Case Study of Building Machine Learning for Cyber-Physical Systems

Cyber-physical Systems (CPS) integrate computing, network communication, and control to facilitate smart-world systems. Nonetheless, the interconnection of sensing and actuating devices in the Internet of Things (IoT) creates new uncertainties and countermeasures must be developed. Deep Learning is an emerging tool with the power to conduct data analysis to address uncertainties. In this research, we have conducted a survey of deep learning platforms and applications and applied deep learning to handle typical CPS functions, including monitoring and control (energy demand prediction) and security (malware detection and prediction), developing appropriate system frameworks and evaluating a variety of learning mechanisms and relevant technologies.