Modeling and Parameter Estimation for Nonstationary Signals and Its Spectrum Representation

2014 Fiscal Year Final Research Report

• Project/Area Number: 24500342
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Statistical science
• Research Institution: Wakayama University
• Principal Investigator: IJIMA Hiroshi 和歌山大学, 教育学部, 教授 (90397604)
• Project Period (FY): 2012-04-01 – 2015-03-31
• Keywords: 非定常信号 / 時変モデル / 推定

Outline of Final Research Achievements

A great deal of interest has been paid to develop models of random signals such as auto-regression(AR) model, moving-average(MA)model, ARMA model and so on. In addition, the estimation methods for model parameter have been proposed. Such model can be applied to detection, estimation and prediction of signals. Conventionally, many models represent the stationary random signal. However, many signals in nature should be represented as nonstationary models. The purpose of this research is to develop mathematical models of signals having nonstationary properties and to propose estimation metods of time-varying parameters included in the models.
Results of the reserch are asfollows : i) As mathematical model of signals, new model with time-varying parameters has been proposed. ii) Estimation methods of nonstationary signal model have been proposed. iii) Proposed methods have been applied to the spectrum analysis, speech signal processing and seismic signals．

Free Research Field

信号処理