|
(See ISLA for the new IAS website. Most maintenance activities on this old website have stopped.)
Robocup Simulation Soccer
  |
Robocup Rescue Simulation |
Objective
The RoboCupRescue project is an international cooperation to promote research and development in this socially significant domain.
The Simulation League concentrates on emergency decision support in case of a
disaster by integration of information, prediction, planning, and human interaction, which are classic AI research topics.
A generic urban disaster simulation environment is constructed on network computers. Heterogeneous intelligent agents such as fire fighters, commanders, victims, volunteers, etc. conduct search and rescue activities in this virtual disaster world.
This problem involves advanced and interdisciplinary research themes. As AI/robotics research, for example, behavior strategy (e.g. multi-agent planning, realtime/anytime planning, heterogeneity of agents, robust planning, mixed-initiative planning) is a challenging problem. For disaster researchers, RoboCupRescue works as a standard basis in order to develop practical comprehensive simulators adding necessary disaster modules.
Research group members
Prof. dr. ir. F.C.A. Groen
Drs. A. Visser
Alumni
Drs. S.B.M Post
Drs. M.L. Fassaer
Research Achievements
The Rescue Simulation environment is rich platform for multi-agent research.
We concentrated our research on the following topics:
-
The central idea of our proposed communications system is to provide all
relevant agents with the same information, in order to make them perform
more effectively as a team. Our communication protocol others a reliable
method for sharing information in a synchronized manner. We have shown
that this communication method greatly contributes to the overall amount
of knowledge at agents' disposal.
-
We designed the coordination process to coordinate agent behavior in
a way that showed cooperation and understanding of the problem space,
without using the communication channels for coming to an agreement.
To attain this, we based it on well known principals of game theory while
assuming the common knowledge provided by the communication scheme.
We created a new layer of qualitatively describing agent actions in the form
of behaviors and a new layer of describing the problem in the form of a
situation model. We showed that using a game tree to structure an agent's
thought process we could coordinate the joint actions of the rescue agents
in the near future.
References
-
S.B.M. Post, M.L. Fassaert, A. Visser "The high-level communication model for multiagent coordination in the RoboCupRescue Simulator", in D. Polani, B. Browning, A. Bonarini, K. Yoshida (Eds.), RoboCup 2003, Lecture Notes on Artificial Intelligence, Springer Verlag, Berlin. (Portable Document Format 143 Kb)
-
M.L. Fassaert, S.B.M. Post, A. Visser "The common knowledge model of a team of rescue agents", 1th International Workshop on Synthetic Simulation and Robotics to Mitigate Earthquake Disaster, Padova, Italy, 6 July 2003. (Portable Document Format 251 Kb)
This platform was also used for educational purposes and research
on the undergraduate level:
-
A. Abbo, S. Peelen, 'Progressive Deepening for GameTrees: An application for RoboRescue', June 2004, Bsc-thesis University of Amsterdam (255 Kb)
More information
-
Detailed information (including LabBook) can be found on the local Project Homepage.
|
|
