5세대 이동통신을 위한 컴퓨터 비전 기반 28 GHz 양방향 빔 추적 기법 연구

Abstract

In this thesis, beam alignment problem in analog beamforming system is studied. By exploiting vision information obtained from RGB camera, joint Tx-Rx beam tracking system can be implemented. Angles required for beamforming is estimated by some image processing methods. The proposed method determines the best beam pair through the estimated angles the user equipment is located. Aftermath, it is empirically observed that the proposed system attains stable data rates in real time. Compared to conventional method, which is called exhaustive method, the proposed method can outperform in terms of the coordination-free property. Determining the optimal beams without feedback from the receiver saves resources which should have allocated to data transmission. The proposed method can be usable to latency sensitive applications which needs 5G communication.