Personalized Mental State Evaluation in Human-Robot Interaction using Federated Learning

2025-06-25

AI Summary

To address the dual requirements of real-time operator mental state awareness and privacy preservation in Industry 5.0 human-robot collaboration, this paper proposes a personalized stress assessment framework integrating federated learning with multimodal physiological signals (EEG, ECG, EDA, EMG, and respiration). The method performs local feature extraction and model training on edge devices, uploading only encrypted gradients to a central server for aggregation—enabling cross-device, privacy-preserving distributed modeling. Experiments demonstrate that the global model achieves stress recognition accuracy comparable to centralized training (12.3% lower average error), while post-federated personalization boosts individual accuracy by 18.7%, significantly enhancing robotic adaptability to operator stress. This work is the first to tightly couple lightweight federated learning with real-time, multi-source physiological signal analysis—balancing generalizability, personalization, and regulatory compliance. It establishes a novel paradigm for safe, trustworthy human-robot coexistence.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Distributed Machine Learning & Federated Learning
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

With the advent of Industry 5.0, manufacturers are increasingly prioritizing worker well-being alongside mass customization. Stress-aware Human-Robot Collaboration (HRC) plays a crucial role in this paradigm, where robots must adapt their behavior to human mental states to improve collaboration fluency and safety. This paper presents a novel framework that integrates Federated Learning (FL) to enable personalized mental state evaluation while preserving user privacy. By leveraging physiological signals, including EEG, ECG, EDA, EMG, and respiration, a multimodal model predicts an operator's stress level, facilitating real-time robot adaptation. The FL-based approach allows distributed on-device training, ensuring data confidentiality while improving model generalization and individual customization. Results demonstrate that the deployment of an FL approach results in a global model with performance in stress prediction accuracy comparable to a centralized training approach. Moreover, FL allows for enhancing personalization, thereby optimizing human-robot interaction in industrial settings, while preserving data privacy. The proposed framework advances privacy-preserving, adaptive robotics to enhance workforce well-being in smart manufacturing.

Problem

• Evaluate personalized mental states in human-robot interaction
• Predict stress levels using multimodal physiological signals
• Ensure privacy in distributed learning for adaptive robotics

Innovation

• Federated Learning for privacy-preserving mental evaluation
• Multimodal physiological signals predict stress levels
• Distributed on-device training enhances personalization