Resource allocation method for unmanned aerial vehicle-assisted edge computing network based on intelligent reflecting surface assistance

Abstract

A resource allocation method for an unmanned aerial vehicle-assisted edge computing network based on intelligent reflecting surface assistance, comprising: constructing an unmanned aerial vehicle-assisted edge computing network system model based on intelligent reflecting surface assistance; constructing a communication model, a time delay model and an energy consumption model; according to the communication model, the time delay model and the energy consumption model, constructing a maximum average throughput function; optimizing the maximum average throughput function by using successive convex approximation and block coordinate descent algorithms, and according to the optimized maximum average throughput function, calculating an optimal resource allocation result. The present application uses an intelligent reflecting surface to assist in an unmanned aerial vehicle-assisted network system, thereby increasing the channel gain of links between automobiles and unmanned aerial vehicles.

Description

A UAV-assisted edge computing network resource allocation method based on intelligent reflective surface assistance

[0001] This application claims priority to the Chinese patent application filed with the Patent Office of China on July 24, 2023, with application number 2023109129642 and invention name “A method for allocating network resources of drone-assisted edge computing based on intelligent reflective surface assistance”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the technical field of drone resource allocation, and in particular to a drone-assisted edge computing network resource allocation method based on intelligent reflective surface assistance. Background Art

[0003] The data obtained by the vehicle's sensors cannot fully analyze the environment around the vehicle. By coordinating the fusion and offloading of multi-vehicle perception data through drones, a complete and timely real-time and comprehensive perception map can be provided to the vehicle, thereby providing guarantees for the safe driving decisions of intelligent vehicles.

[0004] However, given the numerous buildings, roads, and other obstacles in urban environments, these obstacles can affect the task offloading link between UAVs and ground vehicles, reducing the offloading rate. Especially when the distance between the UAV and ground vehicle is long, signal attenuation caused by occlusion is more pronounced, further exacerbating the decline in offloading rate. Therefore, ensuring the offloading rate of UAV-assisted intelligent vehicles in the considered obstacle-occluded scenario is a worthy research issue.

[0005] Smart reflectors offer low power consumption, high energy efficiency, and the ability to reconfigure wireless communication transmission environments. They can effectively assist drones in improving the quality of their communication links with ground terminals, ensuring the latency requirements of edge computing. Therefore, this application proposes a smart reflector-enabled drone-assisted vehicle edge computing offloading strategy. This strategy aims to maximize the average throughput of vehicle terminals by jointly optimizing task offloading, the smart reflector phase shift matrix, vehicle transmit power, and drone computing resource allocation.

[0006] Summary of the Invention

[0007] To solve the problems existing in the above-mentioned prior art, this application proposes a method for allocating network resources of drone-assisted edge computing based on intelligent reflective surface assistance, including:

[0008] S1: Construct a UAV-assisted edge computing network system model assisted by intelligent reflective surfaces;

[0009] S2: Construct communication model, delay model and energy consumption model based on the UAV-assisted edge computing network system model assisted by the intelligent reflective surface;

[0010] S3: Construct the maximum average throughput function based on the communication model, delay model and energy consumption model;

[0011] S4: Optimize the maximum average throughput function using continuous convex approximation and block drop algorithm, and calculate the optimal resource allocation result based on the optimized maximum average throughput function.

[0012] Preferably, the intelligent reflective surface-assisted drone-assisted edge computing network system model includes a drone, an intelligent reflective panel and a user terminal; wherein the user terminal is used to generate computing tasks; and the computing tasks are uploaded to the drone through the intelligent reflective panel.

[0013] Preferably, constructing a communication model includes: obtaining the transmission distance between the user terminal and the drone and the channel gain of the link between the user terminal and the drone based on the position information of the user terminal and the drone; obtaining the transmission distance between the user terminal and the smart reflecting surface and the channel gain between the user terminal and the smart reflecting surface link based on the position information of the user terminal and the smart reflecting surface; obtaining the transmission distance and the channel gain between the drone and the smart reflecting surface link based on the position information of the smart reflecting surface and the drone.

[0014] Furthermore, the expression of the communication model is:

[0015] Where Θ represents the phase shift matrix of the smart reflector, β n Represents the reflection amplitude of the smart reflective surface element, θ n Represents the phase coefficient of the smart reflective surface element, h UR represents the channel gain between the UAV and the smart reflective surface, α0 represents the path loss when the reference distance is 1 meter, and d UR represents the transmission distance between the UAV and the intelligent reflective surface, λ is the carrier wavelength, N is the number of reflective units, φ un is the cosine value of the departure angle between the UAV and the smart reflector link, represents the channel gain between the UAV and the vehicle, represents the straight-line distance between the UAV and vehicle k, is a complex Gaussian distribution with a mean of 0 and a variance of 1, η1 represents the path loss index between links, represents the channel gain between the smart reflective surface and vehicle k, d RV represents the straight-line distance between the intelligent reflective surface and the vehicle, β is the Rice index, They represent the time-of-day gain part and the non-line-of-sight gain part between the intelligent reflective surface and the vehicle, r k represents the transmission rate of vehicle k’s computing task uploaded to the UAV, B k represents the bandwidth of a single vehicle, P k represents the transmission power of vehicle k, σ 2 represents the Gaussian white noise power.

[0016] Preferably, the expression of the delay model is:

[0017] in, Indicates the amount of tasks handled locally by the vehicle; C k Indicates the number of cycles required for the CPU to calculate each bit; represents the computing resources of vehicle k for processing local computing tasks; Represents the computing resources of the drone to process user k's computing tasks.

[0018] Preferably, the expression of the energy consumption model is:

[0019] in, represents the energy consumed by vehicle k to upload mission data to the UAV; Calculate energy consumption locally; T k The time delay consumed by the vehicle to upload mission data to the UAV; P k is the transmission power of the UAV; ε is the effective capacitance coefficient of the CPU; Provide local computing resources for vehicles; Calculate latency locally for task data.

[0020] Preferably, the optimized maximum average throughput function expression is:

[0021] stC1:

[0022] C2:

[0023] C3:

[0024] C4:

[0025] C5:

[0026] C6:

[0027] C7:

[0028] C8:0≤θ n ≤2π

[0029] Among them, K represents the number of vehicles, T k represents the total delay of the kth vehicle uploading the computing task, r k Indicates the transmission rate of the vehicle uploading computing tasks, The C1 constraint represents the range of computing resources allocated by the UAV to vehicle k; the C2 constraint represents the range of total computing resources allocated by the UAV to the vehicle; the C3 constraint represents the delay constraint for the UAV to process computing tasks; the C4 constraint represents whether the computing tasks uploaded by vehicle k to the UAV can be fully calculated on the UAV side; the C5 constraint represents whether the computing tasks transmitted by vehicle k to the UAV can be fully transmitted; the C6 constraint represents the upper and lower limits of the vehicle's signal transmission power; the C7 constraint represents the upper limit of the vehicle's energy consumption for processing perception data; and the C8 constraint represents the phase shift adjustment range constraint for the intelligent reflective surface.

[0030] Preferably, the process of optimizing the maximum average throughput function using continuous convex approximation and block landing algorithm includes: fixing task unloading, vehicle generation power and UAV computing resource allocation, converting the maximum average throughput function P3 into a sub-problem P3.1 about the variable Θ according to the fixed task unloading, vehicle generation power and UAV computing resource allocation, solving the sub-problem to obtain the optimal phase shift matrix of the smart reflection surface; fixing the smart reflection surface phase shift matrix, converting the maximum average throughput function P3 into a sub-problem P3.2 about the variables D, P, and f according to the fixed smart reflection surface phase shift matrix, where D is task unloading, P is the transmission power of the vehicle terminal, and f is the computing resource allocation of the UAV; solving the sub-problem P3.2 to obtain the results of task unloading, the transmission power of the vehicle terminal and the computing resource allocation of the UAV.

[0031] Beneficial effects of this application:

[0032] This application uses intelligent reflective surfaces to assist the UAV network system, thereby improving the channel gain of the link between the vehicle and the UAV. This application maximizes the average vehicle throughput by jointly optimizing the computing task volume, computing resources and offloading decisions, intelligent reflective surface phase shift, and vehicle transmission power allocation of the UAV-assisted edge computing system. By solving the maximum average throughput optimization problem, the optimal resource allocation solution is obtained; compared with the existing baseline method, the method proposed in this application effectively improves the performance of the system, and also achieves the effect of dynamically dividing computing resources. This application uses UAVs as aerial base stations and can be used in wireless network scenarios with limited infrastructure; considering the application of intelligent reflective surfaces to UAV network systems, system resources are reasonably scheduled to improve system performance, which is highly practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] FIG1 is a flow chart of a resource allocation method for drone-assisted edge computing based on intelligent reflective surface assistance in this application;

[0034] FIG2 is a schematic diagram of the IoT device system model for drone-assisted edge computing in this application;

[0035] Figure 3 shows the relationship between the number of users and average throughput under different smart reflective surface elements;

[0036] Figure 4 shows the relationship between the number of smart reflective surface elements and the average vehicle throughput. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0038] A method for allocating network resources of a UAV-assisted edge computing system based on intelligent reflective surface assistance is shown in FIG1 . The method includes:

[0039] S1: Construct a UAV-assisted edge computing network system model assisted by intelligent reflective surfaces;

[0040] S2: Construct communication model, delay model and energy consumption model based on the UAV-assisted edge computing network system model assisted by the intelligent reflective surface;

[0041] S3: Construct the maximum average throughput function based on the communication model, delay model and energy consumption model;

[0042] S4: Optimize the maximum average throughput function using continuous convex approximation and block drop algorithm, and calculate the optimal resource allocation result based on the optimized maximum average throughput function.

[0043] In this embodiment, the specific structure of a reflective surface-assisted drone-assisted edge computing network system model is disclosed. As shown in Figure 2, the structure includes: a drone, an intelligent reflective panel, and multiple ground vehicles; part of the computing tasks generated by the vehicle is retained for local calculation, and part is uploaded to the drone for calculation with the assistance of the intelligent reflective surface.

[0044] Figure 3 shows the relationship between the number of users and average throughput for different numbers of smart reflective surface elements. The experimental results in the figure show that as the number of vehicle terminals increases, the average throughput of the vehicle terminals decreases. This is because as the number of terminal devices increases, the computing resources available to each vehicle terminal decrease, which limits the processing rate of computing tasks and thus reduces the average vehicle throughput. Furthermore, as the number of vehicle terminals increases, network congestion and interference issues also increase, further affecting system performance and vehicle terminal throughput.

[0045] Figure 4 shows the relationship between the number of smart reflective surface elements and the average vehicle throughput. As shown in the figure, when the number of smart reflective surface sub-surfaces remains constant, the vehicle terminal throughput increases as the number of vehicle terminals decreases. Notably, despite the difference in the number of vehicle terminals, the average vehicle throughput continues to rise in all cases as long as the number of smart reflective surface sub-surfaces increases. This is because increasing the number of smart reflective surface sub-surfaces more effectively controls signal reflection and propagation, improving the communication link gain between the drone and vehicle terminals, and thus increasing the efficiency of vehicle data offload transmission. Specifically, when the number of smart reflective surface sub-surfaces differs by 10, the throughput difference is approximately 8%. Experiments show that the introduction of smart reflective surfaces allows signals to bypass numerous obstacles, effectively avoiding potential attenuation in signal transmission between the drone and vehicle terminals, thereby effectively improving the rate and effectiveness of data offload transmission.

[0046] In this embodiment, the process of constructing a communication model includes: obtaining a phase shift matrix of the smart reflective surface based on the amplitude coefficient and phase coefficient of the reflective element of the smart reflective surface; obtaining a channel gain between the UAV and the vehicle based on the transmission distance between the UAV and the vehicle; obtaining a channel gain between the vehicle and the smart reflective surface based on the transmission distance between the vehicle and the smart reflective surface; obtaining a channel gain between the UAV and the smart reflective surface based on the transmission distance between the UAV and the smart reflective surface; and calculating the information transmission rate of the vehicle computing task based on the channel gain between the UAV and the vehicle, the channel gain between the UAV and the smart reflective surface, the channel gain between the vehicle and the smart reflective surface, and the smart reflective surface phase shift matrix.

[0047] The expression of the communication model is:

[0048] Where Θ represents the phase shift matrix of the smart reflector, β n Represents the reflection amplitude of the smart reflective surface element, θ n Represents the phase coefficient of the smart reflective surface element, h UR represents the channel gain between the UAV and the smart reflective surface, α0 represents the path loss when the reference distance is 1 meter, and d UR represents the transmission distance between the UAV and the intelligent reflective surface, λ is the carrier wavelength, N is the number of reflective units, φ un =(x u -x n ) / d UR is the cosine value of the link departure angle between the UAV and the smart reflector, x u is the horizontal coordinate of the UAV, x n is the horizontal coordinate of the smart reflective surface, h UVk represents the channel gain between the UAV and the vehicle, represents the straight-line distance between the UAV and vehicle k, is a complex Gaussian distribution with a mean of 0 and a variance of 1, η1 represents the path loss index between links, represents the channel gain between the smart reflective surface and vehicle k, d RV represents the straight-line distance between the intelligent reflective surface and the vehicle, β is the Rice index, They represent the time-of-day gain part and the non-line-of-sight gain part between the intelligent reflective surface and the vehicle, r k represents the transmission rate of vehicle k’s computing task uploaded to the UAV, B k represents the bandwidth of a single vehicle, P k represents the transmission power of vehicle k, σ 2 represents the Gaussian white noise power.

[0049] In this embodiment, the process of constructing the delay model includes: calculating the local computing processing delay of the task based on the task volume processed locally by the vehicle and the local computing resources of the vehicle; calculating the task processing delay on the drone side based on the task volume unloaded by the vehicle and the computing resources of the drone.

[0050] The expression of the delay model is:

[0051] in, Indicates the amount of tasks handled locally by the vehicle; C k Indicates the number of cycles required for the CPU to calculate each bit; represents the computing resources of vehicle k for processing local computing tasks; Represents the computing resources of the drone to process user k's computing tasks.

[0052] In this embodiment, the process of constructing the energy consumption model includes: calculating the task transmission energy consumption from vehicle k to the drone based on the delay consumed by the vehicle uploading the data task to the drone and the transmission power of the drone; and obtaining the task local processing energy consumption based on the vehicle calculation delay.

[0053] The expression of the energy consumption model is:

[0054] in, represents the energy consumed by vehicle k to upload mission data to the UAV; Calculate energy consumption locally; T k The time delay consumed by the vehicle to upload mission data to the UAV; P k is the transmission power of the UAV; ε is the effective capacitance coefficient of the CPU; Provide local computing resources for vehicles; Calculate latency locally for task data.

[0055] In this embodiment, the process of constructing the maximum average throughput problem of a vehicle includes: calculating the total delay of vehicle data transmission based on a delay model; constructing a maximum average throughput objective function based on the delay model; constructing constraints based on task offloading decisions, the phase shift matrix of the intelligent reflector, the vehicle transmission power, and the allocation of UAV computing resources; and constructing the maximum average throughput problem based on the maximum average throughput objective function and the constraints.

[0056] The maximum average throughput problem is expressed as:

[0057] Among them, K represents the number of vehicles, T k represents the total delay of the kth vehicle uploading the computing task, rk Indicates the transmission rate of the vehicle uploading computing tasks, The amount of tasks that the vehicle remains to perform local calculations.

[0058] The optimization problem is constructed based on the large average throughput problem, where the optimization problem expression is:

[0059] stC1:

[0060] C2:

[0061] C3:

[0062] C4:

[0063] C5:

[0064] C6:

[0065] C7:

[0066] C8:0≤θ n ≤2π

[0067] Among them, K represents the number of vehicles, T k represents the total delay of the kth vehicle uploading the computing task, r k Indicates the transmission rate of the vehicle uploading computing tasks, The C1 constraint represents the range of computing resources allocated by the UAV to vehicle k; the C2 constraint represents the range of total computing resources allocated by the UAV to the vehicle; the C3 constraint represents the delay constraint for the UAV to process computing tasks; the C4 constraint represents whether the computing tasks uploaded by vehicle k to the UAV can be fully calculated on the UAV side; the C5 constraint represents whether the computing tasks transmitted by vehicle k to the UAV can be fully transmitted; the C6 constraint represents the upper and lower limits of the vehicle's signal transmission power; the C7 constraint represents the upper limit of the vehicle's energy consumption for processing perception data; and the C8 constraint represents the phase shift adjustment range constraint for the intelligent reflective surface.

[0068] Due to the tight coupling and non-convexity between variables in the optimization problem, it is difficult to solve it directly using traditional methods. To address this issue, this application uses Block Coordinate Descent (BCD) to decouple the complex non-convex optimization problem into two easily solvable subproblems. Successive Convex Approximation (SCA) is then used to solve the sub-optimal solution. This approach is followed by alternating optimizations, ultimately yielding an approximate sub-optimal solution to the original optimization problem.

[0069] In this embodiment, the process of optimizing the maximum average throughput function using the continuous convex approximation and block landing algorithm includes: fixing task offloading, vehicle generation power, and UAV computing resource allocation, converting the maximum average throughput function P3 into a subproblem P3.1 regarding the variable Θ based on the fixed task offloading, vehicle generation power, and UAV computing resource allocation, solving the subproblem to obtain the optimal phase shift matrix of the smart reflective surface; fixing the smart reflective surface phase shift matrix, converting the maximum average throughput function P3 into a subproblem P3.2 regarding the variables D, P, and f based on the fixed smart reflective surface phase shift matrix, where D is task offloading, P is the transmission power of the vehicle terminal, and f is the computing resource allocation of the UAV; solving subproblem P3.2 to obtain the results of task offloading, vehicle terminal transmission power, and UAV computing resource allocation.

[0070] The specific calculation process includes:

[0071] First, we fix the variables D, P, and f, and the P3 problem is transformed into a subproblem about the variable Θ. This subproblem can be expressed as:

[0072] stC8

[0073] According to the triangle inequality, the objective function in P3.1 satisfies the following conditions:

[0074] When the equation is equal, the objective function of P3.1 reaches its maximum value. Then, by adjusting the phase shift matrix of the RIS, the phase of the reflected link and the direct link are kept consistent, thereby achieving the optimal reflection phase shift:

[0075] in, is the phase of the UV link; and h RV,k are the kth element of the UR link channel vector and the kth element of the RV link channel vector, respectively.

[0076] After solving the RIS variable Θ, the subproblems of the target problem P3 with respect to the variables D, P, and f can be constructed as follows:

[0077] stC1:

[0078] C2:

[0079] C3:

[0080] C4:

[0081] C5:

[0082] C6:

[0083] C7:

[0084] Among them, the left side of C5 is about P k The left side of C7 is a non-convex function. Since C5 and C7 are both non-convex constraints, the optimization problem is still a non-convex problem. Definition Performing a first-order Taylor expansion on C5 gives:

[0085] Among them, P ki is the local optimal point produced at the i-th iteration. Constraint C5 can be reformulated as:

[0086] For non-convex terms with product structures on the left side of the C7 constraint definition Since h1(D) and h2(P) are both non-negative and convex, by giving the feasible solution of the i-th iteration available Convex approximation of Its expression is:

[0087] Constraint C7 can be reformulated as:

[0088] Subproblem P3.2 can be restated as:

[0089] P3.2

[0090] stC1:

[0091] C2:

[0092] C3:

[0093] C4:

[0094] C6:

[0095] C5′:

[0096] C7′:

[0097] Subproblem P3.2 is a standard convex problem and can be solved using the convex optimization solver CVX. By iterating these two subproblems using an alternating optimization strategy until convergence, a suboptimal solution to the original optimization problem can be obtained.

[0098] The above embodiments further illustrate the purpose, technical solutions and advantages of the present application. It should be understood that the above embodiments are only preferred implementation methods of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made to the present application within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for allocating network resources of UAV-assisted edge computing based on intelligent reflective surface, characterized in that: include: S1: Construct a UAV-assisted edge computing network system model assisted by intelligent reflective surfaces; S2: Construct communication model, delay model and energy consumption model based on the UAV-assisted edge computing network system model assisted by the intelligent reflective surface; S3: Construct the maximum average throughput function based on the communication model, delay model and energy consumption model; S4: The maximum average throughput function is optimized using continuous convex approximation and block drop algorithm, and the optimal resource allocation result is calculated based on the optimized maximum average throughput function.

2. According to claim 1, a method for allocating network resources of drone-assisted edge computing based on intelligent reflective surface assistance is characterized in that: The intelligent reflective surface-assisted drone-assisted edge computing network system model includes a drone, an intelligent reflective panel and a user terminal; the user terminal is used to generate computing tasks; and the computing tasks are uploaded to the drone through the intelligent reflective panel.

3. According to claim 1, a method for allocating network resources of drone-assisted edge computing based on intelligent reflective surface assistance is characterized in that: Building a communication model includes: obtaining the transmission distance between the user terminal and the UAV and the channel gain of the link between the user terminal and the UAV according to the position information of the user terminal and the UAV; obtaining the transmission distance between the user terminal and the intelligent reflecting surface and the channel gain between the user terminal and the intelligent reflecting surface according to the position information of the user terminal and the intelligent reflecting surface; obtaining the transmission distance and the channel gain between the UAV and the intelligent reflecting surface according to the position information of the intelligent reflecting surface and the UAV.

4. The method for allocating network resources of a drone-assisted edge computing system based on intelligent reflective surface assistance according to claim 3 is characterized in that: The expression of the communication model is: Where Θ represents the smart reflection surface phase shift matrix, β n represents the reflection amplitude of the smart reflective surface element, θ n represents the phase coefficient of the smart reflector element, h UR represents the channel gain between the UAV and the smart reflective surface, α0 represents the path loss when the reference distance is 1 meter, and d UR represents the transmission distance between the UAV and the intelligent reflective surface, λ is the carrier wavelength, N represents the number of reflective units, φ un is the cosine value of the link departure angle between the UAV and the smart reflector, represents the channel gain from the UAV to the vehicle, represents the straight-line distance between the UAV and vehicle k, is a complex Gaussian distribution with a mean of 0 and a variance of 1. η1 represents the path loss index between links. represents the channel gain between the smart reflective surface and vehicle k, d RV represents the straight-line distance between the intelligent reflective surface and the vehicle, β is the Rice index, They represent the time-to-distance gain part and the non-line-of-sight gain part between the intelligent reflective surface and the vehicle, r k represents the transmission rate of the vehicle k computing task uploaded to the UAV, B k represents the bandwidth of a single vehicle, P k represents the transmission power of vehicle k, σ 2 represents the Gaussian white noise power.

5. The method for allocating network resources of UAV-assisted edge computing based on intelligent reflective surface assistance according to claim 1 is characterized in that: The expression of the delay model is: in, represents the amount of tasks handled locally by the vehicle; C k Indicates the number of cycles required for the CPU to calculate each bit; represents the computing resources of vehicle k for processing local computing tasks; Represents the computing resources of the drone for processing user k's computing tasks.

6. The method for allocating network resources of UAV-assisted edge computing based on intelligent reflective surface assistance according to claim 1, characterized in that: The expression of energy consumption model is: in, represents the energy consumed by vehicle k to upload mission data to the UAV; for Local computing energy consumption; T k The delay consumed by the vehicle uploading mission data to the drone; P k is the transmission power of the drone; ε is the effective capacitance coefficient of the CPU; Provide local computing resources for vehicles; Calculate latency locally for task data.

7. The method for allocating network resources of UAV-assisted edge computing based on intelligent reflective surface assistance according to claim 1 is characterized in that: The optimized maximum average throughput function expression is: C8:0≤θ n ≤2π Where K represents the number of vehicles, T k represents the total delay of the kth vehicle uploading the computing task, r k Indicates the transmission rate of the vehicle uploading computing tasks, is the amount of tasks that the vehicle leaves to calculate locally; constraint C1 indicates the range of computing resources that the UAV allocates to vehicle k; constraint C2 indicates the range of total computing resources that the UAV allocates to the vehicle; C3 indicates the delay constraint for the UAV to process computing tasks; C4 indicates that the amount of computing tasks uploaded by vehicle k to the UAV can be fully calculated on the UAV side; constraint C5 indicates that the computing tasks that vehicle k transmits to the UAV can be fully transmitted; C6 indicates the upper and lower limits of the vehicle's signal transmission power; C7 indicates the upper limit constraint on the energy consumption of the vehicle in processing perception data; and C8 is the phase shift adjustment range constraint for the intelligent reflective surface.

8. The method for allocating network resources of UAV-assisted edge computing based on intelligent reflective surface assistance according to claim 1, characterized in that: The process of optimizing the maximum average throughput function using continuous convex approximation and block landing algorithm includes: fixing task unloading, vehicle power generation and UAV computing resource allocation, converting the maximum average throughput function P3 into a sub-problem P3.1 about variable Θ according to the fixed task unloading, vehicle power generation and UAV computing resource allocation, solving the sub-problem to obtain the optimal phase shift matrix of the smart reflection surface; fixing the smart reflection surface phase shift matrix, converting the maximum average throughput function P3 into a sub-problem P3.2 about variables D, P, f according to the fixed smart reflection surface phase shift matrix, where D is task unloading, P is the transmission power of the vehicle terminal, and f is the computing resource allocation of the UAV; solving sub-problem P3.2 to obtain the results of task unloading, vehicle terminal transmission power and UAV computing resource allocation.

Images