Elektrotechnik & Informationstechnik (2012) 129/4: 293–298. DOI 10.1007/s00502-012-0015-2 ORIGINALARBEITEN Relieved commissioning and human behavior detection in Ambient Assisted Living Systems D. Bruckner OVE, G.Q. Yin, A. Faltinger The goal of Ambient Assisted Living Systems is to provide automated technological aids for the elderly to allow for longer independent living in one’s own premises without the need for transition to stationary care. Such systems target to overcome problems introduced by particular risks of the target group like falling down, risk of illnesses, risk of dementia, etc. Current systems, however, still impose substantial effort in commissioning the system and they lack accuracy in detecting serious problems of the resident. In this article we present methods for relieved commissioning, i.e. automatic detection of the sensors’ types and topology, for added fault tolerance, and for modeling and evaluating human activity patterns with the goal of launching meaningful alarms. Keywords: Ambient Assisted Living; automatic comissioning; human behavior model; topology learning; fault tolerance Vereinfachte Kommissionierung von und Erkennung menschlichen Verhaltens mit Ambient Assistend Living-Systemen. Das Ziel von Ambient Assisted Living-Systemen ist es, automatisierte technische Hilfestellung für ältere Personen zur Verfügung zu stellen, um längeres unabhängiges Wohnen im eigenen, gewohnten Wohnumfeld zu ermöglichen, ohne dass eine Aufnahme in die stationäre Betreuung nötig wird. Diese Systeme zielen darauf ab, den Problemen durch typische altersbedingte Risken wie Demenz, Stürzen oder Krankheit entgegenzuwirken. Aktuell erhältliche Systeme haben aber die Nachteile, dass sie großen Kommissionierungs- aufwand benötigen und dass die Detektionsrate bei kritischen Fällen noch zu niedrig ist. In diesem Artikel zeigen wir daher Methoden für eine vereinfachte, teilautomatische Kommissionierung, in diesem Fall durch automatische Erkennung der Sensoren und deren Ty- pen, durch automatisches Lernen der Topologie des Systems und durch Einführen von Fehlertoleranz. Weiters wird eine Methode zur Modellierung und Evaluierung von Tagesablaufmustern vorgestellt, mit dem Ziel, sinnvollere Alarme zu generieren. Schlüsselwörter: Ambient Assisted Living; automatische Kommissionierung; Verhaltensmodell; Lernen der Topologie; Fehlertoleranz Received January 23, 2012, accepted March 6, 2012, published online August 23, 2012 © Springer-Verlag 2012 1. Introduction Research in the area of Ambient Assisted Living accounts for the fact that people tend to live longer and the implications this fact entails. It is also true that people tend to advance in years more healthy than in the past, however, increasing health care possibilities and less family care result in almost exponential growth of public ex- penditures (Giannakouris 2010). Aside of medical care, the elderly, as any other target group of the population, demand more ser- vices as comfort, security, wellness, social inclusion, technical gad- gets and many more. Thereby, the elderly have a strong focus on economic implementations thereof (Bruckner et al. 2012). Hence, the elderly, or more, aging-related implications, are increasingly not merely seen as a financial burden to society but also as a huge mar- ket for many kinds of targeted products. Thus, AAL solutions are specific user-centered solutions. The EU Support Action support of aC ommon A wareness and knowledge P latform for S tudying and enabling I ndependent L iving, CAPSIL, 1 is currently developing a Eu- ropean research road map for AAL solutions, mainly targeting inde- pendent living, which is only a subset of all demands of the elderly. The following is a list of topics that are within the focus of AAL technologies and research: 1 capsil.org. • Entertainment: Memory training, entertaining robots, robotic pets (Tao et al. 2008), brain training including computer assistance (Barnes et al. 2010), video telephony, senior mobile phones, social networks including accessibility • Autonomy: Barrier-free architecture, mobility including driving systems, mobility services, and home mobility (Urbano et al. 2008), domestic robots (Luo et al. 2005), exoskeletons, personal reminder systems, home security (Kistler et al. 2008) • Monitoring Systems: Behavior monitoring (Bruckner et al. 2007; Bruckner 2007), physiological monitoring (Ghasemzadeh and Ja- fari 2011), cognitive monitoring, detection of unintended or un- wanted situations (Sallans et al. 2005, 2006). In the following, this article concentrates on aspects of monitor- ing systems, in particular methods for easier and more cost-efficient installation and initialization of monitoring systems, since this effort poses a major hurdle for widespread implementation of these sys- tems. Juni/Juli/August 2012 129. Jahrgang © Springer-Verlag heft 4.2012 Bruckner, Dietmar, Univ.-Ass. Dipl.-Ing. Dr., Institute of Computer Technology, Vienna University of Technology, Gußhausstraße 25-29, 1040 Vienna, Austria (E-mail: bruckner@ict.tuwien.ac.at); Yin, Guo Qing, Dr., Institute of Computer Technology, Vienna University of Technology, Gußhausstraße 25-29, 1040 Vienna, Austria; Faltinger, Armin, Dipl.-Ing., Institute of Computer Technology, Vienna University of Technology, Gußhausstraße 25-29, 1040 Vienna, Austria 293