Human-Aware Edge Computing Scheduling for Cyber-Physical Systems

Abstract

-

This work delves into the realm of cyber-physical systems (CPS), focusing on their integration within factory environments, where human actuators are also responsible for performing tasks in the physical world as parts of the process.CPS involves the fusion of computer-based algorithms and physical processes, wherein computations and physical actions mutually influence each other. However, the computational demands of monitoring and control in CPS can challenge centralized systems due to potential data overload. Edge computing emerges as a promising solution, allowing data processing closer to edge devices and facilitating workload distribution among nodes.

Incorporating humans into CPS, termed Human-in-the-Loop (HiL) tasks, demands a careful design to avoid potential accidents. Furthermore, human factors introduce unpredictability, posing challenges such as variable response times.I began by introducing fundamental concepts widely utilized in CPS and task scheduling and described the fundamental ideas behind edge computing. I designed and implemented a CPS environment that simulated a factory production scenario. Through a literature study, I investigated the integration of edge computing into CPS and explored various scheduling techniques at both the kernel and application levels. The objective was to mitigate latency and enhance the speed of Human-in-the-Loop callbackchains. Lastly, I integrated the insights gleaned from the literature study into the simulated environment to validate the findings and identify potential areas for improvement.