SSP'05 -- Proceedings

	IEEE/SP 13th workshop on Statistical Signal Processing July, 17-20, 2005 - Bordeaux - France

Information regarding the paper

Title

Parameter estimation in a class of nonlinear state - space models

Author(s)

Mihai Enescu	Helsinki University of Technology
Visa Koivunen	Helsinki University of Technology

Get the paper in PDF format

To obtain Acrobat Reader (version 5 minimum required) necessary to his read.

Abstract

Many applications rely on a state-space description of the input-output dynamics of the system. In this paper we address the problem of parameter estimation in a class of nonlinear state-space models. The state transition coefficient and the process noise variance are estimated from the received observations in a recursive manner. Some properties of the characteristic function (in this case for normally distributed variables) are exploited. Two applications are presented, estimation of the carrier frequency offset in OFDM and propagation path parameter estimation in channel sounding. Illustrative examples show the reliable performance of the introduced parameter estimation technique.