An immune-based radio frequency fingerprint authentication method

Abstract

The application relates to an immune-based radio frequency fingerprint identity authentication method. The method comprises the following steps: first, acquiring WiFi signal data and performing pretreatment; then, performing radio frequency fingerprint extraction based on the WiFi signal data after the pretreatment to obtain a radio frequency fingerprint vector; then, determining a variable radius self based on the radio frequency fingerprint vector; then, determining a mature detector set by using an evolution-based detector generation method; finally, detecting the variable radius self by using the mature detector set to obtain an identity authentication result. A safe and light identity authentication method is provided, the consumption of computing resources is reduced, and the accuracy of identity authentication is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless device identity authentication, in particular to an immune-based radio frequency fingerprint identity authentication method. BACKGROUND

[0002] Radio frequency fingerprint technology has been used as an additional security layer for wireless devices, with a unique fingerprint used to identify wireless devices to avoid spoofing or impersonation attacks. Radio frequency fingerprint identification is a technology that identifies individual radiation source devices by signal processing means to extract features from collected wireless signals. With its high security, non-contact identification, strong environmental adaptability and wide application scenarios, it is gradually changing people's way of life and work.

[0003] In the prior art, radio frequency fingerprint identification is usually achieved by signal processing, deep learning and physical modeling to realize identity authentication. However, due to the rapid expansion of the number of Internet of Things access devices in recent years, and due to the low power consumption and low computing resources of Internet of Things terminal devices, the demand for a secure and lightweight identity authentication method is increasing. SUMMARY

[0004] Therefore, it is necessary to provide an immune-based radio frequency fingerprint identity authentication method to solve the above technical problems.

[0005] In a first aspect, the present application provides an immune-based radio frequency fingerprint identity authentication method. The method comprises:

[0006] Obtaining WiFi signal data and preprocessing;

[0007] Extracting radio frequency fingerprints based on the preprocessed WiFi signal data to obtain a radio frequency fingerprint vector;

[0008] Determining a variable radius self based on the radio frequency fingerprint vector;

[0009] Determining a mature detector set using an evolution-based detector generation method;

[0010] Detecting the variable radius self using the mature detector set to obtain an identity authentication result.

[0011] Optionally, in an embodiment of the present application, the preprocessing includes frame synchronization, energy normalization, frequency offset estimation and compensation, and symbol alignment.

[0012] Optionally, in an embodiment of the present application, the radio frequency fingerprint vector includes frequency offset, symbol amplitude error, symbol amplitude error root mean square, symbol phase error, and symbol phase error root mean square.

[0013] Optionally, in an embodiment of the present application, the determining the variable-radius self-body based on the radio frequency fingerprint vector comprises:

[0014] based on the inter-class distance between different self-bodies in the self-body set, the inter-class distance between different self-body sets, and a preset empirical value.

[0015] Optionally, in an embodiment of the present application, the determining the mature detector set by using the evolution-based detector generation method comprises:

[0016] randomly generating an initial detector seed, and guiding the initial detector seed to be distributed around the self-body set;

[0017] evolving and self-tolerating the initial detector seed distributed around the self-body set based on the self-body density to obtain a mature detector set.

[0018] Optionally, in an embodiment of the present application, the detecting the variable-radius self-body by using the mature detector set to obtain an identity authentication result comprises:

[0019] identifying an illegal device by using the mature detector set, and identifying a legal device by using a legal device identity authentication algorithm based on mature detector distance.

[0020] Optionally, in an embodiment of the present application, the legal device identity authentication algorithm based on mature detector distance comprises:

[0021] identifying the authentication identity of the legal device based on the mean value of the distance between the mature detector and the unknown device feature vector.

[0022] In a second aspect, the present application also provides an immune-based radio frequency fingerprint identity authentication device. The device comprises:

[0023] a data acquisition module, configured to acquire WiFi signal data and perform preprocessing;

[0024] a radio frequency fingerprint extraction module, configured to perform radio frequency fingerprint extraction based on the WiFi signal data after preprocessing to obtain a radio frequency fingerprint vector;

[0025] a variable-radius self-body determination module, configured to determine a variable-radius self-body based on the radio frequency fingerprint vector;

[0026] a detector generation module, configured to determine a mature detector set by using an evolution-based detector generation method;

[0027] an identity authentication module, configured to detect the variable-radius self-body by using the mature detector set to obtain an identity authentication result.

[0028] In a third aspect, the present application provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor executes the steps of the method in each of the above embodiments.

[0029] In a fourth aspect, the present application provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the method in each of the above embodiments.

[0030] The above-mentioned immune-based radio frequency fingerprint identity authentication method first acquires WiFi signal data and performs preprocessing. Then, radio frequency fingerprint extraction is performed based on the preprocessed WiFi signal data to obtain a radio frequency fingerprint vector. Then, a variable radius self-body is determined based on the radio frequency fingerprint vector. Then, a mature detector set is determined using an evolution-based detector generation method. Finally, the variable radius self-body is detected using the mature detector set to obtain an identity authentication result. That is, by extracting the radio frequency fingerprint of the transmitted signal of the sending device, regarding the radio frequency fingerprint of each device as a self-body vector, setting the radius of each self-body according to the proposed variable self-body radius method, and then generating a mature detector set for different devices through the evolution of cloning, mutation and inhibition, and performing identity authentication. A safe and lightweight identity authentication method is proposed, which reduces the consumption of computing resources and improves the accuracy of identity authentication. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 An application environment diagram of an immune-based radio frequency fingerprint identity authentication method in an embodiment;

[0032] Figure 2 A flowchart of an immune-based radio frequency fingerprint identity authentication method in an embodiment;

[0033] Figure 3 A packet structure diagram of a WiFi signal in an embodiment;

[0034] Figure 4 A diagram of amplitude error and phase error in an embodiment;

[0035] Figure 5 A diagram of a variable radius self-body in an embodiment;

[0036] Figure 6 A diagram of cloning, mutation and inhibition in an evolution-based detector generation method in an embodiment;

[0037] Figure 7 A diagram of the change process of detector seed distribution based on distribution evolution in an embodiment;

[0038] Figure 8 Fig. 14 is a schematic diagram of a process of changing a seed distribution of a quantity evolution-based detector in an embodiment;

[0039] Figure 9 Fig. 15 is a schematic diagram of an overall process of an identity authentication stage in an embodiment;

[0040] Figure 10 Fig. 16 is a schematic diagram of a process of steps of an immune-based radio frequency fingerprint identity authentication method in an embodiment;

[0041] Figure 11 Fig. 17 is a structural block diagram of an immune-based radio frequency fingerprint identity authentication device in an embodiment;

[0042] Figure 12 Fig. 18 is an internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0043] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0044] The immune-based radio frequency fingerprint identity authentication method provided by the embodiments of the present application can be applied to an application environment as shown in Figure 1 . The terminal 102 communicates with the server 104 through a network. The data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things device can be a smart speaker, a smart television, a smart air conditioner, a smart vehicle-mounted device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0045] In an embodiment, as shown in Figure 2 , an immune-based radio frequency fingerprint identity authentication method is provided. Taking the server in Figure 1 as an example, the method includes the following steps:

[0046] S201: Obtain WiFi signal data and perform preprocessing.

[0047] In the embodiments of the present application, first, the transmission signal of the sending device, i.e., WiFi signal data, is collected, and the signal data is preprocessed, such as frame synchronization, frequency offset estimation and compensation, normalization, etc. Specifically, the WLAN product based on the IEEE 802.11 standard is defined as a WiFi device, and the packet structure of the signal based on the IEEE 802.11g standard is as shown in FIG. 3. Figure 3 As shown in FIG. 3, the 802.11g works in the 2.4G frequency band, and uses the orthogonal frequency division multiplexing (OFDM) technology as the modulation technology of the wireless local area network (WLAN) system. The length of each OFDM symbol is 4us, of which the length of the guard interval GI is 0.8us, and the length of the data part is 3.2us. These information will be used for the extraction of the subsequent signal frame. The function of inserting the guard interval is to eliminate the inter-symbol interference (ISI) caused by the multipath channel. Under the 20M bandwidth, there are 64 subcarriers in each symbol. In the 802.11g mode, only 52 subcarriers are valid subcarriers in the 64 subcarriers, of which 48 are data subcarriers, and 4 are pilot subcarriers. The 802.11g packet signal is composed of a preamble and a data part. The preamble includes a short training sequence (STF), a long training sequence (LTF) and a signal field (SIGNAL).

[0048] The short training sequence occupies 12 of the 52 non-zero subcarriers in each OFDM symbol, and is composed of 10 repeated parts, each part lasting 0.8us. Under the 20M sampling rate, the short training sequence has a total of 160 sampling points. The short training sequence can be used for coarse time synchronization and coarse frequency offset estimation due to its repetition. The long training sequence occupies all 52 non-zero subcarriers, and is composed of a guard interval GI2 and two repeated segments, with a total duration of 8us. Under the 20M sampling rate, it also contains 160 sampling points. The long training sequence can be used for fine time synchronization and fine frequency offset estimation, as well as channel response estimation. The signal field occupies a length of 1 symbol, and is modulated using BPSK. The receiver can read the data length, speed and other information from it.

[0049] Assuming that the baseband signal transmitted by the transmitter is x(t), the baseband signal will be modulated by hardware such as digital-to-analog converter, filter, mixer, power amplifier, etc., and finally transmitted through the antenna. Assuming that the overall damage caused by these hardware in the process is defined as f(·). Through the wireless channel transmission, the collected signal can be represented as:

[0050] y(t) = h(t) * f(x(t)) + n(t)

[0051] Where h(t) is the channel response, and n(t) is the additive white Gaussian noise. The received signal y(t) is obtained by using the discrete signal, and assuming that the sampling frequency is T, the sampled received signal can be represented as y[nT], which can be abbreviated as y[n].

[0052] After that, pre-processing operation is performed on the received signal, in one embodiment of the present application, the pre-processing includes frame synchronization, energy normalization, frequency offset estimation and compensation, symbol alignment.

[0053] In one embodiment of the present application, frame synchronization is performed based on the autocorrelation of the short training sequence. Due to the cyclic property of the short training sequence, the autocorrelation is high at the beginning of the frame. With this cyclic property, the autocorrelation value a of the received signal y[n] with a lag of 16 samples is calculated as follows (here, y* represents the complex conjugate of y):

[0054]

[0055] where N win The window length of the summation, plays the role of a low-pass filter. In order to be independent of the absolute level of the input samples, the autocorrelation is normalized with the average power p, and the autocorrelation coefficient c is calculated as:

[0056]

[0057] If three consecutive samples exceed a set threshold, the beginning of the frame is detected, and frame synchronization is achieved.

[0058] Since the transmission power of different transmitters can be different, and the classification should not distinguish the devices according to the power of the received signal, energy normalization is needed for the extracted frame signal y[n], i.e. each sample is divided by the root mean square value of the amplitude signal, and then the energy normalization is:

[0059]

[0060] where Y represents the root mean square value of the amplitude of the frame signal.

[0061] Since there is a difference in the carrier frequency between the up-conversion in the transmitter and the down-conversion in the receiver of the baseband signal, a carrier frequency offset (CFO) is generated. The CFO will cause an inter-carrier interference effect and affect the performance of the OFDM system, so frequency offset estimation and compensation are needed.

[0062] First, a rough frequency offset estimation is performed using the short training sequence, and the specific formula is as follows:

[0063]

[0064] The rough carrier frequency offset estimation value f stf is obtained by the short training sequence, and the long training sequence LTF is compensated, and then the long training sequence S LT ​Fine carrier frequency offset estimation is performed, and the long training sequence is processed according to the method of coarse frequency offset estimation:

[0065]

[0066] The final frequency offset estimation value is F = f stf +f ltf , and the signal after carrier frequency compensation is y[n] = y[n]e -jF .

[0067] The task of symbol alignment is to calculate the position of the start of the symbol, extract the data symbol, and provide it for subsequent FFT transformation and symbol demodulation. Symbol alignment is achieved using the long training sequence, since the sequence is composed of a 64-sample long pattern repeated 2.5 times, the correlation is calculated and the three highest peaks are extracted:

[0068]

[0069] Where N pre corresponds to the length of the preamble sequence, L is the 64-sample repeated pattern of the long training sequence, and argmax3 returns the first three indices that maximize the expression. Therefore, the first data symbol starts at sample index: n p = max(N p )+64, thereby achieving symbol synchronization.

[0070] S203: Extract the radio frequency fingerprint vector based on the WiFi signal data after preprocessing.

[0071] In the embodiments of the present application, after preprocessing, the radio frequency fingerprint of the signal data is extracted, and a plurality of features of each signal are extracted to form a radio frequency fingerprint vector of the device.

[0072] In an embodiment of the present application, the radio frequency fingerprint vector includes frequency offset, symbol amplitude error, symbol amplitude error root mean square, symbol phase error, and symbol phase error root mean square.

[0073] In an embodiment of the present application, for the preprocessed signal, the following five features are extracted to form a radio frequency fingerprint vector of the device: frequency offset, symbol amplitude error, symbol amplitude error root mean square, symbol phase error, and symbol phase error root mean square. Specifically, the deviation of the carrier frequency of the receiver down-conversion and the carrier frequency of the transmitter up-conversion causes the received signal to have a frequency offset, the frequency offset will affect the orthogonality of the subcarriers, and will also cause amplitude distortion, phase distortion, demodulation error, etc. The frequency offset is caused by the mixer error of the transmitter and the receiver, which reflects the hardware characteristics of the device, so the frequency offset CFO is taken as one of the features of the radio frequency fingerprint, and its solving process is shown in the preprocessing process.

[0074] The physical layer of IEEE 802.11 standards uses I / Q modulation techniques to transmit bit information over the wireless channel, the symbols of I / Q modulation scheme can be defined using a constellation diagram, where different symbols can be represented as points in the I / Q space or modulation domain, one symbol can encode multiple data bits. The original modulated ideal signal will be distorted when passing through the hardware impairment of the transmitter, channel influence, etc., so that when the demodulation received signal is mapped to the I / Q plane, the modulation error between the measured signal and the ideal signal is obtained, and the modulation error reflects the radio frequency fingerprint of the device. The preprocessed signal is demodulated, and the amplitude error (ME) and the phase error (PE) are selected as the 2nd and 3rd dimensional features of the radio frequency fingerprint, as shown in Figure 4 Since the error measurement of only one symbol in a frame is meaningless for the experiment, 38 symbols are extracted for each frame, and the amplitude error (ME) and the phase error (PE) are used to describe the average error of all symbols in a frame.

[0075] Assuming that the modulation symbol point of the ideal signal in the I / Q plane is I, and the measured symbol point obtained by demodulating the received signal is R, the amplitude error (ME) can be represented as:

[0076]

[0077] The phase error (PE) can be represented as:

[0078]

[0079] Where angle is the angle of the symbol.

[0080] The root mean square of the amplitude error (ME) and the phase error (PE) is taken as the 4th and 5th dimensional features of the radio frequency fingerprint. The root mean square of the symbol amplitude error (MES) is solved as:

[0081]

[0082] The root mean square of the phase error (PES) is solved as:

[0083]

[0084] In summary, the radio frequency fingerprint vector can be represented as RFF = {CFO, ME, PE, MES, PES}.

[0085] S205: Determine the variable radius self based on the radio frequency fingerprint vector.

[0086] In the embodiments of the present application, the artificial immune system (AIS) is introduced into the radio frequency fingerprint identity authentication, and the radio frequency fingerprint of each device is regarded as a self vector, and the radius of each self of the device is determined according to the proposed variable self radius method. The artificial immune algorithm (AIS) is an intelligent evolutionary algorithm simulating the biological immune system (BIS), which recognizes self antigens by generating antibody substances and excludes non-self antigens. The artificial immune algorithm realizes the recognition of normal data and abnormal data by generating detectors. The negative selection algorithm (NSA) is the most popular algorithm in the current artificial immune system, and the principle is to simulate the immune process in which T lymphocytes in the biological immune system mature through self tolerance. By simulating the process, the radio frequency fingerprint of the current device is regarded as a self set, and the radio frequency fingerprints of all other devices are regarded as a non-self set, and the identity authentication process of the current device is converted into the recognition process of self and non-self through the training of antibody detectors.

[0087] The following definitions are made:

[0088] Self: the feature string sample from the legal device:

[0089] Self={<x,r s >|x∈U,r s ∈R +}

[0090] The self is composed of two parts, x is an n-dimensional feature vector representing the center of the self, and r s is the self radius. For any feature vector within the self radius, it is considered to be a self element. The non-self is as defined above.

[0091] Antibody: abstracted as a detector, which is a basic unit for detecting antigens, and the detector set can be defined as:

[0092] D={<y,r s >|y∈NoneSelf,r s ∈R +}

[0093] Where y is the center vector of the detector, and r s is the radius of the detector. The area covered by the detector is considered to be a non-self area, and the area not covered is considered to be a self area.

[0094] Affinity: expressed as the distance between feature vectors:

[0095]

[0096] The Minkowski distance is the most commonly used distance formula between real number vectors, which becomes the Euclidean distance when ε = 2. The Euclidean distance is the most commonly used affinity calculation method for real-valued negative selection algorithms.

[0097] The performance of artificial immune algorithms depends largely on the quality of detectors, and the performance of detector sets is also affected by the self-radius. In traditional NSA, the self-radius is fixed, and is usually set by experiments and experience. A too small self-radius will cause too many small holes between self-bodies and self-bodies, which cannot be covered by detectors, resulting in an increased false negative rate (reduced coverage). A too large self-radius may cause some non-self-body regions to be incorrectly covered by self-bodies, resulting in an increased false positive rate. Therefore, a suitable self-radius plays an important role in the generation of detectors and the overall model effect. Based on the idea of variable self-radius, the distribution characteristics of different self-body sets and the distances between different self-body sets are comprehensively considered to set the radius of each self-body.

[0098] In an embodiment of the present application, determining the variable-radius self-body based on the radio frequency fingerprint vector comprises:

[0099] Based on the inter-class distance between different self-bodies in the self-body set, the inter-class distance between different self-body sets, and a preset empirical value.

[0100] In an embodiment of the present application, the radius of each self-body is determined by the following three values: the inter-class distance between different self-bodies in the current self-body set, the inter-class distance between different self-body sets, and an empirical value. Assuming that there are k devices in total, each device extracts n radio frequency fingerprints as a self-body set, and the empirical value is set to rs.

[0101] The radius of the jth self-body of device i is specifically set as follows:

[0102] Calculate the minimum distance from the jth self-body to all self-bodies of device i, denoted as Indis min

[0103] Calculate the minimum distance between all self-bodies of device i and all self-bodies of the remaining devices, denoted as outdis min

[0104] Then the self-body radius is set as:

[0105] Self i .x j .r s =min(min(Indis min ,outdis min / 2),rs)

[0106] The radius of each device is calculated by the above rules, and the variable radius self-body is finally obtained, which has the effect as shown in Figure 5 The variable radius improves the effect of the model, and optimizes the false positive rate and the false negative rate, compared with the fixed radius.

[0107] S207: Determine a mature detector set by using an evolution-based detector generation method.

[0108] In an embodiment of the present application, the evolution mechanism of cloning, mutation and suppression is used to generate corresponding mature detector groups for different devices. In the traditional negative selection algorithm, detectors are first generated randomly to ensure their diversity, and then the self-tolerant process is used to delete illegal detectors that recognize self-bodies, and finally the mature detectors are obtained. If the expected coverage rate of the mature detector set to the non-self-body space reaches a preset threshold, the training process is terminated. In the above process, for the narrow non-self-body region formed by the interlacing of self-bodies in the feature space, it is difficult for the candidate detector to be randomly generated to this space, so that the detector cannot effectively distinguish the coverage of this region; while in the large area of non-self-body region, the randomly generated detector is easy to gather, resulting in large area overlap of the detector coverage region, thereby reducing the efficiency of the algorithm. At the same time, different self-body sets have different distribution characteristics, and the traditional detector generation method does not take this into account. In view of these problems, an evolution-based detector generation method is used, and the evolution mainly includes three steps of cloning, mutation and suppression, as shown in Figure 6 Cloning is to make the target quantity more, mutation is to produce excellent target individuals, and suppression is to delete individuals with poor quality. The evolution-based detector generation method includes two evolution processes. First, an initial detector seed is randomly generated, and the detector seed is guided to distribute around the self-body through the first distribution evolution process. The detector seed is subjected to the second quantity evolution process, so that the number of seeds around the self-body distribution is large, and the number of seeds in the large area of non-self-body region is small. Finally, the detector seed is self-tolerant to form a mature detector set.

[0109] Specifically, in an embodiment of the present application, the method for determining a mature detector set by using an evolution-based detector generation method comprises:

[0110] S301: Randomly generate an initial detector seed, and guide the initial detector seed to distribute around the self-body set.

[0111] S303: Perform self-body density-based evolution and self-tolerance on the initial detector seed distributed around the self-body set to obtain a mature detector set.

[0112] In one embodiment of the present application, first, a distribution evolution process is performed, the main purpose of this stage is to let the randomly generated detector seeds evolve to be distributed around the self set. The specific process is as follows:

[0113] Step 1: initialize detector seeds.

[0114] Generate a set of n random detector seeds B in the feature space:

[0115] B = {b | b = (x1, x2, x3…xn), b e U} n

[0116] Calculate the distance matrix M between the seeds and the self set, find the minimum distance between each seed and all the self bodies, and sort M in ascending order according to the minimum distance: min

[0117] M min = (b i , d i ), i = 1, 2, 3…n

[0118] Where b i is the i-th seed, d i is the shortest distance between it and all the self bodies, and M min is selected according to M mean The k individuals with the closest distance are selected to form the initial antibody seed.

[0119] Step 2: cloning.

[0120] Calculate the average of the entire distance matrix M to obtain M mean , and clone the antibody seeds according to the distance M min , with the cloning target being the closer to the self body, i.e. the greater the affinity, the more the number of clones. Set the threshold sp = 0.5 * M mean , and the rules are as follows:

[0121]

[0122] Cloning makes the number of seeds around the self body increase, so that a sufficient number of seeds can cover the area around the self body faster.

[0123] Step 3: mutation.

[0124] Since the cloned antibody seeds and the cloned individuals are located at the same position, the distribution of the antibody seed set after cloning is the same as before cloning. In order to make the seeds distribute around the self set, a mutation operation is needed, so that the seeds have a greater chance of distributing around the self set. M min ​​Small seeds are near the edge of the self distribution, so only need a suitable variation to make them around other positions of the self distribution, too small variation will not make obvious effect, and too large variation will make the seeds far away from the self distribution, so set the variation threshold to be no more than the distance M of the nearest self min . M min Large seeds need to move to the periphery of the self distribution, so need a larger variation, so set the variation threshold to be a multiple of the distance threshold sp. The rule of defining the variation coefficient var is as follows:

[0125]

[0126] where, is the ceiling, set to an integer multiple of sp.

[0127] For the variation of antibody seed b = (x1, x2, x3…x n ), each dimension of its attribute is independently randomly changed:

[0128] x i = x i + rand(var)

[0129] Step 4: Inhibition.

[0130] After cloning and variation, theoretically, more seeds can be obtained around the self set. Since the variation is a random process within a certain range, the seeds after cloning and variation may be far apart, or may be very close due to small variation. At the same time, unrelated seeds may also be very close due to variation. Overly similar antibody seeds will have a large overlap in the subsequent detector generation process, so inhibition operation is needed to delete overly similar individuals. Calculate the distance matrix between all antibody seeds after variation, find the average distance d_avg of the distance matrix, set the inhibition threshold to sp = 0.3 * d_avg, when the distance between any two seeds is less than the threshold, delete the one that is close to the self set.

[0131] Step 5: Loop.

[0132] Recalculate the affinity matrix M between the seed set after inhibition and the self set, find the minimum distance between each seed and all self, sort M min in ascending order according to the minimum distance, and according to M minOnly the k individuals with the highest affinity are retained. The resulting k seeds are now somewhat optimized in terms of distribution compared to the initial k seeds. To make the optimization even higher, an iteration number Gen can be set, and the retained k seeds are brought into step 2 for iteration again. When the iteration number Gen is met, the last retained k seeds form the antibody seed set and are brought into the subsequent process.

[0133] As shown in Figure 7 , the detector seed distribution change process based on distribution evolution is shown. By evolving the detector seeds to better gather around the self distribution in spatial distribution, the generation efficiency of the subsequent mature detectors is improved.

[0134] After the distribution evolution process, the detector seeds around the self set distribution are obtained, but the distribution may not be uniform. The goal is to have more detectors in the area with high self density and small radius, and fewer detectors in the large area with low self density and large radius. Therefore, the detector seeds are again evolved based on the number, so that the number of seeds increases in areas with high self density and decreases in areas with low self density. Finally, the detector seeds are trained into mature detectors that meet the coverage rate. The specific process is as follows:

[0135] Step 1: Initial antigen density.

[0136] Calculate the distance matrix M between the seed set and the self set, calculate the mean value d_avg of M as the average distance between the seed set and the self set, and use this as the threshold. Therefore, the calculation method of the antigen density of each detector seed is:

[0137]

[0138] where Minkowski(d i ,x i ) is the distance between the i-th seed and the i-th self, and if the distance is less than the threshold, the antigen density value of the seed is increased by 1.

[0139] Step 2: Cloning.

[0140] Clone the detector seeds according to the antigen density. The purpose of cloning is to have more detector seeds in areas with high antigen density, so that mature detectors have a chance to enter narrow non-font spaces, and to have fewer detector seeds in areas with low antigen density, so as to reduce the redundancy between mature detectors. Calculate the mean value of the initial antigen density of all detector seeds to get den mean , which is used as the threshold, and the cloning rules are as follows:

[0141]

[0142] Through the above cloning method, the detector seed with high antigen density will get more quantity.

[0143] Step 3: Mutation.

[0144] Since the cloned detector seed and the cloned individual are located at the same position, the cloned detector seed needs to be mutated to make the detector seed have a greater chance of surrounding the self distribution. At this time, den i Large seeds are located at the edge of the self distribution, and den i Small seeds are relatively far away from the self, so the mutation rule should be den i The larger, the smaller the mutation. According to the distance matrix M between the above calculated detector seed set and the self set, the maximum distance dis max And the minimum distance dis min of each detector seed to the self set is calculated, which is used as the mutation range of the mutation coefficient var, and the mutation rule is as follows:

[0145]

[0146] The mutation of the detector seed is the same as the random change of each dimension attribute of the detector seed.

[0147] Step 4: Inhibition.

[0148] In order to cover the non-self space in the narrow self dense area with the detector, while reducing the redundancy between the detectors in the large area non-self area, the cloned detector seed needs to be deleted and inhibited. According to the above step, the distance matrix M, the average value d_avg of M, the antigen density den i of each detector seed and the average value den mean of all seed antigen densities, the average distance between all selves in the self set is calculated as the reference distance dself_avg, and the threshold for inhibition is set as sp:

[0149]

[0150] According to the inhibition threshold sp, the distance between any two detector seeds is calculated, and if it is less than the threshold sp, the individual closer to the self set is deleted to realize the inhibition operation. sp is an inhibition threshold with reference to the antigen density. For the area with high antigen density, in order to cover the narrow non-self space, the detector is allowed to have a larger overlap. For the area with small antigen density, in order to reduce redundancy, the overlap between detectors is reduced.

[0151] Step 5: Self-tolerance.

[0152] The detector seeds are evolved based on self-density to generate mature detectors, and the mature detector set is finally obtained by self-tolerance of the seeds. Here, the generation of mature detectors is based on the idea of V-detector. Since the inhibition process has been introduced in the above evolution process, the overlap problem of mature detectors is no longer considered in the detector generation stage, and only the problem of whether self-tolerance is legal is considered. The purpose of self-tolerance is to make the mature detector only cover the non-self region and cannot cover the self. The specific process is to find the nearest self for each seed, and if the distance dis of the seed to the self is greater than the radius x of the self, then the seed can become a mature detector, and the center of the mature detector is the seed and the radius is d. s s s , otherwise self-tolerance fails and the seed is deleted.

[0153] As shown in Figure 8 , the effect of detector seeds through quantity evolution is shown, and quantity evolution further optimizes the effect of detector seeds compared with distribution evolution.

[0154] S209: detecting the variable radius self using the mature detector set to obtain an identity authentication result.

[0155] In the embodiments of the present application, in the immune stage, the corresponding mature detector set is generated for the subsequent identity authentication by the proposed immune method based on the extracted radio frequency fingerprints of each device. Since the immune stage only performs the tolerance process on the self set of the legal device to generate the corresponding mature detector set, the immune process is equivalent to the registration process of the legal device. The identity authentication stage includes two processes of illegal device identification and legal device classification, which is realized by extracting the radio frequency fingerprints of the unknown device through all the mature detector sets generated in the immune stage. For illegal devices, since they have not gone through the immune stage, they will be identified and marked as illegal. For legal devices, classification is realized by a detector distance-based identity authentication method. As shown in Figure 9 , it is a schematic diagram of the whole process of the identity authentication stage.

[0156] Specifically, in an embodiment of the present application, the detection of the variable radius self using the mature detector set to obtain an identity authentication result includes:

[0157] The mature detector set is used to identify illegal devices, and a legal device identity authentication algorithm based on the distance of the mature detector is used to identify legal devices.

[0158] ​​In an embodiment of the present application, the radio frequency fingerprint is extracted from the device signal as the self-body first, and is submitted to the immune recognition system, then the immune recognition system makes the self-body pass through the mature detector set of all legal devices in turn. For illegal devices, its feature set is not in the self-body set of any legal device, so it will be identified by all mature detector sets, and is judged as illegal device. For legal devices, its self-body vector is distributed in its own self-body set space, so its feature vector will be identified by other mature detector sets except its own class, and the class of the mature detector set which does not identify it is the identity of the legal device.

[0159] Since the self-body set of some legal devices will inevitably overlap in the feature space, and then produce some confusion areas, these confusion areas will not be covered by the mature detector set of multiple devices at the same time, so as to cause the unknown legal device to be judged as multiple identities. For this problem, a legal device identity authentication algorithm based on mature detector distance is proposed to identify the legal device.

[0160] Specifically, in an embodiment of the present application, the legal device identity authentication algorithm based on mature detector distance includes:

[0161] Identifying the authentication identity of the legal device based on the mean value of the distance between the mature detector and the feature vector of the unknown device.

[0162] In an embodiment of the present application, when the unknown legal device is judged as multiple identities (assuming n (n>1) identities), a threshold k is set, and for the feature vector d of the current unknown legal device i , k nearest detectors are found from the n mature detector sets respectively, and the mean value of the distance between the k detectors of each class and the feature vector d i is counted. Since there are n identities, n mean values can be obtained, and the class of the maximum value in the n mean values is the authentication identity of the legal device.

[0163] The above-mentioned immune-based radio frequency fingerprint identity authentication method first acquires WiFi signal data and performs preprocessing. Then, radio frequency fingerprint extraction is performed based on the preprocessed WiFi signal data to obtain a radio frequency fingerprint vector. Next, a variable radius self-body is determined based on the radio frequency fingerprint vector. Then, a mature detector set is determined using an evolution-based detector generation method. Finally, the variable radius self-body is detected using the mature detector set to obtain an identity authentication result. That is, by extracting the radio frequency fingerprint of the transmitted signal of the sending device, the radio frequency fingerprint of each device is regarded as a self-body vector, the radius of each self-body is set according to the proposed variable self-body radius method, and then a mature detector set is generated for different devices through the evolution of cloning, mutation and suppression, and identity authentication is performed. A safe and lightweight identity authentication method is proposed, which reduces the consumption of computing resources and improves the accuracy of identity authentication.

[0164] The specific implementation steps of the immune-based radio frequency fingerprint identity authentication method of the present application are described below with a specific embodiment. As shown in Figure 10 first, S401, WiFi signal data is acquired and preprocessed. The preprocessing includes frame synchronization, energy normalization, frequency offset estimation and compensation, and symbol alignment. Then, S405, radio frequency fingerprint extraction is performed based on the preprocessed WiFi signal data to obtain a radio frequency fingerprint vector. S407, the radio frequency fingerprint vector includes frequency offset, symbol amplitude error, symbol amplitude error root mean square, symbol phase error and symbol phase error root mean square.

[0165] Then, S409, the variable radius self-body is determined based on the radio frequency fingerprint vector. S411, the inter-class distance between different self-bodies in the self-body set, the class-class distance between different self-body sets and the preset empirical value are used. Then, S413, a mature detector set is determined using an evolution-based detector generation method. S415-S417, an initial detector seed is randomly generated, and the initial detector seed is guided to distribute around the self-body set; the initial detector seed distributed around the self-body set is evolved based on the self-body density and self-body tolerance to obtain a mature detector set.

[0166] Finally, S419, the variable radius self-body is detected using the mature detector set to obtain an identity authentication result. S421, the mature detector set is used to identify illegal devices, and a legal device identity authentication algorithm based on the mature detector distance is used to identify legal devices. S423, the mean value of the distance between the mature detector and the unknown device feature vector is used to identify the authentication identity of the legal device.

[0167] It should be understood that although the steps in the flowcharts involved in the embodiments described above are shown in sequence according to the arrows, the steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the embodiments described above can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of the steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0168] Based on the same inventive concept, the embodiments of the present application also provide an immune-based radio fingerprint identity authentication device for implementing the above-mentioned immune-based radio fingerprint identity authentication method. The problem-solving implementation scheme provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more immune-based radio fingerprint identity authentication device embodiments provided below can refer to the limitations of the immune-based radio fingerprint identity authentication method in the above text, which will not be repeated here.

[0169] In one embodiment, as shown in Figure 11 An immune-based radio fingerprint identity authentication device 1100 is provided, comprising a data acquisition module 1101, a radio fingerprint extraction module 1103, a variable radius self-determination module 1105, a detector generation module 1107, and an identity authentication module 1109, wherein:

[0170] The data acquisition module 1101 is configured to acquire WiFi signal data and perform preprocessing.

[0171] The radio fingerprint extraction module 1103 is configured to perform radio fingerprint extraction based on the preprocessed WiFi signal data to obtain a radio fingerprint vector.

[0172] The variable radius self-determination module 1105 is configured to determine a variable radius self based on the radio fingerprint vector.

[0173] The detector generation module 1107 is configured to determine a mature detector set using an evolution-based detector generation method.

[0174] The identity authentication module 1109 is configured to detect the variable radius self using the mature detector set to obtain an identity authentication result.

[0175] In an embodiment of the present application, the preprocessing includes frame synchronization, energy normalization, frequency offset estimation and compensation, and symbol alignment.

[0176] In one embodiment of the present application, the radio frequency fingerprint vector comprises frequency offset, symbol amplitude error, root mean square of symbol amplitude error, symbol phase error and root mean square of symbol phase error.

[0177] In one embodiment of the present application, the variable radius self-determination module is further configured to:

[0178] based on the inter-class distance between different self-bodies in the self-body set, the class-class distance between different self-body sets and a preset empirical value.

[0179] In one embodiment of the present application, the detector generation module is further configured to:

[0180] randomly generate an initial detector seed, and guide the initial detector seed to distribute around the self-body set;

[0181] evolve and self-tolerate the initial detector seed distributed around the self-body set based on self-body density to obtain a mature detector set.

[0182] In one embodiment of the present application, the identity authentication module is further configured to:

[0183] identify illegal equipment by using the mature detector set, and identify legal equipment by using a legal equipment identity authentication algorithm based on mature detector distance.

[0184] In one embodiment of the present application, the identity authentication module is further configured to:

[0185] identify the authentication identity of the legal equipment based on the mean value of the distance between the mature detector and the unknown equipment feature vector.

[0186] The above-mentioned various modules in the above-mentioned immune-based radio frequency fingerprint identity authentication device can be all or part realized by software, hardware and combinations thereof. The above-mentioned various modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned various modules.

[0187] In one embodiment, a computer device is provided, which can be a terminal, and the internal structure diagram thereof can be as shown in Figure 12As shown in the figure. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used for wired or wireless communication with external terminals. Wireless mode can be achieved through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to implement an immune-based radio frequency fingerprint identity authentication method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad provided on the shell of the computer device. It can also be an external keyboard, touchpad or mouse, etc.

[0188] Those skilled in the art can understand that, Figure 12 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0189] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in the above method embodiments.

[0190] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the steps in the above method embodiments.

[0191] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to implement the steps in the above method embodiments.

[0192] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0193] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0194] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0195] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An immune-based radio frequency fingerprint authentication method, characterized by, The method comprises: acquiring WiFi signal data and preprocessing; extracting radio frequency fingerprints based on the preprocessed WiFi signal data to obtain a radio frequency fingerprint vector; determining a variable radius self based on the radio frequency fingerprint vector; determining a mature detector set using an evolution-based detector generation method; detecting the variable radius self using the mature detector set to obtain an identity authentication result; The method comprises: randomly generating an initial detector seed and guiding the initial detector seed to distribute around a self set; evolving and tolerating the self based on the density of the initial detector seed distributed around the self set to obtain a mature detector set.

2. The immune-based radio frequency fingerprint authentication method of claim 1, wherein, The preprocessing includes frame synchronization, energy normalization, frequency offset estimation and compensation, and symbol alignment.

3. The radio frequency fingerprint identity authentication method based on immunity according to claim 1, characterized in that: The radio frequency fingerprint vector includes frequency offset, symbol amplitude error, symbol amplitude error root mean square, symbol phase error, and symbol phase error root mean square.

4. The immune-based radio frequency fingerprint authentication method of claim 1, wherein, The method comprises: based on the inter-class distance between different selves in the self set, the class-class distance between different self sets, and a preset empirical value.

5. The immune-based radio frequency fingerprint authentication method of claim 1, wherein, The method comprises: using the mature detector set to identify illegal devices and using a legal device identity authentication algorithm based on mature detector distance to identify legal devices.

6. The immune-based radio frequency fingerprint authentication method of claim 5, wherein, The legal device identity authentication algorithm based on mature detector distance comprises: identifying the authentication identity of a legal device based on the mean value of the distance between the mature detector and the unknown device feature vector.

7. An immune-based radio frequency fingerprint authentication device, comprising: The device comprises: a data acquisition module for acquiring WiFi signal data and preprocessing; a radio frequency fingerprint extraction module for extracting radio frequency fingerprints based on the preprocessed WiFi signal data to obtain a radio frequency fingerprint vector; a variable radius self determination module for determining a variable radius self based on the radio frequency fingerprint vector; a detector generation module for determining a mature detector set using an evolution-based detector generation method; an identity authentication module for detecting the variable radius self using the mature detector set to obtain an identity authentication result; The method comprises: randomly generating an initial detector seed and guiding the initial detector seed to distribute around a self set; evolving and tolerating the self based on the density of the initial detector seed distributed around the self set to obtain a mature detector set. 8.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-7. The processor executes the computer program to implement the steps of the method of any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

Images