AI and Digital Twin Federation-Based Flexible Safety Control for HumanRobot Collaborative Work Cell

초록

This paper proposes a safety response technology that leverages multi-domain vision sensors to monitor the positions of workers and robots in real-time, enabling dynamic calculation of Flexible Protective Separation Distances (FPSD) for adaptive robot velocity control and safety stops. Compliant with ISO/TS 15066 technical specifications for speed and separation monitoring, the method integrates AI-driven 3D pose estimation and Digital Twin visualization to enhance safety and productivity in human-robot collaborative (HRC) work cells. Experimental results demonstrate that FPSD significantly outperforms static Protective Separation Distance (PSD) methods, achieving collaboration times of 2.50 s (98.04%) and 5.40 s (99.08%) in gantry and collaborative robot experiments, respectively, compared to 0.25 s (9.80%) and 2.50 s (45.87%) for PSD method. The system reduces unnecessary robot pauses with average response times of 0.021 s for deceleration and 0.032 s for stops, ensuring robust collision prevention. Validated using multi-domain sensor data in dynamic HRC tasks, this approach confirms its potential for scalable deployment in smart manufacturing, offering a flexible and efficient safety framework.

키워드