MMSE Turbo Equalization and Detection for Multicarrier Faster-Than-Nyquist Signaling

In this paper, two minimum mean-squared error (MMSE) turbo equalization and detection schemes for spectral efficient multicarrier faster-than-Nyquist (MFTN) signaling system are presented. First, the two-dimensional (2-D) soft input and soft output (SISO) MMSE equalization intended for magnetic recording channels, i.e., 2-D intersymbol interference (ISI), is extended to the detection of MFTN signals. Although the 2-D MMSE turbo equalization exhibits an attractive bit-error-rate (BER) performance under certain time-frequency spacing packing between adjacent symbols/subcarriers, the BER performance will be degraded when introducing serious intercarrier interference. Second, a 1-D SISO MMSE equalization combined with soft successive interference cancellation (SIC) is further proposed for the detection of MFTN signals, and it shows that the MMSE equalization combined with SIC could asymptotically approach the maximum a posterior equalization combined with SIC with negligible BER performance loss. Our computational complexity analysis and numerous simulation results demonstrate that the proposed MMSE turbo equalization schemes may be more preferred choice in the practical detection of MFTN signals.