Michalis Savva’s research while affiliated with University of Cyprus and other places

Internet of Things (IoT) devices face unique security challenges due to their inherent limitations such as limited storage, low computational power, and energy-efficient wireless communication. Traditional security measures, designed for the legacy Internet, fail to adequately protect IoT devices and networks. Particularly vulnerable are Wireless Sensor Networks (WSN) and IoT networks that are susceptible to jamming—a type of attack that significantly threatens wireless networks due to their open nature and the simplicity of launching such attacks. Perpetrators can initiate jamming without specialized hardware or in-depth knowledge of the targeted system. Despite advances in wireless technologies, the ability to thwart jamming attacks in real-world scenarios remains limited, as evidenced by the vulnerability of current security protocols of cellular and Wi-Fi networks. This thesis addresses the critical need for practical anti-jamming strategies to enhance the security of wireless networks, particularly against intelligent jammers that employ advanced machine-learning algorithms to adapt to more sophisticated attack methods such as constant, deceptive, random, or reactive jamming. These intelligent attackers can adjust their strategies and even manipulate detection systems to evade identification. To counter these threats, this dissertation introduces a novel lightweight security framework that utilizes fuzzy logic algorithms to enhance the detection, localization, and recovery mechanisms against jamming attacks in IoT networks. The framework employs network layer metrics to detect jamming at the node level, utilizes a modified multilateration technique to accurately locate jammers, and implements recovery strategies by blacklisting the affected nodes and rerouting traffic within the RPL network. This thesis makes several noteworthy contributions representing a significant IoT security advancement. By applying fuzzy logic to combine crucial metrics from the data link and network layers, the proposed framework not only detects jamming incidents, but also precisely pinpoints their origin, which is essential for effective mitigation. This thesis performs accurate real-time detection and localization using data link and network-layer metrics collected and processed at the edge. Furthermore, the framework's capability to blacklist and recover from compromised network paths introduces a dynamic recovery mechanism that enhances network resilience. Additionally, this thesis introduces a novel jammer called the complex jammer, in which the proposed framework has been accurately identified. Moreover, the framework effectively demonstrates the suitability of fuzzy logic for accurately recognizing multiple jamming attacks in diverse situations, with high accuracy, low memory usage, and quick execution. The effectiveness of this framework was validated through extensive simulations, demonstrating its capability to handle multiple jammers and adapt to evolving jamming strategies, thus significantly improving the resilience of IoT networks against these pervasive threats.

In recent years, the concept of the Internet of things (IoT) has become ubiquitous among several applications. However, traditional localization methods face challenges in accurately measuring sensor node positions, particularly in scenarios involving wireless signal instability, energy harvesting, anchor mobility, latency, and dense environmental conditions. However, these problems open the new ways to utilize machine learning (ML) and self-calibration for localization. To defeat these problems, in this paper, we have proposed a new localization technique, which reduces the localization error caused by the iterations in a triangulation method and drastically shortens the computation time for full network coverage despite noisy environments. First, we present a localization technique based on a looped network. The link quality induction (LQI) is used to formulate multiple feature vectors for the localization problem, solved using linear regression. Second, we examine the appropriateness of a few algorithms under different trade-offs by testing the localization accuracy for different network parameters, including the anchor node density, additive noise, and quality of the wireless channel. The simulation achieves exciting results that the adoption of a support vector machine (SVM) and nonlinear regression model applied with a radial basis function (RBF) kernel leads to an enhanced localization error with Root-Mean Square Error (RMSE) of 0.25m. Furthermore, we gain 48.5% accuracy in the localization error by implementing two models.