A multi-innovation state and parameter estimation algorithm for a state space system with d-step state-delay

•Study the state and parameter estimation problem of state-delay systems.•Present a joint multi-innovation state and parameter estimation algorithm.•Expand the scalar innovation in the gradient algorithm into an innovation vector.•Use the multi-innovation identification theory the state observer. Th...

Full description

Saved in:
Bibliographic Details
Published in:Signal processing Vol. 140; pp. 97 - 103
Main Authors: Xu, Ling, Ding, Feng, Gu, Ya, Alsaedi, Ahmed, Hayat, Tasawar
Format: Journal Article
Language:English
Published: Elsevier B.V 01.11.2017
Subjects:
Multi-innovation theory
Parameter estimation
Signal filtering
State space system
State space system
Signal filtering
Multi-innovation theory
Parameter estimation
ISSN:0165-1684, 1872-7557
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Study the state and parameter estimation problem of state-delay systems.•Present a joint multi-innovation state and parameter estimation algorithm.•Expand the scalar innovation in the gradient algorithm into an innovation vector.•Use the multi-innovation identification theory the state observer. This paper considers the state and parameter estimation problem of a state-delay system. On the basis of the stochastic gradient algorithm (i.e., the gradient based search estimation algorithm), this work extends the scalar innovation into an innovation vector and presents a multi-innovation gradient parameter estimation algorithm for a state-space system with d-step state-delay by means of the multi-innovation identification theory. For thesystems whose states are unknown, we use the states of the state observer for the parameter estimation and use the estimated parameters for the state estimation. This forms a joint multi-innovation state and parameter estimation algorithm for the state-delay systems with immeasurable states. The simulation results indicate that the proposed algorithms can work well.
ISSN:0165-1684
1872-7557
DOI:10.1016/j.sigpro.2017.05.006