Open AccessSparse channel estimation for filtered multi‐tone in time domain and subband domain based on matched filtering demodulation
Author(s) -
Qu Fengzhong,
Zhang Minhao,
Wang Zhenduo,
Qian Caijie,
Lu Xuesong,
Wei Yan
Publication year - 2019
Publication title -
iet communications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.355
H-Index - 62
eISSN - 1751-8636
pISSN - 1751-8628
DOI - 10.1049/iet-com.2018.5931
Subject(s) - channel (broadcasting) , computer science , demodulation , orthogonality , interference (communication) , fading , signal (programming language) , time domain , frequency domain , tone (literature) , algorithm , pilot signal , underwater acoustics , speech recognition , telecommunications , acoustics , underwater , mathematics , computer vision , physics , art , oceanography , geometry , literature , programming language , geology
The interference between filtered multi‐tone (FMT) symbols exists and becomes conspicuous especially in frequency selective fading channel, like an underwater acoustic channel. This leads to the necessity of channel estimation and equalisation. In this study, the authors consider two approaches to utilise the channel sparsity to improve FMT channel estimation performance and further reduce bit error rate in underwater acoustic communications. While the first studies FMT sparse channel estimation in the time domain and is based on received FMT signal and known transmitted FMT signal that is reconstructed by pilots, the second is based on matched filtering demodulated FMT symbols and pilots. Theoretical analysis is performed followed by simulation and experiment accordingly. It is found that the first method is able to successfully estimate the sparse channel while the estimation performance of the second method is unsatisfactory due to poor orthogonality between bases of the channel estimation matrix. Moreover, the lake experiment is carried out with the first method and its results prove that sparse channel estimation has better performance than least square channel estimation.