E-R Homie: A Smart Companion System for Emotional Well-being at Home

Abstract

This paper presents E-R Homie, an emotion-aware smart home system designed to enhance the emotional well-being of users by adapting the home environment in real time. The system aims to overcome the limitations of traditional smart homes by integrating emotional intelligence through multimodal sensing techniques. Voice patterns, facial expressions and optional physiological signals are captured through IoT-enabled devices and analyzed using machine learning algorithms such as CNN, SVM and LSTM. The recognized emotional states are mapped to environmental responses, including adjustments in lighting, temperature and music to promote comfort and relaxation. Experimental observations show that the system achieves an average emotion recognition accuracy of approximately 85% and executes environmental changes within two seconds. The results indicate a significant improvement in user comfort, stress reduction and emotional support. The proposed system demonstrates how emotion-aware automation can contribute to a human-centric living environment and offers a scalable solution for integrating affective computing into everyday life.