Contact us
 |
[email protected] |
 |
3275638434 |
 |
 |
| Paper Publishing WeChat |
Useful Links
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
Article
Navigation Model for a Lego Robot Using a Backpropagation Neural Network
Author(s)
Erick Cervantes Chirinos, Jose Antonio Castan Rocha, Salvador Ibarra Martinez, Julio Laria Menchaca, Javier Guzman Obando, Mayra Trevino Berrones and Emilio Castan Rocha
Full-Text PDF
XML 1258 Views
DOI:10.17265/1548-7709/2015.04.007
Affiliation(s)
ABSTRACT
The paper introduces a novel approach for the autonomous navigation to an unmanned vehicle based on neural networks. A robot from the Lego family is used to represent the vehicle used with the proposal. The main objective of this robot is to navigate in an autonomous way in different scenarios avoiding crash with objects, robots and people in the same environment. To do this, the robot is endowed with ultrasonic, color and push sensors. These devices provide information about the configuration of the scenario in real time. With this data and using a backpropagation neural network, the robot can react and think about which maneuver must do. In order to perform the training phase, a specific simulator was generated in MatLab language. Some experiments were executed to corroborate the performance of the two neural networks and the extrapolation of their models to a free development language dedicated to Lego robots (Bricx Command Center). Finally, from the obtained results, some conclusions are discussed.
KEYWORDS
Index Terms!aLego robotics, neural netork, autonomous navigation approach
Cite this paper
References