Turbo Frequency Domain Equalization and Detection for Multicarrier Faster-Than-Nyquist Signaling

Multicarrier faster-than-Nyquist (MFTN) signaling is a spectral efficient modulation scheme for broadband and broadcasting applications. However, due to the intentionally introduced intersymbol interference (ISI) and intercarrier interference (ICI) by time-frequency packing, the detection of MFTN signals is long considered to be a challenging problem. In this paper, we propose two low complexity detection schemes for MFTN signals based on turbo frequency domain equalization (FDE). Firstly, we extend the one dimensional FDE to two dimensional model to deal with the ISI/ICI interference induced by MFTN signaling, and validate its practical biterror- rate (BER) performance combined with turbo equalization. Secondly, a one dimensional FDE combined with soft successive interference cancellation (SIC) is further proposed to detect the MFTN signals, and it shows that the FDE combined with SIC could asymptotically approach the maximum a posterior (MAP) equalization coupled with SIC with negligible BER performance loss under moderate ISI and ICI. Computational complexity analysis and numerical results show that the FDE combined with SIC is better than 2-D FDE and may be a better choice than MAP equalization combined with SIC in the practical detection of MFTN signals.