A method for monitoring suspicious users with IRSs assistance in internet of things

Abstract

The application discloses an IRSs-assisted suspicious user method in an Internet of Things, comprising the following steps: taking a distributed intelligent reflecting surface IRSs phase shift matrix Phi0 and Phi1 of a reverse pilot transmission stage and a data transmission stage as variables, taking a node availability rate as an optimization target, and establishing an optimization problem P1; and solving the optimization problem P1 to obtain the optimal distributed intelligent reflecting surface IRSs phase shift matrix Phi0 and Phi1 of the reverse pilot transmission stage and the data transmission stage. The method can improve the total received signal strength, the worst received signal strength, the node availability rate and the distance, and ensures concealment in a harsh environment.

Description

Technical Field

[0001] This invention belongs to the field of wireless information security and relates to a method for identifying suspicious users assisted by IRSs in the Internet of Things. Background Technology

[0002] The rapid development of wireless communication networks has brought immense convenience to global communication, but it has also posed significant challenges to public safety. Some criminals exploit wireless communication networks for malicious activities, such as online fraud and spreading rumors. How to prevent these activities has attracted considerable attention from experts in the field of information security. Some have proposed first using big data collection and analysis technologies to screen suspicious users, and then legalizing their communication links at the wireless physical layer. Stable, covert, and continuous monitoring of suspicious communications can prevent and detect potentially harmful illegal and criminal activities at the signal source in the first instance. Thanks to the development of intelligent reflective surface technology in recent years, large-scale deployment of intelligent reflective surfaces is expected to become one of the important technical means to further improve performance.

[0003] Based on whether the legitimate party participates in suspicious communications, they can be divided into two categories: passive and active. Passive communication refers to a party only acquiring suspicious information without actively intervening in the suspicious receiver. However, considering that in general scenarios, both parties in suspicious communications are in relatively concealed locations, and the receiver is often far away, resulting in poor conditions and making it impossible to decode suspicious information with minimal errors, the most common intervention measure in this case is to send jamming signals. This technique of actively intervening in suspicious communications is called active technology, but this technology requires additional energy, and sending artificial noise signals for interference also incurs significant costs for self-interference cancellation. How to solve these problems is key to technological development. However, current work has not considered combining distributed intelligent reflective surfaces with this technology to ensure information security. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for identifying suspicious users assisted by IRSs in the Internet of Things, which can improve the availability of nodes.

[0005] To achieve the above objectives, the method for identifying suspicious users assisted by IRSs in the Internet of Things according to the present invention includes:

[0006] Based on the legitimate location, find the optimal location to deploy distributed IRSs. For the IRS-LMs associated group with a pre-defined LOS with a direct path, take the phase shift matrices Φ0 and Φ1 of the distributed intelligent reflective surface IRSs in the reverse pilot transmission stage and the data transmission stage as variables, take maximizing node availability as the optimization objective, and maximize the worst signal-to-noise ratio of the users in this group to the maximum extent, and establish optimization problem P1.

[0007] Solving the optimization problem P1 yields the optimal distributed intelligent reflective surface IRSs phase shift matrices Φ0 and Φ1 for the reverse pilot transmission stage and the data transmission stage.

[0008] The optimization problem P1 is:

[0009]

[0010] stγ SR ≥γ0

[0011] ||w|| 2 ≤P data

[0012] Where γ0 is a given threshold, γ SR For the signal-to-noise ratio of the suspected receiver's SR, γ i For each legitimate LMs signal-to-noise ratio, P data The data transmission power of the suspected sender ST.

[0013] By using the MM and ADMM algorithms to solve the optimization problem P1, the optimal phase shift matrices Φ0 and Φ1 of the distributed intelligent reflective surface IRSs for the reverse pilot transmission stage and the data transmission stage are obtained.

[0014] In the optimization of the phase shift matrix Φ0 of the distributed IRSs during the reverse pilot transmission stage, maximizing the channel estimation bias is equivalent to optimizing problem P2:

[0015]

[0016] st|v i |=1,i=1,...,N

[0017] Expanding the above equation and removing the constant term, it can be equivalently expressed as:

[0018]

[0019] st|v i |=1,i=1,...,N

[0020] in,

[0021] Solving optimization problem P3 based on the MM maximization and minimization algorithm;

[0022] During the optimization of the phase shift matrix Φ1 of the distributed IRSs in the data transmission phase, maximizing the received signal-to-noise ratio of the worst channel quality side is equivalent to optimizing problem P5:

[0023]

[0024] st|vi |=1,i=1,...,N

[0025] in, and Representing matrix H respectively TM , The k-th row of the matrix, Let be the signal power received by the kth legitimate party.

[0026] Solving optimization problem P5 based on the alternating direction multiplier method of ADMM.

[0027] The present invention has the following beneficial effects:

[0028] The method for detecting suspicious users in the Internet of Things (IoT) assisted by IRSs, as described in this invention, uses the phase shift matrices Φ0 and Φ1 of the distributed intelligent reflective surfaces (IRSs) in the reverse pilot transmission stage and the data transmission stage as variables. Maximizing node availability is the optimization objective, and an optimization problem P1 is established. This optimization problem is then solved, and data transmission is performed accordingly. Simulation tests show that this invention outperforms traditional methods for maximizing total signal strength (Maxsum scheme) and passive (Without IRSs) schemes in terms of total received signal strength, worst-case received signal strength, node availability, and distance. Furthermore, the device remains silent, requiring no transmitting antenna. Active detection is achieved by controlling IRSs pilot deception, reducing energy consumption and cost, and ensuring concealment in harsh environments. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the system model;

[0030] Figure 2 This is a model diagram of the system simulation scenario;

[0031] Figure 3 The graph shows the total signal-to-noise ratio and worst signal-to-noise ratio of the valid LMs as a function of the number of valid LMs, K.

[0032] Figure 4 The graph shows the total signal-to-noise ratio and worst signal-to-noise ratio of the legitimate end LMs as a function of the total number of reflective elements N;

[0033] Figure 5 The cumulative distribution function of the signal-to-noise ratio (SNR) of the legitimate LMs is plotted under the following schemes: maximizing the total SNR (Maxsum), maximizing the worst SNR (Maxmin), random phase, and passive (Without IRS).

[0034] Figure 6 The graph shows the average number of successful nodes for legitimate LMs as a function of the total number of reflecting elements N.

[0035] Figure 7 The probability density distribution of the number of successful nodes in the legitimate LMs schemes is shown in the following diagrams: Maxsum (maximum signal-to-noise ratio), Maxmin (maximum worst signal-to-noise ratio), Random phase, and Without IRS (passive) schemes.

[0036] Figure 8 The graph shows the worst signal-to-noise ratio of legitimate LMs as a function of the location of the distributed intelligent reflective surface IRS, and compares it with the case of a single IRS. Detailed Implementation

[0037] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, not all embodiments, and are not intended to limit the scope of the present invention. Furthermore, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion regarding the concepts disclosed in the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort should fall within the scope of protection of the present invention.

[0038] The accompanying drawings show structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not drawn to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0039] refer to Figure 1 The method for identifying suspicious users assisted by IRSs in the Internet of Things according to the present invention includes the following steps:

[0040] 1) System Model Establishment: The system includes a multi-antenna suspicious transmitter (ST), a single-antenna suspicious receiver (SR), multiple mobile IoT devices as legitimate receivers (LMs), and multiple intelligent reflective surface (IRSs). The purpose of the distributed IRSs is to ensure that there is a direct path between each party and the intelligent reflective surface IRS as much as possible, while improving node availability without increasing the total number of reflective units (deployment cost). The IRSs are deployed in a distributed manner in fixed locations within the environment, and their operation is controlled by a centralized IRS controller. For a specific IRS-LMs association group with a direct path (LOS), the surface phase shift is designed to maximize the minimum received signal-to-noise ratio (SNR) for that group of users. During communication between the suspicious parties, the optimal location for deploying the distributed intelligent reflective surface IRSs is identified and invoked to help the silent party achieve deception during the SR to ST uplink pilot transmission phase. To prevent non-sudden factors from disrupting one party and causing an interruption, the distributed intelligent reflective surfaces (IRSs) also need to assist the silent LMs. During the suspicious information transmission phase from the suspicious sender (ST) to the suspicious receiver (SR), the worst-case scenario's performance is maximized, improving node availability. This invention employs joint optimization of the phase shift matrix of the distributed intelligent reflective surfaces (LMs) in both uplink and downlink phases to achieve sustainable and stable covert operation across multiple nodes. It is assumed that all channels within the system are quasi-static fading channels, and the channel state remains unchanged within a single pilot decoy transmission and data transmission time block. The changes in channel parameters across different transmission time blocks are independent and identically distributed. This system operates in TDD mode, where channel heterogeneity exists. In the simulation, we assume that the relevant links of the legitimate LMs are unknown to the suspicious communication parties ST and SR. This assumption is reasonable because the LMs remain silent during both the pilot transmission and data transmission phases between ST and SR, without transmitting any signals. Furthermore, the phase shift matrix design does not interfere with or suppress the received signal of the suspicious receiver (SR). Therefore, detecting the presence of the LMs is difficult for the suspicious communication parties ST and SR.

[0041] 2) Pilot signal deception

[0042] During this phase, the LMs remain silent, and the phase shift matrix Φ0 of the distributed IRSs, consisting of N total reflection units, is designed to reflect the pilot sequence transmitted from the suspected receiver SR to the suspected transmitter ST. M represents the number of suspected ST antennas at the transmitter, P pilot If the pilot transmit power of the suspected receiver SR is given, then the signal received by the suspected transmitter ST is:

[0043]

[0044] in, and These are the fading channels from the suspected receiver (SR) and distributed IRSs to the suspected transmitter (ST), respectively. For channels from suspicious receivers (SRs) to distributed IRSs, Its diagonal elements are the phase shifts corresponding to the surface reflection units of the distributed IRSs, 0≤|v i |≤1,1≤i≤N,n TR The additive complex white Gaussian noise received by the suspected transmitter ST has a mean of 0 and a variance of .

[0045] 3) Channel estimation

[0046] For the signal received by the suspicious sender ST, the suspicious sender ST uses the least squares method to estimate the Channel State Information (CSI) between the suspicious sender ST and the suspicious receiver SR. Due to pilot spoofing, the channel estimated by the suspicious sender ST is deviated, resulting in a reflection link between ST-IRSs-SR, which leads to information leakage to the distributed IRSs during the next stage of data transmission.

[0047] Suspicious transmitter ST has known pilot transmission power P pilot and pilot sequence u H , The ST channel estimation results for the suspected transmitter.

[0048]

[0049] 4) Data transmission

[0050] During uplink pilot transmission and downlink data transmission, legitimate LMs remain silent. Existing methods for detecting pilot pollution by LMs on ST, including random modulation, artificial noise / random data-assisted methods, random orthogonal pilot methods, and statistical characteristic methods, are all ineffective. The first three methods introduce a certain degree of randomness into the standard pilot pollution phase, making it impossible for the legitimate party to know the pilot sequence to be transmitted in advance. It should be noted that the signal sequence reflected from the distributed IRSs is the same as the pilot sequence transmitted by the suspicious receiver SR, thus rendering the first three schemes ineffective, as it requires the LMs and SR to transmit the same signal sequence. The last method requires prior knowledge of the channel probability distribution of the suspicious receiver SR and the legitimate LMs to estimate the instantaneous channel coefficients. In practice, obtaining such data is not easy, especially for the legitimate LMs, as they do not cooperate with the suspicious communication system. Therefore, the link is unknown to both parties in the suspicious communication, making it impossible to detect the existence of the legitimate party, and the last scheme also fails. Therefore, the suspicious transmitter ST will use maximum ratio transmission (MRT) to transmit legitimate data under the maximum transmit power limit. data The data transmission power of the suspected sender ST. Let ST be the transmit beamforming vector of the suspected transmitter, where,

[0051]

[0052] 5) Data reception

[0053] This is the direct channel between the suspected communicating parties, ST and SR. This is a reflection channel between the suspected sender ST, distributed IRSs, and suspected receiver SR. This represents the channel from the suspicious sender ST to the legitimate sender LMs, where K is the number of legitimate sender LMs. Let Φ1 be the channel from distributed IRSs to legitimate LMs, Φ1 be the phase shift matrix of the distributed IRSs during the data transmission phase, and X be a suspicious information signal with unit power. and The additive complex white Gaussian noise signals from the suspected receiver SR and the legitimate receiver LMs are respectively, with a mean of 0 and variances of respectively. and The signals received by the suspicious receiver SR and the legitimate receiver LMs They are respectively:

[0054]

[0055]

[0056] The signal-to-noise ratios received by the suspicious receiver SR and each legitimate receiver LMs are:

[0057]

[0058]

[0059] γ SR For the signal-to-noise ratio of the suspected receiver's SR, γ i The signal-to-noise ratio of each legitimate LMs is given, assuming the received signal reaches a threshold γ. m Information interpretation can then proceed, indicating success, when γ i ≥γ m When γ is high, the node can successfully decode the suspicious information, and the rate is the suspicious communication rate; when γ is high... i <γ m If the node cannot guarantee error-free decoding of suspicious information, it will fail, and the rate will be 0. The legitimate LMs event can be represented by the following indicator function:

[0060]

[0061] R=1 indicates success, and R=0 indicates failure.

[0062] 6) System optimization objectives

[0063] To maximize the performance of the LMs, the optimization of the phase shift matrix Φ0 in the pilot transmission stage aims to perform pilot deception and maximize the ST channel estimation deviation of the suspicious transmitter. The natural optimization objective of the phase shift matrix Φ1 in the data transmission stage is to maximize the total received signal power. However, non-burst factors can disrupt certain aspects, causing a sharp decrease in the number of successful nodes. Therefore, the optimization objective of the phase shift matrix Φ1 in the data transmission stage of this invention is to maximize the signal-to-noise ratio of the worst-case received signal in order to increase the number of effective nodes.

[0064] This invention maximizes the worst-case reception signal-to-noise ratio by jointly optimizing the phase shift matrices Φ0 and Φ1 of the distributed intelligent reflective surfaces IRSs in the reverse pilot transmission stage and the data transmission stage, thereby improving node availability.

[0065] The overall optimization goal of the system is:

[0066]

[0067] stγ SR ≥γ0

[0068] ||w|| 2 ≤P data

[0069] Constraint γ SR The meaning of ≥γ0 is: ensure that the signal-to-noise ratio (SNR) of the suspicious receiver (SR) is higher than a given threshold γ0, that is, ensure that the suspicious receiver (SR) can correctly decode the information and cannot detect its existence. This can only be achieved if the SNR of each legitimate receiver (LMs) is greater than a certain given threshold, constraining ||w||. 2 ≤P T To limit the power of the suspicious transmitter, when the total power cannot meet the signal-to-noise ratio requirement of the suspicious receiver (SR), the suspicious transmitter (ST) stops transmitting suspicious information. At this time, the signal-to-noise ratio of both the suspicious receiver (SR) and the legitimate receiver (LMs) is 0.

[0070] Optimization of the phase shift matrix Φ0 of distributed IRSs in the reverse pilot transmission stage: maximizing the channel estimation bias is equivalent to a P2 optimization problem.

[0071]

[0072] st|v i |=1,i=1,...,N

[0073] Expanding the above equation and removing the constant term, we get

[0074]

[0075] st|v i|=1,i=1,...,N

[0076] in,

[0077] Because of the non-convex constraint of the unit modulus, P3 is a non-convex problem. Therefore, the minimization-maximum (MM) algorithm is used to solve P3. The core idea of ​​the minimization-maximum (MM) algorithm is to first find the approximate upper bound of the objective function, and then iteratively find the optimal value of the upper bound under the constraints. The convergence point is the local optimum.

[0078] k is the iteration number, let v k To satisfy a feasible solution to problem P3, the next iteration point v k+1 The upper bound of the objective function is

[0079]

[0080] in, Removing the constant term from the above equation, the optimization problem P3 can be equivalently represented as:

[0081]

[0082] Wherein, if and only if v i With α i When they are equal, Therefore, the closed-form optimal solution to optimization problem P4 is:

[0083]

[0084] Let k = k + 1, and update v k The value is calculated until the objective function converges. The feasible point v0 is initialized and the MM algorithm is applied to iteratively solve the optimization problem P3, thus obtaining the local optimal solution Φ0 of the optimization problem P3.

[0085] Optimizing the phase shift matrix Φ1 of distributed IRSs during data transmission: Maximizing the signal-to-noise ratio of the receiver with the worst channel quality is equivalent to optimizing problem P5.

[0086]

[0087] st|v i |=1,i=1,...,N

[0088] They are matrices H TM and The k-th row of the matrix, This represents the signal power received by the kth legitimate party.

[0089] After the first-stage phase shift matrix Φ0 is optimized, the beamforming vector w of the suspected transmitter ST is obtained. With w fixed, the phase shift matrix Φ1 for the second-stage data transmission is optimized. First, the optimization problem P5 is transformed into a suitable form, and the update expression is directly given. With w fixed, the optimization problem P5 can be written as:

[0090]

[0091] st|v i |=1,i=1,...,N

[0092] in,

[0093] We need to find the lower bound of problem P6, and then apply the above equation to... Expand the points to obtain the lower linear bound of the objective function.

[0094]

[0095] At point When the above equation holds true, let:

[0096]

[0097] in, The function is a linear lower bound of the objective function, and at the point... hour, The linear ADMM algorithm needs to be used to maximize the distributed IRSs surface phase shift modulus of 1 under the constraint. The function, and the corresponding optimization problem P7, is:

[0098]

[0099] st|v i |=1,i=1,...,N

[0100] Solving constrained optimization problems is more complex than solving unconstrained ones; therefore, let the feasible region with unit modulus constraints be:

[0101]

[0102] lower bound function Inverting the result gives:

[0103]

[0104] Then problem P7 is equivalent to

[0105]

[0106] The Alternating Direction Multiplier Method (ADMM) is an important method for solving convex optimization problems with separable structures, used to solve equality-constrained optimization problems with multiple separable variables. Therefore, the above equation requires variable separation; let the function... in, [x] k Let x represent the k-th element of the vector x. Then the optimization problem P8 is equivalent to:

[0107]

[0108] in

[0109]

[0110] By introducing an auxiliary variable r, the equivalent ADMM expression is:

[0111]

[0112] stSv-r=0

[0113] The augmented Lagrangian form of problem P10 is:

[0114]

[0115] Where ρ>0 is the penalty parameter, λ is the dual variable of the constraint Sv-r=0, and in the i-th iteration, given v i ,r i ,λ i Apply the linear ADMM method to iteratively update v, r, and λ alternately.

[0116] Linearized ADMM is a variant of ADMM, the difference being that it typically uses... Replace with To modify the update process of v, we need to linearize the quadratic term and add new quadratic regularization. The details of updating each variable are as follows:

[0117] Update v

[0118]

[0119] The optimization problem of equivalent optimization v is:

[0120]

[0121] st|v i |=1,i=1,...,N

[0122] The minimum value of the above problem is:

[0123]

[0124] Where, μ≥ρ||S|| 2 And define proj(x) = [x1 / |x1|,...,x N / |x N |], and set when x i When = 0, [proj(x)] N =1;

[0125] Update r

[0126]

[0127] set up Then we have:

[0128]

[0129] Here, the real and imaginary parts of r are decoupled, so the real and imaginary parts are optimized separately:

[0130]

[0131]

[0132] Find using the binary search algorithm

[0133] Then the real part of r is updated as follows:

[0134]

[0135] Update λ

[0136] λ i+1 =λ i +ρSv i+1 -ρr i+1

[0137] The above update expression is iteratively updated until convergence is achieved, resulting in a suboptimal solution to optimization problem P5. The specific process of the alternating joint optimization AO_MM_ADMM algorithm is shown in Table 1:

[0138] Table 1

[0139]

[0140]

[0141] Simulation test

[0142] Figure 2The given system simulation scenario model shows that the suspected transmitter (ST) has M=5 antennas located at (20m, 0m, 30m). Multiple legitimate single-antenna mobile IoT devices (LMs) are randomly distributed in a non-cooperative XOY domain with a radius of 20m centered at (20m, 160m, 0m). Two distributed intelligent reflective surfaces (IRSs) are located at IRS1 (0m, 20m, 3m) and IRS2 (40m, 20m, 3m), respectively. The suspected receiver (SR) is located at (20m, 100m, 0m). The pilot transmission and data transmission power are set to P. pilot =P data =20dBm, the noise variance of the suspicious sender ST, the suspicious receiver SR, and the legitimate receiver LMs is all 0. To include as many cases as possible, a total of 10 were conducted. 5 In this simulation, the path loss model is derived from PL = (PL0 - 10ρlog) 10 (d / d0) is given. The path loss at a reference distance d0 = 1m is PL0 = -30dB. In the simulation, suspicious communication parties ST and SR are set up, along with suspicious sender ST and legitimate receiver LMs, suspicious sender ST and distributed intelligent reflective surfaces IRSs, and the path loss exponents between distributed IRSs and suspicious receiver SR and LMs are respectively ρ. TR =ρ TM =4, ρ TI =ρ IR =ρ IM =2, because the location of IRSs is selected, and communication links assisted by IRSs can have a transmission environment close to free space, experiencing free space path loss. Channels related to distributed IRSs are dominated by LOS paths, and the Rice factor of the channel between the suspected sender ST and the IRSs is ε. TR =2, for a specific IRSs-LMs association group with a direct path LOS, the relevant channel Rice factor ε RM =10.

[0143] The experimental simulation verification of this invention is respectively represented as follows: Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8 . Figure 3 and Figure 4 The figures show the changes in the total signal-to-noise ratio (SNR) and worst-case SNR of the legal receiver LMs as a function of the number of reflective elements K and the total number of reflective elements N. It can be seen that this invention outperforms the Maxsum scheme in both total received SNR and worst-case SNR. In other words, this invention achieves higher SNR and worst-case SNR gain. Figure 5Let represent the cumulative distribution functions of the legitimate LMs, signal-to-noise ratio (SNR), and worst SNR under the following schemes: maximizing total SNR (Maxsum), maximizing worst SNR (Maxmin), random phase shift, and passive (Without IRS). It can be seen that this invention has significant advantages over other schemes in terms of total SNR and worst SNR. Figure 6 and Figure 7 Experimental results show that the present invention can increase the number of successful nodes, thereby ensuring stable and continuous real-time operation in harsh environments. The worst signal-to-noise ratio of legitimate LMs varies with the number of ST antennas M of the suspected transmitter and the total number of reflective elements N. The experimental results show that increasing the number of reflective elements on the surface of distributed IRSs is more effective in improving the performance of the legitimate LMs than increasing the number of antennas. Moreover, since IRSs are passive devices, deploying large-scale IRSs is more energy-efficient than installing more energy-consuming RF chains and power amplifiers. The experimental results clearly demonstrate that IRSs have the advantages of traditional system designs in terms of communication performance and energy consumption. Figure 8 The experiment compares the worst signal-to-noise ratio of LMs with the location of distributed intelligent reflective surface IRS with that of a single IRS. The results show that the performance improves as IRSs are deployed near LMs, signal hotspots are formed near LMs, and the performance of distributed intelligent reflective surface IRSs is better than that of centralized (single IRS) deployment.

Claims

1. A method for monitoring suspicious users with IRSs assistance in Internet of Things, characterized in that, Comprising: Based on the legal monitoring position, find the optimal position to deploy distributed IRSs, for the preset LOS IRS-LM association group that exists direct path, take the distributed intelligent reflecting surface IRSs phase shift matrix Φ0 and Φ1 in the reverse pilot transmission stage and data transmission stage as variables, take the maximum monitoring node availability as the optimization goal, maximize the worst monitoring signal to noise ratio of the group of monitoring users, establish optimization problem P1; Solve the optimization problem P1, get the optimal distributed intelligent reflecting surface IRSs phase shift matrix Φ0 and Φ1 in the reverse pilot transmission stage and data transmission stage; Optimization problem P1 is: s.t.γ SR ≥γ0 ||w| 2 ≤P data wherein γ0 is a given threshold, γ SR is a signal-to-noise ratio of the suspicious receiving end SR, γ i is a signal-to-noise ratio of each legal monitoring end LM, P data is a data transmission power of the suspicious transmitting end ST, and w is a transmitting beamforming vector of the suspicious transmitting end ST. In the distributed IRSs phase shift matrix Φ0 optimization process of the reverse pilot transmission stage, maximize the channel estimation bias, which is equivalent to optimization problem P2: s.t. |v i | = 1, i = 1,..., N where v i is the ith iteration point; Solve the optimization problem P2 by using MM algorithm, get the optimal distributed intelligent reflecting surface IRSs phase shift matrix Φ0 in the reverse pilot transmission stage; In the distributed IRSs phase shift matrix Φ1 optimization process of the data transmission stage, maximize the worst monitoring signal to noise ratio of the channel quality, which is equivalent to optimization problem P5: s.t. |v i | = 1, i = 1,..., N wherein and respectively represent the matrix H TM , the kth row of the matrix, is the signal power received by the kth legitimate listener, H and h both represent the channel; Solve the optimization problem P5 based on ADMM alternating direction multiplier method, get the optimal distributed intelligent reflecting surface IRSs phase shift matrix Φ1 in the data transmission stage.

Images