Machine Learning–Driven Real-Time Monitoring and Control of Cyber-Physical Systems Using Edge Gateways and Cloud Networks
The growing use of cyber-physical systems (CPS) in business processes such as smart manufacturing, healthcare, transport, and energy has raised major concerns regarding real-time monitoring, control, scalability, and system trustworthiness. The established centralised cloud-based applications are characterised by low responsiveness and high latency (usually 4060 ms), whereas systems that are based solely on the edges do not have the intelligence and coordination across the globe. The following paper will propose a machine-learning-based system to monitor and control CPS in real time, with very low decision-making latency, and to perform scaled system intelligence using edge gateways and cloud networks. In the suggested scheme, edge gateways will process real-time data, identify anomalies, implement control measures using lightweight machine learning models, coordinate their operations globally, optimise their models, and conduct long-term analytics, all provided through the cloud layer. The system has been able to promote distributed learning without disseminating raw data, hence enhancing efficiency and privacy. The comparison of the proposed method, based on the experimental evaluation of over 100 monitoring cycles, shows that the necessary detection accuracy (approximately 91%), compared to the centralised (approximately 80%) and edge-only (approximately 72%) methods, is rather good. Also, response latency decreases to 21 ms from 48 ms, while communication overhead remains at about 85 MB per cycle, making deployment practical and scalable.