Enhancing Quality of Service for IoT Application in Smart Cities: A Hybrid Split Learning and Optimized Routing Approach

Technological advancements in the Internet of Things (IoT) have transformed traditional Consumer Electronics (CE) into next-generation devices with enhanced connectivity and intelligence. This networked connectivity among sensors, actuators, appliances, and other consumer devices improves data availability and enables automatic control within the CE network. However, the diversity, decentralization, and proliferation of CE devices have exponentially increased data traffic. Additionally, traditional static network infrastructure approaches require manual configuration and exclusive management of these devices. Our research introduces a new way to improve the Quality of Service (QoS) in Consumer Electronics. The proposed work uses advanced technologies like Hopfield Networks for managing resources in real-time, Split Learning for secure distributed learning, and Enhanced Split Hyena Optimized Routing for faster data transfer. Hopfield Networks help allocate resources on the fly, resulting in better speed, efficiency, and less power usage. While Split Learning (SL) has been suggested for training models with limited resources, it is not widely used in decentralized and resource-limited IoT devices. Our proposed approach tackles the challenges in CE using Hopfield Networks, Split Learning, and Enhanced Split Hyena Optimized Routing. Our method achieves an impressive 95% success rate in delivering data, showing that it's effective for reliable and efficient IoT communication.