Abstract

At the forefront of digital innovation, Mixed Reality (MR) applications seek to blend the digital and physical worlds seamlessly, providing users with an immersive experience across all devices, regardless of location. Yet, it encounters a formidable challenge: the bottleneck of remote rendering. This research presents the Dynamic Split Computing System, a multi-layered framework designed to address the computational constraints and latency issues that currently limit real-time MR experiences. At the core of this system lies a strategic utilization of edge computing capabilities coupled with the 5G and 6G technologies. The thesis proposed that a distributed approach to computational task management is the key to unlocking responsive, real-time MR interactions. By leveraging a growing network of edge devices, from IoT sensors to autonomous vehicles, the Dynamic Split Computing System orchestrates a more efficient allocation and execution of rendering tasks. This is achieved through a novel resource management and task-splitting mechanism that dynamically assigns computational loads to the most suitable edge nodes, ensuring optimal use of available resources. Through comprehensive evaluations, including mathematical modeling, simulations, and expert feedback, the research validates the effectiveness of the Dynamic Split Computing System. The evaluations underscore the Dynamic Split Computing System as a practical, innovative solution that drives the mobile edge computing field towards a future where real-time applications thrive, free of the limitations of remote rendering.

Results/Implementation/Project Description

Conclusion

[ PDF (optional) ] 

[ Slides Kickoff/Final (optional)]