Research projects

Plug & Playスマートシティとは、インフラ機能がフレキシブルかつスマートに連携可能な都市である。これは、サイバーシステム、物理システム、ヒューマンの複雑な相互作用を有する大規模統合動的ネットワークシステムと見なすことが出来る。 多数のステークホルダーが利己的に振る舞う状況下で、CO2排出抑制など環境に配慮しながら社会システムの健全性を保証する社会設計法の確立は喫緊の課題である。 本研究では、制御学と経済学の知見を総動員し、Plug & Playスマートシティのモデリング、分散協調最適化アルゴリズムの設計、経済効果と費用便益までを一気通貫で実施し、動的メカニズムデザインの本質を明らかにする。

The objective of this study is understanding of adaptive function to environmental change of social insects and construction of adaptive strategies that can control swarm robots based on the understanding. We obtained following results. (1) We analyze behaviors of ants by biological experiments and behaviors of multi-agent systems based on Boids model. We design a control law of swarm robots and carry out experimental validation of formation control of multi-UAVs. (2) The division of labor based on social communication of ants between two nests with large distance can be found in experimental observation. (3) Estimation function of global information based on local behaviors is analyzed. Based on the derived decentralized control strategies experimental demonstration using modular legged robots is carried out. (4) An agent dynamic model is constructed based on Bayesian and Inverse-Bayesian estimation methods.

The distributed and optimal control for large-scale dynamical systems with hierarchical structures and multi time-scale properties was investigated and some application technologies of the proposed control algorithms for distributed energy management systems were studied. First, a novel direct load control method to stabilize the system frequency was proposed. The proposed method utilized the consumers' controllable loads for load frequency control. Second, a load frequency control of a microgrid system was investigated based on static H-infinity control algorithm considering time-scale properties of generators. Then, hierarchical and cooperative model predictive control methodology for load frequency of large-scale electrical power network was studied. An efficient frequency control method was proposed considering economic dispatch and load frequency regulation. Finally, numerical simulation results and experimental results showed the effectiveness of the proposed approaches.