หน้าแรก / วารสารวิชาการ / Low-complexity gated recurrent unit-based channel ...
Low-complexity gated recurrent unit-based channel estimation for space–frequency block coding multiple-input multiple-output orthogonal frequency-division multiplexing in high-mobility tapped-delay-line-D channels
จาตุรงค์ สาระวงศ์  |  N/A  
จำนวนผู้เข้าชม
21
Low-complexity gated recurrent unit-based channel estimation for space–frequency block coding multiple-input multiple-output orthogonal frequency-division multiplexing in high-mobility tapped-delay-line-D channels
เปิดลิงก์เอกสาร
DOI Link
Collection RMUTK Research Repository (RMUTK IR)
วารสารวิชาการ — e-Journal Articles
ID RMUTK Digital RMUTK000203
Title Low-complexity gated recurrent unit-based channel estimation for space–frequency block coding multiple-input multiple-output orthogonal frequency-division multiplexing in high-mobility tapped-delay-line-D channels
Alternative title Low-complexity gated recurrent unit-based channel estimation for space–frequency block coding multiple-input multiple-output orthogonal frequency-division multiplexing in high-mobility tapped-delay-line-D channels
Authors รศ.ดร.ธนัยรัตน์ มาต๊ะ Corresponding
จาตุรงค์ สาระวงศ์ (มหาวิทยาลัยราชภัฏบ้านสมเด็จเจ้าพระยา) ผู้แต่งหลัก
รศ.ดร.พิสิฐ บุญศรีเมือง (สถาบันเทคโนโลยีพระจอมเกล้าเจ้าคุณทหารลาดกระบัง)
Faculty คณะวิศวกรรมศาสตร์
Journal Title ETRI Journal
ISSN 2233-7326
Volume / Issue / Pages Vol.x | No.x | pp.xx-xx
Published 2026-08-11
Year 2569
Level ระดับนานาชาติ (SCOPUS)
Quartile Q2
DOI 10.4218/etrij.2025-0416
Abstract N/A
Abstract (EN) Reliable operation in space–frequency block coding (SFBC) multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) systems requires accurate channel estimation, which is challenging in high-mobility scenarios owing to rapid channel variations. Traditional pilot-assisted estimators, such as least-squares (LS) and linear minimum mean-square error (LMMSE), suffer severe performance loss under these conditions. Long short-term memory (LSTM) networks and other deep-learning-based methods are more accurate but computationally expensive. This study proposes a low-complexity gated recurrent unit (GRU)-based channel estimator that reduces computational overhead while maintaining high performance. Simulation results over tapped-delay-line-D (TDL-D) high-mobility channels demonstrate that the proposed GRU- and conventional LSTM-based estimators outperform LS and LMMSE over the primary evaluated carrier-to-noise ratio (CNR) range of 0–20 dB in terms of mean-square error and bit error rate. Whereas LSTM slightly outperforms the proposed GRU at extremely low CNRs, the proposed GRU achieves superior overall performance, particularly at high CNRs, where LSTM may saturate.
Keywords (TH) N/A
Keywords (EN) channel estimationdeep learninggated recurrent unithigh mobilityMIMO-OFDM
Access Level Open Access (เปิดสาธารณะ)
ลิงก์แหล่งเผยแพร่
เอกสารฉบับเต็ม (ลิงก์ภายนอก)
https://doi.org/10.4218/etrij.2025-0416
เปิดลิงก์