Abstract
This research aims to develop a system that estimates the health status of the elderly using the environmental sensor data from a monitoring system and the dialogue between the robot and the elderly. This paper describes the development of a robot voice-calling system that actively senses the health condition of the elderly based on the response of the elderly to the voice-calling of the robot. Since it is sufficient to elicit some utterance from the elderly, it is decided that the robot system uses scenario-based dialogue. To examine the effectiveness of this system, we have conducted a new user study in which elderly and a robot make a dialogue via video calls. In this paper, a detailed investigation on the duration of dialogue, the response to scenario-based dialogue, and the naturalness of dialogue, have been reported. A scenario-based dialogue system has been evaluated in comparison to a chat dialogue system with voice recognition. The participants responded well to the scenario-based dialogue and wanted to continue the dialogue with the robot. Even though the participants are actively engaged in dialogue, it is found that there is no significant difference in the naturalness of the two dialogue systems according to the evaluation made by the participants.
