Receiver Design for Faster-Than-Nyquist Signaling: Deep-Learning-Based Architectures

Peiyang Song,Fengkui Gong,Haiyang Ding,Guo Li,Qiang Li

doi:10.1109/access.2020.2986679

Abstract

Faster-than-Nyquist (FTN) is a promising paradigm to improve bandwidth utilization at the expense of additional intersymbol interference (ISI). In this paper, we apply state-of-the-art deep learning (DL) technology into receiver design for FTN signaling and propose two DL-based new architectures. Firstly, we propose an FTN signal detection based on DL and connect it with the successive interference cancellation (SIC) to replace traditional detection algorithms. Simulation results show that this architecture can achieve near-optimal performance in both uncoded and coded scenarios. Additionally, we propose a DL-based joint signal detection and decoding for FTN signaling to replace the complete baseband part in traditional FTN receivers. The performance of this new architecture has also been illustrated by simulation results. Finally, both the proposed DL-based receiver architecture has the robustness to signal to noise ratio (SNR). In a nutshell, DL has been proved to be a powerful tool for the FTN receiver design.

Highlights

The last couple of decades have seen the exponential growth of wireless devices and data traffic
THE PROPOSED FTN RECEIVER DESIGN WITH deep learning (DL)-BASED JOINT SIGNAL DETECTION AND DECODING We propose a hybrid DL-based architecture, which has been illustrated in Fig. 4, to replace both the signal detection and channel decoding in conventional FTN receivers
ROBUSTNESS TO signal to noise ratio (SNR) MISMATCHING It is very important for the proposed DL-based detection to be robust to the SNR values, without which the proposed DL-based detection will be trained and employed for different SNR values independently and suffer from high complexity resulting from SNR estimation and the store of massive DL network parameters corresponding to different SNR values

Summary

INTRODUCTION

The last couple of decades have seen the exponential growth of wireless devices and data traffic. The associate editor coordinating the review of this manuscript and approving it for publication was Zhanyu Ma. interference increases the complexity to recovery original signals in the receiver, the Mazo limit [1] proves that without the expansion of bandwidth and loss of BER performance, the FTN signaling can achieve an up to 25% higher transmission rate than conventional Nyquist-criterion design in the additive white Gaussian noise (AWGN) channel. Traditional receiver design focuses on the detection algorithms to eliminate the ISI caused by the smaller symbol interval. Signal detection for FTN based on sequence estimation (as well as the channel decoding) can be regarded as a classification problem which aims to divide a multiple dimension space into several parts. With the development of artificial intelligence (AI) chips [30], the DL-based algorithms may show their advance in future communication systems These facts have inspired us to employ DL into FTN receiver designs.

SYSTEM MODEL

SIGNAL RECONSTRUCTION BY SIC

DL-BASED JOINT SIGNAL DETECTION AND DECODING

SIMULATION RESULTS

CONCLUSION