of quantification of the influencing trust attributes and the weights associated with these parameters, subsequently, quantifying the aggregated trust. Later, the time-varying analysis of both trust evaluation frameworks, i.e., the one mentioned in the first contribution and the one indicated in this contribution has been presented.

The fourth main contribution is employing machine learning techniques to mitigate the need (a) to assign weights to the contributing parameters manually and (b) to define an optimal threshold value. It thus computes the feature matrix for four parameters in two different ways, (a) all of the parameters computed by each trustor for a trustee are treated as individual features, and (b) the mean of each single parameter computed by all of the trustors for a trustee is regarded as a collective feature. Different machine learning algorithms were employed for classifying vehicles as trustworthy and untrustworthy.