Search

Federated Learning (FL) has emerged as a significant paradigm for training machine learning models. This is due to its data-privacy-preserving property and its efficient exploitation of distributed computational resources. This is achieved by conducting the training process in parallel at distributed users. However, traditional FL strategies grapple with difficulties in evaluating the quality of received models, handling unbalanced models, and reducing the impact of detrimental models. To resolve these problems, we introduce a novel federated learning framework, which we call federated testing for federated learning (FedTest). In the FedTest method, the local data of a specific user is used to train the model of that user and test the models of the other users. This approach enables users to test each other's models and determine an accurate score for each. This score can then be used to aggregate the models efficiently and identify any malicious ones. Our numerical results reveal that the proposed method not only accelerates convergence rates but also diminishes the potential influence of malicious users. This significantly enhances the overall efficiency and robustness of FL systems.

Covert communication in wireless networks ensures that transmissions remain undetectable to adversaries, making it a potential enabler for privacy and security in sensitive applications. However, to meet the high performance and connectivity demands of sixth-generation (6G) networks, future wireless systems will require larger antenna arrays, higher operating frequencies, and advanced antenna architectures. This shift changes the propagation model from far-field planar-wave to near-field spherical-wave which necessitates a redesign of existing covert communication systems. Unlike far-field beamforming, which relies only on direction, near-field beamforming depends on both distance and direction, providing additional degrees of freedom for system design. In this paper, we aim to utilize those freedoms by proposing near-field Frequency Diverse Array (FDA)-based transmission strategies that manipulate the beampattern in both distance and angle, thereby establishing a non-covert region around the legitimate user. Our approach takes advantage of near-field properties and FDA technology to significantly reduce the area vulnerable to detection by adversaries while maintaining covert communication with the legitimate receiver. Numerical simulations show that our methods outperform conventional phased arrays by shrinking the non-covert region and allowing the covert region to expand as the number of antennas increases.

Next-generation mobile networks require evolved radio access network (RAN) architectures to meet the demands of high capacity, massive connectivity, reduced costs, and energy efficiency, and to realize communication with ultra-low latency and ultra-high reliability. {Meeting such} requirements for both mobile users and vertical industries in the next decade {requires novel solutions. One of the potential solutions that attracted significant research attention in the past 15 years} is to redesign the radio access network (RAN). In this survey, we present a comprehensive survey on distributed antenna system (DAS) architectures that address these challenges and improve network performance. We cover the transition from traditional decentralized RAN to DAS, including cloud radio-access networks (C-RAN), fog radio-access networks (F-RAN), virtualized radio-access networks (V-RAN), cell-free massive multiple-input multiple-output (CF-mMIMO), and {the most recent advances manifested in} open radio-access network (O-RAN). In the process, we discuss the benefits and limitations of these architectures, including the impact of limited-capacity fronthaul links, various cooperative uplink and downlink coding strategies, cross-layer optimization, and techniques to optimize the performance of DAS. Moreover, we review key enabling technologies for next-generation RAN systems, such as multi-access edge computing, network function virtualization, software-defined networking, and network slicing; in addition to some crucial radio access technologies, such as millimeter wave, massive multi-input multi-output, device-to-device communication, and massive machine-type communication. Last but not least, we discuss the major research challenges in DAS and identify several possible directions for future research., Comment: 45 pages, 14 figures, survey paper

Key generation, a pillar in physical-layer security (PLS), is the process of the exchanging signals from two legitimate users (Alice and Bob) to extract a common key from the random, common channels. The drawback of extracting keys from wireless channels is the ample dependence on the dynamicity and fluctuations of the radio channel, rendering the key vulnerable to estimation by Eve (an illegitimate user) in low-entropy environments because of insufficient randomness. Added to that, the lack of channel fluctuations lower the secret key rate (SKR) defined as the number of bits of key generated per channel use. In this work, we aim to address this challenge by using a reconfigurable intelligent surface (RIS) to produce random phases at certain, carefully curated intervals such that it disrupts the channel in low-entropy environments. We propose an RIS assisted key generation protocol, study its performance, and compare with benchmarks to observe the benefit of using an RIS while considering various important metrics such as key mismatch rate and secret key throughput. Furthermore, we characterize a scaling law as a function of the rate of change of RIS phase switching for the average secret information rate under this protocol. Then, we use both the key throughput and information rate to optimize the overall secrecy rate. Simulations are made to validate our theoretical findings and effectiveness of the proposed scheme showing an improvement in performance when an RIS is deployed., Comment: arXiv admin note: text overlap with arXiv:2308.04726

A reconfigurable intelligent surface (RIS) is an essential component in the architecture of the next generation of wireless communication systems. An RIS is deployed to provide a controllability to the multi-path environment between the transmitter and the receiver, which becomes critical when the line-of-sight signal between them is blocked. In this work, we design an electrically tunable linearly polarized RIS at 2.5 GHz that yields a controllable reflection phase and phase-frequency slope; in other words, we add tunability of the phase-frequency slope to the tunability of the resonance center frequency. The proposed design consists of two layers of unit cells placed over a ground plane, with dog-bone-shaped elements in the top layer and patch elements in the bottom layer. Each patch and dog-bone element is loaded with a varactor, whose reverse bias voltage is controlled to provide a phase-frequency profile with a slope value of 9 degrees/MHz or 0.95 degrees/MHz, and a phase shift range of 320 degrees., Comment: 2 pages, 4 figures, submitted to 2024 IEEE International Symposium on Antennas and Propagation and ITNC-USNC-URSI Radio Science Meeting

Next-generation communication networks are expected to exploit recent advances in data science and cutting-edge communications technologies to improve the utilization of the available communications resources. In this article, we introduce an emerging deep learning (DL) architecture, the transformer-masked autoencoder (TMAE), and discuss its potential in next-generation wireless networks. We discuss the limitations of current DL techniques in meeting the requirements of 5G and beyond 5G networks, and how the TMAE differs from the classical DL techniques can potentially address several wireless communication problems. We highlight various areas in next-generation mobile networks which can be addressed using a TMAE, including source and channel coding, estimation, and security. Furthermore, we demonstrate a case study showing how a TMAE can improve data compression performance and complexity compared to existing schemes. Finally, we discuss key challenges and open future research directions for deploying the TMAE in intelligent next-generation mobile networks., Comment: accepted to appear at IEEE Communications Magazine

This work studies a multi-cell one-bit massive multiple-input multiple-output (MIMO) system that employs one-bit analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) at each base station (BS). We utilize Bussgang decomposition to derive downlink signal-to-quantization-plus-interference-plus-noise ratio (SQINR) and ergodic achievable rate expressions under one-bit quantized maximum ratio transmission (MRT) and zero-forcing (ZF) precoding schemes considering scenarios with and without pilot contamination (PC) in the derived channel estimates. The results are also simplified for the mixed architecture that employs full resolution (FR) ADCs and one-bit DACs, and the conventional architecture that employs FR ADCs and DACs. The SQINR is shown to decrease by a factor of $2/\pi$ and $4/\pi^2$ in the one-bit setting compared to that achieved in the mixed setting and conventional setting respectively under MRT precoding without PC. Interestingly, the decrease in SQINR is less when we consider PC, which is shown to adversely impact the conventional system more than the one-bit system. Similar insights are obtained under ZF precoding with the decrease in the SQINR with the use of one-bit ADCs and DACs being more pronounced. We utilize the derived expressions to yield performance insights related to power efficiency, the numbers of antennas needed by the three architectures to achieve the same sum-rate, and energy efficiency., Comment: Accepted in IEEE Transactions on Communications

With the expected proliferation of delay constrained applications, future communication technologies are pushed towards using short codes. The performance using short codes cannot be inferred through classical channel capacity analysis, which intrinsically assumes long codes and vanishing frame error rate (FER). This paper studies the performance of an uplink large-scale network in the finite blocklength regime. Bounds on the spatially averaged rate outage probability as well as the coding rate meta distribution are derived. The results reveal the exact achievable rate for a given blocklength and FER, and demonstrate the discrepancy between the actual network rate and idealistic classical channel capacity., Comment: Accepted for publications as a Correspondence in the IEEE Transactions on Vehicular Technology. (6 Pages)

Stacked intelligent metasurface (SIM) is an emerging programmable metasurface architecture that can implement signal processing directly in the electromagnetic wave domain, thereby enabling efficient implementation of ultra-massive multiple-input multiple-output (MIMO) transceivers with a limited number of radio frequency (RF) chains. Channel estimation (CE) is challenging for SIM-enabled communication systems due to the multi-layer architecture of SIM, and because we need to estimate large dimensional channels between the SIM and users with a limited number of RF chains. To efficiently solve this problem, we develop a novel hybrid digital-wave domain channel estimator, in which the received training symbols are first processed in the wave domain within the SIM layers, and then processed in the digital domain. The wave domain channel estimator, parametrized by the phase shifts applied by the meta-atoms in all layers, is optimized to minimize the mean squared error (MSE) using a gradient descent algorithm, within which the digital part is optimally updated. For an SIM-enabled multi-user system equipped with 4 RF chains and a 6-layer SIM with 64 meta-atoms each, the proposed estimator yields an MSE that is very close to that achieved by fully digital CE in a massive MIMO system employing 64 RF chains. This high CE accuracy is achieved at the cost of a training overhead that can be reduced by exploiting the potential low rank of channel correlation matrices.

The reflection characteristics of a reconfigurable intelligent surface (RIS) depend on the phase response of the constituent unit cells, which is necessarily frequency dependent. This paper investigates the role of an RIS constituting unit cells with different phase-frequency profiles in a wide-band orthogonal frequency division multiplexing (OFDM) system to improve the achievable rate. We first propose a mathematical model for the phase-frequency relationship parametrized by the phase-frequency profile's slope and phase-shift corresponding to a realizable resonant RIS unit cell. Then, modelling each RIS element with $b$ control bits, we propose a method for selecting the parameter pairs to obtain a set of $2^b$ phase-frequency profiles. The proposed method yields an RIS design that outperforms existing designs over a wide range of user locations in a single-input, single-output (SISO) OFDM system. We then propose a low-complexity optimization algorithm to maximize the data rate through the joint optimization of (a) power allocations across the sub-carriers and (b) phase-frequency profile for each RIS unit cell from the available set. The analysis is then extended to a multi-user multiple-input single-output (MISO) OFDM scenario. Numerical results show an improvement in the coverage and achievable rates under the proposed framework as compared to single-slope phase-frequency profiles.

Physical layer security (PLS) is superior to classical cryptography techniques due to its notion of perfect secrecy and independence to an eavesdropper's computational power. One form of PLS arises when Alice and Bob (the legitimate users) exchange signals to extract a common key from the random common channels. The drawback of extracting keys from wireless channels is the ample dependence on the dynamicity and fluctuations of the radio channel. However, some radio channels are constant such as line of sight (LoS) and can be estimated by Eve (an illegitimate user), or can be quite static in behaviour due to the presence low-mobility users thus restricting the amount of randomness. This in turn lowers the secret key rate (SKR) defined as the number of bits of key generated per channel use. In this work, we aim to address this challenge by using a reconfigurable intelligent surface (RIS) to produce random phases at certain carefully curated intervals such that it disrupts the channel in low entropy environments. We propose an RIS assisted key generation method, study its performance, and compare with benchmarks to observe the benefit of using an RIS while considering various important metrics such as key mismatch rate and average secret key throughput. Simulations are made to validate our theoretical findings showing an improvement in performance when an RIS is deployed.

This work studies the net sum-rate performance of a distributed reconfigurable intelligent surfaces (RISs)-assisted multi-user multiple-input-single-output (MISO) downlink communication system under imperfect instantaneous-channel state information (I-CSI) to implement precoding at the base station (BS) and statistical-CSI (S-CSI) to design the RISs phase-shifts. Two channel estimation (CE) protocols are considered for I-CSI acquisition: (i) a full CE protocol that estimates all direct and RISs-assisted channels over multiple training sub-phases, and (ii) a low-overhead direct estimation (DE) protocol that estimates the end-to-end channel in a single sub-phase. We derive the asymptotic equivalents of signal-to-interference-plus-noise ratio (SINR) and ergodic net sum-rate under both protocols for given RISs phase-shifts, which are then optimized based on S-CSI. Simulation results reveal that the low-complexity DE protocol yields better net sum-rate than the full CE protocol when used to obtain CSI for precoding. A benchmark full I-CSI based RISs design is also outlined and shown to yield higher SINR but lower net sum-rate than the S-CSI based RISs design.

The development of delay-sensitive applications that require ultra high reliability created an additional challenge for wireless networks. This led to Ultra-Reliable Low-Latency Communications, as a use case that 5G and beyond 5G systems must support. However, supporting low latency communications requires the use of short codes, while attaining vanishing frame error probability (FEP) requires long codes. Thus, developing codes for the finite blocklength regime (FBR) achieving certain reliability requirements is necessary. This paper investigates the potential of Convolutional Neural Networks autoencoders (CNN-AE) in approaching the theoretical maximum achievable rate over a Gaussian channel for a range of signal-to-noise ratios at a fixed blocklength and target FEP, which is a different perspective compared to existing works that explore the use of CNNs from bit-error and symbol-error rate perspectives. We explain the studied CNN-AE architecture, evaluate it numerically, and compare it to the theoretical maximum achievable rate and the achievable rates of polar coded quadrature amplitude modulation (QAM), Reed-Muller coded QAM, multilevel polar coded modulation, and a TurboAE-MOD scheme from the literature. Numerical results show that the CNN-AE outperforms these benchmark schemes and approaches the theoretical maximum rate, demonstrating the capability of CNN-AEs in learning good codes for delay-constrained applications., Comment: This paper is submitted in IEEE International Black Sea Conference on Communications and Networking 2023 (Status: Accepted). This is a 6-pages paper with 3 figures and two tables

Some emerging 5G and beyond use-cases impose stringent latency constraints, which necessitates a paradigm shift towards finite blocklength performance analysis. In contrast to Shannon capacity-achieving codes, the codeword length in the finite blocklength regime (FBR) is a critical design parameter that imposes an intricate tradeoff between delay, reliability, and information coding rate. In this context, this paper presents a novel mathematical analysis to characterize the performance of large-scale downlink networks using short codewords. Theoretical achievable rates, outage probability, and reliability expressions are derived using the finite blocklength coding theory in conjunction with stochastic geometry, and compared to the performance in the asymptotic regime (AR). Achievable rates under practical modulation schemes as well as multilevel polar coded modulation (MLPCM) are investigated. Numerical results provide theoretical performance benchmarks, highlight the potential of MLPCM in achieving close to optimal performance with short codewords, and confirm the discrepancy between the performance in the FBR and that predicted by analysis in the AR. Finally, the meta distribution of the coding rate is derived, providing the percentiles of users that achieve a predefined target rate in a network., Comment: This paper is submitted in IEEE Transactions on Wireless Communications (Status: Accepted). This is a 32-pages paper with 11 figures

The paper provides a new perspective on peak- and average-constrained Gaussian channels. Such channels model optical wireless communication (OWC) systems which employ intensity-modulation with direct detection (IM/DD). First, the paper proposes a new, capacity-preserving vector binary channel (VBC) model, consisting of dependent binary noisy bit-pipes. Then, to simplify coding over this VBC, the paper proposes coding schemes with varying levels of complexity, building on the capacity of binary-symmetric channels (BSC) and channels with state. The achievable rates are compared to capacity and capacity bounds, showing that coding for the BSC with state over the VBC achieves rates close to capacity at moderate to high signal-to-noise ratio (SNR), whereas simpler schemes achieve lower rates at lower complexity. The presented coding schemes are realizable using capacity-achieving codes for binary-input channels, such as polar codes. Numerical results are provided to validate the theoretical results and demonstrate the applicability of the proposed schemes., Comment: 33 pages, to appear in IEEE TCOM

This work investigates the downlink performance of a multi-cell massive multiple-input multiple-output (MIMO) system that employs one-bit analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) in the receiving and transmitting radio frequency (RF) chains at each base station (BS) in order to reduce the power consumption. We utilize Bussgang decomposition to derive the minimum mean squared error (MMSE) channel estimates at each BS based on the quantized received uplink training signals, and the asymptotic closed-form expressions of the achievable downlink rates under one-bit quantized zero-forcing (ZF) precoding implemented using the estimated channels. The derived expressions explicitly show the impact of quantization noise, thermal noise, pilot contamination, and interference, and are utilized to study the number of additional antennas needed at each BS of the one-bit MIMO system to perform as well as the conventional MIMO system. Numerical results verify our analysis, and reveal that despite needing more antennas to achieve the same sum average rate, the one-bit massive MIMO system is more energy-efficient than the conventional system, especially at high sampling frequencies.

The dirty paper channel (DPC) under a peak amplitude constraint arises in an optical wireless broadcast channel (BC), where the state at one receiver is the transmitted signal intended for the other receiver(s). This paper studies a class of peak-constrained DPC that is applicable to the optical wireless BC, where the channel state (i.e, `dirt') takes values from some evenly-spaced grid. For the discrete-state DPC studied this paper, a capacity upper bound is obtained from its state-free counterpart. To lower bound its capacity, classical dirty paper coding schemes are revisited, including Costa's coding for DPC and Tomlinson-Harashima (TH) precoding, which serves as benchmark schemes. To improve the benchmark performance, two new precoding schemes are proposed for the discrete-state DPC. Although the proposed schemes do not achieve the state-free capacity contrary to what is known about the Costa's DPC, achievable rates within a small gap to the state-free capacity are demonstrated for the discrete-state DPC. Using the proposed precoding scheme in a two-user peak-constrained Gaussian BC, a new capacity inner bound (IB) is obtained, and is shown to outperform the truncated Gaussian (TG) based IB and is comparable to the best-known IB., Comment: Accepted in IEEE WCNC 2023

Communication with unmanned aerial vehicles (UAVs) in current terrestrial networks suffers from poor signal strength due to the down-tilt of the access points (APs) that are optimized to serve ground users ends (GUEs). To solve this, one could tilt the AP antenna upwards or allocate more power to serve the UAV. However, this negatively affects GUE downlink (DL) rates. In this paper, we propose to solve this challenge using a reconfigurable intelligent surface (RIS) to enhance the UAV communication while preserving the 3GPP- prescribed downwards antenna tilt and potentially improving the DL performance of the GUE. We show that under conjugate beamforming (CB) precoding and proper power split between GUEs and the UAV at the APs, an RIS with phase-shifts configured to reflect radio signals towards the UAV can significantly improve the UAV DL throughput while simultaneously benefiting the GUEs. The presented numerical results show that the RIS- aided system can serve a UAV with a required data rate while improving the GUEs DL performance relative to that in a CF- MIMO system without a UAV and an RIS. We support this conclusion through simulations under a varying numbers of RIS reflecting elements, UAV heights, and power split factor.