Learning Vision Algorithms for Real Mobile Robots with Genetic Programming
BibTeX
@MISC{Barate_learningvision,
author = {Renaud Barate and Antoine Manzanera},
title = {Learning Vision Algorithms for Real Mobile Robots with Genetic Programming},
year = {}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
We present a genetic programming system to evolve vision based obstacle avoidance algorithms. In order to develop autonomous behavior in a mobile robot, our purpose is to designautomaticallyanobstacleavoidancecontroller adaptedtothecurrentcontext. Wefirstrecordshortsequenceswherewemanuallyguidethe robottomoveaway fromthewalls. Thissetofrecordedvideoimagesandcommandsisourlearningbase. Geneticprogrammingisused asasupervisedlearningsystemtogeneratealgorithmsthat exhibitthiscorridorcenteringbehavior. Weshowthatthe generatedalgorithmsareefficientinthecorridorthatwas usedtobuildthelearningbase,andthattheygeneralizeto someextentwhentherobotisplacedinavisuallydifferent corridor. More, theevolutionprocesshasproducedalgorithms thatgopastalimitationofoursystem,thatisthe lackofadequateedgeextractionprimitives. This is a good indication of the ability of this method to find efficient solutions for different kinds of environments.
