Research

Control of Fish School

A typical mathematical model of schooling fish involves an attraction/alignment/repulsion law, which is not necessarily well-considered in the current systems and control theory. In our research project, we are developing theories for the analysis and control of fish schooling models.

References

Shepherding

Shepherding problems refer to a problem of guiding a flock of agents from an initial location to an objective location by fewer external entities, and have several potential applications such as robotic agents to herd sheep and crowd control. We are developing novel algorithms for the shepherding of heterogeneous agents.

Proposed algorithm

Farthest Agent Targeting

References

Deep Unfolding for Model-Based Control (TAF，CREST）

We are developing model-based controller synthesis framework by utilizing the novel machine learning technique called deep unfolding. We are finding that our approach is quite effective for the control of various classes of dynamical systems including systems with saturations, delay systems, and systems with stochastic noises.

References

Epidemics Control in Social Networks (JSPS Grant-in-Aid for Early-Career Scientists)

Due to the current pandemic of COIVD-19, the containment of epidemic processes taking place in social networks is a critical problem of both theoretical and practical interest. Because the vaccines are not yet fully available (as of Mar 16th, 2021), we still need to rely on taking social distances. This social distancing strategy can be regarded as a feedback control on the epidemic processes, as the degree of social distancing is in positive corelation with the disease spread. In our research group, we are leveraging systems and control theory to develop novel theoretical frameworks for the prediction and containment of epidemic spreading processes over complex networks.

Slides

References

Positive Systems（NII Open Collaborative Research）

Positive systems refer to, roughly speaking, the class of dynamical systems, whose response signals to nonnegative input signals are constrained to be nonnegative and have applications in various areas, including pharmacology, epidemiology, population biology, and communication networks. In our research group, we are developing a novel mathematical framework for optimizing positive systems with Geometric Programming. The framework is applicable to several problems of theoretical interest such as H2 and H-infinity optimizations as well as optimization under time-delays and structured uncertainties.

Slides

References

Control Theory meets with Design Engineering（NII Open Collaboration Research）

Projects are indispensable and central in most of the industries for performing several types of work (PMI 2013). For this reason, project management has been one of the major research themes in the feld of engineering design during the last half-century. In our research project, by leveraging theoretical results about positive systems and switched systems in the systems and control theory, we are developing an optimization framework for making a cost-efcient investment in design rules when the underlying dependency structure between modules is changing over time.

Slides

References