Multimodal QOL Estimation During Human–Robot Interaction

Accurately assessing the Quality of Life (QOL) of the older adults is crucial for personalizing interactions with assistive robots and advancing welfare initiatives through Information and Communication Technology (ICT). This study proposes a multimodal QOL estimation system that integrates visual, auditory, and textual data, implemented within a communication robot for interaction experiments with older participants. Experiments with both young and older participants demonstrated that the multimodal learning approach significantly outperforms unimodal methods in major QOL metrics. Results indicate that the proposed method enables real-time dialogue analysis and QOL estimation, contributing to personalized care for the older adults. This research not only enhances the QOL of the older adults but also expands the practical application of AI technologies in digital healthcare, including remote care and home-based medical services.