Joint Precoding and Pre-equalization for Faster-Than-Nyquist Transmission over Multipath Fading Channels
Shan Wen,Guanghui Liu,Chengxiang Liu,Huiyang Qu,Lei Zhang,Muhammad Ali Imran
DOI: https://doi.org/10.1109/tvt.2022.3146423
IF: 6.8
2022-01-01
IEEE Transactions on Vehicular Technology
Abstract:Faster-than-Nyquist signaling (FTNS) has emerged as a promising technique to increase communication capacity in bandwidth-limited channels. However, the presence of FTN-induced inter-symbol interference (FTN-ISI) in the received observations, is detrimental to channel estimation (CE) and data detection in terms of computational complexity and performance. This paper copes with these problems by incorporating linear pre-equalization (LPE) and composite precoding formed by linear spectral precoding and Tomlinson-Harashima precoding (THP), into the FTNS. Specifically, LPE completely pre-equalizes the FTN-ISI, while spectral precoding resolves the LPE-caused signal spectral broadening by introducing proper artificial ISI, which is pre-equalized by THP. Channel-induced ISI, as the only ISI component in the observations, is estimated and equalized using the classical frequency-domain low-complexity schemes. We show that there are four advantages of the LPE-aided CE over the CE designed for the FTN transmissions without FTN-ISI pre-equalization, namely lower pilot overhead, simpler yet optimal pilot sequence design, lower mean-squared error of CE, and more robust against the FTN-ISI. Simulation results show that our scheme improves the performance of CE and detection, compared to existing FTN frequency-domain CE and equalization schemes.
telecommunications,engineering, electrical & electronic,transportation science & technology