A space-air-ground cooperative scheduling method for guaranteeing task priority and system throughput

Abstract

The application claims a space-air-ground cooperative scheduling method for guaranteeing task priority and system throughput, belonging to the technical field of wireless communication. Aiming at the problems of task resource mismatch and variable priority in space-air-ground emergency communication, a multi-layer task cooperative scheduling method is proposed. Aiming at the problem of variable task priority, a time-varying task priority is defined to quantify the dynamic benefits of the task according to the timeliness of disaster data and user priority. Aiming at the problem of task resource mismatch, the task priority service is realized and the transmission rate is maximized by cooperatively designing the trajectory of the unmanned aerial vehicle, user association scheduling and dynamic allocation of backhaul bandwidth. Aiming at the problem of limited computing capacity of edge servers, a particle swarm algorithm based on limited quadratic neighborhood search and a partial offloading strategy based on priority sorting are proposed to improve the system throughput. Finally, a task scheduling model is constructed based on the above problems, and the optimal space-air-ground cooperative scheduling scheme is obtained by using the block coordinate descent algorithm for alternating iteration.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology. Specifically, it relates to a space-air-ground coordinated scheduling method that ensures task priority and system throughput. Background Technology

[0002] Emergency communication networks generally refer to specialized communication networks constructed to ensure emergency response and necessary communication during major natural disasters or sudden emergencies by comprehensively utilizing various communication resources. With continuous technological advancements, emergency communication networks have evolved from fixed or mobile terrestrial emergency communication networks and dual-layer satellite-ground emergency communication networks to integrated space-air-ground emergency communication networks. Compared to traditional emergency communication networks, integrated space-air-ground emergency communication networks offer advantages such as comprehensive information service support capabilities, scalable and flexible networking capabilities, highly efficient and reliable disaster relief capabilities, and more efficient resource utilization.

[0003] The integrated air-space-ground emergency communication network architecture is characterized by diverse node types, heterogeneous network interconnection, complex spatiotemporal behavior, dynamic topology changes, numerous service types, and vastly different requirements, making problems such as low data transmission efficiency and difficulty in guaranteeing service quality particularly prominent. One of the key technologies for solving this problem is task scheduling technology. Currently, satellite task scheduling based on task resource matching, UAV task scheduling based on location deployment and path planning, computational offloading scheduling based on MEC, and dynamic task scheduling based on multi-objective decision-making and machine learning have been extensively studied. However, the trend of multi-task fusion, rapid response, and collaborative scheduling in air-space-ground emergency communication networks is becoming increasingly prominent. Traditional task scheduling schemes face problems such as task resource mismatch caused by node heterogeneity, variable task priorities due to the high real-time nature of disaster data, and reduced system throughput due to limited computing power of edge servers. Therefore, how to conduct multi-level, multi-resource, and multi-task collaborative scheduling under the integrated air-space-ground emergency communication network architecture, and design multi-layer collaborative task scheduling schemes to maximize network utility and resource utilization, has become a critical issue that urgently needs to be addressed.

[0004] To address these issues, this invention proposes a space-air-ground collaborative scheduling method that guarantees both task priority and system throughput. For the problem of variable task priorities, a time-varying task priority is defined based on the timeliness of disaster data and user priorities to quantify dynamic task benefits. For the problem of mismatched task resources, task priority service is achieved through collaborative design of UAV trajectories and user-related scheduling, and transmission rate is maximized by dynamically allocating backhaul bandwidth. To address the problem of reduced system throughput due to limited edge server computing power and excessive incoming data, a particle swarm optimization algorithm based on finite quadratic neighborhood search is proposed to optimize the uplink connection between satellites and high-altitude platforms, and a partial offloading strategy based on priority ranking is proposed to optimize offloading decision variables. Finally, a task scheduling model is constructed based on the above problems, and the optimal space-air-ground collaborative scheduling scheme is obtained by iteratively solving the block coordinate descent algorithm. Summary of the Invention

[0005] This invention aims to solve the problems of the prior art. It proposes a space-air-ground coordinated scheduling method that guarantees task priority and system throughput. The technical solution of this invention is as follows:

[0006] A space-air-ground coordinated scheduling method that guarantees task priority and system throughput includes the following steps:

[0007] S1: Construct a task model in the air-space-ground emergency communication network, characterize task requirements, and define a time-varying task priority to quantify the dynamic benefits of the task.

[0008] S2: Modeling the task scheduling process in air-space-ground emergency communications by coordinating UAV trajectory design, user association, high-altitude platform bandwidth allocation, uplink connection selection, and computational offload decisions;

[0009] S3: Calculate the transmission rate and system throughput of the communication links between nodes;

[0010] S4: Construct a high-altitude platform queue model and calculate the queue length and storage capacity;

[0011] S5: The task scheduling problem in the air-space-ground emergency communication network is constructed as a collaborative task scheduling model that maximizes the weighted sum of the product of task priority and system throughput by collaboratively designing UAV user association, trajectory design, high-altitude platform bandwidth allocation, optimizing uplink selection and computation offloading between high-altitude platforms and satellites.

[0012] S6: Applying the block coordinate descent method, the cooperative task scheduling model is decomposed into a data collection scheduling model that maximizes the priority weighted rate sum and a computation offloading scheduling model that maximizes the system throughput.

[0013] S7: For the data collection and scheduling model, a UAV user association scheme is designed using a linear relaxation combined with CVX solution; a UAV trajectory and high-altitude platform backhaul bandwidth allocation scheme is designed using a successive convex approximation of non-convex problem transformation method combined with CVX solution; and a model convergence acceleration algorithm is proposed.

[0014] S8: For the computational offloading scheduling model, a particle swarm algorithm based on finite quadratic neighborhood search is proposed to optimize the uplink connection between satellites and high-altitude platforms, and a partial offloading strategy based on priority ranking is proposed to optimize the offloading decision.

[0015] S9: Use the block coordinate descent algorithm to alternately solve the data collection scheduling model and the computational unloading scheduling model, and reconstruct infeasible solutions to generate the optimal air-space-ground coordinated scheduling scheme.

[0016] The advantages and beneficial effects of this invention are as follows:

[0017] This invention addresses the issues of task resource mismatch caused by node heterogeneity and variable task priorities due to the high real-time nature of disaster data in emergency air-space-ground communication. It proposes an air-space-ground collaborative scheduling method to ensure both task priority and system throughput. The main innovations of this invention are: 1) To address the variable task priority problem, a time-varying task priority is defined to quantify dynamic task benefits based on the timeliness of disaster data and user priorities; 2) To address the task resource mismatch problem, an air-space-ground collaborative scheduling method is proposed, utilizing a block coordinate descent method to collaboratively optimize UAV trajectory design, user association and high-altitude platform bandwidth allocation, uplink selection, and computational offloading scheduling; 3) To address the limited computing power of edge servers, a particle swarm optimization algorithm based on finite quadratic neighborhood search and a partial offloading strategy based on priority ranking are proposed to optimize computational offloading decisions. Existing research mostly sets task priorities to fixed values ​​without considering the variable task priorities caused by the high real-time nature of disaster data. Therefore, the time-varying task priority definition and calculation method proposed in this invention are not easily conceived by those skilled in the art. Furthermore, existing research in related fields, such as satellite mission scheduling and UAV scheduling, mostly focuses on single-scenario, single-variable task scheduling schemes, which struggle to effectively address the task resource mismatch problem during scheduling. Therefore, the air-space-ground collaborative scheduling method proposed in this invention possesses completeness and originality. Finally, this invention fully analyzes the time-varying nature of uplink connections between high-altitude platforms and LEO satellites, as well as the limitations of edge server computing power and the reduction in system throughput caused by excessive incoming data volume. It develops a particle swarm optimization algorithm based on finite quadratic neighborhood search to optimize the uplink transmission connection of the high-altitude platform and proposes a priority-based partial offloading strategy to optimize computational offloading decisions, thereby maximizing system throughput. In existing research, most studies utilize matching algorithms for uplink connection design and planning, which are less effective than particle swarm optimization algorithms that employ multiple neighborhood selections. Furthermore, using a full offloading mode in computational offloading scheduling leads to reduced resource utilization. Therefore, this invention possesses both creativity and feasibility in its solution. Attached Figure Description

[0018] Figure 1 This is a network model diagram constructed according to a preferred embodiment of the present invention;

[0019] Figure 2 This is a diagram of a high-altitude platform queue model constructed according to a preferred embodiment of the present invention;

[0020] Figure 3 This is a flowchart of a space-air-ground collaborative scheduling method for ensuring task priority and system throughput, as described in this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0022] The technical solution of the present invention to solve the above-mentioned technical problems is:

[0023] A space-air-ground collaborative scheduling method to ensure both task priority and system throughput. It collaboratively considers the data collection, computation, and transmission processes in emergency scenarios, and utilizes block coordinate descent technology to alternately optimize UAV trajectory design, user association, high-altitude platform bandwidth allocation, uplink selection, and computation offloading scheduling to maximize the weighted sum of the product of task priority and system throughput. The specific steps are as follows:

[0024] Step 1: Set scene parameters, including: number of nodes, 3D coordinates, number of regions; divide time slots according to the task scheduling cycle and capture scene parameters within any time slot; import task information, including required time windows, priority, etc.

[0025] Step 2: Model all variables to be optimized and their constraints in the scheduling process, calculate the transmission rate and system throughput between nodes, and construct the Cooperative Task Scheduling Model (CTSM).

[0026] Step 3: By applying the block coordinate descent method, the CTSM model is decomposed into two sub-models: the data collection scheduling model and the computation offloading scheduling model.

[0027] Step 4: For the data collection and scheduling model SP1, a linear relaxation method is used to process the binary variable a in the user association sub-model. i,j,k [n] Relaxation is performed as a continuous variable, thus transforming the non-convex constraint (C7) into a convex constraint that can be effectively solved by the CVX solver. For convex constraints in the trajectory optimization and bandwidth allocation sub-models, a successive convex approximation method is used to relax them.

[0028] Step 5: Apply the block coordinate descent algorithm to iteratively optimize the trajectory, bandwidth allocation, and user-related variables until the target value changes within the set threshold. To accelerate the convergence of the data collection and scheduling model, a circular coverage-based initial UAV trajectory design and a priority-ranking-based initial bandwidth allocation algorithm are adopted to obtain higher-quality initial trajectory and bandwidth allocation schemes.

[0029] Step 6: For the computational offloading scheduling model SP2, a particle swarm optimization algorithm based on finite quadratic neighborhood search is proposed to optimize the uplink transmission connection selection on the high-altitude platform, and a partial offloading strategy based on priority ranking is proposed to optimize the computational offloading decision.

[0030] Step 7: Apply the block coordinate descent method to collaboratively optimize UAV trajectory design, user association and high-altitude platform bandwidth allocation, uplink selection and computation offloading scheduling. Alternately solve the data collection scheduling model and the computation offloading scheduling model until the target value change is within the set threshold. Then verify and reconstruct the feasible solution to generate the optimal air-ground collaborative scheduling scheme.

[0031] The main symbols and parameters involved in this invention and their meanings are listed in Table 1.

[0032] Table 1. Main Symbols, Parameters, and Their Meanings

[0033]

[0034]

[0035] Preferably, in the second step, let the trajectory variable Q = {q} a,i,k [n]}, bandwidth allocation variable α={α i,k [n]}, User-related variable A={a i,j,k [n]}, uplink connection variable B = {b k,l [n]}, calculate the unloading scheduling variable C = {c k,l [n]}. Using the weighted sum of the product of priority and system throughput as the objective function, the cooperative task scheduling model can be constructed as follows:

[0036]

[0037] Here, the objective function, together with (C1), represents maximizing the minimum average realizable weighted rate of the system, aiming to find the most conservative solution across the entire constraint space. (C2) indicates that the starting and ending positions of any UAV are the same within the scheduling period T, to provide periodic data acquisition services. (C3) and (C4) represent the maximum speed V of the UAV, respectively. max and minimum collision avoidance distance d min The constraints are as follows: (C5)-(C7) represent user association constraints for UAVs, ensuring that each UAV can serve at most one rescue device simultaneously in any time slot, and each rescue device can be served by at most one UAV simultaneously. (C8) and (C9) are bandwidth allocation constraints, ensuring that the bandwidth allocated by the high-altitude platform to each user is not less than zero and the sum of the bandwidth allocated to all users does not exceed the total available bandwidth. (C10)-(C12) represent connection constraints between the high-altitude platform and LEO satellites, ensuring that the high-altitude platform transmits data to at most one LEO satellite in any time slot, and one LEO satellite can connect to multiple LEO satellites. H There are three high-altitude platforms. (C13) represents the storage capacity constraint for any high-altitude platform. (C14) represents the constraint for calculating the unloading decision variable.

[0038] Preferably, in the third step, a data collection scheduling model and a computation offloading scheduling model are constructed respectively:

[0039]

[0040]

[0041] Preferably, in the fourth step, different theorems are applied to relax the original function according to the actual situation, thereby transforming the non-convex problem into a convex problem. Theorem 1 is used to obtain the convex lower bound (or upper bound) of a convex (or concave) function, and Theorems 2 and 3 are used to obtain the convex upper bound of a non-convex function:

[0042] Theorem 1: For a convex function f(x), its lower bound can be obtained by performing a first-order Tate expansion:

[0043]

[0044] Here, x0 is any point on the domain of f(x) if and only if x = x0, then f(x) = f(x0).

[0045] Theorem 2: Let q and s are three-dimensional coordinate vectors. If is a known positive constant, then its Hessian matrix is... It has the following properties:

[0046]

[0047] Where I is the identity matrix.

[0048] Theorem 3: For a function f(x), if Then, according to its second-order Tate expansion at a given point x0, we can obtain:

[0049]

[0050] Preferably, in the fifth step, the initial UAV trajectory design algorithm based on circular coverage first clusters the user nodes in the network, and then finds the initial trajectory with the shortest flight distance according to the maximum user coverage criterion; the initial bandwidth allocation algorithm based on priority ranking uses the ratio of the priority of user equipment and UAV to the total priority of all nodes as the bandwidth allocation ratio of the high-altitude platform, so as to transmit more and higher priority data.

[0051] Preferably, in the sixth step, the algorithm first randomly generates m particles B. i (i = 1, 2, ..., m), and calculate the fitness function value f(B) for each particle. i Next, in the first neighborhood operation, a neighborhood solution is searched for each particle by exchanging matrix elements. And the best solution set is selected to replace the original solution set, while updating the neighborhood search count; then, in the second neighborhood operation, for the current solution set {B i Choose any two solutions B p B q Perform a neighborhood search, selecting the best solution and repeating it U times; after the neighborhood search is completed, select the neighborhood set G for each solution. i The neighborhood solution with the largest fitness function value Replace the original solution; check the neighborhood search count of each solution. If it is greater than the threshold, discard it; otherwise, update the local optimum; continue until the maximum number of iterations is reached, and then output the global optimum.

[0052] The partial unloading strategy is optimized based on priority ranking to calculate the unloading decision. First, in each queue... Data packets are sorted by priority ρ i Sort the data according to the amount of data contained in the high-altitude platform queue. Not exceeded the calculation threshold D th If the data exceeds the limit, the calculation is performed locally. Otherwise, the excess data is offloaded to LEO satellites for processing in sequence, while the data within the limit is still calculated locally.

[0053] Preferably, in the seventh step, to ensure the feasibility of the solution, the possible non-binary solutions are reconstructed. Each time slot of length δ is further divided into τ sub-time slots, i.e., the total number of sub-time slots is N′=τN. Taking the drone user association problem as an example, the number of sub-time slots allocated to the rescue equipment in time slot n for the drone is... Take the closest integer. And as τ increases, N... i,j,k [n] will become a standard binary feasible solution and satisfy all constraints.

[0054] The model involved in this invention is as follows:

[0055] 1. Network Model

[0056] The primary application of this invention is an integrated air-space-ground emergency communication network, such as... Figure 1As shown in the diagram. In this network architecture, communication is disrupted due to complete destruction of infrastructure in the disaster area. The network model consists of three layers: the satellite layer, the airborne layer, and the ground user layer. The ground user layer is divided into two user groups based on different needs: user equipment that needs to communicate with the outside world and rescue equipment that needs to transmit data collected during rescue missions back to the ground command center to assist in rescue decision-making. The airborne layer deploys multiple drones at low altitudes to reach various rescue mission points and collect disaster data stored in rescue equipment. At high altitudes, a high-altitude platform provides continuous and stable communication access covering the entire sub-area for user equipment and drones, serving as a relay center for data transmission in the emergency communication network. At the satellite layer, LEO satellites are deployed to provide backhaul connectivity to the high-altitude platform. These satellites are equipped with multiple transceiver antennas operating in the Ka-band and are connected to ground stations via feeder links, providing access to the dedicated ground emergency cellular network, thereby transmitting data back to the ground emergency command and dispatch center. In addition, the satellites and high-altitude platforms in this network model are equipped with edge computing servers with different computing capabilities, which can preprocess emergency business data before transmitting it back.

[0057] 2. High-altitude platform queue model

[0058] The queue set of the high-altitude platform is For any high-altitude platform k, calculate the queue length at the start of time slot n+1. It can be represented as:

[0059]

[0060] in, This represents the amount of data received by the high-altitude platform k within time slot n, where δ is the length of a single time slot. This represents the amount of data diverted after the offloading decision is calculated. (Transmission queue) The length can be expressed as:

[0061]

[0062] in, This indicates the amount of data that the computation queue distributes to the transmission queue after calculating the offloading decision. This represents the amount of data transmitted from the high-altitude platform k to the LEO satellite within time slot n.

[0063] To avoid data backlog on the high-altitude platform that could lead to insufficient storage capacity at any given time slot, the length of the system queue in high-altitude platform k should not exceed its storage capacity C within any time slot. k ,Right now:

[0064]

[0065] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions.

[0066] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0067] The above embodiments should be understood as illustrative only and not as limiting the scope of protection of the present invention. After reading the description of the present invention, those skilled in the art can make various alterations or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A space-air-ground coordinated scheduling method that guarantees task priority and system throughput, characterized in that, Includes the following steps: S1: Construct a task model in the air-space-ground emergency communication network and define a time-varying task priority to quantify the dynamic benefits of the task. S2: Modeling the task scheduling process in air-space-ground emergency communications by coordinating UAV trajectory design, user association, high-altitude platform bandwidth allocation, uplink connection selection, and computational offload decisions; S3: Calculate the transmission rate and system throughput of the communication links between nodes; S4: Construct a high-altitude platform queue model and calculate the queue length and storage capacity; S5: The task scheduling problem in the air-space-ground emergency communication network is constructed as a collaborative task scheduling model that maximizes the weighted sum of the product of task priority and system throughput by collaboratively designing UAV user association, trajectory design, high-altitude platform bandwidth allocation, optimizing uplink selection and computation offloading between high-altitude platforms and satellites. S6: Applying the block coordinate descent method, the cooperative task scheduling model is decomposed into a data collection scheduling model that maximizes the priority weighted rate sum and a computation offloading scheduling model that maximizes the system throughput. S7: For the data collection and scheduling model, a UAV user association scheme is designed using a linear relaxation combined with CVX solution; a UAV trajectory and high-altitude platform backhaul bandwidth allocation scheme is designed using a successive convex approximation of non-convex problem transformation method combined with CVX solution; and a model convergence acceleration algorithm is proposed. S8: For the computational offloading scheduling model, a particle swarm algorithm based on finite quadratic neighborhood search is proposed to optimize the uplink connection between satellites and high-altitude platforms, and a partial offloading strategy based on priority ranking is proposed to optimize the offloading decision. S9: Use the block coordinate descent algorithm to alternately solve the data collection scheduling model and the computational unloading scheduling model, and reconstruct infeasible solutions to generate the optimal air-space-ground coordinated scheduling scheme.

2. The air-space-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, In step S1, a time-varying task priority is defined to quantify the dynamic benefits of a task, as shown below: The task model is modeled as a four-tuple to describe the task requirements, namely... ; The types of users who generate tasks include four categories: emergency and routine tasks, respectively, originating from user devices and rescue equipment in emergency scenarios. Refers to the task Required service time window; Refers to the task Priority; Refers to the task The amount of data; Set the user's initial priority; In time slot The data timeliness weighting factor for the internal request task. In time slot The horizontal coordinates of the disaster center within the area For time slots The distance between the user and the disaster center.

3. The air-space-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, Step S2 specifically includes the following steps: (1) Modeling of UAV trajectory variables; Within any time slot, the UAV's position can be considered stationary; therefore, the UAV trajectory is represented by N two-dimensional sequences. To approximate this, N is the number of time slots. and Representing sub-regions Chinese drones The horizontal and vertical coordinates; (2) Modeling of user-related variables for unmanned aerial vehicles (UAVs); In a task cycle All drones and ground rescue equipment share the same frequency band for communication, and each drone provides services to its associated users via time-division multiple access; define a binary variable. If in a time slot Inside, drones To rescue equipment Providing data collection services, ,otherwise ; (3) Modeling of bandwidth allocation variables for high-altitude platforms; The high-altitude platform uses frequency division multiple access (FDMA) to dynamically allocate backhaul bandwidth among all ground user equipment and UAVs in the area, meaning the high-altitude platform receives data from multiple ground user equipment and UAVs in a single time slot; a continuous variable is defined. Indicates in time slot Time allocation to ground user equipment and drones The bandwidth ratio; (4) Uplink transmission connection variable modeling; For uplink data transmission between the high-altitude platform and the LEO satellite, orthogonal frequency division multiplexing (OFDM) technology is used to divide the frequency bandwidth of the "high-altitude platform-satellite" uplink into a set of orthogonal sub-channels; different high-altitude platforms connect to the same LEO satellite using different orthogonal sub-channels in the same time slot; a binary variable is defined. Indicates high-altitude platform and LEO satellite The time-varying connection relationship between them; if in time slots Inside, high-altitude platform and LEO satellite If a data transmission link exists, then ,otherwise ; (5) Model the unloading scheduling variables; Design an unloading decision binary variable To describe in time slot Internal high-altitude platform Locally or offloaded to LEO satellite The spaceborne edge computing server computes task data; when the high-altitude platform is in a time slot Calculate queue length Greater than the calculation threshold The excess data will then be offloaded to the LEO satellite edge server for processing. ,otherwise As shown in the following formula: 。 4. The air-space-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, Step S3 specifically includes the following steps: (1) Calculate the transmission rate during the data collection phase; For any sub-region within the disaster area, the ground user layer has individual user devices and Each rescue device transmits its data to the high-altitude platform and the drone in the airspace layer. The channel power gain between the high-altitude platform and the user equipment can be expressed as: in, For reference distance Channel power at that time Indicates in time slot User equipment and the high-altitude platform in this area The distance between them; the channel power gain between the drone and the high-altitude platform, and between the drone and the rescue equipment, are calculated as follows: and ; On the high-altitude platform In the middle, when rescue equipment To drones When transmitting data, the signal-to-noise ratio (SNR) at the receiving end can be expressed as: in, Specify the transmission power for rescue equipment. Indicates time slot Co-channel interference caused by all other drones transmitting data within the same channel. This refers to the AWGN power at the UAV receiver; therefore, in the time slot... Internal drones The achievable data reception rate is: use Indicates high-altitude platform The total available bandwidth, and assuming ground user equipment With the specified transmission power drones To specify the transmission power Upload its data to the high-altitude platform; therefore, when and When transmitting data to an aerial platform, the instantaneous achievable data rate is expressed as follows: The signal-to-noise ratio received at the high-altitude platform is expressed as follows: , , The power spectral density of AWGN; (2) Calculate the data offloading phase transmission rate; high-altitude platform and LEO satellite The channel fading factor is modeled as a circularly symmetric complex Gaussian random variable, i.e. ,in, , Each high-altitude platform is assigned a different sub-channel with a Ka-band bandwidth of W for accessing LEO satellites; therefore, in time slots inner subregion The achievable data rate during data transmission is: in, For high-altitude platforms The specified transmission power The square of the channel gain of the "High Altitude Platform-LEO Satellite" is... For high-altitude platforms and LEO satellite The distance between them This is the path loss index. The AWGN power at the satellite receiver; and These represent the altitudes of the LEO satellite and the high-altitude platform, respectively. and These represent the horizontal coordinates of the LEO satellite and the high-altitude platform, respectively. By integrating uplink selection variables and offloading decisions, the high-altitude platform In the time slot The computation rate at that location is: in, For the local computing speed of the edge server on the high-altitude platform, These respectively indicate deployment on high-altitude platforms The CPU cycle frequency of the edge computing server and the number of CPU cycles required to compute a unit bit of data.

5. The air-space-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, Step S4 specifically includes: representing the queue set of all high-altitude platforms as... For any subregion In the time slot At the beginning, calculate the length of the queue. It can be represented as: in, Indicates high-altitude platform In the time slot The amount of data received internally. The length of a single time slot. This represents the amount of data diverted after calculating the offloading decision; transmission queue. The length can be expressed as: in, This indicates the amount of data that the computation queue distributes to the transmission queue after calculating the offloading decision. Indicates in time slot Internal high-altitude platform The amount of data transmitted to LEO satellites; High-altitude platform within any time slot The length of a system queue should not exceed its storage capacity. ,Right now: 。 6. The air-space-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, The weighted sum of the product of task priority and transmission rate in the data collection phase of step S5 can be expressed as: The system throughput during the computational unloading phase can be expressed as: make , , , , The collaborative task scheduling model in step S5 can then be described as follows: Here, the objective function, together with (C1), represents maximizing the minimum average realizable weighted rate of the system, with the aim of finding the most conservative solution in the entire constraint space; (C2) represents the time required for any UAV to complete the task scheduling cycle. The starting and ending positions are the same to provide periodic data acquisition services; (C3) and (C4) are the maximum speeds of the UAV, respectively. and minimum collision avoidance distance The constraints are as follows: (C5)-(C7) represent user association constraints for UAVs, ensuring that each UAV can serve at most one rescue device at any given time slot, and each rescue device can be served by at most one UAV at any given time slot; (C8) and (C9) are bandwidth allocation constraints, ensuring that the bandwidth allocated to each user by the high-altitude platform is not less than zero and the sum of the bandwidth allocated to all users does not exceed the total available bandwidth; (C10)-(C12) represent connection constraints between the high-altitude platform and LEO satellites, ensuring that the high-altitude platform transmits data to at most one LEO satellite at any given time slot, and one LEO satellite can connect to multiple LEO satellites. (C13) represents the storage capacity constraint of any high-altitude platform; (C14) represents the constraint for calculating the unloading decision variable.

7. The air-space-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, Step S6 decomposes the original Cooperative Task Scheduling Optimization Model (CTSM) into two sub-models by applying the block coordinate descent method: a data collection scheduling model and a computational unloading scheduling model. The former optimizes the bandwidth allocation of the high-altitude platform, as well as the association of UAV users and trajectory planning during the data collection phase by fixing the uplink selection variables and computational unloading decision variables of the high-altitude platform. The latter optimizes the uplink selection variables and calculates the unloading decision variables of the high-altitude platform during the data backhaul stage by allocating bandwidth to a fixed high-altitude platform, as well as associating drone users and planning trajectories.

8. A space-air-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 6, characterized in that, In step S7, the data collection and scheduling model uses a linear relaxation method to handle the binary variables in the user association submodel. Relaxing the non-convex constraint (C7) into a convex constraint that the CVX solver can solve by making it a continuous variable; for the convex constraints in the trajectory optimization and bandwidth allocation sub-models, a successive convex approximation method is used to relax them, that is, in each iteration, the original function is approximated with a more tractable function at a given local point; different theorems are applied to relax the original function according to the actual situation; In step S7, a circular coverage-based initial UAV trajectory design and a priority-based initial bandwidth allocation algorithm are used to obtain a higher quality initial trajectory and bandwidth allocation scheme. The initial UAV trajectory design algorithm based on circular coverage first clusters user nodes in the network, and then finds the initial trajectory with the shortest flight distance according to the maximum user coverage criterion. The priority-based initial bandwidth allocation algorithm uses the ratio of the priority of user equipment and UAVs to the total priority of all nodes as the bandwidth allocation ratio for the high-altitude platform, so as to transmit more data with higher priority.

9. A space-air-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 1, characterized in that, In step S8, a particle swarm optimization algorithm based on finite quadratic neighborhood search is proposed to select the link with the highest uplink rate for each high-altitude platform. The fitness function is: The algorithm first generates random... Particles And calculate the fitness function value for each particle. Next, in the first neighborhood operation, a neighborhood solution is searched for each particle by exchanging matrix elements. Then, in the second neighborhood operation, the original solution set is replaced with the best one, and the neighborhood search count is updated; then, for the current solution set... Choose any two solutions , Perform a neighborhood search, selecting the best solution and repeating it U times; after the neighborhood search is complete, select the neighborhood set of each solution. The neighborhood solution with the largest fitness function value Replace the original solution; check the neighborhood search count of each solution. If it is greater than the threshold, discard it; otherwise, update the local optimum; continue until the maximum number of iterations is reached, and then output the global optimum. Step S8 proposes a partial unloading strategy based on priority sorting to optimize the calculation of unloading decisions. First, in each queue... Prioritize data packets Sort the data according to the amount of data contained in the high-altitude platform queue. Not exceeded the calculation threshold If the data exceeds the limit, the calculation is performed locally; otherwise, the excess data is offloaded to the LEO satellite for processing in sequence, while the data within the limit is still calculated locally.

10. A space-air-ground coordinated scheduling method for ensuring task priority and system throughput according to claim 6, characterized in that, Step S9 specifically includes: Given a fixed uplink selection and computational offload scheduling decision (B, C), a block coordinate descent algorithm is used to optimize high-altitude platform bandwidth allocation, UAV user association, and trajectory planning. (A, Q); then get ( In the case of (A, Q), an improved particle swarm optimization algorithm and a priority-based partial unloading strategy are used to solve (B, C); the two processes are executed alternately until the objective function value of CTSM changes within a specified threshold. In step S9, to ensure the feasibility of the solution, the possible non-binary solutions are reconstructed; each solution of length is... The time slots are further divided into There are 10 sub-time slots, meaning the total number of sub-time slots is 100. .

Images