Human Activity Recognition in Smart Home using Deep Learning Models

Abstract

Applying deep learning to IoT data classification would yield deeper and more useful insights. IoT field is very wide and has several applications. In this paper we focus on smart home, especially human activity recognition within a house equipped with ambient sensors. Firstly, we describe ARAS Human Activity Dataset that are used in models training and testing. Secondly, we apply three deep learning models to it in order to classify the activities carried out by residents within the house. The deep learning models that we use in our experiments are : a Multilayer Perceptron (MLP), a Recurrent Neural Network(RNN) and Long short-term memory(LSTM). The results show best performances with MLP followed by RNN. In addition, it should be noted that there is a strong correlation between the frequency of activities and their recognition rate.