Generative Diffusion Receivers: Achieving Pilot-Efficient MIMO-OFDM Communications

TL;DR

This work reframes the MIMO-OFDM receiver as a diffusion-based generator of channel-and-data pairs, leveraging channel priors through a neural denoiser while preserving traditional signal-processing steps to avoid hallucination. A novel imagination-screening mechanism guides diffusion, using recovery error to select and refine candidate channels, enabling the same pretrained model to adapt to different pilot densities and modulation schemes without retraining. Empirical results on a DeepMIMO dataset show up to a 2x reduction in channel-reconstruction error at low-to-moderate pilot densities (-4 to 0 dB SNR) and demonstrate the tradeoffs between imagination size, diffusion steps, and computational cost. The proposed approach offers a pilot-efficient, adaptable receiver design with potential for broader deployment in dynamic wireless environments, albeit with considerations for latency and hallucination risk that call for further acceleration and reliability enhancements.

Abstract

This paper focuses on wireless multiple-input multiple-output (MIMO)-orthogonal frequency division multiplex (OFDM) receivers. Traditional wireless receivers have relied on mathematical modeling and Bayesian inference, achieving remarkable success in most areas but falling short in their ability to characterize channel matrices. Neural networks (NNs) have demonstrated significant potential in this aspect. Nevertheless, integrating traditional inference methods with NNs presents challenges, particularly in tracking the error progression. Given the inevitable presence of noise in wireless systems, generative models that are more resilient to noise are garnering increased attention. In this paper, we propose re-evaluating the MIMO-OFDM receiver using diffusion models, which is a common generative approach. With diffusion models, we can effectively leverage prior knowledge of channel matrices and incorporate traditional signal estimation components. Specifically, we explore the diffusion system and introduce an imagination-screening strategy to guide the diffusion process. Furthermore, diffusion models enable adaptation to varying noise levels and pilot schemes using the same NN, significantly reducing training and deployment costs. Simulated results reveal that, for pilot densities ranging from 4-6 pilots per 64-subcarrier block and signal-to-noise ratios (SNRs) from -4 dB to 0 dB, our proposed receiver reduces channel-reconstruction error by up to two times compared to leading deep-learning models, with the most pronounced improvements observed in low-pilot conditions. Additionally, performance enhancements can be achieved with a larger imagination size, despite increased computational complexity.

Figures (13)

• Figure 1: The workflow of the proposed diffusion-based receiver, only one iteration is shown.
• Figure 2: The NN used in this paper.
• Figure 3: The loss vs epoch curve during training.
• Figure 4: Results under different SNRs and modulations.
• Figure 5: Channel generation results with and without enabling imagination module.