Bipedal Walking Robot Control Using PMTG Architecture

被引：0

作者：

Danilov, Vladimir ^{[1
,3
]}

Klimov, Konstantin ^{[2
,3
]}

Kapytov, Dmitrii ^{[2
,3
]}

Diane, Sekou ^{[1
]}

机构：

[1] Russian Acad Sci, Inst Control Problems, Moscow, Russia

[2] Moscow MV Lomonosov State Univ, Robot Lab, Inst Mech Lomonosov, Michurinsky Prosp 1, Moscow 119192, Russia

[3] Voltbro LCC, Moscow, Russia

来源：

SYNERGETIC COOPERATION BETWEEN ROBOTS AND HUMANS, VOL 2, CLAWAR 2023 | 2024年 / 811卷

关键词：

Reinforcement learning; Locomotion; CPG; Walking robot; LOCOMOTION;

D O I：

10.1007/978-3-031-47272-5_8

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Reinforcement learning based methods can achieve excellent results for robot locomotion control. However, their serious disadvantage is the long agent training time and large number of parameters defining its behavior. In this paper, we propose a method that significantly reduces training time. It is based on the Policy Modulating Trajectory Generator (PMTG) architecture, which uses Central Pattern Generators (CPG) as a gait generator. We tested this approach on an OpenAI BipedalWalker-v3 environment. The paper presents the results of this algorithm, showing its effectiveness in solving a locomotion problem over challenging terrain.

引用

页码：89 / 100

页数：12

共 20 条

[1] Optuna: A Next-generation Hyperparameter Optimization Framework
Akiba, Takuya
Sano, Shotaro
Yanase, Toshihiko
Ohta, Takeru
Koyama, Masanori
[J]. KDD'19: PROCEEDINGS OF THE 25TH ACM SIGKDD INTERNATIONAL CONFERENCCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2019, : 2623 - 2631
[2] OPTIMIZATION AND GAITS IN THE LOCOMOTION OF VERTEBRATES
ALEXANDER, RM
[J]. PHYSIOLOGICAL REVIEWS, 1989, 69 (04) : 1199 - 1227
[3] Whole-Body MPC and Online Gait Sequence Generation for Wheeled-Legged Robots
Bjelonic, Marko
Grandia, Ruben
Harley, Oliver
Galliard, Cla
Zimmermann, Samuel
Hutter, Marco
[J]. 2021 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2021, : 8388 - 8395
[4] Brockman G, 2016, Arxiv, DOI [arXiv:1606.01540, DOI 10.48550/ARXIV.1606.01540]
[5] Danilov V, 2022, Robotics in Natural Settings: CLAWAR 2022, P276, DOI [10.1007/978- 3- 031-15226-9 27, DOI 10.1007/978-3-031-15226-927]
[6] Di Carlo J, 2018, IEEE INT C INT ROBOT, P7440, DOI 10.1109/IROS.2018.8594448
[7] github, OpenAI Gym Leaderboard
[8] Haarnoja T, 2018, arXiv
[9] Learning agile and dynamic motor skills for legged robots
Hwangbo, Jemin
Lee, Joonho
Dosovitskiy, Alexey
Bellicoso, Dario
Tsounis, Vassilios
Koltun, Vladlen
Hutter, Marco
[J]. SCIENCE ROBOTICS, 2019, 4 (26)
[10] Iscen A., 2018, PMLR, P916

← 1 2 →