A Foundation Model for Massive MIMO Precoding with an Adaptive per-User Rate-Power Tradeoff

2025-07-24

AI Summary

To address high energy consumption in massive MIMO precoding, dynamic user rate requirements, and scarcity of high-quality channel state information (CSI), this paper pioneers the integration of Transformer-based foundation models into precoding design. We propose a zero-shot cross-scenario precoding framework that combines cosine-similarity-guided data augmentation with a pre-trained feature extractor to mitigate small-sample training challenges. The method enables per-user flexible trade-offs between spectral efficiency and power consumption, achieving high energy efficiency while maintaining both spectral efficiency and fairness. Experiments demonstrate that, at identical transmit power, our approach attains performance comparable to the computationally intensive WMMSE algorithm—yet with only 1/8 its complexity—and significantly lower energy consumption than zero-forcing (ZF) precoding. These results validate its practical efficacy under low-data-regime conditions.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Deep learning (DL) has emerged as a solution for precoding in massive multiple-input multiple-output (mMIMO) systems due to its capacity to learn the characteristics of the propagation environment. However, training such a model requires high-quality, local datasets at the deployment site, which are often difficult to collect. We propose a transformer-based foundation model for mMIMO precoding that seeks to minimize the energy consumption of the transmitter while dynamically adapting to per-user rate requirements. At equal energy consumption, zero-shot deployment of the proposed foundation model significantly outperforms zero forcing, and approaches weighted minimum mean squared error performance with 8x less complexity. To address model adaptation in data-scarce settings, we introduce a data augmentation method that finds training samples similar to the target distribution by computing the cosine similarity between the outputs of the pre-trained feature extractor. Our work enables the implementation of DL-based solutions in practice by addressing challenges of data availability and training complexity. Moreover, the ability to dynamically configure per-user rate requirements can be leveraged by higher level resource allocation and scheduling algorithms for greater control over energy efficiency, spectral efficiency and fairness.

Problem

• Minimize transmitter energy consumption in mMIMO precoding
• Adapt dynamically to per-user rate requirements
• Address data scarcity for DL model training

Innovation

• Transformer-based foundation model for mMIMO precoding
• Data augmentation using cosine similarity for adaptation
• Dynamic per-user rate-power tradeoff optimization