Supervised linear feature extraction for mobile robot localization
Authors:
- N. VlassisNVN. Vlassis
- Y. MotomuraYMY. Motomura
- B. KroseBKB. Krose
Open Access
- 7 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3, 2979-2984
Abstract
We are seeking linear projections of supervised high- dimensional robot observations and an appropriate en- vironment model that optimize the robot localization task. V~b show that an appropriate risk function to minimize is the conditional entropy of the robot po- sitions given the projected observations. We propose a method of iterative optimization through a proba- bilistic model based on kernel smoothing. To obtain good starting optimization solutions we use canonical correlation analysis. We apply our method on a real experiment involving a mobile robot equipped with an omnidirectional camera in an otfice setup.Keywords
This publication has 7 references indexed in Scilit:
- Robot environment modeling via principal component regression2003
- Bayesian Landmark Learning for Mobile Robot LocalizationMachine Learning, 1998
- A Mobile Robot That Learns Its PlaceNeural Computation, 1997
- Pattern Recognition and Neural Networks1996
- Kernel Smoothing1995
- What is Projection Pursuit?Journal of the Royal Statistical Society Series A: Statistics in Society, 1987
- Algorithm AS 176: Kernel Density Estimation Using the Fast Fourier TransformJournal of the Royal Statistical Society Series C: Applied Statistics, 1982