Model-Driven Deep Learning Scheme for Adaptive Transmission in MIMO-SCFDE System

Jun Li,Wenxin Li,Bo He,Tongliang Xin,Yuanjian Qiao

doi:10.1109/access.2020.3034744

Abstract

Adaptive transmission (AT) is considered as one of the critical technologies to enhance the effectiveness of communication systems. In this article, we propose a model-driven deep learning (DL) scheme for AT in multiple-input multiple-output single-carrier frequency-domain equalization (MIMO-SCFDE) systems, in which the adaptive modulation network (AMNet) and adaptive demodulation network (ADNet) are adopted to complete the modulation of the signal and the modulation recognition of the receiver. Under the target bit error rate (BER), the adaptive modulation (AM) scheme can adjust the modulation mode selection of different transmitting antennas adaptively according to the estimated channel information to improve the throughput. The features required by the AMNet are extracted from the received signal, and the labels are assigned according to the optimal modulation scheme got by analyzing the signal detection performance. Since the spectral correlation function has a powerful ability to suppress noise and the cyclic spectrum varies with the modulation mode, we take the preprocessed cyclic spectrogram as the input of ADNet to achieve the adaptive modulation recognition (AMR). Comparative experiments demonstrate that the proposed scheme gets better performance in terms of throughput and reliability in MIMO-SCFDE systems than the traditional scheme and the existing DL scheme.

Highlights

A DAPTIVE transmission (AT) refers to the technology that the transmitter utilizes the channel state information (CSI) to adjust the transmission strategy adaptively, including changing the transmission power, adjusting the modulation mode, or adjusting the channel coding scheme so that the system can improve the information transmission rate or reliability [1]
The research on applying deep learning (DL) to the physical layer is mainly divided into two types: data-driven network and model-driven network [5]
In [7], the receiving module after removing the cyclic prefix (CP) in the orthogonal frequency division multiplexing (OFDM) system is regarded as a whole and replaced by the offline trained deep neural networks (DNN) to accomplish the process from the radio frequency (RF) receiver to the sink directly

Summary

INTRODUCTION

A DAPTIVE transmission (AT) refers to the technology that the transmitter utilizes the channel state information (CSI) to adjust the transmission strategy adaptively, including changing the transmission power, adjusting the modulation mode, or adjusting the channel coding scheme so that the system can improve the information transmission rate or reliability [1]. In [11], a novel signal detection scheme based on the adaptive ensemble deep learning algorithm in single-carrier frequency-domain equalization (SC-FDE) systems is proposed, which adopts an integrated LSTM model to replace the channel estimation and frequency domain equalization process. Compared with the traditional physical layer technology, the above datadriven DL schemes improve performance by replacing multiple modules of the communication system in an end-toend way. This kind of model discards these existing wireless communication knowledge, and the slight change of model structure will lead to the decline of accuracy.

SYSTEM MODEL

ResNet-50

ADAPTIVE MODULATION RECOGNITION SCHEME

EXPERIMENT AND ANALYSIS

PERFORMANCE COMPARISON ANALYSIS OF AD SCHEME