Emulación de movimiento de la extremidad superior para determinar posición.

Abstract

The recognition of human activities is a field that is gaining relevance in recent years, with the aim of facilitating the study of human physiology and human motor behavior to perform telemedicine-oriented applications. With this precedent, the present project proposes a tool for the recognition of movements and positions of the upper limb, specifically of a mannequin, which emulates the movements of a person. This detection is achieved with the use of inertial measurement sensors located along the upper limb. Motion determination will be achieved with the use of machine learning. In principle, we will obtain a database made up of signals from the inertial sensors that will later have and data processing and thus proceed to a respective training and validation of these for each classifier that MATLAB provides. Finally, a comparison will be made with the main machine learning classifiers with better results, which in this specific case are KNN, SVM and neural networks; for the determination of one of these models for the final application that will be exposed through an interface made in the MATLAB environment.