Input perturbations for adaptive control and learning

DOI10.1016/J.AUTOMATICA.2020.108950zbMATH Open1441.93148arXiv1811.04258OpenAlexW3015500735MaRDI QIDQ2184498FDOQ2184498

Authors: Mohamad Kazem Shirani Faradonbeh, Ambuj Tewari, George Michailidis 

Publication date: 29 May 2020

Published in: Automatica (Search for Journal in Brave)

Abstract: This paper studies adaptive algorithms for simultaneous regulation (i.e., control) and estimation (i.e., learning) of Multiple Input Multiple Output (MIMO) linear dynamical systems. It proposes practical, easy to implement control policies based on perturbations of input signals. Such policies are shown to achieve a worst-case regret that scales as the square-root of the time horizon, and holds uniformly over time. Further, it discusses specific settings where such greedy policies attain the information theoretic lower bound of logarithmic regret. To establish the results, recent advances on self-normalized martingales together with a novel method of policy decomposition are leveraged.

Full work available at URL: https://arxiv.org/abs/1811.04258

Recommendations

• On adaptive linear-quadratic regulators
• scientific article; zbMATH DE number 440431
• Adaptive regulator for MIMO Hammerstein and Wiener systems
• Low-complexity learning of linear quadratic regulators from noisy data
• Controlling unknown linear dynamics with bounded multiplicative regret

zbMATH Keywords

system identificationlinear-quadraticdecision-making under uncertaintyexploration-exploitationadaptive LQRsfinite-time optimalitygreedy policies

Mathematics Subject Classification ID

Linear systems in control theory (93C05) Multivariable systems, multidimensional control systems (93C35) Adaptive control/observation systems (93C40) Perturbations in control/observation systems (93C73)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• User-friendly tail bounds for sums of random matrices
• Asymptotically efficient adaptive allocation rules
• Title not available (Why is that?)
• Online learning in the embedded manifold of low-rank matrices
• Small sample properties of forecasts from autoregressive models under structural breaks
• Optimal control of LTI systems over unreliable communication links
• Control Techniques for Complex Networks
• The AAstrom-Wittenmark self-tuning regulator revisited and ELS-based adaptive trackers
• Title not available (Why is that?)
• Adaptive Linear Quadratic Gaussian Control: The Cost-Biased Approach Revisited
• Stabilization of discrete-time nonlinear uncertain systems by feedback based on LS algorithm
• A note on the structure of two subsets of the parameter space in adaptive control problems
• On the necessity of identifying the true parameter in adaptive LQ control
• Finite time identification in unstable linear systems
• Optimality of Fast-Matching Algorithms for Random Networks With Applications to Structural Controllability
• Finite-Time Adaptive Stabilization of Linear Systems
• On adaptive linear-quadratic regulators
• Adaptive control of linear time invariant systems: the ``Bet on the best principle

Cited In (8)

• Adaptive Optimal Feedback Control with Learned Internal Dynamics Models
• Remarks on input to state stability of perturbed gradient flows, motivated by model-free feedback control learning
• Enhanced P-Type Control: Indirect Adaptive Learning From Set-Point Updates
• Title not available (Why is that?)
• Performance analysis of the compressed distributed least squares algorithm
• On adaptive linear-quadratic regulators
• Title not available (Why is that?)
• Joint learning of linear time-invariant dynamical systems

This page was built for publication: Input perturbations for adaptive control and learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2184498)