Algoritmos para la Modelización y el Reconocimiento de Actividades de la Vida Diaria

Abstract

The growth of the number of elderly people in the total composition of a population has proven to be in the last decades a process that occurs with a considerable speed. Activity monitoring technologies based on sensors networks
constitute a global solution to provide autonomy and quality of life for adults. The simulation allows to generate synthetic data sets corresponding to years of life of a person with minimum effort and time and thus to achieve the initial training of the algorithms of machine-learning present in the systems of monitoring. The addition of climatic influences to the simulation model through diffuse logic allows to contemplate the variations in the activities of a person. The recognition of daily activities with these influences can be realized through an artificial neural network maintaining a high level of precision.