Low-Cost Emotion-Aware Edge Robotics: An Open AI Framework for Real-Time Human–Robot Interaction

This paper presents a low-cost, edge-based framework for real-time emotion detection in robotic systems using affordable hardware such as the NVIDIA RTX 3050 GPU and Arduino microcontrollers. The proposed model integrates Speech Emotion Recognition (SER) based on transformer architectures with robotic actuation to enable context-aware, human-like interactions in constrained environments, such as space exploration and remote telepresence. The framework demonstrates that high accuracy (95%+) can be achieved without high-end computational resources, bridging the gap between emotional intelligence and edge robotics. Experimental evaluations show that tone, pitch, and spectral features correlate strongly with user affect, enabling robots to adapt their behavioral responses dynamically. The approach is designed to support future human–robot collaboration in space and resource-limited domains.