Evaluation of Feature Extraction and Classification for Lower Limb Motion Based on sEMG Signal

A survey on the state of the art of force myography technique (FMG): analysis and assessment

Precise feedback assures precise control commands especially for assistive or rehabilitation devices. Biofeedback systems integrated with assistive or rehabilitative robotic exoskeletons tend to increase its performance and effectiveness. Therefore, there has been plenty of research in the field of biofeedback covering different aspects such as signal acquisition, conditioning, feature extraction and integration with the control system. Among several types of biofeedback systems, Force myography (FMG) technique is a promising one in terms of affordability, high classification accuracies, ease to use, and low computational cost. Compared to traditional biofeedback systems such as electromyography (EMG) which offers some invasive techniques, FMG offers a completely non-invasive solution with much less effort for preprocessing with high accuracies. This work covers the whole aspects of FMG technique in terms of signal acquisition, feature extraction, signal processing, developing the machine learning model, evaluating tools for the performance of the model. Stating the difference between real-time and offline assessment, also highlighting the main uncovered points for further study, and thus enhancing the development of this technique.

User- and Speed-Independent Slope Estimation for Lower-Extremity Wearable Robots

Wearable robots can help users traverse unstructured slopes by providing mode-specific hip, knee, and ankle joint assistance. However, generalizing the same assistance pattern across different slopes is not optimal. Control strategies that scale assistance based on slope are expected to improve the feel of the device and improve outcome measures such as decreasing metabolic cost. Prior numerical methods for slope estimation struggled to estimate slopes at variable walking speeds or were limited to a single estimation per gait cycle. This study overcomes these limitations by developing machine-learning methods that yield continuous, user- and speed-independent slope estimators for a variety of wearable robot applications using an able-bodied wearable sensor dataset. In a leave-one-subject-out cross-validation (N = 9), four-phase XGBoost regression models were trained on static-slope (fixed-slope) data and evaluated on a novel subject's static-slope and dynamic-slope (variable-slope) data. Using all available sensors, we achieved an average error of 0.88° and 1.73° mean absolute error (MAE) on static and dynamic slopes, respectively. Ankle prosthesis, knee-ankle prosthesis, and hip exoskeleton sensor suites yielded average errors under 2° MAE on static and dynamic slopes, except for the ankle prosthesis and hip exoskeleton cases on dynamic slopes which yielded an average error of 2.2° and 3.2° MAE, respectively. We found that the thigh inertial measurement unit contributed the most to a reduction in average error. Our findings suggest that reliable slope estimators can be trained using only static-slope data regardless of the type of lower-extremity wearable robot.

The Impact of Feature Extraction on Classification Accuracy Examined by Employing a Signal Transformer to Classify Hand Gestures Using Surface Electromyography Signals

Interest in developing techniques for acquiring and decoding biological signals is on the rise in the research community. This interest spans various applications, with a particular focus on prosthetic control and rehabilitation, where achieving precise hand gesture recognition using surface electromyography signals is crucial due to the complexity and variability of surface electromyography data. Advanced signal processing and data analysis techniques are required to effectively extract meaningful information from these signals. In our study, we utilized three datasets: NinaPro Database 1, CapgMyo Database A, and CapgMyo Database B. These datasets were chosen for their open-source availability and established role in evaluating surface electromyography classifiers. Hand gesture recognition using surface electromyography signals draws inspiration from image classification algorithms, leading to the introduction and development of the Novel Signal Transformer. We systematically investigated two feature extraction techniques for surface electromyography signals: the Fast Fourier Transform and wavelet-based feature extraction. Our study demonstrated significant advancements in surface electromyography signal classification, particularly in the Ninapro database 1 and CapgMyo dataset A, surpassing existing results in the literature. The newly introduced Signal Transformer outperformed traditional Convolutional Neural Networks by excelling in capturing structural details and incorporating global information from image-like signals through robust basis functions. Additionally, the inclusion of an attention mechanism within the Signal Transformer highlighted the significance of electrode readings, improving classification accuracy. These findings underscore the potential of the Signal Transformer as a powerful tool for precise and effective surface electromyography signal classification, promising applications in prosthetic control and rehabilitation.

Classification of EMG signals with CNN features and voting ensemble classifier

Electromyography (EMG) signals are primarily used to control prosthetic hands. Classifying hand gestures efficiently with EMG signals presents numerous challenges. In addition to overcoming these challenges, a successful combination of feature extraction and classification approaches will improve classification accuracy. In the current work, convolutional neural network (CNN) features are used to reduce the redundancy problems associated with time and frequency domain features to improve classification accuracy. The features from the EMG signal are extracted using a CNN and are fed to the 'k' nearest neighbor (KNN) classifier with a different number of neighbors ( 1 N N , 3 N N , 5 N N , and   7 N N ) . It results in an ensemble of classifiers that are combined using a hard voting-based classifier. Based on the benchmark Ninapro DB4 database and CapgMyo database, the proposed framework obtained 91.3 % classification accuracy on CapgMyo and 89.5 % on Ninapro DB4.

sEMG-based Sarcopenia risk classification using empirical mode decomposition and machine learning algorithms

Early detection of the risk of sarcopenia at younger ages is crucial for implementing preventive strategies, fostering healthy muscle development, and minimizing the negative impact of sarcopenia on health and aging. In this study, we propose a novel sarcopenia risk detection technique that combines surface electromyography (sEMG) signals and empirical mode decomposition (EMD) with machine learning algorithms. First, we recorded and preprocessed sEMG data from both healthy and at-risk individuals during various physical activities, including normal walking, fast walking, performing a standard squat, and performing a wide squat. Next, electromyography (EMG) features were extracted from a normalized EMG and its intrinsic mode functions (IMFs) were obtained through EMD. Subsequently, a minimum redundancy maximum relevance (mRMR) feature selection method was employed to identify the most influential subset of features. Finally, the performances of state-of-the-art machine learning (ML) classifiers were evaluated using a leave-one-subject-out cross-validation technique, and the effectiveness of the classifiers for sarcopenia risk classification was assessed through various performance metrics. The proposed method shows a high accuracy, with accuracy rates of 0.88 for normal walking, 0.89 for fast walking, 0.81 for a standard squat, and 0.80 for a wide squat, providing reliable identification of sarcopenia risk during physical activities. Beyond early sarcopenia risk detection, this sEMG-EMD-ML system offers practical values for assessing muscle function, muscle health monitoring, and managing muscle quality for an improved daily life and well-being.

Mechanomyography Signal Pattern Recognition of Knee and Ankle Movements Using Swarm Intelligence Algorithm-Based Feature Selection Methods

Pattern recognition of lower-limb movements based on mechanomyography (MMG) signals has a certain application value in the study of wearable rehabilitation-training devices. In this paper, MMG feature selection methods based on a chameleon swarm algorithm (CSA) and a grasshopper optimization algorithm (GOA) are proposed for the pattern recognition of knee and ankle movements in the sitting and standing positions. Wireless multichannel MMG acquisition systems were designed and used to collect MMG movements from four sites on the subjects thighs. The relationship between the threshold values and classification accuracy was analyzed, and comparatively high recognition rates were obtained after redundant information was eliminated. When the threshold value rose, the recognition rates from the CSA fluctuated within a small range: up to 88.17% (sitting position) and 90.07% (standing position). However, the recognition rates from the GOA drop dramatically when increasing the threshold value. The comparison results demonstrated that using a GOA consumes less time and selects fewer features, while a CSA gives higher recognition rates of knee and ankle movements.

Wearable Electromyography Classification of Epileptic Seizures: A Feasibility Study

Accurate diagnosis and classification of epileptic seizures can greatly support patient treatments. As many epileptic seizures are convulsive and have a motor component, the analysis of muscle activity can provide valuable information for seizure classification. Therefore, this paper present a feasibility study conducted on healthy volunteers, focusing on tracking epileptic seizures movements using surface electromyography signals (sEMG) measured on human limb muscles. For the experimental studies, first, compact wireless sensor nodes were developed for real-time measurement of sEMG on the gastrocnemius, flexor carpi ulnaris, biceps brachii, and quadriceps muscles on the right side and the left side. For the classification of the seizure, a machine learning model has been elaborated. The 16 common sEMG time-domain features were first extracted and examined with respect to discrimination and redundancy. This allowed the features to be classified into irrelevant features, important features, and redundant features. Redundant features were examined with the Big-O notation method and with the average execution time method to select the feature that leads to lower complexity and reduced processing time. The finally selected six features were explored using different machine learning classifiers to compare the resulting classification accuracy. The results show that the artificial neural network (ANN) model with the six features: IEMG, WAMP, MYOP, SE, SKEW, and WL, had the highest classification accuracy (99.95%). A further study confirms that all the chosen eight sensors are necessary to reach this high classification accuracy.

Electromyography Parameter Variations with Electrocardiography Noise

Electromyograms (EMG signals) may be contaminated by electrocardiographic (ECG) signals that cannot be easily separated with traditional filters, because both signals have some overlapping spectral components. Therefore, the first challenge encountered in signal processing is to extract the ECG noise from the EMG signal. In this study, the EMG, mixed with different degrees of noise (ECG), is simulated to investigate the variations of the EMG features. Simulated data were derived from the MIT-BIH Noise Stress Test (NSTD) Database. Two EMG and four ECG data were composed with four EMG/ECG SNR to 32 simulated signals. Following Pan-Tompkins R-peak detection, four ECG removal methods were used to remove ECG with different compensation algorithms to obtain the denoised EMG signal. A total of 13 time-domain and four frequency-domain EMG features were calculated from the denoised EMG. In addition, the similarity of denoised EMG features compared to clean EMG was also evaluated. Our results showed that with the ratio EMG/ECG SNR = 10 and 20, the ECG can be almost ignored, and the similarity of EMG features is close to 1. When EMG/ECG SNR = 1 and 2, there is a large variation of EMG features. The results of our simulation study would be beneficial for understanding the variations of EMG features upon the different EMG/ECG SNR.

Analysing the effect of robotic gait on lower extremity muscles and classification by using deep learning

Robotic gait training helps the nervous system recover and strengthen weak muscle groups. Many studies in the literature show that applying robotic gait rehabilitation to patients with neurological disorders such as Multiple Sclerosis (MS), Stroke and Spinal Cord Injection (SCI) effectively restores gait ability. In contrast to the studies in the literature that included only healthy individuals, both the control and patient groups were formed and detailed analyses were carried out for both groups. In this study, EMG signals in GMA, GME, ILP, BF, VM, MG, TA muscles were recorded simultaneously with a different electrode placement during robotic gait for the first time in literature and then a location that prevents a phase shift was presented. The classification performance has also been increased by removing 26 different attribute parameters like time, frequency and statistics from the signals instead of gait studies with a maximum of 12-16 traits extraction. The extracted features were classified with the approaches Multilayer Perceptron Neural Networks (MLP), Support Vector Machines (SVM), K-Nearest Neighbourhood algorithm (KNN), Random Forest Classification Algorithm (RF) and Deep Learning and then a detailed performance comparison have been realized. Among the approaches compared the Stochastic Gradient Optimization Algorithm-based deep learning structure produced the best performance with 98.5714% accuracy. It was also seen that it is essential to plan the exoskeleton and the robotic gait pattern suitable for patients' disease state and muscle activation.

Genetic Algorithm for Feature Selection in Lower Limb Pattern Recognition

Choosing the right features is important to optimize lower limb pattern recognition, such as in prosthetic control. EMG signals are noisy in nature, which makes it more challenging to extract useful information. Many features are used in the literature, which raises the question which features are most suited for use in lower limb myoelectric control. Therefore, it is important to find combinations of best performing features. One way to achieve this is by using a genetic algorithm, a meta-heuristic capable of searching vast feature spaces. The goal of this research is to demonstrate the capabilities of a genetic algorithm and come up with a feature set that has a better performance than the state-of-the-art feature set. In this study, we collected a dataset containing ten able-bodied subjects who performed various gait-related activities while measuring EMG and kinematics. The genetic algorithm selected features based on the performance on the training partition of this dataset. The selected feature sets were evaluated on the remaining test set and on the online benchmark dataset ENABL3S, against a state-of-the-art feature set. The results show that a feature set based on the selected features of a genetic algorithm outperforms the state-of-the-art set. The overall error decreased up to 0.54% and the transitional error by 2.44%, which represent a relative decrease in overall errors up to 11.6% and transitional errors up to 14.1%, although these results were not significant. This study showed that a genetic algorithm is capable of searching a large feature space and that systematic feature selection shows promising results for lower limb myoelectric control.

Development of a Neurodegenerative Disease Gait Classification Algorithm Using Multiscale Sample Entropy and Machine Learning Classifiers

The prevalence of neurodegenerative diseases (NDD) has grown rapidly in recent years and NDD screening receives much attention. NDD could cause gait abnormalities so that to screen NDD using gait signal is feasible. The research aim of this study is to develop an NDD classification algorithm via gait force (GF) using multiscale sample entropy (MSE) and machine learning models. The Physionet NDD gait database is utilized to validate the proposed algorithm. In the preprocessing stage of the proposed algorithm, new signals were generated by taking one and two times of differential on GF and are divided into various time windows (10/20/30/60-sec). In feature extraction, the GF signal is used to calculate statistical and MSE values. Owing to the imbalanced nature of the Physionet NDD gait database, the synthetic minority oversampling technique (SMOTE) was used to rebalance data of each class. Support vector machine (SVM) and k-nearest neighbors (KNN) were used as the classifiers. The best classification accuracies for the healthy controls (HC) vs. Parkinson’s disease (PD), HC vs. Huntington’s disease (HD), HC vs. amyotrophic lateral sclerosis (ALS), PD vs. HD, PD vs. ALS, HD vs. ALS, HC vs. PD vs. HD vs. ALS, were 99.90%, 99.80%, 100%, 99.75%, 99.90%, 99.55%, and 99.68% under 10-sec time window with KNN. This study successfully developed an NDD gait classification based on MSE and machine learning classifiers.

Composition Classification of Ultra-High Energy Cosmic Rays

The study of cosmic rays remains as one of the most challenging research fields in Physics. From the many questions still open in this area, knowledge of the type of primary for each event remains as one of the most important issues. All of the cosmic rays observatories have been trying to solve this question for at least six decades, but have not yet succeeded. The main obstacle is the impossibility of directly detecting high energy primary events, being necessary to use Monte Carlo models and simulations to characterize generated particles cascades. This work presents the results attained using a simulated dataset that was provided by the Monte Carlo code CORSIKA, which is a simulator of high energy particles interactions with the atmosphere, resulting in a cascade of secondary particles extending for a few kilometers (in diameter) at ground level. Using this simulated data, a set of machine learning classifiers have been designed and trained, and their computational cost and effectiveness compared, when classifying the type of primary under ideal measuring conditions. Additionally, a feature selection algorithm has allowed for identifying the relevance of the considered features. The results confirm the importance of the electromagnetic-muonic component separation from signal data measured for the problem. The obtained results are quite encouraging and open new work lines for future more restrictive simulations.