An Electromyography Based Intelligent System for Classification of Sitting and Standing Posture

Abstract

This paper presents an Electromyography (EMG) based system for classification of sitting and standing postures. The posture is classified by a machine learning model applied on the lower limb EMG data of the user. The dataset is collected from eight subjects, each with 8000 samples per channel, where six are used for training and two for testing. Time-domain, frequency-domain, and time-frequency-domain features are extracted for classification of sitting and standing postures. An array of algorithms are used for classification. Among all the classifiers Random Forest provided the highest accuracy at 98.38%.