Signal Segments Research Articles

Semantic Segmentation of 12-Lead ECG Using 1D Residual U-Net with Squeeze-Excitation Blocks

Analyzing biomedical data is a complex task that requires specialized knowledge. The development of knowledge and technology in the field of deep machine learning creates an opportunity to try and transfer human knowledge to the computer. In turn, this fact influences the development of systems for the automatic evaluation of the patient’s health based on data acquired from sensors. Electrocardiography (ECG) is a technique that enables visualizing the electrical activity of the heart in a noninvasive way, using electrodes placed on the surface of the skin. This signal carries a lot of information about the condition of heart muscle. The aim of this work is to create a system for semantic segmentation of the ECG signal. For this purpose, we used a database from Lobachevsky University available on Physionet, containing 200, 10-second, and 12-lead ECG signals with annotations, and applied one-dimensional U-Net with the addition of squeeze-excitation blocks. The created model achieved a set of parameters indicating high performance (for the test set: accuracy—0.95, AUC—0.99, specificity—0.95, sensitivity—0.99) in extracting characteristic parts of ECG signal such as P and T-waves and QRS complex, regardless of the lead.

Real-time modeling and feature extraction method of surface electromyography signal for hand movement classification based on oscillatory theory

Objective. Research of surface electromyography (sEMG) signal modeling and feature extraction is important in human motion intention recognition, prosthesis and exoskeleton robots. However, the existing methods mostly use the signal segmentation processing method rather than the point-to-point signal processing method, and lack physiological mechanism support. Approach. In this study, a real-time sEMG signal modeling and separation method is developed based on oscillatory theory. On this basis, an sEMG signal feature extraction method is constructed, and an ensemble learning method is combined to achieve real-time human hand motion intention recognition. Main results. The experimental results show that the average root mean square difference value of the sEMG signal modeling is 0.3838 ± 0.0591, and the average accuracy of human hand motion intention recognition is 96.03 ± 1.74%. On a computer with Intel (R) Core (TM) i5-8250U CPU running Matlab 2016Rb, the execution time for the sEMG signal with an actual duration of 2 s is 0.66 s. Significance. Compared with several existing methods, the proposed method has better modeling accuracy, motion intention recognition accuracy and real-time performance. The method developed in this study may provide a new perspective on sEMG modeling and feature extraction for hand movement classification.

Accurate 3D Reconstruction of White Matter Hyperintensities Based on Attention-Unet.

White matter hyperintensities (WMH), also known as white matter osteoporosis, have been clinically proven to be associated with cognitive decline, the risk of cerebral infarction, and dementia. The existing computer automatic measurement technology for the segmentation of patients' WMH does not have a good visualization and quantitative analysis. In this work, the author proposed a new WMH quantitative analysis and 3D reconstruction method for 3D reconstruction of high signal in white matter. At first, the author using ResUnet achieves the high signal segmentation of white matter and adds the attention mechanism into ResUnet to achieve more accurate segmentation. Afterwards, this paper used surface rendering to reconstruct the accurate segmentation results in 3D. Data experiments are conducted on the dataset collected from Shandong Province Third Hospital. After training, the Attention-Unet proposed in this paper is superior to other segmentation models in the segmentation of high signal in white matter and Dice coefficient and MPA reached 92.52% and 92.43%, respectively, thus achieving accurate 3D reconstruction and providing a new idea for quantitative analysis and 3D reconstruction of WMH.

Evaluation of sEMG Signal Features and Segmentation Parameters for Limb Movement Prediction Using a Feedforward Neural Network

Limb movement prediction based on surface electromyography (sEMG) for the control of wearable robots, such as active orthoses and exoskeletons, is a promising approach since it provides an intuitive control interface for the user. Further, sEMG signals contain early information about the onset and course of limb movements for feedback control. Recent studies have proposed machine learning-based modeling approaches for limb movement prediction using sEMG signals, which do not necessarily require domain knowledge of the underlying physiological system and its parameters. However, there is limited information on which features of the measured sEMG signals provide the best prediction accuracy of machine learning models trained with these data. In this work, the accuracy of elbow joint movement prediction based on sEMG data using a simple feedforward neural network after training with different single- and multi-feature sets and data segmentation parameters was compared. It was shown that certain combinations of time-domain and frequency-domain features, as well as segmentation parameters of sEMG data, improve the prediction accuracy of the neural network as compared to the use of a standard feature set from the literature.

EEG-based emotion charting for Parkinson's disease patients using Convolutional Recurrent Neural Networks and cross dataset learning

Electroencephalogram (EEG) based emotion classification reflects the actual and intrinsic emotional state, resulting in more reliable, natural, and meaningful human-computer interaction with applications in entertainment consumption behavior, interactive brain-computer interface, and monitoring of psychological health of patients in the domain of e-healthcare. Challenges of EEG-based emotion recognition in real-world applications are variations among experimental settings and cognitive health conditions. Parkinson's Disease (PD) is the second most common neurodegenerative disorder, resulting in impaired recognition and expression of emotions. The deficit of emotional expression poses challenges for the healthcare services provided to PD patients. This study proposes 1D-CRNN-ELM architecture, which combines one-dimensional Convolutional Recurrent Neural Network (1D-CRNN) with an Extreme Learning Machine (ELM), robust for the emotion detection of PD patients, also available for cross dataset learning with various emotions and experimental settings. In the proposed framework, after EEG preprocessing, the trained CRNN can use as a feature extractor with ELM as the classifier, and again this trained CRNN can be used for learning of new emotions set with fine-tuning of other datasets. This paper also applied cross dataset learning of emotions by training with PD patients datasets and fine-tuning with publicly available datasets of AMIGOS and SEED-IV, and vice versa. Random splitting of train and test data with 80 − 20 ratio resulted in an accuracy of 97.75% for AMIGOS, 83.20% for PD, and 86.00% for HC with six basic emotion classes. Fine-tuning of trained architecture with four emotions of the SEED-IV dataset results in 92.5% accuracy. To validate the generalization of our results, leave one subject (patient) out cross-validation is also incorporated with mean accuracies of 95.84% for AMIGOS, 75.09% for PD, 77.85% for HC, and 84.97% for SEED-IV is achieved. Only a 1 − sec segment of EEG signal from 14 channels is enough to detect emotions with this performance. The proposed method outperforms state-of-the-art studies to classify EEG-based emotions with publicly available datasets, provide cross dataset learning, and validate the robustness of the deep learning framework for real-world application of psychological healthcare monitoring of Parkinson's disease patients.

Dynamic brain functional network based on EEG microstate during sensory gating in schizophrenia

Objective. Cognitive impairment is one of the core symptoms of schizophrenia, with an emphasis on dysfunctional information processing. Sensory gating deficits have consistently been reported in schizophrenia, but the underlying physiological mechanism is not well-understood. We report the discovery and characterization of P50 dynamic brain connections based on microstate analysis. Approach. We identify five main microstates associated with the P50 response and the difference between the first and second click presentation (S1-S2-P50) in first-episode schizophrenia (FESZ) patients, ultra-high-risk individuals (UHR) and healthy controls (HCs). We used the signal segments composed of consecutive time points with the same microstate label to construct brain functional networks. Main results. The microstate with a prefrontal extreme location during the response to the S1 of P50 are statistically different in duration, occurrence and coverage among the FESZ, UHR and HC groups. In addition, a microstate with anterior–posterior orientation was found to be associated with S1-S2-P50 and its coverage was found to differ among the FESZ, UHR and HC groups. Source location of microstates showed that activated brain regions were mainly concentrated in the right temporal lobe. Furthermore, the connectivities between brain regions involved in P50 processing of HC were widely different from those of FESZ and UHR. Significance. Our results indicate that P50 suppression deficits in schizophrenia may be due to both aberrant baseline sensory perception and adaptation to repeated stimulus. Our findings provide new insight into the mechanisms of P50 suppression in the early stage of schizophrenia.

A new approach for daily life Blood-Pressure estimation using smart watch

Hypertension has become a major factor affecting people’s health. Timely monitoring and prevention of hypertension are of great significance. Most blood-pressure monitoring devices have redundant accessories and complex operations, which are not suitable for daily life applications. To solve these problems, we designed an application-level smart watch that can calculate the PAT feature using simultaneously collected ECG and PPG signals. By fusing the time–frequency-domain features of PPG and the correlation matrix between pulses, the information related to blood pressure in physiological signals was comprehensively extracted. Most of unstable signal segments can be eliminated and the impact of signal disturbances on overall accuracy will be reduced. This paper proposed two blood-pressure estimation modes: a two-step calib-free mode and a personal calibration mode based on transfer learning. The results on our self-established dataset (36 normotensive subjects and 40 hypertensive subjects) indicated that the two-step calib-free mode could estimate blood pressure with accuracy of 0.23 ± 6.1 mmHg and 0.63 ± 9.35 mmHg for DBP and SBP respectively, which was better than previous calib-free methods. Through the personal calibration mode, the estimation accuracy reached 0.02 ± 5.94 mmHg for DBP and 0.3 ± 7.69 mmHg for SBP, which complied with the Association for the Advancement of Medical Instrumentation standard.

A pace recognition method for exoskeleton wearers based on support vector machine-hidden Markov model

In order to improve the motion fluency and coordination of lower extremity exoskeleton robots and wearers, a pace recognition method of exoskeleton wearer is proposed base on inertial sensors. Firstly, the triaxial acceleration and triaxial angular velocity signals at the thigh and calf were collected by inertial sensors. Then the signal segment of 0.5 seconds before the current time was extracted by the time window method. And the Fourier transform coefficients in the frequency domain signal were used as eigenvalues. Then the support vector machine (SVM) and hidden Markov model (HMM) were combined as a classification model, which was trained and tested for pace recognition. Finally, the pace change rule and the human-machine interaction force were combined in this model and the current pace was predicted by the model. The experimental results showed that the pace intention of the lower extremity exoskeleton wearer could be effectively identified by the method proposed in this article. And the recognition rate of the seven pace patterns could reach 92.14%. It provides a new way for the smooth control of the exoskeleton.

Complex network-based pertussis and croup cough analysis: A machine learning approach

The paper proposes a novel approach to bring out the potential of complex networks based on graph theory to unwrap the hidden characteristics of cough signals, croup (BC), and pertussis (PS). The spectral and complex network analyses of 48 cough sounds are utilized for understanding the airflow through the infected respiratory tract. Among the different phases of the cough sound time-domain signals of BC and PS – expulsive (X), intermediate (I), and voiced (V) - the phase ‘I’ is noisy in BC due to improper glottal functioning. The spectral analyses reveal high-frequency components in both cough signals with an additional high-intense low-frequency spread in BC. The complex network features created by the correlation mapping approach, like number of edges (E), graph density (G), transitivity (Tr), degree centrality (D), average path length (L), and number of components (Nc) distinguishes BC and PS. The higher values of E, G, and Tr for BC indicate its musical nature through the strong correlation between the signal segments and the presence of high-intense low-frequency components in BC, unlike that in PS. The values of D, L, and Nc discriminate BC and PS in terms of the strength of the correlation between the nodes within them. The linear discriminant analysis (LDA) and quadratic support vector machine (QSVM) classifies BC and PS, with greater accuracy of 94.11% for LDA. The proposed work opens up the potentiality of employing complex networks for cough sound analysis, which is vital in the current scenario of COVID-19.

Effects of diffusion signal modeling and segmentation approaches on subthalamic nucleus parcellation

The subthalamic nucleus (STN) is commonly used as a surgical target for deep brain stimulation in movement disorders such as Parkinson's Disease. Tractography-derived connectivity-based parcellation (CBP) has been recently proposed as a suitable tool for non-invasive in vivo identification and pre-operative targeting of specific functional territories within the human STN. However, a well-established, accurate and reproducible protocol for STN parcellation is still lacking. The present work aims at testing the effects of different tractography-based approaches for the reconstruction of STN functional territories.We reconstructed functional territories of the STN on the high-quality dataset of 100 unrelated healthy subjects and on the test-retest dataset of the Human Connectome Project (HCP) repository. Connectivity-based parcellation was performed with a hypothesis-driven approach according to cortico-subthalamic connectivity, after dividing cortical areas into three groups: associative, limbic and sensorimotor. Four parcellation pipelines were compared, combining different signal modeling techniques (single-fiber vs multi-fiber) and different parcellation approaches (winner takes all parcellation vs fiber density thresholding). We tested these procedures on STN regions of interest obtained from three different, commonly employed, subcortical atlases. We evaluated the pipelines both in terms of between-subject similarity, assessed on the cohort of 100 unrelated healthy subjects, and of within-subject similarity, using a second cohort of 44 subjects with available test-retest data. We found that each parcellation provides converging results in terms of location of the identified parcels, but with significative variations in size and shape. All pipelines obtained very high within-subject similarity, with tensor-based approaches outperforming multi-fiber pipelines. On the other hand, higher between-subject similarity was found with multi-fiber signal modeling techniques combined with fiber density thresholding. We suggest that a fine-tuning of tractography-based parcellation may lead to higher reproducibility and aid the development of an optimized surgical targeting protocol.

Automated segmentation of magnetic resonance bone marrow signal: a feasibility study

BackgroundManual assessment of bone marrow signal is time-consuming and requires meticulous standardisation to secure adequate precision of findings.ObjectiveWe examined the feasibility of using deep learning for automated segmentation of bone marrow signal in children and adolescents.Materials and methodsWe selected knee images from 95 whole-body MRI examinations of healthy individuals and of children with chronic non-bacterial osteomyelitis, ages 6–18 years, in a longitudinal prospective multi-centre study cohort. Bone marrow signal on T2-weighted Dixon water-only images was divided into three color-coded intensity-levels: 1 = slightly increased; 2 = mildly increased; 3 = moderately to highly increased, up to fluid-like signal. We trained a convolutional neural network on 85 examinations to perform bone marrow segmentation. Four readers manually segmented a test set of 10 examinations and calculated ground truth using simultaneous truth and performance level estimation (STAPLE). We evaluated model and rater performance through Dice similarity coefficient and in consensus.ResultsConsensus score of model performance showed acceptable results for all but one examination. Model performance and reader agreement had highest scores for level-1 signal (median Dice 0.68) and lowest scores for level-3 signal (median Dice 0.40), particularly in examinations where this signal was sparse.ConclusionIt is feasible to develop a deep-learning-based model for automated segmentation of bone marrow signal in children and adolescents. Our model performed poorest for the highest signal intensity in examinations where this signal was sparse. Further improvement requires training on larger and more balanced datasets and validation against ground truth, which should be established by radiologists from several institutions in consensus.

Testing and estimation for clustered signals

We propose a change-point detection method for large scale multiple testing problems with data having clustered signals. Unlike the classic change-point setup, the signals can vary in size within a cluster. The clustering structure on the signals enables us to effectively delineate the boundaries between signal and non-signal segments. New test statistics are proposed for observations from one and/or multiple realizations. Their asymptotic distributions are derived. We also study the associated variance estimation problem. We allow the variances to be heteroscedastic in the multiple realization case, which substantially expands the applicability of the proposed method. Simulation studies demonstrate that the proposed approach has a favorable performance. Our procedure is applied to an array based Comparative Genomic Hybridization (aCGH) dataset.

Classification of Epileptic and Non-Epileptic Electroencephalogram (EEG) Signals Using Fractal Analysis and Support Vector Regression

Seizures are a common symptom of this neurological condition, which is caused by the discharge of brain nerve cells at an excessively fast rate. Chaos, nonlinearity, and other nonlinearities are common features of scalp and intracranial Electroencephalogram (EEG) data recorded in clinics. EEG signals that aren't immediately evident are challenging to categories because of their complexity. The Gradient Boost Decision Tree (GBDT) classifier was used to classify the majority of the EEG signal segments automatically. According to this study, the Hurst exponent, in combination with AFA, is an efficient way to identify epileptic signals. As with any fractal analysis approach, there are problems and factors to keep in mind, such as identifying whether or not linear scaling areas are present. These signals were classified as either epileptic or non-epileptic by using a combination of GBDT and a Support Vector Regression (SVR). The combined method's identification accuracy was 98.23%. This study sheds light on the effectiveness of AFA feature extraction and GBDT classifiers in EEG classification. The findings can be utilized to develop theoretical guidance for the clinical identification and prediction of epileptic EEG signals. Doi: 10.28991/ESJ-2022-06-01-011 Full Text: PDF

Exploring Physiological Signal Responses to Traffic-Related Stress in Simulated Driving.

In this paper, we propose a relatively noninvasive system that can automatically assess the impact of traffic conditions on drivers. We analyze the physiological signals recorded from a set of individuals while driving in a simulated urban scenario in two different traffic scenarios, i.e., with traffic and without traffic. The experiments were carried out in a laboratory located at the University of Udine, employing a driving simulator equipped with a moving platform. We acquired two Skin Potential Response (SPR) signals from the hands of the drivers, and an electrocardiogram (ECG) signal from their chest. In the proposed scheme, the SPR signals are then processed through a Motion Artifact (MA) removal algorithm such that possible motion artifacts arising during the drive are reduced. An analysis considering the scalogram of the single cleaned SPR signal is presented. This signal, along with the ECG, is then fed to various Machine Learning (ML) algorithms. More specifically, some statistical features are extracted from each signal segment which, after being analyzed through a binary ML model, are labeled as corresponding to a stressful situation or not. Our results confirm the applicability of the proposed approach to identify stress in the two scenarios. This is also in accordance with our findings considering the SPR signal scalograms.

Sequence polymorphism within erythrocyte binding domain of EBA175 in Indian and African field isolates

Invasion of erythrocyte by Plasmodium merozoites is mediated by specific molecular interactions between proteins expressed on merozoite surface and the receptors present on erythrocytes. Erythrocyte binding antigen 175 (EBA175) is one such protein that interacts with the sialic acid residues on glycophorin A present on erythrocytes’ surface during invasion. The FII region (PfFII) of EBA175 has been mapped to be critical for binding to erythrocytes. It is reported that antibodies against FII region blocks binding. Polymorphisms in FII region of EBA175 are already reported. The goal of this study was to investigate whether polymorphism in FII region of African P. falciparum field isolates has any effect on erythrocyte binding and also to find whether antibodies raised against FII region from P. falciparum Malayan Camp strain (Camp) can inhibit erythrocyte binding. Genomic DNA of parasites from the blood samples of P. falciparum infected individuals was isolated and PfFII region from these genomic DNA were amplified, cloned and sequenced. Following sequence analysis, we selected three isolates harboring higher PfFII polymorphisms, expressed them on the surface of COS cells as chimeric proteins using secretory signal and transmembrane segments of Herpex simplex virus glycoprotein D (HSVg D) and tested for their erythrocyte binding ability. We further tested the inhibition of erythrocyte binding of these polymorphic FII regions using anti-campPfF2 antibodies. Our results reveal that the polymorphisms in different field isolates included in this study do not have any significant effect on erythrocyte binding and antibodies raised against FII region of camp strain could inhibit erythrocyte binding by all the polymorphic PfFII. This observation strengthens the possibility that PfFII can be a potential candidate vaccine.

Vibration Separation Methodology Compensated by Time-Varying Transfer Function for Fault Diagnosis of Non-Hunting Tooth Planetary Gearbox.

Due to planetary movement of planet gears, the vibration signal perceived by a stationary sensor is modulated and difficult to diagnose. This paper proposed a vibration separation methodology compensated by a time-varying transfer function (TVTF-VS), which is a further development of the vibration separation (VS) method in the diagnosis of non-hunting tooth planetary gearboxes. On the basis of VS, multi-teeth VS is proposed to extract and synthesize the meshing signal of a planet gear using a single transducer. Considering the movement regularity of a planetary gearbox, the time-varying transfer function (TVTF) is represented by a generalized expression. The TVTF is constructed using a segment of healthy signal and an evaluation indicator is established to optimize the parameters of the TVTF. The constructed TVTF is applied to overcome the amplitude modulation effect and highlight fault characteristics. After that, experiments with baseline, pitting, and compound localized faults planet gears were conducted on a non-hunting tooth planetary gearbox test rig, respectively. The results demonstrate that incipient failure on a planet gear can be detected effectively, and relative location of the local faults can be determined accurately.

Real-time mental stress detection technique using neural networks towards a wearable health monitor

In recent times, detection of mental stress using physiological signals has received widespread attention from technology researchers. Although many interesting research works have been reported in this area, evidence of hardware implementation is occasional. The main challenge in stress detection research is use of the optimum number of physiological signals and real-time detection with low-complexity algorithms. In this work, a real-time stress detection technique is presented which utilizes only a photoplethysmogram (PPG) signal to achieve improved accuracy over multi-signal-based mental stress detection techniques. A short 5 s segment of a PPG signal was used for feature extraction with an autoencoder (AE), and features were minimized using recursive feature elimination (RFE) integrated with a multi-class support vector machine (SVM) classifier. The proposed AE-RFE-SVM-based mental stress detection technique was tested with the WeSAD dataset to detect four levels of mental state, namely baseline, amusement, meditation and stress, and achieved an overall accuracy, F1 score and sensitivity of 99%, 0.99 and 98%, respectively, for 5 s PPG data. The technique provided improved performance over discrete wavelet transformation-based feature extraction followed by classification with either of the five types of classifiers, namely SVM, random forest, k-nearest neighbor, linear regression and decision tree. The technique was translated into quad-core-based standalone hardware (1.2 GHz, 1 GB RAM). The resultant hardware prototype achieved low latency (∼0.4 s) and had a low memory requirement (∼1.7 MB). The present technique can be extended to develop a remote healthcare system using wearable sensors.

Misfire Detection Using Crank Speed and Long Short-Term Memory Recurrent Neural Network

In this work, a new approach was developed for the detection of engine misfire based on the long short-term memory recurrent neural network (LSTM RNN) using crank speed signal. The datasets are acquired from a six-cylinder-inline, turbo-charged diesel engine. Previous works investigated misfire detection in a limited range of engine running speed, running load or misfire types. In this work, the misfire patterns consist of normal condition, six types of one-cylinder misfire faults and fifteen types of two-cylinder misfire faults. All the misfire patterns are tested under wide range of running conditions of the tested engine. The traditional misfire detection method is tested on the datasets first, and the result show its limitation on high-speed low-load conditions. The LSTM RNN is a type of artificial neural network which has the ability of considering both the current input in-formation and the previous input information; hence it is helpful in extracting features of crank speed in which the misfire-induced speed fluctuation will last one or a few cycles. In order to select the engine operating conditions for network training properly, five data division strategies are attempted. For the sake of acquiring high performance of designed network, four types of network structure are tested. The results show that, utilizing the datasets in this work, the LSTM RNN based algorithm can overcome the limitation at high-speed low-load conditions of traditional misfire detection method. Moreover, the network which takes fixed segment of raw speed signal as input and takes misfire or fault-free labels as output achieves the best performance with the misfire diagnosis accuracy not less than 99.90%.

A Study on the Effect of Features Obtained From Signal Segments on Classification Success

Successful classification depends on the selection of the distinctive features and the effective channel subset used in the classification. In this study, novel and practical methods are proposed for determining the distinctive features and detecting effective channel subsets in the multi channel classification systems such as EEG. Two different feature extraction methods are compared in the study. The first one is based on classical Wavelet transform and the second is our proposed approach which used the slope of signal segments. Feature vectors are generated from some signal properties such as the mean, standard deviation, numerical integral of the Wavelet coefficients for classical Wavelet transform based feature extraction method. For our proposed method, only the slopes of signal segments are used for the feature vectors. In the proposed Signal Path Slope (SPS) feature extraction method, differently from the classical Wavelet based method, a Savitzky Golay (S-G) filter with an optimal frame length is applied to the signal before segmentation to make the path of the signal more prominent in time domain. In this way, the distinctive classification features are extracted by using S-G filter. For channel selection, an iterative channel selection method based on the classification results which divide the dataset labelled dataset into two groups as % 90 pre-training and %10 pre-test data is proposed. The dataset provided as dataset-3 in BCI competition IV is used in this study. The feature vectors extracted by using the proposed methods are classified for each method with the Support Vector Machine classifier. The results are given comparatively and it is observed that our proposed method has less computational complexity and more successful classification than Wavelet based classical feature extraction methods. The highest classification accuracies of % 67.74 and % 49.27 for subject-1 and subject-2 respectively are obtained with a low dimensional feature vector by proposed SPS feature extraction method. The classification accuracies achieved in the study are increased by % 8.24 for subject-1 and % 14.97 for subject-2 when compared average of the competition results. The significant increase in the success for both subjects shows the consistency of the proposed methods. By this study, it is observed that there is a subject-specific signal pattern related to motor imagery tasks in the brain. This pattern distinctive features is successfully determined by using the proposed methods.

A Signal Segmentation Approach to Identify Incident/Reflected Traveling Waves for Fault Location in Half-Bridge MMC-HVdc Grids

This article presents a new systematic technique for identifying voltage traveling waves (TWs) to determine the location of line faults in half-bridge modular multilevel converter-based high-voltage dc (HBMMC-HVdc) grids. In this technique, the buffered voltage signal frame around the fault-detection time is first scaled and then segmented via an optimization process. Finally, the incident/reflected TWs arrival times are obtained by executing a simple search algorithm on the reconstructed signal segments’ differences. This article describes how to use this technique in three forms of TW-based fault location schemes, including the single-ended scheme with known TW velocity, the double-ended scheme with known TW velocity, and the double-ended scheme with unknown TW velocity. The application results on a four-terminal HBMMC-HVdc grid simulated with exact component models show the proposed technique’s high capability and accuracy in all the three TW-based fault-location schemes. According to these results, the average fault-location errors are less than 0.5% for all the schemes. The numerical results also confirm that the proposed technique maintains its excellent performance, even in the face of close to terminal faults with distances down to 4 km, faults with high resistances up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$450 \Omega $ </tex-math></inline-formula> , and noisy signals with signal-to-noise ratios down to 55 dB. Moreover, the comparison results confirm that the proposed approach is more tolerant of measurement noise than the wavelet transform.