阅读说明：本技术 一种智能反射表面辅助型mimo隐蔽通信系统及参数优化方法 (Intelligent reflection surface auxiliary MIMO covert communication system and parameter optimization method ) 是由 郑通兴 陈欣 温雅婷 刘浩文 于 2021-07-20 设计创作，主要内容包括：本发明提供的一种智能反射表面辅助型MIMO隐蔽通信系统及参数优化方法,包括智能反射表面和虚拟监听方,所述智能反射表面用于实现发送方和接收方之间的通信,且发送方与接收方和虚拟监听方之间的直射信道均设置有障碍物；所述虚拟监听方用于非法截获智能反射表面向接收方卸载的任务；本发明利用智能反射表面辅助隐蔽通信,由于智能反射表面的每个反射元件都是无源地反射,且其没有配置调制器、解调器等信号处理设备,将减少系统额外的资源消耗；同时,本发明能够在不需要大量射频链路的情况下实现虚拟大规模MIMO以对抗多天线虚拟监听方,提升隐蔽通信性能。(The invention provides an intelligent reflection surface auxiliary MIMO covert communication system and a parameter optimization method, wherein the intelligent reflection surface auxiliary MIMO covert communication system comprises an intelligent reflection surface and a virtual monitoring party, the intelligent reflection surface is used for realizing communication between a sender and a receiver, and direct channels between the sender and the receiver and between the sender and the virtual monitoring party are provided with barriers; the virtual monitor is used for illegally intercepting the task of the intelligent reflection surface unloaded to the receiver; according to the invention, the intelligent reflection surface is used for assisting in covert communication, and because each reflection element of the intelligent reflection surface is passively reflected and is not provided with signal processing equipment such as a modulator and a demodulator, the extra resource consumption of the system is reduced; meanwhile, the invention can realize virtual large-scale MIMO to resist a multi-antenna virtual monitoring party without a large number of radio frequency links, thereby improving the covert communication performance.)

1. An intelligent reflection surface auxiliary MIMO covert communication system is characterized by comprising an intelligent reflection surface, a sender, a receiver and a virtual monitor, wherein the intelligent reflection surface is used for realizing communication between the sender and the receiver, and direct communication channels between the sender and the receiver and between the sender and the virtual monitor are provided with barriers; the virtual monitor is used for illegally intercepting and capturing tasks unloaded by the intelligent reflection surface to the receiving party, and noise uncertainty exists at the virtual monitor.

2. The system of claim 1, wherein the IRS is equipped with M reflective elements.

3. The MIMO covert communication system of claim 1, wherein N is configured for the sender, the receiver and the virtual listener respectively_A、N_BAnd N_WA root antenna.

4. A method for optimizing parameters of an intelligent reflective surface assisted MIMO covert communication system, wherein the intelligent reflective surface assisted MIMO covert communication system according to any one of claims 1 to 3 comprises the following steps:

step 1, optimizing a detection threshold value used by a virtual monitoring party for judging whether to communicate;

step 2, obtaining a minimum detection error probability expression of the virtual monitoring party according to the optimal detection threshold value;

step 3, obtaining the concealment constraint condition of the intelligent reflection surface auxiliary MIMO concealment communication system according to the minimum detection error probability expression;

and 4, optimizing parameters in the intelligent reflection surface auxiliary MIMO covert communication system according to the covert constraint condition.

5. The method of claim 4, wherein in step 1, the detection threshold value used by the virtual listener for determining whether to communicate is optimized, and the specific method is as follows:

s1, setting the detection error of the virtual monitor into a missing detection and a false alarm, wherein an expression of the missing detection probability of the virtual monitor and an expression of the false alarm probability of the virtual monitor are obtained;

s2, obtaining a virtual listener detection error probability expression according to the virtual listener missed detection probability expression and the virtual listener false alarm probability expression;

and S3, deriving the detection error probability expression of the virtual monitoring party to obtain an optimal detection threshold value.

6. The parameter optimization method of the intelligent reflection surface auxiliary MIMO covert communication system of claim 4, wherein in step 2, the minimum detection error probability of the virtual listener is calculated according to the optimal detection threshold value; the specific method comprises the following steps:

and substituting the optimal detection threshold value into the detection error probability expression to obtain the minimum detection error probability of the virtual monitoring party.

7. The method of claim 4, wherein in step 3, the system concealment constraint condition is derived according to the minimum detection error probability, and the specific method is as follows:

and (4) enabling the minimum detection error probability to be larger than or equal to the preset concealment success probability to obtain the system concealment constraint condition.

8. The method of claim 4, wherein in step 4, the parameters in the intelligent reflection surface assisted MIMO covert communication system are optimized according to covert constraint conditions, and the specific method is as follows:

the optimization target of the intelligent reflection surface auxiliary MIMO covert communication system is set as follows: maximizing the concealment rate of the intelligent reflection surface auxiliary MIMO covert communication system;

then an optimization problem expression is obtained:

s.t.R≥0 (10a)

tr(R)≤P (10b)

tr(H_IWQH_AIR(H_IWQH_AI)^H)≤η (10d)

and solving the optimization problem by adopting a DDPG algorithm to obtain the optimal value of the parameter in the intelligent reflection surface auxiliary MIMO covert communication system.

Technical Field

The invention relates to the problem of covert communication, in particular to an intelligent reflective surface auxiliary MIMO covert communication system and a parameter optimization method.

Background

With the rapid development of wireless communication technology, the information security problem is gradually valued by people, high-level encryption based on a secret key is an important feasible means for guaranteeing privacy security at present, but the means also encounters some obstacles in practical application, such as huge system overhead brought by secret key distribution, storage and management. Covert communication, also called low detection probability communication, is used as a secure communication means without a secret key, aims to reduce the probability that communication is detected by an illegal user, improves the confidentiality and the security of information delivery, and provides a novel solution for protecting the information security.

Various techniques are currently used to enhance covert communications, such as multiple antenna techniques, relaying, and interference. However, these techniques inevitably have some drawbacks that limit the enhancement of covert communication performance. For example, multi-antenna techniques are channel adaptive in nature, with performance being affected by the quality of the wireless channel. Antenna arrays of smaller size provide less enhancement of covert communication performance, while increasing the number of antennas requires more radio frequency links, resulting in higher power consumption and complexity. The relay includes processes of decoding and regenerating signals, and additional resources of the system are consumed. The implementation of interference usually requires additional friendly nodes and the interfering signal may interfere with the target user if not effectively controlled.

Covert communication performance is largely dependent on the wireless propagation environment, and the covert communication technology can adapt to a time-varying wireless channel but cannot effectively improve the time-varying wireless channel. Different from the technology, the intelligent reflecting surface has higher application value in covert communication due to the characteristics of low power consumption and channel reconstruction. The intelligent reflective surface, also called reconfigurable intelligent surface, is a software controlled super surface consisting of a large number of passive reflective elements. Each reflecting element can independently change the phase of an incident signal to cooperatively generate a fine reflected beam, so that the reflected signal is superposed and enhanced in the direction of a target user and offset and weakened in the direction of an illegal user, the intelligent reconstruction of a channel is realized, and the power consumption in the whole process is very low. These advantages make the intelligent reflective surface have wide application prospects in other communication scenarios, such as multi-cell, large-scale D2D, wireless information power transmission, and the like.

At present, researches on intelligent reflection surface auxiliary covert communication mostly focus on a scene that a target user and a virtual monitoring party are single-antenna receivers, and system parameters are optimized by adopting an alternative optimization algorithm. In fact, the MIMO and the intelligent reflection surface are simultaneously introduced into a covert communication system, the directional beam forming of the MIMO is utilized to concentrate the emission energy on the target direction, the emission signal power is reduced, the intelligent reflection surface is utilized to realize channel reconstruction, and the channel performance is improved, so that the virtual large-scale MIMO is realized under the condition that a large number of radio frequency links are not needed, and the covert communication performance can be further improved.

However, current research on intelligent reflective surface assisted MIMO covert communications is quite deficient. When the intelligent reflection surface is considered to assist MIMO covert communication, the difficulty of the system parameter optimization problem of covert rate maximization is increased sharply, because only optimizing a transmitting beam forming vector is no longer the optimal choice, but a transmitting signal covariance matrix needs to be optimized, and meanwhile, an objective function becomes a complex logarithm determinant function, so that the problem is difficult to be solved well by the existing method. Therefore, how to efficiently solve the parameter optimization problem of the intelligent reflection surface auxiliary MIMO covert communication system is an important topic worthy of research.

Disclosure of Invention

The invention aims to provide an intelligent reflection surface auxiliary MIMO covert communication system and a parameter optimization method, which solve the defects in the prior art.

In order to achieve the purpose, the invention adopts the technical scheme that:

the invention provides an intelligent reflection surface auxiliary MIMO covert communication system which comprises an intelligent reflection surface and a virtual monitor, wherein the intelligent reflection surface is used for realizing communication between a sender and a receiver, and direct channels between the sender and the receiver and between the virtual monitor are provided with barriers; the virtual monitoring party is used for illegally intercepting and capturing tasks unloaded by the intelligent reflection surface to the receiving party.

Preferably, the intelligent reflective surface IRS is equipped with M reflective elements.

Preferably, the sender, the receiver and the virtual listener are respectively configured with N_A、N_BAnd N_WA root antenna.

A parameter optimization method of an intelligent reflection surface auxiliary MIMO covert communication system is based on the intelligent reflection surface auxiliary MIMO covert communication system and comprises the following steps:

step 1, optimizing a detection threshold value used by a virtual monitoring party for judging whether to communicate;

step 2, obtaining a minimum detection error probability expression of the virtual monitoring party according to the optimal detection threshold value;

step 3, obtaining the concealment constraint condition of the intelligent reflection surface auxiliary MIMO concealment communication system according to the minimum detection error probability expression;

and 4, optimizing parameters in the intelligent reflection surface auxiliary MIMO covert communication system according to the covert constraint condition.

Preferably, in step 1, optimizing a detection threshold value used by the virtual listener to determine whether to communicate is performed, where the specific method is as follows:

s1, setting the detection error of the virtual monitor into a missing detection and a false alarm, wherein an expression of the missing detection probability of the virtual monitor and an expression of the false alarm probability of the virtual monitor are obtained;

s2, obtaining a virtual listener detection error probability expression according to the virtual listener missed detection probability expression and the virtual listener false alarm probability expression;

and S3, deriving the detection error probability expression of the virtual monitoring party to obtain an optimal detection threshold value.

Preferably, in step 2, the minimum detection error probability of the virtual listener is calculated according to the optimal detection threshold value; the specific method comprises the following steps:

and substituting the optimal detection threshold value into the detection error probability expression to obtain the minimum detection error probability of the virtual monitoring party.

Preferably, in step 3, the system concealment constraint condition is obtained according to the minimum detection error probability, and the specific method is as follows:

and (4) enabling the minimum detection error probability to be larger than or equal to the preset concealment success probability to obtain the system concealment constraint condition.

Preferably, in step 4, parameters in the intelligent reflective surface auxiliary MIMO covert communication system are optimized according to a covert constraint condition, and the specific method is as follows:

setting the optimization target of the MIMO covert communication system assisted by the intelligent reflecting surface as follows: maximizing the concealment rate of the intelligent reflection surface auxiliary MIMO covert communication system;

then an optimization problem expression is obtained:

s.t.R≥0 (10a)

tr(R)≤P (10b)

tr(H_IWQH_AIR(H_IWQH_AI)^H)≤η (10d)

and solving the optimization problem by adopting a DDPG algorithm to obtain the optimal value of the parameter in the MIMO covert communication system assisted by the intelligent reflection surface.

Compared with the prior art, the invention has the beneficial effects that:

the parameter optimization system of the intelligent reflection surface auxiliary MIMO covert communication system provided by the invention simultaneously introduces the MIMO and the intelligent reflection surface into covert communication, and further improves the concealment rate of the system by utilizing the channel reconstruction of the intelligent reflection surface and the directional beam formation of the MIMO; that is, the directional beam forming of MIMO is utilized to concentrate the transmitted energy on the target direction, reduce the transmitted signal power, and the intelligent reflection surface is utilized to realize channel reconstruction and improve the channel performance, thereby realizing virtual large-scale MIMO to confront the multi-antenna virtual monitor without a large number of radio frequency links, and improving the covert communication performance; compared with the traditional technology for realizing covert communication based on relay, interference and the like, the method utilizes the intelligent reflection surface to assist in covert communication, and because each reflection element of the intelligent reflection surface is passively reflected and is not provided with signal processing equipment such as a modulator and a demodulator, the extra resource consumption of the system is reduced.

The invention provides a parameter optimization method of an intelligent reflection surface auxiliary MIMO covert communication system, provides a covert communication scheme capable of effectively resisting a multi-antenna virtual listener, and has wider applicability. Because of the constraint of the unit module value of the optimization variable and the high coupling of the optimization variable in the objective function and the constraint condition, the optimization problem provided by the invention is a complex non-convex problem, and the traditional convex optimization method is difficult to be directly used for solving the problem; meanwhile, although the problem can be solved by the alternative optimization algorithm, the optimal solution of the problem cannot be obtained, so that the improvement of the covert communication performance of the system is limited to a great extent; in fact, the optimization problem can be viewed as a decision-making problem of continuous motion and state space, and deep reinforcement learning is a powerful tool for solving such problems. Therefore, the invention adopts a depth deterministic strategy gradient (DDPG) algorithm in the deep reinforcement learning, can effectively find a more optimal solution of the problem and furthest improves the concealment rate of the system.

Drawings

FIG. 1 is a covert communication system model to which the present invention relates;

FIG. 2 is a DDPG based deep reinforcement learning network model;

FIG. 3 is a simulation plot of the blind rate as a function of maximum transmit power given a random phase shift comparing the present optimization algorithm to the alternative optimization algorithm;

Detailed Description

The invention is described in further detail below with reference to the figures and examples.

As shown in fig. 1, the MIMO covert communication system assisted by an intelligent reflective surface provided by the present invention includes a sender Alice, an intelligent reflective surface IRS, and a receiver Bob, wherein the sender Alice communicates with the receiver Bob through the intelligent reflective surface IRS, and a direct channel between the sender Alice and the receiver Bob is provided with an obstacle.

It is assumed that the sender Alice knows the global channel state information CSI.

The intelligent reflective surface IRS is equipped with M reflective elements.

The sender Alice, the receiver Bob and the virtual listener Willie are respectively configured with N_A、N_BAnd N_WA root antenna.

Suppose there is noise uncertainty at the virtual listener Willie and the probability density function of the virtual listener Willie noise powerComprises the following steps:

wherein the content of the first and second substances,is the true noise power of the virtual listener Willie; ρ ≧ 1 is the noise uncertainty;the noise power is the rated noise power of the virtual listener Willie, that is, the noise power when ρ is 1.

The invention provides a parameter optimization method for an intelligent reflection surface auxiliary MIMO covert communication system, which comprises the following steps:

step 1, the virtual listener Willie adopts an energy detection method to judge whether communication occurs, and a detection threshold value is set and recorded asDetecting the statistic as the mean received power T_W. From LuFrom the design angle of the bar, the virtual monitoring party is set to always select the optimal detection threshold value, so that the detection error probability is minimum.

For covert communication, Willie needs to distinguish two hypotheses, respectively, null hypothesis H₀And alternative hypothesis H₁Wherein zero assumes H₀No message is sent for the sender Alice; alternative hypothesis H₁The prior probabilities for both hypotheses are 1/2 for sender Alice to send a message.

Setting whether there are observation samples of the virtual listener Willie for a plurality of times, and then setting the average received power T of the virtual listener Willie_WCan be expressed as:

wherein, P_W＝tr(H_IWQH_AIR(H_IWQH_AI)^H) Is the useful signal power received by the virtual listener Willie, tr () represents the trace of the matrix; h_IWAnd H_AIRespectively obtaining a channel coefficient matrix from the intelligent reflecting surface IRS to the virtual listener Willie and a channel coefficient matrix from the sender Alice to the intelligent reflecting surface IRS;is an intelligent reflective surface IRS phase shift matrix, diag () represents a diagonal matrix; theta_iIs the phase shift coefficient of the ith reflective element;is the covariance matrix of the MIMO transmitted signal, x is the transmitted signal, R is greater than or equal to 0.

The detection errors of the virtual listener Willie are divided into two types of missed detection and false alarm, wherein the probability P of missed detection_MDComprises the following steps:

probability of false alarm P_FAComprises the following steps:

thus, the detection error probability ξ is expressed as:

wherein Pr () represents a probability; lambda denotes the virtual listener Willie detection threshold value, P_WRepresenting the power of the desired signal received by Willie;representing the true noise power of Willie; max (a, b) represents the maximum of a, b;is a probability density function of Willie noise power.

And secondly, solving an optimal detection threshold value lambda to ensure that the detection error probability xi of the virtual listener Willie is minimum, and from the design angle of robustness, in order to ensure the concealment of communication, requiring that the minimum detection error probability is more than or equal to the preset concealment success probability 1-kappa, and the kappa belongs to [0,1] as a very small number, thereby obtaining a system concealment constraint condition, wherein the specific steps are as follows:

in the formula (5), let xi be derived from lambda to obtain

When in useWhen in useWhen the temperature of the water is higher than the set temperature,thereby obtaining an optimum detection threshold value lambda^*Comprises the following steps:

lambda is measured^*Substituting into xi expression to obtain minimum detection error probability xi_minComprises the following steps:

xi should be required to guarantee covert communication_minNot less than 1-kappa and is further derived

And thirdly, setting the maximum transmission power of the Alice at the sender as P, and meeting the requirement that tr (R) is less than or equal to P. Intelligent reflective surface IRS each reflective element has an amplitude gain of 1 to incident signalThe optimization problem can be expressed as the maximum system concealment rate C by jointly optimizing R and Q under the constraint condition

s.t.R≤0 (10a)

tr(R)≤P (10b)

tr(H_IWQH_AIR(H_IWQH_AI)^H)≤η (10d)

Wherein H_IBIs a channel coefficient matrix from the intelligent reflecting surface IRS to the receiver Bob;is the noise power of the receiver Bob.

Solving the optimization problem by adopting a depth deterministic strategy gradient (DDPG) algorithm in deep reinforcement learning, specifically:

the communication system is set to be an environment and the sender is an agent that selects a transmit covariance matrix R and a phase shift matrix Q of the intelligent reflective surface.

The state space is defined as:

in the formula, C^t-1Represents the concealment rate at t-1; q^t-1Representing a phase shift matrix at t-1; vec () represents stacking the matrix column by column into a column vector; real () denotes the real part of the complex number; imag () represents the imaginary part of the complex number;

the action space is defined as:

in the formula (I), the compound is shown in the specification,representing the phase of diagonal elements of the intelligent reflective surface phase shift matrix;a real diagonal element representing a transmit signal covariance matrix;representing the real part of the lower triangular matrix elements (except the diagonal elements) of the transmit signal covariance matrix;representing the imaginary part of the lower triangular matrix element (except the diagonal elements) of the transmit signal covariance matrix.

The reward function is defined as:

r_tthe initial value is 0, and then whether the constraint condition is satisfied is judged. R if and only if the constraints are all true_t＝C^t，C^tIndicating the current concealment rate. Otherwise if the constraint (10a) does not hold, thenβ_kRepresents the negative eigenvalue of R; if the constraint (10b) does not hold, r_t＝r_t+ P-tr (R); if the constraint (10d) does not hold, r_t＝r_t+η-tr(H_IWQH_AIR(H_IWQH_AI)^H)。

The DDPG-based deep reinforcement learning network model is shown in fig. 2, and includes 4 Deep Neural Networks (DNNs) including an online operator-critical (Q value) network and a target operator-critical network.

The network training process comprises the following steps: initialize 4 DNNs, empty experience poolsAt the beginning of each round, R and Q are randomly generated, and the current state s is obtained_t. Will state s_tInput into the network resulting in action a_tAnd is formed by_tForming variables R and Q, calculating the current prize value R_tAnd the next step status s_t+1. Will(s)_t,a_t,r_t,s_t+1) Store to experience pool of capacity DIn (1). When the experience pool is filled with the samples, the training of the network is started, namely N samples(s) are randomly captured in the experience pool_i,a_i,r_i,s_i+1) For updating network parameters. Each round generates a new sample to update the experience pool. When the network converges, the solutions R and Q of the original optimization problem can be recovered by the output action.

And finally finishing the algorithm.

The invention introduces MIMO and intelligent reflecting surface into the covert communication system at the same time for the first time, utilizes directional beam forming of MIMO to concentrate the transmitting energy on the target direction, reduces the transmitting signal power, realizes channel reconstruction by combining the intelligent reflecting surface, and improves the channel performance, thereby realizing virtual large-scale MIMO to resist a multi-antenna virtual monitoring party under the condition of not needing a large number of radio frequency links, and improving the covert communication performance.

The parameter optimization of the proposed intelligent reflective surface auxiliary type MIMO covert communication system is a problem to be solved. Different from the parameter optimization of covert communication in which an intelligent reflecting surface auxiliary target user and a virtual listener are single-antenna receivers, an objective function and a covert constraint condition in the parameter optimization problem of the proposed intelligent reflecting surface auxiliary MIMO covert communication system are complex matrix functions, optimization variables are highly coupled in the complex matrix functions, and higher requirements are provided for solving the optimization problem.

In this embodiment, all channel gain matrices are set to large-scale fading L₀d^-μ/2And a small scale fading matrix. L is₀-15dB represents the path loss at 1m, and μ -3 represents the path loss exponent. The elements in the small-scale fading matrix are complex gaussian random variables with a mean value of 0 and a variance of 1. The distances from IRS to Alice, Bob and Willie are d_AI＝40m、d_IB＝15m、d_IW＝40m。N_A＝N_B＝N_W＝2，M＝4，κ＝0.1，ρ＝3dB，In the deep reinforcement learning based on the DDPG algorithm, the experience pool capacity D is 20000, N is 128, all DNN networks are 4-layer structures, and the 4-layer structures are an input layer, two hidden layers and an output layer respectively.

The activation functions of the input layer and the two hidden layers of the Actor and critic networks are both ReLU functions, and the activation function of the output layer of the Actor network is a tanh function.

Fig. 3 compares the present optimization algorithm with the alternative optimization algorithm for the change in concealment rate with the maximum transmit power P given a random phase shift. It can be seen from the figure that the concealment rate under the optimization algorithm is obviously greater than that under the alternate optimization algorithm and the scheme of only optimizing the covariance matrix of the transmitted signal by giving random phase shift, and the concealment rate under the optimization algorithm is improved more remarkably as the transmitted power is increased. Therefore, the method and the device better solve the problem of parameter optimization of the intelligent reflection surface auxiliary MIMO covert communication system and improve the covert communication performance of the system.