Project Title: Distributed Edge Intelligence for Immersive Tactile Internet
Description
Distributed edge intelligence is one of the enabling technologies for time-critical applications by allowing devices to offload computation and energy demanding operations to computing resources in the proximity. In Tactile Internet, due to the resource limitation on the robot side, it is desirable to leverage distributed edge intelligence instead of performing computation locally. This project aims to develop a distributed edge intelligence framework for the Tactile Internet, including the development of advanced offloading methods for computation tasks (e.g., Deep Learning real-time inference, 3D reconstruction), balancing Cloud and Edge computation and edge caching. In particular, task-oriented communication, an emerging communication paradigm, will be studied and applied for designing efficient transmission strategies for task data in computation offloading. It is important to understand trade-offs of edge intelligence engine’s location in the network and the communication and computation load of tasks, under different network QoS and user QoE requirements of tactile Internet services, and then perform joint optimization of computation, communication, and haptic control.
