Combining Pervasive Computing with Activity Recognition and Learning

Today, networks, microprocessors, memory chips, smart sensors and actuators are faster, more powerful, cheaper and smaller than ever. Chips are all around, invading everyday objects. Wireless networks enable to easily connect them. Everyday objects can then propose innovative and unexpected interactions (Ullmer & Ishii, 2000). Clothes will transport one’s profile to reconfigure his environment according to his preferences (Abowd et al. 1997). Lamps will help people finding lost objects (Vergnes et al., 2005). Interactive portraits will reflect at distance the mood and health state of one’s beloved relatives (Mynatt et al., 2001). This new technologically enhanced environment will enable finding novel solutions to help people in their everyday life, such as elders that suffer from cognitive deficit and have many difficulties to carry out their activities of daily living (Pigot et al., 2003). Most of these people wish to stay at home, where they feel comfortable and safe, as long as possible. The governments aim to help them for social reasons as well as economical ones. However, keeping cognitively impaired people at home involves many risks that are necessary to control. In order to do that, the physical and human environment must be specifically designed to compensate the cognitive impairments and the loss of autonomy (Ramos et al., 2008). Combining pervasive computing with techniques from artificial intelligence (AI) greatly increases the acceptance of the pervasive assisted living and makes it more capable of providing a better quality of life in a non-intrusive way, where elderly people, with or without disabilities, could clearly benefit from this concept. From the computational perspective, there is a natural association between them. However, research addressing smart environments has in the past largely focused on network and hardware oriented solutions. AI-based techniques (planning and action theory, ontological and temporal reasoning, etc) that promote intelligent behaviour have not been examined to the same extent (Augusto & Nugent, 2006), although notable exceptions can been found in the domain of activity recognition for healthcare. Prior work has been done to use sensors to recognize the execution status of particular types of activities, such as hand washing (Mihailidis et al., 2007), meal preparation (Barger et al., 2002), and movements around town (Liao et al., 2004). Additionally, several projects have attempted to do more general activity 22