Machine learning algorithms for blockchain-based security mechanisms in UAVs: a review

Within the last two decades, communication technologies have evolved and continue evolving rapidly, and this caused significant advancements in Internet of Things applications and services. Unmanned aerial vehicles (UAVs), popularly known as drones, have numerous emerging applications in various domains of life and have attracted the interest of researchers because of their fundamental attributes such as mobility, flexibility, reliability, and energy efficiency in wireless networks. Similar to wireless sensor networks and other remote sensing applications, these devices are generally not designed with integrated security mechanisms. Furthermore, the existing solutions in which the main focus is on drone communication security are quite limited in the literature. Blockchain (BC) is defined as a list of data blocks as a publicly distributed ledger, which is linked together using cryptography such as hash functions and public key infrastructures. In this matter, cryptographic security issues have increased. Conventional UAVs mostly depend upon centralized servers, where the authentication mechanisms run. Therefore, in recent years, new, advanced authentication/privacy protection techniques have been designed and developed for the diversity of computer security requirements resulting in the introduction of different kinds of security models. One of these suggestions is to use a decentralized BC-based server to secure information. In order to make precise decisions, machine learning (ML) techniques are required to analyze and decide on secure data in the BC-based server. There are various traditional ML techniques, such as support vector machines, convolutional neural networks, and long short-term memory to analyze the attacks on a BC-based network. However, there are a limited number of detailed studies in the literature on ML adoption for making BT-based networks more resilient against attacks. In this paper, various ML algorithm applications based on BC-based cryptographic algorithms for UAVs are discussed in detail, with some future research directions given.