Localization of Wheeled Mobile Robots from Slip Ratio Estimation with Simple Model

Abstract

This paper proposes a localization method by estimating the slip ratio of wheeled mobile robots. The slip ratio estimator is designed based on the equations of motion. Input torque and encoders data are used, and driving force, vehicle velocity, and slip rate are estimated. This method can estimate the robot's position accurately without visual odometry (VO) when the driving resistance is obtained. When VO is available, even if the obtained driving resistance is different from its true value, instantaneous speed observer updates the resistance and the robot's velocity. The sampling time of torque is very short, which allows us to obtain position estimation much faster than VO. The experiments were conducted with a robot on alumina balls, and the proposed methods showed improvements in the accuracy of localization compared to wheel odometry.