http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
mmWave대역 다중 사용자 MIMO시스템에서의 에너지 효율 향상을 위한 하이브리드 프리코딩 기법 연구
신범식(Beom-Sik Shin),임승우(Seung-Woo Im),오지혜(Ji-Hye Oh),송형규(Hyoung-Kyu Song) 대한전자공학회 2021 대한전자공학회 학술대회 Vol.2021 No.6
In this paper, we propose hybrid analog and digital precoding scheme for energy efficiency enhancement in millimeter wave (mmWave) band. Systems based on mmWave band rely on the use of large antenna arrays. Conventional full-connected hybrid beamforming structures shows comparable performance to the full-digital beamforming structures, but require a large number of high-resolution phase shifter to implement an analog precoder, which requires high power consumption and hardware complexity. To solve this problem, we employ a constant phase shifter to implement analog precoder based on switch network. As a result, the proposed structure showed higher energy efficiency than conventional hybrid beamforming structure for mmWave MU-MIMO systems.
5G 기반의 SU-MIMO 시스템을 위한 채널 추정 기법 연구
오지혜(Ji-Hye Oh),임승우(Seung-Woo Im),신범식(Beom - Sik Shin),정지성(Ji-Sung Jung),송형규(Hyoung-Kyu Song) 대한전자공학회 2021 대한전자공학회 학술대회 Vol.2021 No.6
Compared to the existing LTE (Long Term Evolution), 5G RS (Reference Signal) uses DM-RS (demodulation reference signal) for data demodulation. Since the LTE macro cell base station continuously transmits a cell-specific reference signal (CRS), time-frequency resources are consumed and interference in terms of the small cell base station occurs. To solve this problem, the DMRS is transmitted with the necessary information for each channel. In this paper, we propose a multiantenna radio channel estimation technique using 5G DM-RS structure for signal detection and beamforming in mobile communication system. It supports multiple layers by utilizing CDM (Code Division Multiplexing) and FDM (Frequency Division Multiplexing) methods.
Massive MIMO 시스템에서 딥 러닝을 이용한 모델 기반 신호 검출 기법 연구
장준용(Jun-Yong Jang),신범식(Beom-Sik Shin),오지혜(Ji-Hye Oh),임승우(Seung-Woo Im),송형규(Hyoung-Kyu Song) 대한전자공학회 2021 대한전자공학회 학술대회 Vol.2021 No.6
In this paper, an additional study about the exiting MMNet for data detection using Deep Learning (DL) is provided. Existing detection methods are unable to apply massive Multiple Input Multiple Output (MIMO) systems due to high complexity. Therefore, new detection methods with moderate complexity and high error rate performance have been studied until a recently date. In order to solve above problem, DL based detection method has been studied and MMNet is successful model-driven detection method with high adaptive capability. This paper presents a new approach to the denoiser of the nonlinear process for MMNet.