Home // ICCGI 2017, The Twelfth International Multi-Conference on Computing in the Global Information Technology // View article

Terrain Classification Using a Radial Basis Function Network

Authors:
Tiny Du Toit
Hennie Kruger

Keywords: IMU; inertial measurement unit; Radial basis function network; sensor; terrain classification.

Abstract:
In this paper, inertial contact sensor based terrain classification is performed with a Radial basis function network (RBFN). Compared to the more popular Multilayer perceptrons, RBFNs are also intelligent techniques and universal approximators, but with a much simpler structure and shorter training time. It has been shown that RBFNs are efficient classifiers and consequently may be used for terrain classification. For the experiments, a mobile robot platform recorded vibration training data with an inertial measurement unit (IMU) while traversing five different terrains: asphalt, carpet, dirt, paving, and tiles. The composition of these terrains induces specific vibrations in the mobile platform which are measured by the IMU. The vibration signatures are comprised of the mobile robot’s linear acceleration, orientation, and the earth’s magnetic field. In contrast to most terrain classification techniques found in literature, no pre-processing of the data is performed. This reduces the computational overhead needed for real-time classification. A RBFN is then trained using a hybrid conjugate gradient descent method and k-fold cross- validation. Identification of the terrain is performed in real-time. The results are compared to those obtained by a Naïve Bayes method and a Support Vector Machine, which have also been successfully applied to terrain classification in literature. It was found that the RBFN outperformed these other techniques by a relatively large margin. Consequently, the RBFN with no pre-processing of the input data may be used as a contact sensor based terrain classification method.

Pages: 11 to 16

Copyright: Copyright (c) IARIA, 2017

Publication date: July 23, 2017

Published in: conference

ISSN: 2308-4529

ISBN: 978-1-61208-571-5

Location: Nice, France

Dates: from July 23, 2017 to July 27, 2017