By Topic

Analysis of motion searching based on reliable predictability using recurrent neural network

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

5 Author(s)
Shun Nishide ; Department of Intelligence Science and Technology, Graduate School of Informatics, Kyoto University, Japan ; Tetsuya Ogata ; Jun Tani ; Kazunori Komatani
more authors

Reliable predictability is one of the main factors that determine human behaviors. The authors developed a model that searches and generates robot motions based on reliable predictability. Training of the model consists of three phases. In the first phase, the model trains a sequential learner, namely recurrent neural network with parametric bias, to self-organize robot and object dynamics. In the second phase, steepest descent method is utilized to search for robot motion that induces the most predictable object motion. In the third phase, a hierarchical neural network is trained to link object image with the searched motion. Experiments were conducted with cylindrical objects. Analysis of the results have shown that the robot has acquired the most reliable robot motion, shifting it according to the posture of the object. Twenty motion generation experiments have resulted in generation of robot motion that induces consistent rolling motion of the objects.

Published in:

2009 IEEE/ASME International Conference on Advanced Intelligent Mechatronics

Date of Conference:

14-17 July 2009