Monitoring and detecting outliers for elder's life activities in a smart home: A case study

Monitoring the progress or deterioration of the elder's life activities is vital for a medical assistant or care helper to provide an early alert when an unusual activity may be happening in other days. In this study, we have considered a technique for monitoring and identifying abnormal behavior of health data collected from a smart home to help an elderly person to live safely in his/her home longer away from hospitals. We approach this by using statistical techniques. The proposed technique is applied in an intelligent environment where a network of wireless sensors is installed. To validate our approach, we use a real sensor data extracted from a smart home where an elderly lady living alone during one year. The experiment results show that the statistical test Modified Thompson tau successfully distinguishes the normal and abnormal behavioral patterns.