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

A signal detection method based on deep learning

A signal detection and deep learning technology, applied in the field of signal detection based on deep learning, can solve problems such as high false alarm probability and missed detection probability, limited application, etc., to enhance detection performance, strong nonlinear mapping fitting ability, improve The effect of detection performance

Active Publication Date: 2020-10-13
ZHEJIANG UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the high probability of false alarm and missed detection in the existing neural network, its application is limited.

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
  • A signal detection method based on deep learning
  • A signal detection method based on deep learning
  • A signal detection method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] The specific implementation manner of the present invention will be described in detail below in conjunction with the accompanying drawings and examples.

[0033] figure 1 A single-node detection neural network structure diagram is given, which is named "DetectNet" below. The DetectNet consists of two convolutional layers (Conv1 and Conv2) with 60 filters of size 10, a 128 neuron The fully connected layer (FC1), an input combination layer (Concatenated layer), two LSTM layers of 128 neurons (LSTM1 and LSTM2), a fully connected layer of 128 neurons (FC2) and a fully connected layer of 2 neurons (FC3) composition. Except for the last fully connected layer that uses SOFTMAX as the activation function, the rest of the network layers use ReLU. The input of the neural network is a time-domain sampled signal vector, and the output is two types of probability vectors of the existence and non-existence of the main signal, and the larger one is the single-node detection result....

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 belongs to the field of wireless communication, and relates to a signal detection method based on deep learning: after energy normalization is performed on the time-domain signal sampling vector, it is input into a single-node detection neural network, and two types of probability vectors of the existence or non-existence of the main signal are output , take the larger one, and get the single-node detection result as the signal detection result; or, after energy normalization is performed on the time-domain signal sampling vectors of different nodes, input the single-node detection neural network, and output whether the main signal exists or not The two types of probability vectors are used as the input of the multi-node cooperative detection neural network, and the multi-node detection result is obtained as the signal detection result; wherein, the single-node detection neural network is trained by a two-stage training method. The invention effectively excavates and utilizes the structural information of the modulated signal and the soft information of a single node, improves the detection performance of a single node and the cooperative detection performance of multiple nodes, and improves the above defects of the existing traditional signal detection method.

Description

technical field [0001] The invention belongs to the field of wireless communication, and in particular relates to a signal detection method based on deep learning. Background technique [0002] Cognitive radio is one of the important technologies to solve the problem of spectrum scarcity, and has attracted great attention from both industry and academia. One of the characteristics of cognitive radio is that the secondary user determines whether there is an available spectrum hole by detecting the signal of the primary transmitter. The detection performance directly affects the spectral efficiency of the secondary user and whether the primary user can communicate normally. Therefore, how to achieve reliable detection with the lowest possible false alarm probability and missed detection probability at the lowest possible signal-to-noise ratio is crucial for cognitive radio development is crucial. [0003] Energy detection is a common detection technology, but in actual syste...

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
Patent Type & Authority Patents(China)
IPC IPC(8): H04B17/382G06K9/62G06N3/04
CPCH04B17/382G06N3/045G06F18/214
Inventor 钟财军高佳宝易雪梅陈晓明张朝阳
Owner ZHEJIANG 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