A Sim-to-Real Learning-Based Framework for Contact-Rich Assembly by Utilizing CycleGAN and Force Control

被引:10
|
作者
Shi, Yunlei [1 ,2 ]
Yuan, Chengjie [3 ]
Tsitos, Athanasios [2 ]
Cong, Lin [1 ]
Hadjar, Hamid [2 ]
Chen, Zhaopeng [4 ]
Zhang, Jianwei [1 ]
机构
[1] Univ Hamburg, Dept Informat, Tech Aspects Multimodal Syst, D-20146 Hamburg, Germany
[2] Agile Robots AG, Dept Intelligent Solut & Applicat, D-81477 Munich, Germany
[3] Tech Univ Munich, Dept Elect & Comp Engn, D-80333 Munich, Germany
[4] Agile Robots AG, D-81477 Munich, Germany
基金
美国国家科学基金会;
关键词
Robot sensing systems; Domain adaptation; force control; Peg-in-Hole (PiH); reality gap; sim-to-real;
D O I
10.1109/TCDS.2023.3237734
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep reinforcement learning (RL) has succeeded in robotic manipulation applications. However, training robots in the real world is challenging due to sample efficiency and safety concerns. Sim-to-real transfer has been proposed to address the aforementioned concerns but introduces the reality gap. In this work, we introduce a sim-to-real learning framework for vision-based assembly tasks and perform training in a simulation environment by employing raw image inputs from a single camera to address the aforementioned issues. We build a robotic Peg-in-Hole (PiH) training environment that requires low-level simulation knowledge. We also present a domain adaptation method based on a cycle-consistent generative adversarial network (CycleGAN) and a force control transfer approach to bridge the reality gap. The proposed framework, trained in a simulation environment with different environmental scenes, can be successfully transferred to a real PiH setup with a UR5e robot. We then reproduce these results with a Diana7 robot and different peg shapes to verify the generalization ability of the framework.
引用
收藏
页码:2144 / 2155
页数:12
相关论文
共 20 条
  • [11] A Conformal Mapping-based Framework for Robot-to-Robot and Sim-to-Real Transfer Learning
    Gao, Shijie
    Bezzo, Nicola
    2021 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2021, : 1289 - 1295
  • [12] Zero-shot sim-to-real transfer of reinforcement learning framework for robotics manipulation with demonstration and force feedback
    Chen, Yuanpei
    Zeng, Chao
    Wang, Zhiping
    Lu, Peng
    Yang, Chenguang
    ROBOTICA, 2023, 41 (03) : 1015 - 1024
  • [13] Sim-to-Real Learning-Based Nonlinear MPC for UAV Navigation and Collision Avoidance in Unknown Cluttered Environments
    Doukhi, Oualid
    Lee, Deok-Jin
    IEEE ACCESS, 2025, 13 : 46249 - 46262
  • [14] Variable-Impedance and Force Control for Robust Learning of Contact-rich Manipulation Tasks from User Demonstration
    Enayati, Nima
    Mariani, Stefano
    Wahrburg, Arne
    Zanchettin, Andrea M.
    IFAC PAPERSONLINE, 2020, 53 (02): : 9834 - 9840
  • [15] Maximizing the Use of Environmental Constraints: A Pushing-Based Hybrid Position/Force Assembly Skill for Contact-Rich Tasks
    Shi, Yunlei
    Chen, Zhaopeng
    Cong, Lin
    Wu, Yansong
    Craiu-Mueller, Martin
    Yuan, Chengjie
    Chang, Chunyang
    Zhang, Lei
    Zhang, Jianwei
    2021 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND BIOMIMETICS (IEEE-ROBIO 2021), 2021, : 496 - 501
  • [16] Unified HumanRobotEnvironment Interaction Control in Contact-Rich Collaborative Manipulation Tasks via Model-Based Reinforcement Learning
    Liu, Xing
    Liu, Yu
    Liu, Zhengxiong
    Huang, Panfeng
    IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, 2023, 70 (11) : 11474 - 11482
  • [17] Deep learning-based control framework for dynamic contact processes in humanoid grasping
    Cheng, Shaowen
    Jin, Yongbin
    Wang, Hongtao
    FRONTIERS IN NEUROROBOTICS, 2024, 18
  • [18] Learning-based Optimization Algorithms Combining Force Control Strategies for Peg-in-Hole Assembly
    Zou, Peng
    Zhu, Qiuguo
    Wu, Jun
    Xiong, Rong
    2020 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2020, : 7403 - 7410
  • [19] A Learning-Based Model Predictive Control Framework for Real-Time SIR Epidemic Mitigation
    She, Baike
    Sundaram, Shreyas
    Pare, Philip E.
    2022 AMERICAN CONTROL CONFERENCE, ACC, 2022, : 2565 - 2570
  • [20] Reinforcement learning-based real-time intelligent energy management for hybrid electric vehicles in a model predictive control framework*
    Yang, Ningkang
    Ruan, Shumin
    Han, Lijin
    Liu, Hui
    Guo, Lingxiong
    Xiang, Changle
    ENERGY, 2023, 270