Seminars and Events at automatic control
All seminars are held at the Department of Automatic Control, in the seminar room M 3170-73 on the third floor in the M-building, unless stated otherwise.
Seminar by Lantian Zhang: Adaptive Learning and Control with Binary-Valued Output Observations
Seminarium
From:
2025-03-20 16:45
to
17:25
Place: Seminar Room M 3170-73 at Dept. of Automatic Control, LTH
Contact: anders [dot] rantzer [at] control [dot] lth [dot] se
Date & Time: March 20, 16:45-17:25
Location: Seminar Room M 3170-73 at Dept. of Automatic Control, LTH
Speaker: Lantian Zhang, KTH Royal Institute of Technology
Title: Adaptive Learning and Control with Binary-Valued Output Observations
Abstract: Dynamical systems with nonlinear observations are widely encountered in control systems, signal processing, and machine learning. This talk considers online learning and control problems for finite-dimensional linear systems under binary-valued and randomly disturbed output observations. The main challenge lies in the inherent nonlinearity and the unavailability of traditional regression vectors required for constructing adaptive algorithms, as only binary output information is available. In this talk, we first study the adaptive estimation problem of the corresponding infinite-impulse response (IIR) dynamical systems. By leveraging double-array martingale theory, we establish global convergence results for both the adaptive prediction regret and the parameter estimation error, without resorting to such stringent data conditions as persistent excitation and bounded system signals that have been used in almost all existing related literature. Based on this, we design an adaptive control law that can effectively combine adaptive learning and feedback control. Finally, we are able to show that the closed-loop adaptive control system is optimal in the sense that the long-run average tracking error is minimized almost surely for any given bounded reference signals. Numerical examples are provided to illustrate the performance of the proposed algorithms.
Biography: Lantian Zhang received the Ph.D. degree in system theory from Academy of Mathematics and Systems Science at Chinese Academy of Sciences, in July 2024, under the direction of Prof. Lei Guo. She is currently a postdoc at the KTH Royal Institute of Technology, Sweden. Her research interests include the identification and adaptive control of nonlinear stochastic systems, adaptive estimation under quantized observations, machine learning in autonomous systems, and judicial sentencing computation.