International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | A Survey on Emotion-Aware Deep Reinforcement Learning for Human-Robot Interaction |
|---|---|
| ABSTRACT | Robots are moving beyond industrial roles into homes, hospitals, factories, and public spaces, where they must work closely with people. For safe and comfortable interaction, robots need to sense human emotions and adjust their behaviour in real time. Emotion-aware deep reinforcement learning (EA-DRL) brings these abilities together by combining emotion detection with adaptive decision-making. Emotion sensing uses signals such as speech, facial expressions, body movement, and physiological data to identify human states like stress, calm, or engagement. Deep reinforcement learning allows robots to learn through trial and feedback, improving actions over time rather than following fixed scripts. This paper surveys advances in EA-DRL for human-robot interaction, examining its methods, applications, and challenges. It reviews how emotion-aware robots support fields such as healthcare, education, social companionship, industrial collaboration, and navigation in public spaces. It explores key challenges including fast emotion detection, large training needs, safety during learning, processing limits on small robots, generalization across environments, and privacy risks. A generalized process model is presented, showing how sensing, feature extraction, emotion recognition, and reinforcement learning combine to create adaptive behavior. Traditional machine learning approaches are compared with DRL-based systems, highlighting the trade-off between simplicity and adaptability. |
| AUTHOR | SUSHMITA SHEEBA DSA, ESWIN PRIA ANGEL |
| VOLUME | 176 |
| DOI | DOI: 10.15680/IJIRCCE.2025.1311075 |
| pdf/75_A Survey on Emotion-Aware Deep Reinforcement Learning for Human-Robot Interaction.pdf | |
| KEYWORDS |
