Signal Processing Perspective Research Articles

Developing Laboratory Experiments to teach Undergraduate Engineering Students Multi-User Massive MIMO Technology for 5G Cellular Networks

An effective method of facilitating student learning in a laboratory environment is “practice by doing”. The study of modules related to signal-processing for digital communications requires deep mathematical and theoretical foundations but the practice goal is not really emphasized in the undergraduate curriculum at the University of Mauritius. This causes the students to lose interest in the corresponding modules resulting in high rate of failures. In this work, we propose to develop laboratory experiments with a view to bridge the gap between theoretical and practical aspects in the field of Massive Multiple-Input Multiple-Output (MIMO) systems for 5G cellular networks. Various laboratory scenarios are set up that consider a Maximum Ratio Combining (MRC) receiver in the uplink with an uncorrelated Rayleigh fading channel. Moreover, from a signal processing perspective to enhance the student’s understanding, we analyze the efficiency and error performance of Massive MIMO systems with MPSK and MQAM modulation schemes as well as perfect and imperfect channel estimates. The leading industry software package MATLAB R2022 is used to develop all laboratory experiments and the codes are elaborated in this analysis. Data collected from the experiments are used to generate spectral efficiency and error performance curves which can be used for future research. The findings underscore the significance of accounting for both scenarios and illuminate promising avenues for future research in the realm of massive MIMO education and learning. The assimilation of MATLAB® flowcharts for each MRC receiver, MPSK, and MQAM with perfect and imperfect Channel State Information (CSI) adds further depth to the study, ensuring a comprehensive understanding of the intricacies of massive MIMO systems. Ultimately, this contribution helps in the nurturing of expertise such that future generations of wireless communication pioneers can be inspired.

Federated recurrent-based adaptive battle royale algorithm for privacy-preserving data aggregation in industrial IoT: a signal, image, and video processing perspective

Federated recurrent-based adaptive battle royale algorithm for privacy-preserving data aggregation in industrial IoT: a signal, image, and video processing perspective

Data Science Education: The Signal Processing Perspective [SP Education

Data Science Education: The Signal Processing Perspective [SP Education

Real-time computer-generated integral imaging light field displays: revisiting the point retracing rendering method from a signal processing perspective.

Integral imaging light field displays (InIm-LFDs) can provide realistic 3D images by showing an elemental image array (EIA) under a lens array. However, it is always challenging to computationally generate an EIA in real-time with entry-level computing hardware because the current practice that projects many viewpoints to the EIA induces heavy computations. This study discards the viewpoint-based strategy, revisits the early point retracing rendering method, and proposes that InIm-LFDs and regular 2D displays share two similar signal processing phases: sampling and reconstructing. An InIm-LFD is demonstrated to create a finite number of static voxels for signal sampling. Each voxel is invariantly formed by homogeneous pixels for signal reconstructing. We obtain the static voxel-pixel mapping through arbitrarily accurate raytracing in advance and store it as a lookup table (LUT). Our EIA rendering method first resamples input 3D data with the pre-defined voxels and then assigns every voxel's value to its homogeneous pixels through the LUT. As a result, the proposed method reduces the computational complexity by several orders of magnitude. The experimental rendering speed is as fast as 7 to 10 ms for a full-HD EIA frame on an entry-level laptop. Finally, considering a voxel may not be perfectly integrated by its homogeneous pixels, called the sampling error, the proposed and conventional viewpoint-based methods are analyzed in the Fourier domain. We prove that even with severe sampling errors, the two methods negligibly differ in the output signal's frequency spectrum. We expect the proposed method to break the long-standing tradeoff between rendering speed, accuracy, and system complexity for computer-generated integral imaging.

Deep Learning Meets Sparse Regularization: A signal processing perspective

Deep Learning Meets Sparse Regularization: A signal processing perspective

The Foundations of Computational Imaging: A signal processing perspective

The Foundations of Computational Imaging: A signal processing perspective

Extending compositional data analysis from a graph signal processing perspective

Traditional methods for the analysis of compositional data consider the log-ratios between all different pairs of variables with equal weight, typically in the form of aggregated contributions. This is not meaningful in contexts where it is known that a relationship only exists between very specific variables (e.g. for metabolomic pathways), while for other pairs a relationship does not exist. Modeling absence or presence of relationships is done in graph theory, where the vertices represent the variables, and the connections refer to relations. This paper links compositional data analysis with graph signal processing, and it extends the Aitchison geometry to a setting where only selected log-ratios can be considered. The presented framework retains the desirable properties of scale invariance and compositional coherence. A real data example from bioinformatics underlines the usefulness of this approach.

Physics-Guided Terahertz Computational Imaging: A tutorial on state-of-the-art techniques

Visualizing information inside objects is an everlasting need to bridge the world from physics, chemistry, and biology to computation. Among all tomographic techniques, terahertz (THz) computational imaging has demonstrated its unique sensing features to digitalize multidimensional object information in a nondestructive, nonionizing, and noninvasive way. Applying modern signal processing and physics-guided modalities, THz computational imaging systems have now been launched in various application fields in industrial inspection, security screening, chemical inspection, and nondestructive evaluation. In this article, we review recent advances in THz computational imaging modalities in the aspects of system configuration; wave propagation and interaction models; and physics-guided algorithms for digitalizing interior information of imaged objects. Several image restoration and reconstruction issues based on multidimensional THz signals are further discussed, which provides a crosslink among material digitalization, functional property extraction, and multidimensional imager utilization from a signal processing perspective.

Hyperspectral Anomaly Detection Based on Chessboard Topology

Without any prior information, hyperspectral anomaly detection is devoted to locating targets of interest within a specific scene by exploiting differences in spectral characteristics between various land covers. Traditional methods originated from the signal processing perspective, and most of them rely heavily on specific model assumptions. Due to the model-driven attributes, such methods cannot mine the deep-level features of data to adapt to the variability of scenes, and cannot fully extract the information of land covers contained in images to accurately separate anomalies from background. By independently designing a chessboard-shaped topological framework that avoids making any distribution assumptions but directly mines high-dimensional data features to break through the limitations of traditional detectors, this paper proposes a novel chessboard topology-based anomaly detection (CTAD) method, which constructs a chessboard-shaped topology to dissect images and extract detailed information of land covers adaptively, thereby enabling highly accurate detection. Extensive experimental results on HSIs in real scenes demonstrate that the proposed CTAD can be adapted to the variability of scenes by autonomously learning data features, and exhibits strong generalization and detection capabilities, facilitating practical applications.

Design and Analysis of Convolutional Neural Layers: A Signal Processing Perspective

Convolutional layers (CLs) are ubiquitous in contemporary deep neural network (DNN) models, commonly used for automatic feature extraction. A CL performs cross-correlation between the input to the layer and a set of learnable kernels to produce the layer output. Typically, kernel weights are randomly initialized and automatically learned during model training using the backpropagation and gradient descent algorithms to minimize a specific loss function. Modern DNN models comprise deep hierarchical stacks of CLs and pooling layers. Despite their prevalence, CLs are perceived as a magical tool for feature extraction without solid interpretations of their underlying working principles. In this work, we advance a method for designing and analyzing CLs by providing novel signal processing interpretations of the CL by exploiting the correlation and equivalent convolution functions of the layer. The proposed interpretations enable the employment of CLs to develop finite impulse response (FIR) filters, matched filters (MFs), short-time Fourier transform (STFT), discrete-time Fourier transform (DTFT), and continuous wavelet transform (CWT) algorithms. The main idea is to pre-assign the CL kernel weights to implement a specific convolution- or correlation-based DSP algorithm. Such an approach enables building self-contained DNN models in which CLs are utilized for various preprocessing and feature extractions tasks, enhancing the model portability, and cutting down the preprocessing computational cost. The proposed DSP interpretations provide an effective means to analyze and explain the operation of automatically trained CLs in the time and frequency domains by reversing the design procedures. The presented interpretations are mathematically established and experimentally validated with a comprehensive machinery fault diagnosis application example illustrating the potential of the proposed methodology.

Graph Signal Smoothness Based Feature Learning of Brain Functional Networks in Schizophrenia.

In this paper we study the brain functional network of schizophrenic patients based on resting-state fMRI data. Different from the region of interest (ROI)-level brain networks that describe the connectivity between brain regions, this paper constructs a subject-level brain functional network that describes the similarity between subjects from a graph signal processing (GSP) perspective. Based on the subject graph, we introduce the concept of graph signal smoothness to analyze the abnormal brain regions (feature brain regions) in which schizophrenic patients produce abnormal functional connections and to quantitatively rank the degree of abnormality of brain regions. We find that in the patients' brain networks, many new connections appear and some common connections are strengthened. The feature brain regions can be easily found according to the value of connection differences. Finally, we validate the learned feature brain regions by the results of two types of statistical analyses (ROI-to-ROI analysis and seed-to-voxel analysis), and the feature brain regions derived from graph signal smoothness are indeed the brain regions with significant differences in the statistical analysis, which illustrates the potential of graph signal smoothness for use in quantitative analysis of brain networks.

Common‐offset domain reflection–diffraction separation with deep learning

AbstractSeparation of diffracted from reflected events in seismic data is still challenging due to the relatively low amplitude of the diffracted wavefield compared to the reflected wavefield as well as the overlap in the kinematics of reflection and diffraction events. A workflow based on deep learning can be a simple and fast alternative, but using training data made by physics‐based modelling is expensive and lacks diversity in terms of noise, amplitude, frequency content and wavelet. This results in poor generalization beyond the training data without retraining and transfer learning. In this paper, we demonstrate successful applications of reflection–diffraction separation using a conventional U‐net architecture. The novelty of our approach is that we do not use synthetic data created from physics‐based modelling, but instead use only synthetic data built from basic geometric shapes. Our domain of application is the pre‐migration common‐offset domain where reflected events resemble local geology and the diffracted wavefield consists of downward convex hyperbolic diffraction patterns. Both patterns were randomly perturbed in many ways while maintaining their intrinsic features. This approach is inspired by the common practice of data augmentation in deep learning for machine vision applications. Since many of the standard data augmentation techniques lack a geophysical motivation, we have instead perturbed our synthetic training data in ways to make more sense from a signal processing perspective or given our ‘domain knowledge’ of the problem at hand. We did not have to retrain the network to show good results on the field data set. The large variety and diversity in examples enabled to trained neural networks to show encouraging results on synthetic and field data sets that were not used in training.

Attenuation of seismic migration smile artifacts with deep learning

Attenuation of migration artifacts on Kirchhoff migrated seismic data can be challenging due to the relatively low amplitude of migration artifacts compared to reflections as well as the overlap in the kinematics of reflection and migration smiles. Several ‘conventional’ filtering methods exist and recently deep learning based workflows have been proposed. A deep learning workflow can be a simple and fast alternative to existing methods. In case of supervised training of a deep neural network using training data made by physics-based modelling or actual migrations is expensive and lacks diversity in terms of noise, amplitude, frequency content and wavelet. This can result in poor generalization beyond the training data without re-training and transfer learning. In this paper we demonstrate successful applications of migration smile separation using a conventional U-net architecture. The novelty in our approach is that we do not use synthetic data created from physics-based modelling, but instead use only synthetic data build form basic geometric shapes. Our domain of application is the migrated common offset domain, or simply the stack of the pre-stack migrated data, where reflections resemble local geology and migration smiles are upward convex hyperbolic patterns. Both patterns were randomly perturbed in many ways while maintaining their intrinsic features. This approach is inspired by the common practice of data augmentation in deep learning for machine vision applications. Since many of the standard data augmentation techniques lack a geophysical motivation, we have instead perturbed our synthetic training data in ways to make more sense for a signal processing perspective or given our ‘domain knowledge’ of the problem at hand. We did not have to retrain the network to produce good results on the field dataset. The large variety and diversity in examples enabled the trained neural network to show encouraging results on synthetic and field datasets that were not used in training.

On embeddings and inverse embeddings of input design for regularized system identification

Input design is an important problem for system identification and has been well studied for the classical system identification, i.e., the maximum likelihood/prediction error method. For the emerging regularized system identification, the study on input design has just started, and it is often formulated as a non-convex optimization problem minimizing a scalar measure of the Bayesian mean squared error matrix subject to certain constraints. Among the state-of-art input design techniques for regularized system identification is the so-called quadratic mapping and inverse embedding (QMIE) method. Based on the quadratic mapping between the input and its covariance, the QMIE method is first to obtain the optimal autocovariance by solving a transformed convex optimization problem and then to find all the inputs corresponding to the optimal autocovariance by the time domain inverse embedding (TDIE). In this paper, we report some new results on the embeddings/inverse embeddings of the QMIE method. Firstly, we present a general result on the frequency domain inverse embedding (FDIE) that is to find the inverse of the quadratic mapping described by the discrete-time Fourier transform. Then we show the relation between the TDIE and the FDIE from a graph signal processing perspective. Finally, motivated by this perspective, we further propose a graph induced embedding and its inverse, which include the previously introduced embeddings as special cases. This deepens our understanding of input design from a broader perspective beyond the time domain and frequency domain viewpoints.

On acoustically modulated jet shear layers and the Nyquist–Shannon sampling theorem

The goal of this paper is to present the behavior of a jet shear layer in response to acoustic excitation from a signal processing perspective. The main idea is that the vortices that roll-up in the jet shear layer are similar to the discrete samples of a digital control system, and, hence, that the Nyquist–Shannon sampling theorem should apply. We further hypothesize that the strength of a vortex is determined by the mean amplitude of the excitation waveform during its creation. We also argue that, at least in some cases, demodulation occurs as a result of the vorticity signal generated by the convection of discrete vortices past a point in the shear layer. This vorticity signal is related to the amplitude modulation (AM) excitation waveform by a half-wave rectification operation, a common implementation of an AM demodulator. To investigate these ideas, a free, round jet that is excited upstream of the nozzle is studied using particle image velocimetry. Experiments are conducted that confirm that the sampling theorem applies, and an aliased response is observed when the Nyquist limit is exceeded. Previous authors have attributed demodulation to a vortex merging mechanism, but we demonstrate that merging is not always required for demodulation and suggest that it is one of two mechanisms at play.

AGAR: Array-Geometry-Aided Ambiguity Resolution for Baseline Growing With Global Navigation Satellite Systems

Ambiguity resolution (AR) is a fundamental problem for carrier phase based signal processing tasks to leverage the superhigh precision of wavelength-level range and velocity measurements. With the elaborately designed waveform and coordinated running of the space-based satellite system, the antenna-array observations of global navigation satellite system (GNSS) signals feature a phase measurement model. In this article, the AR of baseline estimation with GNSS carrier phase measurements only in the AR step is examined from an array signal processing perspective. The array-geometry-aided ambiguity resolution approach, coined as AGAR, is proposed for growing the baseline estimation provided by a search-free algorithm to the accuracy of the aperture level in an effective way. First, the single-epoch and search-free $2q$-order AR method is further investigated in size and statistical independence. Second, the conventional phase beampattern is defined to characterize the similarity of carrier phase measurement vector of signals from different directions. Third, a simple and effective ambiguity-lookup-table approach after the conventional phase beampattern is proposed which fulfills the goal of baseline growing to the array aperture level. Fourth, the identifiability, success rate, Cramér–Rao bound (CRB), and computation complexity are analyzed. As a result, the phase-based and complex-based processings are distinguished, resulting in an alternative analytical prediction of outlier probability that well approximates AR’s success rate. The relationship between the success rate of AR and CRB is also clarified. Numerical simulations are carried out to verify the proposed analytical prediction and AR approach. The AR success rate increased from 10% to 93% at a relatively large measurement error of 0.05 cycle.

Post Hoc Explainability for Time Series Classification: Toward a signal processing perspective

Time series data correspond to observations of phenomena that are recorded over time <xref ref-type="bibr" rid="ref1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</xref> . Such data are encountered regularly in a wide range of applications, such as speech and music recognition, monitoring health and medical diagnosis, financial analysis, motion tracking, and shape identification, to name a few. With such a diversity of applications and the large variations in their characteristics, time series classification is a complex and challenging task. One of the fundamental steps in the design of time series classifiers is that of defining or constructing the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">discriminant features</i> that help differentiate between classes. This is typically achieved by designing novel <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">representation techniques</i> <xref ref-type="bibr" rid="ref2" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[2]</xref> that transform the raw time series data to a new data domain, where subsequently a classifier is trained on the transformed data, such as one-nearest neighbors <xref ref-type="bibr" rid="ref3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[3]</xref> or random forests <xref ref-type="bibr" rid="ref4" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[4]</xref> . In recent time series classification approaches, deep neural network models have been employed that are able to jointly learn a representation of time series and perform classification <xref ref-type="bibr" rid="ref5" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[5]</xref> . In many of these sophisticated approaches, the discriminant features tend to be complicated to analyze and interpret, given the high degree of nonlinearity.

Federated Learning: A signal processing perspective

The dramatic success of deep learning is largely due to the availability of data. Data samples are often acquired on edge devices, such as smartphones, vehicles, and sensors, and in some cases cannot be shared due to privacy considerations. Federated learning is an emerging machine learning paradigm for training models across multiple edge devices holding local data sets, without explicitly exchanging the data. Learning in a federated manner differs from conventional centralized machine learning and poses several core unique challenges and requirements, which are closely related to classical problems studied in the areas of signal processing and communications. Consequently, dedicated schemes derived from these areas are expected to play an important role in the success of federated learning and the transition of deep learning from the domain of centralized servers to mobile edge devices.

Collaborative Human Decision Making With Heterogeneous Agents

While there has been extensive work on modeling of human decision-making both for individuals and groups from a cognitive psychology point of view, research on this topic from a signal processing and information fusion perspective is relatively recent. In this work, we consider a distributed detection problem consisting of a number of human local decision makers and a fusion center (FC). Signal detection theory is exploited to answer why promoting heterogeneity could improve the performance of collaborative human decision-making. We consider the following two scenarios: 1) the local decision makers are independent and the level of heterogeneity is measured in terms of the variability of human expertise and 2) humans make correlated local decisions due to their perceptual and behavioral similarities and heterogeneity is measured by the amount of correlation. In both cases, we show that the detection performance of the FC can be improved with the increase of heterogeneity. In particular, in the second scenario, we develop a portfolio theory-based framework to select participants from correlated human agents so that heterogeneity is enhanced resulting in improved decision-making performance. Simulations are provided for illustration and performance comparison.

Spatio-Temporal Graph Neural Networks for Multi-Site PV Power Forecasting

Accurate forecasting of solar power generation with fine temporal and spatial resolution is vital for the operation of the power grid. However, state-of-the-art approaches that combine machine learning with numerical weather predictions (NWP) have coarse resolution. In this paper, we take a graph signal processing perspective and model multi-site photovoltaic (PV) production time series as signals on a graph to capture their spatio-temporal dependencies and achieve higher spatial and temporal resolution forecasts. We present two novel graph neural network models for deterministic multi-site PV forecasting dubbed the graph-convolutional long short term memory (GCLSTM) and the graph-convolutional transformer (GCTrafo) models. These methods rely solely on production data and exploit the intuition that PV systems provide a dense network of virtual weather stations. The proposed methods were evaluated in two data sets for an entire year: 1) production data from 304 real PV systems, and 2) simulated production of 1000 PV systems, both distributed over Switzerland. The proposed models outperform state-of-the-art multi-site forecasting methods for prediction horizons of six hours ahead. Furthermore, the proposed models outperform state-of-the-art single-site methods with NWP as inputs on horizons up to four hours ahead.