Safe reinforcement learning with online filtering for fatigue-predictive human-robot task planning and allocation in production

TL;DR

This paper introduces PF-CD3Q, a safe reinforcement learning method that online estimates fatigue parameters to optimize human-robot task planning while ensuring worker safety in manufacturing.

Contribution

It develops a real-time fatigue estimation approach integrated into a constrained RL framework for safer, more efficient human-robot collaboration.

Findings

PF-CD3Q effectively tracks fatigue and updates parameters during production.

The method ensures task assignments stay within safe fatigue limits.

Experimental results demonstrate improved safety and efficiency in task planning.

Abstract

Human-robot collaborative manufacturing, a core aspect of Industry 5.0, emphasizes ergonomics to enhance worker well-being. This paper addresses the dynamic human-robot task planning and allocation (HRTPA) problem, which involves determining when to perform tasks and who should execute them to maximize efficiency while ensuring workers' physical fatigue remains within safe limits. The inclusion of fatigue constraints, combined with production dynamics, significantly increases the complexity of the HRTPA problem. Traditional fatigue-recovery models in HRTPA often rely on static, predefined hyperparameters. However, in practice, human fatigue sensitivity varies daily due to factors such as changed work conditions and insufficient sleep. To better capture this uncertainty, we treat fatigue-related parameters as inaccurate and estimate them online based on observed fatigue progression during production. To address these challenges, we propose PF-CD3Q, a safe reinforcement learning (safe RL) approach that integrates the particle filter with constrained dueling double deep Q-learning for real-time fatigue-predictive HRTPA. Specifically, we first develop PF-based estimators to track human fatigue and update fatigue model parameters in real-time. These estimators are then integrated into CD3Q by making task-level fatigue predictions during decision-making and excluding tasks that exceed fatigue limits, thereby constraining the action space and formulating the problem as a constrained Markov decision process (CMDP).