콘텐츠 본문
논문 국내 국내전문학술지(KCI급) 무선통신시스템에서 딥러닝에 기반한 미래 신호대잡음비 예측
- 학술지 구분 국내전문학술지(KCI급)
- 게재년월 2022-12
- 저자명 정의림공동(교신),선중규,윤동환,최재웅,오정은,조아민
- 학술지명 차세대융합기술학회논문지
- 발행처명 국제차세대융합기술학회
- 발행국가 국내
- 논문언어 한국어
- 전체저자수 6
논문 초록 (Abstract)
This paper proposes a deep-learning based signal to noise ratio (SNR) prediction technique for wireless communication environments. The communication system considered in this paper uses time division duplex (TDD), and receives signal using multiple antennas while transmits with only one antenna. Based on the SNR measurements when receiving in the past, we proposed a convolutional neural network (CNN) model to predict the SNRs for all antennas at the time of future transmission. If there is no received signal nor SNR measurement, the SNR measurements are filled by linear interpolation of neighboring two received SNRs. According to the simulation results, the wideband signals show better prediction performance than the narrowband signals. In the case of wideband, the proposed technique is about 0.37～0.98 dB superior to the conventional method for 20 km/h. For narrowband, the proposed one is better by 0.29～0.88 dB.
