IoT System Based on Deep Learning for the Identification and Feedback of Work Postures When Using a Computer

Abstract

It is common for office workers, mostly dedicated to IT, to present musculoskeletal pain in the back, neck and shoulders due to poor posture practices they adopt while doing their work in front of the computer for long periods, this is known as forced postures. Our main work seeks to implement an IoT system with force sensors, model RP-S40-ST, based on the use of classification algorithms and deep learning techniques for the identification and correction of postures through feedback. Ten classification algorithms were used for training and validation of the model, with the Logistic Regression algorithm achieving the highest accuracy rate being .8794 and .9052 respectively.