Real-Time Self-Localization of a Mobile Robot by Vision and Motion System

An autonomous robot with an omni-vision camera and omni-moving platform is designed to satisfy the requirement of the Federation of International Robot-soccer Association and RoboCup robot soccer competitions. To obtain the robot’s location on the field, we use a white-line pattern match localization algorithm. However, when the white-line information is incomplete during the matching process, the observed data differ significantly from a pre-built database. Thus, the localization result causes errors. In this study, we introduce an encoder localization algorithm to obtain a robot’s moving direction and distance. We propose an algorithm that integrates white-line pattern match localization and encoder localization. In the integration process, we use a fuzzy system to search the surrounding points to localize the robot. The results demonstrate that integration localization outperforms localization with only white-line pattern match localization or encoder localization. With the proposed algorithm, we can obtain the robot’s location within 30 ms at an error of less than 10 cm. The integration localization algorithm is compared to other methods to demonstrate its performance.