A downlink fairness transmission rate optimization method for a 5G / 6G-oriented UAV-IRS assisted NOMA-MEC system

Abstract

The application discloses a downlink fairness transmission rate optimization method for a 5G / 6G-oriented UAV-IRS assisted NOMA-MEC system, the core of which is to maximize the minimum transmission rate of a mobile device (MD) by jointly optimizing the 3D position of a UAV, an IRS reflection matrix and power allocation of each mobile node while considering a line-of-sight (LoS) and a non-line-of-sight (NLoS) propagation environment.

Description

Technical Field

[0001] This invention relates to the field of network technology, and in particular to a method for optimizing downlink fair transmission rate in a UAV-IRS-assisted NOMA-MEC system for 5G / 6G. Background Technology

[0002] With the rapid advancement of communication technologies and the decline in operating costs, unmanned aerial vehicles (UAVs) have become a key component of the evolution of 5G / 6G networks. UAVs are gaining increasing recognition for their ability to enhance network performance by providing flexible, fast, and scalable wireless connectivity solutions. However, despite their immense potential, deploying UAVs in urban and remote environments still faces several challenges. Obstacles such as buildings and tall structures, as well as the inherent limitations of line-of-sight (LoS) communication links, can lead to significant signal attenuation, ultimately reducing communication quality and performance. Overcoming these challenges is crucial to fully realizing the potential of UAVs in 5G / 6G networks. As a key technology for 5G / 6G networks, intelligent reflective surfaces (IRS), equipped with passive reflective elements and software controllers, can precisely adjust the phase of reflected signals, enabling dynamic management and adaptive control of wireless channels. This significantly improves communication quality and expands network coverage, attracting widespread attention.

[0003] IRS deployment can be divided into two main scenarios: terrestrial IRS (GIRS) and airborne IRS (AIRS). In the GIRS scenario, IRS panels are mounted on various surfaces, such as building exteriors, outdoor billboards, and interior walls. In contrast, the AIRS scenario involves drones equipped with IRS, acting as relay nodes for reflected signals, establishing a Line-of-Sight (LoS) link between the access point (AP) and ground mobile device (GMD). Compared to GIRS, AIRS leverages the mobility of drones to provide a more flexible and robust solution, especially in non-line-of-sight (NLoS) or challenging signal coverage areas. AIRS, capable of omnidirectional reflection, is increasingly becoming a research focus. However, in the AIRS scenario, the 3D aerial position of the drone directly affects the state of the IRS reflection channel. Furthermore, the presence of non-line-of-sight conditions in AIRS significantly impacts the transmission rate. These factors pose significant challenges to optimizing IRS panel angles and improving overall system performance.

[0004] Most existing AIRS studies simplify computation by focusing on ideal line-of-sight channels and ignoring the effects of non-line-of-sight conditions, or primarily concentrate on one-to-one UAV services using Time Division Multiple Access (TDMA), failing to fully leverage Non-Orthogonal Multiple Access (NOMA) to improve system throughput. In ultra-dense wireless networks, these methods struggle to meet the Quality of Service (QoS) requirements for large-scale access and can cause severe interference. Furthermore, many existing studies focus primarily on optimizing global system performance metrics, such as task execution time and energy consumption, often neglecting the Quality of Experience (QoE) of individual mobile terminals. This oversight can lead to poor performance for some terminals and create fairness issues across the network.

[0005] Gao et al. deployed MEC servers as base stations on UAVs and used multiple IRSs to assist ground users in defending against UAV eavesdropping, thereby achieving secure computation offloading. To address the severe challenge of energy constraints affecting UAV sustainability and performance, Wang et al. jointly optimized IRS phase shift, UAV trajectory, flight time, and resource allocation, along with GMD scheduling, to minimize processing time. However, the simplified line-of-sight master channel model did not fully leverage the potential of UAV 3D maneuverability. Recent research has considered the advantages of UAV 3D spatial positioning in enhancing system performance. Asim et al. addressed total cost, including energy consumption, mission completion time, and UAV maintenance costs, and integrated multiple UAVs to serve numerous GMDs with the support of multiple IRSs. However, in multi-user air-to-ground communication scenarios, the fixed nature of IRS panels combined with UAV mobility limits coverage, making it difficult for UAVs to maintain continuous proximity to IRS equipment.

[0006] Chinese patent CN119342607A discloses a resource allocation method for a non-orthogonal multiple access and wireless power-carrying communication system based on intelligent reflectors and UAV-assisted systems. This method maximizes the system's sum rate under constraints of UAV trajectory, base station transmit power allocation, IRS reflection coefficient, minimum transmission rate of IoT devices, and energy harvesting. First, an IRS-UAV-assisted multi-cluster NOMA-SWIPT communication system model is established, and a multi-objective optimization problem is proposed. For the non-convex optimization problem, based on an alternating iteration strategy, it is decomposed into multiple single-objective optimization sub-problems, which are solved step-by-step using methods such as continuous convex approximation. Compared to a UAV system without an IRS, this method effectively improves the system's sum rate and has better practicality and feasibility.

[0007] Chinese patent CN120602995A discloses a dynamic resource allocation method for a UAV-assisted IRS-MEC system with NCO-NOMA assistance. This method targets an MEC system comprising an IRS with M reflective elements, several users, and several UAVs. The UAVs act as airborne MEC servers, providing computing services to the users. There is no communication link between the users and the UAVs; instead, task offloading is performed with the assistance of the IRS. The method groups users, with each group containing two users and occupying an independent subchannel, and performs task offloading using non-completely overlapping NOMA technology. All user devices are equipped with a single antenna to meet lightweight deployment requirements. The goal of this method is to minimize the task completion latency for all users in the non-completely overlapping NOMA-based UAV-assisted IRS-MEC system, while satisfying the task latency constraints of each user and the energy consumption constraints of each user and UAV. This is achieved by jointly optimizing user transmit power, transmission time, and IRS reflection phase shift to allocate computing resources, implement UAV selection strategies, and determine flight trajectories.

[0008] The disadvantages of the existing technology are: (1) Using drones as small MECs, it is assumed that they can only serve one GMD at a time. This approach has two main limitations. First, it does not fully consider the limited computing power of drones, which may not guarantee the quality of service. Second, it fails to effectively utilize NOMA to improve system throughput. In ultra-dense wireless networks, it may not be able to meet the QoS expectations of many users.

[0009] (2) The AP-GMD or UAV-GMD link is simplified to a LoS model. This simplification ignores the impact of NLoS randomness on the channel, limiting the applicability of the research results in complex urban environments. In actual complex communication environments, especially in densely built-up urban areas or scenarios with many obstacles, signal propagation is often affected by multiple factors such as obstruction, reflection, diffraction, and scattering, with non-line-of-sight components accounting for a significant proportion. When the relevant algorithms are directly applied to real networks, their robustness and reliability are difficult to guarantee, severely limiting the promotion and application value of the research conclusions in actual complex urban environments.

[0010] (3) While focusing primarily on GIRS's enhancement of system performance, it fails to adequately address the fairness requirements among edge users. In multi-user communication scenarios, different users exhibit significant differences in geographical location, channel conditions, and service demands. Pursuing only optimal overall performance can easily lead to resource allocation bias towards central users with better channel conditions, leaving users at the network edge or with poor channel conditions at a long-term disadvantage. This type of approach fails to adequately characterize and guarantee the service quality and fairness requirements of edge users, potentially causing uneven user experience or even system service failures. Therefore, how to improve system performance while simultaneously ensuring fairness among users, especially the minimum service guarantee for edge users, remains a critical issue that has not yet been fully resolved in existing technologies. Summary of the Invention

[0011] To address the problems existing in the prior art, the purpose of this invention is to provide a downlink fair transmission rate optimization method for 5G / 6G UAV-IRS-assisted NOMA-MEC systems.

[0012] To solve the above problems, the present invention adopts the following technical solution.

[0013] The core of this invention is to construct a UAV-IRS-assisted multi-user non-orthogonal multiple access (NOMA) edge computing system. This system consists of a multi-antenna ground access point (AP), a UAV equipped with an intelligent reflector (IRS), and multiple ground mobile devices (GMDs). The UAV provides enhanced reflective communication links to ground users through flexible three-dimensional deployment, while the IRS reconstructs the wireless channel through adjustable reflective phase shifts, thereby improving the overall system transmission performance. The system is applicable to various complex communication scenarios, including line-of-sight and non-line-of-sight scenarios, and possesses strong environmental adaptability.

[0014] In this system, the AP employs a power-domain non-orthogonal multiple access mechanism to simultaneously transmit downlink data to multiple ground mobile devices within the same frequency band. To achieve a synergistic improvement in system fairness and resource utilization efficiency, this invention requires joint decision-making regarding the three-dimensional spatial position of the UAV. IRS reflection phase shift matrix and downlink transmission power allocation for mobile devices in various regions Its optimization objective is to maximize the minimum downlink transmission rate for all users in the system, while satisfying system power constraints, UAV flight altitude constraints, and minimum quality of service requirements for users.

[0015] The above joint optimization objective can be expressed as the following optimization model: in Indicates the first Equivalent downlink achievable speed for a ground mobile device The minimum rate threshold for users, For the total power budget of the system, and These represent the minimum and maximum permitted flight altitudes for the drone, respectively.

[0016] To address the aforementioned high-dimensional, strongly coupled, and non-convex joint optimization problem, this invention introduces auxiliary variables and rate lower bound approximation to transform the original max-min fairness problem into a tractable equivalent form. Based on the block coordinate descent method, the joint optimization problem is decomposed into three sub-problems: IRS reflection phase shift matrix optimization, UAV 3D position optimization, and downlink transmission power allocation optimization. Through alternating iterative solutions, the system's fair transmission rate is continuously improved until the algorithm converges.

[0017] The overall process of this solution can be summarized into the following seven key steps: 1. Define system variables and parameters: Define the optimization variables and fixed parameters in the system, including the UAV's three-dimensional position, IRS phase shift matrix, user power allocation coefficients, and related communication and calculation parameters.

[0018] 2. Construction of Joint Fairness Optimization Model: Based on downlink NOMA transmission and UAV-IRS channel model, a joint optimization problem is constructed with the objective of fair system transmission rate.

[0019] 3. Problem Convexification and Block Coordinate Descent Framework Design: The non-convex optimization model is made convex through equivalent transformation and sequential convex approximation, and an iterative solution framework based on block coordinate descent is constructed.

[0020] 4. IRS Reflection Phase Shift Matrix Optimization: Optimize the phase shift matrix of the smart reflector given the transmission power allocation P and the UAV position Q.

[0021] 5. Optimization of UAV 3D Deployment Position: Optimize the 3D position of the UAV given the power allocation P and the IRS phase shift matrix θ.

[0022] 6. Downlink transmission power allocation optimization: Given the UAV position Q and the IRS phase shift matrix θ, optimize the downlink transmission power allocation for each user.

[0023] 7. Alternating Iterative Updates and Convergence Determination: The variables are iteratively updated according to the block coordinate descent strategy, and the convergence of the algorithm is determined based on the change in the objective function.

[0024] Beneficial effects of the present invention Compared with the prior art, the advantages of this invention are: 1. A max-min optimization model for downlink transmission rate in UAV-IRS-assisted multi-access NOMA-MEC networks is proposed. This model considers the motion of the IRS mounted on the UAV, thus introducing strong coupling between 3D spatial location, IRS channels, and GMD power allocation. By leveraging the mobility of the UAV, the IRS can be dynamically deployed to areas that enhance communication performance. This makes the model particularly suitable for applications such as emergency communications, where rapid network deployment is crucial.

[0025] 2. An efficient alternating iterative algorithm was developed to maximize the minimum transmission rate, thereby improving fairness for users in MEC networks. The core challenge lies in solving a high-dimensional non-convex optimization problem in a stochastic NLoS communication environment. To address this, an approximation based on the expected stochastic transmission rate is first performed, and the BCD technique is applied to decompose the problem into three distinct subproblems. This effectively eliminates the coupling between the UAV's 3D position, the IRS reflection matrix, and the GMD transmit power.

[0026] 3. For the reflection matrix optimization subproblem, the kernel norm and spectral norm are introduced to transform the rank-one non-convex constraint into a non-convex optimization problem with a convexity-discrepancy (DC) structure. Further, the penalized SCA (PSCA) method and first-order Taylor expansion are used to transform the original problem into a semidefinite programming (SDP) problem, which can then be solved efficiently. For the UAV 3D localization optimization subproblem, to decouple the spatial distances between communication nodes in the optimization objective, auxiliary variables are introduced and an algorithm based on the SCA method is designed using convex optimization techniques. In addition, the Lagrange dual transformation technique and the subgradient method are used to solve the power allocation optimization subproblem. Attached Figure Description

[0027] Figure 1 This is a diagram of the NOMA-MEC network structure based on AIRS according to the present invention.

[0028] Figure 2 This is a flowchart of the algorithm of the present invention. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0030] Please see Figures 1 to 2 A downlink fair transmission rate optimization method for 5G / 6G UAV-IRS-assisted NOMA-MEC systems includes: 1. Variable Definition and System Modeling: The system consists of a ground access point (AP), a drone (UAV) equipped with a smart reflector (IRS), and multiple ground mobile devices (GMDs). Let the number of GMDs in the system be... Its index set is represented as Ground access points (APs) are equipped with The root antenna, whose antenna index set is defined as The intelligent reflective surface carried by the drone is made of It consists of several adjustable reflective units, and the set of reflective unit indices is represented as follows: Let the first The position of each GMD in three-dimensional space is represented as: The location of the ground access point (AP) is represented as The three-dimensional position of the drone is represented as To ensure the flight safety of drones, their flight altitude must meet certain constraints. ,in This indicates the permissible flight altitude range for the drone. The reflection phase shift matrix of the IRS is defined as... ,in Indicates the first Phase shift coefficient of each reflecting unit, Thus satisfying the unit modulus constraint The definitions of the system variables and parameters described above form the basis for the subsequent joint optimization modeling and solution in this invention.

[0031] In a UAV-IRS-assisted NOMA-MEC system, the ground access point employs a power-domain non-orthogonal multiple access (NOMA) mechanism to simultaneously transmit downlink data to multiple ground mobile devices within the same frequency band. The transmitted signal of the access point can be represented as... ,in Indicates assignment to the first The transmission power of each GMD This represents the corresponding unit energy information signal, satisfying... The total system transmit power satisfies the constraints. ,in This indicates the maximum transmit power of the access point. The downlink signals received by each GMD are represented as follows: ,in This represents the equivalent channel gain, which is composed of the direct link and the IRS reflection link. This represents additive white Gaussian noise.

[0032] Therefore, the first The signal-to-interference-plus-noise ratio (SINR) of a GMD can be expressed as: Its corresponding downlink achievable rate is ,in This represents the system bandwidth. Considering the optimizability of the rate under random channel conditions, an equivalent channel average gain is introduced. And by approximating the user rate using its upper bound form, we obtain... Based on the above definition, a joint fairness optimization problem is constructed with the objective of maximizing the minimum user downlink transmission rate in the system. Let the UAV position variable be... The power allocation variable is The IRS phase shift variable is Then the joint optimization problem can be expressed as The aforementioned optimization problem achieves a fair rate improvement in the system's downlink by jointly optimizing the UAV's 3D deployment location, the IRS reflection phase shift matrix, and the downlink transmission power allocation.

[0033] 2. Problem Convexification and Block Coordinate Descent Framework Design: The constructed maximum-minimum fair rate optimization problem has an objective function and constraints that include logarithmic functions and fractional structures related to the signal-to-interference-plus-noise ratio (SINR), making the original problem a highly non-convex optimization problem. To reduce the solution complexity, auxiliary variables are introduced. The original maximum-minimum problem is equivalently transformed into the following form: Due to the rate expression The function contains nonlinear coupling terms. To further improve solvability, a sequential convex approximation method is used to approximate the rate function with a lower bound. Based on the convexity-induced inequality, the following is applied: Perform linearization, given the previous iteration... Under the conditions, the first The lower bound of the rate for a single user can be expressed as follows: ,in .

[0034] Based on the aforementioned rate lower bound, the original problem (P2) can be further approximated as follows: To solve the aforementioned non-convex optimization problem, a Block Coordinate Descent (BCD) method is employed. This method divides the optimization variables into multiple sub-variable blocks, and then alternately optimizes each variable block while keeping the other variables constant. Specifically, the optimization problem is decomposed into three sub-problems: UAV 3D position optimization, IRS reflection phase shift matrix optimization, and downlink transmission power allocation optimization. By iteratively solving these sub-problems and updating the relevant variables, the algorithm continues until it meets the convergence condition, thereby obtaining a feasible solution to the original optimization problem.

[0035] 3. Block Coordinate Descent Iterative Update and Convergence Criterion: First, under the condition of fixed UAV position Q and downlink power allocation P, the reflection phase shift matrix of the Intelligent Reflector (IRS) is optimized. Since the rate expression and rank constraint are both non-convex, a sequential convex approximation method is used to approximate the interference term with a first order, and a nuclear norm penalty term is introduced to relax the rank constraint, thereby transforming the original non-convex problem into a semidefinite programming form. Under the given penalty factor, the matrix variables are continuously updated by iteratively solving the convex optimization problem. Continue until the algorithm converges.

[0036] Secondly, with a fixed IRS reflection phase shift matrix Given the downlink power allocation P, the three-dimensional deployment position Q of the UAV is... Optimization is then performed. By introducing a sequential convex approximation method, the non-convex constraints are approximated to a first order, transforming the UAV position optimization subproblem into a convex optimization problem. Given the solution from the previous iteration, the UAV's 3D position is updated by solving the convex subproblem, gradually approximating the feasible solution of the original problem during the iteration process. This iterative process is repeated until the objective function converges, yielding the optimized result of the UAV's 3D deployment position.

[0037] Finally, fix the UAV's three-dimensional position Q and the IRS reflection phase shift matrix. Under these conditions, the downlink transmission power allocation P for mobile devices in various locations is as follows: Optimization is then performed. At this point, the fair rate optimization problem can be transformed into a problem concerning only the power allocation variable P and auxiliary variables. The subproblem aims to maximize the minimum user rate of the system. For the convex power allocation problem, a dual decomposition method is employed. By constructing a Lagrangian function and introducing corresponding dual variables, the optimal power allocation solution for each user is obtained under the condition of fixing the dual variables. Subsequently, by updating the dual variables and iteratively solving, joint optimization of power allocation and system fair rate is achieved. This iterative process is repeated until the power allocation results and the objective function value converge, yielding the downlink transmission power allocation scheme for various ground-based mobile devices.

[0038] The aforementioned sub-problems are executed alternately within a unified BCD framework. In each iteration, the system sequentially updates the IRS reflection matrix, UAV position, and user power allocation, and terminates based on changes in the objective function or the convergence of variables. Simulation results show that the method converges stably within a finite number of iterations and achieves a good balance between system performance and computational complexity.

[0039] In summary, this invention proposes a multivariate joint optimization method for UAV-assisted intelligent reflective communication systems. It constructs a closed-loop optimization framework that integrates system modeling, problem decomposition, iterative solution, and convergence determination, aiming to solve the non-convex optimization problem caused by the high coupling between UAV deployment, IRS configuration, and power allocation in multi-user scenarios.

[0040] This solution first formalizes the system performance optimization problem into a joint optimization model aimed at maximizing the minimum achievable rate for users by uniformly modeling the UAV's 3D position, user transmission power, and IRS reflection coefficient. Building upon this, and addressing the non-convexity and strong coupling of variables prevalent in the original problem, the solution employs Block Coordinate Descent (BCD) to decompose the overall problem into three clearly structured sub-problems: IRS reflection matrix optimization, UAV 3D position optimization, and user power allocation optimization. This significantly reduces the solution complexity.

[0041] In the specific solution process, this invention combines optimization tools such as Sequential Convex Approximation (SCA), semidefinite relaxation, penalty function method, and Lagrange duality decomposition to perform equivalent transformations and efficient solutions to each subproblem. This allows non-convex constraints, which were originally difficult to handle directly, to be gradually approximated into forms solvable by standard convex optimization tools. Simultaneously, by alternately updating various decision variables within a unified iterative framework, continuous improvement in system performance and stable convergence are achieved.

[0042] Ultimately, this technical solution establishes a joint optimization mechanism capable of adaptively coordinating UAV deployment, IRS reflection control, and power resource allocation. While ensuring algorithm convergence and feasibility, it effectively improves the system's spectral efficiency and user fairness. This method possesses good versatility and scalability, providing strong technical support for the design and deployment of future complex air-to-ground integrated communication systems and intelligent reflection communication networks.

[0043] The above description is merely a preferred embodiment of the present invention; however, the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and its improved concepts, should be covered within the scope of protection of the present invention.

Claims

1. A downlink fairness transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G, characterized in that, Includes the following steps: Step 1: Define system variables and parameters: Define the optimization variables and fixed parameters in the system, including the UAV's three-dimensional position, IRS phase shift matrix, and user power allocation coefficients; Step 2, Construction of Joint Fairness Optimization Model: Based on the downlink NOMA transmission and UAV-IRS channel model, a joint optimization problem is constructed with the system fair transmission rate as the objective. Step 3, Problem Convexification and Block Coordinate Descent Framework Design: The non-convex optimization model is made convex through equivalent transformation and sequential convex approximation, and an iterative solution framework based on block coordinate descent is constructed. Step 4: Optimize the phase shift matrix of the IRS reflection: Given the transmission power allocation P and the UAV position Q, optimize the phase shift matrix of the smart reflector. Step 5: Optimize the 3D deployment position of the UAV: ​​Given the transmission power allocation P and the IRS phase shift matrix θ, optimize the 3D position of the UAV. Step 6, Downlink Transmission Power Allocation Optimization: Given the UAV position Q and the IRS phase shift matrix θ, optimize the downlink transmission power allocation for each user; Step 7: Alternating Iterative Update and Convergence Determination: Iteratively update each variable according to the block coordinate descent strategy, and determine whether the algorithm has converged based on the change in the objective function.

2. The downlink fair transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G as described in claim 1, characterized in that: In step 1, the key variables and parameters of the UAV-IRS-assisted NOMA-MEC system are first uniformly defined and modeled to provide a foundation for subsequent fairness optimization; Consider a multi-user access UAV-IRS assisted NOMA-MEC network. The system consists of a ground access point (AP), a UAV equipped with a smart reflector (IRS), and multiple ground mobile devices (GMDs). Let the number of GMDs in the system be... Its index set is represented as ; Ground access point AP is equipped with The root antenna, whose antenna index set is defined as The intelligent reflective surface carried by the drone is made of It consists of several adjustable reflective units, and the set of reflective unit indices is represented as follows: Let the first The position of each GMD in three-dimensional space is represented as: The location of the ground access point (AP) is represented as follows: The three-dimensional position of the drone is represented as To ensure the safety of drone flight, its flight altitude must meet certain constraints. in Indicates the permitted flight altitude range for the drone. The reflection phase shift matrix of IRS is defined as in Indicates the first Phase shift coefficient of each reflecting unit, Thus satisfying the unit modulus constraint .

3. The downlink fair transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G as described in claim 2, characterized in that: In step 2, in the UAV-IRS-assisted NOMA-MEC system, the ground access point adopts a power-domain non-orthogonal multiple access (NOMA) mechanism to simultaneously transmit downlink signals to multiple ground mobile devices within the same frequency band. The transmission signal of the access point can be represented as... in Indicates assignment to the first The transmission power of each GMD This represents the corresponding unit energy information signal, satisfying... The total transmit power of the system meets the constraints. in Indicates the maximum transmit power of the access point. No. The downlink signals received by each GMD are represented as follows: in This represents the equivalent channel gain, which is composed of the direct link and the IRS reflection link. This represents additive white Gaussian noise. Therefore, the first The signal-to-interference-plus-noise ratio (SINR) of a GMD can be expressed as: Its corresponding downlink reachable rate is in Indicates system bandwidth. Considering the optimizability of the rate under random channel conditions, an equivalent channel average gain is introduced. And by approximating the user rate using its upper bound form, we obtain... Based on the above definition, a joint fairness optimization problem is constructed with the objective of maximizing the minimum user downlink transmission rate in the system. Let the UAV position variable be... The power allocation variable is The IRS phase shift variable is Then the joint optimization problem can be expressed as 。 4. The downlink fair transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G as described in claim 3, characterized in that: In step 3, the maximum-minimum fair rate optimization problem constructed in step 2 has an objective function and constraints that include a logarithmic function and a fractional structure of signal-to-interference-plus-noise ratio, making the original problem a highly non-convex optimization problem. To reduce the solution complexity, auxiliary variables are introduced. The original maximum-minimum problem is equivalently transformed into the following form: Due to the rate expression The function contains nonlinear coupling terms. To further improve solvability, a sequential convex approximation method is used to approximate the rate function with a lower bound. Based on the convexity-induced inequality, the following is applied: Perform linearization, given the previous iteration... Under the conditions, the first The lower bound of the rate for a single user can be expressed as follows: in Based on the aforementioned lower bound on the rate, the original problem P2 can be further approximated as follows: To solve the aforementioned non-convex optimization problem, a block coordinate descent method is adopted. The optimization variables are divided into multiple sub-variable blocks, and each variable block is optimized alternately while keeping the other variables fixed. The optimization problem is decomposed into the following three sub-problems: UAV three-dimensional position optimization, IRS reflection phase shift matrix optimization, and downlink transmission power allocation optimization. The above sub-problems are solved iteratively, and the relevant variables are updated until the algorithm meets the convergence condition, thereby obtaining a feasible solution to the original optimization problem.

5. The downlink fair transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G as described in claim 4, characterized in that: For step 4, at the fixed drone position and downlink power allocation Under these conditions, optimization is performed on the reflection phase shift matrix of the intelligent reflector IRS. In this case, the joint fair rate optimization problem can be transformed into an optimization problem concerning only the IRS phase shift variable. and auxiliary variables The subproblem of is represented in the form of . It also satisfies the IRS unity modulus constraint and the user rate constraint. To facilitate optimization, an equivalent channel matrix is ​​introduced. , will the Average channel gain of individual users Rewritten as a quadratic form in terms of the IRS phase shift variable: in For constant terms that are independent of IRS, Accordingly, the unit modulus constraint of the IRS phase shift matrix can be equivalently expressed as: Under the above representation, the first Approximate downlink rate for each user Can be written as in and Let these represent the logarithmic functions corresponding to the signal term and the interference term, respectively. Since both the rate expression and the rank constraint are non-convex, a sequential convex approximation method is used to approximate the disturbance term with a first order, and a nuclear norm penalty term is introduced to relax the rank constraint. This transforms the original non-convex problem into a semidefinite programming problem. Given the penalty factor, the convex optimization problem is solved iteratively, continuously updating the matrix variables. Continue until the algorithm converges.

6. The downlink fair transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G as described in claim 5, characterized in that: In step 5, the IRS reflection phase shift matrix is ​​fixed. and downlink power allocation Under these conditions, the three-dimensional deployment location of the drone With optimization, the fair rate optimization problem can now be transformed into a problem that depends only on the drone's position variable. and auxiliary variables The subproblem of is represented in the form of . It also meets the constraints of UAV flight altitude and user rate. Because the user rate expression includes distance terms between the drone and the access point, and between the drone and the user, the optimization variables... The objective function and constraints exhibit strong nonlinear coupling, making the original problem a nonconvex optimization problem. To reduce the solution complexity, a set of auxiliary variables is introduced to characterize the upper and lower bounds of the distance between the UAV and various ground nodes, thus reconstructing the original rate expression into an equivalent form with respect to the auxiliary variables. Based on this, an approximate expression for user rate is given. Perform an equivalent transformation and construct a feasible lower bound on the drone's position to obtain... in and This represents the upper and lower bounds of the equivalent channel gain related to the drone's location. By introducing a sequential convex approximation method to perform first-order approximation on non-convex constraints, the UAV position optimization subproblem can be transformed into a convex optimization problem. Given the solution of the previous iteration, the UAV's 3D position is updated by solving the convex subproblem, and the feasible solution of the original problem is gradually approximated during the iteration process. The above iteration process is repeated until the objective function value converges, and the optimization result of the UAV's 3D deployment position is obtained.

7. The downlink fair transmission rate optimization method for a UAV-IRS-assisted NOMA-MEC system for 5G / 6G as described in claim 5, characterized in that: For step 6, fix the three-dimensional position of the UAV. and IRS reflection phase shift matrix Under these conditions, downlink transmission power allocation for mobile devices on various ground surfaces Optimization is then performed, at which point the fair rate optimization problem can be transformed into one that depends only on the power allocation variable. and auxiliary variables The subproblem is to maximize the minimum user rate of the system. Due to the approximate expression for user rate It still contains non-convex terms. To reduce the solution complexity, variable substitution is introduced. Rewrite the rate function as The above expression consists of linear terms, logarithms, and functions, and the whole expression is about the variable. It is a concave function. Based on this variable transformation, the power allocation subproblem can be transformed into a convex optimization problem, with constraints including minimum user rate constraints and total system power constraints. For the convex power allocation problem, a dual decomposition method is employed. By constructing a Lagrangian function and introducing corresponding dual variables, the optimal power allocation solution for each user is obtained under the condition of fixing the dual variables. Subsequently, by updating the dual variables and iteratively solving the problem, joint optimization of power allocation and system fair rate is achieved. This iterative process is repeated until the power allocation results and the objective function value converge, thus obtaining the downlink transmission power allocation scheme for various ground-based mobile devices.

Images