Filtering of Systems with Nonlinearities
Abstract
Estimation problems, and filtering among them, are basically concerned with extracting the best information from inaccurate observation of signals. Perhaps the earliest roots of this type of problems go back to the least squares estimation at the time of Galileo Galilei in 1632 and Gauss in 1795. The relatively modern and more general development of least-squares estimation in stochastic processes is marked by the work of A.N. Kolmogorov and N. Wiener in the 1940's. Most recently, and due to vast research and development of the space age, the estimation theory experienced a new outlook. This was marked by the work of P. Swerling in 1958 and 1959 in connection with satellite tracking, and the work of R. Kalman using state space approach. Kalman's work had the impact of greatly popularizing and spreading the estimation theory in different fields of applications. Also, works by Stratonovich and Kushner are among the recent developments of the subject. In this document the nonlinear filtering problem is treated using a new approach. The approach consists of unifying a system model approximation technique with the filtering solution based on the approximate model. As a result, this paper describes three new practically implementable filters for stochastic dynamic systems which include nonlinearities in their structure.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1982
- Accession Number
- ADA124635
Entities
People
- Hosam E. Emara-shabaik
Organizations
- University of California, Los Angeles