A radio access identification method based on radio frequency fingerprints in a MIMO scenario

Abstract

The application provides a kind of MIMO scene based on radio frequency fingerprint air interface access identification method, comprising: establishing the communication architecture of multiple antenna base station and several terminal wireless devices;Terminal wireless device sends signal;Accept and separate signal;Feature extraction and identification;Restore original data.In feature extraction, the application uses the characteristics of multiple antenna equipment in MIMO scene, proposes the concept of relative radio frequency fingerprint, calculates the hardware mismatch relationship between multiple antennas as the radio frequency characteristics of equipment, and classifies and identifies it;And propose a communication protocol mechanism, both sides send pilot to obtain channel parameters, one side uses channel parameters to construct equalization matrix to process received pilot signal, and eliminate the influence of radio channel on radio frequency fingerprint characteristics.The application can obtain stable radio frequency fingerprint characteristics and effectively improve the identification accuracy.

Description

Technical Field

[0001] This invention belongs to the field of information security technology and relates to the application technology of relative radio frequency fingerprinting and signal backhaul protocol in Internet of Things terminal devices. Specifically, it relates to an air interface access identification method based on radio frequency fingerprinting in MIMO scenarios. Background Technology

[0002] The Internet of Things (IoT) is one of the most popular technologies in the era of digital transformation. While IoT technology is developing rapidly, it faces many security problems due to limited device resources, large device scale, diversified attacks, and lack of standardized identity authentication. IoT device access authentication security also faces many challenges.

[0003] While numerous authentication schemes and security protocols exist to enhance communication security, these schemes often suffer from issues such as high computational load, complex hardware node authentication processes, limited scalability, and communication mode incompatibility. Meanwhile, physical layer authentication, as a crucial aspect of physical layer security, plays an indispensable role in secure wireless communication. Currently, RF fingerprint-based physical layer authentication technology has attracted widespread attention and research both domestically and internationally.

[0004] Radio frequency (RF) fingerprinting technology leverages the hardware differences between wireless devices caused by their internal electronic components. It collects RF signals, analyzes and extracts these unique hardware features as a fingerprint to uniquely identify a specific RF device. Because RF fingerprints possess excellent characteristics such as uniqueness, universality, independence, robustness, and long-term invariance, and because these characteristics are independent of the specific data they carry, they have a wide range of applications.

[0005] However, existing radio frequency fingerprint extraction technologies are not mature, and accurately extracting device fingerprints is not easy; moreover, an accurate device fingerprint is unnecessary during the identification process. Furthermore, because radio frequency fingerprint recognition technology relies on radio frequency features embedded in wireless signals for device authentication, the received signal is mixed with wireless channels and RFF (Radio Frequency Imaging), which is detrimental to feature extraction. Existing technologies mainly focus on improving classifier accuracy, but none can completely eliminate the influence of wireless channels.

[0006] Therefore, how to solve the difficulty of extracting radio frequency fingerprints from IoT devices and how to separate the wireless channel from the radio frequency fingerprint are urgent problems to be solved. Summary of the Invention

[0007] To address the issue that radio frequency fingerprints are parasitic on signal transmission and their features are easily affected by channel parameters, leading to unstable fingerprint features, this invention proposes a communication protocol mechanism. Both communicating parties send pilot signals to obtain channel parameters. One party then uses the channel parameters to construct an equalization matrix to process the received pilot signals, eliminating the influence of the wireless channel on the radio frequency fingerprint features, thereby obtaining stable fingerprint features unaffected by the channel.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for air interface access identification based on radio frequency fingerprinting in a MIMO scenario includes the following steps:

[0010] S1, establish a communication architecture between a multi-antenna base station and several terminal wireless devices. Each terminal wireless device needs to send a data request to the base station to communicate with the base station.

[0011] S2, the terminal wireless device sends a signal. The transmitted data is encoded, modulated, inserted with pilots, and has defects added before reaching the base station receiver section via the wireless channel. After adding receiver defects, the received signal is obtained.

[0012] S3, receive and separate the signal; the base station separates the pilot and data of the received signal.

[0013] S4, Feature Extraction and Recognition: The pilot section obtains the transmitter's radio frequency features through the radio frequency fingerprint extraction module, and then obtains the device recognition result through the classification module;

[0014] S5, recover the original data. The data part is demodulated and decoded to obtain the original data that was sent.

[0015] Furthermore, step S4 includes the following sub-steps:

[0016] S41, establish normalized calibration matrices for base stations and user equipment respectively;

[0017] S42, the base station first sends downlink pilot signals to the user;

[0018] S43, the user receives a signal that includes a downlink radio channel;

[0019] S44, the user performs channel estimation based on the received signal;

[0020] S45, within the coherence time, the user sends an uplink pilot signal to the base station, and the base station receives the signal, which includes the uplink radio channel;

[0021] S46, the user feeds back the downlink channel estimation results to the base station within the coherence time;

[0022] S47, the base station receives the downlink channel estimation result fed back by the user, and constructs a post-equalization matrix by using the transpose function of channel estimation;

[0023] S48, the base station multiplies the received signal by its own relative defect characteristic matrix, then multiplies the result by the post-equalization matrix, processes and simplifies the signal according to mathematical calculation and a channel model. The simplified result does not include the wireless channel, and the coupling influence of the wireless channel on the radio frequency fingerprint is eliminated;

[0024] S49, the base station extracts the radio frequency fingerprint characteristics of user equipment from the signal for authentication, the user performs LS channel estimation on the processed signal to obtain the defect characteristic matrix of the UE, and uses the characteristic matrix to identify different UEs for identity authentication.

[0025] Further, said step S41 comprises the following process: the base station constructs a database independently through self-transmission and self-reception; first, a reference antenna transmits a pilot to all other antennas of the device, and the other antennas receive the signal for channel estimation operation of the other antennas; then, the other antennas transmit pilots to the reference antenna, and the reference antenna receives the signal for channel estimation operation of the reference antenna; the relative defects of all antennas are obtained through the absolute defect of the reference antenna, and finally a normalized calibration matrix of the base station is obtained; a normalized calibration matrix of the user is obtained by adopting the same method.

[0026] Further, said step S41 specifically comprises the following sub-steps:

[0027] (1.1) selecting the first antenna of a base station BS as a reference antenna, and transmitting a pilot to the other m-1 antennas:

[0028] the pilot s is transmitted by the first base station antenna BS,1→m , 1<m≤M, the remaining m-1 antennas are in a listening state, and the pilot signal listened to is recorded as y BS,1→m , the expression thereof is:

[0029] y BS,1→m =h BS,1→m s BS,1→m +z BS,1→m ,

[0030] wherein is the channel from base station reference antenna 1 to base station antenna m, t BS,1 is the hardware gain of the first transmitting antenna of the base station, r BS,m is the hardware gain of the m-th receiving antenna of the base station, z BS,1→m is Gaussian white noise;

[0031] (1.2) the reference antenna performs a listening behavior, and the other m-1 antennas transmit pilots thereto:

[0032] A pilot signal S is transmitted from base station antenna m to base station antenna 1 BS,m→1 , where 1<m≤M, and the m-1 pilot signals received by the reference antenna are:

[0033] y BS,m→1 =h BS,m→1 s BS,m→1 +z BS,m→1 ,

[0034] wherein represents the channel from base station antenna m to base station reference antenna 1;

[0035] (1.3) Each antenna performs channel estimation according to the pilot signal it receives, obtaining:

[0036]

[0037]

[0038] (1.4) The base station radio frequency impairment characteristic matrix C is used BS to obtain:

[0039]

[0040] C BS =diag(c BS,1 , ……, c BS,M ),

[0041] wherein c BS,m represents the impairment of the m-th antenna, which is an element on the diagonal of the impairment matrix T BS represents the radio frequency impairment at the transmitting end of the base station, represents the radio frequency impairment at the receiving end of the base station, and the estimation results of each group of channels are divided to obtain the impairment relationship between the m-th antenna and the first reference antenna:

[0042]

[0043] wherein represents the absolute impairment of the first antenna, so the absolute impairment of the m-th antenna is expressed as:

[0044] c BS,m =c BS,1 c BS,m→1 ,

[0045] (1.5) A calibration matrix is obtained through normalization:

[0046] Assume that the absolute impairment of the first antenna is 1, that is, c BS,1 =1, then the absolute impairment of the m-th antenna is exactly its relative impairment, expressed as: c BS,m→1 =cBS,m Therefore, the relative characteristic matrix of a base station is defined as:

[0047]

[0048] The relative defect matrix obtained by the user equipment (UE) is similar to that of the base station equipment, where... The user's radio frequency defect feature matrix is ​​calculated as follows:

[0049]

[0050] Furthermore, in step S48, the signal processed according to mathematical calculations and the channel model is represented as follows:

[0051] Y = WC' BS Y U =WC' BS H U S U +WC' BS Z U

[0052] Simplifying Y yields the following equation:

[0053]

[0054] Where W is the post-equilibrium matrix. for The transpose function, To ensure that the UE feeds back the downlink channel estimation results to the BS within the coherent time, C' BS Y is the relative defect feature matrix of the base station (BS). U For the signal received by BS, H U For uplink wireless channel, S U Z is the uplink pilot signal. U For additive white Gaussian noise, c BS,1 This indicates the absolute defect of the first antenna.

[0055] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0056] 1. The radio frequency fingerprinting technology used in this invention, as an important non-cryptographic authentication technology, possesses unforgeable wireless radio frequency characteristics. Its core idea is to ensure the information security of wireless networks from an information theory perspective, rather than simply by increasing computational complexity. Radio frequency fingerprinting technology provides a method for authenticating device identity at the physical layer, without involving higher-level protocols, reducing message interaction and authentication message overhead between higher-level protocols, and effectively reducing authentication latency. Furthermore, the computation involved in this invention does not require high computing power, thus it can be deployed on low-performance devices with limited computing power, solving the problem that low-performance devices lack sufficient computational resources for asymmetric encryption authentication calculations. Therefore, this invention has the advantages of low latency, strong security, and lightweight design.

[0057] 2. This invention proposes for the first time the concept of relative radio frequency fingerprinting. It uses the characteristics of multiple antennas in a MIMO scenario to calculate the hardware mismatch relationship between multiple antennas as the radio frequency feature of the device. This feature is independent of location and can be used for classification and identification.

[0058] 3. This invention utilizes the relative hardware mismatch between different antennas as a feature of each UE. The BS and UE communicate using the protocol proposed herein, eliminating common channels through the incomplete reciprocity of the channels while retaining the features of the device to be authenticated. Authentication eliminates the coupling effect between the channel and the features. Under this authentication scheme, the features are not affected by the wireless channel and are not easily altered by the channel, exhibiting the advantage of feature stability. Therefore, stable radio frequency fingerprint features can be obtained, effectively improving identification accuracy. Attached Figure Description

[0059] Figure 1 This is an overall flowchart of base station user authentication according to an embodiment of the present invention;

[0060] Figure 2 This is a flowchart of a self-transmitting and self-receiving acquisition device relative to radio frequency fingerprints according to an embodiment of the present invention;

[0061] Figure 3 This is a schematic diagram of a base station reference antenna transmitting pilot signals according to an embodiment of the present invention;

[0062] Figure 4 This is a schematic diagram of a base station reference antenna receiving pilot signal according to an embodiment of the present invention;

[0063] Figure 5 This is a schematic diagram of the communication process according to an embodiment of the present invention;

[0064] Figure 6 This invention presents the recognition accuracy performance of different numbers of user devices under different signal-to-noise ratios in one embodiment of the present invention.

[0065] Figure 7This invention presents the recognition accuracy performance under different signal-to-noise ratios and different numbers of device antennas in one embodiment of the present invention. Detailed Implementation

[0066] The technical solutions provided by the present invention will be described in detail below with reference to specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention.

[0067] This invention provides an air interface access identification scheme based on radio frequency fingerprinting technology in MIMO scenarios. This scheme leverages the inherent uniqueness and unforgeability of radio frequency fingerprints, along with the characteristics of physical layer authentication, to ensure secure authentication of IoT devices. Consider a multi-antenna base station with M antennas in a certain area. Within the base station's coverage area, smart connected devices in different locations need to communicate with it. Assume there are K legitimate users, each with N antennas. Each device transmits information through a multipath channel. An attacker attempts to access the base station and launch an attack. To ensure secure communication between the base station and legitimate users, authentication of access devices is required. The system achieves ultra-low latency physical layer authentication while ensuring secure access.

[0068] This invention provides an air interface access identification method based on radio frequency fingerprinting technology in a MIMO scenario, which mainly includes the following steps:

[0069] S1, the system first simulates and generates several terminal wireless devices, which need to send data requests to the base station to communicate with the base station.

[0070] S2, Transmit Signal. The signal transmission function simulates the transmission process of the terminal device. The transmitted data is encoded, modulated, has pilots inserted, and defects added before reaching the base station receiver via the wireless channel. After adding receiver defects, the received signal is obtained.

[0071] S3, receive and separate the signal; the base station separates the pilot and data of the received signal.

[0072] S4, Feature Extraction and Recognition: The pilot section obtains the transmitter's radio frequency (RF) characteristics through the RF fingerprint extraction module, and then the device identification result is obtained through the classification module. In terms of RF fingerprint design, this invention, utilizing the multi-antenna characteristics of devices in MIMO scenarios, proposes for the first time the concept of relative RF fingerprints. It calculates the hardware mismatch relationship between multiple antennas as the device's RF characteristics and performs classification and recognition based on these characteristics.

[0073] Specifically, the process for this step is as follows: Figure 1 As shown, it includes the following sub-steps:

[0074] (1) The device performs self-transmitting self-receiving independent calibration matrix construction: first, the reference antenna transmits pilots to all other antennas of the device, and the other antennas receive the signals for channel estimation operation of the other antennas; after that, the other antennas transmit pilots to the reference antenna, and the reference antenna receives the signals for channel estimation operation of the reference antenna. Thus all antennas can obtain corresponding absolute impairments, the relative impairments of all antennas can be obtained through the absolute impairments of the reference antenna, and finally a normalized calibration matrix, that is, the relative impairment matrix, is obtained.

[0075] Step (1) comprises the following sub-steps, as Figure 2 shown, taking base station equipment as an example:

[0076] (1.1) Select the first antenna of base station BS as the reference antenna, and transmit pilots to the other m-1 antennas:

[0077] as Figure 3 shown, the first base station antenna transmits pilot s BS,1→m (1<m≤M), the remaining m-1 antennas are in a listening state, and the received pilot signal is recorded as y BS,1→m , the expression of which is:

[0078] y BS,1→m =h BS,1→m s BS,1→m +z BS,1→m ,

[0079] wherein is the channel from the base station reference antenna 1 to the base station antenna m, t BS,1 represents the hardware gain of the first transmitting antenna of the base station, r BS,m represents the hardware gain of the m-th receiving antenna of the base station, z BS,1→m is Gaussian white noise;

[0080] (1.2) The reference antenna performs listening, and the other m-1 antennas transmit pilots to it:

[0081] as Figure 4 shown, the base station antenna m transmits pilot signal S BS,m→1 (1<m≤M), the reference antenna (that is, the first antenna) receives m-1 pilot signals as follows:

[0082] y BS,m→1 =h BS,m→1 s BS,m→1 +z BS,m→1 ,

[0083] wherein is the channel from the base station antenna m to the base station reference antenna 1;

[0084] (1.3) Each antenna makes a channel estimate based on the pilot signal it receives, resulting in:

[0085]

[0086]

[0087] (1.4) Using the base station radio frequency defect feature matrix C BS ,have to:

[0088]

[0089] C BS =diag(c BS,1 ,……,c BS,M ),

[0090] Where c BS,m The defect of the m-th antenna is represented by the elements on the diagonal of the defect matrix. T BS This indicates a radio frequency defect at the base station's transmitting end. These represent the radio frequency defects at the base station receiver, and they are all diagonal arrays. Dividing the estimation results of each group of channels yields the defect relationship between the m-th antenna and the first reference antenna:

[0091]

[0092] in Let m represent the absolute defect of the first antenna. Therefore, the absolute defect of the m-th antenna is represented as:

[0093] c BS,m =c BS,1 c BS,m→1 .

[0094] (1.5) Normalization yields the calibration matrix:

[0095] Assume the absolute defect of the first antenna is 1, i.e., c BS,1 =1, then the absolute defect of the m-th antenna is its relative defect, expressed as: c BS,m→1 =c BS,m Therefore, the relative characteristic matrix of a base station is defined as:

[0096]

[0097] The relative defect matrix obtained by the user equipment (UE) is similar to that of the base station equipment, where... The user's radio frequency defect feature matrix is ​​calculated as follows:

[0098]

[0099] (2) Start interactive communication. The communication process is illustrated as follows: Figure 5 As shown, the BS first sends a downlink pilot signal S to the UE. D ;

[0100] (3) The UE receives a signal of Y. D =H D S D +Z D Z D It is additive white Gaussian noise, following a Gaussian distribution, H D For downlink wireless channels;

[0101] (4) UE according to Y D Channel estimation is performed, and the calculation method for channel estimation is as follows:

[0102]

[0103] in, For S D The pseudo-inverse matrix;

[0104] (5) During the coherence time, the UE sends the uplink pilot signal S to the BS. U BS received signal Y U =H U S U +Z U Z U It is additive white Gaussian noise, following a Gaussian distribution, H U For uplink wireless channel;

[0105] This invention designs a Signal Backhaul Protocol (SEP), with the base station and the user as the main components. The user sends the channel estimation result back to the base station. The base station processes the received signal through a series of operations to obtain the user's relative radio frequency fingerprint characteristics. The base station then compares these characteristics with fingerprints in a fingerprint database to identify the user equipment. Specifically:

[0106] (6) The UE feeds back the downlink channel estimation results to the BS within the coherence time.

[0107] (7) The BS receives feedback from the UE. use transpose function The post-equilibrium matrix is ​​constructed and calculated as follows:

[0108]

[0109] (8) BS multiplies the received signal by its own relative defect feature matrix C'. BS After multiplying by the equalization matrix, the processed signal is obtained, which is represented as follows:

[0110] Y = WC' BS Y U =WC' BS H U S U +WC' BS Z U ,

[0111] Simplifying Y yields the following equation:

[0112]

[0113] (9) Because R UE With T UE Since it is a diagonal matrix, we can obtain the following equation:

[0114]

[0115] Substituting the result into Y, we get the following formula:

[0116]

[0117] Therefore, the calibration matrix of the user UE defect characteristics is obtained as follows:

[0118]

[0119] Normalization is performed using the first value of the defect feature matrix as a benchmark to obtain the defect feature matrix C' of the UE. UE This feature matrix contains the UE's radio frequency fingerprint features. Using this feature matrix C' UE It can identify different user interfaces (UEs) and perform authentication.

[0120] We use the relative hardware mismatch relationship of the device antenna chain as the feature of each user equipment. Then, the base station and the user equipment communicate using the above process. First, the influence of the wireless channel on the radio frequency fingerprint feature is eliminated. Then, the base station obtains the channel estimate and extracts the fingerprint feature to identify the user. By matching the feature with the feature in the fingerprint database, it is determined whether the current authenticated user is a legitimate user.

[0121] like Figure 6 As shown, under the TDL-C channel, when the number of users is less than 20 and the signal-to-noise ratio (SNR) is greater than 15 dB, the identification accuracy of the R-RFF scheme can reach 97.8%, and it can maintain above 91.4% even at 5 dB. The identification accuracy decreases as the number of users increases. When the number of users is 30 and the SNR is greater than 10 dB, the identification accuracy for 30 users can reach over 93.9%. This demonstrates that the proposed R-RFF method is suitable for user classification in multi-antenna devices with four antennas.

[0122] like Figure 7 The figure shows the identification accuracy for users with different numbers of antennas. It can be seen that, because the proposed R-RFF scheme relies on the RF defect relationship between the device's antenna chains, the identification accuracy at four antennas is higher than at two antennas. In a dual-antenna device, one antenna serves as the reference antenna, meaning only the RF defect relationship of one antenna can be used as a classification feature, leading to poorer classification for dual-antenna devices. However, this also indicates that as the number of terminal antennas increases, this scheme relying on the relationship between antennas will become increasingly effective with the widespread adoption of antennas in the future.

[0123] S5, recover the original data. The data part is demodulated and decoded to obtain the original data that was sent.

[0124] The technical means disclosed in this invention are not limited to those disclosed in the above embodiments, but also include technical solutions composed of any combination of the above technical features. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications are also considered within the scope of protection of this invention.

Claims

1. A method for air interface access identification based on radio frequency fingerprinting in a MIMO scenario, characterized in that, Includes the following steps: S1, establish a communication architecture between a multi-antenna base station and several terminal wireless devices. Each terminal wireless device needs to send a data request to the base station to communicate with the base station. S2, the terminal wireless device sends a signal. The transmitted data is encoded, modulated, inserted with pilots, and has defects added before reaching the base station receiver section via the wireless channel. After adding receiver defects, the received signal is obtained. S3, receive and separate the signal; the base station separates the pilot and data of the received signal. S4, Feature Extraction and Recognition: The pilot section obtains the transmitter's radio frequency characteristics through the radio frequency fingerprint extraction module, and then obtains the device identification result through the classification module; including the following sub-steps: S41, establish normalized calibration matrices for base stations and user equipment respectively; S42, the base station first sends downlink pilot signals to the user; S43, the user receives a signal that includes a downlink radio channel; S44, the user performs channel estimation based on the received signal; S45, within the coherent time, the user sends an uplink pilot signal to the base station, and the base station receives the signal, which includes the uplink radio channel; S46, the user feeds back the downlink channel estimation results to the base station within the coherence time; S47, the base station receives the downlink channel estimation result from the user and constructs the post-equalization matrix using the transpose function of the channel estimation; S48, the base station multiplies the received signal by its own relative defect feature matrix, and then multiplies it by the subsequent equalization matrix. Based on mathematical calculations and channel models, the signal is processed and simplified. The simplified result does not include the wireless channel, thus eliminating the coupling effect of the wireless channel on the radio frequency fingerprint. S49, the base station extracts the radio frequency fingerprint features of the user equipment from the signal for authentication. The user performs LS channel estimation on the processed signal to obtain the defect feature matrix of the UE. The feature matrix is ​​used to identify different UEs and perform authentication. S5, recover the original data. The data part is demodulated and decoded to obtain the original data that was sent.

2. The air interface access identification method based on radio frequency fingerprinting in a MIMO scenario according to claim 1, characterized in that, Step S41 includes the following process: The base station independently builds its own database. First, the reference antenna sends pilot signals to all other antennas of the device. The other antennas receive the signals and use them for channel estimation calculations. Then, other antennas send pilot signals to the reference antenna, which receives the signals and uses them for channel estimation calculations. The relative defects of all antennas are obtained by using the absolute defects of the reference antenna, and finally the normalized calibration matrix of the base station is obtained; the normalized calibration matrix of the user is obtained by using the same method.

3. The air interface access identification method based on radio frequency fingerprinting in a MIMO scenario according to claim 2, characterized in that, Step S41 specifically includes the following sub-steps: (1.1) Select the first antenna of base station BS as the reference antenna and send pilot signals to the other m-1 antennas: Pilot signals are transmitted from the first base station antenna. The remaining m-1 antennas are in listening mode, and the detected pilot signal is recorded as... Its expression is: , in This represents the channel from base station reference antenna 1 to base station antenna m. This indicates the hardware gain of the base station's first transmitting antenna. This represents the hardware gain of the m-th receiving antenna at the base station. It is Gaussian white noise; (1.2) The reference antenna performs a listening action, and the other m-1 antennas send pilot signals to it: Pilot signals are sent from base station antenna m to base station antenna 1. The reference antenna receives m-1 pilot signals as follows: , in , which represents the channel from base station antenna m to base station reference antenna 1; (1.3) Each antenna makes a channel estimate based on the pilot signal it receives, resulting in: , (1.4) Using the base station radio frequency defect feature matrix ,have to: , in The defect of the m-th antenna is represented by the elements on the diagonal of the defect matrix. , This indicates a radio frequency defect at the base station's transmitting end. To represent the radio frequency defects at the base station receiver, the estimated results for each group of channels are divided to obtain the defect relationship between the m-th antenna and the first reference antenna: , in Let represent the absolute defect of the first antenna, then the absolute defect of the m-th antenna is expressed as: , (1.5) Normalization yields the calibration matrix: Assume the absolute defect of the first antenna is 1, that is Then the absolute defect of the m-th antenna is its relative defect, expressed as: Therefore, the relative characteristic matrix of a base station is defined as: ； The relative defect matrix obtained by the user equipment (UE) is similar to that of the base station equipment, where... The user's radio frequency defect feature matrix is ​​calculated as follows: 。 4. The air interface access identification method based on radio frequency fingerprinting in a MIMO scenario according to claim 1, characterized in that, In step S48, the signal processed according to mathematical calculations and the channel model is represented as follows: right After simplification, we get the following equation: in The post-equilibrium matrix, , for The transpose function, The UE feeds back the downlink channel estimation results to the BS within the coherent time. This is the relative defect feature matrix of the base station (BS). The signal received by BS For uplink wireless channel, Uplink pilot signal It is additive white Gaussian noise. This indicates the absolute defect of the first antenna.

Images