CogWatch: Intelligent agent-based system to assist stroke survivors during tea-making

Stroke survivors often have difficulties performing Activities of Daily Living (ADL). When they try to complete a task, they tend to rely on caregivers who give them cues when necessary. However, this reliance on caregivers' support may affect their ability to live an independent life. Hence, to tackle this issue, while giving them the guidance they need during ADL, the development of a specific Intelligent Assistive Planning System (AIPS) may be beneficial. In this paper, the design of CogWatch, which is a system implemented to assist stroke survivors during tea-making, is described. The Markov Decision Process framework that can be used in such a context, and which can potentially model any sequential tasks, is also explained. Finally, preliminary results involving twelve patients interacting with the system are analysed, and future plans are discussed.