analiza aiqbench
This study demonstrates that reinforcement learning can effectively personalize collaborative robot control strategies to accommodate human musculoskeletal constraints, reducing task completion times by 38% while eliminating pain risk. Although a simulation-to-real gap was observed, a fine-tuning phase successfully bridged this divide, validating the approach across diverse anthropometries. The findings highlight the potential for RL-driven cobots to create safer, more inclusive industrial environments by integrating ergonomic optimization directly into robotic control policies.
A Human-Sensitive Controller: Adapting to Human Musculoskeletal Disorder-Related Constraints via Reinforcement Learning
arXiv:2504.10102v2 Announce Type: replace Abstract: Work-Related Musculoskeletal Disorders continue to be a major challenge in industrial environments, leading to reduced workforce participation, increased healthcare costs, and long-term disability. This study introduces a human-sensitive robotic system aimed at reintegrating individuals with a history of musculoskeletal disorders into standard job roles, while simultaneously optimizing ergonomic conditions for the broader workforce. This research leverages reinforcement learning (RL) to develop a human-aware control strategy for collaborative robots, focusing on optimizing ergonomic conditions and preventing pain during task execution. Two RL approaches, Q-Learning and Deep Q-Network (DQN), were implemented and tested to personalize control strategies based on individual user characteristics. Although experimental results revealed a simulation-to-real gap, a fine-tuning phase successfully adapted the policies to real-world conditions. DQN outperformed Q-Learning by completing tasks faster while maintaining zero pain risk and safe ergonomic levels, achieving on average 38% shorter task completion times across all tested anthropometries. The structured testing protocol confirmed the system's adaptability to diverse human anthropometries, underscoring the potential of RL-driven cobots to enable safer, more inclusive workplaces.
Authors:Vitor Martins (1), Sara M. Cerqueira (1), Mercedes Balcells (2 and 3), Elazer R Edelman (2 and 4), Cristina P. Santos (1 and 5) ((1) Center for MicroElectroMechanical Systems (CMEMS), University of Minho, Guimarães, Portugal, (2) IMES, Massachusetts Institute of Technology, Cambridge, MA, USA, (3) GEVAB, IQS School of Engineering, Barcelona, Spain, (4) Brigham and Women's Hospital, Harvard Medical School Boston, MA, USA, (5) LABBELS-Associate Laboratory, University of Minho, Guimarães, Portugal)
Abstract:Work-Related Musculoskeletal Disorders continue to be a major challenge in industrial environments, leading to reduced workforce participation, increased healthcare costs, and long-term disability. This study introduces a human-sensitive robotic system aimed at reintegrating individuals with a history of musculoskeletal disorders into standard job roles, while simultaneously optimizing ergonomic conditions for the broader workforce. This research leverages reinforcement learning (RL) to develop a human-aware control strategy for collaborative robots, focusing on optimizing ergonomic conditions and preventing pain during task execution. Two RL approaches, Q-Learning and Deep Q-Network (DQN), were implemented and tested to personalize control strategies based on individual user characteristics. Although experimental results revealed a simulation-to-real gap, a fine-tuning phase successfully adapted the policies to real-world conditions. DQN outperformed Q-Learning by completing tasks faster while maintaining zero pain risk and safe ergonomic levels, achieving on average 38% shorter task completion times across all tested anthropometries. The structured testing protocol confirmed the system's adaptability to diverse human anthropometries, underscoring the potential of RL-driven cobots to enable safer, more inclusive workplaces.
Robotics (cs.RO); Systems and Control (eess.SY)
arXiv:2504.10102 [cs.RO]
(or arXiv:2504.10102v2 [cs.RO] for this version)
Text integral sub licența deschisă a publicației · 254 cuvinte