Abstract:To mitigate the difficulties associated with the extraction of Mechanomyography (MMG) signals from raw Accelerometer (ACC) data and the subsequent classification of human lower limb activities based on MMG signals, the Feature Mode Decomposition (FMD) algorithm has been utilized for the isolation of the MMG signal. Simultaneously, surface Electromyography (sEMG) signals were recorded to perform correlation analyses, thereby validating the effectiveness of the extracted Mechanomyography (MMG) signals. The results demonstrate that the envelope entropy derived from the FMD was the lowest among the observed values, and the composite signal obtained via FMD displayed the highest correlation with the sEMG signal. This indicates that FMD is capable of efficiently isolating the MMG signal while maintaining the maximal quantity of muscle contraction data. To address the challenge of classifying human lower limb activities, a comprehensive feature extraction procedure was implemented, resulting in the derivation of 448 unique features from multi-channel mechanomyography (MMG) signals. Subsequently, Kernel Principal Component Analysis (KPCA) was employed to diminish the feature set's dimensionality. This was succeeded by the deployment of a Temporal Convolutional Network integrated with an Attention mechanism (TCN-Attention) to train the classification model. Additionally, an enhanced Northern Goshawk Optimization Algorithm was leveraged for optimization purposes. The findings indicate that FMD exhibited the minimum envelope entropy value of 8.13, concurrently attaining the maximum correlation coefficient of 0.87 between MMG and sEMG signals. Significantly, the SCNGO-TCN-Attention model demonstrated superior classification accuracy, attaining an exceptional accuracy rate of 98.44%.

What problem does this paper attempt to address?

This paper aims to solve the difficulties in extracting myomechanical graph (MMG) signals from raw accelerometer (ACC) data and classifying human lower - limb activities based on MMG signals. Specifically: 1. **MMG Signal Extraction**: - The paper uses the Feature Mode Decomposition (FMD) algorithm to isolate MMG signals. The FMD algorithm can effectively extract MMG signals from raw accelerometer data and has shown good performance in multiple fields such as multi - sensor fault diagnosis and bearing fault diagnosis. - To verify the effectiveness of the FMD algorithm, the study also recorded the surface electromyography (sEMG) signal and carried out a correlation analysis. The results show that the correlation coefficient between the composite signal obtained by FMD and the sEMG signal is the highest, indicating that FMD can efficiently separate MMG signals while retaining the most muscle contraction data. 2. **Lower - Limb Activity Classification**: - The paper proposes a comprehensive feature extraction method and extracts 448 unique features from multi - channel MMG signals. - Kernel Principal Component Analysis (KPCA) is used to reduce the dimension of the feature set to improve the efficiency of model training. - The Sine - Cosine Northern Goshawk Optimization (SCNGO) is adopted to optimize the parameters of the Temporal Convolutional Network with Attention (TCN - Attention) model to improve the classification accuracy. - The experimental results show that the minimum envelope entropy of the FMD algorithm is 8.13, the maximum correlation coefficient between MMG signals and sEMG signals is 0.87. In addition, the classification accuracy of the SCNGO - TCN - Attention model reaches 98.44%. In conclusion, this paper mainly solves the problem of efficiently extracting MMG signals from raw accelerometer data and achieves high - precision classification of human lower - limb activities through an optimized deep - learning model.

An Investigation into Mechanomyography for Signal Extraction and Classification of Human Lower Limb Activity

Feature Selection and Reduction of Lower Limb Activity Recognition Based on Surface Electromyography and Motion Data

Prediction of Lower Limb Action Intention Based on Surface EMG Signal

Variational Mode Decomposition and Empirical Wavelet Transform-Based Feature Extraction and Ensemble Classifier for Lower Limb Movement Prediction With Surface Electromyography Signal

Denoising Surface Electromyography Signals Based on Complementary Ensemble Empirical-Mode Decomposition and Multi-Scale Entropy

Measuring Motor Unit Discharge, Myofiber Vibration, and Haemodynamics for Enhanced Myoelectric Gesture Recognition.

Classification of Individual Finger Motions Using Single-Site Mechanomyography by Optimized Long Short-Time Memory

Enhance Decoding of Functional Lower-Limb Movements by Combining Sensory Motor Rhythm and Movement-Related Cortical Potential Features

Hand-motion patterns recognition based on mechanomyographic signal analysis

Hardware Implementation for Lower Limb Surface EMG Measurement and Analysis Using Explainable AI for Activity Recognition

Effectiveness of Surface Electromyography in Pattern Classification for Upper Limb Amputees

Multi-modality deep forest for hand motion recognition via fusing sEMG and acceleration signals

A novel feature extraction method for machine learning based on surface electromyography from healthy brain

Development of a Hybrid Surface EMG and MMG Acquisition System for Human Hand Motion Analysis.

Multi-Attention Feature Fusion Network for Accurate Estimation of Finger Kinematics from Surface Electromyographic Signals

Classification of the Mechanomyogram Signal Using a Wavelet Packet Transform and Singular Value Decomposition for Multifunction Prosthesis Control

Multichannel Finger Pattern Recognition Using Single-Site Mechanomyography

Decoding finger movement patterns from microscopic neural drive information based on deep learning

Multimodal Human-Exoskeleton Interface for Lower Limb Movement Prediction Through a Dense Co-Attention Symmetric Mechanism.

Non-Invasive Analysis of Motor Unit Activation During Simultaneous and Continuous Wrist Movements

Kernel Density Estimation of Electromyographic Signals and Ensemble Learning for Highly Accurate Classification of a Large Set of Hand/Wrist Motions