In this paper, constrained optimization methods with a sparsity-aware Newton-type direction update are applied to adaptive filtering. Different versions of the proposed algorithms, which include a sparsity-aware RLS and a CG algorithm, present lower computational complexity than traditional algorithms based on LMS-Newton, since the proposed algorithms require the use of a smaller Hessian estimate. The proposed algorithms are tested in acoustic echo cancellation and time-varying channel identification. Some of them present a computational reduction of 25% of that of the traditional LMS-Newton in the best case scenario. Most of the proposed algorithms present faster convergence than the benchmark algorithms in several scenarios.

Prosthetic hands are of paramount importance in the rehabilitation of upper limbs amputees. Gesture recognition using surface electromyography (sEMG) data has emerged as a great option for controlling prosthetic devices, since these data are acquired by non-intrusive sensors. This work presents a real-time classification system based on artificial neural networks with individualized muscle activity segmentation and using dynamic time warping (DTW) based features. For real-time classification, we modify the size of the sliding windows so that their length is sufficient to fully capture the muscle activity signal. For data segmentation, we propose an enhanced muscle activity detection in which validation is used to fine-tune the thresholds needed to determine the beginning and end of muscle contraction. We used two validation methods: cross-validation and multi-holdout. Moreover, we propose a post-processing technique to choose the most representative class when there are multiple classifications for a given data. By combining all the proposed techniques, the accuracy of the resulting system was (97.2 ± 0.3)% in the classification of 6 hand gestures from 10 healthy people, representing an increase of 7.1% in the mean accuracy compared with the baseline model.

Prosthetic hands play a vital role in the rehabilitation of upper limb amputees. Gesture recognition using surface electromyography (sEMG) data has emerged as an excellent option for controlling such prosthetic devices since one does not require invasive methods to obtain these data. In order to improve gesture recognition, we must extract the muscle activity from the raw data before classification, as each gesture has its own patterns. In this paper, we use an artificial neural network classifier with individualized data segmentation based on gesture detection to identify six hand movements. We used data from ten healthy volunteers. By combining data segmentation and crossvalidation, we were able to refine the amplitude thresholds used to determine the beginning and end of muscle contractions for each person. We designed several experiments using different types of cross-validation. The performance achieved by the proposed model using 4-fold cross-validation was (93.6 ± 0.7)%, which represents 3.5% more than the mean accuracy of the baseline model, in which there is a single arbitrarily-chosen segmentation threshold for all volunteers.

Machine learning techniques have shown success in classifying hand gestures. As the prevalence of prosthetic devices continues to rise, the adoption of non-invasive technologies, such as surface electromyography (sEMG), becomes paramount. This study systematically assesses the isolated influence of classification algorithms within hand gesture recognition (HGR) systems using sEMG data and dynamic time warping (DTW) based features. This approach effectively handles temporal variations in sEMG signals by leveraging DTW, ensuring input features are invariant to gesture speed. Six supervised learning classifiers were evaluated: the multilayer perceptron, support vector machine, logistic regression, linear discriminant analysis, k-nearest neighbors, and decision tree. Cross-validation was employed to fine-tune the segmentation hyperparameters, significantly improving results. To ensure reproducibility, the source code has been made available, the proposed system design has been detailed, and the evaluation protocols have been described. Our findings indicate that logistic regression outperformed other classifiers in this setup, achieving 95.2% accuracy in classifying six hand movements from ten healthy individuals, representing a 1.6% improvement over the best previously reported performance using the same publicly available dataset. Future research will assess the proposed HGR system’s generalization capability on larger datasets suitable for training more complex classifiers, including deep learning models.

This paper provides a review of user selection algorithms for massive multiple-input multiple-output (MIMO) systems under the line-of-sight (LoS) propagation model. Although the LoS propagation is extremely important to some promising technologies, like in millimeter-wave communications, massive MIMO systems are rarely studied under this propagation model. This paper fills this gap by providing a comprehensive study encompassing several user selection algorithms, different linear precoders and simulation setups, and also considers the effect of partial channel state information (CSI). One important result is the existence of practical cases in which the LoS propagation model may lead to significant levels of interference among users within a cell; these cases are not satisfactorily addressed by the existing user selection algorithms. Motivated by this issue, a new user selection algorithm based on inter-channel interference (ICI) called ICI-based selection (ICIBS) is proposed. Unlike other techniques, the ICIBS accounts for the ICI in a global manner, thus yielding better results, especially in cases where there are many users interfering with each other. In such scenarios, simulation results show that when compared to the competing algorithms, the proposed approach provided an improvement of at least 10.9% in the maximum throughput and 7.7% in the 95%-probability throughput when half of the users were selected.

Resumo-Comparamos duas versões do algoritmo MUSIC já consolidadas na literatura, formuladas para localização de uma única fonte estática, quando aplicadas ao caso de múltiplas fontes sonoras móveis em ambiente aberto no campo distante.A literatura informa que elas apresentam desvantagens na análise da localização de mais de uma fonte emitindo sinais de natureza não-estacionária.Decidimos verificar numa série de experimentos as limitações dos algoritmos na prática.Expõem-se considerações e particularidades sobre o desempenho desses algoritmos em relação à definição e classificação de picos máximos em funções de saída e às características dos sinais

Massive multiple-input multiple-output (MIMO) enables increased throughput by using spatial multiplexing. However, the throughput may severely degrade when the number of users served by a single base station increases, especially under line-of-sight (LoS) propagation. Selecting users is a possible solution to deal with this problem. In the literature, the user selection algorithms can be divided into two classes: small-scale fading aware (SSFA) and large-scale fading aware (LSFA) algorithms. The LSFA algorithms are good solutions for massive MIMO systems under non LoS propagation since the small-scale fading does not affect the system performance under this type of propagation. For the LoS case, the small-scale fading has a great impact on the system performance, requiring the use of SSFA algorithms. However, disregarding the large-scale fading is equivalent to assuming that all users are equidistant from the base station and experience the same level of shadowing, which is not a reasonable approximation in practical applications. To address this shortcoming, a new user selection algorithm called the fading-ratio-based selection (FRBS) is proposed. FRBS considers both fading information to drop those users that induce the highest interference to the remaining ones. Simulation results considering LoS channels show that using FRBS yields near optimum downlink throughput, which is similar to that of the state-of-the-art algorithm, but with much lower computational complexity. Moreover, the use of FRBS with zero-forcing precoder resulted in 26.28% improvement in the maximum throughput when compared with SSFA algorithms, and 35.39% improvement when compared with LSFA algorithms.

Learn to solve the unprecedented challenges facing Online Learning and Adaptive Signal Processing in this concise, intuitive text. The ever-increasing amount of data generated every day requires new strategies to tackle issues such as: combining data from a large number of sensors; improving spectral usage, utilizing multiple-antennas with adaptive capabilities; or learning from signals placed on graphs, generating unstructured data. Solutions to all of these and more are described in a condensed and unified way, enabling you to expose valuable information from data and signals in a fast and economical way. The up-to-date techniques explained here can be implemented in simple electronic hardware, or as part of multi-purpose systems. Also featuring alternative explanations for online learning, including newly developed methods and data selection, and several easily implemented algorithms, this one-of-a-kind book is an ideal resource for graduate students, researchers, and professionals in online learning and adaptive filtering.

This article proposes versions of the set-membership affine-projection algorithm that estimate an adequate threshold and a proper constraint-vector for a system identification scenario.A novel constraint-vector is proposed for a non-white input, since the Simple-Choice Constraint Vector is asymptotically optimum for white inputs.Non-white inputs are commonplace in communication systems when there is channel encoding or precoding in the transmitter.A fixed and a time-varying threshold set-membership version are used.Simulations show that the proposed algorithm presents lower mean-square error and updates in fewer iterations than the traditional Simple-Choice and Exponential-Decay Constraint Vectors.For the timevarying threshold, the computational complexity of the algorithm is similar to the Simple-Choice Constraint Vector.

The paper deals with the identification of nonlinear systems with adaptive filters. In particular, adaptive filters for functional link polynomial (FLiP) filters, a broad class of linear-in-the-parameters (LIP) nonlinear filters, are considered. FLiP filters include many popular LIP filters, as the Volterra filters, the Wiener nonlinear filters, and many others. Given the large number of coefficients of these filters modeling real systems, especially for high orders, the solution is often very sparse. Thus, an adaptive filter exploiting sparsity is considered, the improved proportionate NLMS algorithm (IPNLMS), and an optimal step-size is obtained for the filter. The optimal step-size alters the characteristics of the IPNLMS algorithm and provides a novel gradient descent adaptive filter. Simulation results involving the identification of a real nonlinear device illustrate the achievable performance in comparison with competing similar approaches.

Adaptive filters exploiting sparsity have been a very active research field, among which the algorithms that follow the "proportional-update principle", the so-called proportionate-type algorithms, are very popular. Indeed, there are hundreds of works on proportionate-type algorithms and, therefore, their advantages are widely known. This paper addresses the unexplored drawbacks and limitations of using proportional updates and their practical impacts. Our findings include the theoretical justification for the poor performance of these algorithms in several sparse scenarios, and also when dealing with non-stationary and compressible systems. Simulation results corroborating the theory are presented.

We have been witnessed a growing research activity to advance new strategies to detect and exploit underlying sparsity in the parameters of physical models. In many cases, the sparsity is not explicit in the relations among the parameter coefficients requiring some suitable tools to reveal the potential sparsity. This work proposes a family of adaptive filtering algorithms, aimed at exposing some hidden features of the unknown parameters. Although the basic idea applies to any algorithm, we will concentrate the work in the LMS-type algorithms, giving rise to a family collectively named as Feature LMS (F-LMS) algorithms. These algorithms increase the convergence speed and reduce the steady-state mean-squared error, in comparison with the classical LMS solution. The main idea is to apply linear transformations, through the so-called feature matrices, to reveal the sparsity hidden in the coefficient vector, followed by a sparsity-promoting penalty function to exploit the exposed sparsity. For illustration, a few F-LMS algorithms for lowpass, bandpass, and highpass systems are introduced by using simple feature matrices that require either only simple operations or can learn the features. Simulations and real-life experiments demonstrate that the F-LMS algorithms bring about several performance improvements whenever the unknown sparsity of parameters is exposed.

Power allocation techniques, among which the max-min fairness power allocation (MMFPA) is one of the most widely used, are essential to guarantee good data throughput for all users in a cell. Recently, an efficient MMFPA algorithm for massive multiple-input multiple-output (MIMO) systems has been proposed. However, this algorithm is susceptible to the initial search interval employed by the underlying bisection search. Even if the optimal point belongs to the initial search interval, this algorithm may fail to converge to such a point. In this letter, we use the Perron-Frobenius theory to explain this issue and provide search intervals that guarantee convergence to the optimal point. Furthermore, we propose the bound test procedure as an efficient way of initializing the search interval. Simulation results corroborate our findings.

Recently, the improved simple set-membership affine projection (IS-SM-AP) algorithm has been proposed in order to exploit sparsity in system models. Although its update equation resembles that of the set-membership affine projection (SM-AP) algorithm, the IS-SM-AP algorithm has two fundamental advantages over the SM-AP algorithm: (i) it can exploit sparsity in system models and (ii) its computational complexity is lower. Up to now, the properties of this algorithm have been addressed only through numerical simulations, and no analytical study has been presented. To fill this gap, in this study, the authors analyse the steady-state mean squared error of the IS-SM-AP algorithm using the energy conservation method. Furthermore, some important implementation issues are addressed, and a time-varying parameter for the discard function is proposed. Finally, the authors present numerical results corroborating the theoretical analysis and the effectiveness of the proposed time-varying parameter.

Many real systems have inherently some type of sparsity. Recently, the feature least-mean square (F-LMS) has been proposed to exploit hidden sparsity. Unlike the existing algorithms, the F-LMS algorithm performs a linear combination of the adaptive coefficients to reveal and then exploit the hidden sparsity. However, many systems have also plain besides hidden sparsity, and the F-LMS algorithm is not able to exploit the former. In this paper, we propose a new algorithm, named simple sparsity-aware F-LMS (SSF-LMS) algorithm, that is capable of exploiting both kinds of sparsity simultaneously. The hidden sparsity is exploited just like in the F-LMS algorithm, whereas the plain sparsity is exploited by means of the discard function applied to the filter coefficients. By doing so, the proposed SSFLMS algorithm not only outperforms the F-LMS algorithm when plain sparsity is also observed, but also requires fewer arithmetic operations. Numerical results show that the proposed algorithm has faster speed of convergence and reaches lower steady-state mean-squared error (MSE) than the F-LMS and classical algorithms, when the system has plain and hidden sparsity.

The paper addresses adaptive algorithms for Volterra filter identification capable of exploiting the sparsity of nonlinear systems. While the l 1 -norm of the coefficient vector is often employed to promote sparsity, it has been shown in the literature that superior results can be achieved using an approximation of the l 0 -norm.Thus, in this paper, the Geman-McClure function is adopted to approximate the l 0 -norm and to derive l 0 -norm adaptiveVolterra filters. It is shown through experimental results, also involving a real-world system, that the proposed adaptive filters can obtain improved performance in comparison with classical approaches and l 1 -norm solutions.

New approaches have been proposed to detect and exploit sparsity in adaptive systems. However, the sparsity is not always explicit among the system coefficients, thus requiring some tools to reveal it. By means of the so-called feature function, we propose the low-complexity feature stochastic gradient (LF-SG) algorithm to exploit hidden sparsity. The proposed algorithm aims at reducing the computational load of the learning process, as compared to the least-mean-square (LMS) algorithm. We focus on block-lowpass systems, but the proposed approach can easily be adapted to exploit other kinds of features of the unknown system, e.g., highpass and bandpass characteristics. Then, we analyze some properties of the LF-SG algorithm, namely its steady-state mean squared error (MSE), its bias, and the choice of the step-size parameter. Simulation results illustrate the competitive MSE performance of the LF-SG in comparison with the LMS, but the former algorithm requires much fewer multiplication operations to identify lowpass systems. For instance, to identify a measured room impulse response, the LF-SG algorithm realized less than half of the multiplication operations required by the LMS algorithm.

Sparse representations of model parameters have been widely studied. In the adaptive filtering literature, most studies address the cases where the sparsity is directly observed, therefore, there is a growing interest in developing strategies to exploit hidden sparsity. Recently, the feature LMS (F-LMS) algorithm was proposed to expose the sparsity of models with low- and high-frequency contents. In this paper, the F-LMS algorithm is extended to expose hidden sparsity related to models with bandpass spectrum, including the cases of narrowband and broader passband sources. Some simulation results show that the proposed approaches lead to F-LMS algorithms with fast convergence, low misadjustment after convergence, and low computational cost.

Set-membership affine projection (SM-AP) adaptive filters have been increasingly employed in the context of online data-selective learning. A key aspect for their good performance in terms of both convergence speed and steady-state mean-squared error is the choice of the so-called constraint vector. Optimal constraint vectors were recently proposed relying on convex optimization tools, which might sometimes lead to prohibitive computational burden. This paper proposes a convex combination of simpler constraint vectors whose performance approaches the optimal solution closely, utilizing much fewer computations. Some illustrative examples confirm that the sub-optimal solution follows the accomplishments of the optimal one.