DocumentCode
3013347
Title
Application Research on RoboCup 3D Agent Walking Using Improved HEDGER Algorithm
Author
Jun, WANG ; Jing, XU ; Hongyuan, Wang
Author_Institution
Sch. of Inf. Sci. & Eng., Jiangsu Polytech. Univ., Changzhou, China
fYear
2010
fDate
25-27 June 2010
Firstpage
486
Lastpage
489
Abstract
RoboCup 3D simulation provides continuous states and action space. HEDGER is a learning algorithm based on Q-learning which can working on continuous states and action space. For the inefficient of its searching algorithm, we improve it with Kd-tree multi-dimensional space searching technology to solve the problem of the online learning of robot walking.
Keywords
learning (artificial intelligence); mobile robots; multi-robot systems; tree searching; Hedger Algorithm; Kd-tree multidimensional space searching technology; Q-learning; RoboCup 3D agent walking; RoboCup 3D simulation; online learning; Aerospace electronics; Learning; Legged locomotion; Solid modeling; Three dimensional displays; Training; Kd_tree; Q-leaning; RoboCup; robot walking;
fLanguage
English
Publisher
ieee
Conference_Titel
Electrical and Control Engineering (ICECE), 2010 International Conference on
Conference_Location
Wuhan
Print_ISBN
978-1-4244-6880-5
Type
conf
DOI
10.1109/iCECE.2010.125
Filename
5631575
Link To Document