The MOBED(mobile embedded system) research group at Yonsei University has actively been participating in the research and development of a wide range of embedded and target-specific system software. We are currently interested in (1) on-device AI systems, (2) neural-enhanced streaming systems, (3) energy-optimized mobile systems, (4) cross-device technologies, and (5) intelligent IoT systems.

1. On-device System Optimization for AI Applications
· Performance, Energy, Thermal-aware multi-DNN inference for mobile AI applications leveraging CPU, GPU, and NPU
· Omnidirectional 3D perception frameworks for resource-constrained edge devices
· Context-aware inference system for mobile mixed reality applications

2. Neural-enhanced Streaming for Immersive Applications
· Super-resolution enhanced 360° video live streaming for mobile devices
· Mobile volumetric video streaming with photorealistic 3D reconstruction technology
· Neural-enhanced streaming optimization with heterogeneous mobile processors

3. Energy Optimization for Mobile Devices
· Linux kernel and Android framework optimization for energy/thermal-aware devices
· Energy-aware scheduler and governor for heterogeneous multi-core AP
· Energy-optimized AI applications for mobile devices

4. Cross-device Techniques for Heterogeneous Devices
· Fine-grained user interface distribution for cross-device web experiences
· Web-based systems for cross-device I/O sharing
· Applications using cross-device techniques

5. Intelligent Computing for IoT Systems
· Split learning-based IoT applications for energy-constrained sensor devices
· Context-aware input data compression for efficient AI-IoT task offloading
· Machine learning-based sensor fusion for IoT applications