An earlier version of this paper was presented at IWANN-2005. ; We propose a technique to speed up the learning of the inverse kinematics of a robot manipulator by decomposing it into two or more virtual robot arms. Unlike previous decomposition approaches, this one does not place any requirement on the robot architecture and, thus, it is completely general. Parametrized Self-Organizing Maps (PSOM) are particularly adequate for this type of learning, and permit comparing results obtained directly and through the decomposition. Experimentation shows that time reductions of up to two orders of magnitude are easily attained. ; This work was supported by projects: 'Perception, action & cognition through learning of object-action complexes.' (4915), 'Reconstrucció i anàlisi del moviment de grans estructures robòtiques i bioquímiques.' (J-05265), 'Analysis and motion planning of complex robotic systems' (4802), 'Grup de recerca consolidat - ROBÒTICA' (8007). This work was supported in part by the Generalitat de Catalunya under the consolidated Robotics group, by the Spanish Ministry of Science and Education under Project DPI2007-60858, by the “Comunitat de Treball dels Pirineus” under Project 2006ITT-10004, and by the European Commission under Project PACO-PLUS, CogSys Integrated Project FP6-IST-4-27657. ; Peer Reviewed
Learning inverse kinematics: Reduced sampling through decomposition into virtual robots
2008-01-01
doi:10.1109/TSMCB.2008.928232
Article (Journal)
Electronic Resource
English
DDC: | 629 |
General approach for inverse kinematics of nR robots
Online Contents | 2014
|A Deep Reinforcement Learning Approach for Inverse Kinematics of Concentric Tube Robots
BASE | 2019
|Novel Bio-inspired Inverse Kinematics for Fault-Tolerant Multilegged Robots
BASE | 2021
|