In systems operating with extremely large antenna arrays and high-frequency signaling, multiple users often reside in the radiative near-field, and accurate localization becomes essential. Unlike conventional far-field systems that rely solely on direction of arrival (DoA) estimation, near-field localization exploits spherical wavefront propagation to recover both DoA and range information. While subspace-based methods, such as MUltiple SIgnal Classification (MUSIC) and its extensions, offer high resolution and interpretability for near-field localization, their performance is significantly impacted by model assumptions, including non-coherent sources, well-calibrated arrays, and a sufficient number of snapshots. To address these limitations, this work proposes artificial intelligence (AI)-aided subspace methods for near-field localization that enhance robustness to real-world challenges. Specifically, we introduceNF-SubspaceNet, a deep learning-augmented 2D MUSIC algorithm that learns a surrogate covariance matrix to improve localization under challenging conditions, andDCD-MUSIC, a cascaded AI-aided approach that decouples angle and range estimation to reduce computational complexity. We further develop a novel model-order-aware training method to accurately estimate the number of sources, that is combined with casting of near-field subspace methods as AI models for learning. Extensive simulations demonstrate that the proposed methods outperform classical and existing deep-learning-based localization techniques, providing robust near-field localization even under coherent sources, miscalibrations, and few snapshots.

Future wireless systems are envisioned to utilize the large spectra available at THz bands for wireless communications. Extremely massive multiple-input multiple-output (MIMO) antennas can be costly and power inefficient for wideband THz communications. An alternative antenna technology, which can achieve low-cost and power-efficient THz signaling, is based on leaky wave antennas (LWAs). In this paper, we explore the usage of the LWAs for wideband downlink multi-user THz communications. We propose a model for LWA-aided communication systems that faithfully captures the antenna operations. We show that LWAs yield frequency-dependent beams, where the equivalent wideband channel induces a dependence between angle, frequency, and spectral lobe width. We identify the LWA’s inherent frequency-selective beamsteering capabilities as motivating multi-band THz communications, in which subbands are allocated among users based on their relative angles. Then, we propose an alternating optimization algorithm for jointly optimizing the LWA configuration along with the spectral division and power allocation to maximize the achievable sum rate performance. Our numerical results show that a single LWA can generate diverse beampatterns, exhibiting performance comparable to costly MIMO architectures in wideband THz multi-user systems.

Infrastructure-mounted sensors can capture rich environmental information to enhance communications and facilitate beamforming in millimeter-wave systems. This work presents an efficient sensing-assisted long-term beam tracking framework that selects optimal beams from a codebook for current and multiple future time slots. We first design a large attention-enhanced neural network (NN) to fully exploit past visual observations for beam tracking. A convolutional NN extracts compact image features, while gated recurrent units with attention capture the temporal dependencies within sequences. The large NN then acts as the teacher to guide the training of a lightweight student NN via knowledge distillation. The student requires shorter input sequences yet preserves long-term beam prediction ability. Numerical results demonstrate that the teacher achieves Top-5 accuracies exceeding 93% for current and six future time slots, approaching state-of-the-art performance with a 90% reduction of model parameters. The student closely matches the teacher's performance while reducing the number of model parameters by over 1670% and cutting complexity by over 450%, despite operating with 60% shorter input sequences. This improvement significantly enhances data efficiency, reduces latency, and reduces power consumption in sensing and processing.

Future wireless technologies will require accurate localization of multiple users in the radiative near-field. A leading approach employs subspace decomposition of the input covariance and localizes by peak-finding over the MUltiple SIgnal Classification (MUSIC) spectrum, which is suitable for non-coherent sources with sufficient snapshots and calibrated arrays. This work introduces deep-learning-aided cascaded differentiable MUSIC (DCD-MUSIC) that augments MUSIC near-field localization with dedicated deep neural networks (DNNs), allowing it to operate reliably and interpretably. DCD-MUSIC utilizes two DNNs trained to produce surrogate covariances, one from which the angles and number of sources are recovered, and one to compute the range MUSIC spectrum. This is achieved via a novel learning method that (i) facilitates division into signal and noise subspaces; and (ii) converts MUSIC into a differentiable machine learning model. Our results show that DCD-MUSIC successfully localizes multiple coherent near- and far-field sources.

Integrating deep neural networks (DNNs) into wireless receivers can enhance reliability in the presence of hard-to-model channels. However, in order to successfully deploy deep receivers, one must address the rapid channel variations while accounting for the limited availability of data and computing resources. This paper presents a novel framework for rapid online learning of deep receivers that builds on continual Bayesian learning. By modeling the channel variations as a dynamic system in the space of DNN model parameters, we enable efficient single-step updates, supporting the rapid training of Bayesian DNNs using limited data. We propose two online learning algorithms based on extended Kalman filtering and on Bayesian gradients. Unlike typical approaches that avoid catastrophic forgetting, our methods prioritize adapting to current channel realization. Numerical results show that the proposed continual Bayesian learning formulation yields deep receivers that can effectively adapt to varying channels with minimal computational overhead.

In cable-driven robots, the actuation is transmitted via long cables for control and manipulation of the end-effectors. The cables introduce biases due to non-linear tension, creating a challenge for accurate modeling and localization. This paper presents a novel data-driven algorithm that addresses these biases in partially known dynamic systems, focusing on cable-driven robots. The proposed approach integrates classic Kalman filtering with deep learning to enhance tracking and overcome the limitations of existing state-space modeling. Our algorithm learns from data to explicitly track the robot’s end-effector, while implicitly tracking its cable-induced bias. We demonstrate the effectiveness of the algorithm on two different robotic manipulators: a planar two-link robotic manipulator, and the Raven surgical robot, constituting testbeds for dynamic systems where the bias arises from time-varying cable tension. Our experimental results reveal that while conventional methods like the Extended Kalman Filter struggle with pose tracking due to system biases, our algorithm successfully localizes the robot’s end-effector and uncovers underlying biases, showing significant advancements in predictive capabilities and dynamic system understanding.

Joint communications and sensing (JCAS) is expected to be a crucial technology for future wireless systems. This paper investigates beamforming design for a multi-user multi-target JCAS system to ensure fairness and balance between communications and sensing performance. We jointly optimize the transmit and receive beamformers to maximize the weighted sum of the minimum communications rate and sensing mutual information. The formulated problem is highly challenging due to its non-smooth and non-convex nature. To overcome the challenges, we reformulate the problem into an equivalent but more tractable form. We first solve this problem by alternating optimization (AO) and then propose a machine learning algorithm based on the AO approach. Numerical results show that our scheme scales effectively with the number of the communications users and provides better performance with shorter run time compared to conventional optimization approaches.

Hybrid precoding is essential for implementing massive multiple-input multiple-output (MIMO) transceivers in a scalable and power-efficient manner. Due to the frequent change in channel conditions, rapid adaptation in the precoders are needed. However, tuning hybrid precoders for a given channel involves lengthy and computationally heavy optimization. While recent works managed to limit the number of iterations via deep unfolding, the complexity of each iteration may still be too high for rapid tuning. In this work, we provide approximations to the optimization process of projected gradient ascent based hybrid precoders, which drastically reduce the computational complexity. To cope with the errors induced in doing so, we leverage deep learning techniques, tuning the hyperparameters of the solver to achieve reliable hybrid precoders despite the induced approximations. Our numerical study shows that the proposed learned approximated optimizers operate with limited iterations and complexity reduced by up to 98.5%, with inducing only a minor rate loss compared to full non-approximated solvers. These advancements are useful for smart transportation systems, such as autonomous vehicles and connected infrastructure, where massive MIMO enables real- time communication for safety-critical applications.

Sequential Monte Carlo (SMC), or particle filtering, is widely used in nonlinear state-space systems, but its performance often suffers from poorly approximated proposal and state-transition distributions. This work introduces a differentiable particle filter that leverages the unsupervised variational SMC objective to parameterize the proposal and transition distributions with a neural network, designed to learn from high-dimensional observations. Experimental results demonstrate that our approach outperforms established baselines in tracking the challenging Lorenz attractor from high-dimensional and partial observations. Furthermore, an evidence lower bound based evaluation indicates that our method offers a more accurate representation of the posterior distribution.

Federated learning (FL) is a leading approach for iterative learning using possibly private data available at edge devices. The federated operation gives rise to challenges in privacy leakage, which accumulates in learning, and communication latency. These limitations are often individually mitigated by the introduction of privacy preserving noise and user-selection policies, typically at the cost of accuracy. In this work, we propose Privacy-aware Active User SElection (PAUSE), which balances the trade-off between privacy accumulation, communication latency, and optimization of the learned model, via dedicated user selection. This triplet is used to construct a reward (cost function), according to which a multi-armed bandit (MAB)-based algorithm dynamically chooses a subset of users in each round, while guaranteeing bounded accumulated privacy leakage. We establish a theoretical analysis, systematically showing that the reward growth rate of PAUSE follows the best-known rate in MAB literature. While the privacy guarantees hold by the construction of PAUSE, we numerically validate its associated improved latency and accuracy gains in different experimental settings of FL.

Radio frequency wireless power transfer enables charging low-power mobile devices without relying on wired infrastructures. Current existing wireless power transfer systems are typically designed assuming far-field propagation, where the radiated energy is steered to towards given angles, resulting in limited efficiency and possible radiation in undesired locations. An emerging technology for wireless signaling is based on dynamic metasurface antennas (DMAs), which efficiently realize electrically large arrays. When such arrays are employed at high frequencies, wireless power transfer might take place in the radiating near-field (Fresnel) region, where spherical wave propagation holds, providing more degrees-of-freedom and improved performance. In this article, we study wireless power transfer systems charging multiple devices in the Fresnel region, where the energy transmitter is equipped with a DMA, exploring how the antenna configuration can exploit the spherical wavefront to generate focused energy beams. In particular, after presenting a mathematical model for DMA-based radiating near-field wireless power transfer systems, we characterize the weighted sum-harvested energy maximization problem of the considered system, and we propose an efficient solution to jointly design the DMA weights and digital precoding vector. Then, by accounting for hardware constraints, we further extend our study to encompass practical scenarios with discrete phase shifts in DMA elements. Simulation results show that our design generates focused energy beams capable of improving energy transfer efficiency in the radiating near-field with minimal energy pollution.

This survey article focuses on the emerging connections between machine learning and data compression. While the fundamental limits of classical (lossy) data compression are well-established through rate-distortion theory, recent advancements have uncovered new theoretical analyses and application areas inspired by machine learning. We review recent works on task-based and goal-oriented compression, rate-distortion-perception theory, and compression for estimation and inference. Deep learning-based approaches have provided natural, data-driven methods for compression. Accordingly, we survey recent efforts in applying deep learning techniques to task-based or goal-oriented compression, as well as image/video compression and transmission. Additionally, we discuss the potential use of large language models for text compression. Finally, we outline future research directions in this promising field.

Various communication technologies are expected to utilize manet. By combining manet with noma communications, one can support scalable, spectrally efficient, and flexible network topologies. To achieve these benefits of noma manet, one should determine the transmission protocol, particularly the superposition code. However, the latter involves lengthy optimization that has to be repeated when the topology changes. In this work, we propose an algorithm for rapidly optimizing superposition codes in multi-hop noma manet. To achieve reliable tunning with few iterations, we adopt the emergingdeep unfoldingmethodology, leveraging data to boost reliable settings. Our superposition coding optimization algorithm utilizes a small number of projected gradient steps while learning its per-user hyperparameters to maximize the minimal rate over past channels in an unsupervised manner. The learned optimizer is designed for both settings with full csi, as well as when the channel coefficients are to be estimated from pilots. We show that the combination of principled optimization and machine learning yields a scalable optimizer, that once trained, can be applied to different topologies. We cope with the non-convex nature of the optimization problem by applying parallel-learned optimization with different starting points as a form of ensemble learning. Our numerical results demonstrate that the proposed method enables the rapid setting of high-rate superposition codes for various channels. Index terms— NOMA, MANET, deep unfolding.

Synchronous federated learning (FL) is a popular paradigm for collaborative edge learning. It typically involves a set of heterogeneous devices locally training neural network (NN) models in parallel with periodic centralized aggregations. As some of the devices may have limited computational resources and varying availability, FL latency is highly sensitive to stragglers. Conventional approaches discard incomplete intra-model updates done by stragglers, alter the amount of local workload and architecture, or resort to asynchronous settings; which all affect the trained model performance under tight training latency constraints. In this work, we propose stragglers-aware layerwise federated learning (SALF) that leverages the optimization procedure of NNs via backpropagation to update the global model in a layer-wise fashion. SALF allows stragglers to synchronously convey partial gradients, having each layer of the global model be updated independently with a different contributing set of users. We provide a theoretical analysis, establishing convergence guarantees for the global model under mild assumptions on the distribution of the participating devices, revealing that SALF converges at the same asymptotic rate as FL with no timing limitations. This insight is matched with empirical observations, demonstrating the performance gains of SALF compared to alternative mechanisms mitigating the device heterogeneity gap in FL.

Multiple-input multiple-output (MIMO) systems play a key role in wireless communication technologies. A widely considered approach to realize scalable MIMO systems involves architectures comprised of multiple separate modules, each with its own beamforming capability. Such models accommodate cell-free massive MIMO and partially connected hybrid MIMO architectures. A core issue with the implementation of modular MIMO arises from the need to rapidly set the beampatterns of the modules, while maintaining their power efficiency. This leads to challenging constrained optimization that should be repeatedly solved on each coherence duration. In this work, we propose a power-oriented optimization algorithm for beamforming in uplink modular hybrid MIMO systems, which learns from data to operate rapidly. We derive our learned optimizer by tackling the rate maximization objective using projected gradient ascent steps with momentum. We then leverage data to tune the hyperparameters of the optimizer, allowing it to operate reliably in a fixed and small number of iterations while completely preserving its interpretable operation. We show how power efficient beamforming can be encouraged by the learned optimizer, via boosting architectures with low-resolution phase shifts and with deactivated analog components. Numerical results show that our learn-to-optimize method notably reduces the number of iterations and computation latency required to reliably tune modular MIMO receivers, and that it allows obtaining desirable balances between power efficient designs and throughput.

We investigate the application of the factor graph framework for blind joint channel estimation and symbol detection on time-variant linear inter-symbol interference channels. In particular, we consider the expectation maximization (EM) algorithm for maximum likelihood estimation, which typically suffers from high complexity as it requires the computation of the symbol-wise posterior distributions in every iteration. We address this issue by efficiently approximating the posteriors using the belief propagation (BP) algorithm on a suitable factor graph. By interweaving the iterations of BP and EM, the detection complexity can be further reduced to a single BP iteration per EM step. In addition, we propose a data-driven version of our algorithm that introduces momentum in the BP updates and learns a suitable EM parameter update schedule, thereby significantly improving the performance-complexity tradeoff with a few offline training samples. Our numerical experiments demonstrate the excellent performance of the proposed blind detector and show that it even outperforms coherent BP detection in high signal-to-noise scenarios.

Federated learning (FL) is an emerging paradigm that allows a central server to train machine learning models using remote users' data. Despite its growing popularity, FL faces challenges in preserving the privacy of local datasets, its sensitivity to poisoning attacks by malicious users, and its communication overhead, especially in large-scale networks. These limitations are often individually mitigated by local differential privacy (LDP) mechanisms, robust aggregation, compression, and user selection techniques, which typically come at the cost of accuracy. In this work, we presentcompressed private aggregation (CPA), allowing massive deployments to simultaneously communicate at extremely low bit rates while achieving privacy, anonymity, and resilience to malicious users. CPA randomizes a codebook for compressing the data into a few bits using nested lattice quantizers, while ensuring anonymity and robustness, with a subsequent perturbation to hold LDP. CPA-aided FL is proven to converge in the same asymptotic rate as FL without privacy, compression, and robustness considerations, while satisfying both anonymity and LDP requirements. These analytical properties are empirically confirmed in a numerical study, where we demonstrate the performance gains of CPA compared with separate mechanisms for compression and privacy, as well as its robustness in mitigating the harmful effects of malicious users.

Distributed optimization is a fundamental framework for collaborative inference over networks. The operation is modeled as the joint minimization of a shared objective which typically depends on local observations. Distributed optimization algorithms, such as the distributed alternating direction method of multipliers (D-ADMM), iteratively combine local computations and message exchanges. A main challenge associated with distributed optimization, and particularly with D-ADMM, is that it requires a large number of communications to reach consensus. In this work we proposeunfolded D-ADMM, which follows the emerging deep unfolding methodology to enable D-ADMM to operate reliably with a predefined and small number of messages exchanged by each agent. Unfolded D-ADMM fully preserves the operation of D-ADMM, while leveraging data to tune the hyperparameters of each iteration. These hyperparameters can either be agent-specific, aiming at achieving the best performance within a fixed number of iterations over a given network, or shared among the agents, allowing to learn to distributedly optimize over different networks. We specialize unfolded D-ADMM for two representative settings: a distributed sparse recovery setup, and a distributed machine learning learning scenario. Our numerical results demonstrate that the proposed approach dramatically reduces the number of communications utilized by D-ADMM, without compromising on its performance.

The success of deep neural networks (DNNs) is heavily dependent on computational resources. While DNNs are often employed on cloud servers, there is a growing need to operate DNNs on edge devices. Edge devices are typically limited in their computational resources, yet, often multiple edge devices are deployed in the same environment and can reliably communicate with each other. In this work we propose to facilitate the application of DNNs on the edge by allowing multiple users to collaborate during inference to improve their accuracy. Our mechanism, coined edge ensembles, is based on having diverse predictors at each device, which form an ensemble of models during inference. To mitigate the communication overhead, the users share quantized features, and we propose a method for aggregating multiple decisions into a single inference rule. We analyze the latency induced by edge ensembles, showing that its performance improvement comes at the cost of a minor additional delay under common assumptions on the communication network. Our experiments demonstrate that collaborative inference via edge ensembles equipped with compact DNNs substantially improves the accuracy over having each user infer locally, and can outperform using a single centralized DNN larger than all the networks in the ensemble together.

Deep learning is envisioned to facilitate the operation of wireless receivers, with emerging architectures integrating deep neural networks (DNNs) with traditional modular receiver processing. While deep receivers were shown to operate reliably in complex settings for which they were trained, the dynamic nature of wireless communications gives rise to the need to repeatedly adapt deep receivers to channel variations. However, frequent re-training is costly and ineffective, while in practice, not every channel variation necessitates adaptation of the entire DNN. In this paper, we study concept drift detection for identifying when does a deep receiver no longer match the channel, enabling asynchronous adaptation, i.e., re-training only when necessary. We identify existing drift detection schemes from the machine learning literature that can be adapted for deep receivers in dynamic channels, and propose a novel soft-output detection mechanism tailored to the communication domain. Moreover, for deep receivers that preserve conventional modular receiver processing, we design modular drift detection mechanisms, that simultaneously identify when and which sub-module to re-train. The provided numerical studies show that even in a rapidly time-varying scenarios, asynchronous adaptation via modular drift detection dramatically reduces the number of trained parameters and re-training times, with little compromise on performance.

The track-before-detect (TBD) paradigm can enhance radar detection and tracking of weak targets in the presence of noise and clutter. However, TBD gives rise to challenges in computational complexity and reliance on precise mathematical descriptions of the measurement model. This work presents a TBD algorithm combining dynamic programming and deep learning, augmenting the Viterbi algorithm with a dedicated deep neural network (DNN) to address these challenges. Our method alleviates the computational complexity by implementing state-aware pruning while bypassing an explicit use of a measurement model by utilizing a DNN. We demonstrate the effectiveness of our proposed algorithm using physically compliant Range-Doppler measurements.

Various algorithms combine deep neural networks (DNNs) and Kalman filters (KFs) to learn from data to track in complex dynamics. Unlike classic KFs, DNN-based systems do not naturally provide the error covariance alongside their estimate, which is of great importance in some applications, e.g., navigation. To bridge this gap, in this work we study error covariance extraction in DNN-aided KFs. We examine three main approaches that are distinguished by the ability to associate internal features with meaningful KF quantities such as the Kalman gain (KG) and prior covariance. We identify the differences between these approaches in their requirements and their effect on the training of the system. Our numerical study demonstrates that the above approaches allow DNN-aided KFs to extract error covariance, with most accurate error prediction provided by model-based/data-driven designs.

THz communications are expected to play a profound role in future wireless systems. The current trend of the extremely massive multiple-input multiple-output (MIMO) antenna architectures tends to be costly and power inefficient when implementing wideband THz communications. An emerging THz antenna technology is leaky wave antenna (LWA), which can realize frequency selective beamforming with a single radiating element. In this work, we explore the usage of LWAs technology for wideband multi-user THz communications. We propose a model for the LWA signal processing that is physically compliant facilitating studying LWA-aided communication systems. Focusing on downlink systems, we propose an alternating optimization algorithm for jointly optimizing the LWA configuration along with the signal spectral power allocation to maximize the sum-rate performance. Our numerical results show that a single LWA can generate diverse beampatterns at THz exhibiting performance comparable to costly fully digital MIMO arrays.

Lattice quantization implements vector quantization with a simple structured formulation, that is fully determined by the lattice generator matrix and a distance metric. The conventional approach constructs lattices for quantization by minimizing a bound on the rate-distortion tradeoff, which holds for non-overloaded quantizers, while in practice, overloading prevention typically affects performance. In this work we propose a novel technique for constructing lattice that considers possibly overloaded quantizers, for which we learn the lattice generator matrix by directly evaluating the distortion at its output. For training purposes, we convert the continuous-to-discrete quantizer mapping into a differentiable machine learning model, optimized in an unsupervised manner to best fit the data. Subsequently, the data-driven lattice is fixed and ordinarily combined into the quantization process. We provide numerical studies showing that our method attains improved performance compared with alternative lattice designs for various dimensions, and generalizes well to unseen data.

Various multiple-input multiple-output (MIMO) systems, including cell-free massive MIMO and partially connected hybrid MIMO architectures, beamform using multiple similar multi-antenna modules. While this operation enables implementing scalable MIMO in a power and cost effective manner, the setting of the beampattern involves challenging constrained optimization that should be repeatedly solved on each coherence duration. In this work we propose a rapid optimization algorithm for beamforming in uplink modular hybrid MIMO system based on learn-to-optimize methodology. We tackle the rate maximization objective using projected gradient ascent steps with momentum. We then leverage data to tune the hyperparameters of the optimizer, allowing it to operate reliably in a fixed and small number of iterations while completely preserving its interpretable operation. Numerical results show that our learn-to-optimize method notably reduces the number of iterations and computation latency required to reliably tune modular MIMO receivers.

Combining the classical Kalman filter (KF) with a deep neural network (DNN) enables tracking in partially known state space (SS) models. A major limitation of current DNN-aided designs stems from the need to train them to filter data originating from a specific distribution and underlying SS model. Consequently, changes in the model parameters may require lengthy retraining. While the KF adapts through parameter tuning, the black-box nature of DNNs makes identifying tunable components difficult. Hence, we propose Adaptive KalmanNet (AKNet), a DNN-aided KF that can adapt to changes in the SS model without retraining. Inspired by recent advances in large language model fine-tuning paradigms, AKNet uses a compact hypernetwork to generate context-dependent modulation weights. Numerical evaluation shows that AKNet provides consistent state estimation performance across a continuous range of noise distributions, even when trained using data from limited noise settings.

Polar codes have promising error-correction capabilities. Yet, decoding polar codes is often challenging, particularly with large blocks, with recently proposed decoders based on list-decoding or neural-decoding. The former applies multiple decoders, while the latter family learns to decode from data. In this work we introduce a novel polar decoder that combines list-decoding with neural-decoding, by forming an ensemble of multiple weighted belief-propagation (BP) decoders trained with different data. We employ the cyclic-redundancy check code as a proxy for combining the ensemble decoders and selecting the most-likely decoded word after inference, while facilitating real-time decoding. We evaluate our decoder over a wide range of polar codes lengths, empirically showing gains of around 0.25dB in frame-error rate. Our complexity and latency analysis shows that the number of operations approaches that of a single BP decoder at high SNR.

Analog-to-digital converters (ADCs) are key components in digital signal processing systems. Traditional ADCs are designed to accurately represent analog signals. Emerging technologies, such as neuromorphic ADCs, allow tuning the ADC mapping on the device, possibly adapting it to the system task or power considerations. In this work, we study such task-based acquisition using neuromorphic ADCs while jointly accounting for power minimization as well as a generic classification task. We propose a physically compliant model based on resistive successive approximation register ADCs, integrated with memristor components, that can be adjusted to modify the quantization regions. We propose a data-driven algorithm that jointly tunes the neuromorphic ADC along with the digital and analog processing. Our numerical studies demonstrate the efficacy of our design compared to traditional uniform ADCs, simultaneously improving accuracy by up to 1.5× while reducing power consumption by as much as 75%.

People with brain or spinal cord-related paralysis often need to rely on others for basic tasks, limiting their independence. A potential solution is brain-machine interfaces (BMIs), which could allow them to voluntarily control external devices (e.g., robotic arm) by decoding brain activity to movement commands. In the past decade, deep-learning decoders have achieved state-of-the-art results in most BMI applications, ranging from speech production to finger control. However, the 'black-box' nature of deep-learning decoders could lead to unexpected behaviors, resulting in major safety concerns in real-world physical control scenarios. In these applications, explainable but lower-performing decoders, such as the Kalman filter (KF), remain the norm. In this study, we designed a BMI decoder based on KalmanNet, an extension of the KF that augments its operation with recurrent neural networks to compute the Kalman gain. This results in a varying “trust” that shifts between inputs and dynamics. We used this algorithm to predict finger movements from the brain activity of two monkeys. We compared KalmanNet results offline (pre-recorded data, $n=13$ days) and online (real-time predictions, $n=5$ days) with a simple KF and two recent deep-learning algorithms: tcFNN (non-ReFIT version) and LSTM. KalmanNet achieved comparable or better results than other deep learning models in offline and online modes, relying on the dynamical model for stopping while depending more on neural inputs for initiating movements. We further validated this mechanism by implementing a heteroscedastic KF that used the same strategy, and it also approached state-of-the-art performance while remaining in the explainable domain of standard KFs. However, we also see two downsides to KalmanNet. KalmanNet shares the limited generalization ability of existing deep-learning decoders, and its usage of the KF as an inductive bias limits its performance in the presence of unseen noise distributions. Despite this trade-off, our analysis successfully integrates traditional controls and modern deep-learning approaches to motivate high-performing yet still explainable BMI designs.

The wireless channel is a linear input-output relation that depends non-linearly on the RIS configuration: physics-compliant models involve the inversion of an “interaction” matrix. We identify two independent origins of this structural non-linearity:i) proximity-induced mutual coupling between close-by RIS elements;ii) reverberation-induced long-range coupling between all RIS elements arising from multi-path propagation in complex radio environments. Mathematically, we cast the “interaction” matrix inversion as the sum of an infinite Born series [fori)] or Born-like series [forii)] whoseKth term physically represents paths involvingKbounces between the RIS elements [fori)] or wireless entities [forii)]. We identify the key physical parameters that determine whether these series can be truncated after the first and second term, respectively, as tacitly done in common cascaded models of RIS-parametrized wireless channels. We also quantify the non-linearity of a channel’s RIS parametrization in diverse numerical and experimental radio environments ranging from an anechoic (echo-free) chamber to rich-scattering reverberation chambers to corroborate our analysis. Our findings raise doubts about the reliability of existing performance analyses and channel-estimation protocols for cases in which cascaded models poorly describe the physical reality.

The Kalman filter (KF) is a widely-used algorithm for tracking dynamical systems that can be faithfully captured by state space (SS) models. The need to fully describe an SS model limits its applicability under complex settings, e.g., when tracking based on visual or graphical data. This challenge can be treated by mapping the measurements into latent features obeying some postulated closed-form SS model, and applying the KF in the latent space. However, the validity of this approximated SS model may constitute a limiting factor. In this work we tackle the challenges associated with tracking from high-dimensional measurements by jointly learning the KF along with the latent space mapping. Our proposed approach combines a learned encoder while tracking in the latent space using the recently proposed data-driven Kalman-Net, and having both modules jointly tuned from data. Our empirical results demonstrate that the proposed approach achieves improved performance over both model-based and data-driven techniques, by learning a surrogate latent representation that most facilitates tracking.

Stochastic control deals with finding an optimal control signal for a dynamical system in a setting with uncertainty, playing a key role in numerous applications. The linear quadratic Gaussian (LQG) is a widely-used setting, where the system dynamics is represented as a linear Gaussian state-space (SS) model, and the objective function is quadratic. For this setting, the optimal controller is obtained in closed form by the separation principle. However, in practice, the underlying system dynamics often cannot be faithfully captured by a fully known linear Gaussian SS model, limiting its performance. Here, we present LQGNet, a stochastic controller that leverages data to operate under partially known dynamics. LQGNet augments the state tracking module of separation-based control with a dedicated trainable algorithm. The resulting system preserves the operation of classic LQG control while learning to cope with partially known SS models without having to fully identify the dynamics. We empirically show that LQGNet outperforms classic stochastic control by overcoming mismatched SS models.

Pairs trading is a family of trading policies based on monitoring the relationships between pairs of assets. A common pairs trading approach relies on state space (SS) modeling, from which financial indicators can be obtained with low complexity and latency using a Kalman filter (KF), and processed using classic policies such as Bollinger bands (BB). However, such SS models are inherently approximated and mismatched, often degrading the revenue. In this work we propose KalmanNet Bollinger Trading (KalmanBOT), a dataaided policy that preserves the advantages of KF-aided BB policies while leveraging data to overcome the approximated nature of the SS model. We adopt the recent KalmanNet architecture, and approximate BB with a differentiable mapping, converting the policy into a trainable model. We empirically demonstrate that KalmanBOT yields improved rewards compared with model-based and data-driven benchmarks.

Dynamic metasurface antennas (DMAs) provide a new paradigm to realize large-scale antenna arrays for future wireless systems. In this paper, we study the downlink millimeter wave (mmWave) DMA systems with limited number of radio frequency (RF) chains. By using the specific DMA structure, an equivalent mmWave channel model with hybrid beamforming is first explicitly characterized. Based on that, we propose an effective joint microstrip selection and beamforming scheme to accommodate for the limited number of RF chains. A low-complexity digital beamforming solution with channel gain-based microstrip selection is developed, while the analog beamformer is obtained via a coordinate ascent method. The proposed scheme is numerically shown to approach the performance of DMAs without RF chain reduction, verifying the effectiveness of the proposed schemes.

Federated learning (FL) allows a central server to train a model using remote users’ data. FL faces challenges in preserving the local datasets privacy and in its communication overhead; which is considerably dominant in large-scale networks. These limitations are often mitigated individually by local differential privacy (LDP) mechanisms, compression, and user-selection techniques, which often come at the cost of accuracy. In this work we present compressed private aggregation (CPA), which allows massive deployments to simultaneously communicate at extremely low bit-rates while achieving privacy, anonymity, and resilience to malicious users. CPA randomizes a code-book for compressing the data into a few bits, ensuring anonymity and robustness, with a subsequent perturbation to hold LDP. We provide both a theoretical analysis and a numerical study, demonstrating the performance gains of CPA compared with separate mechanisms for compression and privacy.

Electrocardiographic signals (ECG) are used in many healthcare applications, including at-home monitoring of vital signs. These applications often rely on wearable technology and provide low quality ECG signals. Although many methods have been proposed for denoising the ECG to boost its quality and enable clinical interpretation, these methods typically fall short for ECG data obtained with wearable technology, because of either their limited tolerance to noise or their limited flexibility to capture ECG dynamics. This paper presents HKF, a hierarchical Kalman filtering method, that leverages a patient-specific learned structured prior of the ECG signal, and integrates it into a state space model to yield filters that capture both intra- and inter-heartbeat dynamics. HKF is demonstrated to outperform previously proposed methods such as the model-based Kalman filter and data-driven autoencoders, in ECG denoising task in terms of mean-squared error, making it a suitable candidate for application in extramural healthcare settings.

Hybrid beamforming (HBF) is a key enabler for millimeter-wave (mmWave) communications systems, but HBF optimizations are often non-convex and of large dimension. In this paper, we propose an efficient deep unfolding-based HBF scheme, referred to as ManNet-HBF, that approximately maximizes the system spectral efficiency (SE). It first factorizes the optimal digital beamformer into analog and digital terms, and then reformulates the resultant matrix factorization problem as an equivalent maximum-likelihood problem, whose analog beamforming solution is vectorized and estimated efficiently with ManNet, a lightweight deep neural network. Numerical results verify that the proposed ManNet-HBF approach has near-optimal performance comparable to or better than conventional model-based counterparts, with very low complexity and a fast run time. For example, in a simulation with 128 transmit antennas, it attains 98.62% the SE of the Riemannian manifold scheme but 13250 times faster.

Distributed optimization arises in various applications. A widely-used distributed optimizer is the distributed alternating direction method of multipliers (D-ADMM) algorithm, which enables agents to jointly minimize a shared objective by iteratively combining local computations and message exchanges. However, D-ADMM often involves a large number of possibly costly communications to reach convergence, limiting its applicability in communications-constrained networks. In this work we propose unfolded D-ADMM, which facilitates the application of D-ADMM with limited communications using the emerging deep unfolding methodology. We utilize the conventional D-ADMM algorithm with a fixed number of communications rounds, while leveraging data to tune the hyperparameters of each iteration of the algorithm. By doing so, we learn to optimize with limited communications, while preserving the interpretability and flexibility of the original D-ADMM algorithm. Our numerical results demonstrate that the proposed approach dramatically reduces the number of communications utilized by D-ADMM, without compromising on its performance.

We consider the problem of recovering random, time-varying graph processes in a nonlinear dynamic system. The Extended Kalman filter (EKF) is a suitable estimator for such dynamics, but its implementation tends to be complex and possibly unstable when tracking high-dimensional graph signals. To tackle this, we propose the graph signal processing (GSP)-EKF, which replaces the Kalman gain in the EKF with a graph filter that aims to minimize the computed prediction error. The resulting structure of the GSP-EKF Kalman gain increases the numerical stability and reduces the computational burden compared with the standard EKF, particularly when dealing with bandlimited graph processes. We show that for a measurement model with orthogonal graph frequencies, the GSP-EKF coincides with the EKF. The GSP-EKF is evaluated for graph signal tracking in power system state estimation. It is shown that in this case, the proposed GSP-EKF 1) attains the EKF under the accurate model; and 2) outperforms the EKF under a model mismatch, while being notably less complex in both cases.

Direction of arrival (DoA) estimation is a fundamental task in array processing. A popular family of DoA estimation algorithms are subspace methods, which operate by dividing the measurements into distinct signal and noise subspaces. Subspace methods, such as Root-MUSIC, require the sources to be non-coherent, and are considerably degraded when this does not hold. In this work we propose Deep Root-MUSIC (DR-MUSIC); a data-driven DoA estimator which augments Root-MUSIC with a deep neural network applied to the empirical autocorrelation of the input. DR-MUSIC learns how to divide the observations into distinguishable subspaces, thus leveraging data to cope with coherent sources, low SNR and limited snapshots, while preserving the interpretability and the suitability of the model-based algorithm.

The Kalman filter (KF) is a widely-used algorithm for tracking the latent state of a dynamical system from noisy observations. For systems that are well-described by linear Gaussian state space models, the KF minimizes the mean-squared error (MSE). However, in practice, observations are corrupted by outliers, severely impairing the KF’s performance. In this work, an outlier-insensitive KF (OIKF) is proposed, where robustness is achieved by modeling a potential outlier as a normally distributed random variable with unknown variance (NUV). The NUV’s variance is estimated online, using both expectation-maximization (EM) and alternating maximization (AM). The former was previously proposed for the task of smoothing with outliers and was adapted here to filtering, while both EM and AM obtained the same performance and outperformed the other algorithms, the AM approach is less complex and thus requires 40% less runtime. Our empirical study demonstrates that the MSE of our proposed outlier-insensitive KF outperforms previously proposed algorithms, and that for data clean of outliers, it reverts to the classic KF, i.e., MSE optimality is preserved.

Sixth generation (6G) cellular communications are expected to support enhanced wireless localization capabilities. The widespread deployment of large arrays and high-frequency bandwidths give rise to new considerations for localization applications. Emerging antenna architectures, such as dynamic metasurface antennas (DMAs), are expected to be frequently utilized thanks to the achievable high angular resolution and low hardware complexity. Further, wireless localization is likely to take place in the radiating near-field (Fresnel) region, which provides new degrees of freedom, because of the adoption of arrays with large apertures. While current studies mostly focus on the use of costly fully-digital antenna arrays, in this paper we investigate how DMAs can be applied for near-field localization of a single user. We use a direct positioning estimation method based on curvature-of-arrival of the impinging wavefront to obtain the user location, and characterize the effects of DMA tuning on the estimation accuracy. Next, we propose an algorithm for configuring the DMA to optimize near-field localization, by first tuning the adjustable DMA coefficients to minimize the estimation error using postulated knowledge of the actual user position. Finally, we propose a sub-optimal iterative algorithm that does not rely on such knowledge. Simulation results show that the DMA-based near-field localization accuracy could approach that of fully-digital arrays at lower cost.

Signal processing, communications, and control have traditionally relied on classical statistical modeling techniques. Such model-based methods utilize mathematical formulations that represent the underlying physics, prior information, and additional domain knowledge. Simple classical models are useful but sensitive to inaccuracies and may lead to poor performance when real systems display complex or dynamic behavior. On the other hand, purely data-driven approaches that are model-agnostic are becoming increasingly popular as datasets become abundant and the power of modern deep learning pipelines increases. Deep neural networks (DNNs) use generic architectures that learn to operate from data and demonstrate excellent performance, especially for supervised problems. However, DNNs typically require massive amounts of data and immense computational resources, limiting their applicability for some scenarios. In this article, we present the leading approaches for studying and designing model-based deep learning systems. These are methods that combine principled mathematical models with data-driven systems to benefit from the advantages of both approaches. Such model-based deep learning methods exploit both partial domain knowledge, via mathematical structures designed for specific problems, and learning from limited data. Among the applications detailed in our examples for model-based deep learning are compressed sensing, digital communications, and tracking in state-space models. Our aim is to facilitate the design and study of future systems at the intersection of signal processing and machine learning that incorporate the advantages of both domains.

Hybrid precoding plays a key role in realizing massive multiple-input multiple-output (MIMO) transmitters with controllable cost. MIMO precoders are required to frequently adapt based on the variations in the channel conditions. In hybrid MIMO, where precoding is comprised of digital and analog beamforming, such an adaptation involves lengthy optimization and depends on accurate channel state information (CSI). This affects the spectral efficiency when the channel varies rapidly and when operating with noisy CSI. In this work we employ deep learning techniques to learn how to rapidly and robustly optimize hybrid precoders, while being fully interpretable. We leverage data to learn iteration-dependent hyperparameter settings of projected gradient sum-rate optimization with a predefined number of iterations. The algorithm maps channel realizations into hybrid precoding settings while preserving the interpretable flow of the optimizer and improving its inference speed. To cope with noisy CSI, we learn to optimize the minimal achievable sum-rate among all tolerable errors, proposing a hybrid precoder based on the projected conceptual mirror prox minimax optimizer. Numerical results demonstrate that our approach allows using over ten times less iterations compared to that required by conventional optimization with shared hyperparameters, while achieving similar and even improved performance.

Reconfigurable Intelligent Surfaces (RISs) are envisioned to play a key role in future wireless communications, enabling programmable radio propagation environments. They are usually considered as almost passive planar structures that operate as adjustable reflectors, giving rise to a multitude of implementation challenges, including the inherent difficulty in estimating the underlying wireless channels. In this paper, we focus on the recently conceived concept of Hybrid Reconfigurable Intelligent Surfaces (HRISs), which do not solely reflect the impinging waveform in a controllable fashion, but are also capable of sensing and processing an adjustable portion of it. We first present implementation details for this metasurface architecture and propose a convenient mathematical model for characterizing its dual operation. As an indicative application of HRISs in wireless communications, we formulate the individual channel estimation problem for the uplink of a multi-user HRIS-empowered communication system. Considering first a noise-free setting, we theoretically quantify the advantage of HRISs in notably reducing the amount of pilots needed for channel estimation, as compared to the case of purely reflective RISs. We then present closed-form expressions for the Mean-Squared Error (MSE) performance in estimating the individual channels at the HRISs and the base station for the noisy model. Based on these derivations, we propose an automatic differentiation-based first-order optimization approach to efficiently determine the HRIS phase and power splitting configurations for minimizing the weighted sum-MSE performance. Our numerical evaluations demonstrate that HRISs do not only enable the estimation of the individual channels in HRIS-empowered communication systems, but also improve the ability to recover the cascaded channel, as compared to existing methods using passive and reflective RISs.

Programmable radio environments parametrized by reconfigurable intelligent surfaces (RISs) are emerging as a new wireless communications paradigm, but currently used channel models for the design and analysis of signal-processing algorithms cannot include fading in a manner that is faithful to the underlying wave physics. To overcome this roadblock, we introduce a physics-based end-to-end model of RIS-parametrized wireless channels with adjustable fading (coined PhysFad) which is based on a first-principles coupled-dipole formalism. PhysFad naturally incorporates the notions of space and causality, dispersion (i.e., frequency selectivity) and the intertwinement of each RIS element’s phase and amplitude response, as well as any arising mutual coupling effects including long-range mesoscopic correlations. The latter are induced by reverberation and yield a highly nonlinear parametrization of wireless channels through RISs, a pivotal property which is to date completely overlooked. PhysFad offers the to-date missing tuning knob for physics-compliant adjustable fading. We thoroughly characterize PhysFad and demonstrate its capabilities for a prototypical problem of RIS-enabled over-the-air channel equalization in rich-scattering wireless communications. We also share a user-friendly version of our code to help the community transition towards physics-based models with adjustable fading.

Recent years have witnessed growing interest in the application of deep neural networks (DNNs) for receiver design, which can potentially be applied in complex environments without relying on knowledge of the channel model. However, the dynamic nature of communication channels often leads to rapid distribution shifts, which may require periodically retraining. This paper formulates a data-efficient two-stage training method that facilitates rapid online adaptation. Our training mechanism uses a predictive meta-learning scheme to train rapidly from data corresponding to both current and past channel realizations. Our method is applicable to any deep neural network (DNN)-based receiver, and does not require transmission of new pilot data for training. To illustrate the proposed approach, we study DNN-aided receivers that utilize an interpretable model-based architecture, and introduce a modular training strategy based on predictive meta-learning. We demonstrate our techniques in simulations on a synthetic linear channel, a synthetic non-linear channel, and a COST 2100 channel. Our results demonstrate that the proposed online training scheme allows receivers to outperform previous techniques based on self-supervision and joint-learning by a margin of up to 2.5 dB in coded bit error rate in rapidly-varying scenarios.

Deep neural networks (DNNs) allow digital receivers to learn to operate in complex environments. To do so, DNNs should preferably be trained using large labeled data sets with a similar statistical relationship as the one under which they are to infer. For DNN-aided receivers, obtaining labeled data conventionally involves pilot signalling at the cost of reduced spectral efficiency, typically resulting in access to limited data sets. In this paper, we study how one can enrich a small set of labeled pilots data into a larger data set for training deep receivers. Motivated by the widespread use of data augmentation techniques for enriching visual and text data, we propose dedicated augmentation schemes that exploits the characteristics of digital communication data. We identify the key considerations in data augmentations for deep receivers as the need for domain orientation, class (constellation) diversity, and low complexity. Following these guidelines, we devise three complementing augmentations that exploit the geometric properties of digital constellations. Our combined augmentation approach builds on the merits of these different augmentations to synthesize reliable data from a momentary channel distribution, to be used for training deep receivers. Furthermore, we exploit previous channel realizations to increase the reliability of the augmented samples. The superiority of our approach is numerically evaluated for training several deep receiver architectures in different channel conditions. We consider both linear and non-linear synthetic channels, as well as the COST 2100 channel generator, for both single-input single-output and multiple-input multiple-output scenarios. We show that our combined augmentations approach allows DNN-aided receivers to achieve gains of up to 1 dB in bit error rate and of up to$\times 3$in spectral efficiency, compared to regular non-augmented training. Moreover, we demonstrate that our augmentations benefit training even as the number of pilots increases, and perform an ablation study on the different augmentations, which shows that the combined approach surpasses each individual augmentation technique.

The ongoing pandemic and the necessity of frequent testing have spurred a growing interest in pooled testing. Conventional recovery methods from pooled tests are based on group testing or compressed sensing tools which rely on simplistic modeling of the pooling process, and may not be reliable in the presence of complex and noisy measurement procedures and highly infected populations. In this work, we propose a strategy for pooled testing designed for noisy settings, which bypasses the need for a tractable acquisition model. This is achieved by combining deep learning, for implicitly learning the measurement relationship from data, with factor graph inference, which exploits the structured known pooling pattern. Learned factor graphs provide a quantitative readout corresponding to the infection severity, as opposed to group testing which only detects the presence of infection. The proposed scheme is shown to achieve improved robustness to noise compared with previous approaches and to reliably estimate in highly infected populations.

Deep neural networks (DNNs) allow digital receivers to operate in complex environments by learning from data corresponding to the channel input-output relationship. Since communication channels change over time, DNN-aided receivers may be required to retrain periodically, which conventionally involves excessive pilot signaling at the cost of reduced spectral efficiency. In this paper, we study how one can obtain data for retraining deep receivers without sending pilots or relying on specific protocol redundancies, by combining self-supervision with active learning concepts. We focus on the recently proposed ViterbiNet receiver, which integrates into the Viterbi algorithm a DNN for learning the channel. To enable self-supervision, we use the soft-output Viterbi algorithm to evaluate the decision confidence for each of the detected symbols in a given word. Then, to overcome learning with erroneous data, we choose a subset of the recovered symbols to be used for retraining via active learning. The proposed method selects decision-directed data whose confidence is not too low to result in inaccurate labeling, yet not too high to preserve sufficient diversity of the data. We demonstrate that self-supervised symbol-level training yields a performance within a small gap of the Viterbi algorithm with instantaneous channel knowledge.

Providing a metric of uncertainty alongside a state estimate is often crucial when tracking a dynamical system. Classic state estimators, such as the Kalman filter (KF), provide a time-dependent uncertainty measure from knowledge of the underlying statistics; however, deep learning based tracking systems struggle to reliably characterize uncertainty. In this paper, we investigate the ability of KalmanNet, a recently proposed; hybrid; model-based; deep state tracking algorithm, to estimate an uncertainty measure. By exploiting the interpretable nature of KalmanNet, we show that the error covariance matrix can be computed based on its internal features, as an uncertainty measure. We demonstrate that when the system dynamics are known, KalmanNet—which learns its mapping from data without access to the statistics—provides uncertainty similar to that provided by the KF; and while in the presence of evolution model-mismatch, KalmanNet provides a more accurate error estimation.

Direction of arrival (DoA) estimation is a crucial task in sensor array signal processing, giving rise to various successful model-based (MB) algorithms as well as recently developed data-driven (DD) methods. This paper introduces a new hybrid MB/DD DoA estimation architecture, based on the classical multiple signal classification (MUSIC) algorithm. Our approach augments crucial aspects of the original MUSIC structure with specifically designed neural architectures, allowing it to overcome certain limitations of the purely MB method, such as its inability to successfully localize coherent sources. The deep augmented MUSIC algorithm is shown to outperform its unaltered version with a superior resolution.

In this work, we consider the acquisition of stationary signals using uniform analog-to-digital converters (ADCs), i.e., employing uniform sampling and scalar uniform quantization. We jointly optimize the pre-sampling and reconstruction filters to minimize the time-averaged mean-squared error (TMSE) in recovering the continuous-time input signal for a fixed sampling rate and quantizer resolution and obtain closed-form expressions for the minimal achievable TMSE. We show that the TMSE-minimizing pre-sampling filter omits aliasing and discards weak frequency components to resolve the remaining ones with higher resolution when the rate budget is small. In our numerical study, we validate our results and show that sub-Nyquist sampling often minimizes the TMSE under tight rate budgets at the output of the ADC.

The smoothing task is the core of many signal processing applications. It deals with the recovery of a sequence of hidden state variables from a sequence of noisy observations in a one-shot manner. In this work we propose RTSNet, a highly efficient model-based and data-driven smoothing algorithm. RTSNet integrates dedicated trainable models into the flow of the classical Rauch-Tung-Striebel (RTS) smoother, and is able to outperform it when operating under model mismatch and non-linearities while retaining its efficiency and interpretability. Our numerical study demonstrates that although RTSNet is based on more compact neural networks, which leads to faster training and inference times, it outperforms the state-of-the-art, data-driven smoother in a non-linear use case.

We propose a convolutional neural network (CNN) aided factor graphs assisted by mutual information features estimated by a neural network for seizure detection. Specifically, we use neural mutual information estimation to evaluate the correlation between different electroencephalogram (EEG) channels as features. We then use a 1D-CNN to extract extra features from the EEG signals and use both features to estimate the probability of a seizure event. Finally, learned factor graphs are employed to capture the temporal correlation in the signal. Both sets of features from the neural mutual estimation and the 1D-CNN are used to learn the factor nodes. We show that the proposed method achieves state-of-the-art performance using 6-fold leave-four-patients-out cross-validation.

The integration of Reconfigurable Intelligent Surfaces (RISs) into wireless environments endows channels with programmability, and is expected to play a key role in future communication standards. To date, most RIS-related efforts focus on quasi-free-space, where wireless channels are typically modeled analytically. Many realistic communication scenarios occur, however, in rich-scattering environments which, moreover, evolve dynamically. These conditions present a tremendous challenge in identifying an RIS configuration that optimizes the achievable communication rate. In this paper, we make a first step toward tackling this challenge. Based on a simulator that is faithful to the underlying wave physics, we train a deep neural network as surrogate forward model to capture the stochastic dependence of wireless channels on the RIS configuration under dynamic rich-scattering conditions. Subsequently, we use this model in combination with a genetic algorithm to identify RIS configurations optimizing the communication rate. We numerically demonstrate the ability of the proposed approach to tune RISs to improve the achievable rate in rich-scattering setups.

Multiple-input multiple-output (MIMO) systems utilize multiple antennas and signal acquisition chains, facilitating multi-user communications with increased spectral efficiency and better coverage via beamforming. MIMO systems are typically costly to implement and consume high power. A commonly used method to reduce the cost of MIMO receivers is to design hybrid analog/digital beamforming (HBF), which reduces the number of RF chains. However, the added analog circuitry involves active components whose consumed power may surpass that saved in RF chain reduction. An additional method to realize power-efficient MIMO systems is to use low-resolution analog-to-digital converters (ADCs), however, compromising signal recovery accuracy. In this work, we propose a power-efficient hybrid MIMO receiver with dedicated beamforming to mitigate spatial interferers in congested environments, utilizing low-quantization rate ADCs, jointly optimizing the analog and digital processing using task-specific quantization techniques. We present an efficient analog pre-processing hardware architecture utilizing sparse low-resolution vector modulators to reduce analog processing power while maintaining recovery accuracy. Supported by numerical simulations and power analysis, our power-efficient MIMO receiver achieves comparable signal recovery performance to power-hungry fully-digital MIMO receivers using high-resolution ADCs. Furthermore, our receiver outperforms the task-agnostic HBF receivers with low-quantization rate ADCs in recovery accuracy at lower power.

Closing the gap between high data rates and low delay in real-time streaming applications is a major challenge in advanced communication systems. While adaptive network coding schemes have the potential of balancing the rate and the delay in real-time, they often rely on a prediction of the channel behavior. In practice, such a prediction is based on delayed feedbacks, making it difficult to acquire causality, particularly when the channel model is unknown. In this work, we propose a deep learning-based noise prediction (DeepNP) algorithm, which augments the recently proposed adaptive and causal random linear network coding scheme with a neural network that learns to carry out noise prediction from data. This neural augmentation is utilized to maximize the throughput while minimizing in-order delivery delay of the coding scheme, and operate in a channel-model-agnostic manner. We numerically show that performance can dramatically increase by the learned prediction of the channel noise rate, demonstrating that DeepNP gains up to a factor of four in mean and maximum delay and a factor of two in throughput compared with statistic-based network coding approaches.

Federated learning (FL) is an emerging paradigm for training machine learning models using possibly private data available at edge devices. Among the key challenges associated with FL are first the need to preserve the privacy of the local data sets, and second the communication load due to the repeated exchange of updated models; both are often tackled individually with methods whose operation distorts the updated models, e.g., local differential privacy (LDP) mechanisms and lossy compres- sion, respectively. In this work we propose a method for joint privacy enhancement and quantization (JoPEQ), unifying lossy compression and privacy enhancement for FL. JoPEQ utilizes universal vector quantization, where distortion is statistically equivalent to additive noise, and augments the compression distortion with dedicated privacy preserving noise to simultaneously achieve compression and a desired privacy level. We numerically demonstrate that JoPEQ reduces the overall distortion compared to individual LDP and compression, which is translated into improved trained models.

Anomaly detection among a large number of processes arises in many applications ranging from dynamic spectrum access to cybersecurity. In such problems one can often obtain noisy observations aggregated from a chosen subset of processes that conforms to a tree structure. The distribution of these observations, based on which the presence of anomalies is detected, may be only partially known. This gives rise to the need for a search strategy designed to account for both the sample complexity and the detection accuracy, as well as cope with statistical models that are known only up to some missing parameters. In this work we propose a sequential search strategy using two variations of the Generalized Log Likelihood Ratio statistic. Our proposed Hierarchical Dynamic Search (HDS) strategy is shown to be order-optimal with respect to the size of the search space and asymptotically optimal with respect to the detection accuracy. An explicit upper bound on the error probability of HDS is established for the finite sample regime. Extensive experiments are conducted, demonstrating the performance gains of HDS over existing methods.

Large antenna arrays and high-frequency bands are two key features of future wireless communication systems. The combination of large-scale antennas with high transmission frequencies often results in the communicating devices operating in the near-field (Fresnel) region. In this paper, we study the potential of beam focusing, feasible in near-field operation, in facilitating high-rate multi-user downlink multiple-input multiple-output (MIMO) systems. As the ability to achieve beam focusing is dictated by the transmit antenna, we study near-field signalling considering different antenna structures, including fully-digital architectures, hybrid phase shifter-based precoders, and the emerging dynamic metasurface antenna (DMA) architecture for massive MIMO arrays. We first provide a mathematical model to characterize near-field wireless channels as well as the transmission pattern for the considered antenna architectures. Then, we formulate the beam focusing problem for the goal of maximizing the achievable sum-rate in multi-user networks. We propose efficient solutions based on the sum-rate maximization task for fully-digital, (phase shifters based-) hybrid and DMA architectures. Simulation results show the feasibility of the proposed beam focusing scheme for both single- and multi-user scenarios. In particular, the designed focused beams provide a new degree of freedom to mitigate interference in both angle and distance domains, which is not achievable using conventional far-field beam steering, allowing reliable communications for uses even residing at the same angular direction.

One of the main challenges in containing the Coronoavirus disease 2019 (COVID-19) pandemic stems from the difficulty in carrying out efficient mass diagnosis over large populations. The leading method to test for COVID-19 infection utilizes qualitative polymerase chain reaction, implemented using dedicated machinery which can simultaneously process a limited amount of samples. A candidate method to increase the test throughput is to examine pooled samples comprised of a mixture of samples from different patients. In this work we study pooling-based COVID-19 tests. We identify the specific requirements of COVID-19 testing, including the need to characterize the infection level and to operate in a one-shot fashion, which limit the application of traditional group-testing (GT) methods. We then propose a multi-level GT scheme, designed specifically to meet the unique requirements of COVID-19 tests, while exploiting the strength of GT theory to enable accurate recovery using much fewer tests than patients. Our numerical results demonstrate that multi-level GT reliably and efficiently detects the infection levels, while achieving improved accuracy over previously proposed one-shot COVID-19 pooled-testing methods.

Graph signals arise in various applications, ranging from sensor networks to social media data. The high-dimensional nature of these signals implies that they often need to be compressed in order to be stored and conveyed. The common framework for graph signal compression is based on sampling, resulting in a set of continuous-amplitude samples, which in turn have to be quantized into a finite bit representation. In this work we study the joint design of graph signal sampling along with the quantization of these samples, for graph signal compression. We focus on bandlimited graph signals, and show that the compression problem can be represented as a task-based quantization setup, in which the task is to recover the spectrum of the signal. Based on this equivalence, we propose a joint design of the sampling and recovery mechanisms for a fixed quantization mapping, and present an iterative algorithm for dividing the available bit budget among the discretized samples. Our numerical evaluations demonstrate that the proposed scheme achieves reconstruction accuracy within a small gap of that achievable with infinite resolution quantizers, while compressing high-dimensional graph signals into finite bit streams.

Dual function radar and communications (DFRC) systems are the focus of growing research attention. The common DFRC setup considers simultaneous probing and information transmission to a remote receiver, typically involving complex radar-oriented waveforms, whose detection can induce a notable burden on the receiver. In many DFRC applications, the communication receivers are devices which are limited in terms of hardware, power, and memory resources. These receivers are required to extract the desired information from the received dual-function waveform, while operating with a given bit budget. In this paper, we design bit constrained communication receivers in dual-function systems, by considering hybrid analog/digital architectures and treating their operation as task-based quantization. We study two forms of analog processing in these hybrid receivers, allowing to combine inputs in different time instances and antennas or only in different antennas at the same time instance. Simulation results demonstrate that the proposed task-based quantization strategy outperforms receivers operating only in the digital domain with the same total number of quantization bits.

The Kalman filter (KF) is a celebrated signal processing algorithm, implementing optimal state estimation of dynamical systems that are well represented by a linear Gaussian state-space model. The KF is model-based, and therefore relies on full and accurate knowledge of the underlying model. We present KalmanNet, a hybrid data-driven/model-based filter that does not require full knowledge of the underlying model parameters. KalmanNet is inspired by the classical KF flow and implemented by integrating a dedicated and compact neural network for the Kalman gain computation. We present an offline training method, and numerically illustrate that KalmanNet can achieve optimal performance without full knowledge of the model parameters. We demonstrate that when facing inaccurate parameters KalmanNet learns to achieve notably improved performance compared to KF.

The success of deep neural networks (DNNs) as an enabler of artificial intelligence (AI) is heavily dependent on high computational resources. The increasing demands for accessible and personalized AI give rise to the need to operate DNNs on edge devices such as smartphones, sensors, and autonomous cars, whose computational powers are limited. Here we propose a framework for facilitating the application of DNNs on the edge in a manner which allows multiple users to collaborate during inference in order to improve their prediction accuracy. Our mechanism, referred to as edge ensembles, is based on having diverse predictors at each device, which can form a deep ensemble during inference. We analyze the latency induced in this collaborative inference approach, showing that the ability to improve performance via collaboration comes at the cost of a minor additional delay. Our experimental results demonstrate that collaborative inference via edge ensembles equipped with compact DNNs substantially improves the accuracy over having each user infer locally, and can outperform using a single centralized DNN larger than all the networks in the ensemble together.

Multiple-input multiple-output (MIMO) radar is known to achieve high performance by probing with multiple orthogonal wave-forms. However, implementing a low cost low power MIMO radar is challenging. In this work we study reduced cost MIMO radar receivers restricted to operate with low resolution ADCs. A hybrid analog-digital architecture, referred to as bit-limited MIMO radar (BiLiMO) receivers, which are capable of accurately recovering their targets while operating under strict resolution constraints is designed. This is achieved by applying an additional analog filter to the acquired waveforms, and designing the overall hybrid analog-digital system to facilitate target identification using task-based quantization methods. Our numerical results demonstrate that the proposed BiLiMO receiver operating with strict bit budget achieves target recovery performance which approaches that of costly MIMO radars operating with unlimited resolution ADCs.

Recently, dynamic metasurface antennas (DMAs) have emerged as a promising technology for realizing massive multiple-input multiple-output (MIMO) wireless systems. The usage of large arrays, jointly with higher transmitted frequencies, often results in the communicating devices operating in the near-field (Fresnel) region, thus requiring different considerations compared to traditional systems, assumed to operate in the far-field regime. In this paper, we study the potential of beam focusing, feasible in near-field operation, for multi-user MIMO systems, where the base station is equipped with a DMA. We introduce a mathematical model for DMA-based near-field MIMO communications. Then, we characterize the sum-rate maximization problem of the considered system, and propose an efficient solution to jointly design the DMA weights and digital precoding vector. Simulation results show that our design generates focused beams such that users residing at the same angular direction can communicate reliably without interfering, which is not achievable using conventional far-field beam steering.

We consider the problem of convergence time minimization for federated learning (FL) implemented in wireless systems. In such setups, each wireless edge device transmits its local FL model parameters to a base station (BS). The BS then uses the received FL parameters to generate a common FL model and broadcasts it to all edge devices. Since the FL parameters must be transmitted over wireless links, the convergence time depends not only on the number of training steps, but also on the FL parameter transmission delay at each training step, which can be substantial when conveying a large number of parameters. In addition, due to limited wireless resources such as spectrum, only a subset of edge devices can participate in each FL training step, which can further increase convergence time. Our goal therefore is to optimize wireless resource management and user selection for FL, as well as limit the volume of transmitted FL parameters. In this paper, three schemes for facilitating communication efficient FL are introduced: First, a probabilistic device selection scheme is designed such that the devices that can significantly improve the convergence speed and training loss have high probabilities for FL parameter transmission. Then, given the subset of participating devices, an efficient wireless resource allocation scheme is developed. Finally, a quantization method is proposed to reduce the data size. Simulation results demonstrate that the proposed FL method can improve handwritten digit identification accuracy and convergence delay by up to 3% and 90% compared to the conventional FL.

The need to recover high-dimensional signals from their noisy low-resolution quantized measurements is widely encountered in communications and sensing. In this paper, we focus on the extreme case of one-bit quantizers, and propose a deep detector network, called LoRD-Net, for signal recovering from one-bit measurements. Our approach relies on a model-aware data-driven architecture, based on a deep unfolding of first-order optimization iterations. LoRD-Net has a task-based architecture dedicated to recovering the underlying signal of interest from the one-bit noisy measurements without requiring prior knowledge of the channel matrix through which the one-bit measurements are obtained. The proposed deep detector has much fewer parameters compared to black-box deep networks due to the incorporation of domain-knowledge in the design of its architecture, allowing it to operate in a data-driven fashion while benefiting from the flexibility, versatility, and reliability of model-based optimization methods. We numerically evaluate the proposed receiver architecture for one-bit signal recovery in wireless communications and demonstrate that the proposed hybrid methodology outperforms both data-driven and model-based state-of-the-art methods, while utilizing small datasets, on the order of merely ~ 500 samples, for training.

The combination of orthogonal frequency modulation (OFDM) and multiple-input multiple-output (MIMO) techniques plays an important role in modern communication systems. In order to meet the growing throughput demands, future MIMO-OFDM receivers are expected to utilize a massive number of antennas, operate in dynamic environments, and explore high frequency bands, while satisfying strict constraints in terms of cost, power, and size. An emerging technology to realize massive MIMO receivers of reduced cost and power consumption is based on dynamic metasurface antennas (DMAs), which inherently implement controllable compression in acquisition. In this work we study the application of DMAs for MIMO-OFDM receivers operating with bit-constrained analog-to-digital converters (ADCs). We present a model for DMAs which accounts for the configurable frequency selective profile of its metamaterial elements, resulting in a spectrally flexible hybrid structure. We then exploit previous results in task-based quantization to show characterized the achievable OFDM recovery accuracy for a given DMA configuration in the presence of bit-constrained ADCs, and propose methods for adjusting the DMA parameters based on channel state information. Our numerical results demonstrate that by properly exploiting the spectral diversity of DMAs, notable performance gains are obtained over existing designs of conventional hybrid architectures, demonstrating the potential of DMAs for realizing high performance massive antenna arrays of reduced cost and power consumption.

Digital receivers are required to recover the transmitted symbols from their observed channel output. In multiuser multiple-input multiple-output (MIMO) setups, where multiple symbols are simultaneously transmitted, accurate symbol detection is challenging. A family of algorithms capable of reliably recovering multiple symbols is based on interference cancellation. However, these methods assume that the channel is linear, a model which does not reflect many relevant channels, as well as require accurate channel state information (CSI), which may not be available. In this work we propose a multiuser MIMO receiver which learns to jointly detect in a data-driven fashion, without assuming a specific channel model or requiring CSI. In particular, we propose a data-driven implementation of the iterative soft interference cancellation (SIC) algorithm which we refer to as DeepSIC. The resulting symbol detector is based on integrating dedicated machine-learning methods into the iterative SIC algorithm. DeepSIC learns to carry out joint detection from a limited set of training samples without requiring the channel to be linear and its parameters to be known. Our numerical evaluations demonstrate that for linear channels with full CSI, DeepSIC approaches the performance of iterative SIC, which is comparable to the optimal performance, and outperforms previously proposed learning-based MIMO receivers. Furthermore, in the presence of CSI uncertainty, DeepSIC significantly outperforms model-based approaches. Finally, we show that DeepSIC accurately detects symbols in non-linear channels, where conventional iterative SIC fails even when accurate CSI is available.

Federated learning (FL) is a framework for distributed learning of centralized models. In FL, a set of edge devices train a model using their local data, while repeatedly exchanging their trained model with a central server, allowing to tune a global model without having the users share their possibly private data. A major challenge in FL is to reduce the bandwidth and energy consumption due to the repeated transmissions of large volumes of data by a large number of users over the wireless channel. Recently, over-the-air (OTA) FL has been suggested to achieve this goal. In this setting, all users transmit their data signal simultaneously over a Multiple Access Channel (MAC), and the computation is done over the wireless channel. In this paper, we develop a novel convergent OTA FL (COTAF) algorithm, which induces precoding and scaling upon transmissions to gradually mitigate the effect of the noisy channel, thus facilitating FL convergence. We analyze the convergence of COTAF to the loss minimizing model theoretically, showing its ability to achieve a convergence rate similar to that achievable over error-free channels. Our simulations demonstrate the improved convergence of COTAF for training using non-synthetic datasets.

Analog signals processed in digital hardware are quantized into a discrete bit-constrained representation. Quantization is typically carried out using analog-to-digital converters (ADCs), operating in a serial scalar manner. In some applications, a set of analog signals are acquired individually and processed jointly. Such setups are referred to as distributed quantization. In this work we propose a distributed quantization scheme for representing a set of sparse time sequences acquired using conventional scalar ADCs. Our approach utilizes tools from secure group testing theory to exploit the sparse nature of the acquired analog signals, obtaining a compact and accurate representation while operating in a distributed fashion. We then show how our technique can be implemented when the quantized signals are transmitted over a multihop communication network providing a low-complexity network policy for routing and signal recovery. Our numerical evaluations demonstrate that the proposed scheme notably outperforms conventional methods based on the combination of quantization and compressed sensing tools.

Dual-function radar communication (DFRC) systems implement both sensing and communication using the same hardware. An emerging DFRC strategy embeds transmission of digital messages into agility-based radar schemes in the form of index modulation (IM). This approach provides the ability to communicate without entailing degradation in radar performance, at the cost of increased decoding complexity at the receiver side. In this work we propose schemes for reducing the decoding complexity associated with IM-based DFRC systems. We first focus on the receiver side, developing a sub-optimal low complexity scheme for recovering IM symbols embedded in radar waveforms. Then, we propose a method to modify the radar waveform to facilitate the recovery of the communicated bits with minimal effect on the radar performance. Our numerical results demonstrate that the proposed techniques allow the receiver to reliably recover the transmitted symbols with an affordable computational burden.

Modern radar systems are expected to operate reliably in congested environments under cost and power constraints. A recent technology for realizing such systems is frequency agile radar (FAR), which transmits narrowband pulses in a frequency hopping manner. To enhance the target recovery performance of FAR in complex electromagnetic environments, and particularly, its range-Doppler recovery performance, multi-Carrier AgilE phaSed Array Radar (CAESAR) was proposed. CAESAR extends FAR to multi-carrier waveforms while introducing the notion of spatial agility. In this paper, we theoretically analyze the range-Doppler recovery capabilities of CAESAR. Particularly, we derive conditions which guarantee accurate reconstruction of these range-Doppler parameters. These conditions indicate that by increasing the number of frequencies transmitted in each pulse, CAESAR improves performance over conventional FAR, especially in complex environments where some radar measurements are severely corrupted by interference.

Traditional deep learning models are trained on centralized servers using labeled sample data collected from edge devices. This data often includes private information, which the users may not be willing to share. Federated learning (FL) is an emerging approach to train such learning models without requiring the users to share their possibly private labeled data. In FL, each user trains its copy of the learning model locally. The server then collects the individual updates and aggregates them into a global model. A major challenge that arises in this method is the need of each user to efficiently transmit its learned model over the throughput limited uplink channel. In this work, we tackle this challenge using tools from quantization theory. In particular, we identify the unique characteristics associated with conveying trained models over rate-constrained channels, and characterize a suitable quantization scheme for such setups. We show that combining universal vector quantization methods with FL yields a decentralized training system, which is both efficient and feasible. We also derive theoretical performance guarantees of the system. Our numerical results illustrate the substantial performance gains of our scheme over FL with previously proposed quantization approaches.

Accurate symbol detection in multiuser multiple-input multiple-output (MIMO) setups, where multiple symbols are simultaneously transmitted, is a challenging task. A family of algorithms capable of reliably recovering multiple symbols is based on interference cancellation. However, these methods assume that the channel is linear, a model which does not reflect many relevant channels, as well as require accurate channel state information (CSI), which may not be available. In this work we propose a multiuser MIMO receiver which learns to jointly detect in a data-driven fashion, without assuming a specific channel model or requiring CSI. In particular, we propose a data-driven implementation of the iterative soft interference cancellation (SIC) algorithm. The resulting detector, referred to as DeepSIC, is based on integrating dedicated machine-learning (ML) methods into the iterative SIC scheme, and learns to carry out joint detection from a limited set of training samples without requiring the channel to be linear and its parameters to be known. Our numerical evaluations demonstrate that for linear channels with full CSI, DeepSIC approaches the performance of iterative SIC, which is comparable to the optimal performance, while being notably more robust to CSI uncertainty. Finally, we show that DeepSIC accurately detects symbols in non-linear channels, where conventional iterative SIC fails even when accurate CSI is available.

Analog-to-digital conversion allows physical signals to be processed using digital hardware. This conversion consists of two stages: Sampling, which maps a continuous-time signal into discrete-time, and quantization, i.e., representing the continuous-amplitude quantities using a finite number of bits. This conversion is typically carried out using generic uniform mappings that are ignorant of the task for which the signal is acquired, and can be costly when operating in high rates and fine resolutions. In this work we design task-oriented analog-to-digital converters (ADCs) which operate in a data-driven manner, namely they learn how to map an analog signal into a sampled digital representation such that the system task can be efficiently carried out. We propose a model for sampling and quantization which both faithfully represents these operations while allowing the system to learn non-uniform mappings from training data. We focus on the task of symbol detection in multiple-input multiple-output (MIMO) digital receivers, where multiple analog signals are simultaneously acquired in order to recover a set of discrete information symbols. Our numerical results demonstrate that the proposed approach achieves performance which is comparable to operating without quantization constraints, while achieving more accurate digital representation compared to utilizing conventional uniform ADCs.

The combination of orthogonal frequency modulation (OFDM) and multiple-input multiple-output (MIMO) systems plays an important role in modern communication systems. In order to meet the growing throughput demands, future MIMO-OFDM receivers are expected to utilize a massive number of antennas, operate in dynamic environments, and explore high frequency bands, while satisfying strict constraints in terms of cost, power, and size. An emerging technology to realize massive MIMO receivers of reduced cost and power consumption is based on dynamic metasurface antennas (DMAs), which inherently implement controllable compression in acquisition. In this work we study the application of DMAs for MIMO-OFDM receivers operating with bit-constrained analog-to-digital converters (ADCs). We exploit previous results in task-based quantization to show how DMAs can be configured to improve recovery in the presence of constrained ADCs, and propose an algorithm for adjusting the DMA parameters based on channel state information. Our numerical results demonstrate that the DMA-based receiver is capable of accurately recovering OFDM signals, and that its performance is comparable to receivers operating without bit limitations, while being significantly less costly and more power efficient.

Many important schemes in signal processing and communications, ranging from the BCJR algorithm to the Kalman filter, are instances of factor graph methods. This family of algorithms is based on recursive message passing-based computations carried out over graphical models, representing a factorization of the underlying statistics. In order to implement these algorithms, one must have accurate knowledge of the statistical model of the underlying signals. In this work we implement factor graph methods in a data-driven manner when the statistics are unknown. In particular, we propose using machine learning (ML) tools to learn the factor graph, instead of the overall system task, which in turn is used for inference by message passing over the learned graph. We apply the proposed approach to learn the factor graph representing a finite-memory channel, demonstrating the resulting ability to implement BCJR detection in a data-driven fashion. We demonstrate that the proposed system, referred to as BCJRNet, learns to implement the BCJR algorithm from a small training set, and that the resulting receiver exhibits improved robustness to inaccurate training compared to the conventional channel-model-based receiver operating under the same level of uncertainty. Our results indicate that by utilizing ML tools to learn factor graphs from labeled data, one can implement a broad range of model-based algorithms, which traditionally require full knowledge of the underlying statistics, in a data-driven fashion.

Federated learning (FL) refers to the adaptation of a central model based on data sets available at multiple remote users. Two of the common challenges encountered in FL are the fact that training sets obtained by different users are commonly heterogeneous, i.e., arise from different sample distributions, and the need to communicate large amounts of data between the users and the central server over the typically expensive up-link channel. In this work we formulate the problem of FL in which different clusters of users observe labeled samples drawn from different distributions, while operating under constraints on the communication overhead. For such settings, we identify that the combination of statistical heterogeneity and communication constraints induces a tradeoff between the ability of the users of each cluster to learn a proper model and the accuracy in aggregating these models into a global inference rule. We propose an algorithm based on multi-source adaptation methods for such communication-aware clustered FL scenarios which allows to balance these performance measures, and demonstrate its ability to achieve improved inference over conventional federated averaging without inducing additional communication overhead.

Non-orthogonal communications play an important role in future digital communication architectures. In such scenarios, the received signal is corrupted by an interfering communications signal, which is much stronger than the thermal noise, and is often modeled as a cyclostationary process in continuous-time. To facilitate digital processing, the receiver typically samples the received signal synchronously with the symbol rate of the information signal. If the period of the statistics of the interference is synchronized with that of the information signal, then the sampled interference is modeled as a discrete-time (DT) cyclostationary random process. However, in the common interference scenario, the period of the statistics of the interference is not necessarily synchronized with that of the information signal. In such cases, the DT interference may be modeled as an almost cyclostationary random process. In this work we characterize the capacity of DT memoryless additive noise channels in which the noise arises from a sampled cyclostationary Gaussian process. For the case of synchronous sampling, capacity can be obtained in closed form. When sampling is not synchronized with the symbol rate of the interference, the resulting channel is not information stable, thus classic information-theoretic tools are not applicable. Using information spectrum methods, we prove that capacity can be obtained as the limit of a sequence of capacities of channels with additive cyclostationary Gaussian noise. Our results allow to characterize the effects of changes in the sampling rate and sampling time offset on the capacity of the resulting DT channel. In particular, it is demonstrated that minor variations in the sampling period, such that the resulting noise switches from being synchronously-sampled to being asynchronously-sampled, can substantially change the capacity.

Symbol detection plays an important role in the implementation of digital receivers. In this work, we propose ViterbiNet, which is a data-driven symbol detector that does not require channel state information (CSI). ViterbiNet is obtained by integrating deep neural networks (DNNs) into the Viterbi algorithm. We identify the specific parts of the Viterbi algorithm that depend on the channel model, and design a DNN to implement only those computations, leaving the rest of the algorithm structure intact. We then propose a meta-learning based approach to train ViterbiNet online based on recent decisions, allowing the receiver to track dynamic channel conditions without requiring new training samples for every coherence block. Our numerical evaluations demonstrate that the performance of ViterbiNet, which is ignorant of the CSI, approaches that of the CSI-based Viterbi algorithm, and is capable of tracking time-varying channels without needing instantaneous CSI or additional training data. Moreover, unlike conventional Viterbi detection, ViterbiNet is robust to CSI uncertainty, and it can be reliably implemented in complex channel models with constrained computational burden. More broadly, our results demonstrate the conceptual benefit of designing communication systems that integrate DNNs into established algorithms.

This paper proposes to use a deep neural network (DNN)- based symbol detector for mmWave systems such that channel state information (CSI) acquisition can be bypassed. In particular, we consider a sliding bidirectional recurrent neural network (BRNN) architecture that is suitable for the long memory length of typical mmWave channels. The performance of the DNN detector is evaluated in comparison to that of the Viterbi detector. The results show that the performance of the DNN detector is close to that of the optimal Viterbi detector with perfect CSI, and that it outperforms the Viterbi algorithm with CSI estimation error. Further experiments show that the DNN detector is robust to a wide range of noise levels and varying channel conditions, and that a pretrained detector can be reliably applied to different mmWave channel realizations with minimal overhead.

Massive multiple-input multiple-output (MIMO) communications are the focus of considerable interest in recent years. While theoretical gains of such massive MIMO have been established, implementing MIMO systems with large-scale antenna arrays in practice is challenging. Among the practical difficulties associated with massive MIMO implementations are increased cost, power consumption, and physical size. In this work we study the implementation of massive MIMO antenna arrays using dynamic metasurface antennas (DMAs), an emerging technology which inherently handles the aforementioned challenges. DMAs realize planar large-scale arrays of tunable antenna elements, and can adaptively incorporate compression and analog combining in the physical antenna structure, thus reducing cost and power consumption. We first propose a mathematical model for massive MIMO systems with DMAs and discuss their constraints compared to ideal antenna arrays. Then, we characterize the fundamental limits of the resulting systems, and propose an algorithm for designing practical DMAs to approach these limits. Our numerical results indicate that the performance of practical DMA-based massive MIMO systems is comparable with ideal antenna arrays.

Massive multiple-input multiple-output (MIMO) systems have been drawing considerable interest. In the uplink, massive MIMO systems are commonly studied assuming that each base station (BS) decodes the signals of its user terminals separately and linearly while treating all interference as noise. Although this approach provides improved spectral efficiency (SE) in favorable channel conditions, it is generally sub-optimal from an information-theoretic perspective. In this work we characterize the SE of massive MIMO when the BSs are allowed to jointly decode the received signals. We consider two schemes for handling the interference, and derive their SEs for both finite and asymptotic number of antennas. Simulation tests of the proposed methods illustrate their gains in SE compared to standard separate linear decoding, and show that the standard approach fails to capture the actual achievable rates of massive MIMO systems, particularly when the interference is dominant.

Massive multiple-input multiple-output (MIMO) systems are the focus of increasing research attention. In such setups, there is an urgent need to utilize simple low-resolution quantizers, due to power and memory constraints. In this work we study massive MIMO channel estimation with quantized measurements, when the quantization system is designed to minimize the channel estimation error, as opposed to the quantization distortion. We first consider vector quantization, and characterize the minimal error achievable. Next, we focus on practical systems utilizing scalar uniform quantizers, and design the analog and digital processing as well as the quantization dynamic range to optimize the channel estimation accuracy. Our results demonstrate that the resulting massive MIMO system which utilizes low-resolution scalar quantizers can approach the minimal estimation error dictated by rate-distortion theory, achievable using vector quantizers.

Quantizers play a critical role in digital signal processing systems. In practice, quantizers are typically implemented using scalar analog-to-digital converters (ADCs), commonly utilizing a fixed uniform quantization rule which is ignorant of the task of the system. Recent works have shown that the performance of quantization systems utilizing scalar ADCs can be significantly improved by properly processing the analog signal prior to quantization. However, the implementation of such systems requires complete knowledge of the underlying model, which may not be available in practice. In this work we design task-oriented quantization systems with scalar ADCs using deep learning, focusing on the task of multiple-input multiple-output (MIMO) channel estimation. By utilizing deep learning, we construct a task-based quantization system, overcoming the need to explicitly recover the system model and to find the proper quantization rule for it. Our results indicate that the proposed method results in practical MIMO systems with scalar ADCs which are capable of approaching the optimal performance limits dictated by indirect rate-distortion theory, achievable using vector quantizers and requiring complete knowledge of the underlying statistical model.

Quantization allows physical signals to be processed using digital devices. Quantizers are commonly implemented using analog-to-digital converters (ADCs), which operate in a serial and scalar manner and are designed to yield an accurate digital representation of the observed signal. However, in many practical scenarios quantization is part of a system whose task is not to recover the observed signal, but some function of it. Recent works have shown that properly designed task-based quantizers, which include pre-quantization analog combining as well as digital processing, can achieve notable gains in recovering linear functions of the observations. In this work we focus on quantization for the task of recovering quadratic functions. Our analysis is based on principal inertia components (PICs), which form a basis for decomposing the statistical dependence between random quantities. Using PICs, we identify a practical structure of the pre-quantization mapping for recovering quadratic functions, which allows us to design a task-based quantization system capable of accurately estimating these functions. Our numerical study demonstrates that, when using scalar ADCs, notable performance gains that can be achieved using the proposed design over intuitive approaches such as quantizing the quadratic function directly as well as task-ignorant quantization.

Interference-limited communications plays an important role in future digital communication architectures. In such scenarios, the received signal is corrupted by an interfering communications signal, which is typically modeled as a cyclostationary process in continuous-time. To facilitate digital processing, the receiver typically samples the received signal synchronously with the symbol rate of the information signal. The sampled received signal thus contains an interference component which is either cyclostationary or almost cyclostationary in discrete-time (DT), depending on whether the symbol rate of the interference is synchronized with the sampling rate, or it is not. In this work we characterize the capacity of DT interference-limited communications channels, in which the interference is modeled as an additive sampled cyclostationary Gaussian noise. For the case of synchronous sampling, capacity can be obtained in closed form as a direct application of our previous work. When sampling is asynchronous, the resulting channel is not information stable, thus classic information-theoretic tools are not applicable. Using information spectrum methods, we prove that capacity can be obtained as the limit of a sequence of capacities of DT channels with additive cyclostationary noise. Our results facilitate the characterization of the impact of variations in the sampling rate and sampling time offset on the capacity of the resulting DT channel. In particular, it is demonstrated that minor variations in the sampling period can have a notable effect on capacity.

Sampling enables physical signals to be processed using digital hardware. When multiple signals are sampled, the spatial correlation between them may be utilized to reduce the overall reconstruction error. In this work we study joint sampling and reconstruction of multiple correlated stochastic sources, exploiting their correlation to improve recovery. We derive the achievable reconstruction error and the corresponding sampling system for arbitrary sampling rates and spectral structures. The proposed system minimizes the error when sampling below the Nyquist rate by preserving only the most dominant spatial eigenmodes aliased to each frequency. Using this characterization, we obtain sufficient conditions for error free recovery. We also discuss a distributed sampling setting, where each signal is acquired separately, while reconstruction is performed jointly. We characterize conditions under which distributed sampling performs as well as joint sampling. Our numerical results illustrate that joint sampling can achieve negligible reconstruction error using low sampling rates when the signals exhibit notable spatial correlation, and demonstrate that properly exploiting this correlation can dramatically improve reconstruction accuracy.

Massive multiple-input-multiple-output (MIMO) communications are the focus of considerable interest in recent years. While the theoretical gains of massive MIMO have been established, implementing MIMO systems with large-scale antenna arrays in practice is challenging. Among the practical challenges associated with massive MIMO systems are increased cost, power consumption, and physical size. In this paper, we study the implementation of massive MIMO antenna arrays using dynamic metasurface antennas (DMAs), an emerging technology which inherently handles the aforementioned challenges. Specifically, DMAs realize large-scale planar antenna arrays and can adaptively incorporate signal processing methods such as compression and analog combining in the physical antenna structure, thus reducing the cost and power consumption. First, we propose a mathematical model for massive MIMO systems with DMAs and discuss their constraints compared to ideal antenna arrays. Then, we characterize the fundamental limits of uplink communications with the resulting systems and propose two algorithms for designing practical DMAs for approaching these limits. Our numerical results indicate that the proposed approaches result in practical massive MIMO systems whose performance is comparable to that achievable with ideal antenna arrays.

Massive multiple-input multiple-output (MIMO) systems have been drawing considerable interest due to the growing throughput demands on wireless networks. In the uplink, massive MIMO systems are commonly studied assuming that each base station (BS) decodes the signals of its user terminals separately and linearly while treating all interference as noise. Although this approach provides improved spectral efficiency which scales with the number of BS antennas in favorable channel conditions, it is generally sub-optimal from an information-theoretic perspective. In this paper, we characterize the spectral efficiency of massive MIMO when the BSs are allowed to jointly decode the received signals. In particular, we consider four schemes for treating the interference, and derive the achievable average ergodic rates for both finite and asymptotic number of antennas for each scheme. Simulation tests of the proposed methods illustrate their gains in spectral efficiency compared with the standard approach of separate linear decoding, and show that the standard approach fails to capture the actual achievable rates of massive MIMO systems, particularly when the interference is dominant.

Communications over power lines in the frequency range above 2 MHz, commonly referred to as broadband (BB) power line communications (PLC), is a central communications scenario for smart power grids. BB-PLC channels are characterized by a dominant colored non-Gaussian additive noise, as well as by periodic variations of the channel impulse response and the noise statistics, induced by the mains voltage. In this work we study the fundamental rate limits for multiple input-multiple output (MIMO) BB-PLC channels, modeled as periodic channels with additive non-Gaussian noise and finite memory. We present bounds on the capacity of these channels by exploiting a bijection with time-invariant MIMO channels of extended dimensions. We illustrate the resulting fundamental limits in a numerical analysis corresponding to practical MIMO BB-PLC channels.

In many communications scenarios, the channel exhibits periodic characteristics, e.g., power line communications and interference-limited communications. Additionally, certain approximations for mobile radio channels over finite time intervals may result in periodic channel models. In this paper, we study pilot-aided joint estimation of the channel impulse response (CIR) and of the carrier frequency offset (CFO) for linear periodic channels, in which both the CIR and the noise statistics vary periodically in time. We first consider the joint maximum likelihood estimator (JMLE) for the CIR and the CFO, and discuss the practical drawbacks associated with this estimator. When the coefficients of the delay-Doppler spread function of the CIR are approximately sparse, we propose two estimation schemes with higher spectral efficiency and lower computational complexity compared with the JMLE, which are obtained by exploiting both the periodicity and the sparsity of the channel, without requiring a priori knowledge of the sparsity pattern. Finally, we study the design of pilot sequences aimed at improving the estimation performance in sparse periodic channels. Simulation studies corresponding to practical scenarios of the proposed estimators demonstrate that substantial benefits can be obtained by properly accounting for the sparsity and periodicity in the design of estimation schemes.

In[1], the first equation inTheorem 2on page 1195 contains an error, where$\rho $should be replaced with$N_{0} \cdot \rho $.

Communications over power lines in the frequency range above 2 MHz, commonly referred to as broadband (BB) power line communications (PLC), has been the focus of increasing research attention and standardization efforts in recent years. BB-PLC channels are characterized by a dominant colored non-Gaussian additive noise, as well as by periodic variations of the channel impulse response and of the noise statistics. In this paper, we study the fundamental rate limits for BB-PLC channels by bounding their capacity while accounting for the unique properties of these channels. We obtain explicit expressions for the derived bounds for several BB-PLC noise models, and illustrate the resulting fundamental limits in a numerical analysis.

In the phase retrieval problem, the observations consist of the magnitude of a linear transformation of the signal of interest (SOI) with additive noise, where the linear transformation is typically referred to as measurement matrix. The objective is then to reconstruct the SOI from the observations up to an inherent phase ambiguity. Many works on phase retrieval assume that the measurement matrix is a random Gaussian matrix, which in the noiseless scenario with sufficiently many measurements guarantees uniqueness of the mapping between the SOI and the observations. However, in many applications, e.g., optical imaging, the measurement matrix corresponds to the underlying physical setup, and is therefore a deterministic matrix with structure constraints. In this work we study the design of deterministic measurement matrices, aimed at maximizing the mutual information between the SOI and the observations. We characterize necessary conditions for the optimal measurement matrix, and propose a practical design method for measurement matrices corresponding to masked Fourier measurements. Simulation tests of the proposed method show that it achieves the same performance as random Gaussian matrices for various phase recovery algorithms.

Adaptive filters are commonly used in many signal processing and communications systems. In many practical digital communications scenarios, including, for example, interference-limited wireless and wireline communications, as well as narrowband power line communications, the considered signals are jointly cyclostationary. Yet, most works on adaptive filtering of cyclostationary signals used ad hoc application of adaptive algorithms designed for stationary signals, e.g., the least-mean-squares (LMS). It is known that these algorithms may not converge for jointly cyclostationary signals. In this paper, we rigorously study the optimal adaptive filtering of jointly cyclostationary signals. We first identify the relevant objective as the time-averaged mean-squared error criterion (TA-MSE), and obtain an adaptive algorithm as the stochastic approximation of the TA-MSE minimizer. When the considered signals are jointly stationary, the algorithm specializes to the standard LMS algorithm. We provide a comprehensive transient and steady-state performance analysis without imposing a specific distribution on the considered signals, and derive conditions for convergence and stability. The algorithm, which we call time-averaged LMS, is applied to practical scenarios in a simulations study, and an excellent agreement between the theoretical and the empirical performance is observed.

In this paper, we study the secrecy capacity of Gaussian multiple-input multiple-output (MIMO) wiretap channels (WTCs) with a finite memory, subject to a per-symbol average power constraint on the MIMO channel input. MIMO channels with finite memory are very common in wireless communications as well as in wireline communications (e.g., in communications over power lines). To derive the secrecy capacity of the Gaussian MIMO WTC with finite memory, we first construct an asymptotically equivalent block-memoryless MIMO WTC, which is then transformed into a set of parallel, independent, memoryless MIMO WTCs in the frequency domain. The secrecy capacity of the Gaussian MIMO WTC with finite memory is obtained as the secrecy capacity of the set of parallel, independent, memoryless MIMO WTCs, and is expressed as maximization over the input covariance matrices in the frequency domain. Finally, we detail two applications of our result: First, we show that the secrecy capacity of the Gaussian scalar WTC with finite memory can be achieved by waterfilling, and obtain a closed-form expression for this secrecy capacity. Then, we use our result to characterize the secrecy capacity of narrowband powerline channels, thereby resolving one of the major open issues for this channel model.

In many communications scenarios the channel exhibits periodic characteristics. Periodicity may be expressed as a periodically time-varying channel transfer function as well as an additive noise with periodically time-varying statistics. Examples for such scenarios include interference-limited communications, both wireless and wireline, and also power line communications (PLC). In this work, we characterize the capacity of discrete-time, finite-memory Gaussian multiple-input multiple-output (MIMO) channels with periodic characteristics. The derivation transforms the periodic MIMO channel into an extended time-invariant MIMO channel, for which we obtain a closed-form capacity expression. It is shown that capacity can be achieved by an appropriate waterfilling scheme. The capacity expression obtained is numerically evaluated for practical PLC scenarios, and compared to the achievable rate of an ad-hoc orthogonal frequency division multiplexing based scheme, and the gains from optimally handling the periodicity of the channel are quantified.

Narrowband power line communications (NB-PLC) is the central communications technology for the realization of smart power grids. For this reason, NB-PLC channels have been receiving substantial attention in recent years. These channels are characterized by periodic short-term variations of the channel transfer function (CTF) and strong noise with periodic statistics. In this work, we derive the capacity of discrete-time NB-PLC channels, accounting for the periodic properties of both the CTF and the noise. As part of the capacity derivation, we characterize the capacity achieving transmission scheme, which leads to guidelines for constructing a practical code that approaches the capacity as the blocklength increases. The capacity derived in this work is numerically evaluated and the results show that the optimal scheme achieves a substantial rate gain over previously proposed ad-hoc scheme. This gain is due to optimally accounting for the periodic properties of the channel and the noise.

Privacy is a critical issue when communicating over shared mediums. A fundamental model for the information-theoretic analysis of secure communications is the wiretap channel (WTC), which consists of a communicating pair and an eavesdropper. In this work we study the secrecy capacity of Gaussian multiple-input multiple-output (MIMO) WTCs with finite memory. These channels are very common in wireless communications as well as in wireline communications (e.g., in power line communications). We derive a closed-form expression for the secrecy capacity of the MIMO Gaussian WTC with finite memory via the analysis of an equivalent block-memoryless model, which is transformed into a set of parallel independent memoryless MIMO WTCs. The secrecy capacity is expressed as the maximization over the input covariance matrices in the frequency domain. Finally, we show that for the Gaussian scalar WTC with finite memory, the secrecy capacity can be obtained by waterfilling.

Power line communications (PLC) is the central communications technology for the realization of smart power grids. As the designated band for smart grid communications is the narrowband (NB) power line channel, NB-PLC has been receiving substantial attention in recent years. Narrowband power line channels are characterized by cyclic short-term variations of the channel transfer function (CTF) and strong noise with periodic statistics. In this paper, modeling the CTF as a linear periodically time-varying filter and the noise as an additive cyclostationary Gaussian process, we derive the capacity of discrete-time NB-PLC channels. As part of the capacity derivation, we characterize the capacity achieving transmission scheme, which leads to a practical code construction that approaches capacity. The capacity derived in this work is numerically evaluated for several NB-PLC channel configurations taken from previous works, and the results show that the optimal scheme achieves a substantial rate gain over a previously proposed ad-hoc scheme. This gain is due to optimally accounting for the periodic properties of the channel and the noise.

Power line communications (PLC) has been drawing considerable interest in recent years due to the growing interest in smart grid implementation. In smart grids, network control and grid applications are allocated the frequency band of 0-500 kHz, commonly referred to as the narrowband PLC channel. This channel is characterized by strong periodic noise and low signal to noise ratio (SNR). In this work we propose a receiver which uses frequency shift filtering to exploit the cyclostationary properties of both the narrowband PLC noise, as well as the information signal, digitally modulated using orthogonal frequency division multiplexing. The results show that the new receiver obtains a substantial performance gain over previously proposed receivers, without requiring any coordination with the transmitter.

Power line communications (PLC) has been drawing considerable interest in recent years due to the growing interest in smart grid implementation. Specifically, network control and grid applications are allocated the frequency band of 0-500 kHz, commonly referred to as the narrowband PLC channel. This frequency band is characterized by strong periodic noise which results in low signal to noise ratio (SNR). In this work we propose a receiver which uses frequency shift filtering to exploit the cyclostationary properties of both the narrowband power line noise, as well as the information signal, digitally modulated using orthogonal frequency division multiplexing. An adaptive implementation for the proposed receiver is presented as well. The proposed receiver is compared to existing receivers via analysis and simulation. The results show that the receiver proposed in this work obtains a substantial performance gain over previously proposed receivers, without requiring any coordination with the transmitter.