Stochastic Systems

On error evaluation of the quasilinear filtering method

M. L. Dashevskii, I. D. Siluyanova, V. I. Shin

Moscow

Abstract: The paper is concerned with errors of the linear filtering method which statisticall y linearizes the nonlinear functions in the right hand sides of differential equations which describe the useful signals. The numerical results of the studies demonstrate that application of Kalman filters to a statistically linearized nonlinear system can result in biased estimates and significant deflection of the model signal filtering error variance from that of the actual useful signal.

UDC: 621.391.272


Received: 16.03.1983

 English version: Automation and Remote Control, 1984, 45:6, 731–737

Bibliographic databases:

•