A Secure Communication Method for Intelligent Metasurface UAVs Based on Reflection / Transmission

Abstract

A secure communication method for unmanned aerial vehicles (UAVs) based on anti-reflection / transmission intelligent metasurfaces belongs to the field of UAV wireless communication network resource allocation optimization technology. First, the optimal hovering position of the UAV is obtained through the principle of minimizing path loss. Second, given the UAV hovering position, in each iteration, three variables—active beamforming vector, STAR-RIS passive beamforming vector, and power allocation coefficient—are alternately optimized, keeping the other variables constant while optimizing one variable at a time. Third, the three sub-optimization problems are transformed into convex problems through successive convex approximation and Taylor expansion, and the optimal solution is obtained. Finally, the process is iterated until convergence to obtain the optimal solution for the entire problem. This invention is a design method for alternately optimizing parameters such as power allocation coefficient, active beamforming vector, and anti-reflection / transmission coefficient matrix during UAV hovering. This method can maximize the achievable speed of the system in the UAV network while meeting the minimum security communication rate requirements of the user.

Description

Technical Field

[0001] This invention belongs to the field of resource allocation optimization technology for unmanned aerial vehicle (UAV) wireless communication networks, and relates to a joint allocation strategy for maximizing achievable data rate with the assistance of a smart metasurface that simultaneously transmits and reflects light. Specifically, it refers to a method of jointly optimizing the active beamforming vector, the reflection / transmission coefficient matrix, and the power allocation coefficient when the UAV is hovering, thereby maximizing the achievable data rate. Background Technology

[0002] In recent years, researchers have proposed a novel type of intelligent reflective surface called the Simultaneously Transmitting And Reflecting Reconfigurable Intelligent Surface (STAR-RIS). STAR-RIS can simultaneously transmit and reflect received incident signals, enabling wireless signals to cover the entire space. Similar to RIS, by adjusting the transmission and reflection coefficients of each element, STAR-RIS can reconfigure transmitted and reflected signals throughout the wireless space. Objectively speaking, this also provides a new degree of freedom for further improving the performance of wireless communication networks.

[0003] Unmanned Aerial Vehicles (UAVs) are widely used in wireless networks due to their low cost, high mobility, and line-of-sight (LoS) transmission capabilities. Compared to ground-based communication or communication based on high-altitude stable platforms, UAVs can deploy and reach their positions much faster, which is a major advantage of UAVs in wireless communication networks. Furthermore, the ease with which UAVs can establish short-range line-of-sight links can improve some communication channel conditions. UAVs can also serve as mobile temporary base stations, providing temporary signal transmission in areas with challenging geographical conditions. Due to the open nature of wireless transmission, the importance of preventing security attacks such as eavesdropping or interference during information transmission is increasingly prominent, and security issues in wireless communication are receiving more and more attention. In fact, although UAVs can easily establish line-of-sight channels with any user on the ground, which benefits communication speeds in most scenarios, it also makes it easier for potential eavesdroppers to intercept UAV broadcasts. Therefore, how to achieve secure information transmission in wireless communication in UAV scenarios has become a key research issue. It is worth mentioning that the STAR-RIS's ability to extend signals to the entire space also brings the risk of information leakage due to all-space eavesdropping.

[0004] Furthermore, Non-Orthogonal Multiple Access (NOMA) has been implemented as a key design principle for fifth-generation mobile communication technologies. The combined use of NOMA and STAR-RIS also helps overcome the randomness of wireless channels, thereby further improving spectrum performance. Using Successive Interference Cancellation (SIC) technology, NOMA can eliminate some co-channel interference among receivers. However, as the number of users increases and the scale of the wireless network expands, the complexity of NOMA systems also increases, further complicating the use of SIC technology. To address this issue, users can be divided into different groups for processing; that is, group-based NOMA technology can be used to reduce communication complexity. However, this approach also introduces inter-group interference.

[0005] Based on the above problems, this invention studies the problem of secure transmission of user data in the STAR-RIS assisted UAV communication scenario. Under the premise of meeting the minimum requirements for secure communication rate of legitimate users, this invention aims to achieve the maximum achievable rate by jointly optimizing the system parameters. Summary of the Invention

[0006] The purpose of this invention is to solve the user data transmission problem in STAR-RIS-assisted UAV communication scenarios. This invention considers wireless communication scenarios where both the reflecting and transmitting sides have legitimate users and eavesdroppers. Addressing the risk of eavesdropping and data leakage on both the reflecting and transmitting sides, and while meeting the minimum security communication rate requirements for users, this invention jointly optimizes parameters such as UAV position, power allocation coefficient, active beamforming vector, and STAR-RIS reflecting / transmitting coefficient matrix to maximize the achievable speed of the system in the UAV network.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A secure communication method for intelligent metasurface UAVs based on anti-reflection / transmission is proposed. First, the optimal hovering position of the UAV is obtained by minimizing path loss. Second, given the UAV hovering position, three variables—active beamforming vector, STAR-RIS passive beamforming vector, and power allocation coefficient—are alternately optimized in each iteration, keeping the others constant while optimizing one variable. Third, the three sub-optimization problems are transformed into convex problems through successive convex approximation and Taylor expansion, and then the optimal solution is obtained using a CVX solver. Since semidefinite relaxation is used in the optimization problem, Gaussian randomization is employed to recover the vector. Finally, the process is iterated until convergence to obtain the optimal solution for the entire problem. The specific steps include:

[0009] The first step is to build a system model:

[0010] 1) such as Figure 1 A secure downlink communication system for unmanned aerial vehicles (UAVs) based on STAR-RIS comprises a multi-antenna UAV, four single-antenna legitimate users, two single-antenna eavesdroppers, and a planar STAR-RIS with M components. The UAV, acting as the base station, has N antennas and is the transmitter, simultaneously sending information to the legitimate users on both sides via the planar STAR-RIS operating in ES mode. The planar STAR-RIS has two users each on its reflector and transmitter sides. The four single-antenna legitimate users are divided into two groups based on their location, using NOMA technology. Due to the open nature of the wireless channel, eavesdroppers attempting to intercept legitimate user information may exist on both sides of the STAR-RIS, potentially leading to information leakage during transmission.

[0011] 2) The information received by the reflecting user comes from two sources: the direct link signal from the UAV and the signal emitted by the UAV and then transmitted / reflected by STAR-RIS. Therefore, the channel formula for the signal received by the reflecting user is as follows, using the more universal Ricean channel, the expression for the Ricean channel is given:

[0012]

[0013]

[0014] in, This refers to the channel transmitted by the UAV that ultimately reaches the user. It includes two parts: the channel transmitted by the UAV and reaching the user via STAR-RIS transmission / reflection, and the direct link channel transmitted by the UAV directly to the user. The direct link channel uses h... l,d Indicates; h l,k This represents the channel of the planar STAR-RIS reaching the user; Φ k Let F represent the transmission / reflection coefficient matrix of the planar STAR-RIS; F represents the channel transmitted from the UAV to the planar STAR-RIS, which uses a Ricean channel, hence F LoS F represents the line-of-sight transmission portion of the channel. NLoS This represents the non-line-of-sight transmission portion, where K1 is the Rice factor; d UI α represents the distance from the drone to STAR-RIS, and α0 represents the channel gain when the reference distance is 1m.

[0015] Among them, F LoSThe vector is related to both the departure angle (AoD) when the signal leaves the drone and the arrival angle (AoA) when arriving at STAR-RIS. The formula is as follows:

[0016]

[0017]

[0018] Where j represents an imaginary number; d1 and d2 represent the antenna spacing between STAR-RIS and the UAV, respectively; and λ represents the wavelength. η represents the angle of arrival AoA at STAR-RIS; θ represents the departure angle AoD from the UAV to STAR-RIS; the UAV has N antennas, and STAR-RIS has a total of M antennas. y ×M z There are n elements, where 1 ≤ n ≤ N, 1 ≤ m y ≤M y ,1≤m z ≤M z .

[0019] definition, η I The azimuth and elevation angles, θ, represent the angle of arrival from the drone to STAR-RIS. I Representing the departure angle from the drone to STAR-RIS, the following formula exists for m. y ∈{1,…,M y}, m z ∈{1,…,M z The following holds true for n ∈ {1,…,N}:

[0020]

[0021] To increase the signal differentiation between reflecting and transmitting users, an active beamforming vector is used at the UAV to help distinguish users between groups and reduce inter-group interference. This variable is set as ω. k , k∈{T,U}, where ω k This represents the active beamforming vector of the UAV's transmitted signal, with T and U used to distinguish users in the transmission and reflection directions, respectively. To address intra-group interference, taking legitimate user A (the reflecting party) as an example, different power coefficient values ​​are assigned to different legitimate users, thus obtaining the interaction between legitimate user A and the eavesdropping eavesdropper E. S The received signal expression is:

[0022]

[0023] Among them, yA This indicates the signal received by legitimate user A; y e,U This indicates the signal received by the eavesdropper on the same side as the legitimate user A; This represents the channel vector from which the signal originates from the UAV and arrives at legitimate user A; Let represent the channel vector of the eavesdropper; for ease of representation, let A, B, C, and D be used to refer to the four legitimate users, and let s be the channel vectors transmitted to the four users respectively. A ,s B ,s C ,s D It means that p A ,p B ,p C ,p D σ represents the power allocation coefficient for legitimate users; 2 This represents the variance of Gaussian white noise in the channel.

[0024] 3) Under the system's default decoding order, the formula for calculating the user's achievable rate is:

[0025] R = log2(1 + γ) (7)

[0026] Where R represents the reachable rate of the user; γ represents the signal-to-interference-plus-noise ratio of the user.

[0027] Taking a legitimate user A as an example, list the decoded values ​​for user A. A Signal-to-interference-plus-noise ratio (SINR)

[0028]

[0029] User B needs to decode user A's information first, and then decode their own information, so user B decodes s A Signal-to-interference-to-noise ratio User B Decodings B Signal-to-interference-plus-noise ratio (SIR) γ B The expression is as follows:

[0030]

[0031]

[0032] In summary, the reachable rates for users A and B are:

[0033]

[0034] Next, consider the eavesdropper's eavesdropping rate C. e,A ,get

[0035]

[0036] Therefore, the secure communication rate for each legitimate user is:

[0037] R S,l =[R l -C e,l ] + (13)

[0038] Among them, R S,l The secure communication rate of a user is represented by the formula R, which is the user's achievable rate. l The eavesdropping rate C of the eavesdropper in that direction e,l The difference, [x] + = max{x, 0}.

[0039] The second step is to determine the objective function and list the optimization problem:

[0040] For ease of description, some explanations of the parameters are provided here. P represents the power allocation coefficient, which restricts the decoding order of user information, where the active beamforming vector ω k It does not affect the power allocation ratio for users. Based on the secure communication rate requirements of each legitimate user, Γ S,l This represents the minimum secure communication quality requirement. Q represents the drone's hovering position, and L restricts the possible hovering position area of ​​the drone. The optimization objective of this invention is to obtain the maximum achievable system speed, so the optimization problem can be expressed as:

[0041]

[0042]

[0043] ||ω k ||≤1 (14c)

[0044] p A ≥p B ,p C ≥p D (14d)

[0045] p A +p B +p C +p D ≤P S (14e)

[0046] R S,l ≥Γ S,l (14f)

[0047] Q∈L (14g)

[0048] Among them, Rsum This represents the total reachability rate of all legitimate users in the system; ω represents the transmission / reflection amplitude coefficient and phase shift coefficient at element m on the STAR-RIS, respectively; M represents the total number of elements in the STAR-RIS, 1≤m≤M; k P represents the active beamforming vector for transmission / reflection; S The power allocation coefficient matrix represents the power allocation coefficients for legitimate users; Γ S,l This indicates the minimum required rate of secure communication for legitimate users.

[0049] The third step is to design an algorithm to solve the optimization problem:

[0050] In the optimization problem (P1) obtained from the modeling, there are many optimization variables, including the UAV hovering position Q and the UAV's active beamforming vector ω. k STAR-RIS's anti-transmission matrix Φ k The power allocation coefficient matrix P for legitimate users. The original optimization problem is non-convex and cannot be solved using convex optimization methods. An alternating optimization algorithm is proposed below, which simplifies the original optimization problem into a sub-optimization problem under specific conditions where other variables are known. The specific process is as follows:

[0051] 1) Optimization of the drone hovering position Q

[0052] To simplify the scenario of this problem, d l The distance from the drone to each authorized user is represented by the formula d. l =||Qq l ||, where q l This represents the user's three-dimensional coordinate vector. The formula for calculating path loss (in dB) is:

[0053]

[0054] Therefore, the optimization problem regarding the drone's hovering position Q is as follows:

[0055]

[0056] stQ∈L (17)

[0057] Next, the constraints of the original problem are expanded using Taylor's formula to obtain a convex optimization problem. This invention limits the hovering position of the UAV to the range of L.

[0058] 2) Optimization of power allocation coefficient P

[0059] First, we organize the expression for the reachability rate of legitimate users, and then perform a Taylor expansion on the reachability rate of each legitimate user to obtain the lower bound of the reachability rate. Let's take user A, the reflecting party, whose channel is relatively complex, as an example to illustrate this process.

[0060]

[0061] Since the logarithmic function log2(x) is nonconvex, it is first decomposed into a subtraction of two terms, and then the expression Ψ is introduced. U =p A +p B ,Ψ T =p C +p D Expanding the expression at two points, we get R. A The lower bound is denoted as The expression is as follows:

[0062]

[0063]

[0064] in, p represents the power allocation coefficient for the transmitting side user at the τth iteration. C +p D Given the value, expand at this point; due to the excessive length of the expression, it was split into parts, one part using R... A2 express.

[0065] The reachable rates for the other three legitimate users can also be obtained using the above method. obtain in this way The following formulas use This indicates the secure communication rate for each legitimate user.

[0066] The optimization problem (P1) is processed to obtain the following convex optimization problem:

[0067]

[0068] stp A ≥p B ,p C ≥p D (21b)

[0069] p A +p B +p C +p D ≤P S (21c)

[0070]

[0071] 3) STAR-RIS inverse / transmission coefficient matrix Φk Optimization

[0072] To solve the above optimization problem (P1), we construct a new auxiliary matrix, again using user A as an example.

[0073]

[0074] in, It is the auxiliary matrix that is constructed.

[0075] This leads to a matrix related to the channel vector.

[0076]

[0077] Next, construct vectors.

[0078]

[0079] Furthermore, the introduced auxiliary matrix should always satisfy V. U ≥0,rank(V U The condition is ) = 1, otherwise The conditions will not be met.

[0080] After the above processing, Rewritten as follows, the above optimization problem (P1) is rewritten as a rank-constrained semi-definite programming (SDP) problem under the condition that the power allocation coefficient and the active beamforming vector are determined, resulting in the following formula.

[0081]

[0082] Similarly, construct the channel vector correlation matrix. get:

[0083]

[0084] Based on the above, the reachable rate expression for user A is obtained as follows:

[0085]

[0086] Since the legitimate user communication rate of the reflector is only related to the STAR-RIS reflection correlation matrix, a first-order Taylor expansion of equation (27) yields R. A The lower bound of the trace function. Since the trace function satisfies the linearity condition, the order in which the trace functions are calculated does not affect the final result. Therefore, the calculation of the trace function of the matrix can be moved to the outermost part of the right side of the equation, resulting in R. A The lower bound expression:

[0087]

[0088]

[0089] in, This represents the lower bound of the reachable rate of a legitimate user A. Due to the excessive length of the formula, L is used. A1 ,L A2 The formula is broken down into its constituent expressions, as shown above.

[0090] Next, taking the legitimate user C of the transmitting side as an example, an auxiliary matrix related to the eavesdropper is constructed. Since the transmitting user does not have a direct link with the UAV, the dimensions and composition of the channel matrix related to the channel vector constructed are not exactly the same as the user channel matrix of the reflecting side, as shown in the following formula:

[0091]

[0092]

[0093] Therefore, we can conclude that:

[0094]

[0095] Introducing auxiliary vectors The lower bound of user C is obtained by performing a Taylor expansion on the reachable rate of user C. Due to the excessive length of the formula, it is broken down and expressed as follows:

[0096]

[0097]

[0098] After simplification, the optimization problem (P1) yields the following convex optimization subproblems:

[0099]

[0100] stV k ≥0(35b)

[0101] rank(V k )=1(35c)

[0102]

[0103]

[0104] The constraints in the convex optimization subproblem (P4) are relaxed using semi-definite relaxation (SDR). Finally, the processed convex optimization problem can be solved using MATLAB's built-in CVX solver.

[0105] 4) Active beamforming vector ω k Optimization

[0106] Active beamforming vector ω k Optimization and STAR-RIS inverse / transmission coefficient matrix Φ k The optimization methods are the same, both using successive convex approximation and semi-positive definite relaxation to optimize the active beamforming vector ω of the UAV. k The sub-optimization problem is rewritten as an SDP problem to complete the solution process. First, the channel vector is constructed.

[0107]

[0108] Introducing auxiliary variable matrix

[0109] W k =ω k ω k H (37)

[0110] Therefore, we can conclude that:

[0111]

[0112] Therefore, legitimate user A can use the Taylor expansion formula in... The lower bound of the reachable rate of legitimate user A is obtained by expanding the two points. Lower bound of secure communication rate Moreover, the matrix W formed by the active beamforming vectors k The following condition must always be met for W to be able to perform such an operation. k =ω k ω k H The condition always holds true:

[0113] W k ≥0,rank(W k )=1 (39)

[0114] After finding the optimal W k After obtaining the value, Gaussian randomization is used to recover the corresponding active beamforming vector, and the solution is iteratively obtained. After processing, the original optimization problem (P1) is rewritten as a subconvex optimization problem, as shown in the following expression:

[0115]

[0116] stTr(W k )≤1,k∈{T,U} (40b)

[0117]

[0118] W k ≥0,rank(W k )=1(40d)

[0119] After SDR, the aforementioned convex optimization problem can be optimized using the CVX solver.

[0120] 5) Alternating optimization algorithm design

[0121] This invention proposes an alternating optimization algorithm for UAV wireless secure communication problems based on STAR-RIS assistance. First, the optimal hovering position of the UAV is obtained. Given the UAV hovering position, the original optimization problem (P1) is decomposed into three sub-optimization problems according to different optimization variables: the power allocation coefficient, the STAR-RIS passive beamforming vector, and the active beamforming vector, respectively. To address the non-convex nature of these three sub-optimization problems, successive convex approximation, semi-definite relaxation, and Gaussian randomization methods are used to obtain sub-convex optimization problems (P3), (P4), and (P5), respectively. The optimal solution is then obtained through a CVX solver. In each iteration, the active beamforming vector, the STAR-RIS passive beamforming vector, and the power allocation coefficient are alternately optimized. This process is repeated until convergence, yielding the optimal solution for the entire problem, i.e., the maximum total reachable rate of the entire system.

[0122] The beneficial effects of this invention are:

[0123] This invention addresses the secure transmission problem in a STAR-RIS-assisted unmanned aerial vehicle (UAV) communication system. It proposes a deployment scheme that maximizes the achievable data rate of the communication system by rationally designing the UAV's position, active beamforming vector, STAR-RIS passive beamforming vector, and power allocation coefficient. This invention provides a novel approach to achieving secure transmission in UAV systems and maximizing the achievable data rate. Attached Figure Description

[0124] Figure 1 A model diagram of the STAR-RIS-assisted UAV downlink transmission secure communication system.

[0125] Figure 2 A schematic diagram showing the optimal hovering position for a drone.

[0126] Figure 3The impact of the number of STAR-RIS components and total transmit power on the achievable speed of the system.

[0127] Figure 4 The impact of STAR-RIS deployment location on system achievable speed.

[0128] Figure 5 The changes in achievable speeds for Schemes 1 through 4 as the number of STAR-RIS components increases.

[0129] Figure 6 To investigate the impact of employing different types of intelligent reflectors on achievable rates.

[0130] Figure 7 This compares the total reachable rate with the reachable rate of a single legitimate user under different total transmit power. Detailed Implementation

[0131] To better understand the above technical solution, a detailed analysis is provided below in conjunction with the accompanying drawings and specific implementation methods. The specific steps of this embodiment include:

[0132] The first step is to build a system model:

[0133] like Figure 1 This embodiment presents a STAR-RIS-assisted UAV communication system model. The UAV's fixed hovering altitude is set to 100m, and the UAV is equipped with four antennas spaced 0.5λ apart. The STAR-RIS antennas are positioned at [20, 0, 20] and consist of 36 components. In this embodiment, the UAV serves four single-antenna users. Due to the use of NOMA technology, the users on both sides are divided into different groups based on reflection / transmission. Each group contains two users and one eavesdropper. The horizontal positions of the users are [25, -5, 0], [35, -10, 0], [55, 10, 0], and [50, 25, 0], respectively, and the eavesdropper's positions are [270, -260, 0] and [250, 270, 0]. The variance of the Gaussian white noise σ is... 2 = -100dBm, total transmit power P S =0.7w, the path loss of all indirect links at a reference distance d0=1m is -20dB.

[0134] The second step involves determining the objective function based on the specific parameter settings and relevant formulas from the first step. The optimization objective of this invention is to maximize the achievable system rate while ensuring the user's minimum secure communication rate. The optimization problem is expressed as follows:

[0135]

[0136]

[0137] ||ω k ||≤1 (14c)

[0138] p A ≥p B ,p C ≥p D (14d)

[0139] p A +p B +p C +p D ≤P S (14e)

[0140] R S,l ≥Γ S,l (14f)

[0141] Q∈L (14g)

[0142] The third step is to design an algorithm to solve the optimization problem:

[0143] In the original optimization problem, there are numerous optimization variables, including the UAV's hovering position Q and the UAV's active beamforming vector ω. k STAR-RIS's anti-transmission matrix Φ k The power allocation coefficient matrix P for legitimate users. The original optimization problem is non-convex and cannot be solved using convex optimization methods. The proposed alternating optimization algorithm is used to simplify the original optimization problem into sub-optimization problems under specific conditions where other variables are known. These sub-problems are processed and optimized in the order described above.

[0144] 1) Optimization of drone hovering position

[0145] As shown in (P2), the optimization problem for the drone's hovering position Q is solved by expanding the constraints of the original problem using Taylor's formula, resulting in a convex optimization problem. We limit the drone's hovering position to the range L, and the optimal hovering position and convergence graph are shown below. Figure 2 As shown.

[0146] 2) Optimization of power allocation coefficient P

[0147] First, we reorganize the expression of the objective function. Then, we use Taylor expansion to calculate the reachable rate for each legal user, which gives us the lower bound of the reachable rate. Following the above processing method, the reachable rate and secure communication rate of each legitimate user are decomposed into two terms that are subtracted from each other. Then, the expression is expanded in two points to obtain the corresponding lower bound. Finally, a convex optimization problem that can be solved by CVX is obtained.

[0148] 3) STAR-RIS inverse / transmission coefficient matrix Φ k Optimization

[0149] Construct a new matrix V k =ν k (ν k ) H Rewrite the above problem as a rank-constrained SDP problem, considering the coefficient matrices V for reflection and transmission respectively. k On the one hand, we can expand the equation; on the other hand, we can obtain a lower bound on the secure communication rate by constructing an auxiliary matrix related to the eavesdropper. Finally, after Gaussian randomization, we can recover Φ. k The optimal solution is found. Therefore, the original problem is transformed into the form (P4). The processed convex optimization problem can be solved by MATLAB's built-in CVX solver.

[0150] 4) Active beamforming vector ω k Optimization

[0151] Using an optimization method similar to the STAR-RIS inverse / transmission coefficient matrix, the sub-optimization problem concerning the active beamforming vector of the UAV is rewritten into an SDP problem through successive convex approximation and semi-definite relaxation. Then, Gaussian randomization is employed to recover the active beamforming vector. The final processed convex optimization problem is transformed into the form (P5). After relaxing the constraint that the rank of the matrix equals 1, the aforementioned convex optimization problem can be optimized using the CVX solver.

[0152] 5) Alternating optimization algorithm design

[0153] Using the above method, the subproblems can be transformed into convex optimization problems, and the results are as follows:

[0154] Figure 3 The achievable rate for legitimate users varies with the number of STAR-RIS components and the total transmit power P. S The changes show that, with other parameters remaining constant, the reachable rate of legitimate users increases with the number of STAR-RIS components and the total transmit power P. S The increase is due to the increase in the number of STAR-RIS components. The reason for the improved performance of the analysis system model is that as the number of STAR-RIS components increases, more components can receive signals from the UAV, resulting in a greater passive beamforming gain of STAR-RIS.

[0155] Figure 4 The data shows how the achievable speed of STAR-RIS varies with the number of STAR-RIS components under three different conditions: near the reflecting user, near the transmitting user, and far from both groups of users. This is related to which user STAR-RIS prioritizes to serve.

[0156] To demonstrate the advantages of this invention, this embodiment presents the invention as Scheme 1, and also provides Schemes 2 to 6 for comparison. This invention alternately optimizes the active beamforming vector ω of the UAV. k STAR-RIS's anti-transmission matrix Φ k Power allocation coefficient matrix P S These three optimization variables, from Scheme 2 to Scheme 4, can be considered part of this invention. All three schemes lack the optimization step for one of the variables in this invention. Specifically, Scheme 2 only alternately optimizes the STAR-RIS inversion / transmission matrix Φ. k With the power allocation coefficient matrix P S Scheme 3 alternately optimizes the active beamforming vector ω k With the power allocation coefficient matrix P S Scheme 4 alternately optimizes the active beamforming vector ω k The anti-transmission matrix Φ of STAR-RIS k .

[0157] Figure 5 By comparing the performance of Schemes 1 to 4 under different STAR-RIS component counts, it can be seen that under the same STAR-RIS component count, the total achievable rate of Schemes 2, 3, and 4 is lower than that of Scheme 1 compared to the present invention (Scheme 1), thus verifying the superiority of the present invention in improving the system achievable rate.

[0158] Figure 6 We will compare the performance advantages of STAR-RIS compared to traditional RIS. Figure 6 The performance of Scheme 1 (using STAR-RIS assisted communication) and Scheme 5 (using traditional RIS assisted communication) were plotted. It can be seen that under the same conditions where the number of UAV antennas and the number of intelligent reflector elements are equal, the achievable speed of the present invention is higher than that of the traditional RIS scheme, thus proving the performance improvement brought about by the use of STAR-RIS in the present invention.

[0159] Figure 7 This section compares the changes in the total reachability rate of the system and the reachability rate of a single legitimate user within the system. It can be seen that both the total reachability rate and the reachability rate of a single legitimate user increase with the number of STAR-RIS components, verifying the reasonable understanding that the total reachability rate of the system mainly depends on strong users with better channel conditions.

[0160] Combination Figures 2 to 7It can be seen that the present invention has superior achievable speed due to the algorithm that alternately optimizes the active beamforming vector, the STAR-RIS anti / transmission matrix, and the power allocation coefficient, along with the STAR-RIS-assisted communication. It can also be seen that the achievable speed of the STAR-RIS-assisted UAV safety communication system studied in this invention is also affected by the deployment location of the STAR-RIS, the number of STAR-RIS components, and the total transmit power P. S The influence of factors such as the number of antennas N of the drone and the communication status of strong users in the system.

[0161] The above-described embodiments are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A secure communication method for intelligent metasurface unmanned aerial vehicles based on reflection / transmission, characterized in that, The anti-transmission / transmission smart metasurface refers to a smart metasurface that simultaneously transmits and reflects light. The specific method for secure communication between a smart metasurface UAV is as follows: The optimization objective is determined, the optimization problem is listed, and the optimal hovering position of the UAV is obtained through the principle of minimizing path loss. Based on the hovering position of the UAV, the active beamforming vector, the STAR-RIS passive beamforming vector, and the power allocation coefficient are alternately optimized in each iteration. When optimizing one variable, the other variables are kept unchanged. The optimization problem is transformed into a convex optimization subproblem through successive convex approximation and Taylor expansion, and the optimal solution is obtained by solving it. The optimization objective is to obtain the maximum achievable system speed, and the optimization problem is expressed as: (P1): (14a) (14b) (14c) (14d) (14e) (14f) (14g) in, This represents the hovering position of the drone; Indicates the active beamforming vector; P Represents the power distribution factor; The matrix representing the transmission / reflection coefficients of a planar STAR-RIS; This represents the total reachability rate of all legitimate users in the system; Indicates STAR-RIS m Transmission / reflection amplitude coefficient at the component, subscript Indicates the user category, with T and U representing users in the transmission and reflection directions, respectively; Indicates STAR-RIS m Phase shift coefficient at the component; This indicates the total number of components in STAR-RIS. ; These represent the power allocation coefficients for legitimate users A, B, C, and D, respectively. A matrix representing the power allocation coefficients for legitimate users; This indicates the minimum required rate of secure communication for legitimate users. Limit the possible hovering area of ​​the drone.

2. The secure communication method for intelligent metasurface UAVs based on reflection / transmission according to claim 1, characterized in that, Includes the following steps: The first step is to build a system model: 1.1) In a STAR-RIS-based UAV downlink secure communication system, there are multiple-antenna UAVs, four single-antenna legitimate users, two single-antenna eavesdroppers, and one containing... A planar STAR-RIS system with individual components; among which, drones have N Each antenna transmits information simultaneously to legitimate users on both sides via a planar STAR-RIS; the planar STAR-RIS has two users on each of the reflecting and transmitting sides, and the four single-antenna legitimate users are divided into two groups according to their location. 1.2) The information received by the reflecting user includes the direct link signal from the UAV and the signal emitted by the UAV and then transmitted / reflected by STAR-RIS, thus obtaining the channel of the signal received by the reflecting user. ; Active beamforming vectors are used at the UAV. Distinguishing between users in different groups, legitimate user A and the eavesdropper. E S The received signal expression is: (6) in, This indicates the signal received by legitimate user A; This indicates the signal received by the eavesdropper on the same side as the legitimate user A; This represents the channel vector from which the signal originates from the UAV and arrives at legitimate user A; Let A represent the channel vector of the eavesdropper; A, B, C, and D represent four legitimate users. This represents the information transmitted by four legitimate users. These represent the power allocation coefficients for legitimate users; This represents the variance of the Gaussian white noise in the channel; 3) Under the system's default decoding order, the secure communication rate for each legitimate user is: (13) in, Indicates the user's secure communication rate; Indicates the achievable rate for the user; Indicates the eavesdropping rate ; ; The second step is to define the optimization goals and list the optimization problems. The third step is to design an algorithm to solve the optimization problem: In the optimization problem obtained from the modeling, the optimization variables include the drone hovering position. Active beamforming vector of UAV STAR-RIS's anti-transmission matrix Power allocation coefficient matrix for legitimate users The alternating optimization algorithm is used to simplify the optimization problem into a convex optimization subproblem under the condition that other variables are known.

3. A secure communication method for intelligent metasurface UAVs based on reflection / transmission according to claim 2, characterized in that, The specific process of the third step is as follows: 3.1) Hovering position of the drone Q The optimization yields a convex optimization subproblem: (P2): (16) (17) in, The distance from the drone to each authorized user is represented by: 3.2) Power allocation coefficient The optimization yields a convex optimization subproblem: (P3): (21a) (21b) (21c) (21d) in, express The lower bound; express The lower bound; express The lower bound; express The lower bound; Indicates the secure communication rate for each legitimate user; 3.3) STAR-RIS Inverse / Transmission Coefficient Matrix The optimization yields a convex optimization subproblem: (P4): (35a) (35b) (35c) (35d) (35e) 3.4) Active Beamforming Vector The optimization yields a convex optimization subproblem: (P5): (40a) (40b) (40c) (40d) in, This represents the matrix formed by the active beamforming vectors; 3.5) Alternating optimization algorithm design; First, the optimal hovering position of the UAV is obtained. Given the hovering position of the UAV, the optimization problem is decomposed into three convex sub-optimization problems according to different optimization variables. The power allocation coefficient, the STAR-RIS passive beamforming vector, and the active beamforming vector are optimized respectively, and the optimal solutions are obtained by the CVX solver. In each iteration, the active beamforming vector, the STAR-RIS passive beamforming vector, and the power allocation coefficient are optimized alternately. Finally, the iteration continues until convergence, and the optimal solution of the entire problem is obtained, which is the maximum total reachable rate of the entire system.

Images