A Deep Learning Based Receiver for Wireless Communications Systems With Unknown Channel Models

In this paper, we consider the signal detection problem for practical communication systems where the transmission suffers from non-idealities effects such as I/Q imbalance, non-Gaussian noise, and amplifier nonlinearity. In this context, traditional model-based receiver algorithms are not applicable because the exact model for the end- to-end channel is unavailable. To deal with this challenge, a deep learning based receiver is developed, which learns the detection criterion from data and realizes signal detection without knowing the underlying channel model. To improve the generalizability of the proposed detector, domain knowledge in channel estimation and equalization is used to guide the design of the receiver structure. Moreover, the idea from transfer learning and adversarial learning is adopted to assist the training of the network. The performance of the proposed detector is evaluated in terms of bit error rate (BER), and the superiority of the proposed design is shown compared to the state-of-the-art technologies.

I. Introduction

SIGNAL detection which is used to recover information at the receiver plays a crucial role in communication systems. The research on traditional receiver design schemes is also relatively numerous and mature. Literature [1], [2] studies the design of TR receiver systems for Ultra-wideband (UWB) communication systems. This research is based on traditional methods to design suitable receivers for specific communication systems. [3] proposed a low-complexity improved wiener filter coefficient algorithm to estimate the OFDM channels considering a digital radio mondiale (DRM) system. In [4], symbol error performance analysis of OFDM receiver with pulse blanking over frequency selective fading channel has been done. [5] developed an efficient receiver structure. Incorporated MIMO (multiple-input-multiple-output) system with FTN (faster-than-Nyquist) techniques are considered in the paper. To improve performance and throughput, [6] gives the analysis of performance degradation using convex optimization for a mismatched receiver. However, traditional signal detection algorithms are highly rely on exact channel information. Nevertheless, in many practical scenarios, the exact channel model may not be available. This is due to several reasons. For example, device non-ideality such as I/Q imbalance [7] and nonlinear amplifier [8] adds nonlinearity to the end-to-end channel, and it is very difficult to characterize this nonlinearity using analytical models. In addition, the emergence of new scenarios are giving birth to new communication systems such as molecular communications [9], [10], where the channel model is highly complex to mathematically describe or even completely unknown.