A smart reflective surface assisted secure communication method for wireless energy-carrying RSMA network

Abstract

The present invention provides a smart reflective surface assisted secure communication method for a wireless energy-carrying RSMA network, comprising the following steps: S1: constructing an initial target optimization model; S2: iteratively optimizing to obtain an optimal beamforming vector and a public safety rate by given user power division coefficients and an IRS reflection phase shift matrix; iteratively optimizing to obtain an optimal user power division coefficient by given beamforming vectors, a public safety rate, and an IRS reflection phase shift matrix; iteratively optimizing to obtain an optimal IRS reflection phase shift matrix by given beamforming vectors, a public safety rate, and a user power division coefficient; S3: performing wireless secure communication according to the optimal beamforming vector, the public safety rate, the user power division coefficient, and the IRS reflection phase shift matrix. The present invention provides a smart reflective surface assisted secure communication method for a wireless energy-carrying RSMA network, which solves the problem that the current wireless energy-carrying RSMA network cannot take into account the energy consumption of the system while enhancing the secure communication performance of the wireless communication system.

Description

Technical Field

[0001] The present invention relates to the technical field of secure communications, and more specifically, to an intelligent reflective surface-assisted secure communication method for a wireless energy-carrying RSMA network. Background Art

[0002] Intelligent reflecting surface (IRS) is a revolutionary new technology that can effectively improve the resource utilization of wireless communication systems and expand communication coverage. It is composed of a large number of low-cost passive reflective elements. By adjusting the coefficients of the reflective elements at the software level, the phase shift of the reflected signal can be dynamically adjusted, thereby reconfiguring the wireless propagation channel and improving the performance of the communication system.

[0003] The dual purpose of radio signals for wireless information and power transmission (Simultaneous wireless information and power transfer, SWIPT) has always been a research hotspot in wireless communications, that is, wireless energy transmission. A typical SWIPT working scenario is that the base station broadcasts radio signals with a certain power to heterogeneous user terminals located near the base station. These users can be mainly divided into three categories: one is information receiving users who only require decoding information from received radio signals, the other is energy receiving users who only require converting received radio signals into usable electrical energy, and the last is mixed receiving users who require both decoding information from radio signals and converting radio signals into electrical energy. Hybrid receiving users are the most common type of users, and they mainly have two special signal receivers. One is a time-splitting (TS) receiver, that is, within a given working time span, the radio signals received in a part of the time are all used for information decoding, and the radio signals received in the other part of the time are all used for power conversion; the other is a power-splitting (PS) receiver, that is, the received radio signal will be divided into two parts, one part of the signal is used for information decoding, and the other part of the signal is used for power conversion, and the two processes are carried out simultaneously.

[0004] Rate-splitting Multiple Access (RSMA) is a more flexible and powerful multiple access technology in multi-antenna multi-user downlink communication systems. The existing common space-division multiple access (SDMA) and non-orthogonal multiple access (NOMA) technologies are both special cases. SDMA is "total interference as noise", NOMA is "total interference decoding and elimination", and RSMA is a compromise solution, that is, "only part of the interference is considered as noise, and only part of the interference is decoded and eliminated". Therefore, in different communication environments, RSMA can provide more flexible design space for improving communication rate and service quality, as well as reducing implementation complexity.

[0005] In wireless communication systems, due to the broadcast characteristics of wireless channels, the information sent by the transmitter to legitimate users is easily received by illegal eavesdroppers, resulting in information leakage. In traditional wireless communication systems, secure communication can usually be achieved through the following three methods: first, using the multiple antennas of the transmitter to achieve destructive superposition of the transmitted signal at the eavesdropper, that is, to achieve zero-force (ZF) transmission; second, using a specific interference signal to reduce the strength of the transmitted signal received by the eavesdropper; third, using artificially generated random noise, which also interferes with the eavesdropper to reduce the strength of its received transmitted signal. In the SWIPT system, since the energy receiving user is usually closer to the transmitter than the information receiving user or the mixed receiving user, it has a better channel and is easier to eavesdrop on the information sent to the information receiving user or the mixed receiving user.

[0006] Currently, wireless energy-carrying RSMA networks can enhance the secure communication performance of wireless communication systems by deploying IRS near legitimate users, but it is often impossible to take into account the energy consumption of the system at the same time. Summary of the invention

[0007] The present invention provides an intelligent reflective surface assisted secure communication method for a wireless energy-carrying RSMA network in order to overcome the technical defect that the current wireless energy-carrying RSMA network cannot enhance the secure communication performance of the wireless communication system while taking into account the energy consumption of the system.

[0008] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0009] A smart reflective surface assisted secure communication method for a wireless energy carrying RSMA network comprises the following steps:

[0010] S1: Build the initial target optimization model;

[0011] The target optimization model aims to maximize the minimum security energy efficiency in the wireless communication system, and jointly optimizes the beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix in the wireless communication system;

[0012] S2: Given the user power division coefficient and the IRS reflection phase shift matrix, the optimal beamforming vector and public safety rate are obtained through iterative optimization.

[0013] By giving the beamforming vector, public safety rate and IRS reflection phase shift matrix, the optimal user power division coefficient is obtained through iterative optimization.

[0014] By giving the beamforming vector, public safety rate and user power division coefficient, the optimal IRS reflection phase shift matrix is ​​obtained through iterative optimization.

[0015] S3: Perform wireless secure communication based on the optimal beamforming vector, public security rate, user power division coefficient, and IRS reflection phase shift matrix.

[0016] In the above scheme, with the goal of maximizing the minimum security energy efficiency in the wireless communication system, the beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix in the wireless communication system are jointly optimized to obtain the optimal beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix for wireless security communication. It is suitable for any number of legitimate users, any number of potential eavesdroppers, any energy collection requirements of legitimate users and any energy collection requirements of potential eavesdroppers, and not only improves the security communication rate between legitimate users in the wireless communication system, but also saves the energy consumption of the wireless communication system.

[0017] Preferably, the beamforming vectors include {w k} and w c ;

[0018] Among them, {w k} represents the set of beamforming vectors corresponding to all private symbol streams,

[0019] w k represents the private symbol stream s transmitted by AP to the kth legal user pk The corresponding beamforming vector, w c Represents the public symbol stream s transmitted by AP to all legitimate users c The corresponding beamforming vector.

[0020] Preferably, the initial target optimization model is

[0021]

[0022]

[0023]

[0024]

[0025] E e ≥γ e ,

[0026]

[0027]

[0028]

[0029] in, represents the set of public security rates assigned by AP to each user, {ρ k} represents the set of user power division coefficients, ρ k represents the power division coefficient of the kth user, Φ represents the reflection phase shift matrix of IRS, represents the private security rate of the kth legal user, K represents the number of legal users, represents the public security rate assigned by the AP to the kth legal user, represents the public security rate assigned by AP to the i-th legal user, R c Indicates the rate at which the AP sends super public information. represents the rate at which a potential eavesdropper decodes super public information, E k represents the energy received by the kth legitimate user, γ k represents the minimum energy collected by each legitimate user, γ e represents the minimum energy collected by a potential eavesdropper, E e represents the energy received by the potential eavesdropper, θ m represents the phase of the mth reflection unit of IRS, P max Indicates the maximum transmit power actually available to the AP.

[0030] Preferably, a slack variable q is introduced to represent the lower bound of the target optimization model, and the target optimization model is equivalent to:

[0031]

[0032]

[0033]

[0034]

[0035]

[0036] E e ≥γ e ,

[0037]

[0038]

[0039]

[0040] Preferably, given {ρ k} and Φ, the target optimization model is transformed into the optimization w c , {w k}and Submodel of:

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] E e ≥γ e ,

[0047]

[0048] Preferably, it also includes optimizing w c , {w k}and The sub-model is transformed into the optimization w c , {w k}and The joint convex optimization model of:

[0049]

[0050] stf ≥ q (t) u (t) +q (t) (uu (t) )+u (t) (qq (t) )

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058] E e ≥γ e ,

[0059]

[0060] rank(W)=1

[0061] Where (t) represents the number of iterations, the slack variable u represents the upper bound of the transmit power, and the slack variable δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of the kth legitimate user, and the slack variable ζ e represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, the slack variable f represents the minimum safe rate of a legitimate user, and q (t) 、u (t) They represent the corresponding first-order Taylor expansion points, F 1k (W) is a concave function with respect to W, is a convex function F 2k The lower bound of (W) is Denotes a concave function F 1ek The upper bound of (W), F 2ek (W) represents a convex function with respect to W, F 3k (W) is another concave function with respect to W, is a convex function F 4k The lower bound of (W) is Denotes a concave function F 3e The upper bound of (W), F 4e (W) represents a convex function with respect to W, and the set Indicates w k The conjugate transpose of represents the conjugate transpose of the matrix w1, Indicates w c , tr(·) represents the trace of the matrix, and rank(·) represents the rank of the matrix.

[0062] Preferably, the optimal user power division factor is:

[0063]

[0064] Among them, η k represents the energy harvesting efficiency of the kth legal user, G k represents a mixed channel, Represents G k The conjugate transpose of Represents the matrix w i The conjugate transpose of Indicates w c The conjugate transpose of .

[0065] Preferably, given w c , {w k}、 and {ρ k}, the target optimization model is transformed into a sub-model for optimizing Φ:

[0066]

[0067]

[0068]

[0069]

[0070] E e ≥γ e ,

[0071]

[0072] The slack variable f represents the minimum safe rate of a legitimate user.

[0073] Preferably, the method further includes converting the sub-model for optimizing Φ into a joint convex optimization model for optimizing Φ:

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080] E e ≥γ e ,

[0081]

[0082]

[0083] Among them, the matrix S = ss H ,s H represents the conjugate transpose of vector s, the vector Vector v = [v1, ..., v m , ..., v M ] H , the mth element of vector v m={1,2,...,M},slack variable δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of the kth legitimate user, and the slack variable ζ e represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, J 1k (S) is a concave function with respect to S, Denotes the convex function J 2k The lower bound of (S), Denotes the concave function J 1ek The upper bound of (S), J 2ek (S) is a convex function with respect to S, J 3k (S) is another concave function with respect to S, Denotes the convex function J 4k The lower bound of (S), Denotes the concave function J 3e The upper bound of (S), J 4e (S) is a convex function with respect to S, (S) m，m represents the m-th diagonal element of the matrix S, rank(·) represents the rank of the matrix, and M represents the number of reflection units of the IRS.

[0084] Preferably, an interior point method is used to obtain the optimal solution of the joint convex optimization model.

[0085] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0086] The present invention provides an intelligent reflective surface-assisted secure communication method for a wireless energy-carrying RSMA network. The method aims to maximize the minimum secure energy efficiency in a wireless communication system, jointly optimizes beamforming vectors, public security rates, user power division coefficients, and an IRS reflection phase shift matrix in the wireless communication system, and obtains the optimal beamforming vectors, public security rates, user power division coefficients, and an IRS reflection phase shift matrix for wireless secure communication. The method is suitable for any number of legitimate users, any number of potential eavesdroppers, any energy collection requirements of legitimate users, and any energy collection requirements of potential eavesdroppers, and not only improves the secure communication rate between legitimate users in the wireless communication system, but also saves the energy consumption of the wireless communication system at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 It is a flowchart of the implementation steps of the technical solution of the present invention;

[0088] Figure 2 A schematic diagram of a wireless communication system in the present invention;

[0089] Figure 3 It is a comparative schematic diagram of the minimum safe energy efficiency of the system in the present invention as the transmission power of the AP end changes;

[0090] Figure 4 It is a comparative schematic diagram of the minimum security energy efficiency of the system in the present invention as the position of the potential eavesdropper changes;

[0091] Figure 5 It is a comparative schematic diagram of the minimum safe energy efficiency of the system in the present invention and the change of the energy collection demand of the legal user;

[0092] Figure 6 It is a comparative schematic diagram of the minimum security energy efficiency of the system in the present invention as the energy collection requirements of potential eavesdroppers change. DETAILED DESCRIPTION

[0093] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0094] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0095] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0096] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0097] Example 1

[0098] like Figure 1As shown, a smart reflective surface assisted secure communication method for a wireless energy carrying RSMA network includes the following steps:

[0099] S1: Build the initial target optimization model;

[0100] The target optimization model aims to maximize the minimum security energy efficiency in the wireless communication system, and jointly optimizes the beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix in the wireless communication system;

[0101] S2: Given the user power division coefficient and the IRS reflection phase shift matrix, the optimal beamforming vector and public safety rate are obtained through iterative optimization.

[0102] By giving the beamforming vector, public safety rate and IRS reflection phase shift matrix, the optimal user power division coefficient is obtained through iterative optimization.

[0103] By giving the beamforming vector, public safety rate and user power division coefficient, the optimal IRS reflection phase shift matrix is ​​obtained through iterative optimization.

[0104] S3: Perform wireless secure communication based on the optimal beamforming vector, public security rate, user power division coefficient, and IRS reflection phase shift matrix.

[0105] In the specific implementation process, with the goal of maximizing the minimum security energy efficiency in the wireless communication system, the beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix in the wireless communication system are jointly optimized to obtain the optimal beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix for wireless security communication. It is suitable for any number of legitimate users, any number of potential eavesdroppers, any energy collection requirements of legitimate users and any energy collection requirements of potential eavesdroppers, and not only improves the security communication rate between legitimate users in the wireless communication system, but also saves the energy consumption of the wireless communication system.

[0106] Example 2

[0107] A smart reflective surface assisted secure communication method for a wireless energy-carrying RSMA network,

[0108] It includes a RSMA (Rate Splitting Multiple Access Technology)-based SWIPT (dual use of radio signals for wireless information and power transmission) downlink wireless communication system assisted by IRS, such as Figure 2As shown in FIG, the wireless communication system includes a wireless access point (AP), an IRS, K (K>1) legal users (User) and a potential eavesdropper (Eve). The K legal users are all mixed receiving users in SWIPT, that is, they have both information receiving and energy collection requirements; the potential eavesdropper is an energy collection user, which only has energy collection requirements; the AP is equipped with N r (N T >1) transmitting antennas, all legitimate users and potential eavesdroppers are equipped with only one antenna, and the IRS has M reflecting units.

[0109] Define the channel between AP and IRS as The channel between AP and legitimate user k is The channel between AP and potential eavesdropper is The channel between IRS and legitimate user k is The channel between IRS and potential eavesdroppers is Here represents a set of x×y dimensional complex valued matrices, and all channel state information is known. The reflection phase shift matrix of IRS is expressed as Φ=diag(v H ), v = [q1, q2, ..., q M ] H (X H represents the conjugate transpose of the matrix X), where β m ∈[0, 1], θ m ∈[0,2π),β m and θ m Respectively represent the reflection coefficient of the mth reflection unit, i.e., amplitude and phase, where m∈{1,...,M}. In order to achieve the maximum reflection power gain, let |β m |=1, so q m Satisfy |q m |=1.

[0110] In order to implement rate splitting (RS) at the AP, the uncoded information to be sent needs to be segmented, that is, the uncoded information u planned to be sent to the legitimate user k is k Will be divided into two parts, namely the public part u ck and the private part u pk , the public parts of all legitimate users’ uncoded information are packaged into a super public information u c =(u c1 , ..., u cK ). When performing linear precoding, the super public information u c Encoded into a public symbol stream The private part u of each legitimate user pk Encoded separately into a private symbol stream Without loss of generality, we can let E[|s c | 2 ]=1 and Therefore, the signal transmitted by the AP is expressed as in represents the private symbol stream s transmitted by AP to the kth legal user pk The corresponding beamforming vector, Represents the public symbol stream s transmitted by AP to all legitimate users c The corresponding beamforming vector.

[0111] Each legitimate user uses the power splitting method to harvest energy, that is, the receiver equipped by the legitimate user divides the received signal into two parts, one for energy harvesting (EH) and the other for information decoding (ID). Since potential eavesdroppers have no need to receive information, all received signals will be used for energy harvesting.

[0112] Use ρ k represents the power splitting coefficient of the legitimate user k performing power splitting, then the part of the signal received by the legitimate user k that is actually used for information decoding is

[0113] n k It means that the mean value received by the kth user is 0 and the variance is Additive Gaussian white noise. For potential eavesdroppers, since the power of the received information is not divided, when they eavesdrop on the information, the signal that can be used for information decoding is

[0114] where n e The potential eavesdropper receives a mean of 0 and a variance of Additive white Gaussian noise.

[0115] Each legitimate user needs to perform SIC once, that is, decode the super public information, extract the required public information from it, and then remove the super public information from the received information. In this way, when the legitimate user decodes the private information, there is no interference caused by the super public information, and only interference caused by the private information of other legitimate users exists. The signal-to-interference-to-noise ratio of the kth legitimate user decoding the super public information is expressed as:

[0116]

[0117] When the kth legitimate user performs SIC to eliminate the influence of super public information, the signal-to-interference-noise ratio of the decoded private information is expressed as:

[0118] Therefore, the super public information rate that the kth legitimate user can achieve is

[0119] The achievable private information rate is

[0120] For potential eavesdroppers, assume that they also try to decode the super public information first, then perform SIC and then try to decode the private information. The rate at which potential eavesdroppers decode the super public information is

[0121]

[0122] In order to make the public beam interfere with potential eavesdroppers, the super public information rate that legitimate users can achieve is required. and the super public information rate that potential eavesdroppers can achieve With the following relationship:

[0123]

[0124]

[0125] That is, when the AP is at rate R c When sending super public information, all legitimate users can successfully decode s c , and a potential eavesdropper cannot successfully decode s c , which will result in potential eavesdroppers being unable to successfully apply SIC. Define the public security rate assigned to the kth legitimate user as have

[0126]

[0127]

[0128] in When a potential eavesdropper attempts to decode the private information s of legitimate user k pk When , since SIC cannot be applied to the beam of super public information, it is still interfered by the beam, and the corresponding rate is

[0129]

[0130] Therefore, the private security rate of legitimate user k is

[0131] During the entire communication process, the actual security rate that the kth legitimate user can achieve is

[0132]

[0133] Considering that both legitimate users and potential eavesdroppers adopt the linear energy collection model, the energy received by the kth legitimate user is:

[0134]

[0135] Here η k ∈(0, 1] represents the energy collection efficiency of the kth legitimate user. The energy received by the potential eavesdropper is:

[0136]

[0137] Here η e ∈(0,1] Similarly η k Represents the energy harvesting efficiency.

[0138] This embodiment includes the following steps:

[0139] S1: Build the initial target optimization model;

[0140] The target optimization model aims to maximize the minimum security energy efficiency in the wireless communication system, and jointly optimizes the beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix in the wireless communication system;

[0141] More specifically, the beamforming vectors include {w k} and w c ;

[0142] Among them, {w k} represents the set of beamforming vectors corresponding to all private symbol streams,

[0143] w k represents the private symbol stream s transmitted by AP to the kth legal user pk The corresponding beamforming vector, w c Represents the public symbol stream s transmitted by AP to all legitimate users c The corresponding beamforming vector.

[0144] More specifically, the initial target optimization model is

[0145]

[0146]

[0147]

[0148]

[0149] E e ≥γ e ,

[0150]

[0151]

[0152]

[0153] in, represents the set of public security rates assigned by AP to each user, {ρ k} represents the set of user power division coefficients, ρ k represents the power division coefficient of the kth user, Φ represents the reflection phase shift matrix of IRS, represents the private security rate of the kth legal user, K represents the number of legal users, represents the public security rate assigned by the AP to the kth legal user, represents the public security rate assigned by AP to the i-th legal user, R c Indicates the rate at which the AP sends super public information. represents the rate at which a potential eavesdropper decodes super public information, E k represents the energy received by the kth legitimate user, γ k represents the minimum energy collected by each legitimate user, γ e represents the minimum energy collected by a potential eavesdropper, E e represents the energy received by the potential eavesdropper, θ m represents the phase of the mth reflection unit of IRS, P max Indicates the maximum transmit power actually available to the AP.

[0154] In the specific implementation process, constraints It means that all legitimate users can successfully decode the super public information and can prevent the eavesdropper Eve from decoding the corresponding information. The RSMA mechanism can work normally and obtain the corresponding public security rate. Indicates that the public security rate value assigned to each legitimate user should be a non-zero value to avoid leakage and eavesdropping of public information; Constraint E k ≥γ k , Indicates that the energy collected by each legitimate user should not be less than γ k ; Constraint E e ≥γ e Indicates that the energy collected by the eavesdropper Eve should not be less than γ e ; Constraint 0≤ρ k ≤1, Indicates that the power division coefficient should conform to the actual energy conservation situation; constraint 0≤θ m ≤2π, Indicates the actual adjustable phase range of each reflector unit on the IRS; constraint Indicates that the maximum transmit power actually available to the AP is P max .

[0155] More specifically, the slack variable q is introduced to represent the lower bound of the target optimization model, and the target optimization model is equivalent to:

[0156]

[0157]

[0158]

[0159]

[0160]

[0161] E e ≥γ e ,

[0162]

[0163]

[0164]

[0165] S2: Given the user power division coefficient and the IRS reflection phase shift matrix, the optimal beamforming vector and public safety rate are obtained through iterative optimization.

[0166] More specifically, given a set of user power splitting factors {ρ k} and the reflection phase shift matrix Φ of IRS, the target optimization model is transformed into the optimization w c , {w k}and Submodel of:

[0167]

[0168]

[0169]

[0170]

[0171]

[0172] E e≥γ e ,

[0173]

[0174] Constraints To transform:

[0175] The slack variable f is introduced to represent the minimum safe rate of the legal user, and the slack variable u is introduced to represent the upper bound of the transmission power, that is, Then we can get the constraint The lower bound of Then Constraint It can be expressed equivalently using the new constraints as

[0176]

[0177]

[0178]

[0179] For the constraints Use the first-order Taylor expansion to approximate qu and convexify it into an affine function, that is, where q (t) and u (t) is the corresponding first-order Taylor expansion point, and other constraints will be convexified later.

[0180] Constraints To transform:

[0181] make Then constraint E e ≥γ e It can be expressed as

[0182] Similarly, constraints It can be expressed as

[0183] For constraint E k ≥γ k , and E e ≥γ e To transform:

[0184] make and Then there is

[0185]

[0186] Similarly

[0187]

[0188] Constraints To transform:

[0189] for The expression of , temporarily removing the symbol have

[0190] Introducing slack variables δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of legitimate user k, that is, Then Constraint It can be expressed as

[0191]

[0192] also, It can be expressed as

[0193]

[0194] Where W represents the set And there is

[0195]

[0196]

[0197] Then Constraint It can be further expressed as:

[0198]

[0199] because In a constraint The left side of the inequality sign “≥” in , according to the requirements of convex constraints, should be a joint concave function. Obviously, is an affine function, which is both a convex function and a concave function. 1k (W) is a concave function with respect to W, F 2k (W) is a convex function with respect to W, so F 2k The (W) function needs further processing. To this end, the continuous convex approximation method is used to solve the constraint by iteration. An approximate solution of 2k (W) Perform a first-order Taylor expansion to find F 2k The lower bound of (W) as follows:

[0200]

[0201] in, Here W (t)represents the feasible point of the first-order Taylor expansion, and the superscript (t) represents the number of iterations of the optimization variable. Alternative F 2k (W), the constraint Convert to

[0202]

[0203] because is an affine function, so the above constraint is a convex constraint, and the solution W that satisfies the above constraint also satisfies the constraint

[0204] Similarly, for the newly added upper bound constraint on the eavesdropping rate have

[0205]

[0206] Then the above constraints can be expressed as F 1ek (W)+F 2ek (W)≤δ k , Easy to know F 1ek (W) is a concave function with respect to W, F 2ek (W) is a convex function with respect to W, then for F 1ek (W) in W (t) Performing a first-order Taylor expansion gives the upper bound

[0207]

[0208] in, Here W (t) represents the feasible point of the first-order Taylor expansion, and the superscript (t) represents the number of iterations of the optimization variable. Alternative F 1ek (W), the constraint Convert to

[0209]

[0210] Constraints To transform:

[0211] Introducing the slack variable ζ e represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, that is, Then Constraint It can be expressed as

[0212] because The above can be transformed into

[0213] right Deformation

[0214]

[0215] in

[0216]

[0217]

[0218] Then Constraint It can be expressed as:

[0219]

[0220] Likewise, because The left side of the inequality sign “≥” in constraint (3c) should be a jointly concave function according to the requirement of convex constraints, because F 3k (W) is a concave function with respect to W, F 4k (W) is a convex function with respect to W, so F 4k The (W) function needs further processing. To this end, the continuous convex approximation method is also used to solve the constraint by iteration. An approximate solution of 4k (W) Perform a first-order Taylor expansion to find F 4k The lower bound of (W) as follows:

[0221]

[0222] in, Here W (t) represents the feasible point of the first-order Taylor expansion, and the superscript (t) represents the number of iterations of the optimization variable. Alternative F 4k (W), the constraint Convert to

[0223]

[0224] because is an affine function, so the above constraint is a convex constraint, and the solution W that satisfies the above constraint also satisfies the constraint

[0225] The upper bound of the rate of eavesdropping on public information Empathy Constraint The transformation of

[0226]

[0227] F 3e (W) is expanded by the first order Taylor.

[0228]

[0229] in, use Alternative F 3ek (W), the constraint Convert to

[0230]

[0231] More specifically, after the above convexification process, it also includes optimizing w c , {w k}and The sub-model is transformed into the optimization w c , {w k}and The joint convex optimization model of:

[0232]

[0233] stf ≥ q (t) u (t) +q (t) (uu (t) )+u (t) (qq (t) )

[0234]

[0235]

[0236]

[0237]

[0238]

[0239]

[0240]

[0241] E e ≥γ e ,

[0242]

[0243] rank(W)=1

[0244] Where (t) represents the number of iterations, the slack variable u represents the upper bound of the transmit power, and the slack variable δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of the kth legitimate user, and the slack variable ζe represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, the slack variable f represents the minimum safe rate of a legitimate user, and q (t) 、u (t) They represent the corresponding first-order Taylor expansion points, F 1k (W) is a concave function with respect to W, is a convex function F 2k The lower bound of (W) is Denotes a concave function F 1ek The upper bound of (W), F 2ek (W) represents a convex function with respect to W, F 3k (W) is another concave function with respect to W, is a convex function F 4k The lower bound of (W) is Denotes a concave function F 3e The upper bound of (W), F 4e (W) represents a convex function with respect to W, and the set Indicates w k The conjugate transpose of represents the conjugate transpose of the matrix w1, Indicates w c The conjugate transpose of , tr(·) represents the trace of the matrix, and rank(·) represents the rank of the matrix. W ≥ 0 means that every element in the set W is a semi-positive definite matrix, that is, W m ≥0, m∈{c, 1, ..., k, ..., K}, constraint (4k) means that After deformation, W m After obtaining the optimal solution of W, the standard Gaussian random method is applied to restore the W matrix to the w vector.

[0245] By giving the beamforming vector, public safety rate and IRS reflection phase shift matrix, the optimal user power division coefficient is obtained through iterative optimization.

[0246] More specifically, the optimal user power split factor is:

[0247]

[0248] Among them, η k represents the energy harvesting efficiency of the kth legal user, G k represents a mixed channel, Represents G k The conjugate transpose of Represents the matrix w i The conjugate transpose of Indicates w c The conjugate transpose of .

[0249] In the specific implementation process, the optimal user power division coefficient is proved by contradiction: Taking the kth user as an example, if there is an optimal solution Make the constraint E k ≥γ k The inequality is strictly established, that is, Then we can always find another feasible solution satisfy because and It is about ρ k is a strictly monotonically increasing function, so from the above assumptions we can know that By constraint and constraints We can know the feasible solution Able to achieve more feasible solutions The larger the objective function value, the It is not the optimal solution to the problem, ρ k The optimal solution must be such that the constraint E k ≥γ k The value that holds true in the form of an equation, that is The proof is complete.

[0250] By giving the beamforming vector, public safety rate and user power division coefficient, the optimal IRS reflection phase shift matrix is ​​obtained through iterative optimization.

[0251] More specifically, given w c , {w k}、 and {ρ k}, the target optimization model is transformed into a sub-model for optimizing Φ:

[0252]

[0253]

[0254]

[0255]

[0256] E e ≥γ e ,

[0257]

[0258] The slack variable f represents the minimum safe rate of a legitimate user.

[0259] For constraint E k ≥γk , and E e ≥γ e To transform:

[0260] First, mix the channel G k can be re-expressed as:

[0261]

[0262] in, Then constraint E k ≥γ k , can be written as:

[0263]

[0264] Among them are

[0265]

[0266] in S=ss H ,

[0267] The same can be said

[0268]

[0269] Constraint E k ≥γ k , So it can be rewritten as

[0270]

[0271] For the mixed channel G e , using the same method, it can be equivalently expressed as

[0272]

[0273] Then there is

[0274]

[0275]

[0276] in Then constraint E e ≥γ e It can be expressed as

[0277]

[0278] Constraints To transform:

[0279]

[0280] in

[0281]

[0282]

[0283] For potential eavesdroppers to eavesdrop on private information, a slack variable is introduced to represent the upper bound of its rate, that is, Then Constraint It can be expressed as

[0284] Obviously J 1k (S) is a concave function with respect to S and J 2k (S) is a convex function with respect to S, so in order to construct a convex constraint, the constraint needs to be further transformed.

[0285] To J 2k (S) also uses the continuous convex approximation method to solve the problem of satisfying the constraints through iteration. An approximate solution of 2k (S) Perform a first-order Taylor expansion to find J 2k The lower bound of (S) as follows:

[0286]

[0287] in, Here S (t) represents the feasible point of the first-order Taylor expansion, and the superscript (t) represents the number of iterations of the optimization variable. Alternative J 2k (S), the constraint Convert to

[0288]

[0289] Because J 1k (S) is a concave function and is an affine function, so the above constraint is a convex constraint, and the solution S that satisfies the above constraint also satisfies the constraint

[0290] For the newly added constraints The same method can be used to transform

[0291]

[0292] in

[0293]

[0294]

[0295]

[0296] in,

[0297] Constraints Perform deformation.

[0298] Adoption and Constraints The same method is used to convexify.

[0299]

[0300] in

[0301]

[0302]

[0303] Among them J 3k (S) is a concave function with respect to S, J 4k (S) is a convex function of S. The introduction of slack variables indicates that the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user is Then Constraint It can be expressed as

[0304]

[0305] in For J 4k (S) in a feasible solution S (t) The first-order Taylor expansion at , which indicates its lower bound, is

[0306]

[0307] in

[0308] For adding constraints It can be expressed as

[0309]

[0310] in

[0311]

[0312]

[0313]

[0314] in,

[0315] For the constraint 0≤θ m ≤2π, Perform deformation.

[0316] Constrain 0≤θ m ≤2π, It shows the adjustable range of the phase of each reflector unit on the IRS, which is β m =1, we know that the constraint 0≤θ m ≤2π, It can be equivalently expressed as

[0317]

[0318] And v=[q1,q2,...,q M ] H , S=ss H , then the constraint 0≤θ m ≤2π, It can be further expressed equivalently as

[0319]

[0320] It means that the value of all elements on the diagonal of matrix S is equal to the real number 1.

[0321] More specifically, after the above convexification process, the sub-model for optimizing Φ is also transformed into a joint convex optimization model for optimizing Φ:

[0322]

[0323]

[0324]

[0325]

[0326]

[0327]

[0328] E e ≥γ e ,

[0329]

[0330]

[0331] Among them, the matrix S = ss H ,s Hrepresents the conjugate transpose of vector s, the vector Vector v = [v1, ..., v m , …, v M ] H , the mth element of vector v m={1,2,...,M},slack variable δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of the kth legitimate user, and the slack variable ζ e represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, J 1k (S) is a concave function with respect to S, Denotes the convex function J 2k The lower bound of (S), Denotes the concave function J 1ek The upper bound of (S), J 2ek (S) is a convex function with respect to S, J 3k (S) is another concave function with respect to S, Denotes the convex function J 4k The lower bound of (S), Denotes the concave function J 3e The upper bound of (S), J 4e (S) is a convex function with respect to S, (S) m，m represents the m-th diagonal element of the matrix S, rank(·) represents the rank of the matrix, and M represents the number of reflection units of the IRS.

[0332] More specifically, an interior point method is used to obtain the optimal solution of the joint convex optimization model.

[0333] S3: Perform wireless secure communication based on the optimal beamforming vector, public security rate, user power division coefficient, and IRS reflection phase shift matrix.

[0334] Example 3

[0335] In this embodiment, the intelligent reflective surface assisted secure communication method for a wireless energy carrying RSMA network (marked as Proposed) is compared with the following comparison scheme:

[0336] (1) NoIRS: Based on the intelligent reflective surface assisted secure communication method for a wireless energy-carrying RSMA network, the IRS deployment is removed to compare and verify the superiority of deploying the IRS.

[0337] (2) MS: Based on the intelligent reflector-assisted secure communication method for a wireless energy-carrying RSMA network, the objective function is changed to maximize That is, only the maximum and minimum safe rates of the system are considered to compare and explore the safe energy efficiency performance that the system can achieve under different objective functions.

[0338] (3) KJam: Based on the intelligent reflector-assisted secure communication method for a wireless power-carrying RSMA network, the RSMA multiple access scheme is replaced with an SDMA multiple access scheme, and an interference signal (Jamming) is introduced to assist in achieving secure communication. It is assumed that the legitimate user knows the interference signal and can remove the interference caused by the interference signal in the received signal, so as to compare and verify the superiority of RSMA in achieving secure communication.

[0339] (4) KJam-NoIRS: Remove the deployment of IRS based on KJam to verify the superiority of deploying IRS.

[0340] (5) KJam-MS: Based on KJam, the objective function is changed to maximization

[0341] (6) UkJam: Based on KJam, it is assumed that users cannot know the interference signal and therefore cannot remove the interference it brings. This is done to verify whether the interference signal can eliminate the impact on security energy efficiency under the SDMA multiple access method.

[0342] (7) NoJam: Based on KJam, the introduction of any interference signal is not considered to verify the superiority of applying interference signals in achieving secure communications under the SDMA multiple access mode.

[0343] like Figure 3 As shown, the following conclusions can be drawn:

[0344] (1) Except for the MS scheme and the KJam-MS scheme, the curves of other schemes have no significant changes, which means that when the available transmit power is increased, the system cannot use this part of energy to obtain a sufficiently large safety rate increase, that is, using this part of energy will lead to a decrease in the system's minimum safety energy efficiency. Therefore, the system chooses not to use the additional available transmit power, resulting in an almost constant safety energy efficiency. Among all the schemes, the Proposed scheme always has the best performance.

[0345] (2) Comparing the proposed scheme with the NoIRS scheme, as well as the KJam scheme and the KJam-NoIRS scheme, it can be proved that IRS is superior in improving the system safety energy efficiency. Furthermore, the curve intervals of the proposed scheme and the NoIRS scheme are larger than those of the KJam scheme and the KJam-NoIRS scheme. This also shows that compared with the traditional SDMA system, the RSMA system can more effectively improve the system safety energy efficiency performance with the assistance of IRS.

[0346] (3) For the MS scheme and the KJam-MS scheme, the efficiency decreases with the increase of available transmission power. This is because although using more transmission power increases the security rate, it also brings more system energy consumption, which leads to a decrease in the system security energy efficiency. The MS scheme always has better security energy efficiency than the KJam-MS scheme, which verifies the superiority of RSMA in the application of secure communications.

[0347] (4) For the UkJam scheme, since the user cannot know the interference signal, the application of the interference signal will bring additional interference to the user's decoding information, thus causing the system performance to deteriorate. The UkJam scheme has only a slight performance improvement compared to the NoJam scheme. This shows that when the user cannot remove the interference caused by the interference signal, the application of the interference signal will not bring much performance improvement to the system's security energy efficiency.

[0348] like Figure 4 As shown in the figure, the increase in the horizontal coordinate of the potential eavesdropper means that it is getting farther and farther away from the AP. The following conclusions can be drawn:

[0349] (1) For the Proposed solution and the KJam solution, in x e When the value increases, the curves first decrease slowly and then decrease rapidly. This is because the legitimate user is deployed near 9m. When the potential eavesdropper is between 4m and 8m, the system needs to meet the legitimate user's security communication needs and energy collection needs. The potential eavesdropper is closer to the AP and can receive a stronger signal. Therefore, its energy collection needs can be more easily met. Therefore, even if x e As the distance from the AP increases, the signal loss increases, but the impact on system performance is small; e When changing from 10m to 18m, the curve of the above scheme drops rapidly. This is because x e The increased distance from the AP makes it difficult to meet the energy collection requirements of potential eavesdroppers, so more security rates need to be sacrificed in exchange for energy collection from potential eavesdroppers, resulting in a decrease in the security energy efficiency of the system. e During the increase process, the curve of the Proposed solution is always larger than the curve of the KJam solution, especially at x e When the distance to the AP is small, this verifies the superiority of RSMA in achieving security energy efficiency; in addition, the curve of the proposed scheme is always larger than that of the NoIRS scheme, especially at x e This also verifies the superiority of deploying IRS when the distance to the AP is small.

[0350] (2) By comparing the Proposed solution and the KJam solution, we can see that when x eWhen the distance from the AP is far enough, the difference between RSMA multiple access and SDMA multiple access is narrowing. Considering that RSMA requires additional encoding and SIC decoding, SDMA with lower complexity can be used to achieve performance similar to RSMA under this condition, but RSMA is still the best choice.

[0351] like Figure 5 , 6 As shown, the following conclusions can be drawn:

[0352] (1) For Figure 5 and Figure 6 ,The curves of all schemes have a certain degree of decline. This is because the increase of user ,energy harvesting requirements will result in less energy being used for information decoding, ,which will lead to a decrease in the achievable security rate for legitimate ,users, and in turn to a decrease in the security energy efficiency of the ,system.

[0353] (2) Comparison Figure 5 and Figure 6 , we can see that under the same energy harvesting demand change, γ e The increase of leads to a faster drop in the curve of the Proposed solution, which illustrates the energy harvesting requirement of the potential eavesdropper γ e The increase in energy harvesting demand compared to legitimate users γ k The increase has a greater impact on the safety energy efficiency performance of the system.

[0354] (3) In all the scheme comparisons, the proposed scheme is always the one that achieves the best system security and energy efficiency.

[0355] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A smart reflective surface assisted secure communication method for a wireless energy carrying RSMA network, characterized in that: The following steps are involved: S1: Build the initial target optimization model; The target optimization model aims to maximize the minimum security energy efficiency in the wireless communication system, and jointly optimizes the beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix in the wireless communication system; S2: Given the user power division coefficient and the IRS reflection phase shift matrix, the optimal beamforming vector and public safety rate are obtained through iterative optimization. By giving the beamforming vector, public safety rate and IRS reflection phase shift matrix, the optimal user power division coefficient is obtained through iterative optimization. By giving the beamforming vector, public safety rate and user power division coefficient, the optimal IRS reflection phase shift matrix is ​​obtained through iterative optimization. S3: Perform wireless secure communication based on the optimal beamforming vector, public security rate, user power division coefficient and IRS reflection phase shift matrix; The beamforming vectors include {w k } and w c ; Among them, {w k } represents the set of beamforming vectors corresponding to all private symbol streams, w k represents the private symbol stream s transmitted by AP to the kth legal user pk The corresponding beamforming vector, w c Represents the public symbol stream s transmitted by AP to all legitimate users c The corresponding beamforming vector; The initial target optimization model is in, represents the set of public security rates assigned by AP to each user, {ρ k } represents the set of user power division coefficients, ρ k represents the power division coefficient of the kth user, Φ represents the reflection phase shift matrix of IRS, represents the private security rate of the kth legal user, K represents the number of legal users, represents the public security rate assigned by the AP to the kth legal user, represents the public security rate assigned by AP to the i-th legal user, R c Indicates the rate at which the AP sends super public information. represents the rate at which a potential eavesdropper decodes super public information, E k represents the energy received by the kth legitimate user, γ k represents the minimum energy collected by each legitimate user, γ e represents the minimum energy collected by a potential eavesdropper, E e represents the energy received by the potential eavesdropper, θ m represents the phase of the mth reflection unit of IRS, P max Indicates the maximum transmit power actually available to the AP; The slack variable q is introduced to represent the lower bound of the target optimization model, and the target optimization model is equivalent to: The optimal user power division factor is: Among them, η k represents the energy harvesting efficiency of the kth legal user, G k represents a mixed channel, Represents G k The conjugate transpose of Represents the matrix w i The conjugate transpose of Indicates w c The conjugate transpose of .

2. The intelligent reflective surface assisted secure communication method of a wireless energy carrying RSMA network according to claim 1, characterized in that: Given {ρ k } and Φ, the target optimization model is transformed into the optimization w c 、{w k }and Submodel of:

3. The intelligent reflective surface assisted secure communication method of a wireless energy carrying RSMA network according to claim 2, characterized in that: It also includes optimizing w c 、{w k }and The sub-model is transformed into the optimization w c 、{w k }and The joint convex optimization model of: Where (t) represents the number of iterations, the slack variable u represents the upper bound of the transmit power, and the slack variable δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of the kth legitimate user, and the slack variable ζ e represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, the slack variable f represents the minimum safe rate of a legitimate user, and q (t) 、u (t) They represent the corresponding first-order Taylor expansion points, F 1k (W) is a concave function with respect to W, is a convex function F 2k The lower bound of (W) is Denotes a concave function F 1ek The upper bound of (W), F 2ek (W) represents a convex function with respect to W, F 3k (W) is another concave function with respect to W, is a convex function F 4k The lower bound of (W) is Denotes a concave function F 3e The upper bound of (W), F 4e (W) represents a convex function with respect to W, and the set Indicates w k The conjugate transpose of represents the conjugate transpose of matrix w1, Indicates w c , tr(·) represents the trace of the matrix, and rank(·) represents the rank of the matrix.

4. The intelligent reflective surface assisted secure communication method of a wireless energy carrying RSMA network according to claim 1, characterized in that: Given w c 、{w k }、 and {ρ k }, the target optimization model is transformed into a sub-model for optimizing Φ: The slack variable f represents the minimum safe rate of a legitimate user.

5. The intelligent reflective surface assisted secure communication method of a wireless energy carrying RSMA network according to claim 4, characterized in that: It also includes transforming the sub-model for optimizing Φ into a joint convex optimization model for optimizing Φ: Among them, the matrix S = ss H , S H represents the conjugate transpose of vector s, the vector Vector v = [v1,...,v m ,...,v M ] H , the mth element of vector v Slack variable δ k represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the private information of the kth legitimate user, and the slack variable ζ e represents the upper bound of the rate at which a potential eavesdropper can eavesdrop on the public information of a legitimate user, J 1k (S) is a concave function with respect to S, Denotes the convex function J 2k The lower bound of (S), Denotes the concave function J 1ek The upper bound of (S), J 2ek (S) is a convex function with respect to S, J 3k (S) is another concave function with respect to S, Denotes the convex function J 4k The lower bound of (S), Denotes the concave function J 3e The upper bound of (S), J 4e (S) is a convex function with respect to S, (S) m,m represents the m-th diagonal element of the matrix S, rank(·) represents the rank of the matrix, and M represents the number of reflection units of the IRS.

6. The intelligent reflective surface assisted secure communication method of a wireless energy carrying RSMA network according to claim 3 or 5, characterized in that: The interior point method is used to obtain the optimal solution of the joint convex optimization model.

Images