Summary

International Symposium on Nonlinear Theory and its Applications

2005

Session Number:3-1-5

Session:

Number:3-1-5-5

Localization of Cyber Rodent Based on Mixture Kalman Filters

Michiyuki Magono,  Junichiro Yoshimoto,  Shin Ishii,  Kenji Doya,  

pp.401-404

Publication Date:2005/10/18

Online ISSN:2188-5079

DOI:10.34385/proc.40.3-1-5-5

PDF download (220.3KB)

Summary:
Self-localization is one of important topics in the field of mobile robotics. This article presents a localization method based on a Bayesian approach for Cyber Rodent, which is a rodent-like mobile robot. In our method, the state transition and the sensory characteristics are formulated as a linear-Gaussian model with a redundant dimension and a probabilistic RBF network, respectively. By applying a recursive Bayesian inference, the localization algorithm is derived as a variant of mixture Kalman filters. Computer simulation results demonstrate the accuracy and the effectiveness of our method, as compared with a Monte Carlo localization based on a particle filter technique.