NOMA network physical layer security optimization method based on STAR-RIS

Abstract

The present invention relates to a physical layer security optimization method for NOMA networks based on STAR-RIS. The purpose of the present invention is to enhance the security of downlink MISONOMA networks under the condition of imperfect CSI of eavesdroppers. The process is as follows: 1. Initialize parameters; 2. Perform mathematical modeling; 3. Use continuous convex approximation algorithm to solve the base station active beamforming parameters, let m = m + 1 and ω l m =ω l , execute four; four: if m ≥ N SCA , let ω l n =ω l m , m=0, proceed to step 5, if m<N SCA , proceed to step 3; Step 5: Use the continuous convex approximation algorithm to solve the STAR‑RIS passive beamforming parameters, let m = m + 1 and Φ l m =Φ l ; Execute six; six: if m ≥ N SCA , let Φ l n+1 =Φ l m , m=0,n=n+1, if m<N SCA , proceed to step 5; step 7: if n≥N BCD , then the algorithm ends and outputs ω l n and Φ l n , otherwise proceed to step 3. The present invention is applied to the field of wireless network security.

Description

Technical Field

[0001] The present invention belongs to the field of wireless network security, and specifically relates to a physical layer security optimization method for a NOMA network. The present invention utilizes wireless access technology, optimization, and computer science and technology to implement a physical layer security optimization method for a NOMA network based on STAR-RIS under the condition of imperfect eavesdropper channel state information. Background Art

[0002] With the booming development of metasurfaces in the field of materials, researchers have proposed the new concept of reconfigurable intelligence surface (RIS), which can configure wireless channels by intelligently reflecting signals. However, for traditional RIS, the base station (BS) and user equipment (UE) must be located on the same side of the traditional RIS, which limits its application in more complex scenarios. Recently, to overcome the above shortcomings of traditional RIS, the simultaneous transmission and reflection RIS (STAR-RIS) has emerged. This reconfigurable smart surface can achieve 360° wireless signal coverage under three working protocols: energy splitting (ES), mode switching (MS), and time switching (TS).

[0003] Non-orthogonal multiple access (NOMA) is expected to meet the spectrum efficiency (SE) requirements of sixth-generation (6G) wireless networks. Many researchers have studied several aspects of STAR-RIS-assisted NOMA networks, such as maximizing network coverage and minimizing network power consumption. In particular, STAR-RIS-based physical layer security (PLS) technologies have been widely explored. Some researchers have studied the problem of maximizing the secure communication rate in multiple-input single-output (MISO) networks under various STAR-RIS protocols. The model considers the base station beamforming vector and STAR-RIS coefficients. Others have further enhanced the security of NOMA networks by utilizing artificial noise and STAR-RIS. However, nearly all of these approaches assume that the eavesdropper's channel state information (CSI) is fully known, which is unrealistic for practical wireless communication systems. In summary, enhancing the physical layer security of downlink MISO NOMA networks based on STAR-RIS under the condition of imperfect CSI of eavesdroppers remains an open problem. Summary of the Invention

[0004] The purpose of the present invention is to enhance the security of downlink MISO NOMA network under the condition of imperfect CSI of eavesdropper, and to propose a NOMA network physical layer security optimization method based on STAR-RIS.

[0005] The specific process of the NOMA network physical layer security optimization method based on STAR-RIS is as follows:

[0006] Step 1: Initialize parameters, including base station-STAR-RIS channel h AS , STAR-RIS-Bob channel STAR-RIS-Eve channel estimate STAR-RIS-Eve channel true value STAR-RIS-Eve channel estimation error The STAR-RIS-Eve channel estimation error range is denoted as ∈ l , non-perfect serial interference cancellation coefficient η, noise variance σ 2 , the number of iterations of the alternating optimization algorithm N BCD , the maximum number of iterations N of the continuous convex approximation algorithm SCA , base station active beamforming parameters STAR-RIS passive beamforming parameters The iteration number index of the alternating optimization algorithm is n=0, and the iteration number index of the continuous convex approximation algorithm is m=0;

[0007] Step 2: Let the legitimate user and illegal eavesdropper be called Bob and Eve respectively, and mathematically model the physical layer security problem of the downlink MISO NOMA network based on STAR-RIS under the condition of imperfect eavesdropper CSI;

[0008] The CSI is channel state information;

[0009] STAR-RIS is a simultaneously transmitting and reflecting reconfigurable smart surface;

[0010] MISO stands for multiple input and single output;

[0011] NOMA stands for non-orthogonal multiple access;

[0012] Step 3: Use the continuous convex approximation algorithm to fix Solve the base station active beamforming parameter ω under the condition of l , let m=m+1 and And execute step 4;

[0013] in, is the STAR-RIS passive beamforming parameter in the nth iteration of the alternating optimization algorithm, is the base station active beamforming parameter in the mth iteration of the continuous convex approximation algorithm;

[0014] Step 4: If the continuous convex approximation algorithm iteration index satisfies m ≥ N SCA ,make m=0, and proceed to step 5. If m<N SCA , proceed to step three;

[0015] in, is the base station active beamforming vector in the nth iteration of the alternating optimization algorithm;

[0016] Step 5: Use the continuous convex approximation algorithm to fix Solve the STAR-RIS passive beamforming parameter Φ under the condition of l , let m=m+1 and Execute step 6;

[0017] in, is the STAR-RIS passive beamforming parameter in the mth iteration of the continuous convex approximation algorithm;

[0018] Step 6: If the continuous convex approximation algorithm iteration index satisfies m ≥ N SCA ,make m=0,n=n+1 and proceed to step 7, if m<N SCA , proceed to step five;

[0019] in, is the STAR-RIS passive beamforming parameter in the n+1th consecutive convex approximation algorithm iteration;

[0020] Step 7: If the alternating optimization algorithm iteration index satisfies n≥N BCD , then the algorithm ends and outputs as well as Otherwise, proceed to step three.

[0021] The beneficial effects of the present invention are:

[0022] The present invention proposes a physical layer security optimization method for NOMA networks based on STAR-RIS under the condition of imperfect eavesdropper CSI. First, the present invention proposes a confidential communication rate modeling method under the condition of imperfect eavesdropper CSI. Secondly, the present invention uses an alternating optimization algorithm to decouple the optimization variables and simplify them into a simpler form. Finally, the present invention uses a continuous convex approximation algorithm to solve the base station active beamforming optimization problem and the non-convex characteristics in STAR-RIS parameter optimization, and realizes physical layer security optimization in the downlink MISO NOMA network. The present invention can enhance the security of the downlink MISO NOMA network and has excellent performance under different numbers of antennas and various channel estimation error parameters.

[0023] The significance of the present invention is to enhance the security of the downlink MISO NOMA network and reduce the hardware overhead under the same security index conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic diagram of the system model, l∈{r,t}, is the true value of the STAR-RIS-Eve reflection channel, is the true value of the STAR-RIS-Eve transmission channel, is the STAR-RIS-Bob reflection channel value, is the STAR-RIS-Bob transmission channel value;

[0025] Figure 2 It is a schematic diagram of algorithm convergence;

[0026] Figure 3 It is a diagram showing the relationship between the minimum confidential communication rate and the eavesdropper’s channel estimation error, δ 2 is the square of the normalized channel estimation error;

[0027] Figure 4 This is a schematic diagram of the relationship between the minimum secure communication rate and the number of antennas. DETAILED DESCRIPTION

[0028] Specific implementation method 1: The specific process of the NOMA network physical layer security optimization method based on STAR-RIS is as follows:

[0029] Step 1: Initialize parameters, mainly including base station-STAR-RIS channel h AS , STAR-RIS-Bob channel STAR-RIS-Eve channel estimate STAR-RIS-Eve channel true value STAR-RIS-Eve channel estimation error The STAR-RIS-Eve channel estimation error range is denoted as ∈ l , the non-perfect serial interference cancellation coefficient η, that is, when using serial interference cancellation to eliminate interference I l When elimination is performed, there exists η×I l The residual (0<η<1), the noise variance σ 2 , the number of iterations of the alternating optimization algorithm N BCD , the maximum number of iterations N of the continuous convex approximation algorithm SCA , base station active beamforming parameters STAR-RIS passive beamforming parameters The iteration number index of the alternating optimization algorithm is n=0, and the iteration number index of the continuous convex approximation algorithm is m=0;

[0030] Step 2: If Figure 1 As shown in the figure, the legitimate user and illegal eavesdropper are called Bob and Eve respectively, and the physical layer security problem of the downlink MISO NOMA network based on STAR-RIS is mathematically modeled under the condition of imperfect eavesdropper CSI (see formula (1), imperfect means that the illegal eavesdropper's CSI has estimation deviation);

[0031] The CSI is channel state information;

[0032] STAR-RIS stands for Simultaneous Transmission and Reflection RIS (STAR-RIS);

[0033] MISO stands for Multiple-input Single-output (MISO);

[0034] NOMA stands for Non-orthogonal Multiple Access (NOMA);

[0035] Step 3: Use the continuous convex approximation algorithm to fix Solve the base station active beamforming parameter ω under the condition of l , let m=m+1 and And execute step 4;

[0036] in, is the STAR-RIS passive beamforming parameter in the nth iteration of the alternating optimization algorithm, is the base station active beamforming parameter in the mth iteration of the continuous convex approximation algorithm;

[0037] Step 4: If the SCA iteration index satisfies m≥N SCA ,make m=0, and proceed to step 5. If m<N SCA , proceed to step three;

[0038] in, is the base station active beamforming vector in the nth iteration of the alternating optimization algorithm;

[0039] Step 5: Use the continuous convex approximation algorithm to fix Solve the STAR-RIS passive beamforming parameter Φ under the condition of l , let m=m+1 and Execute step 6;

[0040] in, is the STAR-RIS passive beamforming parameter in the mth iteration of the continuous convex approximation algorithm;

[0041] Step 6: If the continuous convex approximation algorithm iteration index satisfies m ≥ N SCA ,make m=0,n=n+1 and proceed to step 7, if m<N SCA , proceed to step five;

[0042] in, is the STAR-RIS passive beamforming parameter in the n+1th consecutive convex approximation algorithm iteration;

[0043] Step 7: If the alternating optimization algorithm iteration index satisfies n≥N BCD , then the algorithm ends and outputs as well as Otherwise, proceed to step three.

[0044] Specific implementation method 2: This implementation method is different from the specific implementation method 1 in that the step 2 is as follows Figure 1 As shown in the figure, the legitimate user and illegal eavesdropper are called Bob and Eve respectively, and a mathematical modeling method for the physical layer security problem of the downlink MISO NOMA network based on STAR-RIS under the condition of imperfect eavesdropper CSI is proposed; the specific process is:

[0045] The channel between Eve and STAR-RIS can be expressed as

[0046]

[0047]

[0048] Among them, Ω l is the NOMA decoding order; To define symbols; is an M-dimensional complex vector space; || || is the two-norm;

[0049] Then the signal-to-interference-and-noise ratios (SINRs) of Bob and Eve are

[0050]

[0051]

[0052] in, is the signal-to-interference-and-noise ratio (SINR) of the legitimate user Bob; ω l′ is the base station active beamforming parameter corresponding to the l′th user; ω i is the base station active beamforming parameter corresponding to the i-th user; Ω(i) is the decoding order of the i-th user; Ω(l′) is the decoding order of the l′th user; To eliminate the residual serial interference, ω j is the base station active beamforming parameter corresponding to the j-th user, Ω(j) is the decoding order of the j-th user; () H is the conjugate transpose; l is the user index; l′ is the user index; Φ l The STAR-RIS passive beamforming parameters;

[0053] is the signal-to-interference-and-noise ratio (SINR) of the illegal eavesdropper Eve; ω l is the base station active beamforming parameter corresponding to the lth user;

[0054] The achievable communication rates of Bob and Eve are

[0055] Since Eve’s channel has channel estimation error, the minimum secure communication rate is

[0056]

[0057] Optimize the base station active beamforming parameters and STAR-RIS passive beamforming parameters to maximize the minimum secure communication rate, forming the following optimization problem model

[0058]

[0059] Among them, P max is the maximum transmission power of the base station, C B is Bob’s minimum achievable rate, C E is the maximum achievable rate of Eve, is the STAR-RIS phase, is the STAR-RIS amplitude parameter; The achievable rate when decoding the kth user data for the lth user, The achievable rate when decoding the kth user data for the kth user, is the achievable rate when the lth user decodes the lth user data, Ω(k) is the decoding order of the kth user, and Ω(l) is the decoding order of the lth user;

[0060] Introducing the slack variable τ l→l , and use the S-procedure theorem to transform the problem into the following form:

[0061]

[0062] Among them, P l→l is the variable corresponding to S-procedure, Q l→l as well as is the parameter corresponding to the positive definite constraint, I M is the M-dimensional unit matrix; [] + Defined as [x] + =(|x|+x) / 2, x is a variable; ± is positive semidefinite.

[0063] Other steps and parameters are the same as those in the first embodiment.

[0064] Specific implementation method three: This implementation method is different from specific implementation methods one or two in that the continuous convex approximation algorithm is used in step three to fix the Solve the base station active beamforming parameter ω under the condition of l ; The specific process is:

[0065] When only considering the base station active beamforming, the problem can be simplified to

[0066]

[0067] Let W l =ω l (ω l ) H , W l Satisfy rank(W l )=1 and

[0068] Among them, about W l The rank 1 constraint uses the nuclear norm ||W l || * and the spectral norm ||W l ||2 is replaced, and the non-convex constraints (b)-(c) in the problem are approximated using continuous convex approximation and first-order Taylor expansion, introducing slack variables The final problem can be simplified as:

[0069]

[0070] in,

[0071]

[0072]

[0073]

[0074]

[0075] Δg2(i,l′)=g2(W i ,l′)-g2(W i n ,l′) (14)

[0076] Δg3(j,l′)=g3(W j ,l′)-g3(W j n ,l′) (15)

[0077]

[0078] (||W l ||2) LB =||W l n ||2+Tr(μ max (W l n )μ max (W l n )H (W l -W l n )) (17)

[0079] in, Decode the slack variable of the achievable rate of the l′th user for the lth user; Decode the slack variable of the kth user’s achievable rate for the lth user; W l′ is the auxiliary variable corresponding to the l′th active beamforming vector of the base station; is the optimization result of the auxiliary variable corresponding to the l′th active beamforming vector of the base station in the nth continuous convex approximation algorithm iteration; W i is the auxiliary variable corresponding to the i-th active beamforming vector of the base station; W i n is the optimization result of the auxiliary variable corresponding to the i-th active beamforming vector of the base station in the n-th continuous convex approximation algorithm iteration; W j is the auxiliary variable corresponding to the j-th active beamforming vector of the base station; is the optimization result of the auxiliary variable corresponding to the j-th active beamforming vector of the base station in the n-th continuous convex approximation algorithm iteration, is the optimization result of the auxiliary variable corresponding to the kth active beamforming vector of the base station in the nth continuous convex approximation algorithm iteration; () UB 、() LB , g1(), g2(), g3(), Δg1(), Δg2(), Δg3() are intermediate variables; μ max (·) is the eigenvector corresponding to the largest eigenvalue of the matrix, ||·||2 is the spectral norm of the matrix, ||·|| * is the matrix nuclear norm, ρ≥0 is the penalty factor weight, W l n With (τ l→l ) n is the optimization result of active beamforming and slack variables in the continuous convex approximation algorithm in the nth iteration, is the STAR-RIS parameter of the alternating optimization algorithm in the mth iteration; Tr is the trace of the matrix; rank is the rank of the matrix;

[0080] Since the problem is convex, the convex optimization toolbox can be used to solve W l (Enter Formula 9-17 into the Convex Optimization Toolbox CVX to solve W l );

[0081] And ω l You can use W l The eigenvalue decomposition of the matrix is ​​recovered.

[0082] Other steps and parameters are the same as those in the first or second embodiment.

[0083] Specific embodiment 4: This embodiment differs from any one of the specific embodiments 1 to 3 in that the continuous convex approximation algorithm is used in step 5 to fix the Solve the STAR-RIS passive beamforming parameter Φ under the condition of l ; The specific process is:

[0084] make And V l =v l (v l ) H , where M is the number of STAR-RIS units, then V l Satisfy rank(V l )=1 and When only considering the STAR-RIS passive beamforming, the problem can be simplified to

[0085]

[0086] Among them, v l is Φ l diagonal elements of ; is the energy coefficient of the first STAR-RIS unit; is the energy coefficient of the second STAR-RIS unit; is the energy coefficient of the Mth STAR-RIS unit; is the phase shift parameter of the first STAR-RIS unit; is the phase shift parameter of the second STAR-RIS unit; is the phase shift parameter of the Mth STAR-RIS unit; V l for Optimization auxiliary variables; (τ l→l ) m 、(P l→l ) m and ω l are the slack variables, positive constraint parameters and base station active beamforming optimization results of the alternating optimization algorithm in m iterations respectively; j is the imaginary unit, j 2 =-1;(P l→l ) n is the positive definite constraint parameter of the alternating optimization algorithm in m iterations;

[0087] About V l The rank 1 constraint uses the nuclear norm ||V l || * And the spectral norm ||Vl ||2 is replaced, and the non-convex constraints (e)-(f) in the problem are approximated using continuous convex approximation and first-order Taylor expansion. The final problem can be simplified to:

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095] in,

[0096]

[0097]

[0098]

[0099] Δh1(l′)=h1(V l ,l′)-h1(V l n ,l′) (23)

[0100] Δh2(l′)=h2(V l ,l′)-h2(V l n ,l′) (24)

[0101] Δh3(l′)=h3(V l ,l′)-h3(V l n ,l′) (25)

[0102] (||V l ||2) LB =||V l n ||2+Tr(μ max (V l n )μ max (V l n ) H (V l -V ln )) (26) Among them, h1(), h2(), h3(), Δh1(), Δh2(), Δh3() are intermediate variables; ρ≥0 is the penalty factor weight, V l n is the optimization result of active beamforming in the nth iteration of the continuous convex approximation algorithm, V i n is the optimization result of active beamforming in the nth iteration of the continuous convex approximation algorithm, is the optimization result of active beamforming in the nth iteration of the continuous convex approximation algorithm, W l m is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; W i m is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm;

[0103] Since the problem is convex, the convex optimization toolbox can be used to solve V l n (Enter formula 19-26 into the convex optimization toolbox CVX to solve V l n );

[0104] And Φ l n You can use V l n The eigenvalue decomposition of the matrix is ​​recovered.

[0105] The other steps and parameters are the same as those in the first to third embodiments.

[0106] Table 1 Symbol Description

[0107]

[0108]

[0109]

[0110] The following examples are used to verify the beneficial effects of the present invention:

[0111] Example 1:

[0112] The coordinates of the base station and STAR-RIS are (0, -100, 10) and (0, 0, 10). The legitimate users in the reflection and transmission regions are located at (-5, -5, 2) and (5, 5, 2), while the illegitimate users are located at (0, -50, 2) and (0, 50, 2). The Ricean coefficient of the channel is set to β = 5dB, and the exponential decay coefficient is set to α = 2.2. This example considers the normalized error of the illegitimate user channel estimation, that is, the normalized channel estimation error δ satisfies STAR-RIS has 12 array elements, a maximum base station transmit power of 30dBm, four base station antennas, and a residual coefficient for serial interference cancellation of 0.01. The comparison algorithms are traditional RIS, fixed energy allocation STAR-RIS, and orthogonal frequency division multiple access. All of these comparison schemes can be optimized within the framework of the proposed algorithm.

[0113] like Figure 2 As shown in Figure 2, all the schemes have achieved convergence under the proposed algorithm framework, which proves the compatibility of the algorithm framework proposed in this invention with traditional RIS and OFDMA. Figure 3 As shown in the figure, compared with other schemes, the algorithm proposed in this paper achieves a higher minimum secure communication rate under different normalized channel estimation errors. This shows that for security-required NOMA systems, the independent adjustment of the reflective and transmissive array elements and energy allocation of STAR-RIS are more effective than traditional RIS. However, the limited spectral efficiency of orthogonal frequency division multiple access makes it insufficient to ensure the security requirements of NOMA networks. In addition, the dynamic energy allocation in the algorithm proposed in this paper outperforms fixed energy allocation.

[0114] like Figure 4 As shown in the figure, because multi-antenna base stations result in stronger active beamforming gains for all schemes, the minimum secure communication rate increases with the number of base station antennas. Compared with other schemes, the algorithm proposed in this paper achieves better minimum secure communication performance with different antenna numbers. In terms of engineering practice, for a given secure communication rate, the hardware cost required by this method (such as the number of base station antennas and STAR-RIS array elements) is lower than that of traditional RIS, which greatly improves deployment flexibility and reduces hardware overhead.

[0115] The present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A NOMA network physical layer security optimization method based on STAR-RIS, characterized by: The specific process of the method is: Step 1: Initialize parameters, including base station-STAR-RIS channel h AS , STAR-RIS-Bob channel STAR-RIS-Eve channel estimate STAR-RIS-Eve channel true value STAR-RIS-Eve channel estimation error The STAR-RIS-Eve channel estimation error range is denoted as ∈ l , non-perfect serial interference cancellation coefficient η, noise variance σ 2 , the number of iterations of the alternating optimization algorithm N BCD , the maximum number of iterations N of the continuous convex approximation algorithm SCA , base station active beamforming parameters STAR-RIS passive beamforming parameters The iteration number index of the alternating optimization algorithm is n=0, and the iteration number index of the continuous convex approximation algorithm is m=0; Step 2: Let the legitimate user and illegal eavesdropper be called Bob and Eve respectively, and mathematically model the physical layer security problem of the downlink MISO NOMA network based on STAR-RIS under the condition of imperfect eavesdropper CSI; The CSI is channel state information; STAR-RIS is a simultaneously transmitting and reflecting reconfigurable smart surface; MISO stands for multiple input and single output; NOMA stands for non-orthogonal multiple access; Step 3: Use the continuous convex approximation algorithm to fix Solve the base station active beamforming parameter ω under the condition of l , let m=m+1 and And execute step 4; in, is the STAR-RIS passive beamforming parameter in the nth iteration of the alternating optimization algorithm, is the base station active beamforming parameter in the mth iteration of the continuous convex approximation algorithm; Step 4: If the continuous convex approximation algorithm iteration index satisfies m ≥ N SCA ,make m=0, and proceed to step 5. If m<N SCA , proceed to step three; in, is the base station active beamforming vector in the nth iteration of the alternating optimization algorithm; Step 5: Use the continuous convex approximation algorithm to fix Solve the STAR-RIS passive beamforming parameter Φ under the condition of l , let m=m+1 and Execute step 6; in, is the STAR-RIS passive beamforming parameter in the mth iteration of the continuous convex approximation algorithm; Step 6: If the continuous convex approximation algorithm iteration index satisfies m ≥ N SCA ,make m=0,n=n+1 and proceed to step 7, if m<N SCA , proceed to step five; in, is the STAR-RIS passive beamforming parameter in the n+1th consecutive convex approximation algorithm iteration; Step 7: If the alternating optimization algorithm iteration index satisfies n≥N BCD , then the algorithm ends and outputs as well as Otherwise proceed to step 3; In step 2, the legitimate user and the illegal eavesdropper are respectively referred to as Bob and Eve. A mathematical modeling method for the physical layer security problem of the downlink MISO NOMA network based on STAR-RIS under the condition of imperfect eavesdropper CSI is proposed. The specific process is as follows: The channel between Eve and STAR-RIS is expressed as Among them, Ω l is the NOMA decoding order; To define symbols; is an M-dimensional complex vector space; || || is the two-norm; Then the signal-to-interference-and-noise ratios (SINRs) of Bob and Eve are in, is the SINR of the legitimate user Bob; ω l′ is the base station active beamforming parameter corresponding to the l′th user; ω i is the base station active beamforming parameter corresponding to the i-th user; Ω(i) is the decoding order of the i-th user; Ω(l′) is the decoding order of the l′th user; To eliminate the residual serial interference, ω j is the base station active beamforming parameter corresponding to the j-th user, Ω(j) is the decoding order of the j-th user; () H is the conjugate transpose; l is the user index; l′ is the user index; Φ l The STAR-RIS passive beamforming parameters; is the signal-to-interference-and-noise ratio (SINR) of the illegal eavesdropper Eve; ω l is the base station active beamforming parameter corresponding to the lth user; The achievable communication rates of Bob and Eve are Since Eve’s channel has channel estimation error, the minimum secure communication rate is Optimize the base station active beamforming parameters and STAR-RIS passive beamforming parameters to maximize the minimum secure communication rate, forming the following optimization problem model Among them, P max is the maximum transmission power of the base station, C B is Bob’s minimum achievable rate, C E is the maximum achievable rate of Eve, is the STAR-RIS phase, is the STAR-RIS amplitude parameter; The achievable rate when decoding the kth user data for the lth user, The achievable rate when decoding the kth user data for the kth user, is the achievable rate when the lth user decodes the lth user data, Ω(k) is the decoding order of the kth user, and Ω(l) is the decoding order of the lth user; Introducing the slack variable τ l→l , and use the S-procedure theorem to transform the problem into the following form: Among them, P l→l is the variable corresponding to S-procedure, Q l→l as well as is the parameter corresponding to the positive definite constraint, I M is the M-dimensional unit matrix; [] + Defined as [x] + =(|x|+x) / 2, x is a variable; ≥ is positive semidefinite.

2. The NOMA network physical layer security optimization method based on STAR-RIS according to claim 1 is characterized in that: In step 3, a continuous convex approximation algorithm is used to fix Solve the base station active beamforming parameter ω under the condition of l ; The specific process is: When only considering the base station active beamforming, the problem is simplified to Let W l =ω l (ω l ) H , W l Satisfy rank(W l )=1 and W l ≥0; Among them, about W l The rank 1 constraint uses the nuclear norm ||W l || * and the spectral norm ||W l ||2 is replaced, and the non-convex constraints (b)-(c) in the problem are approximated using continuous convex approximation and first-order Taylor expansion, introducing slack variables The final problem is simplified to: in, in, Decode the slack variable of the achievable rate of the l′th user for the lth user; Decode the slack variable of the kth user’s achievable rate for the lth user; W l′ is the auxiliary variable corresponding to the l′th active beamforming vector of the base station; is the optimization result of the auxiliary variable corresponding to the l′th active beamforming vector of the base station in the nth continuous convex approximation algorithm iteration; W i is the auxiliary variable corresponding to the i-th active beamforming vector of the base station; W i n is the optimization result of the auxiliary variable corresponding to the i-th active beamforming vector of the base station in the n-th continuous convex approximation algorithm iteration; W j is the auxiliary variable corresponding to the j-th active beamforming vector of the base station; is the optimization result of the auxiliary variable corresponding to the j-th active beamforming vector of the base station in the n-th continuous convex approximation algorithm iteration, is the optimization result of the auxiliary variable corresponding to the kth active beamforming vector of the base station in the nth continuous convex approximation algorithm iteration; () UB 、() LB , g1(), g2(), g3(), Δg1(), Δg2(), Δg3() are intermediate variables; μ max (·) is the eigenvector corresponding to the largest eigenvalue of the matrix, ||·||2 is the spectral norm of the matrix, ||·|| * is the matrix nuclear norm, ρ≥0 is the penalty factor weight, W l n With (τ l→l ) n is the optimization result of active beamforming and slack variables in the continuous convex approximation algorithm in the nth iteration, is the STAR-RIS parameter of the alternating optimization algorithm in the mth iteration; Tr is the trace of the matrix; rank is the rank of the matrix; Since the problem is convex, the convex optimization toolbox is used to solve W l ; And ω l Then use W l The eigenvalue decomposition of the matrix is ​​recovered.

3. The NOMA network physical layer security optimization method based on STAR-RIS according to claim 2 is characterized in that: In step 5, a continuous convex approximation algorithm is used to fix Solve the STAR-RIS passive beamforming parameter Φ under the condition of l ; The specific process is: make And V l =v l (v l ) H , where M is the number of STAR-RIS units, then V l Satisfy rank(V l )=1 and V l ≥0, when only considering the STAR-RIS passive beamforming, the problem is simplified to Among them, v l is Φ l diagonal elements of ; is the energy coefficient of the first STAR-RIS unit; is the energy coefficient of the second STAR-RIS unit; is the energy coefficient of the Mth STAR-RIS unit; is the phase shift parameter of the first STAR-RIS unit; is the phase shift parameter of the second STAR-RIS unit; is the phase shift parameter of the Mth STAR-RIS unit; V l for Optimization auxiliary variables; (τ l→l ) m 、(P l→l ) m and ω l are the slack variables, positive constraint parameters and base station active beamforming optimization results of the alternating optimization algorithm in m iterations respectively; j is the imaginary unit, j 2 =-1;(P l→l ) n is the positive definite constraint parameter of the alternating optimization algorithm in m iterations; About V l The rank 1 constraint uses the nuclear norm ||V l || * And the spectral norm ||V l ||2 is replaced, and the non-convex constraints (e)-(f) in the problem are approximated using continuous convex approximation and first-order Taylor expansion. The final problem is simplified to: in, Among them, h1(), h2(), h3(), Δh1(), Δh2(), Δh3() are intermediate variables; ρ≥0 is the penalty factor weight, V l n is the optimization result of active beamforming in the nth iteration of the continuous convex approximation algorithm, V i n is the optimization result of active beamforming in the nth iteration of the continuous convex approximation algorithm, is the optimization result of active beamforming in the nth iteration of the continuous convex approximation algorithm, W l m is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; W i m is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; is the base station active beamforming parameter in the mth iteration of the alternating optimization algorithm; Since the problem is convex, the convex optimization toolbox is used to solve it. and Then use V l n The eigenvalue decomposition of the matrix is ​​recovered.

Images