arXiv subjects

Reinhard Wiesmayr

Publications and source records attributed to Reinhard Wiesmayr.

18 recordsLinked to original sources

arXiv2026

Supervised Device Charting with CSI Measurements from Commercial 5G NR User Equipments

Radio frequency fingerprint identification (RFFI) is a promising approach to distinguish physical wireless devices using hardware-induced signal imperfections. Conventional RFFI methods only provide discrete device labels and no human-interpretable representation of the relations among received signals. We propose supervised device charting, which maps location-insensitive channel-state information (CSI) fingerprints to a low-dimensional chart that visualizes cluster compactness, overlap, and outliers. We evaluate the method with real-world 5G New Radio (5G NR) measurements from six commercial smartphones and introduce the neighbor label error rate (NLER) to quantify class-separation accuracy. Our results demonstrate that two- and three-dimensional device charts provide an interpretable visualization of the learned RFFI representation. For three-dimensional device charts, the NLER is 0.22% for same-day measurements and 7.38% for measurements from the next day. The device charts reveal a cross-day distribution shift and map the held-out device close to the known device of the same model. Increasing the device chart dimensions further improves cluster separation at the expense of interpretability.

eess.SP↗

arXiv2026

On the Impact of Site-Specific Training for a Real-World 5G NR System

Site-specific training can improve wireless receiver performance without increasing computational complexity. However, real-world results have so far focused on fully trainable neural receivers and single-layer transmissions. We study site-specific finetuning of three receiver architectures: fully trainable neural, model-driven neural, and model-based. We train and evaluate these receivers using new measurements from a standard-compliant 5G NR testbed at ETH Zurich with dual-layer uplink transmission, including measurement campaigns conducted more than six months apart. Our results show that site-specific finetuning (i) substantially improves fully trainable and model-driven neural receivers, while resulting in only marginal gains for the less tunable model-based receiver; (ii) enables a single neural receiver jointly finetuned for single- and dual-layer transmission to closely match receivers finetuned separately for each configuration; and (iii) remains effective across measurement campaigns separated by more than six months. We also investigate site-specific linear minimum mean-square error channel estimation using covariance matrices estimated from either synthetic channels or site-specific measurements. When combined with iterative detection and decoding, site-specific channel estimation achieves the lowest error rate observed in our datasets. Our finetuning code and measurement datasets are publicly available at https://github.com/IIP-Group/site_specific_training

cs.IT↗

arXiv2026

Spatial and Temporal Generalization of CSI-based Neural Positioning

Channel state information (CSI)-based neural positioning learns a mapping from CSI measurements to user equipment (UE) positions using neural networks. However, most existing performance evaluations utilize randomly partitioned train/test CSI-dataset splits, which fail to reflect the generalization requirements of practical deployments and present optimistic results. In this paper, we study the spatial and temporal generalization of neural positioning with standard-compliant Wi-Fi and 5G NR systems for three real-world CSI datasets acquired in indoor and outdoor environments. We assess generalization with two different architectures, a conventional multilayer perceptron (MLP) and a novel transformer architecture, to unseen spatial regions, unseen UE trajectories, and CSI measurement campaigns separated by one week. Our experiments show that both architectures generalize well in space and time, and the proposed transformer consistently outperforms the MLP in positioning accuracy while requiring fewer model parameters.

eess.SP↗

arXiv2026

Channel Charting for Position and Orientation

Channel charting (CC) in real-world coordinates is a recently proposed self-supervised machine learning method that maps high-dimensional channel state information (CSI) to user equipment (UE) position. In this paper, we extend CC to also estimate UE orientation, which can further assist tasks such as beamfinding, precoding, and beam- and cell-assignment. To this end, we propose a novel orientation triplet loss that accounts for angle periodicity and an alignment loss that embeds estimated orientations in real-world coordinates in a self-supervised fashion. Using real-world CSI measurements from a standard-compliant 5G NR system, we demonstrate that the proposed method achieves position and orientation estimation accuracy close to that of supervised approaches trained with ground-truth labels.

eess.SP↗

arXiv2026

SALAD: Self-Adaptive Link Adaptation

Adapting the modulation and coding scheme (MCS) to the wireless link quality is critical for maximizing spectral efficiency while ensuring reliability. We propose SALAD (self-adaptive link adaptation), an algorithm that exclusively leverages ACK/NACK feedback to reliably track the evolution of the signal-to-interference-plus-noise ratio (SINR), achieving high spectral efficiency while keeping the long-term block error rate (BLER) near a desired target. SALAD infers the SINR by minimizing the cross-entropy loss between received ACK/NACKs and predicted BLER values. Based on this inference, SALAD selects the MCS via hypothesis testing: if the SINR is likely underestimated, a higher MCS is selected to accelerate link adaptation under improving channel conditions. To prevent BLER drift from its long-term target, SALAD incorporates a feedback control loop that adjusts the instantaneous BLER target. Over-the-air experiments on a 5G testbed demonstrate that SALAD consistently outperforms the industry-standard outer-loop link adaptation (OLLA). With a single set of parameters, SALAD achieves up to 15% higher throughput and spectral efficiency than multiple OLLA variants across different traffic regimes, while meeting the BLER target.

cs.IT↗

arXiv2026

SINR Estimation under Limited Feedback via Online Convex Optimization

We introduce a novel online convex optimization (OCO) framework to estimate the user's signal-to-interference-plus-noise ratio (SINR) from ACK/NACK feedback, channel quality indicator (CQI) reports, and previously selected modulation and coding scheme (MCS) values. Specifically, the proposed approach minimizes a regularized binary cross-entropy loss using mirror descent enhanced with Nesterov momentum for accelerated SINR tracking. Its parameters are tuned online via an expert-advice algorithm, endowing the estimator with continual learning capabilities. Numerical experiments in ray-traced scenarios show that the proposed method outperforms state-of-the-art schemes in estimation accuracy and adapts robustly to time-varying SINR regimes.

cs.IT↗

arXiv2026

Site-Specific Finetuning of Neural Receivers with Real-World 5G NR Measurements

Finetuning wireless receivers to a specific deployment scenario can yield significant error-rate performance improvements without increasing processing complexity. However, site-specific finetuning has so far only been demonstrated on synthetic channel data and lacks real-world benchmarks. In this work, we empirically study site-specific finetuning of neural receivers using real-world 5G NR physical uplink shared channel (PUSCH) data collected with an over-the-air testbed at ETH Zurich across three scenarios: (i) a small laboratory, (ii) a large office floor, and (iii) a high-mobility outdoor environment. Our results confirm substantial error-rate performance improvements from site-specific finetuning, consistent with earlier findings based on synthetic channel data. Moreover, we demonstrate that these improvements generalize across different user-equipment hardware and deployment scenarios.

eess.SP↗

arXiv2025

A 14ns-Latency 9Gb/s 0.44mm$^2$ 62pJ/b Short-Blocklength LDPC Decoder ASIC in 22FDX

Ultra-reliable low latency communication (URLLC) is a key part of 5G wireless systems. Achieving low latency necessitates codes with short blocklengths for which polar codes with successive cancellation list (SCL) decoding typically outperform message-passing (MP)-based decoding of low-density parity-check (LDPC) codes. However, SCL decoders are known to exhibit high latency and poor area efficiency. In this paper, we propose a new short-blocklength multi-rate binary LDPC code that outperforms the 5G-LDPC code for the same blocklength and is suitable for URLLC applications using fully parallel MP. To demonstrate our code's efficacy, we present a 0.44mm$^2$ GlobalFoundries 22FDX LDPC decoder ASIC which supports three rates and achieves the lowest-in-class decoding latency of 14ns while reaching an information throughput of 9Gb/s at 62pJ/b energy efficiency for a rate-1/2 code with 128-bit blocklength.

cs.AR↗

arXiv2025

CSI-Based User Positioning, Channel Charting, and Device Classification with an NVIDIA 5G Testbed

Channel-state information (CSI)-based sensing will play a key role in future cellular systems. However, no CSI dataset has been published from a real-world 5G NR system that facilitates the development and validation of suitable sensing algorithms. To close this gap, we publish three real-world wideband multi-antenna multi-open RAN radio unit (O-RU) CSI datasets from the 5G NR uplink channel: an indoor lab/office room dataset, an outdoor campus courtyard dataset, and a device classification dataset with six commercial-off-the-shelf (COTS) user equipments (UEs). These datasets have been recorded using a software-defined 5G NR testbed based on NVIDIA Aerial RAN CoLab Over-the-Air (ARC-OTA) with COTS hardware, which we have deployed at ETH Zurich. We demonstrate the utility of these datasets for three CSI-based sensing tasks: neural UE positioning, channel charting in real-world coordinates, and closed-set device classification. For all these tasks, our results show high accuracy: neural UE positioning achieves 0.6cm (indoor) and 5.7cm (outdoor) mean absolute error, channel charting in real-world coordinates achieves 73cm mean absolute error (outdoor), and device classification achieves 99% (same day) and 95% (next day) accuracy. The CSI datasets, ground-truth UE position labels, CSI features, and simulation code are publicly available at https://caez.ethz.ch

eess.SP↗

arXiv2025

Neural Positioning Without External Reference

Channel state information (CSI)-based user equipment (UE) positioning with neural networks -- referred to as neural positioning -- is a promising approach for accurate off-device UE localization. Most existing methods train their neural networks with ground-truth position labels obtained from external reference positioning systems, which requires costly hardware and renders label acquisition difficult in large areas. In this work, we propose a novel neural positioning pipeline that avoids the need for any external reference positioning system. Our approach trains the positioning network only using CSI acquired off-device and relative displacement commands executed on commercial off-the-shelf (COTS) robot platforms, such as robotic vacuum cleaners -- such an approach enables inexpensive training of accurate neural positioning functions over large areas. We evaluate our method in three real-world scenarios, ranging from small line-of-sight (LoS) areas to larger non-line-of-sight (NLoS) environments, using CSI measurements acquired in IEEE 802.11 Wi-Fi and 5G New Radio (NR) systems. Our experiments demonstrate that the proposed neural positioning pipeline achieves UE localization accuracies close to state-of-the-art methods that require externally acquired high-precision ground-truth position labels for training.

eess.SP↗

arXiv2024

Optimizing Puncturing Patterns of 5G NR LDPC Codes for Few-Iteration Decoding

Rate-matching of low-density parity-check (LDPC) codes enables a single code description to support a wide range of code lengths and rates. In 5G NR, rate matching is accomplished by extending (lifting) a base code to a desired target length and by puncturing (not transmitting) certain code bits. LDPC codes and rate matching are typically designed for the asymptotic performance limit with an ideal decoder. Practical LDPC decoders, however, carry out tens or fewer message-passing decoding iterations to achieve the target throughput and latency of modern wireless systems. We show that one can optimize LDPC code puncturing patterns for such few-iteration-constrained decoders using a method we call swapping of punctured and transmitted blocks (SPAT). Our simulation results show that SPAT yields from 0.20 dB up to 0.55 dB improved signal-to-noise ratio performance compared to the standard 5G NR LDPC code puncturing pattern for a wide range of code lengths and rates.

cs.IT↗

arXiv2024

Design of a Standard-Compliant Real-Time Neural Receiver for 5G NR

We detail the steps required to deploy a multi-user multiple-input multiple-output (MU-MIMO) neural receiver (NRX) in an actual cellular communication system. This raises several exciting research challenges, including the need for real-time inference and compatibility with the 5G NR standard. As the network configuration in a practical setup can change dynamically within milliseconds, we propose an adaptive NRX architecture capable of supporting dynamic modulation and coding scheme (MCS) configurations without the need for any re-training and without additional inference cost. We optimize the latency of the neural network (NN) architecture to achieve inference times of less than 1ms on an NVIDIA A100 GPU using the TensorRT inference library. These latency constraints effectively limit the size of the NN and we quantify the resulting signal-to-noise ratio (SNR) degradation as less than 0.7 dB when compared to a preliminary non-real-time NRX architecture. Finally, we explore the potential for site-specific adaptation of the receiver by investigating the required size of the training dataset and the number of fine-tuning iterations to optimize the NRX for specific radio environments using a ray tracing-based channel model. The resulting NRX is ready for deployment in a real-time 5G NR system and the source code including the TensorRT experiments is available online.

cs.IT↗

arXiv2024

PyJama: Differentiable Jamming and Anti-Jamming with NVIDIA Sionna

Despite extensive research on jamming attacks on wireless communication systems, the potential of machine learning for amplifying the threat of such attacks, or our ability to mitigate them, remains largely untapped. A key obstacle to such research has been the absence of a suitable framework. To resolve this obstacle, we release PyJama, a fully-differentiable open-source library that adds jamming and anti-jamming functionality to NVIDIA Sionna. We demonstrate the utility of PyJama (i) for realistic MIMO simulations by showing examples that involve forward error correction, OFDM waveforms in time and frequency, realistic channel models, and mobility; and (ii) for learning to jam. Specifically, we use stochastic gradient descent to optimize jamming power allocation over an OFDM resource grid. The learned strategies are non-trivial, intelligible, and effective.

eess.SP↗

arXiv2024

LoFi User Scheduling for Multiuser MIMO Wireless Systems

We propose new low-fidelity (LoFi) user equipment (UE) scheduling algorithms for multiuser multiple-input multiple-output (MIMO) wireless communication systems. The proposed methods rely on an efficient guess-and-check procedure that, given an objective function, performs paired comparisons between random subsets of UEs that should be scheduled in certain time slots. The proposed LoFi scheduling methods are computationally efficient, highly parallelizable, and gradient-free, which enables the use of almost arbitrary, non-differentiable objective functions. System simulations in a millimeter-wave (mmWave) multiuser MIMO scenario demonstrate that the proposed LoFi schedulers outperform a range of state-of-the-art user scheduling algorithms in terms of bit error-rate and/or computational complexity.

cs.IT↗

arXiv2023

Bit Error and Block Error Rate Training for ML-Assisted Communication

Even though machine learning (ML) techniques are being widely used in communications, the question of how to train communication systems has received surprisingly little attention. In this paper, we show that the commonly used binary cross-entropy (BCE) loss is a sensible choice in uncoded systems, e.g., for training ML-assisted data detectors, but may not be optimal in coded systems. We propose new loss functions targeted at minimizing the block error rate and SNR deweighting, a novel method that trains communication systems for optimal performance over a range of signal-to-noise ratios. The utility of the proposed loss functions as well as of SNR deweighting is shown through simulations in NVIDIA Sionna.

cs.IT↗

arXiv2022

DUIDD: Deep-Unfolded Interleaved Detection and Decoding for MIMO Wireless Systems

Iterative detection and decoding (IDD) is known to achieve near-capacity performance in multi-antenna wireless systems. We propose deep-unfolded interleaved detection and decoding (DUIDD), a new paradigm that reduces the complexity of IDD while achieving even lower error rates. DUIDD interleaves the inner stages of the data detector and channel decoder, which expedites convergence and reduces complexity. Furthermore, DUIDD applies deep unfolding to automatically optimize algorithmic hyperparameters, soft-information exchange, message damping, and state forwarding. We demonstrate the efficacy of DUIDD using NVIDIA's Sionna link-level simulator in a 5G-near multi-user MIMO-OFDM wireless system with a novel low-complexity soft-input soft-output data detector, an optimized low-density parity-check decoder, and channel vectors from a commercial ray-tracer. Our results show that DUIDD outperforms classical IDD both in terms of block error rate and computational complexity.

cs.IT↗

arXiv2021

Distributed Joint Multi-cell Optimization of IRS Parameters with Linear Precoders

We present distributed methods for jointly optimizing Intelligent Reflecting Surface (IRS) phase-shifts and beamformers in a cellular network. The proposed schemes require knowledge of only the intra-cell training sequences and corresponding received signals without explicit channel estimation. Instead, an SINR objective is estimated via sample means and maximized directly. This automatically includes and mitigates both intra- and inter-cell interference provided that the uplink training is synchronized across cells. Different schemes are considered that limit the set of known training sequences from interferers. With MIMO links an iterative synchronous bi-directional training scheme jointly optimizes the IRS parameters with the beamformers and combiners. Simulation results show that the proposed distributed methods show a modest performance degradation compared to centralized channel estimation schemes, which estimate and exchange all cross-channels between cells, and perform significantly better than channel estimation schemes which ignore the inter-cell interference.

cs.IT↗

arXiv2019

Model Order Selection in DoA Scenarios via Cross-Entropy based Machine Learning Techniques

In this paper, we present a machine learning approach for estimating the number of incident wavefronts in a direction of arrival scenario. In contrast to previous works, a multilayer neural network with a cross-entropy objective is trained. Furthermore, we investigate an online training procedure that allows an adaption of the neural network to imperfections of an antenna array without explicitly calibrating the array manifold. We show via simulations that the proposed method outperforms classical model order selection schemes based on information criteria in terms of accuracy, especially for a small number of snapshots and at low signal-to-noise-ratios. Also, the online training procedure enables the neural network to adapt with only a few online training samples, if initialized by offline training on artificial data.

eess.SP↗