Location Algorithm Implementation Using Kalman Filter Applied to a Mobile Robot

J.A. Gonzalez-Andrade, A. Diaz Santana, G. Velazquez Carrillo

Abstract:	This paper proposes the implementation of an algorithm using the Kalman filter to achieve the fusion of three different absolute and relative sensors and thus to obtain the location of the mobile robot through the high voltage transmission line with an error less than + / - 50 cm. It is intended to use an encoder as a relative information sensor, and an inclinometer and a barometer as absolute information sensors. In the proposed algorithm, a Kalman filter was used to integrate the data obtained by the odometer model together with the sensors that provide absolute information on the location of the mobile robot. The odometry model previously obtained was used as a transition matrix of the system model. MATLAB was used to simulate the path of the mobile robot through the catenary curve, to then get the model that describes the covariance matrices of both; the prediction stage and the correction stage of the Kalman filter. A methodology to eliminate systematic errors of the three sensors was implemented. Finally the results of simulations are presented. To evaluate the location algorithm performance, a comparison of the error obtained along the path was made with and without the use of the proposed algorithm.
Published in:	Proceedings of the 2011 International Technical Meeting of The Institute of Navigation January 24 - 26, 2011 Catamaran Resort Hotel San Diego, CA
Pages:	1102 - 1108
Cite this article:	Gonzalez-Andrade, J.A., Santana, A. Diaz, Carrillo, G. Velazquez, "Location Algorithm Implementation Using Kalman Filter Applied to a Mobile Robot," Proceedings of the 2011 International Technical Meeting of The Institute of Navigation, San Diego, CA, January 2011, pp. 1102-1108.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In