AR lattice ℓ-step ahead output uncertainty prediction scheme with uncertainty intervals

George Nikolakopoulos, Anthony Tzes

Research output: Contribution to journalArticle

Abstract

In this paper, a lattice ℓ-step ahead output uncertainty prediction algorithm is presented. The proposed prediction scheme is applicable to linear, stable, auto-regressive (AR) systems. The uncertainty predictors utilize the structure of the lattice filters. Subject to the saturation limits of the excitation signal, the measurement of the system output, and the a priori bounds of the reflection coefficients, the set of the feasible predicted output variables is computed. The suggested scheme is recursively computed over an ℓ-step ahead future time window, while the prediction uncertainty intervals are similarly computed over the lattice stages, regarding the forward and the backward prediction errors. Simulation results are presented that prove the efficacy of the employed lattice uncertainty predictors.

Original languageEnglish (US)
Pages (from-to)893-904
Number of pages12
JournalJournal of Circuits, Systems and Computers
Volume17
Issue number5
DOIs
StatePublished - Oct 1 2008

    Fingerprint

Keywords

  • Lattice filters
  • Uncertainty intervals
  • Uncertainty predictors

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this