Developments in the Application of Machine Learning Techniques for Mobile Ad-Hoc Network Security

The self-configuring and decentralised network system basically known as Mobile Ad-hoc Networks or in short MANETs are essential for a number of applications, including smart cities, battlefield operations, disaster mitigation and many more. However, they are vulnerable to several security concerns, including routing-based attacks such as black hole and grey hole acts of terrorism, due to their dynamic topology, inherent lack of centralised management, and a lack of resources. The present review article explores the possible risks that multiple methods of machine learning may offer to MANETs and how they can be mitigated. This research looks at several attack strategies, classifies them as either external or internal, active or passive and evaluates the shortcomings of protection methods that are currently on the market. Going through various runs of evaluation and assessment of recent improvements in domains to study the advantages in addition to the drawbacks that applies algorithmic and deep machine learning (ML) strategies to enhance MANET security. Furthermore, the studies delves into innovative detection methods, such as using a mixture of algorithms and deep neural networks that assesses their impact through different scenarios at the same time as imparting an extensive review of gift improvements and potential future directions in MANET security.