| Project/Area Number |
15500085
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | The University of Tokyo (2005)
The University of Electro-Communications (2003-2004) |
|---|
Principal Investigator |
TAKECHI Ikuo The University of Tokyo, Graduate School of Information Science and Technology, Professor, 大学院・情報理工学系研究科, 教授 (90293109)
|
|---|
| Project Period (FY) |
2003 – 2005
|
| Project Status |
Completed (Fiscal Year 2005)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2005: ¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 2004: ¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 2003: ¥1,200,000 (Direct Cost: ¥1,200,000)
|
|---|
| Keywords | real time / distributed cooperative system / multi-agent system / RoboCup / RoboCupRescue / limited communiction / cognitive science / Robo Cup / エージェント / シュミレーション / ロボカップサッカー / 認知科学的手法 / かけ声 / 3次元サッカーシュミレーション / ロボカップ / ロボカップレスキュー |
|---|
| Research Abstract |
(1) We developed a soccer agent which can detect opponent ability by visual observation. It is needed to make the agent action suitable for various type of opponent agents.
(2) We developed various short shouts, namely very short communication, to activate rather a long term strategic team play of more than two players. This kind of short shouts are more advanced than those which invoke only reflective actions of other players.
(3) We investigated and tested how to remedy inappropriate, unexpected agent actions which are often observed in matches with other teams in RoboCup soccer simulation league. We devised a mechanism to involve another inner agent in the soccer agent ; that is, we made an agent which is in turn a multi-agent system. Newly introduced inner agent monitors the original (base) agent whether it is to issue an inappropriate action command by referring to its action history and the current situation. If the inner agent finds something wrong, it interrupts the base agent to change the command, and reports the relevant log to the developer for debugging. Through our experiments, we could make a new research plan for agent with reflection and team with reflection.
(4) We investigated a new way of the extension of 2D soccer simulation to 3D soccer simulation. Our research focused on the improvement of the current 3D agent model toward more natural and humanlike 3D model with a good physical simulation model.
|
