2018 Fiscal Year Final Research Report
Intelligent Control System Design based on Quantum Particle Swarm Optimization with NUC High-Density Beowulf Cluster
| Project/Area Number |
16K06197
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent mechanics/Mechanical systems
|
|---|
| Research Institution | Toyota Technological Institute |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Keywords | 知的制御系設計 / 非凸最適化 |
|---|
| Outline of Final Research Achievements |
In this research, we use NUC high-density Beowulf cluster computer with fully integrated dense cores capable of calculating large-scale matrix eigenvalues in parallel, and developed intelligent control system design methods with adaptive mechanisms that can cope with unknown environments by using an efficient solution approach with the cluster computer for distributed optimization including quantum particle swarm optimization. In particular, for practical "human-machine systems (semi-automatic driving, walking assist devices)", "multi-agent systems (electric power networks)", "legged robots", etc. that are required to operate properly in unknown environments, we developed methods for designing intelligent control systems and obtained basic results for integrating the method in each application domain into a unified framework.
|
| Free Research Field |
制御工学
|
| Academic Significance and Societal Importance of the Research Achievements |
未知の状況に対して適切に動作することが求められる知的制御システムの設計はデータの学習と推定を含む複雑なアルゴリズムが必要となるため，一般に非凸問題となる．本研究では，「人間機械システム」，「マルチエージェントシステム」，「脚式ロボット」の各応用分野の問題を具体的に扱い，計算機パワー（並列計算機）と分散最適化を適用して非凸問題の解を求めるアプローチの可能性を明らかにすることができた．
今後，制御理論研究のコミュニティの中で，並列計算機と分散最適化の手法により非凸問題が実用的な精度で「解ける」との認識が定着すれば，新たな理論の創生につながる可能性がある．
|
