Dans cet article, nous proposons une approche pour reconnaitre certaines activités physiques en utilisant un réseau d'objets connectés. L'approche consiste à classer certaines activités humaines : marcher, debout, assis et allonger. Cette étude utilise un réseau d'objets connectés usuels: une montre connectée, un smartphone et une télécommande connectée. Ces objets sont portés par les participants lors d'une expérience non contrôlée. Les données des capteurs des trois dispositifs ont été classées par un algorithme du DNN (Deep Neural Networks) sans prétraitement préalable des données d'entrée. Nous montrons que (DNN) fournit de meilleurs résultats par rapport aux autres algorithmes classiques de type arbres de décision (DT) et Support Vector Machine (SVM). Les résultats montrent également que les activités des participants ont été classées avec une précision de plus de 98,53%, en moyenne. ABSTRACT. This paper proposes to study the recognition of certain daily physical activities by using a network of smart objects. The approach consists in the classification of certain participants' activities, the most common ones and those that are carried out with smart objects:Make a phone call (Call), open the door (Open), close the door (Close) and watch its smartwatch (Watch). The study exploits a network of commonly connected objects: a smart watch and a smartphone, transported by participants during an uncontrolled experiment. The sensors' data of the two devices were classified by a deep neural network (DNN) algorithm without prior data pre-processing. We show that DNN provides better results than Decision Tree (DT) and Support Vector Machine (SVM) algorithms. The results also show that some participants' activities were classified with an accuracy of more than 98%, on average. MOTS-CLÉS. Reconnaissance de l'activité, DNN, environnement non contrôlé.
Precise wind energy potential assessment is vital for wind energy generation and planning and development of new wind power plants. This work proposes and evaluates a novel two-stage method for location-specific wind energy potential assessment. It combines accurate statistical modelling of annual wind direction distribution in a given location with supervised machine learning of efficient estimators that can approximate energy efficiency coefficients from the parameters of optimized statistical wind direction models. The statistical models are optimized using differential evolution and energy efficiency is approximated by evolutionary fuzzy rules.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.