Human home daily living activities recognition based on a LabVIEW implemented hidden Markov model

The recognition of human daily living activities within a house represents an efficient tool to model its power consumption and is also a good indicator for monitoring the health status of the inhabitants. The problematic of activities recognition in smart homes has been extensively addressed in several studies. In this paper, we present an original interactive tool developed under LabVIEW environment with a graphical user interface allowing the modeling of the daily living activities, based on a machine learning Hidden Markov Model. After an overview of the advantage for the consideration of this model in current human activities, we examine how the associated scientific problematic can find an interest and a solution by the integration of machine learning tools. Thus, the application based on a Hidden Markov model approach, is presented and evaluated using two sets of experimental data from literature. Comparing with results obtained by other daily living activities recognition methods, we point out the very satisfactory recognition performance of the Hidden Markov Model and the likelihood of our development associated to a user-friendly graphical interface. This work opens the way to applications dedicated to the supervision of human daily living activities and / or to the management of the electrical consumption within a smart home equipped with non-intrusive sensors.