A Proactive Triple-Mutation MTD Architecture Enhancing AI-Driven Intrusion Detection Against Intelligent Adversaries in CIoT Networks
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
With the widespread deployment of intelligent consumer-electronics devices and the rapid expansion of consumer Internet-of-Things (CIoT) ecosystems that support next-generation intelligent systems—including smart homes, edge-AI gateways, and wearable networks—the volume and heterogeneity of device-generated network traffic have surged. To mitigate the limitations of traditional static AI-driven network intrusion detection systems (NIDS), proactive defence paradigms such as Network Moving Target Defence (NMTD) have been incorporated into device–edge–cloud architectures. Unfortunately, sophisticated adversaries may infer defence strategies and circumvent these mechanisms—leading to what we term the drug resistance of intelligent adversarial attacks. To counter this, we propose a reinforcement learning–based proactive Triple-Mutation Moving Target Defense (TM-MTD) framework that simultaneously mutates access proxies, IP addresses, and security strategies. The lightweight edge-level salience optimization ensures accurate detection in highly dynamic traffic environments, while the global policy differential optimization strengthens the defense system’s long-term resilience. Experimental results demonstrate that TM-MTD significantly enhances security performance against a variety of intelligent attack models, providing a robust defense framework for next-generation intelligent consumer-electronics systems.
Fields of Study & Subjects
Network Security and Intrusion Detection
Adversarial Robustness in Machine Learning
Network Packet Processing and Optimization
Publication details
- Journal / source
- IEEE Transactions on Consumer Electronics
- Volume
- 72
- Issue
- 2
- Pages
- 4705–4716
- Publication date
- 30 March 2026
- Language
- EN