O. Zweigle, U.-P. Käppeler, R. Lafrenz, H. Rajaie, F. Schreiber, and P. Levi (Germany)
Situation Recognition, Cooperative Robotics, Behavior
Situation recognition in RoboCup gets more and more im portant to enhance team behavior strategies. In this paper we present an approach for a cooperative situation recogni tion architecture which directly influences the game strat egy of the own team. The presented situation recognition architecture is derived from the human archetype to recog nize situations by a combination of learned knowledge and observation of the environment. Humans only consider a limited number of situation interpretations for certain envi ronmental scenes. We transfer that concept and introduce Situation Libraries as an analogous approach for computer systems. Consequently for every recognized situation a co operative team behavior is defined. This will effectively enhance the team play and the ability of a team to adapt its behavior to the strategies of an opponent and to the current game state.
Important Links:
Go Back
