Method for identifying signal component characteristics based on predictive analysis

Abstract

The invention provides a method for identifying signal component characteristics based on predictive analysis. The method includes the steps: estimating the prediction sequence of an analytical signal according to incremental prediction orders by the aid of a linear prediction model and a nonlinear prediction model; recording prediction errors of the linear prediction model and the nonlinear prediction model under different prediction orders; respectively obtaining relation curves of 'prediction orders and errors' of linear prediction and nonlinear prediction; extracting change trends of the relation curves of 'the prediction orders and the errors' of linear prediction and nonlinear prediction; and identifying the signal characteristics to judge whether the signal is a determinate linear signal or a determinate nonlinear signal or is in a random process. The method for identifying the signal component characteristics is simple and convenient in calculation process, so that the signal characteristics are identified accurately and efficiently.

Description

technical field [0001] The invention relates to the field of signal processing, in particular to an identification method of signal component characteristics based on predictability analysis. Background technique [0002] With the rapid development of nonlinear dynamics, people have a deeper understanding of signal properties. The research of chaos theory in nonlinear research shows that some superficially complex random processes are often produced by some simple regular physical laws. For such signals with regular nonlinear structure, traditional linear signal analysis methods are difficult to directly and effectively deal with them, and specific nonlinear processing methods must be used to observe their essential regularity through complex appearances. [0003] Before effectively processing the signal, it is necessary to clearly identify the characteristics of the signal and judge whether the signal to be analyzed and processed is a deterministic linear signal, a determi...

AI Technical Summary

Problems solved by technology

For the identification method of signal characteristics and parameters of deterministic linear signals, power spectrum estimation is generally used, but such methods cannot distinguish deterministic nonlinear signals from random processes

For the identification of deterministic nonlinear signals, the measurement of geometric regularity in its reconstructed phase space is generally used, but such methods need to accurately obtain the parameter values ​​of embedding dimension and delay time when reconstructing the phase space. It is difficult to estimate accurately in practical applications; at the same time, such methods require as long as possible data length and high calculation costs, which makes it difficult for such methods to guarantee real-time performance in practical applications

Images