Task success transmission rate-oriented low-altitude unmanned aerial vehicle cluster network resource allocation method

Abstract

The invention provides a task success transmission rate-oriented low-altitude unmanned aerial vehicle cluster network resource allocation method, and relates to the technical field of low-altitude communication, and the method comprises the steps: building a low-altitude communication network system model based on an unmanned aerial vehicle cluster; constructing an unmanned aerial vehicle cluster network resource allocation optimization equation based on the low-altitude communication network system model; based on the unmanned aerial vehicle cluster network resource allocation optimization equation, performing analysis by using an unmanned aerial vehicle cluster frequency block sharing method to obtain a frequency block sharing scheme of the unmanned aerial vehicle clusters, and determining the number of sub-channels contained in a frequency block occupied by each unmanned aerial vehicle cluster by using a dynamic frequency block competition algorithm to obtain a resource allocation scheme between the unmanned aerial vehicle clusters; and optimizing the resource allocation scheme in the cluster by using a multi-priority dynamic resource allocation algorithm and taking maximization of the differentiation task successful transmission probability as a target to obtain a network resource allocation result. The problem that reliable transmission of low-altitude key tasks and efficient utilization of resources are difficult to ensure in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of low-altitude communication technology, and in particular to a method for allocating resources in a low-altitude unmanned aerial vehicle (UAV) swarm network based on mission success transmission rate. Background Technology

[0002] Unmanned aerial vehicle (UAV) swarm networks, as a core component of low-altitude communication networks, have attracted significant attention due to their advantages such as flexible deployment, high mobility, and low cost. UAV swarm networks can collaboratively operate and efficiently execute complex tasks, thus providing strong support for IoT applications such as intelligent transportation, environmental monitoring, and emergency rescue, and have received widespread attention from industry and academia. In various IoT applications, how to efficiently optimize the allocation of key resources such as spectrum and power under resource-constrained conditions in UAV networks, ensuring the reliability and fairness of mission transmission, while simultaneously considering overall network performance and quality of service, is a core problem that urgently needs to be solved. Existing research on UAV networks mainly focuses on non-UAV swarm scenarios and UAV swarm scenarios. In non-UAV swarm scenarios, existing research typically optimizes the allocation of spectrum and power resources to improve system throughput, reduce communication latency, expand coverage, and enhance overall network efficiency. Related methods encompass traditional optimization techniques, heuristic algorithms, and machine learning methods to adapt to the resource scheduling needs of different application scenarios. In UAV swarm scenarios, research focuses more on collaborative control, formation flight, and resource allocation, aiming to further reduce energy consumption and latency and improve system performance through collaborative optimization of resource configuration within the swarm.

[0003] While drones have been widely used in low-altitude communication services within the emerging low-altitude economy, drone swarm networks still face several key challenges. First, due to limited resources and a large number of nodes in a drone swarm, traditional communication methods (such as OFDMA or TDMA) often lead to a significant decrease in the available resources for individual drones, while also causing severe inter-node interference. Second, different application scenarios have significantly different requirements for task transmission, and drone swarms exhibit obvious heterogeneity, making it difficult to adapt "one-size-fits-all" resource allocation, task scheduling, or network optimization strategies to actual needs. Furthermore, comprehensively considering task reliability and fairness to ensure overall network performance is also a pressing issue. Therefore, designing a resource allocation strategy for drone swarm networks that can ensure reliable transmission of critical tasks while achieving efficient resource utilization is particularly important. Summary of the Invention

[0004] To address the aforementioned shortcomings in existing technologies, the low-altitude UAV swarm network resource allocation method based on mission success transmission rate provided by this invention solves the problem that existing technologies struggle to ensure reliable transmission of critical low-altitude missions and achieve efficient resource utilization.

[0005] To achieve the aforementioned objectives, the present invention employs the following technical solution: a low-altitude UAV swarm network resource allocation method oriented towards mission success transmission rate, comprising: S1: Establish a low-altitude communication network system model based on low-altitude UAV swarms; S2: Based on the low-altitude communication network system model, construct the optimization equation for the allocation of network resources in the UAV swarm; S3: Based on the optimization equation for network resource allocation of UAV swarms, the frequency block sharing method of UAV swarms is used for analysis to obtain the frequency block sharing scheme of UAV swarms; S4: Based on the frequency block sharing scheme, the dynamic frequency block contention algorithm is used to maximize the fairness of resource allocation among clusters. The number of sub-channels contained in the frequency block occupied by each UAV cluster is determined, and the resource allocation scheme among UAV clusters is obtained. S5: Utilizing a multi-priority dynamic resource allocation algorithm, with the goal of maximizing the success probability of differentiated task transmission, the resource allocation schemes for sub-channels and power of UAVs within the cluster are optimized. By integrating the resource allocation schemes among UAV clusters, the network resource allocation results are obtained, thus completing the network resource allocation for the low-altitude UAV cluster.

[0006] The beneficial effects of the present invention are as follows: The present invention provides a low-altitude UAV swarm network resource allocation method oriented towards mission success transmission rate. Through frequency block sharing strategy, inter-cluster resource allocation and intra-cluster resource allocation method, the resource utilization rate and differentiated service quality of UAV swarm network are optimized in a coordinated manner. (1) By deriving the conditions for interference-free sharing of frequency blocks between different clusters in the UAV swarm network, the sharing of spectrum resources between different UAV clusters is realized, thereby providing an efficient and stable spectrum management scheme for large-scale UAV access and improving network resource utilization and communication performance; (2) The deep reinforcement learning strategy is used to adaptively solve the problem of spectrum resource allocation decision-making between UAV network clusters under unknown environmental conditions. This method can adapt well to the differences between different UAV clusters and make targeted resource allocation decisions, thereby ensuring the fairness of resource allocation between UAV clusters; (3) A heuristic method is used to solve the power and sub-channel allocation problem within the cluster. By utilizing the advantages of fast algorithm convergence speed, low computational complexity and strong global search capability, the traditional method is avoided from being trapped in local optima, thereby maximizing the probability of mission success transmission and the efficient utilization of resources within the cluster.

[0007] Further, S1 includes: A low-altitude communication network consisting of multiple UAV swarms is established. Each swarm consists of a leader UAV equipped with multiple antennas and multiple member UAVs equipped with one antenna. The swarm adopts a star topology. The main controller is deployed at the base station, and the auxiliary controller is deployed at the leader UAV. The member UAVs transmit data to the leader UAV via the uplink. The leader UAV aggregates the data and then forwards it to the base station for processing. A network model is established. A line-of-sight channel model between UAVs is established, and the channel gain between member UAVs and the leading UAV is calculated. The continuous time axis is divided into multiple discrete time slots of equal length. The spectrum resources of the UAV swarm network are divided into multiple identical sub-channels. Power levels are defined, and communication sub-channel occupancy and power level selection indicator variables are provided. The communication rate of the UAVs is calculated using Shannon's formula. Based on the calculated channel gain, discrete time slots, sub-channels, indicator variables, and communication rate, the communication model is obtained. Establish the task data volume, task requirement delay, and task priority; calculate the actual transmission delay of the task based on the communication rate; define the indicator variable for successful task transmission; and establish the task model. Among them, the network model, communication model, and task model belong to the low-altitude communication network system model.

[0008] By accurately constructing network models, communication models, and multi-priority task models for UAV swarms, a theoretical basis is provided for subsequent task scheduling and resource allocation of UAVs in low-altitude areas, ensuring the efficient operation of the system under different environments and requirements.

[0009] Further, S2 includes: Using the inter-cluster frequency block partitioning method, all spectrum resources in the low-altitude communication network system model are divided into non-overlapping frequency blocks, and the partitioning and usage of frequency blocks are obtained. Each frequency block is used as the smallest unit of spectrum used by the cluster. Based on the frequency block partitioning and usage, and utilizing the cluster utility function and resource constraints, a network resource allocation optimization equation for UAV clusters is constructed, which includes optimization equations within the UAV cluster, optimization equations between UAV clusters, and global optimization equations.

[0010] To address the issues of limited low-altitude network resources and significant interference, a novel spectrum resource allocation method is proposed, and a corresponding optimization model is constructed. By refining the spectrum resource allocation and dynamically adjusting the transmission power, a model foundation is provided for effectively improving the success probability of differentiated missions and the utilization rate of spectrum resources.

[0011] Furthermore, the expression for the division and usage of the frequency blocks is as follows: ; ; ; ; ; ; Where Q represents the total number of frequency blocks. Represents a cluster Frequency blocks Indicator variables for usage, Indicates the first One cluster, Represents a cluster set. This indicates the cluster corresponding to the maximum value. Represents a cluster Number of sub-channels used Indicates the first The number of sub-channels contained in each frequency block Indicates the occupation of the first A cluster set of frequency blocks, Represents the first in the network Each frequency block, Represents the first in the network Each frequency block, This represents the first frequency block in the network. This represents the second frequency block in the network. This represents the Q-th frequency block in the network. Represents the set of sub-channels in the network. This represents the a-th sub-channel in the network. Indicates the first in the network Sub-channels, Indicates the number of sub-channels.

[0012] Furthermore, the expression for the cluster utility function is: ; ; The expression for the optimization equation within the drone swarm is: ; ; ; ; ; ; The expression for the optimization equation among drone swarms is: ; ; ; ; ; ; ; ; ; ; The expression for the global optimization equation is: ; ; in, Indicates the first The utility function of a cluster. This represents the set of task priorities, where m represents the UAV task priority index. Indicates the first The weight corresponding to the priority of each task This indicates that the priority within a cluster is The ratio of the number of successfully transmitted tasks to the total number of tasks of that priority. Represents the set of weights. This indicates the weight of task with priority 1. This indicates the priority 2 task weight. Indicates priority Task weight, Indicates each priority as Task weights This represents the function that takes the maximum value, where k represents the sub-channel index. Represents a cluster The set of sub-channels occupied (t) represents the cluster The first Are the individual member drones in the time slot? Occupy sub-channel , Indicates time slot , This indicates the preset SINR threshold. Represents a cluster The set of the i-th power indicator variables, This indicates that corresponding elements within a set are multiplied and added together. Represents a set of power levels. This represents the maximum allowable power value, and j represents the index of the power indicator variable. Represents a set of power levels. (t) represents the cluster The first Are the individual member drones in the time slot? Use power rating , This represents the expectation operation. Indicates the occupied frequency block The corresponding cluster set, Let q represent the frequency block index and Q represent the number of frequency blocks, where Q represents the utility function. Indicates the first individual frequency blocks The corresponding set of sub-channels, Represents the set of sub-channels. Indicates the first Each frequency block, Indicates the first Each frequency block, Indicates the first frequency block. Indicates the second frequency block. This represents the Q-th frequency block. Indicates the first individual frequency blocks The number of sub-channels included. Represents a set index. Indicates the frequency band used A collection of clusters, Represents a cluster set. Represents a cluster Does it occupy a frequency block? , Indicates multiplication. This represents the shortest distance between the member drone of the i-th cluster and the member drone of the l-th cluster within cluster c. Indicates the threshold of no interference. Indicates a pair with Related numbers, J represents the transmit power at power level J, and n represents the number of clusters sharing the frequency band. This indicates the transmit power at power level 1. This represents the farthest distance between the member drone of the i-th cluster within cluster c and the leader drone of that cluster. This represents the function that takes the minimum value.

[0013] Further, S3 includes: Based on the resource allocation optimization equation of UAV swarm network, each swarm in the UAV swarm network is abstracted as a node; Determine whether the inter-cluster frequency block sharing interference-free condition is met between any two nodes. If not, add an edge between the nodes to obtain the network topology and form an undirected graph G. Perform the traditional graph coloring method to color each node in the undirected graph G to obtain all solutions; Iterate through all solutions. If there is a solution where one color is occupied by more than three nodes, then check again whether it meets the no-interference condition. If it does not meet the condition, then move one of the nodes of that color to the cache set. Iterate through the nodes in the cache set, reassign them colors, and remove the node when it meets the no-interference condition, until the cache set is empty, and update all feasible solutions. The optimal solution is the one with the smallest variance in the number of nodes contained in all colors among all feasible solutions, thus obtaining the frequency block sharing scheme for the drone swarm.

[0014] By abstracting multiple drone clusters as nodes and inter-cluster interference as edges, an undirected graph is constructed, thereby transforming inter-cluster interference into a graph coloring problem. An improved coloring graph scheme is then used to achieve frequency block sharing between clusters, solving the problems of limited resources and complex interference in low-altitude networks.

[0015] Furthermore, the expression for the interference-free condition is: ; ; in, Indicates multiplication. This represents the shortest distance between the member drone of the i-th cluster and the member drone of the l-th cluster within cluster c. Indicates the threshold of no interference. Indicates a pair with Related numbers, This indicates the transmit power at power level J. This indicates the transmit power at power level 1. This represents the farthest distance between the member drone of the i-th cluster within cluster c and the leader drone of that cluster. Represents the i-th cluster, Indicates the number of clusters sharing the frequency band. Indicates the signal wavelength.

[0016] An interference-free expression for sharing frequency blocks between clusters is provided. By inputting the maximum and minimum system power levels, the minimum distance between member UAVs in the cluster, and the maximum distance between the cluster leader and member UAVs, the cluster distance condition for interference-free sharing of frequency blocks is obtained, providing a theoretical basis for frequency block sharing between clusters.

[0017] Further, S4 includes: Based on the frequency block sharing scheme, the frequency blocks occupied by each cluster are obtained; Construct the initial primary Q-network and secondary Q-network; By initializing the state space, the Q-values ​​of all current actions are calculated from the Q-network, based on... The greedy algorithm selects the optimal action to obtain the number of sub-channels contained in each frequency block at the current time; Based on the number of sub-channels contained in each frequency block, the corresponding cluster resource allocation strategy is invoked to complete the power and sub-channel allocation within the cluster, thereby obtaining the current status and immediate reward. Based on the current state and immediate reward, the main Q-network and the secondary Q-network are trained iteratively to obtain the trained main Q-network and secondary Q-network. By initializing the state, the trained main Q network and secondary Q network are tested. Each time slot task arrives dynamically. Based on the current state, the success transmission probability of each cluster is obtained, and then the utility function of each cluster is calculated to obtain the reward function at the current moment. Based on the network, the current action and the state at the next moment are obtained. Execute the action, complete the resource allocation for the current time slot, and enter the next time slot. If the reward is less than the threshold, reset the state, determine the number of sub-channels contained in the frequency block occupied by each drone cluster, and obtain the resource allocation scheme between drone clusters.

[0018] Furthermore, the expression for the reward function is: ; in, Represents the reward function, Represents a piecewise function. This represents the expectation operation. Indicates the occupied frequency block A collection of clusters, Represents the utility function.

[0019] Further, S5 includes: Determine the population size, number of member drones, power level, and number of iterations. Based on the resource allocation scheme, determine the number of sub-channels in the cluster to obtain the initial population. Based on the initial population, with the goal of maximizing the probability of successful transmission of differentiated tasks, the first population is generated by calculating the fitness function value of the population and selecting a certain number of chromosomes from the population according to the selection probability. Perform a crossover operation on the first group according to the crossover probability to obtain the second group; The second population is mutated according to the mutation probability to obtain the third population; Calculate the fitness function of each chromosome in the third population, compare the fitness function values ​​of chromosomes in the initial population with those in the third population, retain the chromosomes with higher fitness, and obtain the updated initial population. Based on the updated population, the chromosome corresponding to the maximum fitness function is extracted to obtain the power and sub-channel allocation results within the cluster. By combining the resource allocation schemes among the UAV clusters, the network resource allocation results are obtained, and the network resource allocation within the low-altitude UAV cluster is completed. Attached Figure Description

[0020] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein: Figure 1 This is an exemplary flowchart of a low-altitude unmanned aerial vehicle (UAV) swarm network resource allocation method oriented towards mission success transmission rate, as shown in some embodiments of this specification. Figure 2 This is an exemplary schematic diagram of a low-altitude unmanned aerial vehicle (UAV) swarm network system model according to some embodiments of this specification; Figure 3 This is an exemplary schematic diagram of a low-altitude unmanned aerial vehicle (UAV) swarm network resource allocation method oriented towards task success transmission rate, as shown in some embodiments of this specification. Figure 4 This is an exemplary schematic diagram illustrating, according to some embodiments of this specification, the conditions under which a drone swarm uses the same frequency block without interference. Detailed Implementation

[0021] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0022] Example Figure 1 This is an exemplary flowchart illustrating a low-altitude unmanned aerial vehicle (UAV) swarm network resource allocation method oriented towards mission success transmission rate, according to some embodiments of this specification. Figure 1 and Figure 3 As shown, the process includes the following steps. In some embodiments, the process may be executed by a processor.

[0023] S1: Establish a low-altitude communication network system model based on low-altitude UAV swarms.

[0024] In some embodiments, the processor can establish a low-altitude communication network system model based on a multi-UAV swarm, including a network model, a communication model, and a mission model.

[0025] In some embodiments, such as Figure 2 As shown, the processor can establish a low-altitude communication network composed of multiple UAV swarms. Each swarm includes a leader UAV equipped with multiple antennas and multiple member UAVs equipped with one antenna each. The swarm adopts a star topology. The main controller is deployed at the base station, and the auxiliary controller is deployed at the leader UAV. Member UAVs transmit data to the leader UAV via uplink, where the data is aggregated and then forwarded to the base station for processing. A network model is established; a line-of-sight channel model between UAVs is established, and the channel gain between member UAVs and the leader UAV is calculated; the continuous time axis is divided into multiple discrete time slots of equal length; the spectrum resources of the UAV swarm network are divided into multiple identical sub-channels; power levels are defined, and communication sub-channel occupancy and power level selection indicator variables are provided; the communication rate of the UAVs is calculated using the Shannon formula; based on the calculated channel gain, discrete time slots, sub-channels, indicator variables, and communication rate, a communication model is obtained; the task data volume, task requirement delay, and task priority are established, and the actual transmission delay of the task is calculated based on the communication rate. A task success transmission indicator variable is defined, and a task model is established. Among these, the network model, communication model, and task model belong to the low-altitude communication network system model.

[0026] In some embodiments, the processor can establish a low-altitude communication network composed of multiple drone swarms. Each swarm includes a leader drone equipped with multiple antennas and multiple member drones equipped with one antenna each, and the swarm adopts a star topology. A main controller is deployed at the base station, responsible for interacting with the leader drone to manage the external swarm status and monitor global resources; a secondary controller is deployed at the leader drone, responsible for resource allocation between member drones and the leader drone within the swarm. Member drones transmit data to the leader drone via uplink, where the data is aggregated and then forwarded to the base station for processing, thereby achieving load balancing and improving network scalability. Specifically, the processor can construct a network model based on a low-altitude drone swarm network. This network consists of multiple drone swarms. , This represents the total number of drones in the cluster. Each cluster consists of a leader drone equipped with multiple antennas and multiple member drones equipped with one antenna each, using a star topology. The main controller is deployed at the base station and is responsible for interacting with the leader drone to manage the external cluster status and monitor global resources. The secondary controller is deployed at the leader drone and is responsible for resource allocation between the member drones and the leader drone. Member drones transmit data to the leader drone via the uplink, where it is aggregated and then forwarded to the base station for processing, thereby achieving load balancing and improving network scalability.

[0027] In some embodiments, the processor can establish a line-of-sight channel model between UAVs, calculate the channel gain between member UAVs and the leading UAV; divide the continuous time axis into multiple discrete time slots of equal length; divide the spectrum resources of the UAV swarm network into multiple identical sub-channels; classify power levels, and provide communication sub-channel occupancy and power level selection indicator variables; then, calculate the communication rate of the UAVs using the Shannon formula. Specifically, the processor can divide the channel into multiple equal-length sub-channels of equal length. The sub-channel is denoted as , This represents the total number of sub-channels. The time slot is divided into multiple equal-length time slots, represented as... , This represents the total number of time slots. UAVs communicate using a non-orthogonal transmission method, and each UAV can occupy at most one sub-channel for communication. A cluster channel model for the low-altitude UAV swarm network is constructed. Based on the line-of-sight channel model, the communication channel between member UAVs and the leading UAV is modeled, providing cluster... Inner Channel gain between member drones and the lead drone Give time slot Subchannel occupancy and power level indicator variables The communication rate between the member drones and the first drone was calculated using Shannon's formula. .

[0028] In some embodiments, the processor can establish a task model, including task size, latency, and priority; calculate the actual transmission latency of the task based on the communication rate; define a successful transmission indicator variable; and obtain the proportion of successful task transmissions within the cluster. Specifically, the processor can construct a task model for a low-altitude UAV swarm network. Each task is represented as: Each task includes a priority level. Data volume Delayed demand Calculate the actual transmission delay of the task based on the task data volume and transmission rate. , The current transmission rate is used, and the success of the task transmission is determined by comparing the latency with the task latency requirements. Further define the success transmission rate for tasks with different priorities. This is used as an evaluation metric for cluster task transmission and resource allocation.

[0029] S2: Based on the low-altitude communication network system model, construct the optimization equation for the allocation of network resources in the UAV swarm.

[0030] In some embodiments, the processor can construct an optimization equation for the allocation of network resources in a drone swarm, including: designing an intra-cluster utility function weighted by the probability of successful task transmission and task priority; constructing an intra-cluster resource allocation optimization equation with the goal of maximizing the probability of successful task transmission; providing inter-cluster frequency block related definitions and constructing an inter-cluster resource allocation optimization equation with the goal of minimizing the variance of the probability of successful task transmission between clusters; and constructing a global optimization equation with the goal of maximizing the probability of successful task transmission and the constraint of minimizing the variance of the probability of successful task transmission between clusters.

[0031] In some embodiments, the processor can utilize an inter-cluster frequency block partitioning method to divide all spectrum resources in the low-altitude communication network system model into non-overlapping frequency blocks, obtaining the partitioning and usage of frequency blocks. Each frequency block serves as the smallest unit of spectrum used by the cluster. Based on the partitioning and usage of frequency blocks, and using the cluster utility function and resource constraints, an optimization equation for UAV cluster network resource allocation is constructed, including optimization equations within the UAV cluster, optimization equations between UAV clusters, and global optimization equations. Specifically, the processor can provide an inter-cluster frequency block partitioning method to divide all spectrum resources in the UAV cluster network into non-overlapping frequency blocks, with each frequency block serving as the smallest unit of spectrum used by the cluster. A frequency block is defined as a set of continuous sub-channels. Represents the first in the network Each frequency block is orthogonal to the others and together they constitute the network's spectrum resources.

[0032] In some embodiments, the expression for frequency block allocation and usage is as follows: ; ; ; ; ; ; Where Q represents the total number of frequency blocks. Represents a cluster Frequency blocks Indicator variables for usage, Indicates the first One cluster, Represents a cluster set. This indicates the cluster corresponding to the maximum value. Represents a cluster Number of sub-channels used Indicates the first The number of sub-channels contained in each frequency block Indicates the occupation of the first A cluster set of frequency blocks, Represents the first in the network Each frequency block, Represents the first in the network Each frequency block, This represents the first frequency block in the network. This represents the second frequency block in the network. This represents the Q-th frequency block in the network. Represents the set of sub-channels in the network. This represents the a-th sub-channel in the network. Indicates the first in the network Sub-channels, Indicates the number of sub-channels.

[0033] In some embodiments, the expression for the cluster utility function is: ; ; The expression for the optimization equation within the drone swarm is: ; ; ; ; ; ; The expression for the optimization equation among drone swarms is: ; ; ; ; ; ; ; ; ; ; The expression for the global optimization equation is: ; ; in, Indicates the first The utility function of a cluster. This represents the set of task priorities, where m represents the UAV task priority index. Indicates the first The weight corresponding to the priority of each task This indicates that the priority within a cluster is The ratio of the number of successfully transmitted tasks to the total number of tasks of that priority. Represents the set of weights. This indicates the weight of task with priority 1. This indicates the priority 2 task weight. Indicates priority Task weight, Indicates each priority as Task weights This represents the function that takes the maximum value, where k represents the sub-channel index. Represents a cluster The set of sub-channels occupied (t) represents the cluster The first Are the individual member drones in the time slot? Occupy sub-channel , Indicates time slot , This indicates the preset SINR threshold. Represents a cluster The set of the i-th power indicator variables, This indicates that corresponding elements within a set are multiplied and added together. Represents a set of power levels. This represents the maximum allowable power value, and j represents the index of the power indicator variable. Represents a set of power levels. (t) represents the cluster The first Are the individual member drones in the time slot? Use power rating , This represents the expectation operation. Indicates the occupied frequency block The corresponding cluster set, Let q represent the frequency block index and Q represent the number of frequency blocks, where Q represents the utility function. Indicates the first individual frequency blocks The corresponding set of sub-channels, Represents the set of sub-channels. Indicates the first Each frequency block, Indicates the first Each frequency block, Indicates the first frequency block. Indicates the second frequency block. This represents the Q-th frequency block. Indicates the first individual frequency blocks The number of sub-channels included. Represents a set index. Indicates the frequency band used A collection of clusters, Represents a cluster set. Represents a cluster Does it occupy a frequency block? , Indicates multiplication. This represents the shortest distance between the member drone of the i-th cluster and the member drone of the l-th cluster within cluster c. Indicates the threshold of no interference. Indicates a pair with Related numbers, J represents the transmit power at power level J, and n represents the number of clusters sharing the frequency band. This indicates the transmit power at power level 1. This represents the farthest distance between the member drone of the i-th cluster within cluster c and the leader drone of that cluster. This represents the function that takes the minimum value.

[0034] The number of sub-channels in the frequency blocks occupied by each cluster is jointly optimized. The number of sub-channels, the number of frequency blocks, and their corresponding constraints are used as constraints to minimize the difference in the probability of successful task transmission between clusters, thereby achieving fairness in resource allocation between clusters.

[0035] In some embodiments, the processor can provide an intra-cluster utility function based on a weighted average of the probability of successful task transmission and task priority, providing a theoretical optimization objective for intra-cluster optimization and thus ensuring reliable transmission of critical tasks. An optimization equation for intra-cluster resource allocation is constructed with the goal of maximizing the intra-cluster utility function. This jointly optimizes the transmit power and sub-channel allocation of member drones within the cluster to maximize the probability of successful task transmission within the cluster, ensuring reliable transmission of differentiated tasks. An inter-cluster resource allocation optimization equation is constructed with the goal of minimizing the variance of the utility functions of each cluster. By optimizing the number of frequency block sub-channels occupied by each cluster, the utility function values ​​among different clusters are balanced, thereby ensuring the fairness of overall network resource allocation. A global optimization equation is constructed with the goal of maximizing the sum of the utility functions of each cluster. This jointly optimizes the power level and sub-channel allocation of drones within the cluster, and uses the minimization of the variance of the probability of successful task transmission between clusters as a constraint to achieve overall coordination and optimization across clusters, improving the overall reliability of task transmission and the fairness of resource allocation across the network. Specifically, the processor can provide a utility function weighted by the probability of successful task transmission and task priority, comprehensively considering the differentiated needs of tasks and reliable transmission; and construct optimization equations within the UAV swarm to... To achieve the optimization objectives, the transmit power and sub-channel allocation of member UAVs are jointly optimized, with constraints including sub-channel occupancy, signal-to-interference-plus-noise ratio (SINR), power, and binary variables, to maximize the probability of successful transmission of differentiated tasks within the cluster. An inter-cluster optimization equation is constructed with the objective of minimizing the variance of the utility function of each cluster. A global optimization equation is also constructed with the objective of maximizing the sum of the utility functions of each cluster. Finally, the power of UAVs within the cluster, sub-channel occupancy, and the number of sub-channels in the frequency block are jointly optimized, with the constraint of minimizing the variance of the probability of successful task transmission between clusters, to achieve overall coordination and optimization across clusters, thereby improving the reliability of global task transmission and the fairness of resource allocation in the network.

[0036] S3: Based on the optimization equation for network resource allocation of UAV swarms, the frequency block sharing method of UAV swarms is used for analysis to obtain the frequency block sharing scheme of UAV swarms.

[0037] In some embodiments, the processor can provide a method for sharing frequency blocks in a drone swarm, providing interference-free conditions for shared frequency blocks and designing a corresponding frequency block allocation method to obtain a frequency block sharing scheme for the drone swarm, providing an efficient initial solution for subsequent resource allocation. Specifically, the processor can design an optimized coloring graph method, combined with the derived interference-free sharing conditions, to determine whether each drone swarm can share frequency blocks, thereby obtaining a frequency block sharing scheme for the drone swarm and providing an efficient initial solution for subsequent intra- and inter-swarm resource allocation. First, as... Figure 4 As shown, give The condition that a cluster of drones uses the same frequency block without interference (assuming that the gain of the drone's transmitting antenna and receiving antenna are both 1).

[0038] In some embodiments, the processor can abstract each cluster in the UAV swarm network as a node based on the UAV swarm network resource allocation optimization equation; determine whether the interference-free condition for inter-cluster frequency block sharing is met between any two nodes; if not, add an edge between the nodes to obtain the network topology, forming an undirected graph G; execute the traditional coloring graph method to color each node in the undirected graph G to obtain all solutions; traverse all solutions, if there is a solution where one color is occupied by more than three nodes, determine again whether it meets the interference-free condition; if not, move one of the nodes in that color to the cache set; traverse the nodes in the cache set, reassign colors to them, and remove the node when it meets the interference-free condition, until the cache set is empty, and update all feasible solutions; calculate the optimal solution among all feasible solutions, taking the one with the smallest variance in the number of nodes contained in all colors as the optimal solution, and obtain the frequency block sharing scheme for the UAV swarm.

[0039] In some embodiments, the expression for the interference-free condition is: ; ; in, Indicates multiplication. This represents the shortest distance between the member drone of the i-th cluster and the member drone of the l-th cluster within cluster c. Indicates the threshold of no interference. Indicates a pair with Related numbers, This indicates the transmit power at power level J. This indicates the transmit power at power level 1. This represents the farthest distance between the member drone of the i-th cluster within cluster c and the leader drone of that cluster. Represents the i-th cluster, Indicates the number of clusters sharing the frequency band. Indicates the signal wavelength.

[0040] In some embodiments, the processor can provide conditions for interference-free sharing of frequency blocks between clusters; abstract each cluster in the UAV swarm network as a node, and determine whether the interference-free conditions are met between each node. If not, an edge is added between the nodes to obtain an undirected graph; execute a coloring graph method to color each node in the undirected graph to obtain all solutions (i.e., frequency block sharing schemes); traverse all solutions, and if there is a solution where one color is used by more than three nodes, determine whether it meets the interference-free conditions. If not, move one of the nodes in that color to a cache set; traverse the nodes in the cache set, reassigning colors to them until the interference-free conditions are met and removing the node, until the cache set is empty, and update all feasible solutions; calculate all feasible solutions, select the one with the smallest variance in the number of nodes contained in all colors as the optimal solution, and obtain the final inter-cluster frequency block sharing scheme. Specifically, the processor can abstract each cluster in the UAV swarm network as a node; determine whether the interference-free conditions are met between each node; if not, add an edge between the nodes to obtain the network topology. The traditional coloring method is applied to an undirected graph. Color each node in the solution to obtain all solutions. (Frequency block sharing scheme); Traverse all solutions. If a solution exists where one color is occupied by more than three nodes, determine whether it meets the interference-free condition. If not, move one of the nodes of that color to the cache set. ; Traverse the cache collection For each node in the cache set, recolor it until a no-interference condition is met, then remove the node. For an empty set, update all feasible solutions. Calculate all feasible solutions In the solution, the one with the smallest variance in the number of nodes contained in all colors is considered the optimal solution. Thus, a frequency block sharing scheme was obtained.

[0041] S4: Based on the frequency block sharing scheme, the dynamic frequency block contention algorithm is used to determine the number of sub-channels contained in the frequency block occupied by each UAV cluster with the goal of maximizing the fairness of resource allocation among clusters, thus obtaining the resource allocation scheme among UAV clusters.

[0042] In some embodiments, the processor can provide a dynamic frequency block contention algorithm based on a deep dual-Q network to determine the number of sub-channels contained in the frequency block occupied by each UAV cluster, aiming at fairness in resource allocation among clusters, and realizing resource allocation among UAV clusters. Specifically, the processor can obtain the specific frequency block occupied by each UAV through an adaptive frequency block sharing algorithm of the UAV cluster, and then dynamically determine the number of sub-channels contained in the frequency block occupied by each UAV based on the deep dual-Q network, realizing resource allocation among clusters and providing a spectrum resource foundation for subsequent resource allocation within the cluster.

[0043] In some embodiments, the processor can obtain the frequency blocks occupied by each cluster based on a frequency block sharing scheme; construct an initialized primary Q-network and secondary Q-network; calculate the Q-values ​​of all current actions from the Q-networks by initializing the state space, and according to... The greedy algorithm selects the optimal action to obtain the number of sub-channels contained in each frequency block. Based on the number of sub-channels in each frequency block, it calls the corresponding cluster resource allocation strategy to complete the power and sub-channel allocation within the cluster, thereby obtaining the current state and immediate reward. Based on the current state and immediate reward, iteratively trains the primary Q-network and secondary Q-network to obtain the trained primary Q-network and secondary Q-network. By initializing the state, the trained primary Q-network and secondary Q-network are tested. Tasks arrive dynamically in each time slot. Based on the current state, the success probability of task transmission for each cluster is obtained, and then the utility function of each cluster is calculated to obtain the reward function at the current moment. Based on the network, the current action and the state at the next moment are obtained. The action is executed to complete the resource allocation for the current time slot and enter the next time slot. If the reward is less than the threshold, the state is reset, and the number of sub-channels contained in the frequency block occupied by each UAV cluster is determined to obtain the resource allocation scheme between UAV clusters. Specifically, the processor can provide the system state space: Defined as the number of available sub-channels per frequency block; b. Provides system action space: Defined as the single change in the number of available sub-channels in a frequency block at each step, where , And satisfying: in each There is only one in the middle. Each The sum of the elements satisfies ; ,in represent The number of combinations; for every two actions and ( ), That is, any two actions are not the same. c. Provide a reward function and exploration rate to achieve a balance between exploration and exploitation, and accelerate the convergence of the model.

[0044] In some embodiments, the expression for the reward function is: ; in, Represents the reward function, Represents a piecewise function. This represents the expectation operation. Indicates the occupied frequency block A collection of clusters, Represents the utility function.

[0045] In some embodiments, the processor can obtain the frequency blocks occupied by each cluster using an adaptive frequency block sharing algorithm for UAV swarms; initialize the parameters of the primary Q-network and the secondary Q-network, and set the global step counter to 0; in each training round, first initialize the state space, then calculate the Q-values ​​of all current actions from the Q-network, according to... - A greedy algorithm selects the optimal action; executes the selected action, calculates the number of sub-channels in the frequency block occupied by the cluster in the current state, calls the cluster's resource allocation strategy, and thus obtains a new state and current reward; stores the current state, action, reward, and next state in the experience replay pool, updates the current state, and increments the global step count by 1 for subsequent sampling training; periodically randomly selects a batch of samples from the replay pool to train the main Q network and update its parameters; after reaching a certain number of steps, synchronizes the parameters of the main Q network to the target Q network to ensure the stability and convergence speed of the training process; after meeting the training cutoff condition, saves the neural network parameters and ends the pre-training; initializes the state; each time slot task arrives dynamically, calculates the utility function of each cluster based on the success transmission probability of the task, and then calculates the reward at the current moment and the state at the next moment of the current action; executes the action, completes the resource allocation of the current time slot, and enters the next time slot; if the reward is less than the threshold, the state is reset. Specifically, the processor can obtain the frequency block occupied by each cluster using the UAV cluster frequency block sharing method; the pre-training phase begins by initializing the main Q network parameters. Parameters of the sub-Q network , Set the global step counter global_step=0; in each training round, first initialize the state space. Then, calculate the Q-values ​​of all current actions from the Q-network, based on... Greedy algorithm selects the optimal action ; Obtain the number of sub-channels contained in each frequency block, then invoke the corresponding cluster resource allocation strategy and execute the selected action. Thus, a new state is obtained. and instant rewards ;Will The data is stored in the experience replay pool, and the current state is updated. Simultaneously, the global step count (global_step) is incremented by 1 for subsequent sampling and training. A batch of samples is periodically randomly selected from the replay pool to train the main Q-network and update its parameters. Every certain number of steps, the parameters of the main Q-network are synchronized to the target Q-network. This ensures the stability and convergence speed of the training process; after the training cutoff condition is met, the neural network parameters are saved, and pre-training ends; during the testing phase, the state is initialized. Each time slot task arrives dynamically, based on the current state. The success rate of task transmission for each cluster is obtained, and then the utility function of each cluster is calculated to obtain the reward at the current time step. And obtain the current action based on the Q network. And the state in the next moment. ; Perform actions Complete resource allocation for the current time slot and proceed to the next time slot; if reward If the value is less than the threshold, the state is reset.

[0046] In some embodiments, the expression for the probability of randomly selecting an action is: ; in, This represents the probability of randomly selecting an action. Indicates the number of rounds.

[0047] S5: Utilizing a multi-priority dynamic resource allocation algorithm, with the goal of maximizing the success probability of differentiated task transmission, the resource allocation schemes for sub-channels and power of UAVs within the cluster are optimized. By integrating the resource allocation schemes among UAV clusters, the network resource allocation results are obtained, thus completing the network resource allocation for the low-altitude UAV cluster.

[0048] In some embodiments, the processor can provide a multi-priority dynamic resource allocation algorithm based on a genetic algorithm, aiming to maximize the successful transmission probability of differentiated tasks. This achieves joint optimization of power and sub-channel allocation within the cluster, while also ensuring fairness between clusters, thereby guaranteeing reliable transmission of differentiated tasks and fairness between clusters. Specifically, the processor can leverage the efficiency of genetic algorithms in optimizing complex and discrete resource allocation, enabling each cluster to efficiently optimize and achieve near-optimal power and sub-channel solutions based on its occupied frequency block resources. This includes: a. Encoding: using binary form to encode the cluster... The power within is encoded and represented as ,in , , This represents the floor function. (Cluster) The subchannel coding is represented as: ; Indicates the number within the cluster Do each member's drone occupy the first... Sub-channels, , For clusters Number of drones within the cluster. One chromosome within is represented as: ; b. Fitness function: Defines the cluster The fitness function is ; in It is a penalty function, defined as: ; This represents a binary to decimal conversion function.

[0049] c. Selection, crossover, and mutation: The probability of an individual being selected to enter the next generation is determined by the following formula, ; Indicates the first The fitness function of each chromosome. To enhance the global search capability of the genetic algorithm, the chromosomes are compared using a fitness function. Cross over the probabilities. The probability varies.

[0050] In some embodiments, the processor can determine the population size, the number of member drones, the power level, and the number of iterations, determine the number of sub-channels of the cluster based on the resource allocation scheme, and obtain an initial population; based on the initial population, with the goal of maximizing the probability of successful transmission of differentiated tasks, the processor calculates the fitness function value of the population and selects a certain number of chromosomes from the population according to the selection probability to generate the first population. Crossover is performed on the first population according to the crossover probability to obtain the second population; mutation is performed on the second population according to the mutation probability to obtain the third population; the fitness function of each chromosome in the third population is calculated, and the fitness function values ​​of chromosomes in the initial population and the third population are compared. Chromosomes with higher fitness are retained to obtain the updated initial population; based on the updated population, the chromosome corresponding to the largest fitness function is extracted to obtain the power and sub-channel allocation results within the cluster. The resource allocation scheme between UAV clusters is integrated to obtain the network resource allocation results, thus completing the network resource allocation within the low-altitude UAV cluster.

[0051] In some embodiments, the processor can provide a utility function expression for differentiated task transmission based on a genetic algorithm, and provide a corresponding penalty function; input state information: population size, number of member UAVs in the cluster, number of sub-channels corresponding to the occupied frequency band, power level, number of iterations, crossover probability and mutation probability; initialize the population; construct chromosomes in the genetic algorithm according to the transmit power of member UAVs in the cluster and their sub-channel occupancy with the leading UAV, and initialize the population pop according to the preset population size; in each iteration, calculate the fitness function value of the population; select a certain number of chromosomes according to the selection probability to obtain a new population pop1; cross pop1 according to the crossover probability to obtain pop2; mutate the chromosomes in pop2 according to the mutation probability to obtain pop3; calculate the fitness function of each chromosome in pop and pop3 respectively, compare their size, retain the one with the larger fitness, and update the population pop. Output the largest fitness function and its chromosome, i.e., the power and sub-channel allocation scheme in the cluster. Specifically, the processor can input state information: population size, number of member UAVs, number of sub-channels of cluster c, power level and number of iterations; b. initialize the population. c. In each iteration, calculate the population. fitness function value d. According to the probability of selection Select a certain number of chromosomes from the population to generate a new population. e. According to crossover probability right Perform crossover operations to obtain the population. f. According to the probability of mutation right By performing mutations, a new population is obtained. g. Calculation Fitness function for each chromosome ,Compare and Size, retain those with high fitness, update the population. h. Output: The maximum fitness function and its chromosome, i.e., the power and sub-channel allocation scheme within the cluster.

[0052] The resource allocation method for low-altitude UAV swarm networks, based on mission success rate, fully considers the differentiated needs of missions, effectively ensures reliable mission transmission, and simultaneously maintains network fairness. This improves resource utilization efficiency and overall system performance, thereby promoting the intelligent and large-scale development of low-altitude UAV networks. This method is not only applicable to high-reliability scenarios such as disaster emergency response but can also be extended to low-altitude application fields such as intelligent transportation and industrial internet, demonstrating significant practical value and application prospects.

[0053] In some embodiments of this specification, a low-altitude UAV swarm network resource allocation method oriented towards mission success transmission rate is provided. Through frequency block sharing strategy, inter-cluster resource allocation and intra-cluster resource allocation methods, the resource utilization rate and differentiated service quality of UAV swarm network are optimized in a coordinated manner. (1) By deriving the conditions for interference-free sharing of frequency blocks between different clusters in the UAV swarm network, the sharing of spectrum resources between different UAV clusters is realized, thereby providing an efficient and stable spectrum management scheme for large-scale UAV access and improving network resource utilization and communication performance; (2) The deep reinforcement learning strategy is used to adaptively solve the problem of spectrum resource allocation decision-making between UAV network clusters under unknown environmental conditions. This method can well adapt to the differences between different UAV clusters and make targeted resource allocation decisions, thereby ensuring the fairness of resource allocation between UAV clusters; (3) A heuristic method is used to solve the power and sub-channel allocation problem within the cluster. By utilizing the advantages of fast algorithm convergence speed, low computational complexity and strong global search capability, the defects of traditional methods that are prone to getting trapped in local optima are avoided, thereby maximizing the probability of mission success transmission and the efficient utilization of resources within the cluster.

Claims

1. A resource allocation method for low-altitude UAV swarm networks based on mission success transmission rate, characterized in that, include: S1: Establish a low-altitude communication network system model based on low-altitude UAV swarms; S2: Based on the low-altitude communication network system model, construct the optimization equation for the allocation of network resources in the UAV swarm; S3: Based on the optimization equation for network resource allocation of UAV swarms, the frequency block sharing method of UAV swarms is used for analysis to obtain the frequency block sharing scheme of UAV swarms; S4: Based on the frequency block sharing scheme, the dynamic frequency block contention algorithm is used to maximize the fairness of resource allocation among clusters. The number of sub-channels contained in the frequency block occupied by each UAV cluster is determined, and the resource allocation scheme among UAV clusters is obtained. S5: Utilizing a multi-priority dynamic resource allocation algorithm, with the goal of maximizing the success probability of differentiated task transmission, the resource allocation schemes for sub-channels and power of UAVs within the cluster are optimized. By integrating the resource allocation schemes among UAV clusters, the network resource allocation results are obtained, thus completing the network resource allocation for the low-altitude UAV cluster.

2. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 1, characterized in that, S1 includes: A low-altitude communication network consisting of multiple UAV swarms is established. Each swarm consists of a leader UAV equipped with multiple antennas and multiple member UAVs equipped with one antenna. The swarm adopts a star topology. The main controller is deployed at the base station, and the auxiliary controller is deployed at the leader UAV. The member UAVs transmit data to the leader UAV via the uplink. The leader UAV aggregates the data and then forwards it to the base station for processing. A network model is established. A line-of-sight channel model between UAVs is established, and the channel gain between member UAVs and the leading UAV is calculated. The continuous time axis is divided into multiple discrete time slots of equal length. The spectrum resources of the UAV swarm network are divided into multiple identical sub-channels. Power levels are defined, and communication sub-channel occupancy and power level selection indicator variables are provided. The communication rate of the UAVs is calculated using Shannon's formula. Based on the calculated channel gain, discrete time slots, sub-channels, indicator variables, and communication rate, the communication model is obtained. Establish the task data volume, task requirement delay, and task priority; calculate the actual transmission delay of the task based on the communication rate; define the indicator variable for successful task transmission; and establish the task model. Among them, the network model, communication model, and task model belong to the low-altitude communication network system model.

3. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 1, characterized in that, S2 includes: Using the inter-cluster frequency block partitioning method, all spectrum resources in the low-altitude communication network system model are divided into non-overlapping frequency blocks, and the partitioning and usage of frequency blocks are obtained. Each frequency block is used as the smallest unit of spectrum used by the cluster. Based on the frequency block partitioning and usage, and utilizing the cluster utility function and resource constraints, a network resource allocation optimization equation for UAV clusters is constructed, which includes optimization equations within the UAV cluster, optimization equations between UAV clusters, and global optimization equations.

4. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 3, characterized in that, The expression for the division and usage of the frequency blocks is as follows: ； ； ； ； ； ； Where Q represents the total number of frequency blocks. Represents a cluster Frequency blocks Indicator variables for usage, Indicates the first One cluster, Represents a cluster set. This indicates the cluster corresponding to the maximum value. Represents a cluster Number of sub-channels used Indicates the first The number of sub-channels contained in each frequency block Indicates the occupation of the first A cluster set of frequency blocks, Represents the first in the network Each frequency block, Represents the first in the network Each frequency block, This represents the first frequency block in the network. This represents the second frequency block in the network. This represents the Q-th frequency block in the network. Represents the set of sub-channels in the network. This represents the a-th sub-channel in the network. Indicates the first in the network Sub-channels, Indicates the number of sub-channels.

5. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 3, characterized in that, The expression for the cluster utility function is: ； ； The expression for the optimization equation within the drone swarm is: ； ； ； ； ； ； The expression for the optimization equation among drone clusters is: ； ； ； ； ； ； ； ； ； ； The expression for the global optimization equation is: ； ； in, Indicates the first The utility function of a cluster. This represents the set of task priorities, where m represents the UAV task priority index. Indicates the first The weight corresponding to the priority of each task This indicates that the priority within a cluster is The ratio of the number of successfully transmitted tasks to the total number of tasks of that priority. Represents the set of weights. This indicates the weight of priority 1 task. This indicates the priority 2 task weight. Indicates priority Task weight, Indicates each priority as Task weights This represents the function that takes the maximum value, where k represents the sub-channel index. Represents a cluster The set of sub-channels occupied (t) represents the cluster The first Are the individual member drones in the time slot? Occupy sub-channel , Indicates time slot , This indicates the preset SINR threshold. Represents a cluster The set of the i-th power indicator variables, This indicates that corresponding elements within a set are multiplied and added together. Represents a set of power levels. This represents the maximum allowable power value, and j represents the index of the power indicator variable. Represents a set of power levels. (t) represents the cluster The first Are the individual member drones in the time slot? Use power rating , This represents the expectation operation. Indicates the occupied frequency block The corresponding cluster set, Let q represent the frequency block index and Q represent the number of frequency blocks, where Q represents the utility function. Indicates the first individual frequency blocks The corresponding set of sub-channels, Represents the set of sub-channels. Indicates the first Each frequency block, Indicates the first Each frequency block, Indicates the first frequency block. Indicates the second frequency block. This represents the Q-th frequency block. Indicates the first individual frequency blocks The number of sub-channels included. Represents a set index. Indicates the frequency band used A collection of clusters, Represents a cluster set. Represents a cluster Does it occupy a frequency block? , Indicates multiplication. This represents the shortest distance between the member drone of the i-th cluster and the member drone of the l-th cluster within cluster c. Indicates the threshold of no interference. Indicates a pair with Related numbers, J represents the transmit power at power level J, and n represents the number of clusters sharing the frequency band. This indicates the transmit power at power level 1. This represents the farthest distance between the member drone of the i-th cluster within cluster c and the leader drone of that cluster. This represents the function that takes the minimum value.

6. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 1, characterized in that, S3 includes: Based on the resource allocation optimization equation of UAV swarm network, each swarm in the UAV swarm network is abstracted as a node; Determine whether the inter-cluster frequency block sharing interference-free condition is met between any two nodes. If not, add an edge between the nodes to obtain the network topology and form an undirected graph G. Perform the traditional graph coloring method to color each node in the undirected graph G to obtain all solutions; Iterate through all solutions. If there is a solution where one color is occupied by more than three nodes, then check again whether it meets the no-interference condition. If it does not meet the condition, then move one of the nodes of that color to the cache set. Iterate through the nodes in the cache set, reassign them colors, and remove the node when it meets the no-interference condition, until the cache set is empty, and update all feasible solutions. The optimal solution is the one with the smallest variance in the number of nodes contained in all colors among all feasible solutions, thus obtaining the frequency block sharing scheme for the drone swarm.

7. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 6, characterized in that, The expression for the interference-free condition is: ； ； in, Indicates multiplication. This represents the shortest distance between the member drone of the i-th cluster and the member drone of the l-th cluster within cluster c. Indicates the threshold of no interference. Indicates a pair with Related numbers, This indicates the transmit power at power level J. This indicates the transmit power at power level 1. This represents the farthest distance between the member drone of the i-th cluster within cluster c and the leader drone of that cluster. Represents the i-th cluster, Indicates the number of clusters sharing the frequency band. Indicates the signal wavelength.

8. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 1, characterized in that, S4 includes: Based on the frequency block sharing scheme, the frequency blocks occupied by each cluster are obtained; Construct the initial primary Q-network and secondary Q-network; By initializing the state space, the Q-values ​​of all current actions are calculated from the Q-network, based on... The greedy algorithm selects the optimal action to obtain the number of sub-channels contained in each frequency block at the current time; Based on the number of sub-channels contained in each frequency block, the corresponding cluster resource allocation strategy is invoked to complete the power and sub-channel allocation within the cluster, thereby obtaining the current status and immediate reward. Based on the current state and immediate reward, the main Q-network and the secondary Q-network are trained iteratively to obtain the trained main Q-network and secondary Q-network. By initializing the state, the trained main Q network and secondary Q network are tested. Each time slot task arrives dynamically. Based on the current state, the success transmission probability of each cluster is obtained, and then the utility function of each cluster is calculated to obtain the reward function at the current moment. Based on the network, the current action and the state at the next moment are obtained. Execute the action, complete the resource allocation for the current time slot, and enter the next time slot. If the reward is less than the threshold, reset the state, determine the number of sub-channels contained in the frequency block occupied by each drone cluster, and obtain the resource allocation scheme between drone clusters.

9. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 8, characterized in that, The expression for the reward function is: ； in, Represents the reward function, Represents a piecewise function. This represents the expectation operation. Indicates the occupied frequency block A collection of clusters, Represents the utility function.

10. The low-altitude UAV swarm network resource allocation method based on mission success transmission rate according to claim 1, characterized in that, S5 includes: Determine the population size, number of member drones, power level, and number of iterations. Based on the resource allocation scheme, determine the number of sub-channels in the cluster to obtain the initial population. Based on the initial population, with the goal of maximizing the probability of successful transmission of differentiated tasks, the first population is generated by calculating the fitness function value of the population and selecting a certain number of chromosomes from the population according to the selection probability. Perform a crossover operation on the first group according to the crossover probability to obtain the second group; The second population is mutated according to the mutation probability to obtain the third population; Calculate the fitness function of each chromosome in the third population, compare the fitness function values ​​of chromosomes in the initial population with those in the third population, retain the chromosomes with higher fitness, and obtain the updated initial population. Based on the updated population, the chromosome corresponding to the maximum fitness function is extracted to obtain the power and sub-channel allocation results within the cluster. By combining the resource allocation schemes among the UAV clusters, the network resource allocation results are obtained, and the network resource allocation within the low-altitude UAV cluster is completed.

Images