Wireless network security authentication method and system based on burst signal fingerprint identification

Abstract

The invention belongs to the technical field of network security authentication, and particularly provides a wireless network security authentication method and system based on burst signal fingerprint identification, and the method comprises the steps: employing a maximum likelihood detection algorithm, and carrying out the detection of a transient process of a radiation source burst signal; calculating the instantaneous amplitude of the signal transient process; extracting a radiation source fingerprint feature implied in the instantaneous amplitude of the signal transient process by adopting a least square estimation algorithm; in the training stage, fingerprints of legal radiation sources are trained by using a support vector description algorithm, and a wireless communication network security authentication model is established; and an authentication stage: identifying the radiation source fingerprint by using the trained radiation source identification model to complete the authentication of the equipment identity. According to the method, the transient process of a burst signal is accurately detected, mathematical modeling is carried out on the transient distortion process of a transmitter, and the estimation and identification of the fingerprint characteristics of the radiation source in the transient process are realized, so that the authentication of wireless communication network access equipment is completed.

Description

Technical Field

[0001] The present disclosure relates to the technical field of network security authentication, and in particular to a wireless network security authentication method based on burst signal fingerprint recognition. Background Art

[0002] Wireless communication network security authentication refers to a network security authentication method that uses technical means to promptly detect illegal user attacks, and promptly adopts various measures to prevent attackers from achieving their attack objectives, thereby minimizing their own losses. Traditional wireless communication network security authentication is mainly achieved through access control and encryption. Due to the poor privacy of wireless security protocols and the easy forgery of security credentials, illegal devices can invade the network by simulating legitimate device information or stealing key login, which brings great security risks to wireless communications. How to ensure the security of wireless communication networks and enable the safe and reliable transmission of data between the two communicating parties has become an urgent problem to be solved.

[0003] In recent years, the radiation source fingerprint recognition technology of wireless communication has received widespread attention at home and abroad. Radiation source fingerprint recognition is a technology that uses hardware-related radio frequency fingerprint features to identify radiation source targets. Compared with other individual recognition technologies, its advantage is that it does not rely on information content interpretation, but directly associates wireless signals with transmitting devices. By adding the radiation source fingerprint feature matching and recognition of both communicating parties to the authentication mechanism, the user's identity authentication is realized from the physical layer, further improving the wireless communication network security authentication capability.

[0004] The traditional radiation source fingerprint recognition method is mainly based on the radiation source mechanism model. By establishing a mathematical model for the non-ideal characteristics of some key analog components in the radiation source, the signal is analyzed in different domains by signal processing, and the direct measurement features or dimensionality reduction transformation features of the fingerprint are extracted to realize the individual identification of the radiation source: Reference [Y.Niu, W.Xiong, Z.Li, B.Dong and X.Fu,"On the Identification Accuracy of the I / QImbalance-Based Specific Emitter Identification,"in IEEE Access, vol.11, pp.75462-75473, 2023, doi:10.1109 / ACCESS.2023.3296688.] extracts the IQ imbalance parameters of the radiation source signal as the fingerprint feature of the transmitter, which is mainly for digital phase modulation signals, and has poor effect on frequency modulation or amplitude modulation signals.

[0005] Reference [Y.Yao, L.Yu and Y.Chen. Specific Emitter Identification Based on Square Integral Bispectrum Features[C].2020IEEE 20th International Conferenceon Communication Technology(ICCT), Nanning, China, 2020:1311-1314.] By calculating the high-order cumulants of the signal and its bispectrum, the bispectrum is reduced in dimension and the features are re-extracted as the fingerprint features of the radiation source.

[0006] Compared with traditional methods, deep neural networks can retrieve abstract features through multiple hidden layers with nonlinear activation functions, which is conducive to extracting deep and subtle features of radiation source signals. Reference [Ren Dongfang, Zhang Tao, Han Jie, et al. Specific radiation source identification method based on ITD and texture analysis [J]. Journal of Communications, 2017, 38 (12): 160-168.], through short-time Fourier transform, the one-dimensional signal is converted into a two-dimensional image, and then the convolutional network is used as a feature extractor to obtain deep features and PCA technology is used to reduce the dimension of the features to achieve individual radiation source identification. Reference [Jia Xin, Jiang Lei, Guo Jingjing, Qi Zisen. Communication radiation source individual identification method based on deep clustering [J]. Journal of Air Force Engineering University, 2024, 25 (1): 115-122] The original I / Q data is learned by using the powerful feature extraction and data reconstruction capabilities of the autoencoder network. By jointly optimizing the representation learning process and the feature clustering process, the degree of fit between the two is improved, and the identification of 5 types of ZigBee devices is achieved.

[0007] With the development of radiation source fingerprint recognition technology, people continue to explore and expand its application areas. In the past two years, patents related to radio frequency fingerprint security authentication have been released:

[0008] The radio frequency fingerprint recognition method and system for distributed Internet of Things devices with patent publication number CN 118606827A deploys multiple radio frequency edge nodes, converts radio frequency signals into spectrum graphs to extract radio frequency fingerprints, and uses a federated learning method to coordinate the edge nodes to train the radio frequency fingerprint recognition model, thereby realizing the characteristics of collaborative training between nodes.

[0009] The radio frequency fingerprint efficient and secure authentication technology for wireless multi-heterogeneous terminals with patent publication number CN 118612745A adopts general software defined radio technology and GNU Radio platform, and uses deep learning algorithms to transform the complex distribution of signal samples into a simple and easy-to-process basis distribution based on normalized flow, thereby achieving device security authentication by allocating a higher probability of authorized device samples.

[0010] The patent publication number is CN 118354322 A, which is a radio frequency fingerprint authentication and recognition method based on multi-domain feature fusion and deep learning. It uses signal carrier frequency offset as an important feature representation, time domain feature representation and time-frequency domain feature representation as auxiliary multi-domain feature representation, and uses convolutional neural networks to fuse features, thereby improving recognition performance and stability.

[0011] In the field of radiation source fingerprint recognition, according to the different signal segments used to extract fingerprint features, the radiation source fingerprint extraction technology is divided into fingerprint extraction technology based on signal transient processes and fingerprint extraction technology based on signal steady-state processes. The traditional technical solutions recorded in the above papers and patents mainly adopt fingerprint extraction technology based on signal steady-state processes. Although the steady-state feature recognition method has the characteristics of large data volume and relatively easy feature extraction, it is difficult to achieve comprehensive feature extraction. When the signal communication protocol, modulation parameters, etc. change, the fingerprint feature may become invalid. In the transient process, the signal does not carry data transmission information, so the fingerprint feature is not easily affected by factors such as changes in the communication protocol. It is only related to the hardware characteristics of the radiation source itself. Therefore, this process can be used to extract fingerprint features to identify the identity of the device. However, since the transient process usually lasts for a short time, higher requirements are placed on the algorithm in actual application. Summary of the invention

[0012] The present disclosure aims to solve at least one of the technical problems existing in the prior art, and proposes a wireless network security authentication method based on burst signal fingerprint recognition.

[0013] In a first aspect, the present disclosure provides a wireless network security authentication method based on burst signal fingerprint recognition, comprising:

[0014] S1, using the maximum likelihood detection algorithm to detect the transient process of the radiation source burst signal;

[0015] S2, calculate the instantaneous amplitude of the transient process of the signal;

[0016] S3, using the least squares estimation algorithm to extract the radiation source fingerprint characteristics implicit in the instantaneous amplitude of the signal transient process;

[0017] S4, training phase: using the support vector description algorithm to train the fingerprint of the legal radiation source and establish a wireless communication network security authentication model;

[0018] S5, authentication phase: Use the trained radiation source identification model to identify the radiation source fingerprint and complete the authentication of the device identity.

[0019] Preferably, the maximum likelihood detection amount of the maximum likelihood detection algorithm in S1 is:

[0020]

[0021] The kth complex baseband signal c formed by the received signal and the preamble sequence k Substituting into the above formula, the burst moment can be accurately detected, where the beginning of each burst of the signal is a fixed length of L, P is the oversampling factor, r k is the kth received signal.

[0022] Preferably, the S3 specifically includes:

[0023] A subtle feature extraction algorithm based on the amplitude distortion model of the signal transient process is used to extract multiple parameters to characterize the amplitude distortion caused by the non-ideality of components during the power-on and power-off process of the transmitter. The feature parameter set of known identity samples is trained using a support vector description classifier to distinguish the fingerprint features of the power-on and power-off process of different radiation sources.

[0024] Preferably, the least squares estimation algorithm model in S3 is as follows:

[0025]

[0026] R=[r(1),r(2),....,r(M+4)] T

[0027]

[0028]

[0029] Among them, r(n) is the actual received sampling signal, s(n) is the project burst signal, and n is the time series. Here the transmitter is modeled as a second-order low-pass RC system, and the system function is a 1 、a 2 , b 0 and b 1 is the coefficient of the second-order low-pass system function of the transmitter, and M is the system order of the receiver.

[0030] Preferably, the S1 specifically includes:

[0031] The wireless communication network security authentication system has a total of 5 wireless terminals, including 4 legal devices and 1 illegal device, all of which are from the same manufacturer and model;

[0032] Control 4 legal equipment terminals to transmit burst signals, with single burst duration, modulation mode, signal bandwidth and carrier frequency all randomly selected;

[0033] At the authentication center, according to the transmission signal conditions, the receiver parameters are reasonably set to obtain the baseband signal, set the relevant parameters of the wireless communication network security authentication system, and complete the transient process detection of the signal.

[0034] Preferably, the S4 specifically includes:

[0035] The support vector description classifier is trained using the transient fingerprint feature data of four legitimate devices, and a wireless communication network security authentication model is established.

[0036] Preferably, after S4 and before S5, the method further includes:

[0037] Control any device terminal to transmit a burst signal, randomly select a single burst duration, modulation mode, signal bandwidth and carrier frequency, and repeat steps S1-S4.

[0038] The present invention also provides a wireless network security authentication system based on burst signal fingerprint recognition, which can be used to implement the wireless network security authentication method based on burst signal fingerprint recognition. The system includes:

[0039] A transient detection module is configured to detect a transient process of a burst signal of a radiation source using a maximum likelihood detection algorithm;

[0040] An amplitude calculation module configured to calculate the instantaneous amplitude of a transient process of a signal;

[0041] A signal extraction module is configured to use a least squares estimation algorithm to extract the radiation source fingerprint characteristics implicit in the instantaneous amplitude of the transient process of the signal;

[0042] A training module is configured to train fingerprints of legitimate radiation sources using a support vector description algorithm to establish a wireless communication network security authentication model;

[0043] The authentication module is configured to use the trained radiation source identification model to identify the radiation source fingerprint and complete the authentication of the device identity.

[0044] Beneficial effects:

[0045] (1) The present invention uses the radiation source fingerprint characteristics of the transient process of the burst signal to realize the identity authentication of the wireless terminal device. Different from the existing radiation source fingerprint characteristics extracted by the steady-state process, the radiation source fingerprint characteristics based on the transient process are independent of the communication protocol and modulation parameters of the radiation source, and can deeply characterize the characteristics of the physical layer of the radiation source. Therefore, it has a wider range of applications, especially in wireless communication network security management scenarios where the terminal has adaptive communication functions;

[0046] (2) The present invention adopts the maximum likelihood estimation algorithm to solve the problem that the transient process of the burst signal cannot be accurately extracted, and verifies the influence of the transient process detection accuracy on the accuracy of radiation source fingerprint recognition through experimental analysis;

[0047] (3) The present invention theoretically derives the transmitter transient distortion model in the z domain, establishes a mathematical model, and uses the least squares estimation algorithm to obtain a fingerprint feature estimation method. The advantage of this algorithm is that it has low computational complexity and is suitable for wireless communication network security authentication scenarios with limited computing resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of an application scenario of a wireless network security authentication method based on burst signal fingerprint recognition provided by an embodiment of the present disclosure;

[0049] Figure 2 A curve chart showing the change in individual recognition accuracy with transient process estimation deviation provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0050] In order to enable those skilled in the art to better understand the technical solution of the present disclosure, the present disclosure is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0051] Unless otherwise defined, the technical terms or scientific terms used in the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the present disclosure are not any order, quantity or importance, but are only used to distinguish different components. Similarly, similar words such as "one", "one" or "the" are not limited to quantity, but are at least one. "Including" or "including" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used for relative positional relationships. When the absolute position of the described target changes, the relative positional relationship may also change accordingly.

[0052] In each of the accompanying drawings, the same elements are represented by similar reference numerals. For the sake of clarity, the various parts in the accompanying drawings are not all drawn to scale. In addition, some well-known parts may not be shown in the drawings.

[0053] Many specific details of the present disclosure are described below, such as component structures, materials, dimensions, processing techniques, and technologies, in order to more clearly understand the present disclosure. However, as those skilled in the art will appreciate, the present disclosure may be implemented without following these specific details.

[0054] like Figure 1 As shown, the embodiment of the present invention provides a wireless network security authentication method based on burst signal fingerprint recognition, which is characterized by comprising:

[0055] S1, using the maximum likelihood detection algorithm to detect the transient process of the radiation source burst signal;

[0056] S2, calculate the instantaneous amplitude of the transient process of the signal;

[0057] S3, using the least squares estimation algorithm to extract the radiation source fingerprint characteristics implicit in the instantaneous amplitude of the signal transient process;

[0058] S4, training phase: using the support vector description algorithm to train the fingerprint of the legal radiation source and establish a wireless communication network security authentication model;

[0059] S5, authentication phase: Use the trained radiation source identification model to identify the radiation source fingerprint and complete the authentication of the device identity.

[0060] The embodiment of the present invention adopts the maximum likelihood detection algorithm to solve the problem of accurate estimation of the transient process of the burst signal. A mathematical model of the transient fingerprint feature of the transmitter is established, and a fingerprint feature extraction algorithm based on least squares estimation is given. Finally, the support vector description classifier is used to realize the identification and authentication of the radiation source.

[0061] By accurately detecting the transient process of the burst signal and mathematically modeling the transient distortion process of the transmitter, the fingerprint characteristics of the transient process radiation source are estimated and identified, thereby completing the authentication of the wireless communication network access device. It should be noted that the technical solution disclosed in the present invention does not limit the order of the above steps, that is, the execution order of the steps can be arranged arbitrarily.

[0062] Combination Figure 1 and Figure 2 As shown, the embodiment of the present invention specifically includes the following steps:

[0063] Step 1: There are 5 wireless terminals in the wireless communication network security authentication system, including 4 legal devices and 1 illegal device. All the devices are of the same model and from the same manufacturer.

[0064] Step 2: Control four legal device terminals to transmit burst signals, with the single burst duration, modulation mode, signal bandwidth and carrier frequency all randomly selected;

[0065] Step 3: At the authentication center, according to the transmission signal situation, reasonably set the receiver parameters, obtain the baseband signal, set the relevant parameters of the wireless communication network security authentication system, and complete the transient process detection of the signal;

[0066] Step 4: Calculate the instantaneous amplitude of the transient signal;

[0067] Step 5: Use the least squares estimation algorithm to extract the fingerprint features of the transient process;

[0068] Step 5: Use the transient fingerprint feature data of the four legitimate devices to train the support vector description classifier and establish a wireless communication network security authentication model;

[0069] Step 6: Control any device terminal to transmit a burst signal, with the single burst duration, modulation mode, signal bandwidth and carrier frequency randomly selected, and repeat steps 3 to 5;

[0070] Step 7: Input the transient fingerprint feature data into the wireless communication network security authentication model to obtain the identity information of the device.

[0071] The recognition results of the test samples of the five radio stations are shown in Table 1 below:

[0072] Table 1 Recognition results of test samples of five radio stations

[0073]

[0074] The following is an explanation of the principles of each algorithm one by one.

[0075] 1) The principle of the transient process extraction algorithm based on maximum likelihood detection is as follows:

[0076] Signal transient process detection is the premise and basis for transient distortion feature extraction. Since the switching process of the signal is very short, accurately detecting the transient process of the signal directly affects the effectiveness of subsequent transient feature extraction. The commonly used autocorrelation detection algorithm and double sliding window detection algorithm are limited by the single correlation or multiplication selection window length, and the time resolution of the algorithm is difficult to meet the requirements of transient process detection. Here, the maximum likelihood algorithm is used to use the synchronization code or preamble code of the signal to accurately extract the starting time of the transient process of the burst signal.

[0077] Assume that the burst signal is modulated by FSK, and the beginning of each burst of the signal is a fixed L-length preamble sequence: C = {c 0 ,c 1 ,...,c L-1}, the received signal under Gaussian channel conditions can be expressed as:

[0078]

[0079] Where f(n) is the modulation phase of the information data symbol, ω 0 is the frequency deviation, φ 0 is the initial phase, T 0 is the sampling period, h(t) is the baseband shaped pulse function. Assuming the symbol rate of the signal is T, the oversampling factor P = T / T s .v n is the zero-mean complex Gaussian white noise with normalized signal amplitude and a variance of σ 2 Assume that 0 and 1 are equally distributed in the information data bit stream, and ω 0 T 0 ,φ 0 Uniformly distributed in [-π, π].

[0080] When the transmitted signal s(n) is known, the sample points of the received signal r(n) are independent of each other after sampling, so the conditional probability density function of the received signal r can be expressed as

[0081]

[0082] where r = {r 0 ,r 1 ,...,r PL-1},S={S 0 ,S 1 ,...,S L-1} is the symbol sequence contained in the window, s={s 0 ,s 1 ,...,s PL-1} is the complex baseband signal of the information data contained in the window.

[0083] The essence of transient process detection is to accurately detect the moment when the signal burst begins, which is a binary existence detection. The binary detection model is defined as: H 1 : The current window contains a known leading edge c, that is, the burst start point is detected; H 0 :The current window does not contain C but contains unknown sequence D.

[0084] For equation (2), 0 and ω 0 T 0 Finding the expectation in the interval [-π, π] gives:

[0085]

[0086] in Due to s k It is a constant amplitude complex baseband signal, so Q(r,s) is independent of s and is represented by Q(r) below.

[0087] H 1 Assume that the detection window contains a known leading sequence c, then the conditional probability function is

[0088]

[0089] H 0 Assume that the detection window does not contain the known leading sequence C, but contains the unknown sequence D, then the conditional probability function is

[0090] where {c k} is the complex baseband signal formed by the preamble sequence, {d k} is a complex baseband signal formed by an unknown sequence.

[0091] Since the leading sequence C is known, substituting it into p 1 (r), considering that the information sequences 0 and 1 obey uniform distribution, we have:

[0092]

[0093] After derivation and discarding the constant term, the maximum likelihood detection quantity can be obtained as

[0094]

[0095] The complex baseband signal c formed by the received signal and the preamble sequence k Substituting into the above formula, accurate detection of the sudden moment can be achieved.

[0096] 2) The principle of transient process fingerprint extraction and recognition algorithm based on least squares estimation is as follows:

[0097] Assuming that the receiver is in an AWGN channel, the baseband transient sampling signal obtained after down-conversion can be expressed as

[0098] r(n)=A(n)exp(j((ω 0 +ω n )n+φ(n)))+v(n)

[0099] Where A(n) is the amplitude transient signal, ω 0 is the frequency deviation, φ(n) is the phase transient signal, ω n is frequency modulation, v(n) is the variance σ2 Additive white Gaussian noise.

[0100] When the signal-to-noise ratio is high enough, the above formula can be approximated as

[0101]

[0102] in is the signal amplitude affected by Gaussian white noise, p(n) = v p (n) is the effect of Gaussian white noise on the signal phase.

[0103] Then the instantaneous amplitude of the received signal is:

[0104]

[0105] The instantaneous phase is:

[0106] r P (n) = j(ω 0 +ω n )n+φ(n)+p(n)

[0107] We regard the ideal burst signal s(n) as the input of a system and the actual received sampled signal r(n) as the output of the system, then we can use an N-order linear system to describe the transient change process of the signal. Assume that the system function H(z) of the discrete system can be expressed as follows:

[0108]

[0109] The Z domain of the received signal is expressed as:

[0110] R(z)=S(z)H(z)

[0111] Where S(z) is the z-transform of the ideal constant amplitude burst signal. To simplify the analysis process, the influence of the receiver on the signal is not considered here.

[0112] According to the instantaneous amplitude expression v A (n) is the variance σ 2 The white Gaussian process of the received signal can be expressed as:

[0113]

[0114] Where H(z) is the second-order low-pass RC system function of the transmitter, H(z) = H T (z)H R (z), H R (z) is the system function of the receiver filter, assuming H R (z) is an M-order FIR filter, whose filter coefficients are λ0 ,λ 1 , ...,λ M , assuming that the transmitter is a second-order low-pass RC system function

[0115]

[0116] Assume θ = [λ 0 ,λ 1 ,......,λ M-1 ,a 0 ,a 1 ,b 0 ,b 1 ], and the input signal and output signal are known, from which the Laplace transform mathematical model of the amplitude transient signal can be obtained. Assume

[0117] We use the least squares method to estimate and obtain the estimated values ​​of the characteristic parameters related to transient amplitude distortion:

[0118]

[0119] Where R = [r(1), r(2), ...., r(M+4)] T ,

[0120] Using a subtle feature extraction algorithm based on the amplitude distortion model of signal transient processes, we extracted twelve parameters to characterize the amplitude distortion caused by the non-ideality of components during the on-off process of the transmitter. By training the feature parameter set of known identity samples using a support vector description classifier, we were able to distinguish the on-off processes of different radiation sources.

[0121] The embodiment of the present invention further provides a wireless network security authentication system based on burst signal fingerprint recognition, which can be used to implement the wireless network security authentication method based on burst signal fingerprint recognition. The system includes:

[0122] A transient detection module is configured to detect a transient process of a burst signal of a radiation source using a maximum likelihood detection algorithm;

[0123] An amplitude calculation module configured to calculate the instantaneous amplitude of a transient process of a signal;

[0124] A signal extraction module is configured to use a least squares estimation algorithm to extract the radiation source fingerprint characteristics implicit in the instantaneous amplitude of the transient process of the signal;

[0125] A training module is configured to train fingerprints of legitimate radiation sources using a support vector description algorithm to establish a wireless communication network security authentication model;

[0126] The authentication module is configured to use the trained radiation source identification model to identify the radiation source fingerprint and complete the authentication of the device identity.

[0127] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present disclosure, but the present disclosure is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and substance of the present disclosure, and these modifications and improvements are also considered to be within the scope of protection of the present disclosure.

Claims

1. A wireless network security authentication method based on burst signal fingerprint recognition, characterized in that: include: S1, using the maximum likelihood detection algorithm to detect the transient process of the radiation source burst signal; S2, calculate the instantaneous amplitude of the transient process of the signal; S3, using the least squares estimation algorithm to extract the radiation source fingerprint characteristics implicit in the instantaneous amplitude of the signal transient process; S4, training phase: using the support vector description algorithm to train the fingerprint of the legal radiation source and establish a wireless communication network security authentication model; S5, authentication phase: Use the trained radiation source identification model to identify the radiation source fingerprint and complete the authentication of the device identity.

2. The wireless network security authentication method based on burst signal fingerprint recognition according to claim 1 is characterized in that: The maximum likelihood detection amount of the maximum likelihood detection algorithm in S1 is: The kth complex baseband signal c formed by the received signal and the preamble sequence k Substituting into the above formula, the burst moment can be accurately detected, where the beginning of each burst of the signal is a fixed length of L, P is the oversampling factor, r k is the kth received signal.

3. The wireless network security authentication method based on burst signal fingerprint recognition according to claim 1 is characterized in that: The S3 specifically includes: A subtle feature extraction algorithm based on the amplitude distortion model of the signal transient process is used to extract multiple parameters to characterize the amplitude distortion caused by the non-ideality of components during the power-on and power-off process of the transmitter. The power-on and power-off processes of different radiation sources can be distinguished by training the feature parameter set of known identity samples using a support vector description classifier.

4. The wireless network security authentication method based on burst signal fingerprint recognition according to claim 1, characterized in that: The least squares estimation algorithm model in S3 is as follows: R=[r(1),r(2),....,r(M+4)] T Among them, r(n) is the actual received sampling signal, s(n) is the project burst signal, and n is the time series. Here, the transmitter is modeled as a second-order low-pass RC system, and the system function is a1, a2, b0 and b1 are the coefficients of the second-order low-pass system function of the transmitter, and M is the system order of the receiver.

5. The wireless network security authentication method based on burst signal fingerprint recognition according to claim 1, characterized in that: The S1 specifically includes: The wireless communication network security authentication system has a total of 5 wireless terminals, including 4 legal devices and 1 illegal device, all of which are from the same manufacturer and model; Control 4 legal equipment terminals to transmit burst signals, with single burst duration, modulation mode, signal bandwidth and carrier frequency all randomly selected; At the authentication center, according to the transmission signal conditions, the receiver parameters are reasonably set to obtain the baseband signal, set the relevant parameters of the wireless communication network security authentication system, and complete the transient process detection of the signal.

6. The wireless network security authentication method based on burst signal fingerprint recognition according to claim 5, characterized in that: The S4 specifically includes: The support vector description classifier is trained using the transient fingerprint feature data of four legitimate devices, and a wireless communication network security authentication model is established.

7. The wireless network security authentication method based on burst signal fingerprint recognition according to claim 6 is characterized in that: After S4 and before S5, the method further includes: Control any device terminal to transmit a burst signal, randomly select a single burst duration, modulation mode, signal bandwidth and carrier frequency, and repeat steps S1-S4.

8. A wireless network security authentication system based on burst signal fingerprint recognition, characterized in that: The system can be used to implement the wireless network security authentication method based on burst signal fingerprint recognition as described in any one of claims 1 to 7 above, and the system includes: A transient detection module is configured to detect a transient process of a burst signal of a radiation source using a maximum likelihood detection algorithm; An amplitude calculation module configured to calculate the instantaneous amplitude of a transient process of a signal; A signal extraction module is configured to use a least squares estimation algorithm to extract the radiation source fingerprint characteristics implicit in the instantaneous amplitude of the transient process of the signal; A training module is configured to train fingerprints of legitimate radiation sources using a support vector description algorithm to establish a wireless communication network security authentication model; The authentication module is configured to use the trained radiation source identification model to identify the radiation source fingerprint and complete the authentication of the device identity.

Images