Multimodal Artificial Intelligence for Elderly Care Robots in Perception Interaction Decision Making and System Deployment

Abstract

With the rapid growth of the aging population in the world, elderly care systems are facing increasing pressure and challenges in healthcare services, daily assistance, and emotional support. Artificial intelligence (AI) has been widely integrated into elderly care robots to enhance autonomy, perception, interaction, and decision-making capabilities. This paper reviews recent advances in AI for elderly care robots, focusing on perception, natural language processing, machine learning, affective computing, and decision-making. In addition, it examines key studies, compares primary technical approaches, and identifies key challenges, including privacy risks, ethical issues, system reliability, and human-robot trust. Based on recent studies, this study explains how AI improves elderly care quality and identifies gaps in real-world deployment and long-term performance. The results reveal that AI improves personalization and autonomy, while major challenges remain in safety validation, ethical governance, and multimodal interaction performance, which need to be addressed to enable scalable and sustainable elderly care solutions.