Design of companion aging and early warning service system based on speech emotion recognition

This study proposes a speech emotion recognition-based system to meet the growing demand for intelligent elderly care in China. The system creates an architecture for emotion recognition, companionship, and risk alerts by integrating deep learning and intelligent dialog. The research results show that it can demonstrate good performance in key indicators such as emotion recognition accuracy, interaction naturalness, and warning timeliness, and provide feasible technical solutions to improve the quality of life of the elderly.