Preprints, Working Papers, ... Year : 2025

Regularization of the DREM procedure for parameter estimation under partial excitation

Abstract

This paper investigates estimating unknown constant parameters in linear regression with measurement noise. It introduces two concepts to analyze different types of excitation of the regressor: partial and feeble. The former indicates the absence of persistent or interval excitation, while the latter characterizes the presence of excitation that is inadequate for efficient estimation in a noisy environment. Furthermore, the paper proposes integrating regularization techniques to improve the estimation performance of the dynamic regressor extension and mixing (DREM) method. This enhancement is scrutinized analytically to gauge the extent of improvement achieved. The efficacy of theoretical findings is illustrated in the simulations.

Fichier principal
Vignette du fichier
Regularization_and_unexcited_lre_main.pdf (353) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04894803 , version 1 (17-01-2025)

Identifiers

  • HAL Id : hal-04894803 , version 1

Cite

Jian Wang, Stanislav Aranovskiy, Rosane Ushirobira, Denis Efimov. Regularization of the DREM procedure for parameter estimation under partial excitation. 2025. ⟨hal-04894803⟩
0 View
0 Download

Share

More