Abstract
We describe an on-line machine learning ensemble technique, based on an adaptation of the mixture of experts (ME) model, for predicting terrain in autonomous outdoor robot navigation. Binary linear models, trained on-line on images seen by the robot at different points in time, are added to a model library as the robot navigates. To predict terrain in a given image, each model in the library is applied to feature data from that image, and the models’ predictions are combined according to a single-layer (flat) ME approach. Although these simple linear models have excellent discrimination in their local area in feature space, they do not generalize well to other types of terrain, and must be applied carefully. We use the distribution of training data as the source of the a priori pointwise mixture coefficients that form the soft gating network in the ME model. Single-class Gaussian models are learned during training, then later used to perform density estimation of incoming data points, resulting in pointwise estimates of model applicability. The combined output given by ME thus permits models to abstain from making predictions for certain parts of the image. We show that this method outperforms a less sophisticated, non-local baseline method in a statistically significant evaluation using natural datasets taken from the domain.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Jackel, L., Krotkov, E., Perschbacher, M., Pippine, J., Sullivan, C.: The DARPA LAGR program: Goals, challenges, methodology, and Phase I results. Journal of Field Robotics 23(11-12), 945–973 (2006)
Howard, A., Turmon, M., Matthies, L., Tang, B., Angelova, A., Mjolsness, E.: Towards learned traversability for robot navigation: From underfoot to the far field. Journal of Field Robotics 23(11–12), 1005–1017 (2006)
Procopio, M.J., Mulligan, J., Grudic, G.: Learning terrain segmentation with classifier ensembles for autonomous robot navigation in unstructured environments. Journal of Field Robotics 26(2), 145–175 (2009)
Procopio, M.J., Mulligan, J., Grudic, G.: Long-term learning using multiple models for outdoor autonomous robot navigation. In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2007, pp. 3158–3165 (2007)
Procopio, M.J., Mulligan, J., Grudic, G.: Learning in dynamic environments with Ensemble Selection for autonomous outdoor robot navigation. In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), September 2008, pp. 620–627 (2008)
Jacobs, R.A., Jordan, M.I., Nowlan, S.J., Hinton, G.E.: Adaptive mixtures of local experts. Neural Computation 3(1), 79–87 (1991)
Jordan, M.I., Jacobs, R.A.: Hierarchical mixtures of experts and the EM algorithm. Neural Computation 6, 181–214 (1994)
Jacobs, R.A.: Methods for combining experts’ probability assessments. Neural Computation 7(5), 867–888 (1995)
Woods, K., Kegelmeyer, W., Bowyer, K.: Combination of multiple classifiers using local accuracy estimates. IEEE Trans. Pattern Analysis and Machine Intelligence 19(4), 405–410 (1997)
Sato, M.A., Ishii, S.: On-line EM algorithm for the normalized gaussian network. Neural Computation 12(2), 407–432 (2000)
Moody, J., Darken, C.J.: Fast learning in networks of locally-tuned processing units. Neural Computation 1(2), 281–294 (1989)
Grudic, G., Mulligan, J., Otte, M., Bates, A.: Online learning of multiple perceptual models for navigation in unknown terrain. In: FSR 2007: Proceedings of the International Conference on Field and Service Robotics (2007)
Bishop, C.M.: Pattern Recognition and Machine Learning (Information Science and Statistics). Springer, New York (2006)
Cox, D.R., Snell, E.J.: Analysis of Binary Data, 2nd edn. Chapman Hall, London (1989)
Lin, C.J., Weng, R.C., Keerthi, S.S.: Trust region Newton method for large-scale logistic regression. J. Mach. Learn. Res. 9, 627–650 (2008), http://www.csie.ntu.edu.tw/~cjlin/liblinear/
Grudic, G., Mulligan, J.: Outdoor path labeling using polynomial mahalanobis distance. In: Proceedings of Robotics: Science and Systems (RSS), Philadelphia, PA (2006)
Procopio, M.J.: Hand-labeled DARPA LAGR datasets (2007), http://ml.cs.colorado.edu/~procopio/labeledlagrdata/
Hollander, M., Wolfe, D.A.: Nonparametric statistical methods, 2nd edn. Wiley-Interscience, New York (1999)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2009 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Procopio, M.J., Kegelmeyer, W.P., Grudic, G., Mulligan, J. (2009). Terrain Segmentation with On-Line Mixtures of Experts for Autonomous Robot Navigation. In: Benediktsson, J.A., Kittler, J., Roli, F. (eds) Multiple Classifier Systems. MCS 2009. Lecture Notes in Computer Science, vol 5519. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-02326-2_39
Download citation
DOI: https://doi.org/10.1007/978-3-642-02326-2_39
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-02325-5
Online ISBN: 978-3-642-02326-2
eBook Packages: Computer ScienceComputer Science (R0)