Enabling Large-Scale Channel Sounding for 6G: A Framework for Sparse Sampling and Multipath Component Extraction

TL;DR

This work tackles the bottleneck of acquiring large-scale real-world ISAC channel data for 6G by introducing a sparse, nonuniform frequency sampling framework and a likelihood-rectified SAGE (LR-SAGE) MPC extraction method. The Parabolic Frequency Sampling (PFS) scheme eliminates delay ambiguity and dramatically reduces the number of required frequency samples, while the LR-SAGE algorithm compensates for nonuniform sampling and molecular absorption to yield accurate MPC estimates. Simulation and THz experimental results demonstrate up to 50x faster measurements, around 98% data-volume reduction, and a 99.96% reduction in post-processing complexity, with MPCs and channel characteristics consistent with dense, exhaustive measurements. This framework enables high-throughput construction of large-scale ISAC channel datasets essential for AI-native 6G systems and ISAC research at THz frequencies.

Abstract

Realizing the 6G vision of artificial intelligence (AI) and integrated sensing and communication (ISAC) critically requires large-scale real-world channel datasets for channel modeling and data-driven AI models. However, traditional frequency-domain channel sounding methods suffer from low efficiency due to a prohibitive number of frequency points to avoid delay ambiguity. This paper proposes a novel channel sounding framework involving sparse nonuniform sampling along with a likelihood-rectified space-alternating generalized expectation-maximization (LR-SAGE) algorithm for multipath component extraction. This framework enables the acquisition of channel datasets that are tens or even hundreds of times larger within the same channel measurement duration, thereby providing the massive data required to harness the full potential of AI scaling laws. Specifically, we propose a Parabolic Frequency Sampling (PFS) strategy that non-uniformly distributes frequency points, effectively eliminating delay ambiguity while reducing sampling overhead by orders of magnitude. To efficiently extract multipath components (MPCs) from the channel data measured by PFS, we develop a LR-SAGE algorithm, rectifying the likelihood distortion caused by nonuniform sampling and molecular absorption effect. Simulation results and experimental validation at 280--300~GHz confirm that the proposed PFS and LR-SAGE algorithm not only achieve 50$\times$ faster measurement, a 98\% reduction in data volume and a 99.96\% reduction in post-processing computational complexity, but also successfully captures MPCs and channel characteristics consistent with traditional exhaustive measurements, demonstrating its potential as a fundamental enabler for constructing the massive ISAC datasets required by AI-native 6G systems.

Figures (12)

• Figure 1: Proposed sparse frequency sampling framework for channel sounding.
• Figure 2: $\mathcal{L}_{\text{single}}(\tau)$ for different frequency sampling schemes and molecular absorption (MA) conditions with $\tau_\text{true}=50$ ns ($f_\text{c}=380~\text{GHz}$, $B=10~\text{GHz}$, and $K=35$).
• Figure 3: Proposed parabolic frequency sampling scheme. (a) Local frequency step, $f'(k)$, and (b) Frequency distribution function, $f(k)$. ($f_\text{c}=380~\text{GHz}$, $B=10~\text{GHz}$ and $K=35$).
• Figure 4: Rectified Likelihood function $\mathcal{L}(\tau)$ of the proposed parabolic frequency sampling scheme for a path delay of 50 ns with (a) $K=20$, (b) $K=35$ and (c) $K=50$ ($f_\text{c}=375~\text{GHz}$, $B=10~\text{GHz}$, $T=30^\circ \mathrm{C}$ and $D_\text{VP}=20~\text{g/m}^3$).
• Figure 5: Zoomed-in mainlobe of likelihood function $\mathcal{L}(\tau)$ for a path delay of 50 ns with $K=50$ under (a) Uniform frequency sampling not affected by molecular absorption, (b) Parabolic frequency sampling with molecular absorption rectified by $I(f;\tau)$ and (c) Parabolic frequency sampling with molecular absorption and without rectification. ($f_\text{c}=375~\text{GHz}$, $B=10~\text{GHz}$, $T=30^\circ \mathrm{C}$ and $D_\text{VP}=20~\text{g/m}^3$)