Project Grant 2246698

This $129,887 Project Grant from the National Science Foundation's Computer and Information Science and Engineering program (CFDA 47.070) will fund the development of a new edge-IoT framework called DeepEdge at George Mason University from October 2022 through September 2023. DeepEdge will use deep online learning to allocate resources for heterogeneous IoT applications and dynamic devices across geographically distributed edge clouds. It aims to maximize users' quality of experience. Key outcomes will include a new quality of experience model, a two-stage resource allocation scheme using deep machine learning to adapt application quality of service requirements based on available edge cloud resources while maintaining priorities, and a novel deep Q-learning approach to dynamically select optimal edge nodes to handle multiple application tasks in a way that optimizes task execution delay. The awardee will implement a hardware and software testbed to validate and evaluate the effectiveness, efficiency, and practicality of the proposed research advancing edge computing systems for heterogeneous and dynamic IoT applications.