Optimal control provides a framework for designing controllers that maximize the dynamic performance of a given system configuration. We aim to extend this framework beyond traditional optimal control to establish methodologies for optimal system design, in which both the controller and the structural components of the system are designed simultaneously.
We develop design methodologies that simultaneously optimize the physical structures and controllers of feedback control systems. This integrated approach enables efficient designs by leveraging the strengths and compensating for the weaknesses of both the structural and control components.
Traditionally, system design has followed a two-step process: first defining the physical structure, such as dimensions and material properties, and then designing a controller to operate it. By designing them concurrently, however, we can create structures that are easier to control, reduce control effort, avoid excessive material use, and lower overall costs through more efficient control.
To achieve this, we extend optimal control theory and reinforcement learning beyond conventional use, enabling the joint optimization of physical structures and controllers. Unlike existing methods that often assume linear system responses or focus on short control horizons, our approach improves long-horizon control performance even for nonlinear systems.
These methods can be applied not only to simultaneously optimize the structures and controllers of mechanical systems such as robots but also to integrated system design for energy systems, where operational and infrastructure planning are considered together.
| Joint and Contract Research Achievements | ||
|---|---|---|
| Information on conferences, exhibitions, and other related events | [1] K. Nishimura, H. Hoshino, and E. Furutani, Combined plant and control co-design via solutions of Hamilton-Jacobi-Bellman equation based on physics-informed learning, 2024 Joint 13th International Conference on Soft Computing and Intelligent Systems and 25th International Symposium on Advanced Intelligent Systems (SCIS&ISIS), Himeji, Japan, November 9-12 (2024) [2] H. Hoshino, An approach to integrated design of physical structure and controller based on Galerkin approximations of Hamilton-Jacobi equation, Proceedings of the 2023 International Symposium on Nonlinear Theory and its Applications (NOLTA2023), pp.403-406, Catania, Italy, September 26–29 (2023). |
|
Learn about the cutting-edge research achievements, advanced technical resources, and know-how that the Graduate School of Engineering at the University of Hyogo can provide. We welcome partners who can create the future together with us by jointly generating new ideas and technologies through collaborative research, commissioned research, and technical consultations.
More
