Fear Learning for Flexible Decision Making in RoboCup: A Discussion

In this paper, we address the stagnation of RoboCup com- petitions in the fields of contextual perception, real-time adaptation and flexible decision-making, mainly in regards to the Standard Platform League (SPL). We argue that our Situation-Aware FEar Learning (SAFEL) model has the necessary tools to leverage the SPL competition in these fields of research, by allowing robot players to learn the behaviour profile of the opponent team at runtime. Later, players can use this knowledge to predict when an undesirable outcome is imminent, thus having the chance to act towards preventing it. We discuss specific scenarios where SAFEL’s associative learning could help to increase the positive outcomes of a team during a soccer match by means of contextual adaptation.