Omnidirectional Vision-Based Localization for ERSOW Middle Size League Soccer Robot

EEPIS Robot Soccer On Wheeled (ERSOW) is a robot soccer team representing the Politeknik Elektronika Negeri Surabaya has been designed to compete in national and international robot soccer competitions. A key component of an autonomous soccer robot system is accurate position estimation on the field. The existing method employed by the wheeled soccer team authors relies on odometry sensors, which are prone to significant errors due to wheel slip affecting the rotary encoder's accuracy. To overcome this limitation, this study proposes the use of an omnidirectional camera combined with a color segmentation method to detect the field's boundary fence, serving as a positional reference. The wheeled soccer robot team, equipped with this camera, can detect the field boundary from 20 cm to 360 cm, with optimal stability between 20 cm and 320 cm. The proposed method demonstrated a substantial improvement in position estimation accuracy compared to the odometry-based method, reducing the average error in the X axis by 5 cm (71.67 %) and in the Y -axis by 4 cm (75.49 %).