State machines synchronization for collaborative behaviors applied to centralized robot soccer teams

Abstract

In robot soccer, collaborative behaviors are necessary to establish team coordination. In centralized architectures with global perception, the team coordination is carried out by a making decision system, where the team strategy is programmed out. Finite state machines are an alternative for the making decision systems design in order to assign players roles and behaviors, depending on the game conditions. In this paper a team strategy for robot soccer architectures with global perception and centralized control is proposed, through the use of synchronized state machines for collaborative behaviors among the players by using a synchronization function in some determinate states. This function is used to synchronize one machine state which selects the behavior of one player, with other state which selects the behavior of another player. The synchronization is used, for instance, to coordinate a pass between two players looking for a goal, or blocking an opposite goal by an opposite defender player. Synchronized state machines presented better results than strategies with state machines non-synchronized on different matches played.