Learning and recognizing human activities of daily living(ADL), is very useful and essential to build a pervasive home monitoring system. These monitoring technologies are indispensable for developing the next generation of smart houses. In this paper we describe a fuzzy logic system for recognizing activities in home environment using a set of sensors: physiological sensors (cardiac frequency, activity or agitation, posture and fall detection sensor), microphones, infrared sensors, debit sensors and state-change sensors. Motivated by the fact that fuzzy controllers have been successfully embedded within billions of dollars in commercial products, plus the characteristic of data providing from each sensor, the fusion of the different sensors has been performed by using fuzzy logic. This fuzzy logic approach allowed us to recognize several activities of daily living (ADLs) for ubiquitous healthcare.

Today elderly people are the fastest growing segment of the population in developed countries, and they desire to live as independently as possible. But independent lifestyles come with risks and challenges. Medical in-home telemonitoring (and, more generally, telemedicine) is a solution to deal with these challenges and to ensure that elderly people can live safely and independently in their own homes for as long as possible. In this context we propose an automatic in-home healthcare monitoring system for several uses and to meet the needs identified above. The proposed telemonitoring system is a multimodal platform with several sensors that can be installed at home and enables us to have a full and tightly controlled universe of data sets. It integrates elderly physiological and behavioral data, the acoustical environment of the elderly, environmental conditions and medical knowledge. Each modality is processed and analyzed by specific algorithms. A data fusion approach based on fuzzy logic with a set of rules directed by medical recommendations, is used to fuse the various subsystem outputs. This multimodal fusion increases the reliability of the whole system by detecting several distress situations. In fact this fusion approach takes into account temporary sensor malfunction and increases the system reliability and the robustness in the case of environmental disturbances or material limits (Battery, RF range, etc.). The Fuzzy logic fusion methods brings high flexibility to the telemonitoring platform especially in combining modalities or adding other sensors. The proposed telemonitoring system will ensure pervasive in-home health monitoring for elderly people.

Stress is an increasingly prevalent mental health condition across the world. In Europe, for example, stress is considered one of the most common health problems, and over USD 300 billion are spent on stress treatments annually. Therefore, monitoring, identification and prevention of stress are of the utmost importance. While most stress monitoring is carried out through self-reporting, there are now several studies on stress detection from physiological signals using Artificial Intelligence algorithms. However, the generalizability of these models is only rarely discussed. The main goal of this work is to provide a monitoring proof-of-concept tool exploring the generalization capabilities of Heart Rate Variability-based machine learning models. To this end, two Machine Learning models are used, Logistic Regression and Random Forest to analyze and classify stress in two datasets differing in terms of protocol, stressors and recording devices. First, the models are evaluated using leave-one-subject-out cross-validation with train and test samples from the same dataset. Next, a cross-dataset validation of the models is performed, that is, leave-one-subject-out models trained on a Multi-modal Dataset for Real-time, Continuous Stress Detection from Physiological Signals dataset and validated using the University of Waterloo stress dataset. While both logistic regression and random forest models achieve good classification results in the independent dataset analysis, the random forest model demonstrates better generalization capabilities with a stable F1 score of 61%. This indicates that the random forest can be used to generalize HRV-based stress detection models, which can lead to better analyses in the mental health and medical research field through training and integrating different models.

The Sweet-Home project aims at providing audio-based interaction technology that lets the user have full control over their home environment, at detecting distress situations and at easing the social inclusion of the elderly and frail population. This paper presents an overview of the project focusing on the multimodal sound corpus acquisition and labelling and on the investigated techniques for speech and sound recognition. The user study and the recognition performances show the interest of this audio technology.

Exergames have been proposed as a potential tool to improve the current practice of musculoskeletal rehabilitation. Inertial or optical motion capture sensors are commonly used to track the subject’s movements. However, the use of these motion capture tools suffers from the lack of accuracy in estimating joint angles, which could lead to wrong data interpretation. In this study, we proposed a real time quaternion-based fusion scheme, based on the extended Kalman filter, between inertial and visual motion capture sensors, to improve the estimation accuracy of joint angles. The fusion outcome was compared to angles measured using a goniometer. The fusion output shows a better estimation, when compared to inertial measurement units and Kinect outputs. We noted a smaller error (3.96°) compared to the one obtained using inertial sensors (5.04°). The proposed multi-sensor fusion system is therefore accurate enough to be applied, in future works, to our serious game for musculoskeletal rehabilitation.