A Spectrum Sensing Method Based on Small-Samples Training Neural Network
A neural network and spectrum sensing technology, applied in the field of wireless communication networks, can solve problems such as calculation of a large number of samples, achieve the effects of improving detection performance, simplifying the debugging process, and reducing the probability of leak detection
Active Publication Date: 2022-02-08
NAT UNIV OF DEFENSE TECH
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] The purpose of the present invention is to provide a spectrum sensing method based on small-sample training neural network to solve the technical problems in the prior art that require a large number of samples and a large number of calculations in the neural network-based spectrum sensing method
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0039] In order to further understand the contents of the present invention, the following examples will be described. The present invention is based on the following commonly used and practical assumptions: the fading a, noise ε, and time delay τ in the sensing environment h where the detector is located are distributed according to a certain law within a certain range, that is, h~ρ(a, ε, τ ); define the following parameters: the number of antennas at the sensing end M, the number of samples of the sensing time signal N, so the sampled received signal can be expressed as X M×N; Considering the relevant spectral state of the received signal, the training samples of length K in the perception environment h can be expressed as the y (k) ∈{0,1},y (k) =1 means the spectrum is occupied, otherwise, y (k) =0; θ represents the parameters of the neural network in the neural network detector, θ 0 Denotes the initial parameters of the detector, Indicates the fine-tuning value of t...
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
Login to View More Abstract
The present invention provides a spectrum sensing method based on small sample training neural network, which includes the following steps: pre-training the neural network detector, pre-deploying the detector in the actual operating environment of the system, and iteratively calculating with a small amount of data and a small amount of gradient Find the fine-tuning value of the detector in different perception environments and its test loss value, then calculate the gradient of the initial parameter corresponding to the loss value and update the initial parameter gradient, so that the initial parameter of the detector can quickly adapt to the change of the environment; based on The initial parameters of the pre-trained neural network detector are adjusted online; the sampled received signal is input into the adjusted neural network detector to predict the probability of whether the spectrum is occupied at this time. The spectrum sensing method can achieve detection performance similar to that of the existing neural network detector based on a large number of samples and gradient iterations, and can effectively reduce the amount of calculation and the number of samples required for detector adjustment.
Description
technical field [0001] The invention belongs to the technical field of wireless communication networks, and relates to a spectrum sensing method based on small sample training neural network. Background technique [0002] Spectrum sensing, that is, judging whether the current spectrum is occupied by receiving signals, has received more and more attention. Because of its wide range of applications, spectrum sensing technology is widely used in many communication scenarios. For example, cognitive radio technology and anti-jamming communication technology. [0003] Research on spectrum sensing technology generally assumes certain prior information, such as known communication characteristics or channel characteristics of the sending end. However, such prior information is often difficult to obtain in practical applications, which leads to poor practicability of this type of method. With the rapid development of machine learning technology, existing research gradually focuses...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): H04B17/382G06N3/08G06N3/04
CPCH04B17/382G06N3/084G06N3/045
Inventor 赵海涛魏急波高士顺熊俊张晓瀛周力辜方林唐麒
Owner NAT UNIV OF DEFENSE TECH