Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
EMG-Driven Human-Exoskeleton Interaction Model for Knee Flexion and Extension Rehabilitation
Submission Author:
Luca Borgonovi , SP
Co-Authors:
Luca Borgonovi, Denis César Mosconi Pereira, Adriano Siqueira
Presenter: Luca Borgonovi
doi://10.26678/ABCM.COBEM2021.COB2021-0192
Abstract
The use of computational models is extremely important for robotic rehabilitation, as it allows the simulation of its dynamic behavior before building the device represented by the model, ensuring its safe performance, an essential feature for the use of the instrument in physical therapy treatments. Using this approach of dynamic systems modeling, this work aims to present and simulate a human-exoskeleton interaction model for knee flexion and extension equipped with an active orthosis, estimating its the human torque having EMG signals as inputs. The importance of this work is the development and analysis of a biomechanical model that takes into account biological aspects of the human being, based on the EMG signals, instead of considering a human body as a multibody system subject only to torques, which is, in its nature, purely mechanical. In order to develop biomechanical relationships, the EMG signals are converted into muscle activations and then torques exerted by the muscles. These require an interaction model, developed in OpenSim software, to obtain the angular positions of the system during movement. Finally, the accuracy of the EMG signal method is compared with another more traditional method, that of Inverse Dynamics, and with experimental data collected by sensors. The use of active EMG signals is effective in estimating the dynamic behavior of the movement, but the torque values derived from the Inverse Dynamics are more realistic.
Keywords
Human-Robot Interaction, OpenSim, Wearable Robots
