Robocup-Rescue Simulation: in case of Fire Fighting Planning
RoboCup-Rescue project was proposed, to examine disaster prevention andmitigation using technology from RoboCup. We have implementeda disaster simulator for RoboCup-Rescue, and use it to select the optimal distribution of fire brigades. We found the “Conc
- PDF / 177,010 Bytes
- 6 Pages / 451 x 677.3 pts Page_size
- 0 Downloads / 190 Views
1 Tokyo Institute of Technology Kitano Symbiotic System Project,ERATO, Japan Science and Technology Corporation 3 Chubu University
Abstract. RoboCup-Rescue project was proposed, to examine disaster prevention and mitigation using technology from RoboCup. We have implemented a disaster simulator for RoboCup-Rescue, and use it to select the optimal distribution of fire brigades. We found the ”Concentrate Strategy” is the best in this case.
1
Introduction
Disaster mitigation is attracting attention as a new domain for Multi-Agent System research, and RoboCup-Rescue[1] is proposed. We have been designed and implemented a simulator for the RoboCup-Rescue. SoccerServer[2] has been successful as a test-bed Multi-Agent environment with a few dozen agents. The RoboCup-Rescue Simulation System differs from SoccerServer in that it has to be able to handle much larger number of heterogeneous agents. These differences provide a lot of new research issues for Multi-Agent systems. In this paper, we show an example of process to select the optimal agent distribution, in this simulation system.
2 2.1
RoboCup-Rescue Simulation System Structure of the RoboCup-Rescue Simulation System
As shown in Figure 1, the RoboCup-Rescue Simulation System consists of a core module of the system called Kernel and a number of modules which are plugged into the Kernel. It can simulate many combinations of phenomena when we plug in the necessary disaster simulators. For the sake of modularity, these plug-in components communicate with each other only via the kernel using protocols based upon UDP/IP. These protocols are formed to make it possible to pass only the necessary properties of objects, where the simulation world consists of various kinds of objects, each of which has a certain properties. This protocol does not depend on any particular simulation model and algorithm; this design allows modules to be added or removed easily. More details about the simulation system are shown in the “RoboCup-Rescue Simulator Manual”[4]. P. Stone, T. Balch, and G. Kraetzschmar (Eds.): RoboCup 2000, LNAI 2019, pp. 351-356, 2001. c Springer-Verlag Berlin Heidelberg 2001
352
Masayuki Ohta et al.
request for information viewers
notification about update
requested information
integrated simulation results GIS
integrated simulation results
simulation results kernel
component simulators
gathered commands sensory information
commands
agents
Fig. 1. Structure of RoboCup-Rescue Simulation System
2.2
RoboCup-Rescue Agent
One of the main tasks to consider in rescue activities is developing agents. Agent modules decide the action which intelligent individual is going to take. In every simulation cycle, agents receive sensory information and send back control commands. The perception information is extremely limited, and with this information, agents have to decide how they will act. Civilian, fire brigade, ambulance team and police force are now implemented. 2.3
Rescue vs. Soccer
The SoccerServer is a famous testbed for Multi-Agent research. Because of
Data Loading...
