Research
URLLC (ultra-reliable low-latency communication) in 5G System with Machine Learning
- Low-latency & high-speed communication scheme for 360VR service with computer vision and machine learning
- Viewport prediction with Machine Learning
- Tile-based transmission of predicted viewport with high quality video
- Network slicing design in 5G communication system
Free-viewpoint (pseudo-6DoF) 360VR System based on Social Media Server
In this research, our ultimate goal is to implement a service as though a VR user freely exists in a place. To this end, it is most important to reconstruct a VR space that provides six degree-of-freedom (DOF), namely, yaw, pitch, roll, surge, sway, and heave. However, most currently released VR services that are based on the real world limit users’ movements to three DOF. Even if the services support six DOF, most are highly complex and based on computer graphics. To overcome this problem, we first assume that there is a full Internet of things (IoT) infrastructure for collecting important data for VR space reconstruction. This assumption is realistic because many researchers expect that in the near future, IoT technology will lead to a world that connects not only people to people but also things to things. In this research, we propose an end-to-end system architecture for the VR space that is based on the real world along with the element technologies.
- Image/video processing and reconstruction at each viewpoint location
- Audio Processing and reconstruction at each viewpoint location
- Social media server platform and metadata design
- Low-latency & high-speed communication scheme for 360VR service with computer vision and machine learning
Open-loop Precoding for Spatial Multipleixng in Correlated MIMO Fading Channels
Muliple-input multiple-output (MIMO) systems can offer high data rates through spatial multiplexing (SM) in wireless communications.
However, the performance of SM systems is significantly degraded in transmit correlated MIMO channels. In closed-loop MIMO sys-
tems, precoding can enhance the error performance by exploiting channel state information(CSI), such as channel covariance, at the
transmitter. However, in closed-loop MIMO systems, feedback information, such as CSI, is susceptible to channel variation. This can
be problematic, as the accuracy of feedback information can be critical for a fast time-varying channel in mobile communications. The
existence of a delay or estimation error in the feedback loop causes severe performance degradation. Moreover, broadcasting syste-
ms do not have any feedback informaiton. It is necessary to consider open-loop precoding. Therefore, we research various open-loop
precoding methods for SM systems in transmit correlated MIMO channels.
Device-to-Device Communications