Federated Learning-Based Intrusion Detection System for Mobile Networks

Abstract

In today’s highly connected digital world, mobile phones are no longer just communication tools—they have become essential for accessing online services, managing finances, social networking, and storing personal and professional data. With this increasing reliance on mobile devices, mobile networks have become attractive targets for cybercriminals. Vulnerabilities are caused by malware, phishing attacks, and unauthorized access. Rising at a fast pace, these technologies put users’ data and privacy at risk. Traditional Intrusion Detection Systems (IDS), which monitor the network traffic and detect suspicious activity, are commonly used to address these issues. However, most conventional IDS approaches rely on centralized data collection and analysis, which creates significant challenges, including privacy violations, increased network bandwidth consumption, and a lack of scalability across millions of devices. This thesis presents a novel approach by proposing, in other words, Intrusion Detection Systems (IDS), specifically, this is the network traffic monitoring and detection of a Federated Learning (FL)-based intrusion detection system, which is developed explicitly in mobile networks. Federated Learning is the machine learning method that allows training models, without submitting raw data to a central server, in a decentralized way by operating on user devices That is, Intrusion Detection Systems (IDS), namely, monitoring and determination of the network traffic of a Federated Learning (FL) based intrusion detector, which, in turn, is particularly developed in mobile networks FL-based IDS performs competitively when compared to traditional centralized systems, while offering significant advantages in privacy preservation and communication efficiency. This research contributes a practical, secure, and efficient solution for modern mobile network environments. It also opens the door for future studies on how privacy-preserving technologies like FL can be applied to cybersecurity problems, particularly in areas such as Internet of Things (IoT), smart cities, and mobile cloud computing.