Conventional training objectives for speech enhancement, such as mean squared error (MSE) in the time frequency domain and scale-invariant signal-to-distortion ratio (SI-SDR) in the waveform domain, exhibit limited correlation with human perception. Recent perceptual objectives better align with perceptual metrics, but often rely on auxiliary assessment networks or pretrained encoders, increasing training complexity and potentially causing instability. To address these limitations, this letter proposes a non-learned, closed-form objective, termed the multi-band relative contrastive loss (MBRCL). MBRCL enforces a relative margin in perceptually informed frequency bands to constrain the enhanced speech to be strictly closer to the clean reference than to the noisy input, and incorporates a lightweight temporal envelope gradient consistency term to preserve temporal dynamics. The loss is non-learned, almost everywhere differentiable, and requires only paired noisy–clean utterances for training. Experiments on BLSTM, CNN-U-Net, and GTCRN show gains over the matched fidelity baseline in terms of PESQ, STOI, SI-SDRi, and DNSMOS-OVRL; for the BLSTM setting, MBRCL also outperforms PESQ-inspired PMSQE, Quality-Net-based learned-loss training, and a recent frozen WavLM feature-space objective.

This paper presents a novel symbol detection method for massive Multiple-Input Multiple-Output (m-MIMO) systems, addressing the challenges posed by unknown additive noise distributions. While the optimal MIMO detector under uniform priors is the Maximum Likelihood (ML) detector, its implementation depends on accurate knowledge of the noise distribution which is often inaccessible. Furthermore, for some types of additive noise, such as impulsive noise, the probability density function (PDF) does not admit a closed-form expression, making ML detection infeasible. In our approach, we exploit the favorable propagation properties of m-MIMO systems to obtain a reliable initial estimate of the transmitted symbol vector using a simple zero-forcing (ZF) detector. We then generate a limited number of random points from the input symbol constellation in a restricted neighborhood around the ZF estimate. These points are subsequently used to obtain samples from the unknown noise distribution, which are mapped to a latent space typically (but not necessarily) characterized by a Gaussian distribution, where approximate ML detection can be performed. We benchmark our proposed detector, called Zero-Forcing based Latent Space Symbol Detector (ZF-LSSD) against existing methods across various m-MIMO configurations and noise scenarios. Numerical simulations show that our detector consistently outperforms these methods in diverse additive noise environments.

In industrial soft-sensor modeling, the scarcity and imbalance of process data often lead to overfitting and poor generalization of predictive models. To address these challenges, this article proposes a transformation-aware diffusion model (TA-DM) that integrates transformed-domain supervision for data-augmented soft sensing. We explore transformation-aware designs and introduce a novel structure-breaking loss framework that enhances the denoising objectives of denoising diffusion implicit model and TimeDDIM by encouraging the model to disrupt redundant patterns and capture richer structural variations. In implementation, our proposed approach formulates loss functions across multiple transformation domains—including discrete Fourier transform, wavelet transform, and principal component analysis (PCA)—to explicitly guide the model in learning complementary and diverse structural features beyond the original time domain, significantly advancing the representational quality and diversity of generated time-series data. To further bridge the discrepancy between the data generated by TA-DM and the real data, we propose a just-in-time learning-based sample selection strategy. This strategy leverages the representation space of the diffusion model to adaptively select local samples relevant to the current operating condition through similarity matching. These samples are then fused with limited real data to improve soft-sensor modeling. This targeted augmentation effectively narrows the synthetic-real domain gap and enhances model robustness under complex conditions. Experimental results on a numerical example and real-world industrial datasets demonstrate that TA-DM significantly outperforms existing augmentation baselines under data-scarce and distribution-shifting scenarios.

In this study, we investigate a downlink multiuser multiple-input multiple-output (MU-MIMO) system employing a two-dimensional (2D) movable antenna (MA) array. We propose a two-timescale optimization framework to jointly optimize antenna position vector (APV) and beamforming for sum-rate maximization, addressing hardware limitations that restrict real-time antenna adjustments. Specifically, APV is updated based on long-term channel statistics at the start of each coherence block, while beamforming is optimized per time slot using short-term information. An efficient stochastic successive convex approximation (SSCA)-based algorithm is developed for joint optimization. To enhance performance and reduce complexity, we propose a deep-unfolding neural network (DUNN) integrating non-linear activations and first-order Taylor approximations for matrix inversions. Furthermore, we introduce a meta-learning approach for improved initialization and rapid convergence, along with an online adaptation mechanism for continuous adjustment to changing channel conditions. Simulation results show that our proposed approach improves sum-rate compared to conventional MIMO systems with uniform arrays, and the DUNN outperforms the SSCA-based algorithm. Meanwhile, meta-learning and online adaptation framework enables network to adapt to channel changes better and converge faster.

We propose PhysMVNet, a physics-inspired end-to-end framework for multichannel speech enhancement that integrates a learnable MVDR beamformer, a Helmholtz-inspired STFT-domain regularizer, and a residual spectral mapping module. The beamformer is trained with a reconstruction loss, while the regularizer encourages local smoothness in the STFT spectrogram to improve robustness to noise and array perturbations. To mitigate spectral distortions introduced by beamforming, we incorporate a three-band residual spectral mapping network to restore fine details. Experiments on CHiME-3/4 show that PhysMVNet achieves state-of-the-art perceptual quality and intelligibility while maintaining a lightweight design suitable for realtime application. It also remains stable under extreme low-SNR conditions and array perturbations. Ablation studies confirm the contribution of each component, highlighting the benefits of physics-inspired priors in deep beamforming networks for robust, high-fidelity speech enhancement.

In this study, we explore a downlink multiuser multiple-input multiple-output (MU-MIMO) system with movable antennas (MA). Our objective is to jointly optimize the MA positions and beamforming to maximize system sum-rate. Due to hardware limitations that make real-time MA position adjustments impractical, we propose a two-timescale optimization scheme. In this scheme, the antenna position vector (APV) is update based on long-term channel statistics at the beginning of each coherence time block, while the beamforming matrix is optimized based on short-term information in each subsequent time slot within the same block. We develop an efficient stochastic successive convex approximation (SSCA)-based algorithm for joint APV and beamforming design. Additionally, in order to improve performance and reduce computational complexity, we propose a deep-unfolding neural network (NN) that preserves the structure of SSCA-based algorithm while incorporating a nonlinear activation function and trainable parameters based on first-order Taylor approximations for matrix inversion. Meanwhile, projection operators are designed to ensure compliance with the design constraints. Simulation results show that the twotimescale algorithm improves sum-rate compared to conventional MIMO systems with uniform linear arrays (ULA), and the deepunfolding NN outperforms the SSCA-based algorithm.

Extremely large-scale massive MIMO (XL-MIMO) is foreseen as a promising technology to achieve ultra-low latency, high data transmission speed, and extremely low error rate in future 6 G networks. The increases in antenna apertures and the use of higher frequencies (millimeter-wave and subTHz) in XL-MIMO significantly extend the Rayleigh distance, thereby enhancing the prevalence of near-field (NF) communications. Unfortunately, existing far-field channel models struggle to accurately capture both line-of-sight (LoS) and non-line-ofsight (NLoS) propagation paths in the presence of NF effects. Moreover, the huge number of antenna elements greatly increases computational demands and complicates channel estimation tasks. In this paper, we propose a federated learning (FL)based near-field channel estimation (NFCE) framework for mixed LoS/NLoS environments. In this framework, we employ a federated deep residual learning (FDRL)-based convolutional neural network (CNN) architecture, where only a subset of local devices participates in the distributed training process. By utilizing the complex convolution and a few residual blocks, this framework reduces the effect of signal noise while minimizing communication overhead and computational complexity. Simulation results demonstrate that our proposed framework significantly improves NFCE performance in terms of normalized mean square error and bit error rate compared to selected benchmark schemes.

This paper proposes a distributed framework for simplifying smart pilot signal assignment in cellular massive Multiple-Input Multiple-Output (MIMO) networks, aiming to reduce the complexity of pilot decontamination that comes with high cell count. The framework is based on the Triangular Lattice (TL) structure, tackling the problem of computing pilot assignments for the entire network by dividing it into ’molecules’. These molecules equate to lattice unit cells in a Triangular Lattice pattern for a network with hexagonal cell geometry. Molecules can constitute independent sub-problems by treating each one as a separate network, over which existing pilot assignment methods can be run. The resulting molecule pilot assignments can be efficiently merged to produce assignments for the entire network. In simulations, we show that the application of the TL framework to pilot assignment methods significantly reduces their computational overhead while preserving their efficacy, a trend which scales well with cell count.

This paper explores a novel Neural Network (NN) architecture suitable for Beamformed Fingerprint (BFF) localization in a millimeter-wave (mmWave) multiple-input multiple-output (MIMO) outdoor system. The mmWave frequency bands have attracted significant attention due to their precise timing measurements, making them appealing for applications demanding accurate device localization and trajectory estimation. The proposed NN architecture captures BFF sequences originating from various user paths, and through the application of learning mechanisms, subsequently estimates these trajectories. Specifically, we propose a method for trajectory estimation, employing a transformer network (TN) that relies on attention mechanisms. This TN-based approach estimates wireless device trajectories using BFF sequences recorded within a mmWave MIMO outdoor system. To validate the efficacy of our proposed approach, numerical experiments are conducted using a comprehensive dataset of radio measurements in an outdoor setting, complemented with ray tracing to simulate wireless signal propagation at 28 GHz. The results illustrate that the TN-based trajectory estimator outperforms other methods from the existing literature and possesses the ability to generalize effectively to new trajectories outside the training dataset.

In this work, we propose a new speech enhancement model referred to as auditory scene-attention model (ASAM), that can adapt dynamically to changes in the auditory scene components, such as speaker gender, input SNR levels, and background noise properties. To this end, a representative set of so-called Universal Scene Models (USM), each associated to a different auditory scene component, are first created, where each model attempts to predict a corresponding ideal ratio mask (IRM). The dynamic adaptation to changes in the auditory scene is then carried by computing the outputs of the USMs and forming a weighted combination of the most relevant scene models. This adaptation process is implemented via a frame-based attention mechanism, allowing to realize a soft selection of USM models, and taking advantage from both scene-dependent and scene-independent models. The evaluation of the proposed ASAM model, under different noise conditions and input SNR levels, shows substantial improvements in terms of standard speech Quality and intelligibility measures.

Motivated by recent advances in deep generative probabilistic modelling, we propose a robust multiple-input multiple-output (MIMO) symbol detector that aims to perform maximum likelihood (ML) detection without knowledge of the noise statistics. While the optimal MIMO detector (under uniform priors) is the ML detector, its implementation requires knowledge of the noise distribution. Furthermore, for some types of additive noise such as impulsive noise, the probability density function (PDF) of the noise does not admit a closed form expression thus making ML detection intractable. To overcome these limitations, our proposed approach learns a score function of the noise distribution directly from data. Subsequently, the learned score function is used to transform the noise distribution to a known (and tractable) prior distribution through the use of a stochastic differential equation. Via numerical simulations, the proposed detector is shown to outperform recent benchmark approaches for various types of additive noise, and to achieve near optimal ML performance where applicable.

Existing design approaches for hybrid precoding with dynamic array of subarrays architecture (DAoSA) implicitly rely on the assumption of infinite-resolution phase shifters (PSs) and digital to analog converters (DACs). However, ideal PSs and DACs are impractical and deviation from this assumption may lead to significant performance degradation. In this paper, we investigate the design of a DAoSA hybrid precoder that employs low-resolution PSs and DACs with adjustable switch connections. The design aims to minimize the power consumption while achieving high spectral efficiency under quantization constraints. To solve this complex problem, we develop a joint optimization approach comprised of three interwined algorithms: element-by-element quantized PS (EBE-QPS), DAC bit allocation (DAC-BA) and switch network design (SND). Numerical simulations show that our proposed approach for DAoSA hybrid precoder design with finite-resolution PSs and DACs leads to reduced power consumption compared to existing benchmarks while maintaining the required spectral efficiency.

As an enabling technology for emerging and future generations of wireless networks, sparse code multiple access (SCMA) offers major improvements in terms of spectral efficiency and massive connectivity. Although the message passing algorithm (MPA) for SCMA decoding at the receiver side can achieve near optimum performance, it entails high computational complexity. In this paper, to address this issue, we propose a novel SCMA decoder based on deep residual neural network (ResNet), wherein the decoder is trained to predict the transmit codewords. In our approach, residual blocks are employed to tackle the problems of accuracy saturation and vanishing gradients with deep learning based decoder, while batch normalization is utilized to enhance the stability and robustness of the decoder. The performance of the proposed ResNet decoder for SCMA is validated by means of simulations over AWGN and Rayleigh fading channels. The results show that besides a much reduced complexity, the proposed decoder leads to improvements in term of bit error rate (BER) over competing deep neural network (DNN) based decoders.

In this paper, we investigate the application of downlink beamforming along with non-orthogonal multiple access (NOMA) in a multi-user multiple-input multiple-output (MIMO) system. The joint optimization framework for user clustering, downlink beamforming and power allocation scheme is formulated as a novel mixed-integer non-linear program (MINLP), where the aim is to minimize the total transmission power while satisfying quality-of-service (QoS), user clustering and power constraints. Owing to the non-convexity and combinatorial nature of the problem, obtaining an optimal solution is challenging. To tackle this issue, we first develop an algorithm based on branch-and-bound (BB), whereby the feasible space is successively partitioned and searched by means of lower and upper bounds on the objective function. While this algorithm is shown to return an$\epsilon $-optimal solution within a finite number of iterations, it entails high computational complexity. Considering this limitation, we then reformulate the original problem into a more tractable form and conceive a low-complexity algorithm for its solution based on the penalty dual-decomposition technique. The proposed joint design algorithms for MIMO NOMA are evaluated by means of simulations over mmWave channels. Results show significant improvements in terms of total transmit power and spectral efficiency compared to benchmark approaches.

Detecting emotions directly from a speech signal plays an important role in effective human-computer interactions. Existing speech emotion recognition models require massive computational and storage resources, making them hard to implement concurrently with other machine-interactive tasks in embedded systems. In this paper, we propose an efficient and lightweight fully convolutional neural network for speech emotion recognition in systems with limited hard-ware resources. In the proposed FCNN model, various feature maps are extracted via three parallel paths with different filter sizes. This helps deep convolution blocks to extract high-level features, while ensuring sufficient separability. The extracted features are used to classify the emotion of the input speech segment. While our model has a smaller size than that of the state-of-the-art models, it achieves a higher performance on the IEMOCAP and EMO-DB datasets. The source code is available https://github.com/AryaAftab/LIGHT-SERNET

Although many state-of-the-art approaches for improving the accuracy of Voice Activity Detection (VAD) have been proposed, their performance under adverse noise conditions with low Signal-to-Noise Ratio (SNR) remains limited. In this paper, we introduce a novel attention model-based deep neural network (DNN) architecture for VAD which takes advantage of complex Ideal Ratio Mask (cIRM). The proposed model, named AM-cIRM, consists of three sequential modules: extraction of cIRM features from the noisy speech using a DNN-based architecture; combination of cIRM with log-Mel spectrogram features along with temporal contextual extension; and VAD using an attention model that exploits the spectro-temporal information in the transformed features. Experimental results show that the proposed AM-cIRM achieves improved VAD performance when compared to state-of-the-art methods under different noise conditions.

We study the problem of direction of arrival (DoA) estimation for cell-free massive MIMO (m-MIMO) systems operating over extremely high frequency (EHF) and terahertz (THF) bands, where the wireless channel can effectively be modeled by a line-of-sight path. For this model, a low-complexity deep neural network (DNN)-based method is proposed to estimate the DoA of a radio wave impinging on an access point (AP) equipped with an antenna array. To train the DNN, a special feature set is proposed obtained from the first superdiagonal entries of the spatial correlation matrix. This selection of features makes it possible to employ a DNN with only a few low-dimensional layers, which considerably speeds up training and processing. More importantly, it is shown that the trained DNN is robust against quantization noise in the array snapshot data. This property makes the centralized implementation of the proposed DNN-based method feasible, which is particularly well-suited for cell-free m-MIMO. Through extensive simulations, the new method is shown to achieve an estimation performance that nearly matches or exceeds that of classical bechmark methods, but with considerably reduced complexity.

In this paper, an adaptive algorithm is proposed for the estimation and tracking of the channel coefficients in peer-to-peer communication through a network of relays. Using the observed signals at the relay and destination nodes, the channel state information (CSI) is estimated centrally by taking advantage of a Markov model for the source-relay and relay-destination channels, and employing the Cubature Kalman Filter (CKF). The estimated CSI is used for solving a robust relay beamforming problem, aiming to minimize the total transmitted power by the relays subject to signal-to-interference-plus-noise ratio (SINR) constraint at each one of the destination nodes. Through simulations, the proposed CSI estimation is shown to be unbiased and converge to the Cramer-Rao-Lower-Bound (CRLB) for low and moderate error levels. Furthermore, the ensuing beamformer design exhibits better performance compared to existing robust beamforming methods.

Recently, there has been much interest in the use of convolutional neural networks (CNN) for mobile user localization in massive multiple-input multiple-output (MIMO) systems operating at millimeter wave (mmWave) frequencies. However, current CNN-based approaches cannot predict the confidence interval bounds for the localization accuracy. While the Bayesian neural network (BNN) method can be employed to estimate the model uncertainty, it entails a high computational cost. In this letter, the Monte Carlo (MC) dropout based method is proposed as a low-complexity approximation to BNN inference for capturing the uncertainty in a CNN-based mmWave MIMO outdoor localization system, without sacrificing accuracy. The proposed method is evaluated by means of simulations using a ray-tracing model of urban propagation at 28GHz. Results show that the localization uncertainty region can be properly determined and that their shape depends on the maximum power received at the user.

Nonlinear time-series modeling is fundamental to a wide variety of control and prediction problems. This letter focuses on the joint parameter and time-delay estimation for an extended version of the nonlinear exponential autoregressive (ExpAR) time-series model. To address the difficulties posed by the unknown time-delay and improve the estimation accuracy, we first employ the redundant rule to transform the ExpAR model into an augmented identification model. Then we invoke the multi-innovation theory to enhance data utilization and propose a new algorithm that combines stochastic gradient descent with discrete search for estimating the unknown model parameters and time-delay. The simulation results show that by properly adjusting the innovation length, the estimation accuracy of the proposed multi-innovation algorithm can significantly exceed that of the single-innovation algorithm.

Achieving high channel estimation accuracy and reducing hardware cost as well as power dissipation constitute substantial challenges in the design of massive multiple-input multiple-output (MIMO) systems. To resolve these difficulties, sophisticated pilot designs have been conceived for the family of energy-efficient hybrid analog-digital (HAD) beamforming architecture relying on adaptive-resolution analog-to-digital converters (RADCs). In this paper, we jointly optimize the pilot sequences, the number of RADC quantization bits and the hybrid receiver combiner in the uplink of multiuser massive MIMO systems. We solve the associated mean square error (MSE) minimization problem of channel estimation in the context of correlated Rayleigh fading channels subject to practical constraints. The associated mixed-integer problem is quite challenging due to the nonconvex nature of the objective function and of the constraints. By relying on advanced fractional programming (FP) techniques, we first recast the original problem into a more tractable yet equivalent form, which allows the decoupling of the fractional objective function. We then conceive a pair of novel algorithms for solving the resultant problems for codebook-based and codebook-free pilot schemes, respectively. To reduce the design complexity, we also propose a simplified algorithm for the codebook-based pilot scheme. Our simulation results confirm the superiority of the proposed algorithms over the relevant state-of-the-art benchmark schemes.

Device-to-device (D2D) communication is an enabling technology for fog computing by allowing the sharing of computation resources between mobile devices. However, temperature variations in the device CPUs affect the computation resources available for task offloading, which unpredictably alters the processing time and energy consumption. In this paper, we address the problem of resource allocation with respect to task partitioning, computation resources and transmit power in a D2D-aided fog computing scenario, aiming to minimize the expected total energy consumption under probabilistic constraints on the processing time. Since the formulated problem is non-convex, we propose two sub-optimal solution methods. The first method is based on difference of convex (DC) programming, which we combine with chance-constraint programming to handle the probabilistic time limitations. Considering that DC programming is dependent on a good initial point, we propose a second method that relies on only convex programming, which eliminates the dependence on user-defined initialization. Simulation results demonstrate that the latter method outperforms the former in terms of energy efficiency and run-time.

Speech exhibits strong dependencies among its samples in both time and frequency domains. In this paper, we propose a low-complexity composite model for speech enhancement (SE) that integrates a convolutional neural network (CNN) and a long short-term memory (LSTM) network. These two modules take full advantage of the spectral and temporal information of input speech and extract in parallel a complementary set of features. The CNN is enabled to capture non-local spectral information via dilated frequency convolutions. It also incorporates an attention mechanism to recalibrate its weights without imposing considerable additional complexity. A grouping strategy is adopted for LSTM implementation to reduce its complexity while keeping performance almost unchanged. Our composite model is carefully designed to address concerns in real-time applications including limited computational resources, low-latency processing, and causal architecture. Through extensive and comparative simulation studies, it is shown that the proposed model significantly outperforms some other DNN-based SE methods in the recent literature.

In this paper, we study the application of spatial user clustering along with downlink beamforming in multiple-input multiple-output sparse code multiple access (MIMO-SCMA) systems. A user clustering algorithm based on a constrained K-means method is proposed to limit the number of users in each cluster. Subsequently, a two-stage beamforming approach is developed in which a cluster beamformer and user-specific beamformer obtained from each stage are combined to form the final beamformer for each user. Specifically, in the first stage, the block diagonalization technique is employed to design cluster beamformers so that the inter-cluster interference is removed. In the second stage, an optimization problem is formulated to determine user-specific beamformers for all the users such that the total transmit power is minimized under signal-to-interference-plus-noise ratio (SINR) constraints. The performance of the proposed user clustering and downlink beamforming approaches in MIMO-SCMA systems is evaluated through simulations. The results provide useful insights into the advantages of the proposed scheme in terms of transmit power, and spectral efficiency over benchmark approaches.

In this paper, we propose a novel power allocation scheme for physical-layer network coding (PNC) in downlink multi-way relay channels (MWRC). The power allocation is formulated as a constrained optimization problem, where the aim is to maximize the success probability under a total power constraint when using Babai estimation for signal detection. Optimizing over this metric allows us to maximize the probability of successfully decoding a chain of network codes, which is of crucial importance in downlink multi-way PNC. Specifically, to meet the different requirements for transmission quality in applications, we consider different aggregate measures of success probability over the participating user terminals, i.e., the arithmetic mean, the geometric mean, and the maximin. For each measure, we formulate a constrained optimization and demonstrate the concavity of the objective, allowing us to obtain solution efficiently via iterative means. The performance of the proposed power allocation schemes for downlink PNC in MWRC is evaluated by means of computer simulations over Raylegih fading channels. The results demonstrate the effectiveness of the proposed schemes in improving the success probability in the reception of a chain of network codes.

In this paper, we investigate secrecy-energy efficient hybrid beamforming (BF) schemes for a satellite-terrestrial integrated network, wherein a multibeam satellite system shares the millimeter wave spectrum with a cellular system. Under the assumption of imperfect angles of departure for the wiretap channels, the hybrid beamformer at the base station and digital beamformers at the satellite are jointly designed to maximize the achievable secrecy-energy efficiency, while satisfying signal-to-interference-plus-noise ratio constraints of both the earth stations (ESs) and cellular users. Since the formulated optimization problem is nonconvex and mathematically intractable, we propose two robust BF schemes to obtain approximate solutions with low complexity. Specifically, for the case of a single ES, we integrate the Charnes-Cooper approach with an iterative search algorithm to convert the original nonconvex problem into a solvable one and obtain the BF weight vectors. In the case of multiple ESs, by exploiting the sequential convex approximation method, we convert the original problem into a linear one with multiple matrix inequalities and second-order cone constraints, for which we obtain a solution with satisfactory performance. The effectiveness and superiority of the proposed robust BF design schemes are validated via simulations using realistic satellite and terrestrial downlink channel models.

In this paper, we propose a novel robust downlink beamforming (BF) design for satellite-terrestrial integrated networks. Under a realistic assumption that the angular information of eavesdroppers is not perfectly known, we establish an optimization framework for hybrid BF at the terrestrial base station and digital BF at the satellite to maximize the secrecy-energy efficiency of the system, while satisfying the quality-of-service constraints of both earth station and cellular user. Since the formulated optimization problem is mathematically intractable, we present an iterative algorithm based on the Charnes-Cooper approach to optimize the BF weight vectors. The effectiveness and superiority of the proposed robust hybrid BF scheme are validated via computer simulations.

In this paper, we study the realization of any given fully-digital precoder (FDP) by hybrid analog/digital precoding (HADP) in wide-band mmWave systems. We first formulate the massive-MIMO OFDM-based HADP system design and then, introduce the notion of perfect reconstruction at sampling point (PRSP) for FDP realization. Furthermore, the minimum number of required RF chains and its trade-off with the bandwidth of the transmitted signal is discussed. Next, we present a novel FDP realization with two RF chains. Finally, simulation results are presented showing the superiority of the proposed hybrid design to recently published works.

In this paper, we present a deep neural network (DNN) based algorithm to restore the high-frequency (HF) component of the enhanced speech processed by Kalman filtering, where the DNN is applied for estimating the magnitude of HF component from the low-frequency (LF) counterpart. The complete HF component is then computed with the estimated magnitude given by the DNN and the phase of the Kalman filtered speech. By incorporating our restoration algorithm into Kalman filter based speech enhancement method, our new speech enhancement system is able to recover the HF component with better perceptual quality and less distortion. Experimental results demonstrate that the proposed method outperforms the state-of-the-art Kalman filter based method in terms of both speech quality and intelligibility.

This paper addresses the problem of channel estimation using pilots in hybrid analog/digital massive multiple-input multiple-output (MIMO) systems for future millimetre wave (mmWave) communications. To further reduce system cost and implementation complexity, we consider an alternative architecture derived from RF switches as opposed to the phase shifters in the conventional literature. The channel estimation is formulated as a combinatorial optimization problem where the aim is to minimize the mean square error (MSE) between the real and estimated channels over a finite set of allowed values for the precoder switches. A genetic algorithm (GA) is developed for solving this problem and obtaining the MIMO channel estimates. Simulations show that the proposed scheme can estimate channels as accurately, if not more, as an existing solution using phase shifters.

In this article, relay-assisted computation offloading (RACO) is investigated, where user A wishes to share the results of computational tasks with another user B with the assistance of a mobile-edge relay server (MERS). To enable this computation offloading, we propose a hybrid relaying (HR) approach employing a pair of orthogonal frequency bands, which are, respectively, used for the amplify-forward relaying of computational results and the decode-forward relaying of the unprocessed raw tasks. The motivation here is to adapt the allocation of computing and communication resources both to dynamic user requirements and to diverse computational tasks. Using this framework, we seek to minimize the weighted sum of the execution delays and the energy consumption in the RACO system by jointly optimizing the computation offloading ratio, the bandwidth allocation, the processor speeds, as well as the transmit power levels of both user A and the MERS, under some practical constraints. By adopting a series of transformations, we first recast this problem into a form amenable to optimization and then develop an efficient iterative algorithm for its solution based on the concave-convex procedure (CCCP). By virtue of the particular problem structure in our case, we propose furthermore a simplified algorithm based on the inexact block coordinate descent (IBCD) method, which leads us to much lower computational complexity. Finally, our numerical results demonstrate the advantages of the proposed algorithms over the state-of-the-art benchmark schemes.

Due to the severe pathloss experienced by electromagnetic waves in the millimeter wave (mmWave) band, a substantial challenge in their design is to have an adequate coverage area. With the objective of improving the coverage area and the sum rate attained, we conceive new full-duplex (FD) mmWave multiple-input multiple-output (MIMO) multiple-relay systems. Specifically, we propose a novel two-timescale analog-digital hybrid beamforming scheme for maximizing the sum rate, while reducing the system's complexity and the channel state information (CSI) signalling overhead, as well as mitigating both the effects of self-interference and that of outdated CSIs caused by the associated delays. In the proposed scheme, the long-timescale analog beamforming matrices are designed based on the available channel statistics and updated in a frame-based manner, where a frame contains a fixed number of time slots. By contrast, the short-timescale digital beamforming matrices are optimized more frequently - namely for each time slot - based on the low-dimensional effective CSI matrices available on a real-time basis. We develop both an efficient analog beamforming algorithm based on the cut-set bound as well as on stochastic successive convex approximation (SSCA) and an innovative digital beamforming algorithm that relies on the theory of penalty dual decomposition (PDD), where our design objective is to maximize the system's sum rate. Both the convergence properties and the computational complexity of the proposed algorithms are also examined. Our simulation results show that the proposed two-timescale hybrid beamforming design significantly outperforms the conventional beamformers both in terms of requiring a lower CSI-signalling overhead and a higher sum rate in the face of realistic outdated CSIs.

In this work, we investigate the problem of secure communications in a downlink large-scale multi-antenna assisted simultaneous wireless information and power transfer (SWIPT) system, where a base station (BS) transmits signals to serve a number of information decoding (ID) and energy harvesting (EH) users. Considering that the EH users can potentially eavesdrop the ID users' confidential information, we study the robust joint design of the hybrid analog-digital (A/D) beamforming (BF) matrices and of the artificial redundant signal (ARS) covariance matrix at the BS, where the aim is to maximize the worst-case sum secrecy rate for the ID users under a transmit power constraint, a nonlinear EH constraint and a unit-modulus constraint on the entries of the analog BF matrix. The corresponding optimization problem is very challenging due to the nonlinear and nonconvex objective function and constraints. Using innovative optimization techniques, we first transform the original problem into an equivalent but more tractable form, and then develop a novel joint iterative algorithm based on the penalty-concave-convex procedure (CCCP) for solving the resultant problem. We show that the proposed penalty-CCCP based algorithm for ARS-aided robust joint hybrid BF design converges to a Karush-Kuhn-Tucker solution of the original problem, and also analyze its computational complexity. Our simulation results verify that the resultant robust joint hybrid BF design algorithm relying on ARS significantly outperforms the conventional hybrid BF benchmark algorithms and efficiently achieves the performance of the fully-digital BF with reduced number of radio frequency chains and energy consumption.

This work studies the joint problem of content placement, remote radio head (RRH) clustering and beamformer design, in a cache-enabled cloud-radio access network (C-RAN). In the considered system, downlink users are cooperatively served by multiple RRHs, in turn connected to a centralized baseband unit (BBU) pool via fronthaul links. Each RRH is equipped with a local cache from which it can directly acquire the requested user contents, without utilizing the fronthaul links. We aim to jointly optimize the aforementioned three aspects, in order to strike a balance between fronthaul traffic reduction and transmission power minimization. To this end, we propose to employ the ratio between these two important system utilities as the objective function, referred to as caching efficiency. Two joint design algorithms are presented to address the resulting nonconvex optimization problem, which features coupling constraints and mixed-integer variables, namely: the penalty concave-convex procedure (P-CCCP) and penalty dual decomposition (PDD) based algorithms. Furthermore, since content placement is usually updated over a larger timescale, we propose a two-timescale joint design algorithm, where the P-CCCP and PDD-based algorithms can be employed for efficient initialization as well as for establishing performance limits. Simulation results validate the efficiency of the proposed algorithms.

In this paper, we propose a novel zero-forcing (ZF)- based beamforming (BF) scheme for a wireless powered cognitive satellite-terrestrial network (CSTN) operated in the millimeter wave band. Assuming that the satellite and base station are equipped with multiple antennas, we aim at maximizing the sum rate of the CSTN while satisfying the signal-to-interference-plus-noise- ratio requirements for both the information receivers (IRs) and earth stations, the energy harvesting requirements of the energy receivers (ERs), and the secrecy constraints at the ERs. Since the resulting optimization problem is mathematically intractable, we propose a novel multi-beam-based ZF BF scheme to generate beamforming vectors to serve the IRs and ERs. Specifically, the original nonconvex problem is decomposed into two independent subproblems. The first subproblem, which features beam orthogonality constraints, leads to closed form solutions for the beamforming vectors. The second subproblem, aiming at finding the optimal power allocation, is solved via the S-procedure. Finally, the effectiveness of the proposed scheme is demonstrated by simulation results.

In this paper, we present novel methods for signal detection in single carrier zero-padded spatial modulation under high mobility conditions. By expressing the doubly selective channel in terms of the basis expansion model (BEM), first a maximum likelihood (ML) method is presented as a processing framework. To reduce the complexity of the exhaustive ML search, two novel methods, respectively the BEM-based partial interference cancellation (BPIC) and BPIC with successive interference cancellation (BPIC-SIC), are then proposed. The complexity of the new methods are compared and their performance is evaluated by simulations in terms of bit error rate. The results indicate that the new schemes can remarkably improve the performance compared with the conventional methods.

In this paper we propose a fully convolutional neural network (CNN) for complex spectrogram processing in speech enhancement. The proposed CNN consists of one-dimensional (1-d) convolution and frequency-dilated 2-d convolution, and incorporates a residual learning and skip-connection structure. Compared with the state-of-the-art, the proposed CNN achieves a better performance with fewer parameters. Experiments have shown that the complex spectrogram processing is effective in terms of phase estimation, which benefits the reconstruction of clean speech especially in the female speech case. It is also demonstrated that the model yields a convincing performance with small memory footprint when the number of parameters is limited.

This work studies the joint problem of optimal content placement, RRH clustering and beamformer design, in a cache-enabled cloud-radio access network (C-RAN). In the considered system, multiple remote radio heads (RRHs) connected to a centralized baseband unit (BBU) pool via fronthaul links, cooperatively serve the downlink users by grouping them into potentially overlapping clusters. Each RRH is equipped with a local cache from which it can directly acquire the requested user contents, without the need to occupy the fronthaul links. We aim to jointly optimize the caching placement, user association and downlink beamforming vector at each RRH, in order to strike a balance between fronthaul traffic reduction and transmission power minimization. To this end, we propose to employ the ratio between these two important system utilities as the objective function, referred to as caching efficiency. A penalty dual decomposition (PDD) based algorithm is presented to address the resulting nonconvex optimization problem, which features coupling constraints and mixed-integer variables. Simulation results validate the efficiency of the proposed algorithm.

In this paper, we present a novel deep neural network (DNN) based Kalman filter (KF) algorithm for speech enhancement, where DNN is applied for estimating key parameters in the KF, namely, the linear prediction coefficients (LPCs). By training the DNN with a large database and making use of the powerful learning ability of DNN, our proposed DNN-KF algorithm is able to estimate LPCs from noisy speech more accurately and robustly, leading to an improved performance as compared to traditional KF based approaches in speech enhancement. Experimental results demonstrate that our DNN-KF method outperforms two existing KF based speech enhancement methods in terms of both speech quality and intelligibility.

Synchronization is a challenging problem especially for distributed systems, such as out-of-coverage D2D networks, as no common reference time is available. In such cases, devices use distributed synchronization algorithms, however, accurately determining when to stop the synchronization process is as challenging as achieving synchronization since they do not have the synchronization status of other devices in the network. From energy efficiency and performance perspective, the synchronization process should be stopped at all devices at the same time. In addition, to counteract the effect of propagation delays during synchronization, timing-advance (TA) clocks should be employed. This could be achieved in the synchronization process, however, after the devices are synchronized, there is no central mechanism to instruct them on TA clocks for transmitting or receiving data packets. In this paper, we propose a synchronization algorithm which, in an energy-efficient manner, allows devices to (i) acquire the synchronization status of others and terminate the synchronization process as soon as all devices in the network are synchronized, (ii) allow the synchronized devices to properly advance/regress their clocks prior to data communication by tracking their relative timing. We numerically demonstrate that the maximum synchronization error over multipath channels is around 0.6μs and it can be maintained during data communication.

In this paper, we study the design of hybrid analog/digital beamformers for uplink connection in massive multiple-input multiple-output (MIMO) systems under imperfect channel state information (CSI). The norm-bounded channel error model is used to capture characteristics of imperfect CSI in practical systems. The objective function is formulated based on the minimum mean squared error (MMSE) worst-case robustness. We consider both single user (SU) and multiuser (MU) reception modes of a millimeter-Wave (mmWave) massive-MIMO base station (BS). For the SU scenario, we study hierarchical beamformer optimization as well as joint precoder/combiner optimization for users with limited and extended computational capabilities, respectively. These optimization techniques are subsequently extended to the MU case where a new hybrid robust combiner design is proposed. Simulation results are presented confirming the superiority of our designs when compared to recent robust hybrid designs in the literature.

In this study, the authors present and investigate a novel cooperative relaying scheme for cognitive radio networks (CRNs), which is based on non‐orthogonal multiple access (NOMA). In the proposed scheme, following the detection of an idle channel, the secondary base station transmits a power domain NOMA signal to a first nearby secondary user (SU). In addition to decoding its own signal, this user applies a decode‐and‐forward strategy to relay the signal intended to a second SU. In contrast to previous works, where the spectrum sensing and transmission phases are treated separately, the authors here consider both phases jointly in the design and analysis of the proposed scheme. To characterise performance of the latter, analytical expressions are derived for the outage probability and the ergodic rate of the two SUs by assuming a flat Rayleigh fading channel model. The performance of two traditional orthogonal multiple access schemes is also analysed for comparison. Simulation and numerical results are presented to demonstrate the effectiveness of the proposed cooperative relaying scheme for CRN, as well as the accuracy of the analytical results.

The joint design of hybrid beamforming matrices is conceived for multiuser mm-wave full-duplex (FD) multiple-input multiple-output (MIMO) relay-aided systems in the presence of realistic channel state information (CSI) errors. Specifically, considering a probabilistic CSI error model, we maximize the system's worst-case sum rate by jointly optimizing the base station's (BS's) analog and digital beamforming matrices, plus the analog receive and transmit beamforming matrices of the relay station (RS) as well as its digital amplify-and-forward beamforming matrix under practical constraints. Explicitly, the transmit power constraints of the BS and RS, the residual self-interference power constraint of the RS, the per-user quality of service constraints, and the unit-modulus constraints on the analog beamforming matrix elements are all taken into account. Since the resultant optimization problem is very challenging due to its highly nonlinear objective function and nonconvex coupling constraints, we first transform it into a more tractable form. We then develop a novel joint optimization algorithm based on the penalty dual decomposition (PDD) technique to solve the resultant problem. The proposed PDD-based algorithm performs double-loop iterations: the inner loop updates the optimization variables in a block coordinate descent fashion, while the outer loop adjusts the Lagrange multipliers and penalty parameter, hence ensuring convergence to the set of stationary solutions of the original problem. Our simulations show that the mm-wave FD hybrid MIMO relay systems relying on our new algorithm significantly outperform both their non-robust FD and conventional half-duplex counterparts.

This paper analyzes the ergodic capacity of a user-centric virtual cell (VC) dense network, where multiple access points (APs) form a VC for each user equipment (UE) and transmit data cooperatively over millimeter wave (mmWave) channels. Different from traditional microwave radio communications, blockage phenomena have an important effect on mmWave transmissions. Accordingly, we adopt a distance-dependent line- of-sight (LOS) probability function and model the locations of the LOS and non-line-of-sight (NLOS) APs as two independent non-homogeneous Poisson point processes (PPP). Invoking this model in a VC dense network, new expressions are derived for the downlink ergodic capacity, accounting for: blockage, small-scale fading and AP cooperation. In particular, we compare the ergodic capacity for different types of fading distributions, including Rayleigh and Nakagami. Numerical results validate our analytical expressions and show that AP cooperation can provide notable capacity gain, especially in low- AP-density regions.

This paper addresses the design of multiuser MIMO amplify-and-forward relaying within a cloud radio access network (C-RAN) from an energy-efficiency perspective. The aim is to jointly select remote radio heads and optimize their transceiver in order to assist the communication between multiple source-destination pairs. We formulate the design problem as an interference leakage minimization subject to per-relay power constraints along with linear signal preserving constraints at the destinations. To obtain an energy efficient relaying solution, the objective function is penalized with a regularization term which promotes group-sparsity among the resultant relaying weights. A low-complexity iterative algorithm based on the alternating direction method of multipliers (ADMM) is then proposed to solve the regularized problem. Simulation results demonstrate the explicit benefits of the proposed algorithm, which results in notably lower power consumption and computational complexity than conventional relaying design methods.

A derivative of orthogonal frequency division multiplexing (OFDM) based on offset quadrature amplitude modulation, OFDM/OQAM, is among the waveform contenders for future wireless networks. Focusing on multiuser scenarios, i.e. uplink of a multiple access system, we propose an improved joint estimation method for carrier frequency offset, sampling time offset and channel impulse response needed for the practical application of OFDM/OQAM. The proposed method offers a pilot-based maximum-likelihood (ML) estimation of the unknown parameters in multiuser OFDM/OQAM systems. Formulation of the estimator is based on the splitting of each received pilot symbol into contributions from surrounding pilot symbols, non-pilot symbols, additive noise and multiuser interference. The proposed estimator is compared with a highly cited previous work of the same focus where the improvements in the results indicate the superiority of the former.

This paper addresses the design of joint beamforming and jamming for a millimeter wave (mmWave) information surveillance system where a suspicious transmitter in the network sends messages to a suspicious receiver under the supervision of a surveillant controller (SC), which not only carries out the duty of a base station or other access point, but also legitimately monitor the suspicious link. Specifically, we seek to maximize the effective monitoring rate for information surveillance by jointly optimizing the analog transmit and receive beamforming vectors of the suspicious link, the analog jamming and monitoring beamforming vectors at the SC and the jamming signal's power level under transmit power, successful monitoring, and self-interference power constraints at the SC, along with unit modulus constraint on the elements of the radio frequency analog beamforming vectors. The resulting optimization problem is quite challenging due to the tight coupling of the design variables in the objective function and constraints. To solve it, we develop a novel algorithm based on the penalty dual decomposition (PDD) technique, where the exacting constraints are penalized and dualized into the objective function as augmented Lagrangian components. The proposed PDD-based algorithm performs double-loop iterations, i.e., the inner loop resorts to the concave-convex procedure to update the optimization variables; while the outer loop adjusts the Lagrange multipliers and penalty parameter of the augmented Lagrangian cost function. We show that the proposed PDD-based joint beamforming and jamming algorithm converges to a stationary solution of the original problem. Based on our simulation results, the proposed algorithm achieves significantly better performance than the conventional beamforming and jamming algorithms.

In this letter, a novel diffusion Volterra (DV) algorithm is proposed for distributed in-network system identification in the presence of α-stable noise. The proposed algorithm is based on the logarithmic least mean pth-power criterion, which makes it robust against impulsive interferences, at the price of increased complexity. To overcome this shortcoming, we further develop the diffusion interpolated Volterra algorithm, which provides computational savings and good performance in comparison with the DV algorithm. Simulations results show that the proposed adaptive algorithms achieve better performance than the state-of-the-art approaches for distributed nonlinear system identification in impulsive noise.

In this paper, we introduce a discriminative training algorithm of the non-negative matrix factorization (NMF) model for single-channel enhancement of convolutive noisy speech. The basis vectors for the clean speech and noises are estimated simultaneously during the training stage by incorporating the concept of classification from machine learning. Specifically, we employ the probabilistic generative model (PGM) of classification, specified by an inverse Gaussian distribution, as a priori structure for the basis vectors. Both the NMF and classification parameters are obtained by using the expectation-maximization (EM) algorithm, which guarantees convergence to a stationary point. Experimental results show that the proposed algorithm provides better enhancement performance than the benchmark algorithms.

Time synchronization is a key aspect of device-to-device (D2D) schemes, particularly in decentralized networks where no reference time is available. Distributed phase-locked loops (DPLL) is a synchronization algorithm well suited for decentralized situations. In this work, we study DPLL in the context of 5G networks, where we include in our analysis several practical aspects of D2D communication, such as propagation delays, multipath propagation, and the use of single-carrier frequency division multiple access (SC-FDMA). We propose practical methods to compensate for their effects, and introduce new performance metrics to evaluate the merits of the synchronization algorithm. Through simulations at the physical layer, which capture the effects of analog-digital conversions, we demonstrate that time synchronization in a decentralized setting is possible under the constraints specified by the 3GPP for D2D applications.

This paper considers the uplink of large-scale multiple-user multiple-input multiple-output (MU-MIMO) millimeter wave (mmWave) systems, where a number of mobile stations (MSs) communicate with a single base station (BS) equipped with a large-scale antenna array, for application to fifth generation (5G) wireless networks. Within this context, the use of hybrid transceivers along with antenna selection can significantly reduce the implementation cost and energy consumption of analog phase shifters and low-noise amplifiers (LNA). We aim to jointly design the MS beamforming vectors, the hybrid receiving matrices (baseband and analog) and the antenna selection matrix at the BS in order to maximize the achievable system sum-rate. By exploiting the special structure of the problem and linear relaxation, we first convert this problem into three subproblems which are solved via an alternating optimization (AO) method. Specifically, the antenna selection matrix is optimized via the concave-convex procedure (CCCP); the weighted mean-square error minimization (WMMSE) approach is used to find the solution for the transmit beamformer; and the hybrid receiver is obtained via manifold optimization (MO). The convergence of the proposed algorithm is analysed and its effectiveness is verified by simulation.

Wireless transmission with lens antenna arrays is becoming more and more attractive for millimeter wave (mmWave) multiple-input multiple-output (MIMO) systems with limited radio frequency (RF) chains due to their energy-focusing capability. In this paper, we consider the joint design of beam selection and precoding to maximize the sum rate of a downlink single-sided lens MU-MIMO mmWave system under transmit power constraints. We first formulate the optimization problem into a tractable form using the popular weighted minimum mean squared error (WMMSE) approach. To solve this problem, we then propose an efficient joint beam selection and precoding algorithm based on the innovative penalty dual decomposition (PDD) method. Simulation results demonstrate that our proposed algorithm can achieve near-optimal performance when compared to the fully digital precoding scheme and thus outperform the competing methods.

Cognitive radio (CR) is an important technology that allows to deal with spectrum congestion, where secondary applications (users) attempt to access a frequency band that is reserved for a primary application. A challenging function for a CR is to sense a frequency band and detect the absence or presence of a licensed user, a task referred to as spectrum sensing. In this paper, we investigate the performance of the Rao-test based detector for wideband spectrum sensing under non-Gaussian noise in a multi-carrier transmission framework. Specifically, we incorporate this detector into the universal filtered multicarrier (UFMC) modulation scheme envisaged for 5G systems. Through numerical simulations, we show that the Rao-test based detector combined with UFMC outperforms the traditional OFDM based system in a realistic non-Gaussian noise environment.

This study investigates the problem of physical layer security for amplify‐and‐forward (AF) multiple‐input multiple‐output (MIMO) relay systems operating in the presence of a passive eavesdropper. Specifically, the authors consider the robust design of an artificial noise (AN)‐assisted non‐linear transceiver employing Tomlinson–Harashima precoding (THP), with imperfect knowledge of the legitimate channel states. The design problem can be reformulated as a two‐level optimisation, where the outer problem aims to optimise the source precoder as a function of the relay precoder, while the inner problem at the relay aims to jointly optimise the relay precoder as well as the power allocation between the AN and the information‐bearing signals. To solve the inner problem, the authors adopt a bisection method which attempts to maximise the AN power level, to confuse the eavesdropper, while satisfying the mean‐squared‐error requirement for the intended user. Some relaxation for the objective function is applied to transform the problem into a standard convex optimisation one. Regarding the outer problem, closed‐form solutions for the precoders can be derived by an iterative method based on the Karush–Kuhn–Tucker conditions. Simulation results illustrate the superior secrecy performance provided by the proposed non‐linear transceiver design with AN and THP.

This paper considers the uplink of large-scale multiple-user multiple-input multiple-output millimeter wave systems, where several mobile stations (MSs) communicate with a single base station (BS) equipped with a large-scale antenna array, for application to fifth generation wireless networks. Within this context, the use of hybrid transceivers along with antenna selection can significantly reduce the implementation cost and energy consumption of analog phase shifters and low-noise amplifiers. We aim to jointly design the MS beamforming vectors, the hybrid receiving matrices (baseband and analog), and the antenna selection matrix at the BS in order to maximize the achievable system sum-rate under a set of constraints. The corresponding optimization problem is nonconvex and difficult to solve, mainly due to the receive antenna selection and constant modulus constraints on the analog receiving matrix. By exploiting the special structure of the problem and linear relaxation, we first convert this problem into three subproblems, which are solved via an alternating optimization method. The latter iteratively updates the antenna selection matrix, the transmit beamforming vectors, and the hybrid receiving matrices by sequentially addressing each subproblem while keeping the other variables fixed. Specifically, the antenna selection matrix is optimized via the concave-convex procedure; the weighted mean-square error minimization approach is used to find the solution for the transmit beamformer; and the hybrid receiver is obtained via manifold optimization. The convergence of the proposed algorithm is analyzed and its effectiveness is verified by simulation.

This paper addresses joint transceiver design for secure downlink communications over a multiple-input multiple-output relay system in the presence of multiple legitimate users and malicious eavesdroppers. Specifically, we jointly optimize the base station (BS) beamforming matrix, the relay station (RS) amplify-and-forward transformation matrix, and the covariance matrix of artificial noise, so as to maximize the system worst-case secrecy rate in the presence of the colluding eavesdroppers under power constraints at the BS and the RS, as well as quality of service constraints for the legitimate users. This problem is very challenging due to the highly coupled design variables in the objective function and constraints. By adopting a series of transformation, we first derive an equivalent problem that is more tractable than the original one. Then, we propose and fully develop a novel algorithm based on the penalty concave-convex procedure (penalty-CCCP) to solve the equivalent problem, where the difficult coupled constraint is penalized into the objective and the resulting nonconvex problem is solved at each iteration by resorting to the CCCP method. It is shown that the proposed joint transceiver design algorithm converges to a stationary solution of the original problem. Finally, our simulation results reveal that the proposed algorithm achieves better performance than other recently proposed transceiver designs.

Jointly detecting the signals from the source and relay in a spatial multiplexing (SM) multiple-input multiple-output (MIMO) relaying system improves the transmit reliability significantly. However, the existing joint detection schemes for SM MIMO relaying systems, which achieve full diversity, such as the near maximum likelihood (ML) decoder, suffer from high complexity. In this paper, we propose a weighted combining (WC) algorithm, which is applied before the detector in the SM MIMO decode-and-forward relaying system. The proposed algorithm merges the received signal vectors from the source and relay into a combined signal without expanding their dimension, and formulates an equivalent MIMO channel matrix for the combined signal, resulting in a lower complexity for the subsequent detection. We analyze the performance of the proposed WC algorithm with ML detection in terms of the diversity order and computational complexity. An approximate upper bound on the symbol error probability (SEP) for the proposed algorithm is also derived. Simulation results show that in symmetric networks, the proposed WC algorithm achieves substantially lower complexity, while maintaining an SEP performance similar to that of the benchmark NML decoder. The consistency of the derived upper bound on the SEP is also verified by simulations.

In this paper, we propose joint transceiver design algorithms for the full-duplex K -pair multiple-input multiple-output interference channel with simultaneous wireless information and power transfer. To mitigate and exploit the complex interference, we consider two important utility optimization problems, i.e., the sum power minimization problem and the sum-rate maximization problem. In the first problem, our aim is to minimize the total transmission power under both transmission rate and energy harvesting (EH) constraints. An iterative algorithm based on alternating optimization (AO) and with guaranteed monotonic convergence is proposed to successively optimize the transceiver coefficients. The algorithm consists of three main steps, where the concave-convex procedure (CCCP), the minimum mean-square error (MMSE) criterion, and the semidefinite relaxation technique are, respectively, employed to compute the vectors of power splitting ratios, the receiving matrices, and the transmitting beamforming vectors. Two simplified algorithms based on fixed beamformers, namely, the maximum ratio transmission and the maximum signal-to-interference-leakage beamformers are also proposed. In the second problem, our aim is to maximize the sum-rate under additional power and EH constraints. Due to the highly non-convex nature of this problem, we first reformulate it into an equivalent-weighted MMSE problem by introducing suitable weight factors, such that the global optima of the two problems are identical. Then, by utilizing the concept of AO and CCCP, we show that the equivalent problem can be efficiently solved. Again, with the aid of the fixed beamformers, two simplified algorithms are provided to reduce the computational complexity. Simulation results are presented to validate the effectiveness of the proposed algorithms.

This paper investigates the design of robust nonlinear transceivers conceived for multiple-input multiple-output full-duplex wireless-powered relay networks in the face of realistic imperfect channel state information (CSI). A novel self-energy recycling aided relaying protocol is employed, whereby the relay node benefits from energy harvesting (EH) gleaned from the self-interfering link in addition to its primary energy. The proposed nonlinear transceiver relies on a Tomlinson-Harashima (TH) precoder along with an amplify-and-forward (AF) relaying matrix and a linear receiver, where the TH precoder is composed of a feedback matrix and a source precoding matrix. Two different criteria are considered for the robust design of the nonlinear transceiver in the presence of channel estimation errors modeled by the Gaussian distribution. The first one aims to minimize the mean-squared-error (MSE) at the destination subject to a transmit power constraint at the source and an EH constraint at the relay. The resultant optimization problem is converted to four subproblems and solved via an alternating optimization (AO) algorithm that iteratively updates the transceiver coefficients by sequentially addressing each subproblem, while keeping the other matrix variables fixed. The second design criterion aims to minimize the transmit power at the source under both MSE and EH constraints. Similarly, an AO-based iterative algorithm is proposed for solving this problem. Our simulation results show that the robust design advocated is capable of alleviating the effects of CSI errors, hence improving the robustness of the system over that of the corresponding linear designs.

In this paper, we propose a basis compensation algorithm for non-negative matrix factorization (NMF) models as applied to supervised single-channel speech enhancement. In the proposed framework, we use extra free basis vectors for both the clean speech and noise during the enhancement stage in order to capture the features which are not included in the training data. Specifically, the free basis vectors of the clean speech are obtained by exploiting a priori knowledge based on a Gamma distribution. The free bases of the noise are estimated using a regularization approach, which enforces them to be orthogonal to the clean speech and noise basis vectors estimated during the training stage. Experimental results show that the proposed NMF algorithm with basis compensation provides better performance in speech enhancement than the benchmark algorithms.

This paper addresses the transceiver design problem for secure downlink communications over a multiple-input multiple-output (MIMO) relay system in the presence of multiple eavesdroppers. A new algorithm based on alternating optimization (AO) is first proposed to maximize the signal-to-noise ratio (SNR) of a legitimate receiver under power constraints at the base station (BS) and the relay station (RS) and a set of secrecy constraints, by using the semidefinite relaxation (SDR) technique. To reduce complexity, a simplified design algorithm based on switched relaying (SR) is also proposed, in which both the BS and the RS are equipped with a codebook of permutation matrices. Based on this codebook, we construct a number of latent transceivers, each consisting of a BS beamforming vector and an optimally scaled RS permutation matrix. We use the bisection search and second-order cone programming (SOCP) techniques to design each latent transceiver and choose the optimal one with the largest SNR. We also develop an efficient approach to construct the codebook of permutation matrices. Our results show that the SR based algorithm significantly reduces the computational complexity while maintaining a similar performance to the AO based algorithm.

In this paper, we present a new dereverberation algorithm based on the weighted prediction error (WPE) method. In contrast to the conventional WPE method which alternatively estimates the reverberation prediction weights and early speech spectral variance, the proposed algorithm estimates the latter efficiently by employing a geometric spectral enhancement approach and a proper estimate for late reverberant spectral variance (LRSV). Hence, our algorithm does not require iterations to estimate the reverberation prediction weights nor needs alternation between the prediction weights and the spectral variance of early speech. Performance assessments demonstrate considerable improvements in terms of speech quality measures and computational load compared to previous WPE-based dereverberation methods.

In this paper, we propose a single channel speech enhancement algorithm using a subband iterative Kalman filter. A wavelet filterbank is first used to decompose the noise corrupted speech into a number of subbands. To achieve the best tradeoff among the noise reduction, speech intelligibility and computational complexity, a partial reconstruction scheme based on consecutive mean squared error is proposed to synthesize the low-frequency (LF) and high-frequency (HF) subbands. An iterative Kalman filter is then applied to the partially reconstructed HF subband speech. Finally, the enhanced HF subband speech is combined with the partially reconstructed LF subband speech to reconstruct the fullband enhanced speech. Experimental results show that the proposed subband iterative Kalman filter based algorithm is capable of reducing adverse environmental noises for a wide range of input SNRs. The overall performance of our method in terms of segmental SNR, perceptual evaluation of speech quality (PESQ) and computational cost is superior to several existing Kalman filter based algorithms.

In this paper, we study the problem of joint multiple-input multiple-output (MIMO) amplify-and-forward (AF) relaying and artificial noise (AN) optimization for secure communication between a source-destination pair in the presence of multiple eavesdroppers (eves). The eves' channel state information (ECSI) is subject to arbitrarily distributed random errors. Assuming that only the first and second moments of the ECSI errors are known, we introduce a probabilistically robust design method, which aims to maximize the received signal-to-interference-plus-noise ratio (SINR) at the destination while satisfying a set of robust intercept probability constraints. Since the resultant optimization problem is non-convex, we propose a solution approach by resorting to a duality-based method along with the semidefinite relaxation (SDR) technique, where a global optimal solution to our design problem can be found. Our simulation results demonstrate the improved secrecy of the proposed robust relaying design against eavesdropping and its robustness against the channel uncertainties.

In this paper, we propose a joint transceiver design algorithm for the full-duplex (FD) K-pair multiple- input multiple-output (MIMO) interference channel with simultaneous wireless information and power transfer (SWIPT). The aim is to minimize the total transmission power under both transmission rate and energy harvesting (EH) constraints. An iterative algorithm based on alternating optimization and with guaranteed monotonic convergence is proposed to successively optimize the transceiver coefficients. The algorithm consists of three main steps, aimed at successively optimizing: 1) the power splitting (PS) vectors of the EH nodes; 2) the receive beamforming vectors; 3) the transmit beamforming vectors.The first step is carried out based on concave-convex procedure (CCCP), the second step is based on the minimum mean square error (MMSE) criterion and the third step resorts to using semidefinite relaxation (SDR). Simulation results are presented to validate the effectiveness of the proposed algorithm.

Jointly detecting signals from the source and relay in a multi-input multi-output (MIMO) relaying system can achieve lower symbol error probability (SEP) and higher diversity order. In the literature, the best detector for spatial multiplexing decode-and-forward (DF) MIMO relaying systems is the near maximum likelihood (NML) decoder. However, both NML decoder and its variation are computationally intensive, especially when high-order modulations and multiple data streams are adopted. In order to develop a more efficient detection scheme, we propose a weighted combining (WC) algorithm which is applied before the final detector. The proposed algorithm merges the signal vectors from the source and relay without expanding their dimension and formulates an equivalent MIMO channel matrix for the combined signal, resulting in a much lower complexity for the subsequent detection. Simulation results show that by using the proposed WC algorithm with the ML detector, the same diversity gain as that of the more complex NML detection scheme can be achieved. In particular, in a symmetric network topology, the performance of the proposed WC algorithm is comparable to that of NML.

In this letter, we introduce a discriminative training algorithm of the basis vectors in the nonnegative matrix factorization (NMF) model for single-channel speech enhancement. The basis vectors for the clean speech and noises are estimated simultaneously during the training stage by incorporating the concept of classification from machine learning. Specifically, we consider the probabilistic generative model (PGM) of classification, which is specified by class-conditional densities, along with the NMF model. The update rules of the NMF are jointly obtained with the parameters of the class-conditional densities using the expectation-maximization (EM) algorithm, which guarantees convergence. Experimental results show that the proposed algorithm provides better performance in speech enhancement than the benchmark algorithms.

In this paper, we investigate a multiuser multiple-input single-output relay system with simultaneous wireless information and power transfer, where the received signal is divided into two parts for information decoding and energy harvesting (EH), respectively. Assuming that both base station (BS) and relay station (RS) are equipped with multiple antennas, we study the joint transceiver design problem for the BS beamforming vectors, the RS amplify-and-forward transformation matrix, and the power splitting (PS) ratios at the single-antenna receivers. The aim is to minimize the total transmission power of the BS and the RS under both signal-to-interference-plus-noise ratio and EH constraints. First, an iterative algorithm based on alternating optimization (AO) and with guaranteed convergence is proposed to successively optimize the transceiver coefficients. This AO-based approach is then extended into a robust transceiver design against norm bounded errors in channel state information (CSI), by using semidefinite relaxation and the S-procedure. Second, a novel design scheme based on switched relaying (SR) is proposed that can significantly reduce the computational complexity and overhead of the AO-based designs while maintaining a similar performance. In the proposed SR scheme, the RS is equipped with a codebook of permutation matrices. For each permutation matrix, a latent transceiver is designed, which consists of BS beamforming vectors, optimally scaled RS permutation matrix, and receiver PS ratios. For the given CSI, the optimal latent transceiver with the lowest total power consumption is selected for transmission. We propose concave-convex procedure-based and subgradient-type iterative algorithms, respectively, to design the latent transceivers under perfect and imperfect CSI. Simulation results are presented to validate the effectiveness of all the proposed algorithms.

We study the performance of diffusion least-mean squares algorithms for distributed parameter estimation in multi-agent networks when nodes exchange information over wireless communication links. Wireless channel impairments, such as fading and path-loss, adversely affect the exchanged data and cause instability and performance degradation if left unattended. To mitigate these effects, we incorporate equalization coefficients into the diffusion combination step and update the combination weights dynamically in the face of randomly changing neighborhoods due to fading conditions. When channel state information (CSI) is unavailable, we determine the equalization factors from pilot-aided channel coefficient estimates. The analysis reveals that by properly monitoring the CSI over the network and choosing sufficiently small adaptation step-sizes, the diffusion strategies are able to deliver satisfactory performance in the presence of fading and path loss.

We investigate the performance of distributed least-mean square (LMS) algorithms for parameter estimation over sensor networks where the regression data of each node are corrupted by white measurement noise. Under this condition, we show that the estimates produced by distributed LMS algorithms will be biased if the regression noise is excluded from consideration. We propose a bias-elimination technique and develop a novel class of diffusion LMS algorithms that can mitigate the effect of regression noise and obtain an unbiased estimate of the unknown parameter vector over the network. In our development, we first assume that the variances of the regression noises are known a priori. Later, we relax this assumption by estimating these variances in real time. We analyze the stability and convergence of the proposed algorithms and derive closed-form expressions to characterize their mean-square error performance in transient and steady-state regimes. We further provide computer experiment results that illustrate the efficiency of the proposed algorithms and support the analytical findings.

Multiple-input multiple-output (MIMO) amplify-and- forward (AF) relaying is designed for secure communication between a source-destination pair in the presence of multiple eavesdroppers. Assuming statistical knowledge of the eavesdroppers' channel state information (ECSI) errors, we introduce a probabilistically robust design method, which aims to optimize the source transmission power and AF relaying matrix by maximizing the received signal-to- interference-plus-noise ratio (SINR) at the destination, while satisfying a set of intercept probability constraints. The resultant optimization problem becomes nonconvex, and hence we propose a conservative two-step solution, where the source transmission power and the relaying matrix are sequentially optimized. Our simulation results demonstrate the improved secrecy of the proposed relaying design against eavesdropping and its robustness against the channel uncertainties.

This paper addresses the problem of secure amplify-and-forward (AF) relaying for multiple-input multiple output (MIMO) relaying networks in the presence of multiple eavesdroppers. Assuming practical imperfect eavesdroppers' channel state information (ECSI), we propose a robust approach to optimize the relay AF matrix, subject to power constraint, in order to maximize the received signal-to-interference-plus-noise ratio (SINR) at the destination while satisfying a set of secrecy constraints. The ECSI errors are assumed to fall within some predefined bounded sets. Since the resultant optimization problem is non-convex and semi-infinite, we transform it into a form constituted by the differences of convex functions (DC) using suitable matrix transformation techniques. Then an algorithmic solution with proven convergence is proposed by resorting to the penalty-DC algorithm (P-DCA). Experimental results show the security of the proposed transceiver design against eavesdropping and the robustness against the channel uncertainties.

In this paper, we present a new algorithm for the estimation of the noise power spectral density (PSD) matrix, as needed for multi-microphone speech enhancement in a general non-stationary noisy environment. First, we propose a recursive scheme for noise PSD estimation in which the current, previous and close subsequent noisy speech frames are properly weighted. The forgetting factor for the recursive updating of the smoothed PSD is obtained based on an overall measure of the SNR across all microphone signals. Since this SNR measure depends on the noise statistics, we choose to iteratively update it using the latest available estimate of the noise PSD matrix. Finally, to obtain better estimation accuracy in the proposed method, we further apply a direct extension of the minimum tracking approach to the estimated noise PSD matrix. Performance of the proposed algorithm is evaluated in terms of objective measures and its superiority is shown with respect to two recent noise PSD estimation methods in the context of speech enhancement.

Among the alternative multicarrier modulation techniques to orthogonal frequency division multiplexing (OFDM), a derivative of OFDM based on offset quadrature amplitude modulation (OFDM-OQAM) has been one of the most prominent to alleviate the sensitivity problem of the former to timing and frequency mismatch. In this paper, we propose an improved joint estimation method for carrier frequency offset (CFO), sampling time offset (STO) and channel impulse response (CIR) in OFDM-OQAM systems. The proposed method instruments a data-aided maximum-likelihood (ML) joint estimation of the unknown parameters, as derived under an assumption of Gaussian noise and independent input symbols by splitting the interference into pilot, non-pilot and noise terms. Performance evaluation is carried out through simulations by comparing the proposed method with a highly-cited previous work which considers all the three types of parameters in one development. The improvements in the results indicate the superiority of the proposed joint ML-based estimator.

In this paper, we consider the power splitting technique for multiple-input single-output (MISO) interference channel where the received signal is divided into two parts for information decoding and energy harvesting (EH) respectively. Specifically, assuming norm-bounded errors (NBE) in the channel state information (CSI), we study the robust joint beamforming and power splitting (JBPS) design problem, where the total transmission power is minimized subject to both signal-to-interference- plus-noise ratio (SINR) and EH constraints. We first propose an efficient approximation method based on semidefinite relaxation (SDR) for solving the highly non-convex JBPS problem, where the latter can be formulated as a semidefinite programming (SDP) problem. Then, a low complexity algorithm is proposed using EH relaxation and cutting-set philosophy, which partitions the original problem into an alternating sequence of optimization and worst-case analysis subproblems with guaranteed convergence. Finally, simulation results are presented to validate the robustness and efficiency of the proposed algorithms.

In this letter, we propose a semiblind channel estimation technique for OFDM/OQAM wireless systems. The proposed technique exploits the real property of transmitted symbols to blindly identify the channel-induced rotation in the received signal, thereby reducing the pilot overhead for estimation purpose. Specifically, the channel phase over each subcarrier can be obtained from the spatial-sign covariance matrix of the received signal, while the channel amplitude can be expressed in terms of the subcarrier power. In effect, the frequency-domain noise can be reduced effectively by block averaging over multiple symbols. Since the channel delay spread is usually much smaller than the symbol duration, channel coefficients obtained from the estimated phase and amplitude can be further refined through low-rank filtering. Simulation results validate the efficacy of the proposed technique in time-dispersive fading channels, showing its robustness under different signal-to-noise ratio (SNR) conditions.

In this work, we propose a novel adaptive reduced-rank receive processing strategy based on joint preprocessing, decimation and filtering (JPDF) for large-scale multiple-antenna systems. In this scheme, a reduced-rank framework is employed for linear receive processing and multiuser interference suppression based on the minimization of the symbol-error-rate (SER) cost function. We present a structure with multiple processing branches that performs a dimensionality reduction, where each branch contains a group of jointly optimized preprocessing and decimation units, followed by a linear receive filter. We then develop stochastic gradient (SG) algorithms to compute the parameters of the preprocessing and receive filters, along with a low-complexity decimation technique for both binary phase shift keying (BPSK) and M-ary quadrature amplitude modulation (QAM) symbols. In addition, an automatic parameter selection scheme is proposed to further improve the convergence performance of the proposed reduced-rank algorithms. Simulation results are presented for time-varying wireless environments and show that the proposed JPDF minimum-SER receive processing strategy and algorithms achieve a superior performance than existing methods with a reduced computational complexity.

We consider a cognitive radio (CR) network consisting of a secondary transmitter (ST), a secondary destination (SD) and multiple secondary relays (SRs) in the presence of an eavesdropper, where the ST transmits to the SD with the assistance of SRs, while the eavesdropper attempts to intercept the secondary transmission. We rely on careful relay selection for protecting the ST-SD transmission against the eavesdropper with the aid of both single-relay and multi-relay selection. To be specific, only the “best” SR is chosen in the single-relay selection for assisting the secondary transmission, whereas the multi-relay selection invokes multiple SRs for simultaneously forwarding the ST's transmission to the SD. We analyze both the intercept probability and outage probability of the proposed single-relay and multi-relay selection schemes for the secondary transmission relying on realistic spectrum sensing. We also evaluate the performance of classic direct transmission and artificial noise based methods for the purpose of comparison with the proposed relay selection schemes. It is shown that as the intercept probability requirement is relaxed, the outage performance of the direct transmission, the artificial noise based and the relay selection schemes improves, and vice versa. This implies a trade-off between the security and reliability of the secondary transmission in the presence of eavesdropping attacks, which is referred to as the security-reliability trade-off (SRT). Furthermore, we demonstrate that the SRTs of the single-relay and multi-relay selection schemes are generally better than that of classic direct transmission, explicitly demonstrating the advantage of the proposed relay selection in terms of protecting the secondary transmissions against eavesdropping attacks. Moreover, as the number of SRs increases, the SRTs of the proposed single-relay and multi-relay selection approaches significantly improve. Finally, our numerical results show that as expected, the multi-relay selection scheme achieves a better SRT performance than the single-relay selection.

This paper addresses the problem of carrier frequency synchronization and time-varying channel equalization in the uplink of a broadband multi-user wireless communication system employing an oversampled perfect reconstruction filter bank (OPRFB) transceiver structure for multi-carrier modulation. Based on the maximum likelihood (ML) principle, a pilot-aided joint estimator of the carrier frequency offsets (CFO) and channel equalizer coefficients of the multiple users is proposed. The performance of the new estimator is examined for various subband allocation schemes by means of numerical simulations under realistic conditions of operation. For mobile users with time-varying fading channels, we also study the effect of using different distributions of pilots over time. Our results show that the proposed estimator can provide accurate estimates of the unknown CFO and channel parameters, which in turn can be used to design effective compensation mechanisms.

Bayesian estimators of short time spectral amplitude (STSA) have received considerable attention in the field of speech enhancement. In this paper, we propose new multi-microphone extensions for the conventional Ephraim and Malah's speech spectral amplitude estimation method. Unlike the conventional estimators where the spectral phase is assumed to be uniformly distributed, the proposed extensions treat the latter as an unknown parameter to be estimated. It is shown that the proposed methods can exploit spectral phase estimates to improve the performance of the current speech STSA estimators and have the potential to provide even further improvement given a more accurate estimate of the spectral phase. Experimental results indicate the superiority of the new approaches in terms of noise reduction and speech distortion measures, in addition to the reduced computational complexity provided by the proposed minimum mean square method as compared to state-of-the-art solutions.

In this paper, cooperative localization of mobile nodes in non-line of sight (NLOS) situation is considered using a constrained square root unscented Kalman filter (CSRUKF). The NLOS measurements are used as quadratic constraints, which form a convex feasible region inside which the positions of the mobile nodes are supposed to be. The CSRUKF consists of two main stages: square root unscented Kalman filter (SRUKF) and sigma point projection. In the former, a conventional SRUKF is used to estimate the state vector and the Cholesky factor of the error covariance matrix. In the latter, a new set of sigma points are generated, and the ones violating the constraints are projected onto the feasible region by solving a set of convex quadratically constrained quadratic programs (QCQP). Each QCQP can be solved independently and in parallel for each sigma point violating the constraint, thus the algorithm is suitable for distributed processing. The simulation results show that our algorithm can perform well in different NLOS scenarios.

In this study, we propose a robust transceiver scheme with switched preprocessing (SP) for K-user multiple-input multiple-output (MIMO) interference channels. The channel state information (CSI) available is assumed to be imperfect under norm-bounded errors (NBE). Each transmitter is provided with a codebook of permutation matrices, so that each arrangement of permutation matrices among the K transmitters will generate a group of K parallel transceivers. The optimum transceiver group within the class of all possible such groups is chosen by a suitable selection mechanism for data transmission. To design each transceiver group, we adopt a worst-case design approach to minimize the maximum per user MSE. We show that the proposed transceiver design problem can be partitioned into an alternating sequence of optimization and worst-case analysis subproblems, which involves solving Second-Order Cone Programming (SOCP) problems. Simulation results show that the performance of the proposed SP-based transceiver is significantly better than existing methods in the presence of imperfect CSI.1.

In this work, we propose a transceiver strategy with switched preprocessing (SP) for interference suppression in K-pair multiple-input multiple-output (MIMO) interference channels. Each transmitter is equipped with a codebook of permutation matrices. For the given MIMO interference channel, all the combinations of permutation matrices among the transmitters can create a number of parallel transceivers. Based on the given channel state information (CSI) and a block of transmit symbols, the optimum transceiver branch is chosen by a suitable selection criterion for transmission. For each branch, we introduce a robust transceiver design algorithm based on minimizing the mean square error (MSE) criterion. The selection criterion is designed to minimize the Euclidean distance between the true transmit symbol vector and the pre-estimated noiseless received vector. Simulation results show that the performance of the proposed technique is significantly better than prior art in the case of imperfect CSI.

In SC-FDMA systems the bit error rate (BER) is very sensitive to channel estimation errors. We propose the use of blind (or semi-blind) subspace decomposition to estimate the channel frequency responses between the transmitter and the multiple antennas of the receiver in the uplink of a 4G system employing SC-FDMA. The proposed subspace- based channel estimation technique requires very little overhead in terms of pilot symbols dedicated to channel estimation. Furthermore, we show through simulations that, when applied to SC-FDMA, it can provide a BER performance comparable to a system with perfect channel estimates.

This paper addresses the problem of amplify-and- forward (AF) relaying for multiple-input multiple-output (MIMO) multiuser relay networks. where each source transmits multiple data streams to its corresponding destination with the assistance of multiple relays. Assuming only imperfect channel state information (CSI) of all the source-relay and relay-destination links, we propose a robust approach to jointly design the source and relay precoders and the receive filters, in which the worst per-stream mean square error (MSE) is minimized subject to source and relay power constraints. The channel uncertainties are assumed to be Gaussian distributed and the well-known Kronecker model is employed to characterize the spatial correlations in the proposed design. The resultant optimization problem is nonconvex and therefore, an algorithmic solution with proven convergence is proposed by resorting to the iterative block coordinate update approach along with matrix transformation and convex conic optimization techniques. Simulation results show that the proposed joint transceiver design can achieve an improved robustness against the channel uncertainties when compared to the non-robust approaches.

Recently, DFT-based oversampled perfect reconstruction filter banks (OPRFB), as a special form of filtered multitone, have shown great promises for applications to multicarrier modulation. Still, accurate frequency synchronization and channel equalization are needed for their reliable operation in practical scenarios. In this paper, we first derive a data-aided joint maximum likelihood (ML) estimator of the carrier frequency offset (CFO) and the channel impulse response (CIR) for OPRFB transceiver systems operating over frequency selective fading channels. Then, by exploiting the structural and spectral properties of these systems, we are able to considerably reduce the complexity of the proposed estimator through simplifications of the underlying likelihood function. The Cramer Rao bound on the variance of unbiased CFO and CIR estimators is also derived. The performance of the proposed ML estimator is investigated by means of numerical simulations under realistic conditions with CFO and frequency selective fading channels. The effects of different pilot schemes on the estimation performance for applications over time-invariant and mobile time-varying channels are also examined. The results show that the proposed joint ML estimator exhibits an excellent performance, where it can accurately estimate the unknown CFO and CIR parameters for the various experimental setups under consideration.

We present a joint estimator of the time of arrival (TOA) and angle of arrival (AOA) for impulse radio ultrawideband (UWB) systems in which an antenna array is employed at the receiver. The proposed method consists of two steps: 1) preliminary estimation of the TOA and the average power delay profile (APDP) using energy-based threshold crossing and log-domain least-squares fitting, respectively; and 2) joint TOA refinement and AOA estimation by local 2-D maximization of a log-likelihood function (LLF) that employs the preliminary estimates from the first step. The derivation of the LLF relies on an original formulation in which the superposition of images from secondary paths is modeled as a Gaussian random process, whose second-order statistical properties are characterized by a wideband space-time correlation function. In addition to the APDP, this function incorporates a special gating mechanism to represent the onset of the secondary paths, thereby leading to a novel form of the LLF. Closed-form expressions for the Cramer-Rao bound on the variance of the TOA and AOA estimators are also derived, which formally take into account pulse overlap through this gating mechanism. In simulation experiments based on multipath UWB channel models featuring both diffuse and directional image fields, our approach exhibits superior performance to that of a competing scheme from the recent literature.

We propose a distributed least-mean squares (LMS) procedure based on a diffusion strategy for localization and tracking of mobile terminals in cellular networks. In the proposed algorithm, collaborating base stations measure two sets of parameters, namely, the received signal strength (RSS) and the signal propagation time (SPT) to estimate mobile locations. The proposed algorithm has a simple operational structure, offers agile tracking performance and helps the network to save energy and radio resources by benefiting from its decentralized and adaptive signal processing features.

A novel joint TOA and AOA estimator is proposed for impulse radio Ultra-Wideband (IR-UWB) systems, in which a uniform linear array of antennas is employed at the receiver. The proposed method consists of two steps: (1) coarse estimation of the TOA and the average power delay profile; (2) joint TOA refinement and AOA estimation by maximization of a novel log likelihood function (LLF) using the coarse estimates from the first step. The derivation of the LLF is based on an original approach in which the pulse image from the primary path is modeled as a deterministic component while the superposition of the images from the secondary paths is modeled as a Gaussian random process. In addition, a special gating mechanism is used to characterize the secondary paths, thereby leading to a previously unknown form of the LLF in step (2). According to simulation experiments based on standard UWB channel models, our approach exhibits superior performance to that of a competing scheme from the recent literature.

In this paper, we study the linear transceiver design for the downlink of a cellular network assisted by a multi-antenna relay. A diagonal scaling scheme is proposed in which multiple single-antenna users apply different complex-valued scaling to their signals before decoding, as represented by an equivalent diagonal equalizer matrix. This equalizer is designed together with a linear precoder at the base station (BS) and a linear processing matrix at the relay. The objective is to minimize the weighted minimum mean square error (MMSE) between the precoder input and the equalizer output, subject to power constraints at the BS and the relay. In particular, the optimal relaying matrix is first derived in closed form as a function of the precoder and the equalizer. The latter two can then be jointly designed in an efficient iterative manner. Simulation results demonstrate lower bit-error rates (BERs) than previous design methods.

We propose a modified diffusion strategy for parameter estimation in sensor networks where nodes exchange information over fading wireless channels. We show that the effect of fading can be mitigated by incorporating local equalization coefficients into the diffusion process. We explain how the equalization coefficients are chosen and show that the (mean) stability of the network continues to be insensitive to the choice of the combination weights and to the network topology. Our computer experiments demonstrate that the performance of the modified diffusion algorithm in fading scenario is nearly identical to that of centralized least-mean square (LMS) with equalized input data.

Multiuser multiple-input multiple-output (MU-MIMO) wireless systems that employ single-carrier frequency-domain equalization (SC-FDE) for uplink transmissions can provide high data rates with increased spectral efficiency under severe channel conditions. However, the availability of accurate channel estimates at the base station (BS) receiver is crucial for achieving peak performance. In this paper, we investigate the use of subspace decomposition and derive a novel algorithm for the blind estimation of MU-MIMO channels in SC-FDE systems, as specified by the 3GPP LTE. By exploiting the long data blocks available in LTE standards, our proposed blind algorithm can obtain accurate estimates of the MU-MIMO channels over every block of transmitted data. This provides for a bandwidth-efficient solution in SC-FDE systems by eliminating (reducing) the need to allocate an entire block of pilot sequences for each active transmitter. Furthermore, since the channel estimation is deployed at the BS, the computational complexity is not considered to be a significant burden for future systems in exchange for the increased spectral efficiency. The results of simulations over fading channels, using realistic system parameters representative of LTE-Advanced, demonstrate the advantages of our proposed blind subspace-based channel estimation algorithm and support the feasibility of the resulting MU-MIMO SC-FDE scheme with reduced training.

Single carrier frequency division multiple access (SC-FDMA), a multiple user access scheme based on the single carrier frequency domain equalization (SC-FDE) technique, has been proposed for the uplink in fourth generation (4G) mobile communications. SC- FDE requires reliable channel estimates to maintain an acceptable bit error rate (BER) performance. Much research focuses on the use of pilot symbols which reduces the spectral efficiency of SC-FDE systems. In this paper, we propose a subspace decomposition approach for semi-blind channel estimation in SC-FDE systems as a means to reduce or eliminate the need for pilot symbols and increase their overall spectral efficiency. Simulation results show that, under the conditions presented in this paper, a BER of 10-3can be achieved with a power loss of roughly 1 dB compared to a system with perfect channel estimates. This result is achieved with channel estimates that have a normalized mean square error (NMSE) less than 1%.

In this work, we propose a low-complexity adaptive transceiver algorithm for the K-pair multiple-input multiple-output (MIMO) interference channels. The proposed algorithm is based on the joint optimization of transmit and receive vectors using the constrained constant modulus (CCM) criterion. We firstly derive CCM-based expressions for the transmit and receiver vectors. Then, we develop recursive least-squares (RLS) adaptive algorithms for their efficient implementation. Unlike earlier block-based transceivers for MIMO interference channels, the proposed algorithms have low computational complexity and can track the time-varying channels and interference as changes occur in the surrounding wireless environment. In particular, simulation results show that the proposed adaptive algorithms achieve the performance of the Sum-MSE algorithm at a much reduced complexity.

In full-duplex communication systems with discrete multi-tone (DMT) modulation, echo cancellers are employed to cancel echo by means of adaptive filters. Generally, the structure present in the DMT signals is used to decrease the computational complexity of these cancellers by splitting the operations between the time and frequency domains. In this work, we introduce a general framework for designing echo cancellers for such systems in an arbitrary mixed domain. This is achieved by introducing a generic decomposition of the Toeplitz data matrix at the transmitter in terms of arbitrary unitary matrices. Then, based on this decomposition, a new mixed-domain echo cancellation structure is derived, which performs an exact instantaneous gradient-type adaptation. This mixed-domain configuration is also extended for realizing constrained adaptation whereby linear constraints are used to ensure the proper mapping of the weight vectors in different domains. The proposed structures offer a unified framework to study existing cancellers and to design new ones with better performance measures. This framework is employed to propose a new canceller based on discrete trigonometric transformations. The analytical and numerical results presented show that this canceller has a faster convergence rate than the existing ones with similar complexity and is more robust.

In this paper, we present an iterative spectrum sensing scheme for multiantenna assisted cognitive radio (CR) using the expectation-maximization (EM) algorithm. Considering a wideband frequency spectrum, the secondary user (SU) performs an EM-based joint estimation and detection (JED), where the channel coefficients and noise variance are estimated jointly with the primary user (PU) signal variance. We also provide a semi-analytical evaluation of the proposed scheme using the Neyman-Pearson criterion. Compared with the conventional Generalized Likelihood Ratio detector (GLRD), the EM-based JED scheme enhances the detection process of the multiple antenna CR with few iterations and modest complexity.

Sub-Nyquist maximum likelihood (ML)-based time of arrival (TOA) estimation methods for ultra-wideband (UWB) signals normally assume a priori knowledge of the UWB channel in the form of the average power delay profile (APDP). In practice however, and despite its importance, the APDP is not always available. To address this issue, we develop in this paper a joint estimator of TOA and APDP. Knowing that the APDP of a UWB channel usually consists of several clusters, each with specific exponential decay rate, a parametric APDP model of this type is employed. The parameters of this model are estimated via a least-squares fitting approach; then the estimated APDP is used to form a likelihood function and obtain a ML estimator of the TOA. Simulations show that the TOA estimated jointly in this way achieves a good accuracy in practical scenarios. The proposed APDP estimate can also help to boost the performance of previously reported TOA estimators that assume a priori APDP knowledge, although the proposed ML scheme generally offers superior performance.

This paper considers a cooperative MIMO relaying system in which the source sends information to the destination with the aid of multiple relays, each equipped with multiple antennas. We design optimal linear relaying matrices that minimize the mean square error between the source and the received signal vectors under a total transmitted power constraint imposed on all the relays. The optimal matrices are obtained by solving the corresponding Karush-Kuhn-Tucker (KKT) conditions. Simulation results show significant advantages of the newly designed relaying matrices over other competing approaches in terms of bit-error rate (BER) performance. Moreover, this strategy enables the self-demultiplexing of the spatial substreams without the need for a MIMO combiner at the destination, so that a simplified receiver structure can be used without performance loss. In addition, the new design can serve as a suboptimal solution for multiuser MIMO relaying applications.

Spectrum sensing is an important functionality of cognitive radio as a means to detect the presence or absence of an existing user (EU), including the primary user or other secondary users, in a certain spectrum band. Energy detector (ED) is a widely used spectrum sensing technique based on the assumption that the EU is either present or absent during the whole sensing period. However, this assumption is not realistic in a dynamic environment where the EU could appear or disappear at any time. The performance of the conventional ED actually deteriorates in the scenario where the EU activity status changes during the sensing period. In this study, an adaptive ED is proposed to improve the detection performance in such dynamic environments. Analytical performance evaluation of the proposed adaptive ED along with simulation results prove its superiority over the conventional ED.

In this letter, we consider wideband spectrum sensing in the presence of correlation between the occupancies of frequency subbands. We begin by formulating the maximum a posteriori (MAP) estimator of channel occupancy based on measurements from multiple frequency subbands. Since the MAP estimator's complexity grows exponentially with the number of subbands, we propose an alternative structure, in which the subband energy measurements are linearly combined according to a minimum mean-square error (MMSE) criterion to form a sufficient statistic for binary detection in each subband. Through analysis and numerical simulations, we show that the proposed frequency-coupled detector can significantly outperform the traditional decoupled one.

In DMT-based communication systems where full-duplex transmission is required, digital echo cancellers are employed to cancel echo by means of adaptive filters. In order to reduce the computational complexity of these cancellers, the structure of the Toeplitz matrix containing the transmitted signal is usually exploited to transform the time domain signals and perform the emulation and adaptive update in a more convenient domain (e.g. frequency domain). In this paper, we consider a recently proposed dual transform domain echo canceller, which is based on the general decomposition of the data Toeplitz matrix. A comprehensive comparative performance evaluation of the proposed method with the existing methods is provided. This evaluation includes the comparison of the convergence curves and computational cost of the algorithms. The comparison shows that the proposed canceller achieves a faster convergence with a low error floor with no increase in the complexity.

In Bayesian short-time spectral amplitude (STSA) estimation for speech enhancement, the spectral components are traditionally assumed uncorrelated. However, this assumption is inexact since some correlation is present in practice. We thus investigate a multi-dimensional STSA estimator that assumes correlated frequency components. Since the closed-form solution of this optimum estimator is not readily available, we previously derived closed-form expressions for an upper and a lower bound on the desired estimator. In this paper, we study the proximity between the upper and the lower bounds and propose a new family of estimators that are derived from these bounds and characterized by a scalar parameter 0 ≤ γ ≤ 1, with γ = 0 corresponding to the lower bound and γ = 1 to the upper bound. Experimental results show that the proposed estimators achieve a better performance than existing estimators, especially at high SNR.

In this paper, we consider a wireless communication scenario with multiple source-destination pairs communicating through several cooperative amplify-and-forward relay terminals. The relays are equipped with multiple antennas that receive the source signals and transmit them to the destination nodes. We develop two iterative relay beamforming algorithms that can be applied in real-time. In both algorithms, the relay beamforming matrices are jointly designed by minimizing the received power at all the destination nodes while preserving the desired signal at each destination. The first algorithm requires the existence of a local processing center that computes the beamforming coefficients of all the relays. In the second algorithm, each relay can compute its beamforming coefficients locally with the help of some common information that is broadcasted from the other relays. This is achieved at the expense of enforcing the desired signal preservation constraints non-cooperatively. We provide two extensions of the proposed algorithms that allow the relays to control their transmission power and to modify the quality of service provided to different sources. Simulation results are presented validating the ability of the proposed algorithms to perform their beamforming tasks efficiently and to track rapid changes in the operating environment.

A new method is proposed for the blind subspace-based identification of the coefficients of time-varying (TV) single-input multiple-output (SIMO) FIR channels. The TV channel coefficients are represented via a finite basis expansion model, i.e. linear combination of known basis functions. In contrast to earlier related works, the basis functions need not be limited to complex exponentials, and therefore do not necessitate the a priori estimation of frequency parameters. This considerably simplifies the implementation of the proposed method and provides added flexibility in applications. The merits of the proposed technique, including asymptotic consistency, are demonstrated by numerical simulations.

In communication systems where full-duplex transmission is required, digital echo cancellers are employed to cancel echo by means of adaptive filtering. In order to reduce the computational complexity of these cancellers, the structure of the Toeplitz matrix containing the transmitted signal is usually exploited to transform the time domain signals and perform the emulation and adaptive update in a more convenient domain (e.g. frequency domain). In this paper, we consider a general decomposition of the Toeplitz matrix and examine the effect of different components of the decomposition on the computational complexity and convergence behaviour of the canceller. Based on this general decomposition, a new dual transform domain canceller is proposed which has improved convergence compared to the current echo cancellers and also does not require the transmission of dummy data on the unused tones.

This paper presents a blind subspace-based tracking scheme for precoded MIMO-OFDM systems over rapidly time-varying wireless channels. Subspace-based tracking is normally considered for slow time-varying channels only. Thanks to the frequency correlation of the wireless channels, the proposed scheme is able to collect data not only from the time but also from the frequency domain to speed up the update of the required second-order statistics. After each update of the statistics, the subspace information is also updated using orthogonal iteration, and then a new channel estimate is computed. The proposed algorithm is verified in 3GPP-SCM suburban macro scenario, in which a mobile station is allowed to move in any direction with a constant speed of 100 km/h. The simulation results show that the root mean square error of the channel estimates converges to the level of 2 times 10-2within less than 5 OFDM symbols even for such a high Doppler rate.

In this paper, we consider a cooperative relaying scenario with multiple sources transmitting to one or more destination nodes through several relay terminals. Each relay is equipped with multiple receive and transmit antennas. We assume that the relays can estimate their uplink (relay-destination) channels with enough accuracy and that they have access to the training sequences transmitted by the sources. We present two adaptive training-based algorithms for multiuser relay beamforming. Both algorithms use Kalman filtering to estimate the beamforming matrices iteratively. The first algorithm is centralized where the relay terminals forward their received data to a processing center that computes the beamforming coefficients and feeds them back to the relays. In the second algorithm, each relay terminal can estimate its beamforming matrix locally using its received data and some common information that is broadcasted by the other relays. We present numerical simulations that validate the good performance of the proposed beamforming algorithms in stationary and nonstationary signal environments.

In this letter, we show that many existing short-time spectral amplitude (STSA) Bayesian estimators for speech enhancement all have a similarly structured cost function. On this basis, we propose a new cost function that generalizes those of existent Bayesian STSA estimators and then obtain the corresponding closed-form solution for the optimal clean speech STSA. The resulting family of estimators, which we will term the generalized weighted family of STSA estimators (GWSA), features a new parameter that acts only on the estimated clean speech STSA. It is found that this new parameter yields an added flexibility in terms of achievable gain curves when compared to those of existing estimators. Moreover, we show that the new estimator family tends to a Wiener filter for high instantaneous signal-to-noise ratios.

In this paper, we develop a beamforming algorithm for multiuser MIMO-relaying wireless systems. We consider a relaying scenario with multiple sources transmitting to one or more destination nodes through several relay terminals. Each relay is equipped with multiple antennas. We jointly design the beamforming matrices of the cooperating relays by minimizing both the noise received at each destination node and the interference caused by the sources not targeting this node. We impose additional constraints that preserve the received signal from each source at its targeted destination node. The relay beamforming problem is shown to be a convex optimization problem and is formulated as a second-order cone program that can be efficiently solved using interior point methods. Numerical simulations are presented showing the superior performance of our beamforming technique compared to previously proposed zero forcing relay beamforming.

In this paper, we first propose a new family of Bayesian estimators for speech enhancement where the cost function includes both a power law and a weighting factor. Secondly, we set the parameters of the estimator based on perceptual considerations by taking into account the masking properties of the ear and the perceived loudness of sound. Our results show that the new estimator achieves better overall performance than existing Bayesian estimators both in terms of objective and subjective measures. Specifically, it shows a segmental SNR improvement of up to 0.65 dB while it obtains the best scores in a MUSHRA test for both white and aircraft cockpit noises.

Among all blind channel estimation problems, subspace-based algorithms are attractive due to its fast- converging nature. It primarily exploits the orthogonality structure of the noise and signal subspaces by applying a signal-noise space decomposition to the correlation matrix of the received signal. In practice, the correlation matrix is unknown and must be estimated through time averaging over multiple time samples. To this end, the wireless channel must be time-invariant over a sufficient time interval, which may pose a problem for wideband applications. We proposed a novel subspace-based blind channel estimation algorithm with short time averaging periods, as obtained by exploiting the frequency correlation among adjacent OFDM subcarriers. In this paper, asymptotic performance bounds of the proposed algorithm are investigated by using perturbation analysis. We also present numerical results of the proposed as well as referenced subspace-based methods, including cyclic prefix and virtual carriers approaches. Based on the asymptotic performance bounds, the proposed scheme is justified in obtaining a desired correlation matrix efficiently by reducing the number of the OFDM blocks for time averaging up to 85%.

In this paper, we investigate the noise robustness of Wang and Shamma's early auditory (EA) model for the calculation of an auditory spectrum in audio classification applications. First, a stochastic analysis is conducted wherein an approximate expression of the auditory spectrum is derived to justify the noise-suppression property of the EA model. Second, we present an efficient fast Fourier transform (FFT)-based implementation for the calculation of a noise-robust auditory spectrum, which allows flexibility in the extraction of audio features. To evaluate the performance of the proposed FFT-based auditory spectrum, a set of speech/music/noise classification tasks is carried out wherein a support vector machine (SVM) algorithm and a decision tree learning algorithm (C4.5) are used as the classifiers. Features used for classification include conventional Mel-frequency cepstral coefficients (MFCCs), MFCC-like features obtained from the original auditory spectrum (i.e., based on the EA model) and the proposed FFT-based auditory spectrum, as well as spectral features (spectral centroid, bandwidth, etc.) computed from the latter. Compared to the conventional MFCC features, both the MFCC-like and spectral features derived from the proposed FFT-based auditory spectrum show more robust performance in noisy test cases. Test results also indicate that, using the new MFCC-like features, the performance of the proposed FFT-based auditory spectrum is slightly better than that of the original auditory spectrum, while its computational complexity is reduced by an order of magnitude.

We propose a new family of Bayesian estimators for speech enhancement where the cost function includes both a power law and a weighting factor. The parameters of the cost function, and therefore of the corresponding estimator gain, are chosen based on characteristics of the human auditory system, namely, the compressive nonlinearities of the cochlea, the perceived loudness and the ear's masking properties. It is found that choosing the parameters in this way results in a decrease of the estimator gain at high frequencies. This frequency dependence of the gain improves the noise reduction while limiting the speech distortion. Experimental results show that the new estimators achieve better enhancement performance than existing Bayesian estimators such as those based on the minimum mean-square error (MMSE) of the short-time spectral amplitude (STSA), the MMSE of the logarithm of the STSA (LSA) or the weighted euclidien (WE) error, both in terms of objective and subjective measures.

In this paper, we present an improved implementation for an auditory-inspired FFT-based model which calculates a noise-robust FFT spectrum. Through the use of characteristic frequency (CF) values of the cochlear filters in an early auditory (EA) model for power spectrum selection, and the use of a pair of running averages for the implementation of self-normalization, the proposed FFT model allows more flexibility in the extraction of audio features. To evaluate the performance of the proposed FFT model, a speech/music/noise classification task is carried out wherein a decision tree learning algorithm (C4.5) is used as the classifier. Audio features used for classification include the mel-frequency cepstral coefficient (MFCC) features, a set of conventional spectral features, and spectral features calculated using the proposed FFT model. Compared to the conventional MFCC and spectral features, the spectral features based on the proposed FFT model show more robust performance in noisy test cases.

Subspace-based blind channel estimation primarily exploits the orthogonality structure of the noise and signal subspaces by applying a signal-noise space decomposition to the correlation matrix of the received signal. In practice, the correlation matrix is unknown and must be estimated through time averaging over multiple symbol blocks. To this end, the wireless channel must be time-invariant over a sufficient time interval, which may pose a problem for wideband applications. In this paper, we propose a novel subspace-based blind channel estimation algorithm with a reduced time averaging interval, as obtained by exploiting the frequency correlation among adjacent OFDM subcarriers. We present simulation results of the proposed as well as referenced subspace-based methods, including Cyclic Prefix and Virtual Carriers approaches, and show that the proposed scheme is able to obtain a desired correlation matrix by reducing the number of the OFDM blocks for time averaging up to 85 %.

To reduce the complexity of space-time multiuser detection, it has been proposed recently to exploit the spatial dimension by forming groups of users and apply the detection individually to each group. In this work we propose a new space-time receiver structure based on the group-optimal MMSE linear detector along with a new grouping algorithm that respects practical hardware limitations. Furthermore, an extension of the proposed structure which allows non-mutually exclusive grouping is presented. The simulation results show that the proposed reduced-complexity receiver structure provides a bit error rate (BER) performance close to the full linear MMSE multiuser detector.

Mobile communication systems based on DS-CDMA suffer from multiple access interference (MAI), which limits the system capacity. Several techniques, such as beamforming with multiple antennas and multiuser detection (MUD), have proved to be effective to mitigate MAI. To reduce the complexity of MUD, it has been proposed to process users in smaller groups. The interference between group is then reduced using beamforming with smart antennas. In this paper, a new group-based block linear successive interference cancellation structure is developed. The structure is shown to provide BER performance close to the full space-time MUD at a lower computational cost.

We derive a block linear MMSE equalizer suitable for zero padded multicarrier systems with insufficient guard length. The proposed equalizer is not limited to systems using the IFFT/FFT pair (like in DMT or OFDM), but to any multicarrier system that can be modelled as a pair of perfect reconstruction filter banks. As the computational complexity of the equalizer can be high, an efficient implementation based on the Cholesky factorization is developed. Compared to the traditional zero-forcing approach, experiments show that in the presence of interblock interference (IBI), the proposed equalizer performs significantly better, regardless of the signal-to-noise ratio (SNR). In the absence of IBI (i.e. when the guard interval is sufficiently long), improvements have also been observed at low SNR

Recently, an early auditory model (K. Wang and S. Shamma, 1994) that calculates a so-called auditory spectrum, has been employed in audio classification where excellent performance is reported along with robustness in noisy environment. Unfortunately, this early auditory model is characterized by high computational requirements and the use of nonlinear processing. In this paper, inspired by the inherent self-normalization property of the early auditory model, we propose a simplified FFT-based spectrum which is noise-robust in audio classification. To evaluate the comparative performance of the proposed FFT-based spectrum, a three-class (i.e., speech, music and noise) audio classification task is carried out wherein a support vector machine (SVM) is employed as the classifier. Compared to a conventional FFT-based spectrum, both the original auditory spectrum and the proposed self-normalized FFT-based spectrum show more robust performance in noisy test cases. Test results also indicate that the performance of the self-normalized FFT-based spectrum is close to that of the original auditory spectrum, while its computational complexity is significantly lower

The increase in the demand for voice and data wireless services creates a need for a more efficient use of the available bandwidth. Current and future generations cellular systems based on DS-CDMA are known to be interference- limited. Several approaches exist to mitigate the multiple access interference including multiuser detection (MUD). Group-based techniques have been proposed to reduce the complexity of the MUD and have been shown to provide a performance-complexity tradeoff between match filtering and full MUD. In this work, we propose to reduce the inter-group interference (IGI), a limiting factor in group-based systems, using linear parallel interference cancellation (PIC). The complete equivalent matrix filter is derived and conditions for its convergence are discussed. The numerical results show that the proposed technique is effective against IGI, at a reduced computational cost.

In this paper, we consider the problem of blind source separation (BSS) applied to speech signals. Due to reverberation, BSS in the time domain is usually expensive in terms of computations. We propose in this paper a subband BSS system based on the use of adaptive feedback de-mixing networks in an oversampled uniform DFT filter bank structure. We show that the computational cost can be significantly decreased if BSS is carried out in subbands due to the possibility of reducing the sampling rate. Experiments with real speech signals, conducted with two-input two-output BSS systems using oversampled 32-subband and fullband adaptation, indicate that separation quality and distortion are similar for both systems. However, the proposed subband system is more than 10 times computationally faster than the fullband one.

A low-complexity adaptive blind subspace channel estimation algorithm is proposed for direct sequence spread spectrum CDMA systems. Compared with so-called hybrid adaptive channel estimation algorithms, where only the subspace estimation is carried out adaptively, the proposed algorithm is fully adaptive in that both subspace and channel estimates are updated recursively. The new algorithm is derived by exploiting common structural properties of plane rotation-based subspace trackers (e.g. Proteus, RO-FST, etc.). It is characterized by a low-complexity of implementation and numerical robustness over long periods of operation, an essential requirement for wireless radio applications. Moreover, we find in the case of a heavily loaded system that the proposed algorithm has better performance than previous hybrid algorithms.

The design of a shared, centralized acoustic echo canceller (AEC) for use in modem digital communication networks faces new challenges. Specially, the performance of conventional approaches for AEC is severely degraded by vocoder nonlinearities along the transmission chain. In this paper, based on the analysis of the nonlinear echo path and the performance of the Wiener-type post-filter in the presence of vocoders, we propose a centralized AEC which incorporates a psychoacoustic post-filter. Computer experiments show that the proposed AEC is very promising for practical use in terms of high acoustic echo suppression, robust performance and ease of implementation.

The major drawback of most noise reduction methods in speech applications is the annoying residual noise known as musical noise. A potential solution to this artifact is the incorporation of a human hearing model in the suppression filter design. However, since the available models are usually developed in the frequency domain, it is not clear how they can be applied in the signal subspace approach for speech enhancement. In this paper, we present a Frequency to Eigendomain Transformation (FET) which permits to calculate a perceptually based eigenfilter. This filter yields an improved result where better shaping of the residual noise, from a perceptual perspective, is achieved. The proposed method can also be used with the general case of colored noise. Spectrogram illustrations and listening test results are given to show the superiority of the proposed method over the conventional signal subspace approach.

The major drawback of most noise reduction methods is what is known as musical noise. To cope with this problem, the masking properties of the human ear were used in the spectral subtraction methods. However, no similar approach is available for the signal subspace based methods. In a previous work, we presented a frequency to eigendomain transformation which provides a way to calculate a perceptually based upper bound for the residual noise. This bound, when used in the signal subspace approach, yields an improved result where better shaping of the residual noise is achieved. In this paper, we further improve this method and provide an easy way to generalize it to the colored noise case. Listening tests results are given to show the superiority of the proposed method.

We extend the single microphone signal subspace approach for speech enhancement, to a multi-microphone design. In the single microphone case, the tradeoff between speech quality and intelligibility is an handicap which limits its performance. This is because it is based on a linear speech model which does not usually offer enough degrees of freedom for noise reduction. In our method, we show how we can easily, and with comparable computational complexity, get more degrees, of freedom by using signals from more than one microphone. Experimental results show that this leads to improvements in the noise reduction performance.

In this paper we present a subband scheme to decrease the computational complexity of simulating room responses to acoustic signals, important for instance in microphone array systems. Besides, the new method offers added flexibility to the well known image method by allowing to choose the reflection coefficients of every frequency subband independently of each others. The efficiency of this method is tested experimentally.

In this paper, the blind estimation of wireless CDMA receiver coefficients from the second order statistics of the signals is considered. Although many algorithms have been proposed so far, their performance analysis has always been carried out assuming perfect receiver coefficients estimation and/or under time-invariant conditions. In this article, we present some decision-directed blind algorithms and use a time-varying vector channel simulator to compare their performance with those of many previously proposed algorithms. It is shown that decision-directed chip-level algorithms can operate without the use of training sequences to avoid catastrophic error propagation, and that one should not expect an increase in performance from using least squares instead of least-mean-square. Furthermore the unpracticability of the bit-level algorithm (Gerlach 1992) and of the least significant algorithm (Liu and Zoltowski 1997) under a time varying environment is outlined. The performance of the principal component (Stanford) algorithm (Naguib and Paulraj 1995) is also studied.

In this paper, we analyse the issue of efficiently using Givens rotations to perform a more accurate SVD-based subspace tracking. We propose an alternative type of decomposition which allows a more versatile use of Givens rotations. We also show the direct effect of the latter on the tracking error, and develop a cross-terms cancellation concept which leads to a class of high performance algorithms with very low complexity: O(N/sup 2/) if signal and noise subspaces are tracked, O(Nr) if only the signal subspace is tracked, where N is the data vector dimension, and r the number of sources. Comparative simulation experiments support the theoretical work.

We present a new, computationally efficient simulator for time-varying multi-path (fast fading) vector channel that can be used to evaluate the performance of antenna array wireless receivers. The development of the simulator is based on the emulation of the spatio-temporal correlation properties of the vector channel. The channel is modeled as a multi-channel FIR system with time-varying coefficients which are obtained via the application of a space-time correlation shaping transformation on some independent random sequences. The various parts of the new simulator are detailed and channel simulation realizations are presented and commented.

In this work, we derive new algorithms for tracking the eigenvalue decomposition (EVD) of a time-varying data covariance matrix. These algorithms have parallel structures, low operation counts and good convergence behavior. Their main feature is the use of Givens rotations to update the eigenvector estimates. As a result, orthonormality of the latter can be maintained at all time, which is critical in the application of certain signal-subspace methods. The comparative performance of the new algorithms is illustrated by means of computer experiments.

Synthetic microphone signals generated with the image model technique are used to study the effects of room reverberation on the performance of the maximum likelihood (ML) estimator of the time delay, in which the estimate is obtained by maximizing the cross correlation between filtered versions of the microphone signals. The results underscore the adverse effects of reverberation on the bias, variance and probability of anomaly of the ML estimator. Explanations of these effects are provided.

Time delay estimation (TDE) between the signals received by two or more spatially separated microphones can be used as a means for the passive localization of the dominant talker in applications such as audio-conference. However, in a recent study, it has been shown that reverberation can have disastrous effects on TDE performance. In this paper, we develop and evaluate a new cepstral prefiltering technique which can be applied on the microphone signals before the actual TDE in order to obtain a more accurate estimate of the position of a source in a typical reverberant environment. The performance of a TDE system with and without cepstral prefiltering is investigated under controlled conditions via Monte-Carlo simulations. The results clearly demonstrate the beneficial effects of the new cepstral prefiltering technique on TDE performance (i.e., reduction of bias, variance and number of anomalous estimates).

Previously, time-delay estimation (TDE) between two or more microphones has been proposed as a means for the passive localization of a talker in an audio-conference room. The authors present a simulation study of the effects of room reverberation on the performance of the maximum likelihood (ML) estimator of the time delay, which is obtained by maximizing the output of a generalized cross-correlator. To this end, synthetic microphone signals are generated with the image model technique and are used to evaluate the bias, variance and probability of anomaly of the ML estimator as a function of the room reflection coefficients and other external parameters of interest. The results clearly demonstrate the adverse effects of room reverberation on MLTDE performance. Qualitative and quantitative explanations of these effects are provided.>

A new adaptive eigendecomposition algorithm that can be used for on-line high-resolution spectral/spatial analysis is presented. The formulation of the algorithm is based on the interpretation of the correction term in the recursive update of the data covariance matrix estimate as a perturbation term, with the forgetting factor playing the role of a perturbation parameter. Following this interpretation, a first-order perturbation analysis is made to obtain a new recursion expressing the eigenstructure estimate of the true data covariance matrix at time k, in terms of the eigenstructure estimate at time k-1. The resulting algorithm can be realized by means of M linear combiners with nonlinear weight-vector adaption equations, where M is the signal-subspace dimensionality. Comparative simulation results for narrowband array data indicate very good performance of the proposed algorithm.>

The structure and performance of the maximum likelihood (ML) estimator of slowly varying time delays when the received signals are contaminated by additive noise containing a strongly directional component, due to the presence of a fixed, localized source of interference in the acoustic environment is investigated. The ML estimator is obtained by maximizing the output of the log-likelihood processor. The latter is shown to consist of a slowly varying noise canceler followed by a minimum mean square error estimator of the source signal and a correlator. Closed-form expressions are obtained for the Cramer-Rao lower bound (CRLB) on the error covariance matrix of time-varying delay estimators. The effects of directional interference on the CRLB are investigated numerically for a simplified configuration consisting of two sensors and linearly varying time delays.>

The problem of optimum space-time processing for multiple Gaussian source signals transmitted through a slowly-varying linear channel and monitored with a passive array of sensors in the presence of spatially correlated noise is addressed. To solve this problem, a new class of linear systems (LS), referred to as semistationary, is introduced. These LS are characterized by time-frequency representations whose variations in time occur over intervals much larger than the corresponding system correlation time. The general conditions under which semistationary LS can be used in array processing are investigated and shown to be satisfied in many applications. By modeling the slowly varying linear channel as a semistationary LS and using the factorization properties of the optimum processor, closed form expressions are obtained for the log-likelihood function of the array output and for the associated Cramer-Rao lower bound on estimator variance.>

The authors present an exact solution to the problem of maximum-likelihood time-delay estimation over arbitrary observation time T. That is, the standard assumption T>> tau /sub c/+d/sub max/ made in the derivation of the asymptotic maximum-likelihood (AML) estimator, where t/sub c/ is the correlation time of the various processes involved and d/sub max/ the maximum permissible delay, is relaxed. The exact maximum-likelihood (EML) processor is shown to consist of a special finite-time beamformer, followed by a scalar postprocessor based on the eigenvalues and eigenfunctions of a certain integral equation. The solution of this integral equation is obtained for the case of stationary signals with rational power spectral densities (PSD). The performance of EML and AML are compared by means of computer simulations for a first-order low-pass PSD. The results show that EML can lead to a significant improvement in performances (bias, variance, large errors) when the condition T>> tau /sub c/+d/sub max/ is not satisfied.>