A Feasibility Study of a Data-Driven Human-Robot Conversational Interface for Reminiscence Therapy

International audienceSocial Assistive Robotics are widely used in healthcare to improve conventional treatments and increase patient engagement. Reminiscence Therapy (RT) is one application where social robots can be incorporated. RT is commonly used with people living with dementia, and it aims to evoke users' memories and stimulate cognitive functioning using nostalgic materials. This paper presents a feasibility study of a datadriven human-robot conversational interface for reminiscing sessions. Ten healthy participants were recruited to evaluate the usability of the interface, user engagement and interaction perception. The results showed that most of the participants followed the conversation, and half of them contributed highly (i.e., interaction/speaking time >51.54%) during the interaction with the robot. Participants perceived the system as being able to generate context-relevant dialogue