International Journal of Science, Engineering and TechnologyAuthors: Ms. Aarti, Dr. V.K. Srivastava
Abstract: The sudden growth of the Internet of Things (IoT) has added pressure on the necessity to have real-time intensive and energy-efficient edge network data processing. The traditional cloud-based designs are dogged by latency, bandwidth and privacy issues rendering them unsuitable in mission-critical Internet of things applications. The study will provide a lightweight Edge-Enhanced IoT model that combines optimized Deep Learning (DL) models with Generative Artificial Intelligence (GenAI) to support autonomous real-time decision-making. The structure uses quantized and pruned neural networks to infer edges efficiently and uses small-scale neural generators to supplement low-quality sensor measurements and restore lost values and model rare anomalies. To maximize performance and reliability, an architecture with multiple layers with local sensing, edge/fog computation, generative enhancement, and selective cloud synchronization is proposed. It has been shown that, through experimental findings, the accuracy, latency, energy consumption, and scalability of the technology have improved in a variety of IoT applications, such as health monitoring, environmental sensing, and industrial condition analysis. The results indicate the opportunities of integrating Edge Computing, Deep Learning, and Generative AI to develop the next generation intelligent IoT infrastructure that can provide secure, fast, and autonomous real-time services.
