Decision-making for 5x5 Very Small Size Soccer Teams

• Thayana Baldão ITA
• Marcos Maximo ITA
• Cecilia Cesar ITA

Resumo

Due to the high dynamism of IEEE Very Small Size Soccer (VSSS) matches, the main differences between the teams are focused on their decisionmaking system, path planning methods, and control and positioning algorithms. A path planning algorithm that fits well in the context of IEEE VSSS differential robots is the univector field navigation. For the positioning of the robots the combination of Delaunay Triangulation and Gouraud Shading proves to be quite efficient. Regarding decision-making, behavior trees are a good approach, due to their modularity and easiness in changing player’s actions. In addition, a system for changing player’s roles is also important, in order to make a dynamic match possible. The techniques shown here are not novel in the context of robot soccer. In fact, this paper contribution lies in integrating them in a full working strategy for 5x5 VSSS, which ended up winning Latin American Robotics Competition (LARC) 2019.

Palavras-chave: Decision making, Games, Educational robots, Robot sensing systems, Navigation, Sports, Trajectory planning