Pose-Driven Dialogue Framework Integrating Agents and Locations for Elderly
Sinan Chen, Masahide Nakamura · 2024
In recent years, as the technology for human-machine interaction (HCI) has rapidly advanced, the challenge of low autonomy in human dialogue remains unimproved and presents a significant challenge. This is particularly presented by the elderly living at home, who may become increasingly isolated from society and less motivated to communicate. Building upon previous studies, we designed a human posture-driven conversational framework. The key idea was to integrate image-based location sensing with virtual agent (VA) technology. The proposed method makes use of Event-Condition-Action (ECA) rules and Pub/Sub routing techniques. Progress in this research facilitates personalized conversational support by inferring human states (i.e., context) and intentions.