Open AccessNeural model of human gait and its implementation in MATLAB and Simulink Environment using Deep Learning Toolbox
Author(s) -
Jerzy Tchórzewski,
Arkadiusz Wielgo
Publication year - 2021
Publication title -
studia informatica
Language(s) - English
Resource type - Journals
ISSN - 1731-2264
DOI - 10.34739/si.2021.25.03
Subject(s) - artificial neural network , computer science , matlab , artificial intelligence , displacement (psychology) , trajectory , toolbox , perceptron , gait , simulation , deep learning , machine learning , physiology , psychology , physics , astronomy , psychotherapist , biology , programming language , operating system
The article presents selected results of research on the modeling of humanoid robots, including the results of neural modeling of human gait and its implementation in the environment MATLAB and Simulink with the use of Deep Learning Toolbox. The subject of the research was placed within the scope of the available literature on the subject. Then, appropriate research experiments on human movement along a given trajectory were developed. First, the method of measuring the parameters present in the experiment was established, i.e. input quantities (displacement of the left heel, displacement of the right heel) and output quantities (displacement of the measurement point of the human body in space). Then, research experiments were carried out, as a result of which numerical data were measured in order to use them for teaching and testing the Artificial Neural Network. The Perceptron Artificial Neural Network architecture was used to build a model of a neural human walk along a given trajectory. The obtained results were discussed and interpreted, drawing a number of important conclusions.