Computer Sciences and Information Technology
Title : | Detection of Active and Passive Security Attacks using Ensemble Learning Technique for 5G Networks |
Area of research : | Computer Sciences and Information Technology |
Focus area : | Cybersecurity |
Principal Investigator : | Dr. Gunaseelan K, Anna University, Chennai, Tamil Nadu |
Timeline Start Year : | 2023 |
Timeline End Year : | 2026 |
Contact info : | guna_2012@yahoo.co.in |
Details
Executive Summary : | Fifth-generation wireless (5G) is the latest generation of cellular technology, engineered to greatly increase the speed and responsiveness of wireless networks. In the meantime, the security and privacy of 5G wireless networks are of the utmost important. However, owing to the broadcast characteristics of wireless communications, 5G wireless networks are vulnerable to physical layer security threats, such as spoofing attack, eavesdropping and Denial of Service. In the spoofing attack, the attacker can pretend to be a legitimate user using a faked identity, such as a Media Access Control (MAC) address and Internet Protocol (IP) address, then it may gain illegal benefits to further perform advanced attacks, like man-in-the-middle attacks and denial-of-service attacks. In this project work, Channel-based physical-layer security technique using Ensemble learning will be used to detect these types of attacks. This introduces a channel-based attack detection scheme based on channel virtual representation. The Principal Components of Channel Virtual Representation (PC-CVR) are extracted as a new channel feature. Compared with traditional channel features, the proposed features can be more sensitive to the location of transmitters and more suitable to 5G networks. Based on PC-CVR, there are two detection strategies to achieve the attack detection by tackling static and dynamic radio environments, respectively. For the static radio environment where the channel correlation is stable, Neyman-Pearson (NP) testing-based attack detection is provided. For the dynamic radio environment where the channel correlation is changing, ensemble learning-based algorithms are used. In the proposed technique, the features are fed to different Base learner algorithms followed by Meta-Learner that uses ensemble learning algorithms, such as Bagging, Stacking and Boosting to detect and classify types of attacks to protect critical data. The optimisation is also performed for efficient learning and classification. The proposed model achieves better performance than the existing approaches in terms of accuracy and computational time as Ensemble Learning is reliable efficient than other individual Machine Learning models. |
Total Budget (INR): | 19,50,690 |
Organizations involved