The goal of this project is to design better adaptive bitrate (ABR) algorithms for high-performance video delivery and evaluate their performance in the field. While there has been long research in this space, our focus is on the following particular topics:
- Extending the existing algorithms for regular video streaming to bitrate and viewport adaptation algorithms for 360°/volumetric videos
- The theoretical foundation of rate adaptation in video streaming with a focus on theoretical analysis of model predictive control (MPC) for video streaming and systematic consideration of switching cost
- Deep-learning enhancement methods, e.g., super-resolution, that improve the quality of video streaming
- Building end-to-end systems for low-latency 3D video conferencing
