A novel framework for designing a multi-DoF prosthetic wrist control using machine learning

Chinmay P Swami,Jiyeon Kang,Nicholas Lenhard

doi:10.1038/s41598-021-94449-1

Chinmay P Swami, Jiyeon Kang + Show 1 more

Open Access

https://doi.org/10.1038/s41598-021-94449-1

Copy DOI

Abstract

Prosthetic arms can significantly increase the upper limb function of individuals with upper limb loss, however despite the development of various multi-DoF prosthetic arms the rate of prosthesis abandonment is still high. One of the major challenges is to design a multi-DoF controller that has high precision, robustness, and intuitiveness for daily use. The present study demonstrates a novel framework for developing a controller leveraging machine learning algorithms and movement synergies to implement natural control of a 2-DoF prosthetic wrist for activities of daily living (ADL). The data was collected during ADL tasks of ten individuals with a wrist brace emulating the absence of wrist function. Using this data, the neural network classifies the movement and then random forest regression computes the desired velocity of the prosthetic wrist. The models were trained/tested with ADLs where their robustness was tested using cross-validation and holdout data sets. The proposed framework demonstrated high accuracy (F-1 score of 99% for the classifier and Pearson’s correlation of 0.98 for the regression). Additionally, the interpretable nature of random forest regression was used to verify the targeted movement synergies. The present work provides a novel and effective framework to develop an intuitive control for multi-DoF prosthetic devices.

Highlights

Prosthetic arms can significantly increase the upper limb function of individuals with upper limb loss, despite the development of various multi-degrees of freedom (DoF) prosthetic arms the rate of prosthesis abandonment is still high
Random forest regression models were trained with inertial measurement unit (IMU) signals as inputs and the measured angular velocity as outputs
The aim of this study is to demonstrate the feasibility of controlling multi-DoF wrist of a prosthetic arm by using the residual upper limb motion

Summary

Introduction

Prosthetic arms can significantly increase the upper limb function of individuals with upper limb loss, despite the development of various multi-DoF prosthetic arms the rate of prosthesis abandonment is still high. The present study demonstrates a novel framework for developing a controller leveraging machine learning algorithms and movement synergies to implement natural control of a 2-DoF prosthetic wrist for activities of daily living (ADL). One of the widely explored methods is the state machine approach which used two EMG signals to control single joint but allowed for switching between different joints by co-activation of both m uscles[13] These approaches lacked intuitive and simultaneous control of multiple DoF which hindered the dexterity of the hand movement during daily living tasks. Performance of EMG based controllers is limited due to electrode shift, variation in the force from the different pose, and transient changes in EMG due to muscle fatigue from long-term u se[17] These limitations have rekindled the search for alternative approaches to control multi-DoF prosthetic devices. Using the movement synergies between the wrist and the shoulder, the controller allowed natural and intuitive control of prosthetic wrist pronation/supination and reduced the cognitive burden on the prosthesis u ser[24]

Objectives

Methods

Results

Discussion

Conclusion