S ¨ NAP: SyNdetic Assistance Processes Jesse Hoey School of Computer Science, University of Waterloo, Waterloo, Ontario, N2L 3G1, CANADA Andrew Monk Department of Psychology, University of York, York, YO10 5DD, UK Thomas Pl ¨ otz, Dan Jackson, Cuong Pham, Patrick Olivier School of Computing Science, University of Newcastle, Newcastle upon Tyne, NE1 7RU, UK Abstract Activity recognition in intelligent environments could play a key role for support- ing people in their activities of daily life. Partially observable Markov decision process (POMDP) models have been used successfully, for example, to assist peo- ple with dementia when carrying out small multi-step tasks such as hand washing. POMDP models are a powerful, yet flexible framework for modeling assistance that can deal with uncertainty and utility in a theoretically well-justified manner. Unfortunately, POMDPs usually require a very labor intensive, manual setup pro- cedure. This paper describes a knowledge driven method for automatically gener- ating POMDP activity recognition and context sensitive prompting systems for a complex tasks. We call the resulting POMDP a S ¨ NAP (SyNdetic Assistance Pro- cess). The method starts with a psychologically justified (syndetic) description of the task and the particular environment in which it is to be carried out that can be generated from empirical data. This is then combined with a specification of the available sensors and effectors to build a working prompting system that tracks a person’s activities and learns their abilities by using sensor data as evidence in the context of the S ¨ NAP POMDP. The method is illustrated by building a system that prompts through the task of making a cup of tea in a real-world kitchen. 1 Introduction Dementia is an important problem with serious effects on the society because of the changing de- mographic towards an aging population The dependency ratio (the ratio of those typically not in the labor force and potentially needing care to those typically in the labor force and thus able to provide care) is increasing. For example, the old age dependency ratio (the projected number of persons aged 65 and over expressed as a percentage of the projected number of persons aged between 15 and 64) in the European Union is projected to increase from 25.9 in 2010 to 50.4 in 2050 [1]. At the same time the number of people with dementia is increasing. The number of persons with Alzheimer’s disease worldwide, for example, is expected to double and will top 100 million by the year 2050 [2]. This means that the burden of care will have to shift from the professional arena (e.g. hospitals and clinics) into the home and community. Many people with dementia wish to remain living in their own homes as long as possible. However, they generally require some assistance in order to do so. Difficulties performing activities of daily living at home, such as preparing food, washing themselves, or cleaning, may trigger the need for personal assistance or relocation to residential care settings [9]. Moreover, it is associated with di- minished quality of life, poor self-esteem, anxiety, and social isolation for the person with dementia and their caregiver [4].