Virtual Interval Sensing: Toward Safe Bounds for Dynamical Systems
Người báo cáo: Đinh Ngọc Thạch (Conservatoire National des Arts et Métiers, Sorbonne University Alliance, Paris, France)
Thời gian: 09h00 - 09h25, thứ Tư ngày 15/7/2026
Địa điểm: Phòng 508 nhà A6 Viện Toán học
Tóm tắt báo cáo: This talk begins with a general introduction to virtual sensors (i.e., real-time algorithms known as observers), followed by an explanation of how the system’s positivity property can be exploited to design interval observers capable of handling uncertainties. I will then present a unified framework for virtual interval sensing for linear systems, based on the Kazantzis–Kravaris/Luenberger (KKL) observer paradigm. The approach relies on transforming the original system into a suitable target form that enables the direct design of a virtual interval sensor. The interval bounds obtained in the transformed coordinates are subsequently mapped back to the original system variables. Owing to the generality of the KKL framework, the proposed methodology offers a systematic and flexible design procedure.
