Adaptive Input Estimation

Anders Ahlén and Mikael Sternad

IFAC Symposium ACASP 89: Adaptive Systems in Control and Signal Processing, Glasgow, UK, pp 631-636, April 1989.

In Pdf

Outline:

The paper studies the problem of estimating the input signal to a scalar discrete-time linear system The system is known, while the noise and input spectra are unknown. (This problem differs from that of blind deconvolution, where the system is unknown.)

Abstract:

An adaptive algorithm for estimating the input to a linear system is presented. This explicit self-tuning filter is based on the identification of an ARMA innovations model. From that model, input and measurement noise descriptions are decomposed.

Main tools in the algorithm are the solution of two linear systems of equations.

The basic algorithm can be used for input signals described by ARMA models and moving average measurement noise. An extension of the algorithm involves the use of on-line model reduction and spectral factorization. Simulation experiments illustrate the filtering performance.

Related publications:

Paper in IEEE Trans. ASSP 1989 on the design of linear scalar deconvolution estimators.
Later Conference paper in SPIE'91 on adaptive deconvolution.
Paper in Automatica 1990, where the identifiability conditions are derived.