研究会 (2026 年 07 月 25 日)

Abstract:
1. (Part I)
   JST ASPIRE is a research support program designed to:
    - Build a network among the world's top research talent.
    - Promote Japan's participation in the international top research community.
    - Nurture and mobilize future research leaders.
   In 2024, we started the project entitled
   "Building Mathematical Foundation for Cyber Physical Dynamical Systems:
   Interdisciplinary Research and Human Resource Development on Control
   with Prediction and Learning" that is abbreviated as "ASPIRE-CPDS."
   In this talk, we report our activities of  ASPIRE-CPDS up to FY 2025.
   (Part II)
   This work addresses the absolute stability analysis problem of feedback
   systems with slope-restricted and repeated nonlinearities.
   Under the integral quadratic constraint framework, we can employ static
   O'Shea-Zames-Falb multipliers to obtain an LMI-based criterion that
   ensures absolute stability. However, this criterion is only a sufficient
   condition for the absolute stability. In other words, if this ``primal''
   LMI is infeasible, no definitive conclusion can be drawn. To overcome
   this limitation, we consider the corresponding dual LMI that becomes
   feasible if and only if the primal LMI is infeasible.
   Our main result establishes that if the dual solution satisfies a certain
   rank condition, then it is possible to construct both a destabilizing
   slope-restricted nonlinearity and a non-zero equilibrium state, thereby
   proving that the system of interest is not absolutely stable. We also
   report our latest extension of these results to other problem settings.

2. This talk introduces a new robust state feedback control scheme for
   Linear Parameter Varying (LPV) systems subject to bounded disturbances
   and noises. Serving as a deterministic, zonotopic counterpart to the
   classical stochastic LQG architecture, the framework features an offline-
   designed Zonotopic Kalman Filter (ZKF) for fast state estimation and a
   Linear Quadratic Zonotopic (LQZ) control framework for feedback synthesis.
   By proving the separation principle for this architecture, the optimal
   controller and estimator gains are computed independently by solving offline
   Linear Matrix Inequalities (LMIs). Finally, the framework’s real-world
   performance is demonstrated through a vehicle lateral dynamics path-tracking
   application validated using real experimental data.

   Bio: Sara IFQIR received the M.Sc. degree in electronics, electrical
   energy and automatic control, and the Ph.D. degree in automatic control
   from the University of Paris-Saclay, France, in 2016 and 2019, respectively.
   From November 2019 to August 2021, she worked as a postdoctoral researcher
   with the University of Bordeaux, France. Since September 2021, she joined
   Centrale Lille Institut and CRIStAL Laboratory as Associate Professor.
   Her current research interests include nonlinear, switched and multi-agents
   systems, robust control, set-membership estimation and fault diagnosis for
   complex physical systems with special focus on the fields of intelligent
   vehicles.