A novel localization method for indoor mobile robot based on odometry and ceiling visual features

Abstract

A localization method for indoor mobile robot which is composed of ceiling visual feature based position and posture corrector and odometry is represented in this paper. The vision-based position and posture corrector takes advantage of simple structure of the ceiling visual feature and correct the error of the position and posture estimated through encoders in every sample moment. The new localization method is efficacious while the deviation between the kinematic math model and the real physical model is obvious in a few of sample times. The result of the novel localization method is verified through simulations and experiments.