Abstract (EN):
This paper describes the strategy and implementation details of a cooperative agent team for the CiberMouse@RTSS08, a, robotics simulation competition. The paper introduces several concepts concerning cooperative robotics, giving all overview of its applications and challenges. It also presents the CiberMouse@RTSS08 competition rules and the associated simulation system. The proposed approach is based oil a deliberative architecture, thus providing an autonomous intelligent behaviour. A probabilistic mapping procedure, based on the Bayes' theorem, is used to maintain all internal world state. Quadtrees and A* are used for path planning and plan execution. A specific methodology was developed for beacon finding. The agents cooperate oil exchanging world and beacon information. In order to overcome the simulator's broadcasting distance limitation, a communication network between the agents was, created. The paper analyses the impact; of cooperation on this particular problem, by comparing the team's performance in four different situations: communication with network and beacon exchanging, communication without network, communication without beacon exchanging and no communication at all.
Language:
English
Type (Professor's evaluation):
Scientific
Contact:
ei05028@fe.up.pt; ei05053@fe.up.pt; ei05058@fe.up.pt; lpreis@fe.up.pt
No. of pages:
12