Cross-Layer Analysis of Machine Learning Models for Secureand Energy-Efficient IoT Networks

The widespread use of Internet of Things (IoT) raises security and energy efficiency 1 concerns, particularly for low-resource devices. In this paper, we analyse a cross-layer IoT architecture 2 using machine learning models and lightweight cryptography. We focus on analysing vulnerabilities 3 and suggest energy-efficient IoT solutions. Our proposed solution is based on a role-based access 4 control ensuring secure authentication in large-scale IoT deployments and blocks undesired access 5 attempts. By combining convolutional neural networks, rule-based systems, and hybrid artificial 6 intelligence, the proposed cross-layer architecture improves the accuracy of anomaly identification 7 while lowering false positives. The system performance is evaluated by simulations as well as testbeds 8 to accomplish attack mitigation. Results show that the proposed system reduces false positives by 9 28–32% and provide improved security by preventing 95% of unwanted access. We found up to 30% 10 power reduction in the proposed lightweight Speck encryption (8Hz ContikiMAC duty cycle) than 11 the traditional AES encryption. For data injection, and sinkhole and jamming attacks, the system’s 12 resilience is confirmed by Cooja/Contiki simulations, which maintain a 95% packet delivery rate. 13 By learning from convolutional neural networks and artificial intelligence, our approach efficiently 14 improves IoT security and energy efficiency in practical scenarios such as smart schools.