Comparing Sensors for Position Estimation of a Small Size League Robot

• João Victor Lourenço Aguiar Centro Universitário FEI https://orcid.org/0009-0009-2544-0161
• Flavio Tonidandel Centro Universitário FEI https://orcid.org/0000-0003-0345-668X

Resumo

Mobile robots are increasingly relevant across various applications, where accurate localization-determining the robot's position within its environment-remains a fundamental and challenging task for effective navigation and decision-making. This paper details a comparative analysis of sensor configurations for an embedded localization system in RoboCup Small Size League robots. Using an Extended Kalman Filter (EKF), we fuse data from wheel encoders, an Inertial Measurement Unit (IMU), the SSL-Vision system, and an omnidirectional robot kinematic model. Results show that combining onboard sensors (encoders and IMU) in the EKF's prediction phase with SSL-Vision for correction significantly improves accuracy and maintains a rapid 3.5 ms update rate. This sensor fusion proves effective in dynamic environments and mitigates temporary vision occlusions. Future work involves implementing this system into robot position control for enhanced gameplay.

Palavras-chave: Location awareness, Accuracy, Wheels, Position control, Kinematics, Sensor fusion, Robot sensing systems, Mobile robots, Kalman filters, Robots, localization, prediction, correction, sensors