Reactive Stepping for Humanoid Robots using Reinforcement Learning: Application to Standing Push Recovery on the Exoskeleton Atalante

被引:1
|
作者
Duburcq, Alexis [1 ,2 ,3 ]
Schramm, Fabian [1 ]
Boeris, Guilhem [1 ]
Bredeche, Nicolas [2 ]
Chevaleyre, Yann [3 ]
机构
[1] Wandercraft, Paris, France
[2] Sorbonne Univ, CNRS, Inst Syst Intelligents & Robot, ISIR, F-75005 Paris, France
[3] Univ Paris 09, CNRS, Lab Anal & Modelisat Syst Aide Decis, PSL, Paris, France
关键词
WALKING; LOCOMOTION;
D O I
10.1109/IROS47612.2022.9982234
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
State-of-the-art reinforcement learning is now able to learn versatile locomotion, balancing and push-recovery capabilities for bipedal robots in simulation. Yet, the reality gap has mostly been overlooked and the simulated results hardly transfer to real hardware. Either it is unsuccessful in practice because the physics is over-simplified and hardware limitations are ignored, or regularity is not guaranteed, and unexpected hazardous motions can occur. This paper presents a reinforcement learning framework capable of learning robust standing push recovery for bipedal robots that smoothly transfer to reality, providing only instantaneous proprioceptive observations. By combining original termination conditions and policy smoothness conditioning, we achieve stable learning, sim-to-real transfer and safety using a policy without memory nor explicit history. Reward engineering is then used to give insights into how to keep balance. We demonstrate its performance in reality on the lower-limb medical exoskeleton Atalante.
引用
收藏
页码:9302 / 9309
页数:8
相关论文
共 14 条
  • [1] Development of Push-Recovery control system for humanoid robots using deep reinforcement learning
    Aslan, Emrah
    Arserim, Muhammet Ali
    Ucar, Aysegul
    [J]. AIN SHAMS ENGINEERING JOURNAL, 2023, 14 (10)
  • [2] Push Recovery Control for Humanoid Robot using Reinforcement Learning
    Seo, Donghyeon
    Kim, Harin
    Kim, Donghan
    [J]. 2019 THIRD IEEE INTERNATIONAL CONFERENCE ON ROBOTIC COMPUTING (IRC 2019), 2019, : 488 - 492
  • [3] Learning Push Recovery Behaviors for Humanoid Walking Using Deep Reinforcement Learning
    Dicksiano C. Melo
    Marcos R. O. A. Maximo
    Adilson Marques da Cunha
    [J]. Journal of Intelligent & Robotic Systems, 2022, 106
  • [4] Learning Push Recovery Behaviors for Humanoid Walking Using Deep Reinforcement Learning
    Melo, Dicksiano C.
    Maximo, Marcos R. O. A.
    da Cunha, Adilson Marques
    [J]. JOURNAL OF INTELLIGENT & ROBOTIC SYSTEMS, 2022, 106 (01)
  • [5] Learning Full Body Push Recovery Control for Small Humanoid Robots
    Yi, Seung-Joon
    Zhang, Byoung-Tak
    Hong, Dennis
    Lee, Daniel D.
    [J]. 2011 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA), 2011, : 2047 - 2052
  • [6] Trajectory-Free Reactive Stepping of Humanoid Robots Using Momentum Control
    Choi, Hyunchul
    Lee, Sukwon
    Jin, Taeil
    Lee, Sung-Hee
    [J]. 2015 IEEE-RAS 15TH INTERNATIONAL CONFERENCE ON HUMANOID ROBOTS (HUMANOIDS), 2015, : 1173 - 1178
  • [7] Learning to Move an Object by the Humanoid Robots by Using Deep Reinforcement Learning
    Aslan, Simge Nur
    Tasci, Burak
    Ucar, Aysegul
    Guzelis, Cuneyt
    [J]. INTELLIGENT ENVIRONMENTS 2021, 2021, 29 : 143 - 155
  • [8] Visual Navigation for Biped Humanoid Robots Using Deep Reinforcement Learning
    Lobos-Tsunekawa, Kenzo
    Leiva, Francisco
    Ruiz-del-Solar, Javier
    [J]. IEEE ROBOTICS AND AUTOMATION LETTERS, 2018, 3 (04): : 3247 - 3254
  • [9] Momentum-Based Reactive Stepping Controller on Level and Non-level Ground for Humanoid Robot Push Recovery
    Yun, Seung-Kook
    Goswami, Ambarish
    [J]. 2011 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS, 2011, : 3943 - 3950
  • [10] Online Learning of Low Dimensional Strategies for High-Level Push Recovery in Bipedal Humanoid Robots
    Yi, Seung-Joon
    Zhang, Byoung-Tak
    Hong, Dennis
    Lee, Daniel D.
    [J]. 2013 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA), 2013, : 1649 - 1655