ISSN 1991-2927
 

ACP № 4 (58) 2019

Towards Robust Riccati Iterations for Lqg-optimal or Parameter-adaptive Estimation and Control Processes

Innokentiy Vasilyevich Semushin, Ulyanovsk State University, Doctor of Science in Engineering, Professor of Information Technology at Ulyanovsk State University. Memberships in Professional Organizations: IEEE Society; IEEE Control Systems Society; “Russian Professorial Assembly”. Author of papers, monographs and textbooks; holds patents for inventions. Field of Interest: systems theory, control. [e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it. ]I. Semushin

Towards Robust Riccati Iterations for Lqg-optimal or Parameter-adaptive Estimation and Control Processes 52_7.pdf

The paper aims at the development of robust and efficient computational algorithms for stochastic linear control based on scalarized square-root implementations. In the capacity of starting point, it uses the classical (formal) solution to the control problem known from the three-decker monography Stochastic models, estimation, and control, Academic Press, 1979-1982, by Peter S. Maybeck. The mutually time-inverse Riccati iterations being a core part of the solution, are interpreted in a uniform notation as the two-stage update processes. For them, a transition to the two kinds of scalarized algorithms - direct and inverse, is performed to introduce into consideration a scalarized filter and scalarized regulator and avail ourselves of the opportunity to go into the question of numerically stable square-root computation designs for Riccati iterations. The paper demonstrates one possible configuration in the form of the Potter-style algorithm. That establishes the new direction in constructing robust LQG-regulators for control as being based on many advances in robust filtering.

Lqg-control, square-root algorithms, finite-horizon discrete-time control, scalarization.

© FRPC JSC 'RPA 'Mars', 2009-2018 The web-site runs on Joomla!