| Project/Area Number |
15500110
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Perception information processing/Intelligent robotics
|
|---|
| Research Institution | Kobe University |
|---|
Principal Investigator |
OHKURA Kazuhiro Kobe University, Dept. of Mechanical Engineering, Associate Professor, 工学部, 助教授 (40252788)
|
|---|
| Project Period (FY) |
2003 – 2004
|
| Project Status |
Completed (Fiscal Year 2004)
|
| Budget Amount *help |
¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 2004: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2003: ¥2,400,000 (Direct Cost: ¥2,400,000)
|
|---|
| Keywords | Autonomous multi-tobot systems / Artificial evolution / Autonomous specialization / Grid evolutionary computing environment / Sociality / 人口進化 / 自立的機能分化 / グリッドコンピューティング / 進化型計算 / マルチロボットシステム / 自律システム / 社会性エージェント |
|---|
| Research Abstract |
Artificial Life (AL) is a paradigm of designing artifacts based on the metaphor of natural evolution. After getting popularity of AL research in the middle of 1990's, this paradigm seems to be in a stable state in recent years. C.Langton, who is one of the founders of this paradigm, asserted simulated evolution is a key technique for "the life as it could be." Nowadays, evolutionary robotics has gradually been recognized as the important field that connects life form in computer simulations to artifacts showing adaptive behavior. The head investigator of this project has an idea that there are two important topics for evolutionary robotics. One is the theory of open-ended artificial evolution as natural evolution does, and the other is the theory of building social autonomous multi-robot systems. This project aimed at the theoretical development in the two areas as well as the practical validation by building autonomous robot systems and the computational grid system specialized for ev
… More
olutionary computation.
The first objective was the theoretical development of open-ended artificial evolution. The following two things were mainly considered. (1)The parameter tuning for the standard genetic algorithms applied to evolutionary robotics was investigated based on the theory of artificial evolution with neutral networks. (2)The characteristics of evolutionary robotics problems were investigated. It was found the possibility of representing the problem difficulty by the two values of the degree of neutrality and the ruggedness in a fitness landscape. The second objective was the theory of how to develop society in an autonomous multi-robot system. We have shown that the adaptive role assignment was emerged in an autonomous multi-robot as a result of autonomous specialization realized by evolving artificial neural networks. The third objective was developing a computational grid system specialized for evolutionary computation. The grid system performed the job scheduling based on the monitoring service specialized for that equalizing the computational cost for artificial evolution. Less
|
