Sensor-based Bayesian detection of anomalous living patterns in a home setting

In this paper, we present an automated behavior analysis system developed to assist the elderly and individuals with disabilities who live alone, by learning and predicting standard behaviors to improve the efficiency of their healthcare. Established behavioral patterns have been recorded using wireless sensor networks composed by several event-based sensors that captured raw measures of the actions of each user. Using these data, behavioral patterns of the residents were extracted using Bayesian statistics. The behavior was statistically estimated based on three probabilistic features we introduce, namely sensor activation likelihood, sensor sequence likelihood, and sensor event duration likelihood. Real data obtained from different home environments were used to verify the proposed method in the individual analysis. The results suggest that the monitoring system can be used to detect anomalous behavior signs which could reflect changes in health status of the user, thus offering an opportunity to intervene if required.