Dead Reckoning of a Mobile Robot in 2-Dimensional Special Euclidean Group

Abstract

This paper describes a Lie group approach to dead reckoning of a mobile robot. Navigation algorithms do not use dead reckoning alone for estimation of location and attitude of a mobile robot. Dead reckoning is used for prediction or rough guess of robot location for a short time. It is combined with other information to find the best possible estimation of the robot location and attitude. More precise dead reckoning improves the localization performance. Moreover, proper formulation of dead reckoning makes the location estimation problem more consistent and robust. This paper formulates dead reckoning of a mobile robot in Lie group. For dead reckoning in 2-dimensional space, 2-dimensional special Euclidean group is used. Simulation results show that the proposed method provides exact prediction of robot location and attitude regardless of the sampling period.