Low-Complexity Detection for FTN Signaling Based on Weighted FG-SS-BP Equalization Method

Abstract

In this paper, a low-complexity turbo detection scheme based on a weighted factor graph (FG) serial-schedule (SS) belief propagation (BP) equalization method is proposed for Faster-than-Nyquist (FTN) signaling. The iterative equalization method is applied to mitigate the severe intersymbol interference (ISI) introduced by FTN signaling. In order to reduce the complexity of the equalization method, Gaussian approximation (GA) is used to calculate the log-likelihood ratio (LLR). Thus, the computational complexity is merely linear with the number of ISI taps. Furthermore, LDPC, as an efficient coding technique, results in performance that approaches the Shannon limit in this turbo detection scheme. The simulation results show that the proposed weighted FG-SS-BP-based turbo detection method performs close to the optimal detector under ISI-free conditions with very low complexity.