Data-Driven Control System Design Using Various Techniques of Deep Learning
| Project/Area Number |
17K06498
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Control engineering/System engineering
|
|---|
| Research Institution | Yamaguchi University |
|---|
Principal Investigator |
Wakasa Yuji 山口大学, 大学院創成科学研究科, 教授 (60263620)
|
|---|
| Project Period (FY) |
2017-04-01 – 2022-03-31
|
| Project Status |
Completed (Fiscal Year 2021)
|
| Budget Amount *help |
¥4,680,000 (Direct Cost: ¥3,600,000、Indirect Cost: ¥1,080,000)
Fiscal Year 2019: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2018: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2017: ¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
|
|---|
| Keywords | 制御系設計 / 深層学習 / データ駆動制御 / データ駆動型制御 / 最適化 |
|---|
| Outline of Final Research Achievements |
In this study, we have developed the latest neural network structure-based controllers and their data-driven tuning methods by applying various deep learning techniques in order to improve their practicality and versatility. We have also developed data-driven control system design methods without a reference model by applying deep reinforcement learning. In particular, we have improved response prediction methods so that control performance evaluation is accelerated in the control system design process. Furthermore, we have proposed a method for constructing an evaluation model using a neural network that can represent the control engineer's knowledge and insight for control characteristics.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究は、データをいかに上手く扱うかという技術である「データサイエンス」、そして人工知能の急速な発展を支える技術である「深層学習」という、科学技術において近年注目された二つの技術を制御工学分野で応用、発展させたという点で学術的意義があると考えられる。また、高度な理論の知識を必要とせず、複雑なシステムを統一的に扱える制御系設計の枠組みを構築することは、実用性の高い技術の開発という観点から社会的意義があると考えている。
|
Report
(6 results)
Research Products
(34 results)
