Anomaly Recognition in Wireless Ad-hoc Network by using Ant Colony Optimization and Deep Learning

Abstract

As a result of lower initial investment, greater portability, and lower operational expenses, wireless networks are rapidly replacing their wired counterparts. The new technology that is on the rise is the Mobile Ad-Hoc Network (MANET), which operates without a fixed network infrastructure, can change its topology on the fly, and requires no centralised administration to manage its individual nodes. As a result, MANETs must focus on network efficiency and safety. It is crucial in MANET to pay attention to outliers that may affect QoS settings. Nonetheless, despite the numerous studies devoted to anomaly detection in MANET, security breaches and performance difficulties keep coming back. There is an increased need to provide strategies and approaches that help networks be more safe and robust due to the wide variety of security and performance challenges in MANET. This study presents outlier detection strategies for addressing security and performance challenges in MANET, with a special focus on network anomaly identification. The suggested work utilises a dynamic threshold and outlier detection to tackle the security and performance challenges in MANETs, taking into account metrics such as end-to-end delay, jitter, throughput, packet drop, and energy usage.