Abstract
This paper explores the use of sensor-based multimodal data fusion and emotion detection technologies in e-VITA, a three-year EU-Japan collaborative project that developed an AI-powered virtual coaching system to support independent living for older adults. The system integrates these technologies to enable individualized profiling and personalized recommendations across multiple domains, including nutrition, physical exercise, sleep, cognition, spirituality, and social health. Following a review of related work, we detail the implementation and evaluation of data fusion and emotion detection in e-VITA. The paper concludes with a summary of the key research findings and directions for future work.