Fingerprint identification technology of communication radiation source

SummaryCommunicationradiation SourceIndividual identification is a recentCommunicationImportant research topics in the field of confrontation, unlike traditionalCommunicationResearch on modulation pattern recognition in signal reconnaissance, communicationradiation SourceIdentify major studies that embody the same kindradiation SourceAnalysis and extraction of the signal fingerprint of individual differenceTechnology。 In this paper, the analysis and extraction of signal fingerprint is studied.TechnologyDifficulties and solutions. The individual fine features of the steady-state communication signal, such as the deviation of the communication emission signal from the carrier frequency and modulation parameter characteristics, and the difference of the signal spurious output components, and consider the above characteristics as a reflection of the communication station individualTechnologyCharacteristics of the signal fingerprint, a communication emitter identification scheme based on evidence theory is presented, which is an important basis for the realization of individual identification in military communication warfare plan.
Keywords communication radiation source; pattern recognition; frequency stability; individual identification

In modern military electronic information equipment, the radio-based communication equipment is responsible for the important military carrier to provide communication, deployment and implementation of military strikes or monitoring, electronic interference important task. At present, in the field of electronic warfare, the method of "fingerprint" of radiation source has appeared, and by means of the characteristic measurement of the receiving signal, the individual of the radiation source that produces the signal is defined as "the ability to relate the unique electromagnetic characteristics of the radiation source to the individual of the radiation source". The communication station's ability to correlate is called the communication Station individual recognition. If we can effectively extract the individual fingerprint characteristics of communication radio station, separate each radio area in the complex battlefield electromagnetic environment, realize the analysis and identification of the radio station, and analyze which stations are the important communication stations of the enemy, then further analyze the nature and properties of the radio, determine the composition of the communication network, the threat level and the interfering object, etc. It can provide important basis for the targeted electronic interception, disturbance and military attack. It can be seen that it is an important basis for military communication combat plan to determine the characteristics of effective communication radio signal fingerprint and realize individual recognition.

1 fingerprint and signal fingerprint mechanism analysis
1. 1 communication radiation Source fingerprint
The subtle feature of the individual signal of a communication emitter is also known as the "fingerprint" of the communication signal, which can be used to identify the identity of the communication device in which the signal is sent. By means of communication signal processing technology, the hardware characteristic information of the communication radiation source which is carried on the communication signal is found to be stable. It can be considered that the signal fingerprint is mainly manifested as a regular change trend of the same communication equipment repeated in all the signals transmitted by the same communications device, and the information of the repeated change rule in the communication signal has the technical characteristic that reflects the individual characteristics of the signal, and can be used as "fingerprint" to identify the individual characteristics of the communication device transmitting the signal.
The characteristic parameters of the communication signal can be used as fingerprint characteristics: (1) The accuracy difference of the carrier frequency of the communication signal. (2) Individual differences in modulation parameters of communication signals. (3) Stray output difference of radio station.
The characteristic parameters that can be extracted from the radiation source can be divided into the following categories:
(1) Technical characteristics. modulation mode, carrier frequency accuracy, stability of frequencies.
(2) Internal characteristics. Information transmission rate of digital communication. (3) Frequency domain characteristics. Signal bandwidth, FM parameter modulation distortion degree. (4) Time-varying characteristics. Instantaneous envelope, frequency, phase.
1. 2 signal fingerprint Mechanism analysis
The individual identification of communication radiation source is to obtain the working parameters and characteristic parameters of the communication equipment in the signal, and then obtain the information of the system, use and model of the communication equipment by using these parameters, and then grasp its working state, and understand its tactical application characteristics, activity law and the process of combat capability.
This paper studies the individual identification technology of multiple radio stations with same model, same batch and working in the same modulation mode and frequency band. The individual identification of communication radiation source consists of 3 processes: pretreatment, feature extraction and classification recognition.

This article refers to address: http://www.eepw.com.cn/article/155187.htm

The essence of fingerprint recognition of communication signals is pattern recognition, and the process of recognition is composed of design and implementation. Design refers to a certain number of samples for the classifier design, the realization refers to the design of the classifier to treat the identified samples for classification decision-making.

2 method for fingerprint identification of communication signals
2. 1 characteristic parameter domain discriminant method
Using the characteristic parameter range discriminant method, firstly, we need to extract one or more parameter values which can reflect the individual characteristics of the communication signals, and use pattern recognition method to identify the individual parameters according to the domain ranges of each characteristic parameter, and the main process of analysis and processing is in time and frequency domain.
The characteristic parameter range discriminant method can also be used to identify the distance of the feature class, and the difficulty lies in how to extract and select the proper and necessary feature vectors from the original measurement data, focusing on the feature extraction and selection algorithm.
2. 2 signal modeled matching recognition method
The basic idea of signal modeled matching recognition method is as follows:
(1) The same communication signal form of n signal extraction parameters, the characteristics of the extracted parameters of M, n a known signal of the same characteristic parameter values are different, so that each signal has a m characteristic parameter value and other signals are different;
(2) The modeled of n signals is stored in the database, and a modeled library containing n known signals is established.
(3) Extracting the m characteristic parameters of the signal to be identified, and comparing with the modeled in the modeled library, if the extracted characteristic parameter matches with a modeled in the modeled library, the individual identification can be completed.

2. 3CommunicationSignalFingerprint IdentificationClassifier
In statistical pattern recognition, the basic task of classifier is to classify the input of a given characteristic vector into an appropriate feature class according to a certain criterion, that is, to realize the transformation from the feature space to the decision space, thus accomplishing the classification task of the feature class.
The commonly used classifiers are parametric and non-parametric classification algorithms based on statistical decision theory, such as linear and generalized linear decision functions, K-Nearest neighbor algorithm (K-NN), and two-tuple classification trees. If the probability density function of the feature to be identified is known or can be accurately estimated by the sample, then these classification algorithms can obtain the best recognition performance, but theCommunicationSignalFingerprint IdentificationProblems, these conditions are difficult to meet, traditional classifier difficult to obtain satisfactory performance of individual recognition. The main disadvantage is the low recognition rate and poor robustness.
The development of the theory of decision-making overcomes the shortcomings of the traditional classifier, puts forward the more advanced theory of uncertainty inference, and on this basis, the classifier recognition performance has been improved remarkably. Among them, the neural network classifier is a kind of advanced adaptive, non-parametric and nonlinear classifier. It opens up a new way for signal pattern recognition. Neural network is a kind of nonlinear signal processing system characterized by self-organization, adaptive and large-scale distributed parallel computing, which has powerful pattern recognition classification and functional approximation ability, and has good fault tolerance.
Combinatorial classifier is a new topic proposed in recent years, refers to the output of different classifiers through a combination of the second decision, because the fusion of multiple classifiers decision, so can be better classification performance, and each of the classifiers are not required to be optimal, forCommunicationThe identification of signals provides a new approach.

31 individual identification schemes based on evidence theory
In this paper, the research of communication signal fingerprint mainly adopts the method of "Mechanism research-feature analysis-feature extraction-classification experiment", in which the signal-to-noise ratio of real communication signals is large and the recognition rate decreases, which directly affects the classification performance of classifier. Based on D-s evidence combination principle, the classifier design can be extracted from different categories of fingerprint characteristic parameter set, and a combination classifier based on evidence theory is given.radiation SourceThe design method of individual identification, which is shown in structure 2. This article refers to address: http://www.eepw.com.cn/article/155187.htm

For the received communication signal, first through the communication signal preprocessing, then carries on the fingerprint feature extraction, through establishes the communication signal the carrier frequency, the modulation parameter and the spurious component characteristic is the steady state signal fingerprint, may use the time domain, the frequency domain analysis method and the modern time frequency domain and the high order spectral method to carry on the characteristic analysis extraction. In the classifier design, the first level classifier can choose the characteristic parameter range discriminant method, the signal modeled matching recognition method to realize the first level classification, the second stage adopts the parallel combination classifier, so the enigmatic grading problem of the high dimensional feature space is transformed into the problem of dividing the different low dimensional space, Second judgment is given to the output of the first class classifier. In this way, the decision of merging multiple classifiers can achieve better classification performance.
The key to the fusion of the independent information is the determination of the basic probability distribution function (BPA) in the D-s evidence combination classifier. It is believed that if each classifier uses different feature sets or training sets, it can be considered that the results of different classifiers are independent. Therefore, the different individual characteristics of radiation sources are used to train different member classifiers, and the combination of D-s evidence theory can be selected according to the need of the appropriate evidence combination rules to determine the type of communication radiation source .

4 concluding remarks
With the rapid development of communication technology , communication system and modulation style more complex and diversified, the number of communication radio stations is increasing, the signal characteristics of conventional electronic reconnaissance, such as carrier frequency, modulation style and modulation parameters, are difficult to meet the demands of modern battlefield in the complex and changeable signal environment. The identification of communication emitter mainly studies the analysis and extraction technology of the signal fingerprint which manifests the individual difference between the similar radiation sources. This paper analyzes the fingerprint selection problem of communication radiation source, analyzes the mechanism of communication radiation source of signal fingerprint, summarizes the commonly used methods of fingerprint recognition of communication signals, and proposes a method of fingerprint identification based on evidence theory for communication radiation source.

Fingerprint identification technology of communication radiation source