Research on Collaborative Quality Assessment Model of Elbow Muscles based on MC-MMG and DRSN

The purpose of our study was to investigate the individual muscle contribution to generated force under four representative of elbow multi-muscle contraction tasks: flexion, extension, pronation, and supination. In this paper, we proposed a collaborative quality assessment model of muscles to elbow generated force based on a multi-channel mechanomyogram (MC-MMG) to explore the relationship between the elbow generated force and the individual muscles under different contraction tasks. Based on the analysis of elbow anatomy, MMG signals of brachial biceps (BB), brachial (BR), triceps (TR), brachioradialis (BRD) were collected by using MC-MMG collection platform. The Kernel Principal Component Analysis (KPCA) algorithm was used to reduce the dimension of the original MMG signal. Then, the Mean Average Value (MAV) feature of the signals was extracted as the input of the Deep Residual Shrinkage Network (DRSN), which is a new deep learning algorithm to establish the relationship between MC-MMG and generated force. Mean Impact Value (MIV) index was used to assess the contribution level of different muscles groups for estimating the generated force. The experimental results show that the single muscle with the highest MIV value can track the change of generated force better than multiple muscles under different contraction tasks. This result can provide effective guidance for estimating generated force and can be further applied to the recognition of motion intention.

[1]  Glen M. Davis,et al.  Estimation of Electrically-Evoked Knee Torque from Mechanomyography Using Support Vector Regression , 2016, Sensors.

[2]  F. J. Alonso,et al.  A comparison among different Hill-type contraction dynamics formulations for muscle force estimation , 2016 .

[3]  Chenguang Yang,et al.  Estimation of EMG-Based Force Using a Neural-Network-Based Approach , 2019, IEEE Access.

[4]  Lijuan Cao,et al.  A comparison of PCA, KPCA and ICA for dimensionality reduction in support vector machine , 2003, Neurocomputing.

[5]  Toshiyo Tamura,et al.  Deep Residual Networks for Sleep Posture Recognition With Unobtrusive Miniature Scale Smart Mat System , 2021, IEEE Transactions on Biomedical Circuits and Systems.

[6]  T. Housh,et al.  Mechanomyographic amplitude and mean power frequency versus torque relationships during isokinetic and isometric muscle actions of the biceps brachii. , 2004, Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology.

[7]  Darryl G. Thelen,et al.  Gauging force by tapping tendons , 2018, Nature Communications.

[8]  Prospects of Mechanomyography (MMG) in Muscle Function Assessment during FES Evoked Contraction: A Review , 2014 .

[9]  Michael Pecht,et al.  Deep Residual Shrinkage Networks for Fault Diagnosis , 2020, IEEE Transactions on Industrial Informatics.

[10]  Taorong Qiu,et al.  An Algorithm for Extracting Entropy Features from EEG Signals Based on T-test and KPCA and Its Application on Driving Fatigue State Recognition , 2020 .

[11]  Jeffrey R. Stout,et al.  Mechanomyographic amplitude and mean power frequency responses during isometric ramp vs. step muscle actions , 2008, Journal of Neuroscience Methods.

[12]  Philippe Poignet,et al.  EMG-to-force estimation with full-scale physiology based muscle model , 2009, 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems.