Abstract
A simple linear time-varying filter for estimating a class of bounded deterministic signals is proposed, which is a solution of a linear autonomous difference equation, from an unknown additive measurement noise. It is composed of a bank of narrow-band filters whose outputs are scaled by a decaying factor. The filter is shown to have the desirable properties of asymptotic stability, asymptotic tracking and asymptotic noise annihilation. The performance of the proposed filter is compared with that of the Kalman filter under various measurement noise conditions. The filter is used in the identification of a linear system subject to disturbances.

This publication has 4 references indexed in Scilit: