Physics Aware Conditional SetGAN for Spatially Consistent Multi User TR 38.901 Channel Generation

Jul 16, 2026

Upload by Aurelia

TR 38.901-based channel models such as Sionna produce reliable multi-user channel realizations, but generating many realizations is computationally expensive. The work studies whether a trained generative model can generate multi-user TR 38.901 channels faster than Sionna while preserving geometry-imposed spatial correlations. A physics-aware, geometry-conditioned SetGAN is trained on Sionna reference data by separating large-scale received power from normalized small-scale fading, compressing fading with PCA, and learning conditional distributions in latent space. Benchmarks show close received-power distributions, accurate spatial-consistency profiles, and substantial runtime and CPU cost reductions versus Sionna.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages6

Reading Time15 min

I. INTRODUCTION

Related Work and Positioning

Physics-Aware Conditional SetGAN for Spatially Consistent Multi-User TR 38.901 Channel Generation Mauro Gonzalo Tarazona-Levano∗ , David Lopez-Perez∗†, Nicola Piovesan‡, and David Gomez-Barquero∗∗ Institute of Telecommunications and Multimedia Applications (iTEAM), Universitat Politècnica de València (UPV), Spain † Beihang Valencia Polytechnic Institute (BVPI), China ‡Huawei Technologies, France [mgtarlev@iteam.upv.es](mgtarlev@iteam.upv.es) arXiv :2607 . 1 1429v 1 [ cs .LG] 13 Jul 2026 Abstract—TR 38.901-based channel models such as Sionna are reliable, but generating many multi-user channel realizations remains expensive. This paper asks a practical question: can a trained generative model produce multi-user TR 38.901 channels faster than Sionna without losing the spatial correlations imposed by user geometry? To answer this question, we propose a physicsaware, geometry-conditioned SetGAN trained on Sionna reference data. The method separates large-scale received power from normalized small-scale fading, compresses the latter with principal component analysis, and learns the conditional channel distribution in a latent space while preserving geometry-dependent correlations. On the UMa/NLoS benchmark, the model keeps the received-power distributions close to the reference, with about 0.41 dB Wasserstein distance, and reproduces spatial-consistency profiles with mean deviations below 0.03 on median curves versus distance. In addition, it reduces elapsed generation time by a factor of 3.45 and CPU-total cost by a factor of 6.15 relative to Sionna under matched user positions in the fixed-position CPUvs-CPU benchmark. These results show that a trained generative model can substantially accelerate TR 38.901 channel generation without breaking the spatial consistency needed to evaluate multiuser systems. Index Terms—channel modeling, wireless channel generation, multi-user MIMO, 3GPP TR 38.901, generative adversarial networks, set transformers, spatial consistency, Sionna. I. INTRODUCTION Generating multi-user multiple-input multiple-output (MIMO) channel realizations is computationally expensive, especially when large numbers of user-equipment (UE) deployments and configurations must be evaluated. At the same time, these realizations cannot be treated as independent channel draws: nearby UE should exhibit correlated largescale effects and coherent small-scale fading structure, since spatial consistency is essential in multi-user MIMO evaluations. This requirement is particularly important when studying multi-user behavior under realistic geometry. Tools implementing the stochastic channel models specified in 3GPP TR 38.901 provide this type of physically grounded channel generation [1] . Among them, Sionna is an open-source This research is supported by the Generalitat Valenciana through the CIDEGENT PlaGenT, Grant CIDEXG/2022/17, Project iTENTE, and the action CNS2023-144333, financed by MCIN/AEI/10.13039/501100011033 and the European Union “NextGenerationEU”/PRTR. wireless simulation library that implements these models and is widely used as a practical reference [2] . Later studies have also examined the spatial-consistency behavior and calibration of these models [3], [4] . In that sense, TR 38 .901-based simulation tools provide a reliable reference for generating spatially consistent multi-user channels. The practical problem is that repeatedly calling such tools to generate very large numbers of realizations can become prohibitively expensive in runtime. The central question of this paper is therefore the following: can a generative model be designed to reproduce the multiuser MIMO channel frequency responses generated by a TR 38.901-compliant reference tool, while generating them faster than the simulator itself? In this work, Sionna is used as that reference implementation. In the multi-user setting, this is not only about reproducing realistic coefficients for each UE separately. The real cha

FAQs

What problem does the paper address in TR 38.901 multi-user channel generation?

Generating multi-user MIMO channel realizations under TR 38.901 is computationally expensive, especially at large scale. The realizations must also preserve spatial consistency, meaning nearby user equipment should show correlated large-scale effects and structured small-scale fading.

How does the proposed physics-aware conditional SetGAN preserve spatial correlations?

The method conditions on user geometry and learns the channel distribution while maintaining geometry-dependent correlations. It separates large-scale received power from normalized small-scale fading, compresses the fading with PCA, and models the conditional distribution in a latent space.

What performance improvements are reported compared with Sionna?

On the UMa/NLoS benchmark, the model keeps received-power distributions close to reference with about 0.41 dB Wasserstein distance and reproduces spatial-consistency profiles with mean deviations below 0.03. It also reduces elapsed generation time by a factor of 3.45 and CPU-total cost by a factor of 6.15 under matched user positions.

Jul 16, 2026

Upload by Aurelia