Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method for multi-channel blind deconvolution on cascaded neural network

A blind deconvolution and neural network technology, applied in the field of signal processing, can solve problems such as poor estimation of strong nonlinear functions

Inactive Publication Date: 2012-07-04
GUANGDONG BAIYUN UNIV
View PDF0 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

A problem with the MMI algorithm is that the Gram-Charliar expansion is used to estimate the marginal entropy, and the Gram-Charliar expansion is poor for strongly nonlinear functions.

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
  • Method for multi-channel blind deconvolution on cascaded neural network
  • Method for multi-channel blind deconvolution on cascaded neural network
  • Method for multi-channel blind deconvolution on cascaded neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0039] In this embodiment, a method for multi-channel blind deconvolution cascaded neural network, such as figure 1 described, including the following steps:

[0040] (1) A modular neural network is composed of an equalization sub-network and a compression sub-network, the equalization sub-network is a processing unit for extracting a signal source, and the compression sub-network is a processing unit for eliminating other redundant information after extracting a signal from a mixed signal ;like figure 2 As shown, the module neural network of this embodiment is composed of an equalization sub-network and a compression sub-network. On the basis of this module neural network, the method of cascading neural networks in this embodiment can be used to obtain from unknown interleaved signals Extract multiple source signals.

[0041] (2) Balance sub-network synapse {w 1j，p} is updated by the constant model algorithm, and its update formula is as follows:

[0042] ...

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 provides a method for multi-channel blind deconvolution on a cascaded neural network. The method comprises the following steps: (1) forming a module neural network by a balanced sub-network and a compressed sub-network; (2) updating a nerve synapse {wlj,p} of the balanced sub-network by using a constant model algorithm; (3) constituting Hebbian and inverse Hebbian learning rules; and (4) after compression, inputting to the balanced sub-network of a next module network. The method disclosed by the invention is an expansion of a method which is novel, simple and individually effective, and can be used for effectively extracting a plurality of source signals online from unknown and staggered mixed signals, namely each module of the cascaded neural network is composed of the balanced sub-network and the compressed sub-network. The method disclosed by the invention can be applicable to any blind equalization algorithm (an extension of signal channel equalization) and further can be applied to a condition that the quantity of the source signals is unknown in advance. The method disclosed by the invention is easy to realize and can be widely applied to aspects of wireless communication, array processing, biomedical signal processing and the like.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to a method for multi-channel blind deconvolution cascade neural network. Background technique [0002] In recent years, blind signal processing technology has been developed rapidly. Blind signal processing techniques can be roughly divided into two categories: blind source separation and blind deconvolution. Blind signal processing technology has been widely used in signal processing fields such as speech signal processing, biomedical signal processing, communication signal processing, and image signal processing. [0003] Blind deconvolution is a basic problem in many researches and applications such as image processing, speech signal processing, communication, system identification and acoustics, and has important theoretical and application value. [0004] At present, various blind deconvolution algorithms for MIMO (Multiple-Input-Multiple-Output) systems are common...

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): H04L1/06H04L25/03
Inventor 刘建成高俊文张文梅徐献灵杨新盛徐存东
Owner GUANGDONG BAIYUN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com