Constructive Radial Basis Function Networks for Mobile Robot Positioning

https://doi.org/10.1016/S1474-6670(17)44193-0Get rights and content

Abstract

This paper presents a comparison of several methods for constructing neural classifiers in the framework of mobile robot positioning. The task of positioning a mobile robot using a ring of sonars or laser sensors is used as a testbed for different radial basis function neural network constructive algorithms that produce topological maps of the environment. The experiments performed using two different experimental systems show the advantages and drawbacks of the compared algorithms.

Keywords

Neural networks
Mobile robots
Navigation
Self organizing systems
Learning algorithms

Cited by (0)

View Abstract