Learning-Based Safety-Aware Task Scheduling for Efficient Human-Robot Collaboration
Marco Faroni, Alessio Spanò, Andrea Maria Zanchettin, Paolo Rocco
AI summary
Problem
Ensuring human safety in collaborative robotics typically forces robots to slow down or stop, causing significant efficiency losses, while existing mitigation methods require inaccessible or complex prior knowledge of manufacturer-specific safety configurations.
Approach
The method trains a neural network to predict safety-induced speed reductions from system state data, then uses greedy or Monte Carlo planning algorithms to select robot actions that minimize these slowdowns in real time.
Key results
- Learns safety speed scaling directly from process data without explicit safety models
- Introduces greedy and Monte Carlo action selection algorithms for real-time scheduling
- Achieves significant cycle time reductions in simulated and real-world pick-and-packaging tasks
- Generalizes across different safety configurations and adapts to layout changes via retraining
Why it matters
It offers a practical, deployment-ready solution for industrial collaborative robotics that boosts operational efficiency without requiring access to proprietary safety system internals.
Abstract
Ensuring human safety in collaborative robotics can compromise efficiency because traditional safety measures increase robot cycle time when human interaction is frequent. This paper proposes a safety-aware approach to mitigate efficiency losses without assuming prior knowledge of safety logic. Using a deep- learning model, the robot learns the relationship between system state and safety-induced speed reductions based on execution data. Our framework does not explicitly predict human motions but directly models the interaction effects on robot speed, simplifying implementation and enhancing generalizability to different safety logics. At runtime, the learned model optimizes task selection to minimize cycle time while adhering to safety requirements. Experiments on a pick-and-packaging scenario demonstrated significant reductions in cycle times.