Solving the Time Lapse from Vision System in a Robot Soccer Game Using Kalman Filter

摘要

This article discusses some effects present in the Small Size League (SSL) category of robot soccer. The dynamic of this category requires a constant update about the state of objects in the field, however, the time to generate an update is smaller than the time between information packets arriving for teams. This latency could have a negative impact in the performance of the software, which could become limited due to the camera's capability. In order to solve this kind of obstacle, a routine was elaborated using a timer to perform the update every 1ms. The update consists of using the dynamic model of the objects combined with the Kalman filter to fill the lack of information while a new package is not available, so the software can have a reliable prediction of the object's position in the field with higher frequency.