Precise Long Passing Skill on Middle Size League Robot Soccer ERSOW Using Teammate Detection and Pivot Motion

Abstract

The ERSOW robot, a competitor in robot soccer leagues, struggles with its current long-distance passing technique. This method, reliant on robot localization to feed robot orientation adjustments, suffers from accumulating odometry errors and unintended deviations in the ball's trajectory. This research proposes a novel approach to optimize the ERSOW robot's long-distance passing skills, addressing these limitations. The proposed solution incorporates two key elements: ball pivoting movements and a vision system. Ball pivoting maneuvers enable the robot to maintain control of the ball's position during orientation adjustments, minimizing deviations caused by robot movement. Additionally, the integration of a vision system allows the robot to continuously track the receiving teammate's location in real-time. This real-time information enables adjustments to the passing trajectory, mitigating the negative effects of odometry errors. The effectiveness of the proposed approach was evaluated through experimentation. The results demonstrate a success rate of 100% for kickoff passes with an average completion time of 1.21 seconds at a distance of 400 cm and 95.83% for corner passes with an average completion time of 2.44 seconds at a distance of 600 cm. This suggests that 600 cm is the optimal passing distance for achieving successful long passes.