Abstract:

Edge Computing and Internet of Things (IoT) are able to enhance the latency, accuracy, and energy usage of real-time systems to make them more performative.  Edge Computing reduces cloud processing by utilizing network edge processing capabilities.  Due to cloud-based processing of data, modern IoT systems are plagued with high latency, energy usage, and accuracy.  In high-performance deployments of IoT, these limitations erode real-time decision-making and resource optimization. An Edge-Based Machine Learning (E-ML) system combines machine learning algorithms within the network edge for real-time decision-making and efficient resource utilization to overcome these limitations.  Machine learning enhances system accuracy and resource utilization depending on context data, whereas the strategy decreases cloud data transfer and speeds up processing. The proposed solution improves IoT systems by providing real-time edge decision-making, decreasing latency, enhancing prediction accuracy, and saving energy. System resources and settings are optimally adjusted to network demands and conditions.  Local decision-making by the E-ML framework minimizes latency, enhances decision making through intelligent models, and optimizes energy usage.  The approach improves IoT applications like smart cities, healthcare, and autonomous systems, and it can be used in high-scale deployments where efficiency and resource usage are critical.