Unlock instant, AI-driven research and patent intelligence for your innovation.

Hammerstein system identification method based on quantized kernel least mean square errors

A minimum mean square error and system identification technology, applied in the field of signal processing, can solve problems that limit the practicability and generalization ability of algorithms

Inactive Publication Date: 2017-10-20
XI AN JIAOTONG UNIV
View PDF0 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0023] These algorithms can only be used in Hammerstein systems with certain specific structures, but in actual situations, the nonlinear part does not only have polynomial forms, which greatly limits the practicability and generalization ability of the algorithm

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Hammerstein system identification method based on quantized kernel least mean square errors
  • Hammerstein system identification method based on quantized kernel least mean square errors
  • Hammerstein system identification method based on quantized kernel least mean square errors

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0055] The present invention will be further described below in conjunction with the accompanying drawings.

[0056] The present invention is based on the Hammerstein system identification method of quantized kernel minimum mean square error, referred to as quantized kernel adaptive Hammerstein filter (QKAHF), now specifically introduced as follows:

[0057] Quantized Kernel Least Mean Square Error Algorithm (QKLMS)

[0058] When learning a continuous nonlinear input-output mapping d=f(u), where u is an m-dimensional input vector, U is A compact input domain in , and d is the output signal. When input-output pairs {u(i),d(i),i=1,2,...} are available, the learning problem can be understood as finding an estimate of the mapping f based on the training data The kernel adaptive filtering algorithm is a kernel-based sequence estimator, and the estimate of f at step i is f i , according to the last estimate f i-1 and the current sample {u(i),d(i)} to update.

[0059] The M...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a Hammerstein system identification method based on quantized kernel least mean square errors (QKLMS). The method uses a QKLMS method to fit the non-linear part of Hammerstein. The method has strong fitting ability and can fit any non-linear mapping. When the quantization parameters are adjusted, a QKAHF algorithm can obtain different performance. When the quantization parameters become larger, the steady-state convergence value of the mean square errors becomes larger, but the network structure becomes smaller. When the quantization parameters become smaller, the steady-state convergence value of the mean square errors becomes smaller, but the network structure becomes larger, so that the method can achieve better fitting performance and fast convergence, and is easier to be promoted and used in practical applications.

Description

【Technical field】 [0001] The invention belongs to the field of signal processing and relates to a Hammerstein system identification method based on the minimum mean square error of a quantization kernel. 【Background technique】 [0002] In recent years, adaptive filtering has developed rapidly as an optimal filtering method. Adaptive filtering is an optimal filtering method developed on the basis of linear filtering such as Wiener filtering and Kalman filtering, because it has stronger adaptability and better filtering performance. Therefore, it has been widely used in engineering practice, especially in information processing technology. [0003] Kernel Adaptive Filtering (KAF) algorithms are a class of online methods that map raw data into a high-dimensional Regenerated Kernel Hilbert Space (RKHS) where traditional linear adaptive algorithms are implemented. Known kernel adaptive filtering algorithms include kernel least mean square error algorithm (KLMS), kernel affine p...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): H03H21/00
CPCH03H21/0043
Inventor 陈霸东董继尧郑南宁
Owner XI AN JIAOTONG UNIV