| Project/Area Number |
09650284
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent mechanics/Mechanical systems
|
|---|
| Research Institution | Hokkaido University |
|---|
Principal Investigator |
KAKAZU Yukinori Grad.School of Eng.Hokkaido Univ., Pro., 大学院・工学研究科, 教授 (60042090)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
YOKOI Hiroshi Gad.School of Eng.Hokkaido Univ., Asso.Pro., 大学院・工学研究科, 助教授 (90271634)
|
|---|
| Project Period (FY) |
1997 – 1998
|
| Project Status |
Completed (Fiscal Year 1998)
|
| Budget Amount *help |
¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1998: ¥800,000 (Direct Cost: ¥800,000)
|
|---|
| Keywords | Autonomous Robot (Agent) / Cooperative Behavior / Competitive System / Iterative Prisoner's Dilemma / Vision |
|---|
| Research Abstract |
The automatic operations of loading and unloading are one of the most important research themes of factory automation in industry. However, this type of problem is classified into the large scale of scheduling problem that is difficult to solve using combinatorial optimization techniques. For these problems, multiple and cooperating schedulers method has been proposed that is based on parallel computation and distributed autonomous computation. Such method is useful for quick computation by the reasonable schedulers that has a distributed function of computation load. So how to design the reasonable schedulers becomes the key point of this automatic operations of loading and unloading problem
In this research, we propose IPD architecture based on mathematical model of competitive systems. For the automatic loading and unloading problem, autonomous small robots are designed that has two functions of scheduling and loading. The scheduling function is improved by the feedback of evaluating value which is sum of efficiency of robot's work. This improvement is a self-learning process of the proposed architecture based on prisoner's dilemma. This report shows theory of proposed IPD architecture and results of loading and unloading problem.
As a result of this research, the hybrid problem of cooperating loading and unloading operation is solved by new concept of competitive computation of IPD architecture. The IPD architecture is applied to the autonomous scheduling robots group, and the result shows self-learning function of improvement of scheduling capability. Basic phenomena of social system of autonomous robot group is shown and published.
|
