A method for reconstructing a topology of a UAV network

Abstract

The application discloses a kind of unmanned aerial vehicle network topology reconstruction methods, comprising: obtaining unmanned aerial vehicle network information, ground user information;Wherein unmanned aerial vehicle is as air base station and provides communication service for ground user;For the recovery unmanned aerial vehicle network connectivity after unmanned aerial vehicle failure and ensure ground user full coverage and fairness requirement, the optimization problem of optimization target is maximized ground user total fairness throughput;The optimization variable of the optimization problem is the two-dimensional position of each unmanned aerial vehicle, the associated indicator variable between ground user and unmanned aerial vehicle, the communication bandwidth allocated to each ground user;Objective function is composed of the sum of the logarithmic value of all ground user actual throughput;Three-layer iteration method is used to approximate the optimal solution of the optimization problem;Optimal optimization variable adjusts the actual flight position of unmanned aerial vehicle, establishes the access relationship between ground user and unmanned aerial vehicle, allocates communication bandwidth for each user, to complete the topology reconstruction of unmanned aerial vehicle network.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) network technology and relates to a UAV network topology reconstruction method for topology reconstruction that prioritizes fairness in throughput for ground users when a UAV malfunctions. Background Technology

[0002] In UAV network-assisted ground user communication, information exchange between UAV nodes can be achieved without relying on fixed infrastructure, effectively completing data transmission and information fusion. However, UAVs may malfunction due to energy consumption or hardware / software failures, leading to network failure and uneven resource distribution. It is necessary to consider not only the impact of UAV malfunctions on the airspace UAV communication network but also the impact on ground user coverage. While increasing power or deploying relay nodes can mitigate the impact of malfunctions to some extent, these methods typically require additional resources or hardware support, making efficient repair difficult. Node mobility methods can utilize existing network resources to achieve self-organized reconfiguration after a topology failure, effectively restoring network functionality.

[0003] In resource allocation within communication networks, traditional linear utility functions typically aim to maximize the total system throughput. However, this method often overlooks ground users with poor channel conditions, resulting in them receiving minimal resources or even being denied service, leading to significant unfairness. To improve system fairness, a logarithmic utility function is employed for resource allocation, allowing low-throughput ground users to obtain higher marginal utility, thereby enhancing fairness among ground users while maintaining overall performance.

[0004] In current research on node mobility methods, the Dual Domain Coordinated Descent (DCD) algorithm achieves joint ground user association and resource allocation for multiple UAV base stations, and uses bipartite graph matching to correct non-unique solutions under backhaul rates, thereby improving system throughput while ensuring fairness. The Coverage-First Ground User Association (BoaRD) method based on backhaul connectivity constraints optimizes the number of UAVs, 3D deployment locations, and ground user association relationships. It uses graph theory modeling and heuristic solving to minimize the required number of UAVs while ensuring full coverage and backhaul connectivity. Although these two methods consider efficient ground user association and resource coordination, they lack dynamic resource allocation and load balancing mechanisms. Summary of the Invention

[0005] The purpose of this invention is to provide a method for reconstructing the network topology of unmanned aerial vehicles (UAVs), which aims to solve the problems of decreased network connectivity and incomplete ground user coverage caused by UAV malfunctions in UAV-assisted ground communication networks.

[0006] To achieve the above objectives, the present invention employs the following technical solution: A method for reconstructing the topology of a drone network, comprising: Acquire network information and ground user information from drones; drones act as airborne base stations to provide communication services to ground users. To address the need for restoring UAV network connectivity after a UAV malfunction while ensuring full coverage and fairness for ground users, an optimization problem is constructed with the objective of maximizing the total fair throughput for ground users. The optimization variables in this problem are the two-dimensional position of each UAV, the correlation indicator variable between ground users and UAVs, and the communication bandwidth allocated to each ground user. The objective function is the sum of the logarithmic values ​​of the actual throughput of all ground users. The constraints of this optimization problem include: Ground user QoS constraints are used to ensure that ground users meet the minimum throughput threshold for QoS guarantees when served by UAVs; bandwidth constraints are used to limit the total bandwidth allocated to all ground users to not exceed the total transmission bandwidth of the UAVs; ground user association constraints are used to limit each ground user to be served by only a single UAV, limit the upper limit of the number of ground users served by a single UAV, and ensure that all ground users are covered; connectivity constraints are used to maintain the connectivity recovery capability of the UAV network after a failure; correlation constraints are used to restrict the connectivity indicators and correlation indicators between UAVs. A three-layer iterative method is used to approximate the optimal solution to the optimization problem, including a user association layer, a bandwidth allocation layer, and a topology reconstruction layer; when the convergence condition is met or the maximum number of iterations is reached, the optimal objective function value and the optimal optimization variables are output; Adjusting the actual flight position of the UAV according to the optimal optimization variables, establishing the access relationship between ground users and the UAV, and allocating communication bandwidth to each user, thereby completing the topology reconstruction of the UAV network.

[0007] Furthermore, the objective function is expressed as follows:

[0008] in, The objective function value; For drones Two-dimensional position; For ground users With drones The correlation indicator variable between them, if ground users With drones Establishing a connection between them, then It is 1 if it is true, otherwise it is 0. For each ground user Allocated communication bandwidth; The number of drones; The number of ground users; For users With drones The actual throughput between them.

[0009] Furthermore, the QoS constraints for the ground users are as follows: ; in, This represents the minimum throughput threshold for each ground user; For ground users; For drones; Bandwidth constraints:

[0010] in This indicates the total transmission bandwidth of the drone; Ground user association constraints: ; in The maximum number of ground users that can serve each drone; Connectivity constraints: ; in, Represents a collection of drones Any non-empty subset, drone belong ; Indicates a collection of drones Remove from The set consisting of the remaining elements; For drones With drones Connectivity indicators between; when At that time, drones With drones Establish a connection; otherwise, the two will not connect. Relevance constraints: .

[0011] Furthermore, users With drones The actual throughput between Represented as: ; The communication channel between the UAV and the ground user is a probabilistic Loss-of-Stake (LoS) channel, and the ground user... by drone Services provided; For drones The transmission power; To interfere with drones The transmission power; , To interfere with drones The corresponding LosS probability and NLoS probability; , To interfere with drones The corresponding LoS path loss and NLoS path loss; For noise power spectral density, For ground users The allocated communication bandwidth.

[0012] Furthermore, the solution process for the user association layer is as follows: The optimization problem can be simplified into a user association problem, represented as: ; The aforementioned user association problem is modeled as a maximum matching problem in a weighted bipartite graph. To facilitate finding augmenting paths, the weight matrix is ​​preprocessed. An initial matching is then attempted on the subgraph containing zero elements. If this initial matching is perfect, the algorithm terminates. If the initial matching is imperfect, an augmenting path is further sought: unmatched vertices are selected, and a path containing zero elements is attempted to increase the matching value. If an augmenting path is found, the matching is adjusted by increasing the number of edges within the matching. If no augmenting path is found, the weights in the matrix are adjusted: unmatched rows and columns are selected, and the smallest non-zero element in these rows and columns is calculated. This minimum value is subtracted from the unmatched row and added to the matched column. After weight adjustment, the search for a matching continues until a perfect matching is found, resulting in the set of values ​​for the association indicator variable.

[0013] Furthermore, the solution process for the bandwidth allocation layer is as follows: Given the two-dimensional position of the drone and the association between the drone and the user node In this case, the optimization problem can be simplified to a bandwidth allocation problem: ; The solution is obtained using the KKT conditions, which are as follows: ; in, For partial derivative operators; It is the objective function of the bandwidth allocation subproblem; To constrain Lagrange multipliers, To constrain Lagrange multipliers; The objective function is The objective function value at point and the two endpoint values ​​in the constraints , By comparing the objective function values, the optimal solution to the bandwidth allocation problem can be obtained:

[0014] in, The objective function value for the bandwidth allocation subproblem; The candidate bandwidths are obtained using the KKT conditions; the optimal ground user is obtained by applying the KKT conditions. With drones Communication bandwidth allocated between .

[0015] Furthermore, the solution process for the topology reconstruction layer is as follows: ; Design the following penalty function: ; in, User With drones The actual throughput between; It is the set minimum throughput threshold;

[0016] in Indicates drone and The distance between them and These are the maximum and minimum communication distances between drones, respectively. Two penalty functions and Add to the original objective function Then, the new objective function is obtained as follows:

[0017] in, and For the penalty function and The weighting coefficients are determined; the new objective function is solved by gradient ascent.

[0018] Furthermore, the three-layer iterative method is repeated until the difference between the objective function value obtained in the current iteration and the objective function value in the previous iteration is less than the iteration threshold or the maximum number of iterations is reached. The objective function value obtained in the current iteration is then output as the optimal objective function value, which is the maximum total fair throughput for ground users. At the same time, the corresponding optimal parameters are obtained, including the optimal two-dimensional position of each UAV, the optimal associated indicator variable, and the optimal communication bandwidth.

[0019] A terminal device includes a processor, a memory, and a computer program stored in the memory; when the processor executes the computer program, it implements the UAV network topology reconfiguration method.

[0020] A computer-readable storage medium storing a computer program; when the computer program is executed by a processor, it implements the UAV network topology reconfiguration method.

[0021] Compared with the prior art, the present invention has the following technical features: This invention addresses the issue of UAV malfunction by targeting the total fair throughput of all ground users. It uses UAV network connectivity and ground user coverage constraints as limitations, starting with three sub-problems: network topology reconstruction, ground user association, and bandwidth allocation. Expressions for the solutions to each sub-problem are derived, and a joint topology reconstruction and ground user association algorithm is proposed based on these derivations. This algorithm achieves both aerial network connectivity reconstruction and ground user coverage restoration. In the optimization design, a logarithmic utility function is introduced to balance system throughput and ground user fairness. While ensuring that ground user throughput meets Quality of Service (QoS) requirements, the algorithm reduces interference between ground users through reasonable bandwidth allocation, achieving a synergistic goal of aerial network structure optimization and ground coverage restoration, thereby improving overall network performance and robustness. This invention achieves a fairness index of approximately 0.90, which is superior to existing DCD and BoaRD algorithms. Attached Figure Description

[0022] Figure 1 This is a network topology diagram before and after a single node failure; where (a) is before the failure and (b) is after the failure. Figure 2 This is a diagram of a drone network topology reconfiguration scheme; Figure 3 This is a graph showing the change in total throughput as a function of the number of reconstructed UAVs under different topology reconstruction algorithms; Figure 4 This is a graph showing how the fairness of ground users changes with the number of reconfigured drones. Detailed Implementation

[0023] This invention aims to address the issues of decreased network connectivity and incomplete ground user coverage caused by UAV malfunctions in UAV-assisted ground communication networks. It proposes a UAV network topology reconstruction method. After a UAV malfunctions, this method targets the total fair throughput of all ground users, using UAV network connectivity and ground user coverage constraints as limitations. Starting from three sub-problems—network topology reconstruction, ground user association, and bandwidth allocation—it derives the expression for the solution to each sub-problem. Based on the sub-problem derivation results, it proposes a joint topology reconstruction and ground user association algorithm to achieve network connectivity reconstruction and ground user coverage restoration.

[0024] Assuming there is a drone network A drone set consisting of drones with the same computing power is denoted as a drone set. ;as well as Let there be a set of ground users. Unmanned aerial vehicles (UAVs) act as aerial base stations, providing communication services to ground users. Communication networks must simultaneously consider communication links between multiple UAVs and between UAVs and ground users. (UAVs) The three-dimensional coordinates are Ground users The three-dimensional position is All drones are flying at the same altitude The communication coverage radius of each drone is Each drone can serve a maximum of Each ground user can be associated with at most one drone. The network link is divided into communication links and interference links. Considering that drones in the drone network use orthogonal frequencies for communication, when a ground user is being served by a certain drone, other drones will interfere with that ground user.

[0025] The air-to-ground communication channel between a drone and a ground user can be either line-of-sight (LOS) or non-line-of-sight (LOS). Considering this communication channel as a probabilistic Loss-of-Stake (LoS) channel, meaning the channel operates with a LoS probability... Presenting line-of-sight propagation characteristics, with NLoS probability It exhibits non-line-of-sight propagation characteristics; for a given established ground user-UAV pair, assuming the ground user... by drone The service requires separate calculations of path loss in both LosS and NLoS states. , The model is as follows: ; Based on this, the average path loss can be calculated as a weighted sum of the Loss path loss and the NLoS path loss according to their respective probabilities. Wherein, and These represent the system losses of the access link's LoS and NLoS, respectively. For carrier frequency, Indicates ground users With drones The three-dimensional distance between them The speed of light; It can be calculated as follows: ; Since the access channel under consideration is a probabilistic Loss (LoS) channel, the LoS probability and NLoS probability are... , It can be represented as: ; in, These are constant parameters based on the environment, and are empirical values ​​that can be obtained by querying standard channel models, such as in a typical urban area. , ; It is a natural constant; For ground users With drones The angle of elevation between them; It can be calculated as follows: ; Available drones The average path loss of the corresponding access link is expressed as: ; Total interference It can be represented as: ; in, Indicates ground users Interference with drones Path loss between; , To interfere with drones The corresponding LosS probability and NLoS probability; , To interfere with drones The corresponding LosS path loss and NLoS path loss. For ground users If ground users by drone The service provided is by drones. This refers to service drones, while the remaining drones are for ground users. It can then be used to interfere with drones.

[0026] Assuming a drone The transmission power is Then ground users The signal-to-interference-plus-noise ratio (SINR) at a given location can be derived from the average path loss model: ; in For noise power, For noise power spectral density, For ground users The allocated communication bandwidth.

[0027] user With drones The actual throughput between them can be expressed as: ; In resource allocation problems in communication networks, linear utility functions typically aim to maximize the total throughput of ground users. To improve fairness, logarithmic utility functions can be used, with the following expression: ; in, For ground users The optimization algorithm allocates more resources to low-throughput ground users, thereby improving overall efficiency.

[0028] Based on the idea of ​​logarithmic utility functions, the objective function of the optimization problem is designed as the sum of the logarithms of the throughput of all ground users, i.e. This objective function is used in solving all subsequent subproblems.

[0029] To restore airborne network connectivity after a drone malfunction, ensure full coverage for ground users, and maintain fairness among ground users, an optimization problem was constructed, with the objective function expressed as follows: ; in, The objective function value is the sum of the logarithmic throughput of all users. The optimization problem represented by the objective function contains three optimization variables: drones. Two-dimensional position Ground users With drones The correlation between the indicator variables and the communication bandwidth allocated to each ground user If ground users With drones Establishing a connection between them, then It is 1 if it is not 0 otherwise; similar to the introduction of drones. With drones Connectivity indicators between ;when At that time, drones With drones Establishing a connection indicates that there is a valid communication link between the two parties, and vice versa.

[0030] The following describes several constraints in UAV-assisted ground user communication networks.

[0031] a. QoS constraints for ground users.

[0032] In UAV-assisted ground user communication networks, throughput is a key indicator for evaluating the QoS requirements of ground users; ground users The QoS constraints can be expressed as: ; in Represents the minimum throughput threshold for each ground user; constraints Regulations for ground users Only by drones It can only operate normally when the service is provided and the throughput threshold requirements are met.

[0033] b. Bandwidth constraints. To mitigate interference between ground users through bandwidth allocation, the bandwidth constraint is expressed as follows: ; in Represents the total transmission bandwidth of the UAV; constraints Ensure that the total allocated bandwidth for all ground users does not exceed the total bandwidth capacity of the system.

[0034] c. Ground user association constraints.

[0035] In UAV-assisted ground user communication networks, UAV malfunctions may lead to loss of ground user coverage, necessitating ground user re-association. The ground user re-association constraint can be expressed as: ; Among the constraints Ensure that each ground user is served by only one drone; constraints The maximum number of ground users served by each drone is limited to [number]. Constraints This ensures that all ground users receive full coverage service.

[0036] d. Connectivity constraints.

[0037] When a drone node fails in the inter-drone communication network, the network will be disconnected and fragmented; to restore network connectivity, the following connectivity constraints must be established: ; in, Represents a collection of drones Any non-empty subset, drone belong ; Indicates a collection of drones Remove from The set of remaining elements, drones belong Among them, constraints Ensure that for any method of partitioning UAV nodes, there is at least one connecting edge between the two partitioned parts, and constrain... Ensure that at least one exists in the communication network Edges are used to support the minimum spanning tree structure of the connected graph; together, they ensure the connectivity of the UAV network.

[0038] Constraints and This ensured the connectivity of the aerial drone network.

[0039] e. Relevance constraints.

[0040] Due to the connectivity indicators between drones Associated indicator variables Since all variables are binary, the correlation constraint is: ; Based on the above constraints and the designed objective function, the optimization problem of maximizing the total fair throughput of ground users can be expressed as:

[0041] Among the constraints Ensure that ground users are served by drones and meet the minimum throughput threshold for QoS guarantees, and constrain... This represents the bandwidth constraint for a single drone, limiting the total bandwidth allocated to all ground users to no more than [a certain value]. ;constraint , and Ensure full coverage, among which The requirement is that each ground user be served by only a single drone. Set the maximum number of ground users a single drone can serve as a limit. , Ensure all ground users are covered; constraints and Maintaining the connectivity recovery capability of the drone network after a failure; constraints This specifies the constraints on the correlation between drones and the correlation between drones and ground users.

[0042] The specific implementation steps of the method of the present invention are as follows: Step 1: Obtain the number of drones in the drone network. Number of ground users Drones act as aerial base stations to provide communication services to ground users; a minimum throughput threshold is set for ground users. Transmission power of each drone Total transmission bandwidth of drones constant parameters and Maximum communication distance between drones Iteration threshold Minimum communication distance between drones and maximum number of iterations Based on the constraints, the optimization problem of maximizing the total fair throughput for ground users can be expressed as: ; Step 2, set initial conditions, including two-dimensional position. ,power Bandwidth per user Iteration counting ,set up , , and initialize throughput ,in Indicates the iteration label.

[0043] Step 3: Use a three-layer iterative method to approximate the optimal solution of the original problem, which includes a user association layer, a bandwidth allocation layer, and a topology reconstruction layer; stop iterating when the convergence condition is met or the maximum number of iterations is exceeded.

[0044] (1) User Association Layer: Each UAV in the UAV network calculates the association indicator variable in the CPU. First, the optimization problem is simplified into a user association problem, which is represented as: ; The above subproblem is a MINLP problem. Since it is necessary to associate the UAV and the ground user, the subproblem can also be modeled as a maximum matching problem of a weighted bipartite graph.

[0045] Assume the weight set of the weighted bipartite graph constructed by the UAV-assisted ground user communication network is used It means that among them Indicates ground users From drone node The received data rate; if the user node No drone nodes If the service is specified, then the weight value will be set to a very large positive number (or...). To facilitate finding augmenting paths, the weight matrix is ​​first preprocessed. This is done by minimizing both rows and columns: subtracting the smallest element from each row to ensure each row has at least one zero, and subtracting the smallest element from each column to ensure each column also has at least one zero. In the preprocessed matrix, an initial matching is sought on the subgraph containing zero elements. This involves selecting several zeros from the rows and columns such that they form a non-conflicting matching (i.e., at most one zero is selected from each row and column). This step is achieved by finding the most probable matching in the bipartite graph. If the matching found at this point is a perfect matching (i.e., the matching contains...), then... The algorithm terminates when the initial match is imperfect (i.e., not covering all rows and columns). If the initial match is not perfect (i.e., not covering all rows and columns), an augmenting path needs to be found. This involves selecting unmatched vertices and attempting to find a path with zero elements that can increase the match size. If an augmenting path is found, the match is adjusted using that path, increasing the number of edges in the match. If no augmenting path is found (i.e., the current match cannot be expanded), the weights in the matrix need to be adjusted. Unmatched rows and columns are selected, and the smallest non-zero element in these rows and columns is calculated. This minimum value is subtracted from the unmatched row and added to the matched column. Through this adjustment, new zero elements may appear, providing opportunities for the next augmenting path and increasing the probability of finding an augmenting path. After weight adjustment, the search for a match continues. This process is repeated until a perfect match is found. This is used to find the association scheme between the drone and the ground user, i.e., the set of values ​​for the association indicator variable.

[0046] (2) Bandwidth allocation layer: When the two-dimensional position of the UAV is given (Network topology) and the relationship between drones and user nodes In this case, the optimization problem can be simplified to a bandwidth allocation problem, as follows: ; After proving that the objective function is concave, we can deduce from its concavity that the optimization problem satisfies certain conditions, allowing it to be solved using the KKT conditions. The KKT conditions are as follows: ; in, For partial derivative operators; It is the objective function of the bandwidth allocation subproblem; To constrain Lagrange multipliers, To constrain Lagrange multipliers.

[0047] The objective function is The objective function value at point and the two endpoint values ​​in the constraints , By comparing the objective function values, the optimal solution to the bandwidth allocation problem can be obtained, which is the final determination of the user. From drones Obtained bandwidth value:

[0048] in, The objective function value for the bandwidth allocation subproblem is the value obtained by substituting the candidate bandwidths into the objective function. The candidate bandwidths (internal solutions) are obtained from the KKT conditions; the optimal ground user is obtained by applying the KKT conditions. With drones Communication bandwidth allocated between .

[0049] (3) Topology Reconstruction Layer: After solving the above user association and bandwidth allocation subproblems, the association between UAV nodes and ground user nodes and the bandwidth allocated to each ground user are given. Therefore, the optimization problem can be written as a topology reconstruction problem: ; Due to the problem's scale and the complexity of its constraints, direct solutions are difficult. Therefore, we introduce two penalty functions and apply the penalty function method to solve the topology reconstruction problem for restoring the connectivity of UAV networks, thereby determining the network topology.

[0050] To ensure that the throughput of each user does not fall below a given threshold, for each ground user To ensure that the allocated communication bandwidth and communication quality are good enough, the following penalty function is designed: ; in, User With drones The actual throughput between This is the set minimum throughput threshold; if a user's throughput is below the threshold... If so, then a penalty will be introduced.

[0051] To ensure that the communication distance between drones does not exceed the maximum communication radius while maintaining network connectivity, the following penalty function is defined:

[0052] in Indicates drone and The distance between them and These are the maximum and minimum communication distances between drones; if the distance between two drones exceeds the maximum communication distance... or below the minimum communication distance If so, then a penalty will be introduced.

[0053] Two penalty functions and Add to the original objective function Then, the new overall objective function is obtained as follows:

[0054] in, and For the penalty function and The weighting coefficients can be empirical values ​​or adaptively adjusted values; for example, an initial value can be set and then adjusted through simulation or experimentation, such as when throughput is constrained. If there are many users violating the rules or the number of violations is large, the penalty will be increased. Conversely, if the constraints are satisfied, the weights can be maintained or reduced.

[0055] Next, we will use gradient ascent to solve this objective function. Gradient ascent is an iterative optimization method based on gradient information, which can be used to solve optimization problems involving continuous variables. Considering that the constraints in the original problem are embedded in the objective function, we can directly optimize the overall objective function.

[0056] Step 4: Repeat step 3 until the objective function value is obtained in the current iteration. Compared with the objective function value of the previous iteration The difference between them is less than the iteration threshold. Or reach the maximum number of iterations Output the objective function value obtained in the current iteration as the optimal objective function value. This represents the maximum total fair throughput for ground users, and also yields the corresponding optimal parameters, including the optimal two-dimensional position of each drone. Optimal correlation indicator variable and optimal communication bandwidth .

[0057] Based on the optimal two-dimensional position of the UAV obtained in step 4 Adjust the actual flight positions of each UAV; according to the optimal correlation indicator variable Establish access between ground users and drones; and based on optimal communication bandwidth By allocating communication bandwidth to each user, the topology of the drone network is reconstructed, enabling the reconstruction of connectivity in the air network and the restoration of coverage for ground users.

[0058] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for reconstructing the topology of an unmanned aerial vehicle (UAV) network, characterized in that, include: Acquire drone network information and ground user information; Among them, drones serve as aerial base stations to provide communication services to ground users; To address the need to restore UAV network connectivity after a UAV malfunction while ensuring full coverage and fairness for ground users, an optimization problem is constructed with the objective of maximizing the total fair throughput for ground users. The optimization variables in this problem are the two-dimensional position of each UAV, the association indicator variable between ground users and UAVs, and the communication bandwidth allocated to each ground user. The objective function is the sum of the logarithmic values ​​of the actual throughput of all ground users; the constraints of the optimization problem include: Ground user QoS constraints are used to ensure that ground users meet the minimum throughput threshold for QoS guarantees when served by UAVs; bandwidth constraints are used to limit the total bandwidth allocated to all ground users to not exceed the total transmission bandwidth of the UAVs; ground user association constraints are used to limit each ground user to be served by only a single UAV, limit the upper limit of the number of ground users served by a single UAV, and ensure that all ground users are covered; connectivity constraints are used to maintain the connectivity recovery capability of the UAV network after a failure; correlation constraints are used to restrict the connectivity indicators and correlation indicators between UAVs. A three-layer iterative method is used to approximate the optimal solution to the optimization problem, including a user association layer, a bandwidth allocation layer, and a topology reconstruction layer; when the convergence condition is met or the maximum number of iterations is reached, the optimal objective function value and the optimal optimization variables are output; Adjusting the actual flight position of the UAV according to the optimal optimization variables, establishing the access relationship between ground users and the UAV, and allocating communication bandwidth to each user, thereby completing the topology reconstruction of the UAV network.

2. The UAV network topology reconstruction method according to claim 1, characterized in that, The objective function is expressed as follows: in, The objective function value; For drones Two-dimensional position; For ground users With drones The correlation indicator variable between them, if ground users With drones Establishing a connection between them, then It is 1 if it is true, otherwise it is 0. For each ground user Allocated communication bandwidth; The number of drones; The number of ground users; For users With drones The actual throughput between them.

3. The UAV network topology reconstruction method according to claim 2, characterized in that, QoS constraints for ground users: ； in, This represents the minimum throughput threshold for each ground user; For ground users; For drones; Bandwidth constraints: in This indicates the total transmission bandwidth of the drone; Ground user association constraints: ； in The maximum number of ground users that can serve each drone; Connectivity constraints: ； in, Represents a collection of drones Any non-empty subset, drone belong ; Indicates a collection of drones Remove from The set consisting of the remaining elements; For drones With drones Connectivity indicators between; when At that time, drones With drones Establish a connection; otherwise, the two will not connect. Relevance constraints: 。 4. The UAV network topology reconstruction method according to claim 2, characterized in that, user With drones The actual throughput between Represented as: ； The communication channel between the UAV and the ground user is a probabilistic Loss-of-Stake (LoS) channel, and the ground user... by drone Services provided; For drones The transmission power; To interfere with drones The transmission power; , To interfere with drones The corresponding LosS probability and NLoS probability; , To interfere with drones The corresponding LoS path loss and NLoS path loss; For noise power spectral density, For ground users The allocated communication bandwidth.

5. The UAV network topology reconstruction method according to claim 3, characterized in that, The process for solving the user association layer is as follows: The optimization problem can be simplified into a user association problem, represented as: ； The aforementioned user association problem is modeled as a maximum matching problem in a weighted bipartite graph. To facilitate finding augmenting paths, the weight matrix is ​​preprocessed. An initial matching is then attempted on the subgraph containing zero elements. If this initial matching is perfect, the algorithm terminates. If the initial matching is imperfect, an augmenting path is further sought: unmatched vertices are selected, and a path containing zero elements is attempted to increase the matching value. If an augmenting path is found, the matching is adjusted by increasing the number of edges within the matching. If no augmenting path is found, the weights in the matrix are adjusted: unmatched rows and columns are selected, and the smallest non-zero element in these rows and columns is calculated. This minimum value is subtracted from the unmatched row and added to the matched column. After weight adjustment, the search for a matching continues until a perfect matching is found, resulting in the set of values ​​for the association indicator variable.

6. The UAV network topology reconfiguration method according to claim 3, characterized in that, The solution process for the bandwidth allocation layer is as follows: Given the two-dimensional position of the drone and the association between the drone and the user node In this case, the optimization problem can be simplified to a bandwidth allocation problem: ； The solution is obtained using the KKT conditions, which are as follows: ； in, For partial derivative operators; It is the objective function of the bandwidth allocation subproblem; To constrain Lagrange multipliers, To constrain Lagrange multipliers; The objective function is The objective function value at the point and the two endpoint values ​​in the constraint conditions , By comparing the objective function values, the optimal solution to the bandwidth allocation problem can be obtained: in, The objective function value for the bandwidth allocation subproblem; The candidate bandwidths are obtained using the KKT conditions; the optimal ground user is obtained by applying the KKT conditions. With drones Communication bandwidth allocated between .

7. The UAV network topology reconstruction method according to claim 3, characterized in that, The solution process for the topology reconstruction layer is as follows: ； Design the following penalty function: ； in, User With drones The actual throughput between; It is the set minimum throughput threshold; in Indicates drone and The distance between them and These are the maximum and minimum communication distances between drones, respectively. Two penalty functions and Add to the original objective function Then, the new objective function is obtained as follows: in, and For the penalty function and The weighting coefficients are determined; the new objective function is solved by gradient ascent.

8. The UAV network topology reconstruction method according to claim 1, characterized in that, Repeat the three-layer iterative method until the difference between the objective function value obtained in the current iteration and the objective function value in the previous iteration is less than the iteration threshold or the maximum number of iterations is reached. Output the objective function value obtained in the current iteration as the optimal objective function value, which is the maximum total fair throughput of the ground user. At the same time, obtain the corresponding optimal parameters, including the optimal two-dimensional position of each UAV, the optimal associated indicator variable, and the optimal communication bandwidth.

9. A terminal device, comprising a processor, a memory, and a computer program stored in the memory; characterized in that, When the processor executes a computer program, it implements the UAV network topology reconfiguration method according to any one of claims 1-8.

10. A computer-readable storage medium storing a computer program; characterized in that, When the computer program is executed by the processor, it implements the UAV network topology reconfiguration method according to any one of claims 1-8.

Images