Robust Signal Processing Algorithms Research Articles

A triangle-shape region of interest in cardiorespiratory estimation during sleep monitoring

Abstract Introduction: With advancements in sensor and communication technologies, sleep monitoring is moving out of specialized clinics and into everyday homes. Extracting sleep-related data using far less complicated tools and procedures is possible than polysomnography. Respiratory and cardiovascular data are extracted from the signals such as the electrocardiogram (ECG), photoplethysmogram (PPG), and ballistocardiogram (BCG) to identify the aberrant respiratory events of apnea/hypopnea as well as to estimate sleep parameters. However, due to the different sleeping positions, such systems lack accuracy and/or complicated sensor network topology. In this work, we proposed an optimal topology of forcesensitive resistor (FSR) sensors to simplify the system design by identifying the region of interest for estimating cardiorespiratory parameters with minimal error. The sensors are deployed under the mattress and located on the bed frame. Methods: We proposed a low-cost, unobtrusive, non-invasive, and reliable solution with robust signal processing algorithms for cardiorespiratory measurements and automatic signal validation based on signal quality. The solution is established based on a multi-physical layer (MPL) and sensor interfaces coping with the embedded system’s specifications, and signal processing is performed onboard with two independent and simultaneous pipelines for heart rate and respiratory rate using discrete wavelet transform (DWT) and bandpass filter, respectively. Results: We identified the three most contributing FSR sensors forming a triangle shape covering the left upper side of the subject (in the supine position) as the region of interest. We reduced the mean absolute error (MAE) to as low as 3.94 and 2.35 for heart rate and respiratory rate. Conclusions: The approach with the topology of triangle-shaped performs appropriately in estimating the cardiorespiratory parameters in all four regular sleeping positions, i.e. supine, prone, left lateral, and right lateral.

Synoris: an Unmanned Underwater Platform Based on Hydrophone Arrays for Detection and Tracking From Sound Signatures

Sonars have been in practical use since the turn of the 20th century. They are considered to be among the most sophisticated engineering systems concerning detection and tracking. Unmanned underwater vehicles have recently evolved to a higher degree of sophistication with enough processing power to run more demanding artificial intelligence (AI) algorithms. This work is focused on the development of a robust signal processing algorithm on board an Autonomous Underwater Vehicle (AUV), named Synoris, enhanced with target identification capabilities. Synoris is able to perform feature extraction, classification and identification using only on board hardware.

An incremental algorithm based on multichannel non-negative matrix partial co-factorization for ambient denoising in auscultation

One of the major current limitations in the diagnosis derived from auscultation remains the ambient noise surrounding the subject, which prevents successful auscultation. Therefore, it is essential to develop robust signal processing algorithms that can extract relevant clinical information from auscultated recordings analyzing in depth the acoustic environment in order to help the decision-making process made by physicians. The aim of this study is to implement a method to remove ambient noise in biomedical sounds captured in auscultation. We propose an incremental approach based on multichannel non-negative matrix partial co-factorization (NMPCF) for ambient denoising focusing on high noisy environment with a Signal-to-Noise Ratio (SNR) ⩽-5 dB. The first contribution applies NMPCF assuming that ambient noise can be modelled as repetitive sound events simultaneously found in two single-channel inputs captured by means of different recording devices. The second contribution proposes an incremental algorithm, based on the previous multichannel NMPCF, that refines the estimated biomedical spectrogram throughout a set of incremental stages by eliminating most of the ambient noise that was not removed in the previous stage at the expense of preserving most of the biomedical spectral content. The ambient denoising performance of the proposed method, compared to some of the most relevant state-of-the-art methods, has been evaluated using a set of recordings composed of biomedical sounds mixed with ambient noise that typically surrounds a medical consultation room to simulate high noisy environments with a SNR from −20 dB to −5 dB. In order to analyse the drop in denoising performance of the evaluated methods when the effect of the propagation of the patient’s body material and the acoustics of the room is considered, results have been obtained with and without taking these effects into account. Experimental results report that: (i) the performance drop suffered by the proposed method is lower compared to MSS and NLMS when considering the effect of the propagation of the patient’s body material and the acoustics of the room active; (ii) unlike what happens with MSS and NLMS, the proposed method shows a stable trend of the average SDR and SIR results regardless of the type of ambient noise and the SNR level evaluated; and (iii) a remarkable advantage of the proposed method is the high robustness of the acoustic quality of the estimated biomedical sounds when the two single-channel inputs suffer from a delay between them.

A 6-Bit 1.5-GS/s SAR ADC With Smart Speculative Two-Tap Embedded DFE in 130-nm CMOS for Wireline Receiver Applications

Implementing wireline receivers with a front-end analog-to-digital converter (ADC) allows for complex, flexible, and robust signal processing algorithms in the digital domain, as well as easy implementation of advanced modulation schemes beyond binary PAM2. However, the power consumption of the ADC and ensuing digital equalization is a key issue for such receivers in high-speed applications. Embedding analog equalization inside the ADC architecture allows for both a lower ADC resolution and a reduced-complexity digital equalizer, resulting in a more power-efficient receiver. This article presents a 6-bit 1.5-GS/s time-interleaved successive approximation register (SAR) ADC with low-overhead two-tap embedded decision-feedback equalizer (DFE). A smart speculative DFE is proposed to reduce additional conversion cycles required for the equalization realization in the ADC. Moreover, DFE functions are efficiently embedded in the capacitive digital-to-analog converter (DAC) references. The prototype ADC with two-tap DFE is implemented in a 130-nm CMOS process and achieves a 5.24-bit peak effective number of bits and 0.59-pJ/conversion-step figure-of-merit (FOM) at a 1.5-GS/s sampling rate while consuming 34.1 mW and occupying a core area of 0.32 mm2. The effectiveness of the embedded DFE in timing margin improvement is verified for 1.5-Gb/s operation over high-loss FR4 channels at a bit error rate (BER) of 10−9.

The Institute of Acoustics A B Wood Medal 2019: Julien Bonnel

A common requirement in ocean acoustic applications is for robust signal processing algorithms that take into account the complex ocean environment, which is variable in space and time. Julien Bonnel's research has provided the most innovative approach to signal processing of broadband acoustic data in the last decade: time and frequency warping. Warping is a non-linear signal re-sampling method, based on a physical model of the shallow water sound channel, which compensates for dispersion in the propagating modes. The technique extracts the propagating modes from broadband signals in shallow water at significantly shorter ranges than previously possible, using data from only a single hydrophone. The technique he has developed has applications in two important areas: geo-acoustic inversion in shallow water environments and passive detection and localization of marine mammals. An example of the latter is the localization based on whale vocalizations, to enable estimation of the animal's depth in the water. The method has been successfully applied to obtaining the range and depth of bowhead whales in the Chuchki Sea and North Atlantic Right Whales in the Bay of Fundy. Julien continues his research in geo-acoustic inversion to develop methods for extracting modal wavenumbers from short horizontal arrays. Of significant general interest is his work on re-formulating the modal filtering technique for sound channels with different dispersion characteristics. Julien Bonnel is a most promising early-career research scientist who has introduced innovative methods that have led to significant advances in ocean acoustics. The research community has benefitted from the use of his methods and he has done this within a few years of completing his PhD. Importantly, he continues to produce innovative research to improve the methods that he has already developed. The IOA is delighted to present Julien Bonnel with the 2019 A B Wood Medal.

Hybrid Fiber Fabry–Perot Interferometer With Improved Refractometric Response

The reflectance of a hybrid fiber Fabry–Perot interferometer modulated by the Fresnel reflection at the fiber tip was used to assess the refractive index (RI) of liquid samples. Notwithstanding the closeness of RI values between samples and fiber core, well-defined interference spectra in Fourier domain were obtained after a robust signal processing algorithm. Then, the analysis of these spectra allowed us to derive a mathematical expression to obtain the RI value of the fiber tip surrounding media. Next, to assess the fiber refractometer measuring capability a series of liquid samples, typically used for RI calibration, with values ranging from 1.3924 to 1.5882 were measured. A standard deviation of 7 × 10−4 between the measured and actual RI values was estimated. Finally, the refractometer was used to monitor the RI of a sample, with a high thermo-optic coefficient, while it was heated. The accurate and compact all-fiber refractometer proposed here exhibited immunity to optical power fluctuations and one of the largest dynamic ranges reported. This set of combined features make it very attractive for real-time label-free biosensing applications.

Motometrics: A Toolbox for Annotation and Efficient Analysis of Motor Evoked Potentials.

Stimulating the nervous system and measuring muscle response offers a unique opportunity to interrogate motor system function. Often, this is performed by stimulating motor cortex and recording muscle activity with electromyography; the evoked response is called the motor evoked potential (MEP). To understand system dynamics, MEPs are typically recorded through a range of motor cortex stimulation intensities. The MEPs increase with increasing stimulation intensities, and these typically produce a sigmoidal response curve. Analysis of MEPs is often complex and analysis of response curves is time-consuming. We created an MEP analysis software, called Motometrics, to facilitate analysis of MEPs and response curves. The goal is to combine robust signal processing algorithms with a simple user interface. Motometrics first enables the user to annotate data files acquired from the recording system so that the responses can be extracted and labeled with the correct subject and experimental condition. The software enables quick visual representations of entire datasets, to ensure uniform quality of the signal. It then enables the user to choose a variety of response curve analyses and to perform near real time quantification of the MEPs for quick feedback during experimental procedures. This is a modular open source tool that is compatible with several popular electrophysiological systems. Initial use indicates that Motometrics enables rapid, robust, and intuitive analysis of MEP response curves by neuroscientists without programming or signal processing expertise.

Adaptive Sojourn Time HSMM for Heart Sound Segmentation.

Heart sounds are difficult to interpret due to events with very short temporal onset between them (tens of milliseconds) and dominant frequencies that are out of the human audible spectrum. Computer-assisted decision systems may help but they require robust signal processing algorithms. In this paper, we propose a new algorithm for heart sound segmentation using a hidden semi-Markov model. The proposed algorithm infers more suitable sojourn time parameters than those currently suggested by the state of the art, through a maximum likelihood approach. We test our approach over three different datasets, including the publicly available PhysioNet and Pascal datasets. We also release a pediatric dataset composed of 29 heart sounds. In contrast with any other dataset available online, the annotations of the heart sounds in the released dataset contain information about the beginning and the ending of each heart sound event. Annotations were made by two cardiopulmonologists. The proposed algorithm is compared with the current state of the art. The results show a significant increase in segmentation performance, regardless the dataset or the methodology presented. For example, when using the PhysioNet dataset to train and to evaluate the HSMMs, our algorithm achieved average an F-score of [Formula: see text] compared to [Formula: see text] achieved by the algorithm described in [D.B. Springer, L. Tarassenko, and G. D. Clifford, "Logistic regressionHSMM-based heart sound segmentation," IEEE Transactions on Biomedical Engineering, vol. 63, no. 4, pp.822-832, 2016]. In this sense, the proposed approach to adapt sojourn time parameters represents an effective solution for heart sound segmentation problems, even when the training data does not perfectly express the variability of the testing data.

Acoustic detection results for a small unmanned aircraft system extrapolated over range

Small unmanned aircraft systems (SUAS) are becoming cheaper with more advanced reconnaissance and surveillance capabilities. Low-cost sensors and countermunitions are needed to defeat this asymmetric threat. The first step is to detect SUAS using low-cost sensors such as an array of microphones with robust signal processing algorithms. An analysis of six detection algorithms was performed using acoustic data generated by a class I SUAS at ranges of 50 to 650 m and measured with a small tetrahedral microphone array with 20-cm length arms. The detection algorithms were based upon the peak power in delay-and-sum, filtered delay-and-sum, and adaptive delay-and-sum beamforming algorithms. To test the performance of the algorithms at longer ranges, the measured signal, modeled as target plus additive noise, was modified. First, the signal from the target was attenuated using Bass’s model and spherical attenuation, and then, the noise was adjusted to maintain an unbiased estimate of the measured power spectrum density function. Receiver operation characteristics (ROC) curves were generated and the performance of the algorithms as a function of range was evaluated.

Performance evaluation of piezoelectric and differential pressure sensor for vortex flowmeters

Piezoelectric and transient differential pressure sensors are two among the most widely employed sensors for vortex flowmeter application. The present study evaluates the performance of these two techniques under fully developed and disturbed flow conditions. Firstly, the location of the transient differential pressure sensor is optimized to obtain high amplitude signals and good linearity in Strouhal number. Empirical mode decomposition method in combination with autocorrelation decay is successfully employed at high Reynolds numbers to identify the vortex shedding frequency in presence of hydrodynamic noise. The performance of the differential pressure sensor deteriorates significantly under disturbed flow conditions at low Reynolds number due to the presence of low frequency components. This deterioration in the signal quality limits the lower operating range of the flowmeter with differential pressure sensor. The output signals of the piezoelectric sensor and differential pressure sensor under no flow condition are compared to obtain the background noise due to piping vibrations and electrical interferences. These results will help a designer to suggest robust signal processing algorithms for vortex frequency detection.

Multisensor Processing Algorithms for Underwater Dipole Localization and Tracking Using MEMS Artificial Lateral-Line Sensors

An engineered artificial lateral-line system has been recently developed, consisting of a 16-element array of finely spaced MEMS hot-wire flow sensors. This represents a new class of underwater flow sensing instruments and necessitates the development of rapid, efficient, and robust signal processing algorithms. In this paper, we report on the development and implementation of a set of algorithms that assist in the localization and tracking of vibrational dipole sources underwater. Using these algorithms, accurate tracking of the trajectory of a moving dipole source has been demonstrated successfully.

Blind Carrier Frequency Tracking for Filterbank Multicarrier Wireless Communications

Filterbank multicarrier modulation (FBMCM) is an attractive technology for high-speed twisted-pair transmission and for broadband wireless communications, as well. In wireline applications, signal transmission takes place at baseband, so the issue of carrier acquisition and tracking for coherent demodulation does not apply. In wireless communications, on the contrary, carrier-frequency recovery reveals the Achille's heel of multicarrier modulation, so that robust signal processing algorithms are needed in this respect. In this paper, we derive a nondata-aided carrier-frequency offset recovery method for wireless FBMCM modems. In particular, we illustrate how to derive a low-complexity closed-loop tracker starting from a maximum-likelihood approach. We then show that the proposed simplifications do not entail large performance losses. In this respect, we derive the standard performance metrics of a closed-loop tracker (S-curve, root mean square estimation error, acquisition time) both on the additive white Gaussian noise channel and on a typical static frequency-selective wireless channel. We also demonstrate by simulation good robustness of the frequency tracker with respect to FBMCM symbol-timing errors.

Measurement of Rayleigh streaming in high-amplitude standing waves using laser Doppler anemometry

Current theories for Rayleigh streaming in a long, cylindrical pipe predict dc velocities which increase as the square of the amplitude of the first harmonic, and exhibit a parabolic radial profile. The velocity field in such a pipe, driven at one end and rigidly capped at the other, has been measured using LDA (laser Doppler anemometry) techniques. Analysis of the data produced by the LDA system can be complicated by two things: (i) the data are randomly sampled at a frequency below the Nyquist rate, resulting in a large, uneven sample spacing which is unsuitable for standard discrete Fourier transform techniques; and (ii) the dc velocity is typically two to three orders of magnitude smaller than the ac velocity, resulting in a very low signal-to-noise ratio. A robust signal processing algorithm is presented which overcomes these two obstacles, permitting fast and accurate estimates of the magnitudes, phases, and uncertainties of the velocity Fourier coefficients. The measured dc velocities agree well with the theoretical predictions when the tube is driven at low amplitudes, but deviate significantly from the theoretical predictions when the tube is driven at high amplitudes. [Work supported by ONR.]

A robust signal processing algorithm for linear displacement measuring optical transmission sensors

Optical triangulation sensors fall into a general category of noncontact height or displacement measurement devices and are widely used for their simple structure, high resolution, and long operating range. However, there are several factors that must be taken into account in order to obtain high accuracy and reliability: Measurement errors from inclinations of an object surface, probe signal fluctuations generated by speckle effects, power variation of a light source, electronic noises, and so on. In this article, we propose a signal processing algorithm, named expanded average square difference function (EASDF), for a newly designed optical triangulation sensor which is composed of an incoherent source, a line scan array detector, a specially selected diffuse reflecting surface, and several optical components. The EASDF, which is a modified correlation function, can calculate displacement between the probe and the object surface effectively even if there are inclinations, power fluctuations, and noises. This optical triangulation sensor system with the EASDF shows an 8 mm linear operation range and 1 μm resolution without averaging and the maximum measurement error is 2.4 μm for ±10° inclinations.

A FuzzyKth ordered statistic

Robust signal processing algorithms rely on integer ordered statistics. This means that such algorithms cannot be used in a fuzzy environment where variables have fuzzy ranks. In this paper, we show how to calculate the k th order statistic for a set of arguments with fuzzy rank and thereby use classical robust signal processing algorithms in a fuzzy environment. The papcr concludes with a simple application of using the fuzzy k th ordercd statistic to filtering a noisy signal.

Computer analysis of the electrocardiogram during esophageal pacing cardiac stress

It has been estimated that 15 to 30% of patients with suspected or known coronary artery disease are unable to perform an adequate exercise stress test due to a variety of reasons such as obesity, poor physical condition, claudication, etc. Transesophageal atrial pacing has been proposed as a noninvasive alternative for inducing cardiac stress in patients who cannot exercise. Although computer analysis is commonly employed to analyze the electrocardiogram (ECG) during the conventional exercise stress test, the surface ECG recorded during transesophageal atrial pacing is contaminated with large pacing artifacts which confound beat identification by standard computer software. We report the development of a robust signal processing algorithm for interpretation of the surface ECG during transesophageal atrial pacing stress. The algorithm employs novel schemes using both linear and nonlinear transformations to detect and differentiate between the pacing artifact and QRS complex even in difficult situations where the pacing artifact is in proximity to or superimposed on the QRS complex. The algorithm uses sophisticated logic for automatic recognition of sustained capture. It subsequently calculates beat-by-beat and average (over five beats) ST segment amplitude and slope. The algorithm also reports the instantaneous heart rate, RR interval, pace-to-R interval, R-wave amplitude, and estimated sinus node recovery time upon loss of sustained capture. The limitations of present exercise ECG computer methods in processing the ECG during transesophageal atrial pacing stress are evaluated and significantly improved performance by our algorithm is demonstrated.