Identificação automática de situações de emergência através de técnicas de fusão de sinais vitais e de movimentos

Abstract

The main aim of this thesis is the development of a multimodal system for monitoring aged people and/or people with cardiac diseases, as well as the development of a physiological data simulation software. A system for automatic identification of urgency situations is proposed, which is based on the application of probabilistic networks (with the use of the Bayesian Network) for the vital signals and movement fusion. These signals are supplied by a telemonitoring system of patients in domicile, called TELEPAT. Besides the methodology used to get the classification of the sensor signals, the advantages of using probabilistic networks are shown. Finally, the application of this system as an important tool to assist aged or patients with cardiac diseases is demonstrated by experiments. For the formation of synthetic physiological databases, the development of a data simulation software by using normal signals disturbed artificially (in accordance with the profiles of alarming situations) was done. The simulated signals can be used in the training and test stages of the Bayesian Network for data fusion, and also attend a number of software, with other purposes. The data simulation and data fusion software had been developed in Matlab environment.