Co-designing ambient-assisted interventions using digital interlocutors for people with dementia

Abstract

Socially assistive robots (SAR) have been used to assist the care of persons with dementia (PwD). We propose to combine problematic behaviors detection and robot interventions in an Ambient-assisted Intervention System (AaIS). To achieve this long-term aim, our first step was to design and evaluate robot-based interventions. We describe the co-design process conducted with caregivers to define realistic application scenarios for a conversational robot, named Eva, to enact interventions. Most recommendations from the caregivers in the co-design session were implemented in Eva. We conducted two sessions of a music therapy intervention with the new version of Eva with six PwD at a geriatric residence. Preliminary results from the intervention showed that the residents enjoyed the sessions, they talked to Eva to request songs, answered questions made by the robot and sang with it.