Hybrid Cyber-Physical Threat Detection in 5G-IoT Networks Using Deep Learning and Blockchain

Abstract

The implementation of 5G technology together with extensive Internet of Things systems provides three communication services which include ultra-reliable low-latency communication and improved mobile broadband and extensive machine-type communication. The diverse types of Internets of Things devices which operate at high densities create multiple new points of attack that enable networks to suffer from advanced persistent threats and distributed denial-of-service attacks and spoofing and data injection attacks. This paper presents a blockchain-based deep learning system which detects threats in 5G Internet of Things networks. The proposed architecture consists of three components which include convolutional neural networks and long short-term memory networks and smart-contract-based decentralized authentication mechanisms. The researchers tested the system by using an established object detection benchmark which showed that it worked successfully in real-world usage scenarios. The results showed that the system achieved better performance in scalability and reliability and compliance with QoS standards and computational efficiency and security and privacy protection and user experience enhancement. The system detects 98.4% of threats while keeping latency below 10 milliseconds which meets the requirements of URLLC standards.