Acquisition of a biped walking pattern using a poincare map

被引：0

作者：

Morimoto, J ^{[1
]}

Nakanishi, J ^{[1
]}

Endo, G ^{[1
]}

Cheng, G ^{[1
]}

Atkeson, CG ^{[1
]}

Zeglin, G ^{[1
]}

机构：

[1] ATR Computat Neurosci Lab, JST, ICORP, Computat Brain Project, Kyoto 6100288, Japan

来源：

2004 4TH IEEE/RAS INTERNATIONAL CONFERENCE ON HUMANOID ROBOTS, VOLS 1 AND 2, PROCEEDINGS | 2004年

关键词：

biped walking; reinforcement learning; Poincare map;

D O I：

暂无

中图分类号：

TP24 [机器人技术];

学科分类号：

080202 ; 1405 ;

摘要：

We propose a model-based reinforcement learning algorithm for biped walking in which the robot learns to appropriately place the swing leg. This decision is based on a learned model of the Poincare map of the periodic walking pattern. The model maps from a state at a single support phase and foot placement to a state at the next single support phase. We applied this approach to both a simulated robot model and an actual biped robot. We show that successful walking patterns are acquired.

引用

页码：912 / 924

页数：13

共 23 条

[1] [Anonymous], 2003, P 2 INT S AD MOT AN
[2] Biped dynamic walking using reinforcement learning
Benbrahim, H
Franklin, JA
[J]. ROBOTICS AND AUTONOMOUS SYSTEMS, 1997, 22 (3-4) : 283 - 302
[3] Dynamic bipedal walking assisted by learning
Chew, CM
Pratt, GA
[J]. ROBOTICA, 2002, 20 : 477 - 491
[4] Reinforcement learning in continuous time and space
Doya, K
[J]. NEURAL COMPUTATION, 2000, 12 (01) : 219 - 245
[5] DYER P, 1970, COMPUTATIONAL THEORY
[6] THE COORDINATION OF ARM MOVEMENTS - AN EXPERIMENTALLY CONFIRMED MATHEMATICAL-MODEL
FLASH, T
HOGAN, N
[J]. JOURNAL OF NEUROSCIENCE, 1985, 5 (07) : 1688 - 1703
[7] GARCIA M, 1998, ASME, V120, P281, DOI DOI 10.1115/1.2798313
[8] HIRAI K, 1998, P 1998 IEEE INT C RO, P160
[9] Inoue, 1999, P 17 ANN C ROB SOC J, P1193
[10] Jacobson D. H., 1970, Differential Dynamic Programming. American

← 1 2 3 →