UAV-Assisted IoT Applications, Cybersecurity Threats, AI-Enabled Solutions, Open Challenges with Future Research Directions
Muhammad Adil,Houbing Song,Spyridon Mastorakis,Hussein Abulkasim,Ahmed Farouk,Zhanpeng Jin
DOI: https://doi.org/10.1109/tiv.2023.3309548
IF: 8.2
2024-01-01
IEEE Transactions on Intelligent Vehicles
Abstract:Unnamed Ariel Vehicle-assisted-Internet of Things (UAV-assisted IoT) applications have emerged as a powerful integrated technology, showcasing remarkable results in many domains with numerous advantages. However, this technology encounters several challenges, and security is one of them. Given that, authentication and verification of legitimate devices with data privacy pose key concerns in a wireless communication environment. Even though, the literature highlights the growing security threats of this technology, where attackers can easily compromise the existing authentication and data preservation schemes. Therefore, it is crucial for all involved stakeholders to address these concerns using Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), and Reinforcement Learning (RL) algorithms, as they offer cost-effective solutions. While some algorithms have been used in the literature to accurately and effectively predict, detect, and prevent vulnerabilities in this technology, they may not adequately handle modern or advanced security threats. Therefore, in this article, we provide a comprehensive survey of the existing literature from 2014 to 2022, specifically focusing on AI, ML, DL, and RL-enabled prototypes. Our goal is to highlight the contributions and limitations of the considered articles. Based on our observations, we will emphasize on the open challenges to set the stage for future research to enhance the security of this emerging technology. Moreover, to bridge this gap of all aspects of this technology, we will discuss layer-wise security threats and countermeasure schemes following the TCP/IP stack. Finally, we will compare our work with existing review articles to demonstrate its novelty, uniqueness, and potential usefulness for the people working in this field.