The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
Online ISSN : 2424-3124
2007
Session ID : 2A1-D05
Conference information
2A1-D05 Acquisition of Action for Wheeled Type Robots with Consideration to Multiple Moving Obstacles Avoidance
Tomohiro YmaguchiYoshio Watanabe
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Details
Abstract
In this paper, action of wheeled type robot is determined from the obstacle configuration and the robot's self-state using a neural network (NN), and multiple moving obstacles avoided. In order to avoid obstacles with the minimum movement time, the design parameters of the NN are optimized by genetic algorithm (GA), using the data collected from several environments, in which each environment has a different obstacle configuration.
Content from these authors
© 2007 The Japan Society of Mechanical Engineers
Previous article Next article
feedback
Top