Abstract
Activities of daily living are good indicators of elderly health status, and activity recognition in smart environments is a well-known problem that has been previously addressed by several studies. In this paper, we describe the use of two powerful machine learning schemes, ANN (Artificial Neural Network) and SVM (Support Vector Machines), within the framework of HMM (Hidden Markov Model) in order to tackle the task of activity recognition in a home setting. The output scores of the discriminative models, after processing, are used as observation probabilities of the hybrid approach. We evaluate our approach by comparing these hybrid models with other classical activity recognition methods using five real datasets. We show how the hybrid models achieve significantly better recognition performance, with significance level p < 0.05, proving that the hybrid approach is better suited for the addressed domain.
Highlights
Population aging is currently having a significant impact on health care systems [1]
The k-Nearest Neighbor (k-NN) has to be parameterized with the number of neighbors (k) used for classification; in our case, our experiments showed that best results are obtained using k = 5
In this paper we have proposed two new approaches to recognize activities of daily living (ADL) from home environments using a network of binary sensors
Summary
Population aging is currently having a significant impact on health care systems [1]. Reference [17], binary sensors measuring the opening or closing of doors and cupboards, the use of electric appliances, as well as motion sensors were used to recognize ADLs of elderly people living on their own This kind of sensors is considered one of the most promising technologies to solve key problems in the ubiquitous computing domain, due to their suitability to supply constant supervision and their inherent non-intrusive characteristics. This paper proposes two new approaches to recognize ADLs from binary sensor streams based on hybrid HMM schemes (combined with either ANN or SVM). This kind of approach has been previously applied for recognizing human activities using wearable devices but not in a wireless sensor network setting, to the best of our knowledge.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
