arXiv subjects

Oways Alsoloh

Publications and source records attributed to Oways Alsoloh.

3 recordsLinked to original sources

arXiv2026

Resilient Cell-Free MIMO ISAC Under Uncertainties: An Impairment-Aware Graph Neural Network Framework

Beamforming designs for integrated sensing and communication (ISAC) in cell-free multiple-input multiple-output (MIMO) systems face the challenge to stay reliable under uncertainties from both channel state information (CSI) as well as clutter. To this end, this paper proposes an impairment-aware heterogeneous graph neural network (GNN) framework for beamforming that obtains the CSI error level and the clutter covariance as edge features, which enables a single trained model to adapt to both impairments. In particular, the GNN is trained on a low quantile of the minimum user rate with a clutter-aware signal-to-clutter-plus-noise ratio, balanced by a normalized Tchebycheff scalarization. Simulation results show that the proposed GNN framework outperforms the considered baselines in sensing-only and communication-only settings. Depending on the operating point, a single GNN model attains approximately 97\% of each baseline's rate while serving both functionalities at once. The model further generalizes to unseen CSI error levels and clutter counts and renders the ISAC system resilient.

eess.SP↗

arXiv2026

Wideband Channel Measurements in the Lower FR3 Band (6-12 GHz) for Industrial Environments

The upper mid-band, also known as Frequency Range 3 (FR3), spans 7.125-24.25 GHz and is a key candidate band for sixth-generation (6G) wireless communications. However, channel measurements in industrial environments remain scarce, and the 3GPP indoor-factory (InF) channel model, developed for omnidirectional antennas, has yet to be validated across the FR3 band. This paper presents a wideband channel measurement campaign in an industrial production hall, covering 135 line-of-sight (LOS) positions at 1-18 m and 30 non-line-of-sight (NLOS) positions behind an industrial robot at 2.5-16.5 m across 6-12 GHz. In LOS, the path-loss exponent ranged from 1.87 to 2.12, and the close-in fits agreed with the 3GPP InF-LOS model to within approximately 2 dB. However, shadow fading was only 1.4-1.7 dB, compared with the model's 4.3 dB. In NLOS, the robot caused an excess loss that increased from 10.3 to 13.0 dB between 6.5 and 11.5 GHz, consistent with diffraction. Blocking the direct path raised the mean directional delay spread from approximately 5 to 25 ns, close to omnidirectional values. These results indicate that the 3GPP InF model captures mean LOS path loss well but overestimates the variability of beam-aligned FR3 links, and that blockage loss in FR3 is frequency dependent. The dataset and processing code are publicly available.

eess.SP↗

arXiv2026

Self-Supervised Noise2Noise-Enhanced Denoising for Continuous-Scan Air-Plasma THz Spectroscopy

Terahertz time-domain spectroscopy (THz-TDS) based on air-plasma generation and balanced air-biased coherent detection offers gap-free broadband coverage, but individual continuous-scan traces are strongly affected by pulse-to-pulse fluctuations and electronic noise. Reaching a useful signal-to-noise ratio therefore requires averaging multiple traces, which directly increases measurement time. We propose a learned denoising approach that recovers high-quality THz waveforms from as few as one complete continuous delay sweep, referred to here as a single-scan trace. A compact one-dimensional residual U-Net is trained using two complementary strategies: a reference-supervised baseline that maps individual noisy traces to long-average reference waveforms, and a Noise2Noise approach that learns from pairs of independently acquired noisy traces without requiring a clean training target. Averaging the predictions of both models reduces systematic bias and yields a trace-reduction factor of approximately $5.4\times$ at $K=1$, meaning that one denoised trace achieves the reconstruction accuracy of averaging approximately five raw traces. The Noise2Noise model alone achieves $4.9\times$, outperforming both the reference-supervised baseline ($4.6\times$) and classical Wiener filtering ($3.2\times$). These results show that self-supervised learning from repeated noisy measurements can support faster continuous-scan THz-TDS without hardware modification.

eess.SP↗