A method for detecting interference attacks in uplink low-precision ADC MIMO systems

Abstract

The present invention discloses an interference attack detection method for an uplink low-precision ADC MIMO system, the process method includes firstly linearizing the low-precision ADC uniform quantization process by Bussgang decomposition; under the assumption of no interference, iteratively solving the maximum likelihood estimation value of the legal channel by using the MM algorithm; under the assumption of interference, respectively solving the maximum likelihood estimation value of the legal channel and the interference correlation matrix by using the MM algorithm; obtaining the interference attack detection amount by using the generalized likelihood ratio test criterion, comparing the detection threshold with the detection amount, and judging whether the uplink MIMO system is attacked by blocking interference. Compared with the existing high-precision ADC MIMO system interference attack detection technology, the present invention is suitable for more complex low-precision ADC MIMO system scenarios, and can achieve excellent interference attack detection performance with lower complexity.

Description

Technical Field

[0001] The present invention belongs to the field of uplink MIMO communication, and particularly relates to a method for detecting interference attacks in an uplink low-precision ADC MIMO system. Background Art

[0002] The fifth-generation mobile communication technology requires base stations to be equipped with a sufficient number of antennas to meet the increasing demand for data throughput. The diversity and multiplexing technologies of MIMO systems can effectively combat channel fading and improve the transmission rate. At the same time, beamforming technology can also be used at the receiving and transmitting ends to control the directivity of antennas, thereby suppressing interference in other directions. However, in the uplink MIMO system, a large receiving antenna array is equipped with high-precision ADCs, which will cause huge energy losses during use. To solve this problem, wireless communication systems use low-precision ADCs with lower power consumption to quantify received signals. Therefore, the interference detection and suppression technology under low-precision ADCs has become a technology worthy of research.

[0003] Compared with high-precision ADCs, although low-precision ADCs can effectively improve energy utilization, the non-linear distortion of the quantized signals is more serious, which poses certain challenges to signal detection. For the problem of interference attack detection in the uplink MIMO system, existing interference attack detection technologies do not consider the impact of non-linear distortion caused by low-precision ADCs. Finding a feasible and well-performing adaptive interference attack detection method under low-precision ADCs has become an urgent problem to be solved. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for detecting interference attacks in an uplink low-precision ADC MIMO system. For the uplink low-precision ADC MIMO system, considering the correlation of interference signals, the legitimate channel and interference correlation matrix are estimated by the MM algorithm to achieve the interference detection function.

[0005] To solve the above technical problems, the specific technical solution of the present invention is as follows:

[0006] A method for detecting interference attacks in an uplink low-precision ADC MIMO system includes the following steps:

[0007] Step S1: Linearize the uniform quantization process of the low-precision analog-to-digital converter (ADC) by using Bussgang decomposition;

[0008] Step S2: Using the result under the no-interference assumption in Step S1 as the received signal model, solve the maximum likelihood estimate value of the legitimate channel by using the MM algorithm;

[0009] Step S2-1: Construct an initialization problem, and under the condition of no interference, take maximizing the log-likelihood function as the optimization objective;

[0010] Step S2-2: Obtain an upper bound of the objective function applicable to the MM algorithm according to the optimization problem determined in Step S2-1.

[0011] Step S2-3: Take minimizing the upper bound determined in Step S2-2 as the optimization objective, and iteratively solve the closed-form maximum likelihood estimate of the legitimate channel until convergence.

[0012] Step S3: Take the result under the interference hypothesis in Step S1 as the received signal model, and use the MM algorithm to separately solve the maximum likelihood estimates of the legitimate channel and the interference correlation matrix.

[0013] Step S3-1: Construct an initialization problem, and take maximizing the log-likelihood function as the optimization objective under the condition of interference.

[0014] Step S3-2: Obtain an upper bound of the objective function applicable to the MM algorithm according to the optimization problem determined in Step S3-1.

[0015] Step S3-3: Take minimizing the upper bound determined in Step S3-2 as the optimization objective, solve the closed-form maximum likelihood estimate of the legitimate channel in each iteration, and obtain the maximum likelihood estimate of the interference correlation matrix by solving a convex problem.

[0016] Step S4: Use the generalized likelihood ratio test criterion to obtain the interference attack detection quantity, and compare it with the detection threshold. If it is greater than the detection threshold, it is determined that the uplink MIMO system is under an interference attack; if it is less than the detection threshold, it is determined that the uplink MIMO system is not under an interference attack.

[0017] Furthermore, in Step S1, the Bussgang decomposition expression of the low-precision ADC uniform quantization process is

[0018]

[0019] In formula (1), H 1 represents the interference hypothesis, and H 0 represents the interference-free hypothesis; the superscript (·) T represents the transpose of a vector or matrix; Y is a complex matrix with M rows and N columns, representing the quantized pilot received signal, where M represents the number of base station receiving antennas and N represents the pilot symbol length; U is a complex matrix with M rows and N columns, representing the interference signal in the pilot stage; N is a complex matrix with M rows and N columns, representing the noise in the pilot stage, and its elements are independent and identically distributed with zero mean and variance of N 0Complex Gaussian random variable; ε is an M×N complex matrix representing quantization noise; h is a complex vector of length M representing the legitimate channel; s is a complex vector of length N representing pilot symbols; q(·) represents the optimal uniform quantization function, which quantizes the real and imaginary parts of each scalar element in the matrix respectively, and its expression is:

[0020]

[0021] In formula (2), Δ is the minimum distortion quantization step of the low-precision ADC, and the relationship between Δ and the number of bits b of the quantizer is as follows: when b = 1, Δ is 1.596; when b = 2, Δ is 0.9957; when b = 3, Δ is 0.586; when

[0022] b = 4, Δ is 0.3352; when b = 5, Δ is 0.1881;

[0023] L = 2 b , representing the number of quantization regions; a represents the Bussgang decomposition coefficient, and its expression is:

[0024]

[0025] Furthermore, in step S2-1, the optimization objective of maximizing the log-likelihood function under the condition of no interference is

[0026] Maximize lnf(Y|h 0 ,H 0 ) (4)

[0027] In formula (4), ln(·) represents the natural logarithm function with the natural constant e as the base; f(Y|h 0 ,H 0 ) represents the likelihood function of the received signal Y given the channel h 0 , under the condition of no interference, and its expression is

[0028]

[0029] In formula (5), y n represents the nth column element of the received signal Y; s n represents the nth element of the pilot signal s; A 0 is represented as A 0 = diag(h 0 ); diag(·) represents a diagonal matrix with the elements of the input vector as the diagonal elements; the superscript (·) H represents the conjugate transpose of a vector or matrix; the superscript (·) -1 represents the inverse of a matrix; exp(·) represents the exponential function with the natural constant e as the base; |·| represents the determinant of a matrix;

[0030] In the step S2-2, the upper bound expression of the objective function applicable to the MM algorithm is

[0031]

[0032] In formula (6), g t+1 (·) represents the upper bound of the objective function at the (t + 1)-th iteration, and h 0,t represents the estimated value of h 0 obtained by solving the MM algorithm at the t-th iteration, where E t , M t (n) and const 0 are expressed as

[0033]

[0034] In formula (7), I M represents the M-dimensional identity matrix; tr{·} represents the trace of a matrix; A 0,t is expressed as A 0,t = diag(h 0,t );

[0035] In the step S2-3, taking the minimization of the upper bound determined in the step S2-2 as the optimization objective, the expression is

[0036] Minimize g t+1 (h 0 |h 0,t , H 0 ) (8)

[0037] The expression of the legitimate channel estimate value solved in the (t + 1)-th iteration is

[0038]

[0039] Furthermore, in the step S3-1, the optimization objective of maximizing the log-likelihood function under the interference condition is

[0040] Maximize lnf(Y|h 1 , Σ u , H 1 ) (10)

[0041] In formula (10), Σ u represents the interference correlation matrix; f(Y|h 1 , Σ u , H 1 ) represents the likelihood function of the received signal Y given the legitimate channel h 1 and the interference correlation matrix Σ u , and the expression is

[0042]

[0043] In formula (11), A 1 is represented as A 1 = diag(h 1 ); is represented as a diagonal matrix composed of the diagonal elements of the input matrix;

[0044] In the step S3-2, the upper bound expression of the objective function applicable to the MM algorithm is

[0045]

[0046] In formula (12), h 1,t represents the estimated value of h 1 obtained by solving the MM algorithm in the t-th iteration; A 1,t is represented as A 1,t = diag(h 1,t ); Σ u,t represents the estimated value of Σ u obtained by the MM algorithm in the t-th iteration; F t , L t (n), const 1 The expression of is

[0047]

[0048] In formula (13), represents the estimated value of obtained by the MM algorithm in the t-th iteration; ||·|| 2 represents the squared norm of a vector;

[0049] In the step S3-3, with minimizing the upper bound determined in the step S3-2 as the optimization objective, the expression is

[0050] Minimize g t+1 (h 1 , Σ u |h 1,t , Σ u,t , H 1 ) (14)

[0051] The constraint condition for the upper bound minimization optimization problem is

[0052] Σ u ≥ 0 (15)

[0053] The expression for the legal channel estimate value solved in the (t + 1)-th iteration is

[0054]

[0055] The interference correlation matrix Σ for the (t + 1)-th iterative solution u,t+1 is obtained by solving the convex problem of formula (14).

[0056] Furthermore, in step S4, the interference attack detection metric expression obtained using the generalized likelihood ratio test criterion is

[0057]

[0058] In formula (17), and respectively represent the maximum likelihood estimates of the legitimate channel and the interference correlation matrix obtained by the iterative convergence of the MM algorithm under the interference hypothesis H 1 ; represents the maximum likelihood estimate of the legitimate channel obtained by the iterative convergence of the MM algorithm under the interference-free hypothesis H 0 ; γ represents the interference detection threshold; formula (17) represents judging the magnitude relationship between the interference detection metric and the detection threshold. If it is greater than the detection threshold, it is judged that the uplink MIMO system is under an interference attack. If it is less than the detection threshold, it is judged that the uplink MIMO system is not under an interference attack.

[0059] Furthermore, in step S4, the interference detection threshold γ satisfies a given false alarm probability and is obtained through offline repeated independent experiments.

[0060] An interference attack detection method for an uplink low-precision ADC MIMO system of the present invention has the following advantages:

[0061] 1. Compared with the existing interference attack detection technologies for high-precision ADC MIMO systems, the present invention develops an interference attack detection technology for MIMO under low-precision ADC. Specifically, different from the existing high-precision ADC signal models, the present invention constructs a linear quantized received signal model in step S1, providing an algorithm design basis for the interference detection technology of low-precision ADC. Low-precision ADC is widely used in large-scale MIMO systems to reduce the hardware cost and power consumption of the system. Therefore, the present invention is more suitable for large-scale MIMO systems.

[0062] 2. The implementation complexity of the present invention is relatively low. The interference attack detection threshold does not need to be manually adjusted and can adaptively change the threshold value for different scenarios, which is beneficial for engineering applications. Specifically, in actual use, the present invention can pre-train the interference detection threshold values for different communication scenarios and select the threshold value matching the current scenario in step S4 to complete adaptive interference detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic flow diagram of an interference attack detection method for an uplink low-precision ADC MIMO system provided in Embodiment 1 of the present invention.

[0064] Figure 2 Simulation experimental result diagram of an interference attack detection method for an uplink low-precision ADC MIMO system provided in Embodiment 1 of the present invention. Detailed implementation manners

[0065] To better understand the purpose, structure and function of the present invention, the following further describes in detail an interference attack detection method for an uplink low-precision ADC MIMO system of the present invention with reference to the accompanying drawings.

[0066] Embodiment 1

[0067] Refer to Figure 1 and Figure 2 , this embodiment provides an interference attack detection method for an uplink low-precision ADC MIMO system, and this method is applied to the uplink pilot phase, that is, a legitimate user sends a signal pilot signal. First, linearize the quantization process of the low-precision ADC, and then solve the channel estimation value and interference estimation value under two hypotheses of having interference and no interference through the MM algorithm. Finally, construct an interference detection quantity and compare it with a detection threshold to determine whether there is an interference attack behavior.

[0068] Specifically, in this embodiment, the flow of this interference attack detection method is as Figure 1 shown, and the specific steps are as follows:

[0069] Step 1: Linearize the uniform quantization process of the low-precision ADC by using Bussgang decomposition, and the expression is

[0070] H 1 : Y = q(hs T + U + N) = ahs T + aU + aN + ε

[0071] H 0 : Y = q(hs T + N) = ahs T + aN + ε

[0072] where, H 1 represents the hypothesis of having interference, and H 0 represents the hypothesis of no interference; the superscript (·) Trepresents the transpose of a vector or matrix; Y is a complex matrix with M rows and N columns, representing the quantized pilot reception signal, where M represents the number of base station receiving antennas and N represents the pilot symbol length; U is a complex matrix with M rows and N columns, representing the interference signal in the pilot stage; N is a complex matrix with M rows and N columns, representing the noise in the pilot stage, whose elements are independent and identically distributed with zero mean and variance of N 0 is a complex Gaussian random variable; ε is a complex matrix with M rows and N columns, representing quantization noise; h is a complex vector of length M, representing a legal channel; s is a complex vector of length N, representing a pilot symbol; q(·) represents the optimal uniform quantization function, which quantizes the real and imaginary parts of each scalar element in the matrix separately, and its expression is:

[0073]

[0074] Where Δ is the minimum distortion quantization step of the low-precision ADC, and the relationship between Δ and the number of quantizer bits b is shown in Table 1; L = 2 b , represents the number of quantized regions; a represents the Bussgang decomposition coefficient, the expression is:

[0075]

[0076] Table 1. Comparison table of optimal quantization step size Δ and uniform quantization bit number b

[0077] b 1 2 3 4 5 Δ 1.596 0.9957 0.586 0.3352 0.1881

[0078] Step 2: Take the result of step S1 under the assumption of no interference as the received signal model and use the MM algorithm to solve the maximum likelihood estimate of the legal channel. First, construct the initialization problem. Under the condition of no interference, maximize the log-likelihood function as the optimization goal. The expression is

[0079] Maximize lnf(Y|h 0 ,H 0 )

[0080] where ln(·) represents the logarithmic function with e as the base, ln(·) represents the logarithmic function with e as the base, f(Y|h 0 ,H 0 ) indicates that under interference-free conditions, given channel h 0 , the likelihood function of the received signal Y is expressed as

[0081]

[0082] Where y n represents the nth column element of the received signal Y; s n represents the nth element of the pilot signal s; A 0 denoted as A 0 =diag(h0 );diag(·) represents a diagonal matrix with the elements of the input vector as its diagonal elements; the superscript (·) H represents the conjugate transpose of a vector or matrix; the superscript (·) -1 represents the inverse of a matrix; exp(·) represents the exponential function with the natural constant e as the base; |·| represents the determinant of a matrix. The upper bound expression of the objective function applicable to the MM algorithm is

[0083]

[0084] where g t+1 (·) represents the upper bound of the objective function at the (t + 1)-th iteration, h 0,t represents the estimated value of h obtained by solving the MM algorithm at the t-th iteration, where E 0 , M t , and the expressions of const t are 0 as follows

[0085]

[0086]

[0087]

[0088] where I M represents the M-dimensional identity matrix; tr{·} represents the trace of a matrix; A 0,t is represented as A 0,t = diag(h 0,t ). With the optimization objective of minimizing the upper bound, the expression is

[0089] Minimize g t+1 (h 0 |h 0,t , H 0 )

[0090] The expression for the legitimate channel estimate value solved in the (t + 1)-th iteration of this problem is

[0091]

[0092] Step 3: Using the result under the interference assumption in Step S1 as the received signal model, use the MM algorithm to solve the maximum likelihood estimate values of the legitimate channel and the interference correlation matrix respectively. First, construct an initialization problem. Under the condition of interference, with the optimization objective of maximizing the log-likelihood function:

[0093] Maximize lnf(Y|h 1 , Σ u , H 1 )

[0094] where Σ u denotes the interference correlation matrix. f(Y|h 1 , Σ u , H 1 ) represents the likelihood function of the received signal Y given the legitimate channel h 1 and the interference correlation matrix Σ u , and the expression is

[0095]

[0096] where, A 1 is denoted as A 1 = diag(h 1 ), is denoted as a diagonal matrix composed of the diagonal elements of the input matrix. The upper bound expression of the objective function applicable to the MM algorithm is

[0097]

[0098] where h 1,t represents the estimated value of h 1 obtained by the MM algorithm in the t-th iteration; A 1,t is denoted as A 1,t = diag(h 1,t ); Σ u,t represents the estimated value of Σ u obtained by the MM algorithm in the t-th iteration; F t , L t (n), const 1 has the expression

[0099]

[0100]

[0101] L t (n) = as n I M + diag(v t )

[0102]

[0103] where,[[]] represents the estimated value of obtained by the MM algorithm in the t-th iteration; ||·|| 2 represents the squared norm of the vector. With minimizing the upper bound as the optimization objective, the expression is

[0104] Minimize g t+1 (h1 , Σ u |h 1,t , Σ u,t , H 1 )

[0105] The constraints for minimizing the upper bound optimization problem are

[0106] Σ u ≥ 0

[0107] The closed - form solution for the (t + 1)-th iteration of the legitimate channel is

[0108]

[0109] The solution Σ of the interference correlation matrix for the (t + 1)-th iteration u,t+1 can be obtained by using the MATLAB CVX tool.

[0110] Step 4: Using the generalized likelihood ratio test criterion, the expression for the interference attack detection metric is

[0111]

[0112] where and respectively represent the maximum likelihood estimates of the legitimate channel and the interference correlation matrix obtained by the MM algorithm iteratively converging under the interference hypothesis (H 1 ); represents the maximum likelihood estimate of the legitimate channel obtained by the MM algorithm iteratively converging under the no - interference hypothesis (H 0 ). γ represents the interference detection threshold, which should satisfy the given false - alarm probability and can be obtained through offline repeated independent experiments. The above formula represents judging the relationship between the interference detection metric and the detection threshold. If it is greater than the detection threshold, it is judged that the uplink MIMO system is under interference attack; if it is less than the detection threshold, it is judged that the uplink MIMO system is not under interference attack;

[0113] This embodiment considers the pilot period of the uplink MIMO system. In order to verify the interference attack detection effect provided by this embodiment, Monte Carlo simulation experiments are carried out. The parameters involved in the simulation experiments are shown in the following table:

[0114] Table 2: Simulation experiment parameter table

[0115]

[0116]

[0117] Specifically, Table 2 is the parameter table of the simulation experiment. The simulation computer is configured with an i5-10400 CPU@2.00GHz and 8.00GB of memory. Where the present invention is not described in detail, it is the well-known technology in the art. Figure 2 The following is the simulation result, which shows that: when the interference signal power ratio (JSR) and other uplink MIMO system parameters are fixed, the higher the signal-to-noise ratio, the greater the probability of correct detection.

[0118] It can be understood that the present invention is described by some embodiments. Those skilled in the art know that, without departing from the spirit and scope of the present invention, various changes or equivalent substitutions can be made to these features and embodiments. Additionally, under the teaching of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application belong to the scope protected by the present invention.

Claims

1. A method for detecting interference attack in an uplink low-precision ADC MIMO system, characterized in that: The steps include: Step S1, linearizing the uniform quantization process of the low-precision analog-to-digital converter ADC by using Bussgang decomposition; Step S2, using the result under the assumption of no interference in step S1 as the received signal model, and using the MM algorithm to solve the maximum likelihood estimate of the legal channel; Step S2-1, constructing an initialization problem, taking maximizing the log-likelihood function as the optimization goal under interference-free conditions; Step S2-2, according to the optimization problem determined in step S2-1, obtaining an upper bound of the objective function applicable to the MM algorithm; Step S2-3, taking minimizing the upper bound determined in step S2-2 as the optimization goal, iteratively solving the closed form maximum likelihood estimate of the legal channel until convergence; Step S3, using the result of step S1 under the assumption of interference as the received signal model, and using the MM algorithm to solve the maximum likelihood estimation values ​​of the legal channel and the interference correlation matrix respectively; Step S3-1, constructing an initialization problem, with maximizing the log-likelihood function as the optimization goal under interference conditions; Step S3-2, according to the optimization problem determined in step S3-1, obtain the upper bound of the objective function applicable to the MM algorithm; Step S3-3, taking minimization of the upper bound determined in step S3-2 as the optimization goal, solving the closed form maximum likelihood estimate of the legal channel in each iteration, and obtaining the maximum likelihood estimate of the interference correlation matrix by solving the convex problem; Step S4: obtain the interference attack detection value by using the generalized likelihood ratio test criterion, and compare it with the detection threshold. If it is greater than the detection threshold, it is determined that the uplink MIMO system is under interference attack; if it is less than the detection threshold, it is determined that the uplink MIMO system is not under interference attack.

2. The interference attack detection method for uplink low-precision ADC MIMO system according to claim 1, characterized in that: In step S1, the Bussgang decomposition expression of the low-precision ADC uniform quantization process is: In formula (1), H1 represents the interference hypothesis, H0 represents the no interference hypothesis; the superscript (·) T represents the transpose of a vector or matrix; Y is a complex matrix of M rows and N columns, representing the quantized pilot reception signal, where M represents the number of base station receiving antennas and N represents the pilot symbol length; U is a complex matrix with M rows and N columns, representing the interference signal in the pilot phase; N is a complex matrix with M rows and N columns, representing the noise of the pilot phase, whose elements are independent and identically distributed complex Gaussian random variables with zero mean and variance N0; ε is a complex matrix of M rows and N columns, representing quantization noise; h is a complex vector of length M, representing a legal channel; s is a complex vector of length N, representing the pilot symbol; q(·) represents the optimal uniform quantization function, which quantizes the real and imaginary parts of each scalar element in the matrix separately, and its expression is: In formula (2), Δ is the minimum distortion quantization step of the low-precision ADC, and the relationship between Δ and the number of quantizer bits b is as follows: when b=1, Δ is 1.596; when b=2, Δ is 0.9957; when b=3, Δ is 0.586; when b=4, Δ is 0.3352; when b=5, Δ is 0.1881; L=2 b , represents the number of quantized regions; a represents the Bussgang decomposition coefficient, and the expression is:

3. The interference attack detection method for uplink low-precision ADC MIMO system according to claim 2, characterized in that: In step S2-1, the optimization goal of maximizing the log-likelihood function under interference-free conditions is Maximize lnf(Y|h0,H0)(4) In formula (4), ln(·) represents the logarithmic function with the natural constant e as the base; f(Y|h0,H0) represents the likelihood function of the received signal Y under interference-free conditions and given channel h0, which is expressed as In formula (5), y n represents the nth column element of the received signal Y; s n represents the nth element of the pilot signal s; A0 is represented by A0=diag(h0); diag(·) represents a diagonal matrix whose diagonal elements are the elements of the input vector; superscript (·) H Indicates the conjugate transpose of a vector or matrix; superscript (·) -1 represents the inverse of a matrix; exp(·) represents an exponential function with the natural constant e as the base; |·| represents the determinant of a matrix; In step S2-2, the upper bound expression of the objective function applicable to the MM algorithm is: In formula (6), g t+1 (·) represents the upper bound of the objective function of the t+1th iteration, h 0,t represents the estimated value of h0 obtained by the MM algorithm at the tth iteration, where E t , M t (n) and const0 are expressed as In formula (7), I M represents the M-dimensional unit matrix; tr{·} represents the trace of the matrix; A 0,t Represented as A 0,t =diag(h 0,t ); In step S2-3, the optimization goal is to minimize the upper bound determined in step S2-2, and the expression is: Minimize g t+1 (h0|h 0,t ,H0) (8) The legal channel estimation value expression solved in the t+1th iteration is:

4. The interference attack detection method for uplink low-precision ADC MIMO system according to claim 3, characterized in that: In step S3-1, the optimization goal of maximizing the log-likelihood function under interference conditions is Maximize lnf(Y|h1,Σ u ,H1)(10) In formula (10), Σ u represents the interference correlation matrix; f(Y|h1,Σ u ,H1) means that under the condition of interference, given the legal channel h1 and the interference correlation matrix Σ u , the likelihood function of the received signal Y is expressed as In formula (11), A1 is expressed as A1=diag(h1); Expressed as represents a diagonal matrix consisting of the diagonal elements of the input matrix; In step S3-2, the upper bound expression of the objective function applicable to the MM algorithm is: In formula (12), h 1,t represents the estimated value of h1 obtained by the MM algorithm at the tth iteration; A 1,t Represented as A 1,t =diag(h 1,t );Σ u,t represents the Σ obtained by the MM algorithm at the tth iteration u The estimated value of F t , L t (n), the expression of const1 is In formula (13), It means that the MM algorithm obtains Estimated value of ;||·|| 2 represents the square of the magnitude of a vector; In step S3-3, the optimization goal is to minimize the upper bound determined in step S3-2, and the expression is: Minimize g t+1 (h1,Σ u |h 1,t ,Σ u,t ,H1) (14) The constraints of the optimization problem of minimizing the upper bound are S u ≥0 (15) The legal channel estimation value expression solved in the t+1th iteration is: The interference correlation matrix Σ solved in the t+1th iteration is u,t+1 It is obtained by solving the convex problem of formula (14).

5. The interference attack detection method for uplink low-precision ADC MIMO system according to claim 4, characterized in that: In step S4, the interference attack detection quantity expression obtained by using the generalized likelihood ratio test criterion is: In formula (17), and They represent the maximum likelihood estimates of the legal channel and interference correlation matrix obtained by iterative convergence of the MM algorithm under the interference hypothesis H1; represents the maximum likelihood estimate of the legal channel obtained by iterative convergence of the MM algorithm under the assumption of no interference H0; γ represents the interference detection threshold; formula (17) represents the relationship between the interference detection amount and the detection threshold. If it is greater than the detection threshold, it is determined that the uplink MIMO system is attacked by interference; if it is less than the detection threshold, it is determined that the uplink MIMO system is not attacked by interference.

6. The interference attack detection method for uplink low-precision ADC MIMO system according to claim 5, characterized in that: In step S4, the interference detection threshold γ satisfies a given false alarm probability and is obtained by repeated independent experiments offline.

Images