Evaluating the odometry error of a mobile robot

Abstract

This paper focuses on issues of odometry for the special case of synchronous drive wheeled mobile robots. In particular, the uncertainty in odometry is modeled by a four parameter statistical model already introduced in previous works together with a strategy to estimate the four parameters from actual data obtained from a given mobile robot. To obtain the data, the robot traverses a path. The error realized along that path is then used to estimate the model parameters. The accuracy on the parameter estimation through this strategy is here discussed by considering paths different for length and shape.