Wearable Full-Body Inertial Measurement with Task Classification Using Deep Learning

CARLOS M SORIA; EUGENIO OROSCO; JEREMIAS GAIA

IEEE LATIN AMERICA TRANSACTIONS

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC

Año: 2021

1548-0992

In this work, an embedded system is developed forthe non-invasive sensing and storage of biomechanical variables ofpeople. It takes advantage of wearable technology, distributingsensors in strategic points of the body, ergonomically andfunctionally. The results are verified by recording and analysingtasks performed by six subjects to form a database. These tasksinclude being stood up, sitting down or standing up from a chair,going upstairs and downstairs and walking. Additionally, aconvolutional neural network is tested for offline taskclassification. This work aims to initiate a process that ends inassistance-oriented applications, for the development of betterinjury rehabilitation techniques and support for elder people,among others. In this way, it seeks to open a path towards animprovement in the living conditions of people with and withoutreduced activities of daily living capacity.