Lightweight AI models for secure and energy-efficient IoT networking in dynamic edge environments

The rapid expansion of Internet of Things (IoT) networks has brought significant challenges, particularly in optimizing energy consumption, maintaining security, and managing dynamic network conditions. Many existing artificial intelligence (AI)-based solutions, while effective, often demand substantial computational resources, making them unsuitable for edge environments with limited energy and processing capabilities. This study introduces an innovative lightweight AI framework specifically designed to enhance both security and energy efficiency in IoT networks operating at the edge. By combining ensemble learning techniques with attention-based neural networks, the framework enables real-time anomaly detection and optimized resource allocation across dynamic IoT systems. Analytical modeling and simulation results indicate that the proposed approach achieves a 30% reduction in energy usage and a 25% improvement in anomaly detection accuracy compared to existing methods. Additionally, its scalable architecture ensures broad applicability across smart cities, critical infrastructure monitoring, and healthcare IoT systems. By addressing these key gaps, this research contributes to advancing the practical deployment of IoT networks in diverse real-world applications.