CN108234033B

CN108234033B - Allan variance-based individual identification method for radiation source

Info

Publication number: CN108234033B
Application number: CN201810019295.5A
Authority: CN
Inventors: 朱胜利; 马俊虎; 甘露; 廖红舒
Original assignee: University of Electronic Science and Technology of China
Current assignee: University of Electronic Science and Technology of China
Priority date: 2018-01-09
Filing date: 2018-01-09
Publication date: 2020-12-29
Anticipated expiration: 2038-01-09
Also published as: CN108234033A

Abstract

The invention belongs to the technical field of communication, and particularly relates to a radiation source individual identification method based on Allan variance. The individual radiation source identification method based on the Allan variance utilizes different nonlinear characteristics brought by nonlinear devices in actual radiation sources. In fact, in the present invention, different radiation sources are treated as different nonlinear systems, and based on this principle, the Allan variance is used to extract the features of these nonlinear systems, thereby identifying the different systems. Because the Allan standard deviations calculated at different correlation times have different meanings, the Allan standard deviations at three correlation times are used for forming the feature vector, and the identification effectiveness is enhanced.

Description

Allan variance-based individual identification method for radiation source

Technical Field

The invention belongs to the technical field of communication, and particularly relates to a radiation source individual identification method based on Allan variance.

Background

Since the last 60 th century when the SEI (Specific identity) technology was proposed, the SEI technology has attracted much attention and has become a research hotspot in the neighborhood of cognitive radio, network security, electronic support systems, and the like. The communication radiation source individual identification technology utilizes individual characteristics extracted from communication signals to realize individual identification of different radiation sources. The technology is often used for military and civil spectrum management, and has important significance for enhancing the security of a wireless network. The existing individual identification methods of radiation sources can be mainly divided into three types: signal transient characteristic-based, signal steady-state characteristic-based and non-linear-based identification methods.

The signal transient-based features are mainly extracted from non-modulated transient signals radiated when the wireless communication device is powered on, and contain abundant device characteristics. Although the transient signal features are obvious and easy to identify, the transient signal generally has a short duration, is difficult to acquire and has high randomness, so that the method based on the transient signal features cannot be widely applied.

Compared with a method based on signal transient characteristics, the method based on signal steady-state characteristics is stable, characteristics are easy to obtain, and the practicability is high. The current methods based on steady-state features are: modulation characteristic-based, spectral characteristic-based, preamble-based, and wavelet transform-based methods, and the like. Although the methods based on steady-state features are more practical than the methods based on transient features, the methods themselves have many disadvantages. For example, methods based on modulation characteristics are not suitable for signals of all modulation types, and the demodulation process has a great influence on the performance of the methods; some methods based on the spectral characteristics of the signal often have certain requirements on the distance between the transmitting and receiving devices; the pilot signal based approach is ineffective for burst signals that do not contain a pilot signal; the performance of wavelet transform based identification methods is often affected by wavelet basis selection.

Power amplifiers are present in all radiation source systems, the non-linear behavior of which has a considerable influence on the signal. Due to the influence of power amplifier nonlinearity, signals emitted by the radiation source also have nonlinear characteristics and can be used as identification characteristics. The method for realizing individual identification by using the phase space analysis method and the displacement Entropy (PE) considers the radiation source as different nonlinear systems, and realizes the individual identification by researching the nonlinear characteristics of the systems. However, the two methods have better effect when the data volume is larger, and the noise has larger influence on the two methods.

Disclosure of Invention

The invention aims to solve the problems and provides a radiation source individual identification method based on Allan variance, which is essentially an analysis method for extracting nonlinear features. Compared with the method based on transient or steady-state characteristics, which needs to determine transient and steady-state signals in advance, the method of the invention does not need the process, and the fingerprint characteristics are extracted from the two signals, so that the method has stronger applicability. In addition, the time complexity of the invention is low, and the invention still has high identification performance under the conditions of small data size and low signal-to-noise ratio.

The technical scheme adopted by the invention is as follows:

a radiation source individual identification method based on Allan variance is characterized in that a communication signal receiving device is adopted to receive communication radiation source signals, the communication radiation source signals are converted into intermediate frequency signals, and burst signal acquisition is carried out on each communication signal receiving device, and the radiation source individual identification method comprises the following steps:

s1: setting reception of a communication signal receiving meansThe burst signal is

Normalizing it:

where N is the signal length.

S2: determining the number of segments based on a given correlation time tau

Divide the data equally into L segments, wherein

Represents rounding down; and averaging the segmented data:

wherein t is a variable in the calculation process;

s3: carrying out difference on the mean value sequence;

s4: calculating the variance of the difference mean sequence under the current segmentation number:

wherein the content of the first and second substances,<·>it is indicated that the average value is taken,

represents the difference mean sequence obtained at S3,

the variance of the sequence of difference means is represented,

indicating that symbols are estimated by the quantity to the right of the symbolThe amount to the left of the number. I.e. used during actual calculation

To estimate

S5: repeating the steps S1 to S4, obtaining Allan variances under P correlation times, and forming a feature vector by using the Allan standard deviation sigma:

V＝[σ₁,σ₂,…,σ_P].

s6: for each burst, performing S1 to S5;

s7: and carrying out individual identification on the radiation source by utilizing a K nearest neighbor classification algorithm.

The method has the advantages that the method for identifying the individual radiation sources based on the Allan variance utilizes different nonlinear characteristics brought by nonlinear devices in actual radiation sources. In fact, in the present invention, different radiation sources are treated as different nonlinear systems, and based on this principle, the Allan variance is used to extract the features of these nonlinear systems, thereby identifying the different systems. Because the Allan standard deviations calculated at different correlation times have different meanings, the Allan standard deviations at three correlation times are used for forming the feature vector, and the identification effectiveness is enhanced.

Drawings

FIG. 1 is a flow chart of an implementation of the present invention;

FIG. 2 is a feature vector corresponding to different radiation sources in embodiment 1 of the present invention;

fig. 3 is a graph of the average correct recognition rate of the radiation source as a function of the signal to noise ratio in example 1 of the present invention.

Detailed Description

The invention is described in detail below with reference to the drawings and examples so that those skilled in the art can better understand the invention.

Examples

The object of this embodiment is to perform fingerprint feature extraction for different radiation sourcesAnd the individual identification is realized, and the identification success rate is simulated. The data in this embodiment comes from four return channel satellite terminals, which are recorded as: e_iAnd i is 1,2,3, 4. The data is received by an agilent E3238S signal detection system. The transmission signal adopts QPSK modulation mode, and the transmission rate is 2 MBoud. After receiving, the radio frequency signal is converted to 70M intermediate frequency. The sampling rate is 400MHz and the signal-to-noise ratio of the received signal is about 35 dB. To control variables, data were collected from laboratory conditions. 200 bursts of data are collected for each terminal for validating the proposed algorithm. It should be noted that the four terminals are produced from the same manufacturer, and E1 and E2 belong to the same batch of products with different serial numbers, and E3 and E4 also belong to the same batch of products with different serial numbers.

In the simulation, 20000 data points are used per burst. For each burst signal, the corresponding Allan standard deviation is found at

correlation time τ

4,6,10 and a feature vector is constructed, as shown in fig. 1. It can be seen that these features have very significant variability. It is stated that the eigenvectors derived from the Allan variance can be used to identify these four radiation sources. In the classification process, during each experiment, one burst signal is used as test data, the rest (199 bursts) are used as training data for identification, another burst data is used as test data in the next experiment, the process is repeated until all the burst data are tested, and the final identification result is obtained, as shown in table 1. As can be seen from Table 1, the recognition accuracy of the method provided by the invention to four radiation sources is over 99.5%.

Table 1 classification and identification results of four radiation sources in example 1

In order to test the performance of the algorithm proposed by the present invention at low snr, the snr is varied between 0 and 20dB by directly adding gaussian white noise of different intensity to the received signal. For each signal-to-noise ratio, 100 monte carlo tests were performed, and the average correct recognition rate of the four radiation sources was obtained, with the results shown in fig. 3. It can be seen that even at 0dB, the average recognition rate of the method provided by the invention is still above 99%.

Claims

1. A radiation source individual identification method based on Allan variance is characterized in that a communication signal receiving device is adopted to receive communication radiation source signals, the communication radiation source signals are converted into intermediate frequency signals, and burst signal acquisition is carried out on each communication signal receiving device, and the radiation source individual identification method comprises the following steps:

s1: setting a burst signal received by a communication signal receiving apparatus to

Normalizing it:

wherein N is the signal length;

s2: determining the number of segments on the basis of a given correlation time τ, i.e. the amount of data in the correlation time τ being n

Divide the data equally into L segments, wherein

Represents rounding down; and averaging the segmented data:

wherein t is a variable in the calculation process;

s3: carrying out difference on the mean value sequence;

represents the difference mean sequence obtained at S3,

the variance of the sequence of difference means is represented,

indicating that the quantity to the left of the symbol is estimated by the quantity to the right of the symbol, i.e. used in the actual calculation

To estimate

S5: repeating the steps S1 to S4, obtaining Allan variance under P correlation times, wherein the value of P is more than half of the number of radiation sources, and forming a feature vector by using the Allan standard deviation sigma:

V＝[σ₁,σ₂,…,σ_P].

s6: for each burst, performing S1 to S5;