| Project/Area Number |
08650504
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
計測・制御工学
|
|---|
| Research Institution | University of the Ryukyus |
|---|
Principal Investigator |
MIYAGI Hayao Univ. of the Ryukyus, Faculty of Eng. Professor., 工学部, 教授 (40112445)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
YAMADA Koji Univ. of the Ryukyus, Faculty of Eng. Lecturer, 工学部, 講師 (90274886)
YAMASHITA Katsumi Univ. of the Ryukyus, Faculty OF Eng. Professor., 工学部, 教授 (60158152)
|
|---|
| Project Period (FY) |
1996 – 1997
|
| Project Status |
Completed (Fiscal Year 1997)
|
| Budget Amount *help |
¥1,700,000 (Direct Cost: ¥1,700,000)
Fiscal Year 1997: ¥600,000 (Direct Cost: ¥600,000)
Fiscal Year 1996: ¥1,100,000 (Direct Cost: ¥1,100,000)
|
|---|
| Keywords | Autonomous robot / Fuzzy control / Robust perturbation technique / Fuzzy relation equation / GA / Lyapunov method / Fuzzy inference / Co-evolution approach / ル-リエ形リアプノフ関数 / ファジイ推論 / 競合共進化モデル / 意思決定 / 残留ファジィ測度 / ファジィ型非線形システム / クラシファイアシステム |
|---|
| Research Abstract |
1.The report presents a pattern classification for systems with arbitrary nonlinear feedbacks. Classification is carried using the robust nonlineear-perturbation technipue, in which nonlinearities are regarded as the non-linear perturbation of linear stable systems. The results show that the autonomous robot system can keep stable if nonlinearities lie in the robust-stable range.
2. We have proposed an algorithm that solves fuzyy relation equations given by Sup・min or Inf・max composite. The solution is useful to establish fuzzy inference.
3. The genetic algorithm(GA) requires setting a fitness function and evaluat-ing it to solve the probrem. then, for the autonomous acquisition problem of game strategy, it is not easy to apply GA because the fitness function can not be fixed in changeable situation of game. The report gives that the proposed co-evolution approach is useful to the problem.
4. The "antecedent" and "consequent" conditions of fuzzy inference systems are well correlated with input and output of conventional nonlinear feedbacks, respectively. The robust perturbation technique shows that if the nonlinearities are within the robust perturbation limit, the fuzzy control system is verified to be stable according to the stability theorm of Lyapunov. The report has investigated the stability of fuzzy contrl systems using thetechnique.
|
