Accelerated convergence with improved robustness for discrete-time parameter estimation

Stanislav Aranovskiy (1, 2) , Rosane Ushirobira (3) , Denis Efimov (3) , Jian Wang (4)

1 CentraleSupélec [campus de Rennes]
2 IETR - Institut d'Électronique et des Technologies du numéRique
3 VALSE - Finite-time control and estimation for distributed systems
4 School of Automation [Hangzhou]

The dynamic regressor extension and mixing (DREM) method provides a fixedtime converging parameter estimator for persistently excited regressor under bounded measurement noises. This note aims to develop this approach for cases with weaker excitation and regressor constraints. Several nonlinear estimation schemes with fixed-time convergence rates and improved measurement noise robustness properties are proposed here.

Domains: Automatic; Automatic Control Engineering; Systems and Control [cs.SY]

Stanislav Aranovskiy, Rosane Ushirobira, Denis Efimov, Jian Wang. Accelerated convergence with improved robustness for discrete-time parameter estimation. Systems and Control Letters, 2022, 55 (12), pp.324-329. ⟨10.1016/j.sysconle.2022.105344⟩. ⟨hal-03760746⟩