Activity monitoring and automatic alarm generation in AAL-enabled homes

In this work, novel contributions towards the emerging field of Ambient Assisted Living (AAL) are introduced. AAL is a concept envisioned in the early 2000s by the European Commission, aiming at supporting specifically senior people by means of technology and thus helping them to lead independent and self-determined lives in their accustomed surroundings as long as possible. Modern home automation technology is believed to be the key to providing various services in the fields of health, safety, comfort, and communication. In the framework of this thesis, health monitoring aspects are of particular interest. Inactivity monitoring is a very promising approach thereto since it allows the detection of potential health threats or cases of emergency without being overly privacy intrusive. Deriving condensed and dependable inactivity profiles representing typical user behaviour is a pivotal prerequisite for automatic emergency monitoring. Several methodologies for computing such patterns are introduced. Based on those inactivity profiles, various alarming criteria (i. e., permissible inactivity thresholds) are utilised to trigger alarms automatically if the users' inactivity levels exceed individual, user-dependant limits. Since false alarms are inevitable in automatic alarming systems, a procedure of handling them is introduced as well. Finally, the real-world application of the devised AAL system is illustrated.