Search

The reconfigurable intelligent surface (RIS) technology is a potential solution to enhance network capacity and coverage without significant investment into additional infrastructure in 6G networks. This work highlights the interest of the mobile communication industry on RIS, and discusses the development of liquid crystal-based RIS for improved energy efficiency and coverage in the millimeter-wave band. Furthermore, the paper discusses perspectives and insights from an industry R&D point of view, addressing relevant use cases, technical requirements, implementation challenges, and practical considerations for RIS deployment optimization in the context of 6G networks. A hardware design of a RIS with liquid crystal at 28 GHz is presented. A propagation model for RIS as a new part in the system architecture is discussed with the approaches of semi-empirical models, geometric models and its combination by the application of artificial intelligence / machine learning. Finally, a channel model for deployment optimization and dimensioning is presented with the findings that rather large RIS is in favor for coverage improvement as well as a greater attenuation at higher frequencies combined with a smaller RIS size.

Low-Rank Adaptation (LoRA) methods have gained popularity in efficient parameter fine-tuning of models containing hundreds of billions of parameters. In this work, instead, we demonstrate the application of LoRA methods to train small-vision models in Federated Learning (FL) from scratch. We first propose an aggregation-agnostic method to integrate LoRA within FL, named FLoCoRA, showing that the method is capable of reducing communication costs by 4.8 times, while having less than 1% accuracy degradation, for a CIFAR-10 classification task with a ResNet-8. Next, we show that the same method can be extended with an affine quantization scheme, dividing the communication cost by 18.6 times, while comparing it with the standard method, with still less than 1% of accuracy loss, tested with on a ResNet-18 model. Our formulation represents a strong baseline for message size reduction, even when compared to conventional model compression works, while also reducing the training memory requirements due to the low-rank adaptation.

We report on simulations of a novel design of optical titanium nitride (TiN)- based Kinetic Inductance Detectors (KIDs) in order to improve their response to optical photons. We propose to separate the meander from the substrate to trap hot phonons generated by optical photons, preventing their rapid propagation through the substrate. These phonons would in turn contribute to the breaking of more Cooper pairs, thereby increasing the response of the detector. In our design, the meander is suspended a few hundred nanometers above the substrate. Furthermore, reflective gold (Au) or aluminum (Al)-based layers can be placed under the meander to improve photon coupling in the optical wavelengths.

This paper tackles the challenges of implementing few-shot learning on embedded systems, specifically FPGA SoCs, a vital approach for adapting to diverse classification tasks, especially when the costs of data acquisition or labeling prove to be prohibitively high. Our contributions encompass the development of an end-to-end open-source pipeline for a few-shot learning platform for object classification on a FPGA SoCs. The pipeline is built on top of the Tensil open-source framework, facilitating the design, training, evaluation, and deployment of DNN backbones tailored for few-shot learning. Additionally, we showcase our work's potential by building and deploying a low-power, low-latency demonstrator trained on the MiniImageNet dataset with a dataflow architecture. The proposed system has a latency of 30 ms while consuming 6.2 W on the PYNQ-Z1 board.

Diabetic retinopathy (DR) is a prevalent complication of diabetes associated with a significant risk of vision loss. Timely identification is critical to curb vision impairment. Algorithms for DR staging from digital fundus images (DFIs) have been recently proposed. However, models often fail to generalize due to distribution shifts between the source domain on which the model was trained and the target domain where it is deployed. A common and particularly challenging shift is often encountered when the source- and target-domain supports do not fully overlap. In this research, we introduce DRStageNet, a deep learning model designed to mitigate this challenge. We used seven publicly available datasets, comprising a total of 93,534 DFIs that cover a variety of patient demographics, ethnicities, geographic origins and comorbidities. We fine-tune DINOv2, a pretrained model of self-supervised vision transformer, and implement a multi-source domain fine-tuning strategy to enhance generalization performance. We benchmark and demonstrate the superiority of our method to two state-of-the-art benchmarks, including a recently published foundation model. We adapted the grad-rollout method to our regression task in order to provide high-resolution explainability heatmaps. The error analysis showed that 59\% of the main errors had incorrect reference labels. DRStageNet is accessible at URL [upon acceptance of the manuscript].

Federated Learning (FL) is a promising distributed method for edge-level machine learning, particularly for privacysensitive applications such as those in military and medical domains, where client data cannot be shared or transferred to a cloud computing server. In many use-cases, communication cost is a major challenge in FL due to its natural intensive network usage. Client devices, such as smartphones or Internet of Things (IoT) nodes, have limited resources in terms of energy, computation, and memory. To address these hardware constraints, lightweight models and compression techniques such as pruning and quantization are commonly adopted in centralised paradigms. In this paper, we investigate the impact of compression techniques on FL for a typical image classification task. Going further, we demonstrate that a straightforward method can compresses messages up to 50% while having less than 1% of accuracy loss, competing with state-of-the-art techniques.

The architecture, engineering and construction (AEC) sector extensively uses documents supporting product and process development. As part of this, organisations should handle big data of hundreds, or even thousands, of technical documents strongly linked together, including CAD design of industrial plants, equipment purchase orders, quality certificates, and part material analysis. However, analysing such records is daunting for users because it gets complicated to sift through hundreds of documents to establish valuable relationships. This paper addresses how knowledge extracted from linked engineering documents contributes to industrial digitalisation under IT/OT convergence. The proposed GraphLED is a system tasked with data processing, graph-based modelling, and colourful visualisation of related documents. The graph-based approach ensures an improved understanding of linked information because the graph structure offers a promising tool to model the underlying data properties of engineering documents. Preliminary system validation indicates quality improvements are possible in the OCR-based data (85.9% of ambiguous text data removed). This work has the potential to benefit the industry by improving the reliability and resilience of industrial production systems through automated summaries of large quantities of documents and their linkage.

Massive multiple-input multiple-output (mMIMO) technology is a way to increase the spectral efficiency of machine-type communications (MTC). To exploit the benefits from large antenna arrays, accurate channel estimation through pilot signals is needed. Massive MTC systems cannot avoid pilot reuse due to the enormous numbers of connected devices. We propose a pilot reuse algorithm based on channel charting (CC) to mitigate pilot contamination in a multi-sector single-cell massive MTC system having spatially correlated channels. We show that after creating an interference map via CC, a simple strategy to allocate the pilot sequences can be implemented. The simulation results show that the CC-based pilot reuse strategy improves channel estimation accuracy, which subsequently improves the symbol detection performance and increases the spectral efficiency compared to other existing schemes. Moreover, the performance of the CC pilot assignment method approaches that of exhaustive search pilot assignment for small network setups., Comment: 17 pages, 11 figures, journal

Constructed facilities should be robust with respect to the loss of load-bearing elements due to abnormal events. Yet, strengthening structures to withstand such damage has a significant impact on construction costs. Strengthening costs should be justified by the threat and should result in smaller expected costs of progressive collapse. In regular frame structures, beams and columns compete for the strengthening budget. In this paper, we present a risk-based formulation to address the optimal design of regular plane frames under element loss conditions. We address the threat probabilities for which strengthening has better cost-benefit than usual design, for different frame configurations, and study the impacts of strengthening extent and cost. The risk-based optimization reveals optimum points of compromise between competing failure modes: local bending of beams, local crushing of columns, and global pancake collapse, for frames of different aspect ratios. The conceptual study is based on a simple analytical model for progressive collapse, but it provides relevant insight for the design and strengthening of real structures., Comment: 36 pages 12 figures

As a key enabler for massive machine-type communications (mMTC), spatial multiplexing relies on massive multiple-input multiple-output (mMIMO) technology to serve the massive number of user equipments (UEs). To exploit spatial multiplexing, accurate channel estimation through pilot signals is needed. In mMTC systems, it is impractical to allocate a unique orthogonal pilot sequence to each UE as it would require too long pilot sequences, degrading the spectral efficiency. This work addresses the design of channel features from correlated fading channels to assist the pilot assignment in multi-sector mMTC systems under pilot reuse of orthogonal sequences. In order to reduce pilot collisions and to enable pilot reuse, we propose to extract features from the channel covariance matrices that reflect the level of orthogonality between the UEs channels. Two features are investigated: covariance matrix distance (CMD) feature and CMD-aided channel charting (CC) feature. In terms of symbol error rate and achievable rate, the CC-based feature shows superior performance than the CMD-based feature and baseline pilot assignment algorithms., Comment: 6 pages, 4 figures

Channel state information (CSI) rapidly becomes outdated in high mobility scenarios, degrading the performance of wireless communication systems. In these cases, time series prediction techniques can be applied to combat the effects of outdated CSI. Recently, it has been shown that recurrent neural networks (RNNs) exhibit outstanding performance in time series prediction tasks. In this paper, we investigate the performance of RNN and long short term memory (LSTM) predictors in a simple Rayleigh flat-fading channel. We conduct numerical experiments to evaluate whether these machine-learning (ML)-based predictors can outperform the optimal linear minimum mean square error Wiener predictor. Our simulation results indicate that the considered neural network predictors outperform the Wiener predictor for small observation window lengths and are more robust under weak channel correlation as well as in the presence of noise. Furthermore, we show that simple shallow RNNs are sufficient to model Rayleigh channels over a wide range of Doppler shifts.

Deploying antenna arrays with an asymptotically large aperture will be central to achieving the theoretical gains of massive MIMO in beyond-5G systems. Such extra-large MIMO (XL-MIMO) systems experience propagation conditions which are not typically observed in conventional massive MIMO systems, such as spatial non-stationarities and near-field propagation. Moreover, standard precoding schemes, such as zero-forcing (ZF), may not apply to XL-MIMO transmissions due to the prohibitive complexity associated with such a large-scale scenario. We propose two novel precoding schemes that aim at reducing the complexity without losing much performance. The proposed schemes leverage a plane-wave approximation and user grouping to obtain a low-complexity approximation of the ZF precoder. Our simulation results show that the proposed schemes offer a possibility for a performance and complexity trade-off compared to the benchmark schemes., Comment: Submitted to Eusipco 2021

We present a novel and low-complexity massive multiple-input multiple-output (MIMO) precoding strategy based on novel findings concerning the subspace separability of Rician fading channels. Considering a uniform rectangular array at the base station, we show that the subspaces spanned by the channel vectors can be factorized as a tensor product between two lower dimensional subspaces. Based on this result, we formulate tensor maximum ratio transmit and zero-forcing precoders. We show that the proposed tensor precoders exhibit lower computational complexity and require less instantaneous channel state information than their linear counterparts. Finally, we present computer simulations that demonstrate the applicability of the proposed tensor precoders in practical communication scenarios., Comment: To appear at Asilomar 2020

New-generation wireless communication systems will employ large-scale antenna arrays to satisfy the increasing capacity demand. This massive scenario brings new challenges to the channel equalization problem due to the increased signal processing complexity. We present a novel low-rank tensor equalizer to tackle the high computational demands of the classical linear approach. Specifically, we propose a method to design a canonical polyadic tensor filter to minimize the mean square error criterion. Our simulation results indicate that the proposed equalizer needs fewer calculations and is more robust to short training sequences than the benchmark., Comment: This revision provides minor corrections to the original paper, which appeared in the Proc. of IEEE ISWCS19. Eq. (7) and below - H in the original paper should read as H_u; Eq. (18) - w_{d,r} in the original paper should read as w_{j,r}; First paragraph of Sec. III-B - the index n is correctly defined as n = n_1 + (n_2-1)*N_1 +... Also, the correct dimensions of tensor W are N_1x...xN_D

We propose practical transceiver structures for double-sided massive multiple-input-multiple-output (MIMO) systems. Unlike standard massive MIMO, both transmit and receive sides are equipped with high-dimensional antenna arrays. We leverage the multi-layer filtering architecture and propose novel layered transceiver schemes with practical channel state information requirements to simplify the complexity of our double-sided massive MIMO system. We conduct a comprehensive simulation campaign to investigate the performance of the proposed transceivers under different channel propagation conditions and to identify the most suitable strategy. Our results show that the covariance matrix eigenfilter design at the outer transceiver layer combined with maximum eigenmode transmission precoding/minimum mean square error combining at the inner transceiver layer yields the best achievable sum rate performance for different propagation conditions and multi-user interference levels.