Containerized Wearable Edge AI Inference Framework in Mobile Health Systems

The proliferation of wearable devices and personal smartphones has promoted smart mobile health (MH) technologies. The MH applications and services are extremely responsive to computation latency. Edge computing is a distinguished form of cloud computing that keeps data, applications, and computing power away from a centralized cloud network or data center. In this work, we design a containerized wearable edge AI inference framework. The cloud computing layer includes two cloud-based infrastructures: The Docker hub repository and the storage as service hosted by Amazon web service. The Docker containerized wearable inference is implemented after training a Deep Learning model on open data set from wearable sensors. At the edge layer, the Docker container enables virtual computing resources instantiated to process data collected locally closer to EC infrastructures. It is made up of a number of Docker container instances. The containerized edge inference provides data analysis framework (DAF) targeted to fulfill prerequisites on latency, and the availability of wearable-based edge applications such as MH applications.