Site-Specific MIMO Channel Generation via Diffusion and Flow Matching: Fidelity, Efficiency, and Downstream Utility

2026-06-18

AI Summary

This work addresses the challenge of acquiring high-fidelity, large-scale MIMO channel data in real-world scenarios, which is often prohibitively expensive. To this end, the authors propose a location-conditioned generative framework that, for the first time, integrates diffusion models and flow matching techniques into site-specific MIMO channel synthesis. Specifically, they develop a conditional Denoising Diffusion Implicit Model (cDDIM) and a conditional Flow Matching Model (cFMM), both leveraging user coordinates as conditioning inputs to generate spatially structured channel matrices. Extensive evaluations across multiple scenarios at 28 GHz and 3.5 GHz demonstrate that cFMM achieves generation quality comparable to cDDIM while offering nearly an order-of-magnitude speedup in inference. Moreover, the synthesized channels significantly enhance the performance of downstream physical-layer tasks, such as channel compression and beam alignment.

Abstract

This paper explores the use of generative models to synthesize high-quality, site-specific multiple-input multiple-output (MIMO) channel data, addressing the high cost of the extensive measurement campaigns required to acquire real-world data for AI-native wireless networks. Two location-conditioned generative paradigms are compared: a conditional denoising diffusion implicit model (cDDIM), and a conditional flow matching model (cFMM). Both these models generate MIMO channel matrices conditioned on user coordinates, to preserve the spatial structure of the deployment site. The approaches are evaluated across three dimensions: statistical fidelity (including beam consistency and effective rank), generation efficiency, and utility in downstream tasks such as channel-state information compression and beam alignment. Results across diverse propagation scenarios (28 GHz and 3.5 GHz, both line-of-sight and non-line-of-sight) demonstrate that both models accurately capture site-specific characteristics, even when trained on scarce ground-truth data. Notably, cFMM achieves a quality comparable to cDDIM with roughly an order of magnitude less inference time. Augmenting scarce site-specific datasets with these synthetic channels yields hefty performance gains in downstream physical layer tasks compared to using scarce data alone or stochastic channels.

Problem

• MIMO channel
• site-specific
• data generation
• measurement cost
• wireless networks

Innovation

• conditional flow matching
• MIMO channel generation
• site-specific modeling
• generative models
• downstream utility