Resource optimization method, system and storage medium for star-ris assisted uav communication system

Abstract

This invention discloses a resource optimization method, system, and storage medium for a STAR-RIS-assisted UAV communication system. The method includes establishing a STAR-RIS-assisted UAV communication system model; constructing an objective optimization problem, aiming to maximize the system's total transmission rate under constraints such as UAV transmit power, spatial multiplexing, user scheduling binary constraints, STAR-RIS energy conservation constraints, and phase shift orthogonality constraints, and optimizing user scheduling, active beamforming, and STAR-RIS transmission and reflection coefficients; employing an alternating optimization framework to decompose the objective optimization problem into user scheduling sub-problems and active beamforming and STAR-RIS configuration sub-problems; using a penalized successive convex approximation method to solve the user scheduling sub-problem; and using a penalized dual decomposition method combined with block coordinate descent to solve the active beamforming and STAR-RIS configuration sub-problems. The method of this invention can significantly improve the overall system transmission rate while suppressing multi-user interference, meeting the high-bandwidth service requirements of surveillance video and other applications.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to a resource optimization method, system, and storage medium for a STAR-RIS-assisted UAV communication system. Background Technology

[0002] The rapid emergence of connected devices and data-intensive applications has placed enormous pressure on traditional cellular networks, often leading to network congestion and poor quality-of-service (QoS). To address these challenges, sixth-generation (6G) networks aim to provide wider coverage, seamless connectivity, and ultra-low latency. Meanwhile, unmanned aerial vehicles (UAVs), due to their flexibility and cost-effectiveness, are seen as an important complement to ground infrastructure, showing potential particularly in urban tasks such as data collection, immersive virtual reality (VR), and aerial surveillance. However, UAV-assisted communication also faces challenges: air-to-ground (A2G) signal transmission is highly susceptible to obstacles such as buildings, which often obstruct line-of-sight (LoS) and severely degrade link quality. To address this issue, reconfigurable intelligent surfaces (RISs) have been proposed. By intelligently controlling the electromagnetic response of their surface cells, RISs can dynamically shape the wireless propagation environment, thereby improving link reliability, spectral efficiency, and energy efficiency. Recent research has explored various RIS-assisted UAV communication frameworks.

[0003] However, traditional RIS technology has fundamental limitations: it can only reflect signals on one side and requires the transmitter and receiver to be located on the same side of the surface, which limits facility deployment flexibility and signal coverage. To overcome this limitation, the Simultaneously Transmitting and Reflecting Smart Surface (STAR-RIS) has emerged. By supporting the simultaneous reflection and transmission of incident signals, STAR-RIS can effectively extend signal coverage to both sides of the surface, achieving full-space signal processing.

[0004] While STAR-RIS enables more flexible signal routing, its performance in multi-user UAV networks is closely tied to user scheduling. Specifically, user scheduling for each time slot determines spatial interference patterns, directly impacting the optimal design of UAV beamforming and STAR-RIS coefficients. As the number and spatial distribution of scheduled users change over time, the system must dynamically allocate limited spatial degrees of freedom while adjusting the STAR-RIS layout. This raises a fundamental trade-off: scheduling more users improves spectral efficiency but increases beam and STAR-RIS alignment complexity, potentially leading to link quality degradation. Conversely, serving fewer users allows for more precise resource allocation but may limit overall throughput. Furthermore, hardware constraints related to directional support make the coupling between user scheduling and STAR-RIS control a critical bottleneck. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a STAR-RIS-assisted UAV communication network framework for airborne surveillance, which, under the constraints of actual conditions, jointly optimizes user scheduling, active beamforming, and STAR-RIS configuration to maximize system performance and rate.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A resource optimization method for a STAR-RIS-assisted UAV communication system includes:

[0008] S1: Establish a STAR-RIS-assisted UAV communication system model, wherein the system model includes a UAV equipped with M antenna groups and a panoramic camera, a group of single-antenna users, and a STAR-RIS with N components. The single-antenna users are divided into indoor users and outdoor users. The UAV performs aerial monitoring tasks along a predetermined flight path and transmits the obtained monitoring video to the single-antenna users. The UAV, in the total time period... The monitoring task was carried out within the period, which was divided into... Each time slot is of equal length. The duration is The location of UAVs and outdoor users remains stationary within each time slot, but may change between different time slots;

[0009] S2: Construct the objective optimization problem P1, where P1 aims to maximize the system's sum and rate, and the optimization variables include user scheduling. Active beamforming and STAR-RIS transmittance and reflection coefficient The constraints include UAV transmit power constraints, spatial multiplexing constraints, user scheduling binary constraints, STAR-RIS energy conservation constraints, and phase shift orthogonality constraints.

[0010] S3: Using an alternating optimization framework, problem P1 is decomposed into user scheduling subproblem P2 and active beamforming and STAR-RIS configuration subproblem P5, and these two subproblems are solved iteratively in an alternating manner.

[0011] S4: With fixed active beamforming and STAR-RIS transmission and reflection coefficients, the user scheduling subproblem P2 is solved using a penalty-based successive convex approximation method to obtain the optimal user scheduling scheme.

[0012] S5: Under the fixed user scheduling scheme, the penalty dual decomposition method is adopted and the block coordinate descent method is combined to solve the active beamforming and STAR-RIS configuration subproblem P5, so as to obtain the optimal active beamforming vector and STAR-RIS transmission coefficient and reflection coefficient.

[0013] S6: Repeat steps S4 and S5 until the system and rate converge, thereby obtaining the globally optimal resource allocation scheme.

[0014] Furthermore, the objective optimization problem P1 in S2 is expressed as:

[0015]

[0016]

[0017]

[0018]

[0019] C4:

[0020]

[0021] C5:

[0022]

[0023] in,

[0024] , It is the channel bandwidth. Indicates user In the time slot achievable rate, , Represents the set of all users. This indicates the total number of users;

[0025] Indicates in time slot user The signal-to-interference-plus-noise ratio at the location; Indicates user In the time slot Scheduling strategy, Indicates user In the time slot If scheduled, then 0; This represents the active beamforming vector of the UAV; For cascaded channels, This indicates that STAR-RIS and users Between time slots The channel, Indicates time slot The STAR-RIS transmission or reflection coefficient matrix, ,when hour, Transmission coefficient matrix ,when hour, Reflection coefficient matrix ; This indicates the channel between the UAV and STAR-RIS; Represents a set of users Excluding users Any user other than Indicates user Noise variance at the location;

[0026] C1 represents the UAV's transmit power constraint, where, Indicates time slot Mid-wave beamforming vector matrix, express The conjugate transpose of . Represents the trace of a matrix. This indicates the maximum transmit power of the UAV;

[0027] C2 represents spatial multiplexing constraints;

[0028] C3 represents the user scheduling binary constraint;

[0029] C4 and C5 represent the energy conservation constraint and phase shift orthogonality constraint for each STAR-RIS unit, respectively.

[0030] in,

[0031] ,

[0032] ,

[0033] ,

[0034] ,

[0035] and Indicates in time slot The first in STAR-RIS The transmission and reflection amplitudes of each unit, and Indicates in time slot The first in STAR-RIS The transmission and reflection phase shifts of each unit .

[0036] Furthermore, given the STAR-RIS transmission and reflection coefficients and the active beamforming vector, the user scheduling subproblem P2 is expressed as:

[0037]

[0038]

[0039] Solving the user scheduling subproblem P2 specifically includes:

[0040] S41: Introducing slack variables and ,in and At a given local point and Applying the first-order Taylor approximation, the linearization rate is obtained. ;

[0041] S42: Binary constraints Equivalent conversion and And add a penalty term to the objective function of problem P2. Question P3 is obtained:

[0042]

[0043]

[0044]

[0045]

[0046] And C2,

[0047] in, It is a punishment factor;

[0048] S43: For non-convex terms At local points Perform a first-order Taylor expansion to linearize the penalty term and obtain the lower bound of the target. Thus, problem P3 is approximated as a convex problem P4:

[0049]

[0050]

[0051] S44: Solve problem P4 using convex optimization tools to obtain the optimal user schedule. .

[0052] Furthermore, given the user deployment scheme, the active beamforming and STAR-RIS configuration subproblem P5 is expressed as:

[0053]

[0054]

[0055] The solution to the active beamforming and STAR-RIS configuration subproblem P5 specifically includes:

[0056] S51: Utilizing the equivalence between rate maximization and weighted mean square error minimization, and introducing a weighted vector. and auxiliary variables and The problem P5 is transformed into a weighted mean square error minimization problem P6:

[0057]

[0058]

[0059]

[0060]

[0061] as well as

[0062] in,

[0063]

[0064] It is an auxiliary equalization variable used to equalize the signal received by the user;

[0065] express The complex conjugate;

[0066] ;

[0067] , ;

[0068] ;

[0069] S52: Employ the penalized dual decomposition method to decompose equality constraints. and Transforming this into a penalty term in the objective function of P6, we obtain the augmented Lagrange problem P7 of problem P6:

[0070]

[0071]

[0072] in, It is a punishment factor. Represents the Lagrange multipliers related to equality constraints. ;

[0073] By analyzing the primal and dual variables in problem P7 The penalty factor is iteratively updated to solve problem P7. The method for solving the original variables in each iteration is the process described in S53, namely:

[0074] S53: The original variables in problem P7 are decomposed into three blocks using the block coordinate descent method. With other blocks fixed, each block is solved alternately. These three blocks are: sectors, sectors and Sector.

[0075] Furthermore, S53 specifically includes:

[0076] S531: Fixed , Including the constant term, problem P7 is simplified to problem P8:

[0077]

[0078]

[0079] in, ,

[0080] S532: Through alternating optimization The amplitude and phase shift are obtained. Closed-form solution;

[0081] S533: Fixed and , directly obtain and The optimal solution;

[0082] S534: Fixed and Transform problem P7 into problem P12:

[0083]

[0084]

[0085] S535: Solve problem P12 using convex optimization tools to obtain... , and The optimal solution;

[0086] S536: Repeat steps S531 to S535 until the objective function in problem P7 converges, thereby obtaining the optimal active beamforming vector and STAR-RIS transmission and reflection coefficients in this iteration.

[0087] Furthermore, S532 specifically includes:

[0088] S5321: Will Decomposed into magnitude vector and phase shift vector , ;

[0089] S5322: Convert problem P8 to problem P9:

[0090]

[0091]

[0092] S5323: Fixed Amplitude and ,definition Problem P9 is broken down into each STAR-RIS unit. Independent subproblems P10:

[0093]

[0094]

[0095] S5324: Based on constraint C11, the equivalent constraint is obtained. Substituting this equivalent constraint into the objective function of problem P10, the optimization problem is simplified to ;

[0096] S5325: The optimal solution is obtained. ;

[0097] S5326: Fixed Phase Shift and ,definition Problem P9 is broken down into each STAR-RIS unit. Independent subproblems P11:

[0098]

[0099]

[0100] S5326: Introducing Polar Coordinate Substitution ,definition and Substituting this into the objective function of problem P11, we obtain... ,in, ;

[0101] S5327: Will Minimization is transformed into minimization within the interval minimize on ,get The optimal value is:

[0102]

[0103] S5328: Obtain the closed-form solution for the amplitude: ;

[0104] S5329: Repeat S5321 to S5328 until the objective function of problem P8 converges, thus obtaining... The final closed-form solution.

[0105] This invention also provides a resource optimization system for a STAR-RIS-assisted UAV communication system, comprising:

[0106] Memory, configured to store computer programs;

[0107] The processor is configured to execute the computer program to implement the resource optimization method for the STAR-RIS-assisted UAV communication system as described above.

[0108] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0109] The beneficial effects of this invention are:

[0110] This invention addresses the highly non-convex, NP-hard problem in UAV communication caused by user scheduling (0 / 1 discrete variables), beamforming, and joint optimization of STAR-RIS phase shift and amplitude. Through a multi-layered alternating iterative framework, the complex original problem is decomposed into user scheduling subproblems and beamforming and STAR-RIS parameter optimization subproblems, which greatly reduces the difficulty of solving the problem and enables the originally difficult joint optimization problem to be solved efficiently through iteration.

[0111] For binary constraints in user scheduling, the penalty-based successive convex approximation (P-SCA) method relaxes non-convex discrete constraints into a solvable convex optimization problem. It automatically approximates 0 / 1 integer solutions during iteration, avoiding non-integer infeasible solutions in traditional relaxation methods and improving the practicality and reliability of scheduling results.

[0112] To address the inherent energy conservation and phase orthogonality constraints of STAR-RIS, a penalty dual decomposition (PDD) combined with block coordinate descent (BCD) is used to optimize the reflection / transmission coefficients, amplitude, and phase shift step by step. A closed-form update formula is derived, which eliminates the need for complex numerical searches, significantly reduces computational complexity, improves convergence speed, and is more suitable for real-time resource allocation of UAVs.

[0113] This invention constructs a mixed integer nonconvex problem by jointly optimizing user scheduling, active beamforming, and STAR-RIS configuration to maximize system and rate. Moreover, to solve this complex problem, this invention proposes an alternating optimization algorithm that combines a penalty-based successive convex approximation (P-SCA) method for user scheduling design with a penalty dual decomposition (PDD) method with closed-loop updates for active beamforming and STAR-RIS control.

[0114] This invention strictly meets the UAV transmit power limits when optimizing user scheduling and beamforming, and simultaneously serves indoor and outdoor users through STAR-RIS, suppressing multi-user interference and enhancing useful signals. It significantly improves the total system transmission rate under limited power, meeting the high bandwidth service requirements such as surveillance video. Moreover, the solution in this invention is based on a time-slotted model design, with resource optimization completed independently in each time slot. This allows it to adapt to aerial communication scenarios with dynamic changes in UAV position and time-varying channels, maintaining stable and efficient communication performance in mobile monitoring tasks.

[0115] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0116] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0117] Figure 1 It is a STAR-RIS-assisted UAV communication network system model;

[0118] Figure 2 This is a schematic flowchart of a resource optimization method for a STAR-RIS-assisted UAV communication system;

[0119] Figure 3 This is a graph showing the convergence analysis results of the method proposed in this invention;

[0120] Figure 4 The evaluation results are the comparison of the method proposed in this invention with four representative benchmark schemes. Detailed Implementation

[0121] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the preferred embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0122] Figure 1 This is a STAR-RIS-assisted UAV communication network system model. For example... Figure 1 As shown, suppose a plane is equipped with A UAV equipped with one antenna array and a panoramic camera performs an aerial monitoring mission along a predetermined flight path. The UAV transmits the acquired monitoring video to a group of single-antenna users, denoted as [username]. , This represents the total number of users. These users are divided into two groups: indoor users and... and outdoor users ,in It is important to note that outdoor users are mobile and their locations change over time, while indoor users remain in a fixed position within the building. Due to obstacles such as buildings and vegetation, the line-of-sight (LOS) link between the UAV and the user is often disrupted. To improve the reliability of A2G communication and extend signal coverage, a STAR-RIS was deployed on the building facade. This STAR-RIS consists of… Composed of several passive units, it assists in establishing outdoor-to-outdoor (O2O) and outdoor-to-indoor (O2I) communication, thereby enhancing connectivity for all users. UAVs operate within a certain time frame. It performs its monitoring tasks within the time frame, which is divided into... There are several equal-length time slots, each with a duration of [duration missing]. ,Right now Transmission strategies and system configurations can be implemented in each time slot. Assuming the locations of the UAV and outdoor users are stationary within each time slot, but may change between different time slots, the activity of the outdoor users can be simulated using a Gaussian-Markov mobility model.

[0123] Outdoor users In the time slot speed and direction It is given by the following formula:

[0124]

[0125] in, and It is the memory weight, with a value ranging from 0 to 1. and Outdoor users Average speed and direction, and Represents the asymptotic standard deviation. Random variable. and Let these represent the variables of velocity and direction, which follow a Gaussian random distribution. Each outdoor user... In the time slot The position is recorded as Its update method is as follows:

[0126]

[0127] STAR-RIS supports the simultaneous transmission and reflection of incident signals through an energy segmentation model. (In time slots) The coefficient matrices for transmission and reflection are respectively expressed as: and , The complex field is represented in the following form: and .in and Indicates time slot No. The transmission and reflection amplitudes of each unit, Indicates time slot No. The transmission and reflection phase shifts of each unit.

[0128] To ensure energy conservation in each STAR-RIS unit, the transmission and reflection amplitudes must meet the following conditions: Following the assumptions of being passive and lossless, and to reflect the physical limitations of the STAR-RIS design, this invention further strengthens the orthogonality constraint of phase shift (i.e., "phase shift"). The channel between the UAV and STAR-RIS is denoted as... STAR-RIS and users Between time slots The channel is denoted as To support multi-user communication, the system employs Space Division Multiple Access (SDMA) technology. However, due to the limited number of UAV antennas, if the number of users served exceeds [a certain threshold], [further limitations may arise]. This would cause severe interference and significantly degrade system performance. Therefore, the system adopts a user scheduling strategy, selecting only a subset of users in each time slot (i.e., from the user set). Service is provided to a selected subset of users. A binary scheduling variable is defined. ,in Indicates user In the time slot Scheduled, otherwise 0. Let For the scheduling vector, Represent the real number field. Spatial multiplexing constraints require... .user In the time slot The received signal is ,in It is the vector of the UAV's active beamforming (for any user). , can also be expressed as ), It is an information symbol. It's noise. Indicates user The noise variance at a given location. Definition ,Right now , Time slot Time user The signal-to-interference-plus-noise ratio (SINR) at a given location is expressed as:

[0129]

[0130] in, This is an effective concatenated channel. (User) In the time slot The achievable rate is calculated as ,in It is the channel bandwidth.

[0131] This invention aims to maximize the system efficiency and rate of STAR-RIS-assisted UAV communication networks while satisfying physical constraints by jointly optimizing active beamforming, user scheduling, and STAR-RIS coefficients.

[0132] Figure 2 This is a schematic flowchart illustrating the resource optimization method for a STAR-RIS-assisted UAV communication system. (For example...) Figure 2 As shown, the method includes the following steps:

[0133] S1: Establish the STAR-RIS-assisted UAV communication system model as described above;

[0134] S2: Construct the objective optimization problem P1, where P1 aims to maximize the system's sum and rate, and the optimization variables include user scheduling. Active beamforming and STAR-RIS transmittance and reflection coefficient The constraints include UAV transmit power constraints, spatial multiplexing constraints, user scheduling binary constraints, STAR-RIS energy conservation constraints, and phase shift orthogonality constraints.

[0135] S3: Using an alternating optimization framework, problem P1 is decomposed into user scheduling subproblem P2 and active beamforming and STAR-RIS configuration subproblem P5, and these two subproblems are solved iteratively in an alternating manner.

[0136] S4: With fixed active beamforming and STAR-RIS transmission and reflection coefficients, the user scheduling subproblem P2 is solved using a penalty-based successive convex approximation method to obtain the optimal user scheduling scheme.

[0137] S5: Under the fixed user scheduling scheme, the penalty dual decomposition method is adopted and the block coordinate descent method is combined to solve the active beamforming and STAR-RIS configuration subproblem P5, so as to obtain the optimal active beamforming vector and STAR-RIS transmission coefficient and reflection coefficient.

[0138] S6: Repeat steps S4 and S5 until the system and rate converge, thereby obtaining the globally optimal resource allocation scheme.

[0139] The objective optimization problem P1 in S2 is expressed as:

[0140] P1:

[0141]

[0142]

[0143]

[0144] C4:

[0145]

[0146] C5:

[0147] Where C1 represents the UAV's transmit power constraint, where, Indicates time slot Mid-wave beamforming vector matrix, express The conjugate transpose of . Represents the trace of a matrix. This indicates the maximum transmit power of the UAV;

[0148] C2 represents spatial multiplexing constraints;

[0149] C3 represents the user scheduling binary constraint;

[0150] C4 and C5 represent the energy conservation constraint and phase shift orthogonality constraint for each STAR-RIS unit, respectively.

[0151] Problem P1 is a mixed-integer nonlinear optimization problem with nonconvex constraints, which is difficult to solve directly. Therefore, in this invention, problem P1 is decomposed into a user scheduling optimization subproblem P2 and an active beamforming and STAR-RIS configuration subproblem P5, and these two subproblems are solved iteratively and alternately.

[0152] A. Subproblem of User Scheduling Optimization

[0153] Given the phase shift, amplitude (i.e., given the STAR-RIS transmission and reflection coefficients), and active beamforming vector, the user scheduling subproblem P2 can be expressed as:

[0154]

[0155]

[0156] Solving the user scheduling subproblem P2 involves the following steps:

[0157] S41: Introducing slack variables and ,in and Therefore, we can obtain At a given local point and Applying the first-order Taylor approximation, the linearization rate is obtained. , Indicates approximation;

[0158] To address the binary constraints (i.e., binary constraints) in equation (6), this invention employs the P-SCA method. The main idea is to add a penalty term to the objective function to limit violations of the binary constraints, while simultaneously using the SCA framework to iteratively solve the resulting non-convex problem. Specifically, the constraints in equation (6) can be represented as the intersection of the following feasible regions:

[0159]

[0160]

[0161] The constraints in equation (6) can be proven to be equivalent to the constraints in equation (9) and equation (10), since any feasible solution that satisfies equation (6) also satisfies equations (9) and (10), and vice versa.

[0162] Therefore, in S42, binary constraints can be implemented. Equivalent conversion and And add a penalty term to the objective function of problem P2. To restrict non-binary solutions, we can obtain problem P3:

[0163]

[0164]

[0165]

[0166]

[0167] And C2,

[0168] in, It is a punishment factor. When When the time comes, problem P3 will produce a binary solution, which is equivalent to (P2).

[0169] S43: For non-convex terms At local points We approximate the target by performing a first-order Taylor expansion and linearizing the penalty term to obtain the lower bound of the target. Thus, problem P3 is approximated as a convex problem P4:

[0170]

[0171]

[0172] S44: Solve problem P4 using a convex optimization tool (such as CVX) to obtain the optimal user schedule. .

[0173] B. Phase shift, amplitude (i.e., STAR-RIS configuration), and active beamforming optimization sub-problems

[0174] Given a user scheduling scheme, the active beamforming and STAR-RIS configuration subproblem P5 can be expressed as:

[0175]

[0176]

[0177] This invention addresses the non-convexity of the rate maximization problem by fully utilizing a known equivalence relationship between rate maximization and weighted mean square error (MSE) minimization. Specifically, the user... In the time slot The MSE at this location is represented as:

[0178]

[0179] in, It is an auxiliary equalization variable used to equalize the signal received by the user. To indicate complex conjugate, express The complex conjugate, .

[0180] Therefore, the solution to the active beamforming and STAR-RIS configuration subproblem P5 specifically includes the following steps:

[0181] S51: Utilizing the equivalence between rate maximization and weighted mean square error minimization, and introducing a weighted vector. and auxiliary variables and The problem P5 is transformed into a weighted mean square error minimization problem P6:

[0182]

[0183]

[0184]

[0185]

[0186] as well as (i.e., equation (4))

[0187] in, ;

[0188] , ;

[0189] In question P6, regarding and The only constraint is equation (14). To address this issue, this invention employs a penalty-based method within the PDD framework.

[0190] Therefore, in S52: the penalized dual decomposition method is adopted, and the equality constraints are constrained by the augmented Lagrangian (AL) formula. and Transforming this into a penalty term in the objective function of P6, we obtain the augmented Lagrange problem P7 of problem P6:

[0191]

[0192]

[0193] in, It is a punishment factor. This represents the Lagrange multiplier associated with equality constraints. When... At that time, the penalty term forces the equality constraint in (14) to hold. Subsequently, by optimizing the original variables (i.e., all the variables to be optimized in problem P7) ), dual variables (i.e. By iteratively updating the penalty factor, the solution can gradually approach the optimal solution of the Karush-Kuhn-Tucker (KKT) condition.

[0194] Specifically, in each iteration, firstly, given the dual variables... and penalty factor In the case of the original variables (i.e., all variables to be optimized), the solution is obtained; then, the dual variables and penalty factors are updated according to the current equality constraints (i.e., equation (14)); by iterating this process continuously, the obtained solution gradually approaches the optimal solution that satisfies the KKT conditions. In each iteration, the solution of the original variables can be obtained by using the Block Coordinate Descent (BCD) method, which can be referred to in step S53.

[0195] In S53: The original variables are decomposed into three blocks using the block coordinate descent method. With other blocks fixed, each block is solved alternately. These three blocks are: Sector (i.e., Sector 1) Sector (i.e., Sector Two) and Section 3 (i.e., Section 3) is described in detail in steps S531 to S536 below.

[0196] The solution for section one may include steps S531 and S532.

[0197] S531: Fixed , and the constant term (i.e.) (Fixed), simplify problem P7 to problem P8 (i.e., solution module one):

[0198]

[0199]

[0200] in, , Although the constraint is non-convex, a high-quality solution can still be obtained by alternately optimizing the magnitude and phase shift.

[0201] Therefore, in S532: optimization can be achieved through alternation. The amplitude and phase shift are obtained. The closed-form solution.

[0202] For any given time slot Expand the expression in question P8 as follows:

[0203]

[0204] Due to constraints (a) holds true. In the objective function, only the term [a] is true. It contains optimization variables, while the rest are constants.

[0205] Therefore, S532 specifically includes the following steps S5321 to S5329.

[0206] S5321: Will Decomposed into magnitude vector and phase shift vector , ;

[0207] S5322: Convert problem P8 to problem P9:

[0208]

[0209]

[0210] S5323: Fixed Amplitude and ,definition Problem P9 is broken down into each STAR-RIS unit. Independent subproblems P10:

[0211]

[0212]

[0213] S5324: According to constraint C11 (meaning...) or (phase difference), to obtain equivalent constraints Substituting this equivalent constraint into the objective function of problem P10, the optimization problem is simplified to ;

[0214] S5325: The optimal solution is obtained. ;

[0215] Next, in S5326: fixed phase shift and ,definition Problem P9 is broken down into each STAR-RIS unit. Independent subproblems P11:

[0216]

[0217] 0

[0218] S5326: Introducing Polar Coordinate Substitution ,definition and (The objective function can be expressed as) Substituting this into the objective function in problem P11, we obtain... ,in, ;

[0219] S5327: Will Minimization is transformed into minimization within the interval minimize on ,get The optimal value is:

[0220]

[0221] S5328: Obtain the closed-form solution with the optimal amplitude: ;

[0222] S5329: Repeat S5321 to S5328 until the objective function of problem P8 converges, thus obtaining... The final closed-form solution. This completes one solution for plate one.

[0223] The solution for section two includes step S533, namely, in S533: fix... and , can be obtained directly according to "SS Christensen, R. Agarwal, E. De Carvalho, andJ. M. Cioffi, "Weighted sum-rate maximization using weighted MMSE for MIMO-BC beamforming design," IEEE Trans. Wireless Commun., vol. 7, no. 12, pp.4792–4799, Dec. 2008" and The optimal solution:

[0224]

[0225]

[0226] The solution for section 3 includes steps S544 and S545.

[0227] S544: Fixed and Transform problem P7 into problem P12:

[0228]

[0229]

[0230] S545: Solve problem P12 using convex optimization tools to obtain... , and The optimal solution;

[0231] S546: Repeat steps S531 to S545 until the objective function in problem P7 converges, thereby obtaining the optimal active beamforming vector and STAR-RIS transmission and reflection coefficients.

[0232] The present invention also provides a resource optimization system for a STAR-RIS-assisted UAV communication system, the system comprising:

[0233] Memory, configured to store computer programs;

[0234] The processor is configured to execute the computer program to implement the resource optimization method for the STAR-RIS-assisted UAV communication system as described above.

[0235] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the resource optimization method for the STAR-RIS-assisted UAV communication system as described above.

[0236] The advantages of the method of the present invention will be illustrated by the following simulation experiments.

[0237] Consider the following scenario: a plane loaded with UAV with root antenna, in a by With the assistance of STAR-RIS, composed of passive units, the service Single-antenna user. STAR-RIS is installed at coordinates. The building facade. This group of users includes... Indoor users and There are 10 outdoor users, all of whom are distributed in an area of ​​1,000 square meters. Within a square urban area. Outdoor users follow a Gauss-Markov movement model, with relevant parameters set to... , , , Unless otherwise specified, the simulation parameters are set as follows: , , .

[0238] Figure 3 This is an analysis of the convergence of the proposed algorithm. Figure 3 Figure (a) shows the convergence behavior of the proposed algorithm under different STAR-RIS configurations. The convergence behavior varies depending on the number of STAR-RIS components. The algorithm converges within approximately 10 iterations. While increasing the number of STAR-RIS components slightly increases the number of iterations, the throughput is significantly improved due to the increased beamforming flexibility. This validates the scalability and efficiency of the proposed method. Figure 3 Figure (b) shows when At that time, the phase offset difference (also known as phase difference) The convergence behavior is such that the phase difference quickly stabilizes at a certain value. or Thus satisfying the orthogonality condition. This result demonstrates that the proposed algorithm can effectively fulfill STAR-RIS's design requirements for simultaneous transmission and reflection.

[0239] Figure 4 Evaluation results of the proposed scheme and four representative benchmark schemes are presented. These four benchmark schemes include: 1) Reflection-only or transmission-only: The STAR-RIS element is statically split, with one half used for reflection and the other half for transmission; 2) Fixed amplitude STAR-RIS: Phase shift is optimized, but the amplitude coefficient remains fixed; 3) Random STAR-RIS configuration: Phase and amplitude coefficients are randomly updated in each time slot; 4) Polling user scheduling: Users are scheduled in a polling manner, and the beamforming vector is optimized.

[0240] Figure 4 Figure (a) shows when Number of STAR-RIS units Impact on system throughput. (This is related to the number of components.) With the increase of , all schemes achieved throughput improvement due to the gain of passive beamforming. However, the proposed scheme consistently achieves higher throughput than all benchmark schemes. It is worth noting that in the proposed scheme, with The increased throughput achieved with the increase in amplitude demonstrates the effectiveness of the joint optimization of beamforming, STAR-RIS configuration, and user scheduling. While the fixed amplitude scheme simplifies implementation, it restricts the degrees of freedom in signal control, while the polling scheme fails to consider channel heterogeneity, leading to inefficient spatial multiplexing.

[0241] Figure 4 (b) explored the situation The figure shows the relationship between system throughput and maximum UAV transmit power. As can be seen from the figure, all schemes benefit from the increase in transmit power. However, the proposed scheme achieves significantly greater throughput across the entire power range. This improvement is attributed to the joint coordination between beamforming and STAR-RIS parameter tuning, which effectively directs power to the target user. In contrast, the reflection-only or transmission-only schemes are fundamentally limited, achieving only suboptimal signal transmission performance due to the inability to fully utilize STAR-RIS. The random STAR-RIS configuration scheme, lacking adaptability, performs poorly, resulting in unstable power allocation and low utilization efficiency. In conclusion, the proposed scheme has better scalability and can more efficiently utilize available space and transmission resources.

[0242] In summary, this invention proposes a STAR-RIS-assisted UAV communication framework for dynamic environments. By deploying STAR-RIS on building facades, the system improves A2G connectivity for both indoor and outdoor users under mobility and occlusion conditions. To maximize system performance and rate, this invention constructs a joint optimization problem involving user scheduling, active beamforming, and STAR-RIS parameter configuration. Given the mixed integer and non-convex characteristics of this problem, it is decomposed into two easily solvable subproblems: the user scheduling problem is solved using the P-SCA method, while the STAR-RIS configuration and beamforming optimization problems are handled using the PDD method with closed-form solutions. These subproblems converge through iterative optimization. Simulation results demonstrate the superiority of the proposed method over benchmark schemes, reveal the trade-off between user scheduling and STAR-RIS coefficient configuration, and highlight the important value of integrated design in reconfigurable UAV networks.

[0243] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A resource optimization method for a STAR-RIS-assisted UAV communication system, characterized in that, include: S1: Establish a STAR-RIS-assisted UAV communication system model, wherein the system model includes a UAV equipped with M antenna groups and a panoramic camera, a group of single-antenna users, and a STAR-RIS with N components. The single-antenna users are divided into indoor users and outdoor users. The UAV performs aerial monitoring tasks along a predetermined flight path and transmits the obtained monitoring video to the single-antenna users. The UAV, in the total time period... The monitoring task was carried out within the period, which was divided into... Each time slot is of equal length. The duration is The location of UAVs and outdoor users remains stationary within each time slot, but may change between different time slots; S2: Construct the objective optimization problem P1, where P1 aims to maximize the system's sum and rate, and the optimization variables include user scheduling. Active beamforming and STAR-RIS transmittance and reflection coefficient The constraints include UAV transmit power constraints, spatial multiplexing constraints, user scheduling binary constraints, STAR-RIS energy conservation constraints, and phase shift orthogonality constraints. S3: Using an alternating optimization framework, problem P1 is decomposed into user scheduling subproblem P2 and active beamforming and STAR-RIS configuration subproblem P5, and these two subproblems are solved iteratively in an alternating manner. S4: With fixed active beamforming and STAR-RIS transmission and reflection coefficients, the user scheduling subproblem P2 is solved using a penalty-based successive convex approximation method to obtain the optimal user scheduling scheme. S5: Under the fixed user scheduling scheme, the penalty dual decomposition method is adopted and the block coordinate descent method is combined to solve the active beamforming and STAR-RIS configuration subproblem P5, so as to obtain the optimal active beamforming vector and STAR-RIS transmission coefficient and reflection coefficient. S6: Repeat steps S4 and S5 until the system and rate converge, thereby obtaining the globally optimal resource allocation scheme.

2. The method according to claim 1, characterized in that, The objective optimization problem P1 in S2 is expressed as: C4： C5： in, , It is the channel bandwidth. Indicates user In the time slot achievable rate, , Represents the set of all users. This indicates the total number of users; Indicates in time slot user The signal-to-interference-plus-noise ratio at the location; Indicates user In the time slot Scheduling strategy, Indicates user In the time slot If scheduled, then 0; This represents the active beamforming vector of the UAV; For cascaded channels, This indicates that STAR-RIS and users Between time slots The channel, Indicates time slot The STAR-RIS transmission or reflection coefficient matrix, ,when hour, Transmission coefficient matrix ,when hour, Reflection coefficient matrix ; This indicates the channel between the UAV and STAR-RIS; Represents a set of users Excluding users Any user other than Indicates user Noise variance at the location; C1 represents the UAV's transmit power constraint, where, Indicates time slot Mid-wave beamforming vector matrix, express The conjugate transpose of . Represents the trace of a matrix. This indicates the maximum transmit power of the UAV; C2 represents spatial multiplexing constraints; C3 represents the user scheduling binary constraint; C4 and C5 represent the energy conservation constraint and phase shift orthogonality constraint for each STAR-RIS unit, respectively. , , ， ， and Indicates in time slot The first in STAR-RIS The transmission and reflection amplitudes of each unit, and Indicates in time slot The first in STAR-RIS The transmission and reflection phase shifts of each unit .

3. The method according to claim 2, characterized in that, Given the STAR-RIS transmission and reflection coefficients and the active beamforming vector, the user scheduling subproblem P2 is expressed as: Solving the user scheduling subproblem P2 specifically includes: S41: Introducing slack variables and ,in and At a given local point and Applying the first-order Taylor approximation, the linearization rate is obtained. ; S42: Binary constraints Equivalent conversion and And add a penalty term to the objective function of problem P2. Question P3 is obtained: And C2, in, It is a punishment factor; S43: For non-convex terms At local points Perform a first-order Taylor expansion to linearize the penalty term and obtain the lower bound of the target. Thus, problem P3 is approximated as a convex problem P4: S44: Solve problem P4 using convex optimization tools to obtain the optimal user schedule. .

4. The method according to claim 2, characterized in that, Given a user deployment scheme, the active beamforming and STAR-RIS configuration subproblem P5 is expressed as: The solution to the active beamforming and STAR-RIS configuration subproblem P5 specifically includes: S51: Utilizing the equivalence between rate maximization and weighted mean square error minimization, and introducing a weighted vector. and auxiliary variables and The problem P5 is transformed into a weighted mean square error minimization problem P6: as well as in, It is an auxiliary equalization variable used to equalize the signal received by the user; express The complex conjugate; ； ， ； ； S52: Employ the penalized dual decomposition method to decompose equality constraints. and Transforming this into a penalty term in the objective function of P6, we obtain the augmented Lagrange problem P7 of problem P6: in, It is a punishment factor. Represents the Lagrange multipliers related to equality constraints. ; By analyzing the primal and dual variables in problem P7 The problem P7 is solved by iteratively updating the penalty factor. The method for solving the original variables in each iteration is the process described in S53, i.e.: S53: The original variables in problem P7 are decomposed into three blocks using the block coordinate descent method. With other blocks fixed, each block is solved alternately. These three blocks are: sectors, sectors and Sector.

5. The method according to claim 4, characterized in that, S53 specifically includes: S531: Fixed , Including the constant term, problem P7 is simplified to problem P8: in, , S532: Through alternating optimization The amplitude and phase shift are obtained. The closed-form solution; S533: Fixed and , directly obtain and The optimal solution; S534: Fixed and Transform problem P7 into problem P12: S535: Solve problem P12 using convex optimization tools to obtain... , and The optimal solution; S536: Repeat steps S531 to S535 until the objective function in problem P7 converges, thereby obtaining the optimal active beamforming vector and STAR-RIS transmission and reflection coefficients in this iteration.

6. The method according to claim 5, characterized in that, S532 specifically includes: S5321: Will Decomposed into magnitude vector and phase shift vector , ; S5322: Convert problem P8 to problem P9: S5323: Fixed Amplitude and ,definition Problem P9 is broken down into each STAR-RIS unit. Independent subproblems P10: S5324: Based on constraint C11, the equivalent constraint is obtained. Substituting this equivalent constraint into the objective function of problem P10, the optimization problem is simplified to: ; S5325: The optimal solution is obtained. ; S5326: Fixed Phase Shift and ,definition Problem P9 is broken down into each STAR-RIS unit. Independent subproblems P11: S5326: Introducing Polar Coordinate Substitution ,definition and Substituting this into the objective function of problem P11, we obtain... ,in, ; S5327: Will Minimization is transformed into minimization within the interval minimize on ,get The optimal value is: S5328: Obtain the closed-form solution for the amplitude: ; S5329: Repeat S5321 to S5328 until the objective function of problem P8 converges, thus obtaining... The final closed-form solution.

7. A resource optimization system for a STAR-RIS-assisted UAV communication system, characterized in that, include: Memory, configured to store computer programs; A processor is configured to execute the computer program to implement a resource optimization method for a STAR-RIS-assisted UAV communication system as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 6.

Images