Outdoor Evacuation Experiment Augmented by Massively Multiagent Simulation

The spread of mobile terminals like cellular phones and PDAs, and positioning systems like GPS, will realize a ubiquitous environment for citydwellers. Ubiquitous/pervasive computing systems in the public space often interact with anonymous users. A large-scale evacuation navigation system, which is an example of socially embedded system, is selected as a target application in this research. Any socially embedded system must be verified from the macro standpoint and the micro standpoint. The former is concerned with how efficient all citizens use system. The latter addresses how a citizen uses the system. Verification will be accurate only when the human subjects feel as if their environment is populated with an adequate number of participants. But it is quite difficult to conduct an experiment with a large number of human subjects. We build an environment for verifying a evacuation navigation system with a small number of human subjects. The goal of this paper is to augment a real world experiments with multiagent simulation for testing socially embedded systems. An augmented environment for outdoor experiment consists of multiagent simulations, GPScapable cellular phones and observation monitor. We conduct proving test and confirm that the environment can create a situation in which human subjects feel like they are participating with a large number of humans. Augmented experiment enhances a real world experiment with multiagent simulation. The augmented experiment system lets human subjects perceive the results of the simulation performed in the virtual space concurrently with the real-world experiment. The human subjects are made to feel that they are participating with many other humans. This approach yields data comparable to that by performing real-world tests with a great many human subjects. In addition, comparing the movements of the human sub-

[2]  M. F.,et al.  Bibliography , 1985, Experimental Gerontology.

[3]  Hirofumi Yamaki,et al.  Disaster Evacuation Guide: Using a Massively Multiagent Server and GPS Mobile Phones , 2007, 2007 International Symposium on Applications and the Internet.

[4]  Hirofumi Yamaki,et al.  Protocol Description and Platform in Massively Multiagent Simulation , 2006 .

[5]  Hideyuki Nakanishi,et al.  FreeWalk/Q: social interaction platform in virtual space , 2004, VRST '04.

[6]  Hirofumi Yamaki,et al.  Caribbean/Q: A Massively Multi-Agent Platform with Scenario Description Language , 2006, SKG.

[7]  Hideyuki Nakanishi,et al.  Transcendent communication: location-based guidance for large-scale public spaces , 2004, CHI.

[8]  Hideyuki Nakanishi,et al.  Augmented Experiment: Participatory Design with Multiagent Simulation , 2007, IJCAI.

[9]  H. Van Dyke Parunak,et al.  Representing Agent Interaction Protocols in UML , 2000, AOSE.

[10]  Zhengping Li,et al.  A Survey of Emergent Behavior and Its Impacts in Agent-based Systems , 2006, 2006 4th IEEE International Conference on Industrial Informatics.

[11]  Award , 2007, The Veterinary record.

[12]  Ishida Toru,et al.  Implementation and Evaluation of Navigation Support System with GPS Mobile Phones , 2007 .

[13]  Hiromitsu Hattori,et al.  Agent modeling with individual human behaviors , 2009, AAMAS.

[14]  Hirofumi Yamaki,et al.  Development of Mega-Scale Navigation Based on Multiagent Simulation , 2005 .

[15]  Peter Stone,et al.  Multiagent traffic management: an improved intersection control mechanism , 2005, AAMAS '05.

[16]  Hideyuki Nakashima,et al.  Evaluation of Usability of Dial-a-Ride Systems by Social Simulation , 2003, MABS.

[17]  Toru Ishida,et al.  Analysis of Pedestrian Navigation Using Cellular Phones , 2007, PRIMA.

[18]  Territoire Urbain,et al.  Convention , 1955, Hidden Nature.

[19]  Hiromitsu Hattori,et al.  Evacuation guide system based on massively multiagent system , 2008, AAMAS.

[20]  Gaku Yamamoto,et al.  Agent Server for a Location-Aware Personalized Notification Service , 2004, MMAS.