Sensor Fusion and Play Strategy Programming for Micro Soccer Robots

Abstract

Our research focused on the programming of the vision system and strategy for a soccer robot system. For the vision system, the varying lighting conditions of the robot soccer field attributed to problems in identify colors of different objects on the field. We developed a light adaptable vision system by applying the YUV threshold method. We also proposed an object position correction approach to deal with the lens distortion in the image. This approach avoided the time-consuming image calibration process, by adjusting the object positions after image processing. Experimental results showed that our vision system had a high recognition rate and accurate position computation even under changing lighting conditions. The development of a strategy program to avoid collision is also covered in this paper.