Fully distributed drone cluster communication method, system, medium and product

Abstract

The present application relates to the field of drone communication technology, and specifically to a fully distributed drone cluster communication method, system, medium, and product, which includes: obtaining initial conditions and using an undirected graph to calculate the transmission information between the bth drone and the #imgabs0#th drone; it includes: defining an undirected graph, whose nodes represent drones; defining initial parameters based on the undirected graph; generating a solution model based on the initial parameters and constraints; calculating solution information based on the solution model; obtaining a new bth drone from the drone cluster, defining the transmission information of the previous bth drone as the initial condition, and calculating new transmission information; until the drones in the drone cluster are traversed. The fully distributed communication method in the present invention can achieve the purpose of decentralization and ensure communication quality while reducing the post-transmission signaling overhead.

Description

Technical Field

[0001] The present invention relates to the field of drone communication technology, and in particular to a fully distributed drone cluster communication method, system, medium and product. Background Art

[0002] When drones perform complex tasks such as wide-area reconnaissance, disaster monitoring, agricultural monitoring, low-altitude logistics, signal coverage, etc., they need large-scale drone clusters to work together. Therefore, higher requirements are currently placed on the communication capabilities of drone clusters.

[0003] However, current drone swarms still face some challenges in providing communication services to users. For example, drone swarm communication links based on high-frequency bands (such as millimeter waves, terahertz, etc.) are easily blocked by obstacles.

[0004] Currently, the traditional solution is to deploy more drones to enhance signal coverage and employ larger array antennas (MIMO) to compensate for severe propagation losses. However, as the number of drones in a swarm increases and the dimensionality of the MIMO channel increases, traditional centralized drone swarm communication systems must send all information to a ground station for processing. This results in a dramatic increase in backhaul signaling overhead and computational complexity, which is particularly noticeable in large-scale drone swarm communication scenarios.

[0005] In addition, patent application CN119341678A discloses an interference power allocation method based on multi-agent deep reinforcement learning, which includes: establishing a collaborative interference model of multiple interference devices against multiple UAV targets in a "many-to-many" communication confrontation scenario; constructing a distributed locally observable Markov decision process to model the interference power allocation problem as a multi-agent fully collaborative model; designing a "decentralized training, distributed decision-making" multi-agent framework to reduce the collaborative decision-making dimension of interference power resources; designing an extractor to pre-process the agent input state information to improve training efficiency; and realizing efficient exploration of power allocation strategies based on a discrete flexible actor-critic algorithm.

[0006] Patent application CN113316091A discloses a method and apparatus for transmitting information in drone clusters. This method addresses the issues of link instability and packet loss caused by frequent topological changes in drone clusters. The method calculates each drone's node degree based on its own attributes and the attributes of multiple neighboring drones. This node degree is then used to identify the drone cluster head and group them into clusters.

[0007] However, the applicant has noticed that these distributed communication methods have difficulty in achieving reliable communication in complex urban environments. Summary of the Invention

[0008] The purpose of the present invention is to provide a fully distributed UAV cluster communication method, system, medium and product, which partially solves or alleviates the above-mentioned deficiencies in the prior art, can reduce the loss in the information transmission process, and ensure the communication quality. Specifically, the present application provides a beamforming design scheme based on the alternating direction multiplier method for the sum rate optimization problem of the fully distributed UAV cluster millimeter wave MIMO communication system, and models the sum rate maximization problem of the centralized UAV cluster millimeter wave MIMO communication system into a distributed consistency optimization problem, so that only a small amount of back-transmission signaling exchange is required between adjacent UAVs to complete the local update and local calculation of the beamforming, and no longer requires information interaction with the ground station. Under the framework of the alternating direction multiplier method, the present invention uses a block gradient descent algorithm and a manifold optimization algorithm to optimize the design of the hybrid beamformer of the UAV and the synthesizer of the user respectively. Compared with the centralized communication system, the fully distributed UAV cluster millimeter wave MIMO communication system proposed by the present invention has basically the same sum rate performance, and can greatly reduce the computational complexity and back-transmission signaling overhead.

[0009] Specifically, the present invention provides a fully distributed UAV cluster communication method, comprising the steps of:

[0010] S1, obtain the initial conditions and use the undirected graph to calculate the bth UAV and the The method comprises the steps of: S11, defining an undirected graph for describing the communication relationship between the drones, wherein the nodes of the undirected graph represent the drones, and one node is connected to another node by at least one edge; S12, defining initial parameters according to the undirected graph, wherein the initial parameters include: channel, undirected graph, incidence matrix, signed incidence matrix, analog beamforming, digital beamforming, and user combiner; S13, defining constraints according to the initial parameters: ;in, 、 are user mergers for the b-th and non-b-th UAVs, for the drone cluster; S14, generating a solution model according to the initial parameters and the constraints; S15, calculating solution information according to the solution model, the solution information including: the optimal analog beamforming, digital beamforming, optimal merging vector of all users, and optimal sum rate of the b-th drone; S16, determining the transmission information according to the solution information; S2, obtaining a new b-th drone from the drone cluster, and defining the transmission information of the previous b-th drone as the initial condition; S3, returning to S1 until the drones in the drone cluster are traversed to complete the current round of communication.

[0011] In some embodiments, S15 includes: setting the relationship between the drone index b and the number of iterations t: , mod represents the modulo operator, B is the number of drones; the bth drone receives the bth drone through the backhaul link. Signaling information sent by drones and ; Update auxiliary variables generated during fraction transformation and auxiliary variables ,in, Represents a user set; the b-th drone transmits signaling information after updating: and ;in, is the complex number at the t+1th iteration , is the complex number at the t-th iteration, 、 is the conjugate transpose and digital beamforming vector of the combiner of the b-th UAV for the j-th user at the t-th iteration, For the channel, is the simulated beamforming at the tth iteration; is the second subsequent command information that does not include the b-th UAV in the t+1th iteration, is the combiner of all users of the b-th drone at the t-th iteration, is a complex matrix; the block gradient descent algorithm is used to optimize the hybrid beamforming matrix of the b-th UAV , simulate the beamforming matrix , digital beamforming matrix ; The b-th drone updates the combiner vector of K users ; Update the dual variables generated during the alternating direction multiplier method ; The bth drone transmits signaling information after updating: and , and sent to adjacent UAVs; after the solution process converges, the output solution information includes: the optimal analog beamforming, digital beamforming, optimal merging vector of all users, and optimal sum rate of the b-th UAV.

[0012] In some embodiments, S11 includes: constructing an undirected graph based on the connection relationship between the drones , wherein the undirected graph has B nodes, and the nodes represent the drones, and the edges Represents the backhaul link between the UAVs; The association matrix Defined as The incidence matrix of The rows and columns correspond to The nodes and edges of Convert to a signed incidence matrix , and for the e-th connection, given edge , to Defined as: ; Wherein, b is the index of the drone, is a complex space.

[0013] In some embodiments, S11 also includes: calculating the degree of communication obstruction of at least one communication partition, wherein a communication partition includes: at least one node; selecting the number of edges connected to the node according to the degree of communication obstruction; and generating or updating the undirected graph according to the number of edges.

[0014] In some embodiments, S11 also includes: when the communication barrier degree of the communication partition is greater than a preset first level, the first node in the communication partition is connected to a second number of second nodes using a second number of edges; and the second node is selected from the communication partition where the communication barrier degree is less than the preset second level.

[0015] In some embodiments, the communication partition includes multiple nodes, and correspondingly, S11 also includes: setting a first sub-number of edges connecting a first node located in a central area of ​​the communication partition; setting a second sub-number of edges connecting a first node located in an edge area of ​​the communication partition, and the first sub-number is greater than or equal to the second sub-number.

[0016] In some embodiments, solving the model includes: ;

[0017] in, is the objective function, 、 For the analog beamforming and digital beamforming of the b-th UAV, is the transmission power of the b-th UAV, is the element in row i1 and column j1 of the simulated beamforming matrix of the b-th UAV, i1 and j1 are the row and column numbers of the elements in the matrix respectively. The i2th element in the user merge vector calculated locally for the bth drone, where i2 is the sequence number of the element in the vector. , , , 、 These are the two types of auxiliary variables introduced accordingly.

[0018] The present invention also provides a fully distributed UAV cluster communication method, comprising the steps of:

[0019] Decentralized module, used to obtain initial conditions and calculate the bth drone and the bth drone using an undirected graph Transmission information between drones; wherein the decentralized module includes: a graph definition unit, used to define an undirected graph for describing the communication relationship between multiple drones, wherein the nodes of the undirected graph represent the drones, and one node is connected to another node through at least one edge; a parameter definition unit, used to define initial parameters according to the undirected graph, wherein the initial parameters include: channel, undirected graph, incidence matrix, signed incidence matrix, analog beamforming, digital beamforming, and user combiner; a constraint definition unit, used to define constraint conditions according to the initial parameters: ;in, 、 are user mergers for the b-th and non-b-th UAVs, The drone cluster comprises: a solution definition unit for generating a solution model according to the initial parameters and the constraint conditions; a solution unit for calculating solution information according to the solution model, wherein the solution information includes: the optimal analog beamforming, digital beamforming, optimal merging vector of all users, and optimal sum rate of the b-th drone; a result unit for determining the transmission information according to the solution information; an update module for obtaining a new b-th drone from the drone cluster, defining the transmission information of the previous b-th drone as the initial condition, and entering a decentralized solution module, a traversal module, and returning to the decentralized module to traverse the drones in the drone cluster to complete the current round of communication.

[0020] The present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the method according to any one of the embodiments is implemented. The present invention further provides a computer program product including the computer program, wherein when the computer program is executed by a computer, the method according to any one of the embodiments is implemented.

[0021] Beneficial Technical Effects: It is worth noting that in traditional centralized communication methods, all drones need to communicate with a single ground base station. In contrast, this invention provides a distributed communication method (also known as a decentralized communication method) that allows for direct, two-way communication without relying on a ground base station.

[0022] Especially in complex scenarios with dense high-rise buildings (such as city centers and high-density residential areas), drone swarms may experience obstruction from high-rise buildings, hindering communication between drones and ground stations. This significantly increases the difficulty of drone communication. To address this, this application can utilize a two-to-two communication method between drones to reduce reliance on ground base stations.

[0023] Specifically, this embodiment provides a solution path defined based on the backhaul link undirected graph, and uses the user combiner as a guiding constraint to ensure that pairwise communication does not cause conflicts. Among them, in order to ensure that the drones complete the overall signal update of the drone cluster through pairwise communication, this application focuses on providing a method that can quickly iterate the communication scheme of the drone cluster based on limited specific factors (such as drone beamforming, user combiner) and with user combiner consistency as the dominant constraint, thereby achieving the purpose of complete decentralization. In particular, this application simplifies the iterative process at least at the following levels:

[0024] 1) With the help of the undirected graph defined by the UAV communication links, the consistency constraints in the beamforming optimization process are simplified. This constraint focuses on utilizing the local information of the UAV (such as ), the capabilities of drone swarms (e.g. The complex matrix of dimension is used to constrain the communication results between two UAVs during the iteration process, so that when the communication results between two UAVs meet the constraint conditions, it means that the handshake between two UAVs is successful.

[0025] 2) Restrictively optimize the communication content between adjacent UAVs. Specifically, during the iterative solution process, the bth UAV only needs to consider its own specific information (such as its own channel information) and the post-transmission information updated by the previous UAV. Furthermore, by limiting the communication content between adjacent UAVs and the dimension of the UAV's local computing content, the computational pressure of the iterative process can be reduced and the iteration efficiency can be improved.

[0026] Furthermore, the present invention uses differential communication density to set or optimize undirected graphs, which can improve the communication capabilities of the decentralized communication method in complex communication environments without causing excessive communication pressure. This is especially true in "low-altitude economy" applications, such as deploying a large number of drones for food delivery in a city. This can pose significant challenges to drone communication in densely populated city centers. The decentralized communication solution based on differentiated communication density in this embodiment can, to a certain extent, reconcile the conflict between communication capabilities and communication pressure.

[0027] Furthermore, for nodes in different areas such as the center and periphery of the same communication partition, differentiated communication density can also be used for communication to ensure that nodes with higher local communication difficulty attempt to communicate with multiple nodes, and the reliability of communication information solution of a single node can be improved through multi-party verification.

[0028] Especially when large-scale drone clusters are performing flight missions in dispersed locations within urban buildings, this differentiated communication density design in local areas can not only enhance the anti-interference communication capabilities of local drones, but also prevent changes in communication density from placing excessive communication pressure on the drone clusters.

[0029] From another perspective, this application chooses a completely decentralized approach, and only completes the signal update of the entire drone cluster through communication between adjacent drones (that is, an adjacent communication mechanism). In order to improve the transmission efficiency and accuracy under this adjacent communication mechanism, this application focuses on selecting limited dual factors such as drone local information and adjacent drones' back-transmitted signaling information to centrally optimize the drone's beamforming during the iteration process. Therefore, on the one hand, this application can use drone local information, that is, the back-transmitted signaling information of adjacent drones (such as the previous drone), to perform collaborative calculations to ensure that the adjacent communication process is correct during the iteration process (to avoid information loss); at the same time, based on local information and the back-transmitted signaling information of adjacent drones, the current beamforming is optimized and solved, and then the optimal communication solution is found through a cyclic iterative process.

[0030] It is also worth noting that through simulation tests, it can be seen that the method adopted in this application for rapid iteration of the communication scheme of the drone cluster based on limited specific factors can not only avoid the interference of obstacles on the communication system, but also the communication performance is almost the same as that of the centralized communication scheme, that is, the communication performance will not be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the embodiments or the description of the prior art. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the various elements or parts are not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without inventive work.

[0032] Figure 1a 1 is a flow chart of a communication method according to an exemplary embodiment of the present invention;

[0033] Figure 1b Schematic diagram of a communication method flow in a preferred embodiment of the present invention;

[0034] Figure 2 Schematic diagram of the centralized communication mode of the UAV cluster millimeter wave MIMO communication system;

[0035] Figure 3Schematic diagram of the distributed communication mode of the UAV cluster millimeter wave MIMO communication system;

[0036] Figure 4 A comparison chart of the sum rate performance of centralized and fully distributed UAV swarm millimeter-wave MIMO communication systems;

[0037] Figure 5 FIG. 4 is a schematic diagram of a communication system module structure in an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In this article, the use of suffixes such as "module", "component" or "unit" to represent elements is only to facilitate the description of the present invention and has no specific meaning in itself. Therefore, "module", "component" or "unit" can be used in a mixed manner. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0039] As used herein, "and / or" includes any and all combinations of one or more listed items. As used herein, "multiple" means two or more, i.e., it includes two, three, four, five, etc. As used herein, the term "approximately" is typically expressed as + / -5% of the value, more typically + / -4% of the value, more typically + / -3% of the value, more typically + / -2% of the value, even more typically + / -1% of the value, even more typically + / -0.5% of the value. In this specification, certain embodiments may be disclosed in a format that is within a certain range. It should be understood that this description of "being within a certain range" is merely for convenience and brevity and should not be interpreted as a rigid limitation on the disclosed range. Therefore, the description of the range should be considered to have specifically disclosed all possible sub-ranges and independent numerical values ​​within this range. For example, description of the range 1-6 should be considered to have specifically disclosed sub-ranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., as well as individual numbers within that range, such as 1, 2, 3, 4, 5, and 6. The above rules apply regardless of the breadth of the range.

[0040] Example 1: See Figure 1a and Figure 1b As shown, the present invention provides a fully distributed UAV cluster communication method, comprising the steps of:

[0041] S1, obtain the initial conditions and use the undirected graph to calculate the bth UAV and the The information transmitted between the drones; The bth drone refers to the drones in the drone cluster except the bth drone.

[0042] For example, a drone cluster includes: a group of drones, and the group of drones communicates with a group of users, wherein the group of drones Included drones, the group of users include users, and each of the drones is equipped with N array antennas and N RF RF chains, each of the users is equipped with M array antennas and 1 RF chain; correspondingly, the initial conditions may include the above-mentioned parameters for describing the communication relationship of the drone cluster, such as the number of drones, the number of users, the number of array antennas, the number of RF chains, etc.

[0043] Wherein, S1 comprises the steps of:

[0044] S11, defining an undirected graph for describing communication relationships between drones, wherein nodes of the undirected graph represent the drones, and a node is connected to another node via at least one edge;

[0045] That is to say, the present invention chooses to use the undirected graph formed by the backhaul links between the drones (or called the backhaul link undirected graph).

[0046] S12, defining initial parameters according to the undirected graph, the initial parameters including: a channel, an undirected graph, an incidence matrix, a signed incidence matrix, an analog beamforming, a digital beamforming, and a user combiner;

[0047] Here, the signal refers to the channel between the drone and the user. In some embodiments, the initial parameters can be obtained by obtaining the initial conditions.

[0048] In this embodiment, the undirected graph, the incidence matrix, and the signed incidence matrix are determined by the undirected graph defined by the backhaul link between the UAVs, and the analog beamforming, the digital beamforming, and the user combiner can have initial parameters obtained from the initial conditions.

[0049] S13, defining constraints based on the initial parameters: ;in, 、 are user mergers for the b-th and non-b-th UAVs, The drone cluster;

[0050] Among them, the constraint condition is used to force two UAVs (such as the bth UAV and the The communication between the two drones (such as the user merger) is consistent, that is, the user information (such as the user merger) calculated by the two is consistent.

[0051] For example, user information could be the number of enabled array antennas (which can be described or defined by a user combiner). In other words, the drone is effectively providing guidance to the user on how to configure antenna usage details. The constraints in this embodiment ensure that the guidance output by the two drones is consistent.

[0052] S14, generating a solution model according to the initial parameters and the constraint conditions;

[0053] Among them, according to the constraints, the centralized system optimization problem is converted into a distributed communication system optimization problem, that is, a decentralized optimization problem. Decentralization means that each drone does not need to communicate with a central base station, but can communicate with adjacent drones in pairs to complete the communication update in the entire drone cluster.

[0054] For example, in some embodiments, the generated solution model is:

[0055] ;

[0056] in, It is the objective function set by the user for the UAV cluster communication system. 、 For the analog beamforming and digital beamforming of the b-th UAV, is the transmission power of the b-th UAV, is the element in row i1 and column j1 of the simulated beamforming matrix of the b-th UAV, i1 and j1 are the row and column numbers of the elements in the matrix respectively. The i2th element in the user merge vector calculated locally for the bth drone, where i2 is the sequence number of the element in the vector. , , , 、 These are the two types of auxiliary variables introduced accordingly.

[0057] S15, calculating solution information (or optimal solution) according to the solution model, wherein the solution information includes: optimal analog beamforming, digital beamforming, optimal combining vectors of all users, and optimal sum rate for the b-th UAV;

[0058] S16, determining the transfer information according to the solution information;

[0059] For example, the current solution information can be used as a transfer message to the next UAV (i.e., the new b-th UAV) to be used as the initial conditions for the next round of calculations. Alternatively, the transfer message can also include other conditional parameters of the current b-th UAV after calculating the optimal solution. Specifically, the data type can be selected and read based on the actual communication solution process.

[0060] S2, obtaining a new b-th drone from the drone cluster, and defining the transmission information of the previous b-th drone as the initial condition;

[0061] S3, return to S1 until the drones in the drone cluster are traversed to complete the current round of communication.

[0062] It is worth noting that traditional centralized communication methods require all drones to communicate with a single ground base station. In contrast, the present invention provides a distributed communication method (also known as a decentralized communication method) that allows for direct, two-way communication without relying on a ground base station.

[0063] Specifically, this embodiment provides a solution path definition based on an undirected graph of backhaul links, and uses a user combiner as a guiding constraint to ensure that no conflicts occur in pairwise communications.

[0064] In some embodiments, S15 includes:

[0065] Set the relationship between drone index b and iteration number t: , mod represents the modulo operator, B is the number of drones;

[0066] The bth UAV receives the Signaling information sent by drones and ;

[0067] Update auxiliary variables generated during fractional transformation and auxiliary variables ,in, Represents a user set (or user group);

[0068] The bth drone transmits signaling information after updating: and ;in, is the complex number at the t+1th iteration , is the complex number at the t-th iteration, 、 is the conjugate transpose and digital beamforming vector of the combiner of the b-th UAV for the j-th user at the t-th iteration, For the channel, is the simulated beamforming at the tth iteration; is the second subsequent command information that does not include the b-th UAV in the t+1th iteration, is the combiner of all users of the b-th drone at the t-th iteration, is a complex matrix;

[0069] Optimize the hybrid beamforming matrix of the b-th UAV using the block gradient descent algorithm , simulate the beamforming matrix , digital beamforming matrix ;

[0070] The b-th drone updates the combiner vector of K users ;

[0071] Update the dual variables generated during the alternating direction multiplication method ;

[0072] The bth drone transmits signaling information after updating: and , and send it to adjacent drones;

[0073] After the solution process converges, the output solution information includes: the optimal analog beamforming and digital beamforming of the b-th UAV, the optimal merging vector of all users, and the optimal sum rate.

[0074] In summary, this application proposes a pairwise communication mechanism to achieve a completely decentralized communication mode. In addition, on the one hand, this application limits the communication content between two drones and simplifies the constraints of the iterative process to improve the computational efficiency of the iterative process while ensuring the effectiveness of the iteration; on the other hand, through the above-mentioned limited communication content and constraints, the beamforming of the drone cluster is quickly optimized, thereby making specific improvements to the form of the communication signal to ensure that the final beamforming result can effectively improve the throughput (or sum rate) of the communication system under the pairwise communication mechanism, while ensuring that the necessary communication content under the pairwise communication mechanism can be accurately transmitted.

[0075] In some embodiments, S11 includes:

[0076] Construct an undirected graph based on the connection relationship between drones , wherein the undirected graph has B nodes, and the nodes represent the drones, and the edges represents the backhaul link between the UAVs;

[0077] The incidence matrix Defined as The incidence matrix of The rows and columns correspond to The nodes and edges of Convert to a signed incidence matrix , and for the e-th connection, given edge , to Defined as:

[0078] ; Wherein, b is the index of the drone, is a complex space.

[0079] In some embodiments, S11 further includes:

[0080] (1) calculating a communication barrier degree of at least one communication partition, wherein a communication partition includes: at least one node;

[0081] For a drone swarm with B drones, an undirected graph with B nodes is generated. This undirected graph can be divided into multiple communication partitions, and each communication partition can have one or more nodes. The communication barrier level is used to define the communication quality or communication difficulty of a communication partition.

[0082] For example, in some embodiments, the degree of communication obstruction can be represented by the number of communication obstructions within a communication zone. For example, a communication zone may contain multiple high-rise buildings, which may cause communication obstruction between two drones. Accordingly, the level of communication obstruction can be defined based on the number of communication obstructions, such as level 3, level 2, level 1, and so on.

[0083] (2) selecting the number of edges connected to the node according to the degree of communication barrier;

[0084] For example, in some embodiments, the higher the communication barrier level of a node's communication zone, the more edges it can be connected to (edges are used to indicate that two drones can establish communication), and different communication zones can also adopt different communication relationships. In other words, the higher the communication barrier level, the greater the communication density of the corresponding communication zone.

[0085] (3) Generate or update the undirected graph according to the number of edges.

[0086] Preferably, in this embodiment, the communication barrier degrees of different communication partitions of an undirected graph may be different, and therefore different communication densities may be set for different communication partitions.

[0087] In summary, the use of differentiated communication relationships in this embodiment to set or optimize an undirected graph can improve the communication capabilities of this decentralized communication method in complex communication environments while simultaneously preventing excessive communication pressure. This is particularly true in "low-altitude economy" applications, such as deploying large numbers of drones for food delivery in densely populated city centers, which can pose significant challenges for drone communication. The decentralized communication solution in this embodiment, based on differentiated communication density, can, to a certain extent, reconcile the conflict between communication capabilities and communication pressure.

[0088] In some embodiments, S11 further includes:

[0089] When the communication barrier degree of the communication partition is greater than a preset first level, the first nodes in the communication partition are connected to the second number of second nodes using a second number of edges; and the second nodes are selected from the communication partitions whose communication barrier degree is less than the preset second level.

[0090] For example, in some embodiments, a first node can be connected to a first number of first nodes using a first number of edges, and at the same time be connected to a second number of second nodes using a second number of edges, that is, the first node can communicate with near nodes (i.e., nodes within the current partition) and far nodes (i.e., nodes in external partitions) at the same time to improve communication reliability.

[0091] For example, in some embodiments, the current communication zone can be used as the first zone, and another communication zone (e.g., a zone adjacent to the first zone) can be selected as the second zone. The first zone includes multiple first nodes, and the second zone includes multiple second nodes. If the communication barrier in the first zone is high, it is preferable to increase communication with the second node outside the first zone.

[0092] In some embodiments, the communication partition includes multiple nodes, and correspondingly, S11 further includes:

[0093] Setting a first subset of edges connecting a first node located in a central area of ​​the communication partition;

[0094] A second sub-number of edges is set for a first node located in an edge area (or a peripheral area) of the communication partition, and the first sub-number is greater than or equal to the second sub-number.

[0095] Preferably, for nodes in different areas such as the center and periphery of the same communication partition, differential communication density can also be used for communication to ensure that nodes with higher local communication difficulty attempt to communicate with multiple nodes, and the reliability of communication information resolution of a single node can be improved through multi-party verification.

[0096] For example, in some embodiments, when multiple nodes are connected sequentially, the edges of the geometric figures formed by the connections can be identified as edge regions (e.g., nodes on the outermost lines can be identified as edge nodes), and correspondingly, the remaining regions can be identified as central regions.

[0097] In some embodiments, the division range of the central area and the edge area can be adjusted according to the number of drone clusters or the size of the communication partition.

[0098] In order to more clearly demonstrate the communication optimization process for decentralized design in the present invention, we will try to obtain the initial conditions for S1 and use an undirected graph to calculate the bth drone and the bth drone. The information transmission between the two drones is described in detail, which includes the following steps:

[0099] S101, providing a communication system, the communication system comprising: a group of drones, wherein the group of drones communicates with a group of users, wherein the group of drones includes drones, the group of users includes users, and each of the drones is equipped with N array antennas and N RF RF chains, each of the users is equipped with M array antennas and 1 RF chain;

[0100] S102, Definition An undirected graph is used to represent the communication relationship of the communication system, wherein the undirected graph has B nodes, and the nodes represent the drones, and the edges Represents the backhaul link between the UAVs. When the original correlation matrix Defined as When the incidence matrix is The rows and columns correspond to The nodes and edges of Convert to a signed incidence matrix , and for the e-th connection, given edge , to Defined as:

[0101] ; Wherein, b is the index of the drone, is a complex space;

[0102] S103, solving a fully distributed UAV cluster communication model based on the initial parameters of the communication system; wherein the initial parameters include: the channel between the UAV and the user , the undirected graph formed by the backhaul links between the drones , correlation matrix , correlation matrix , initial iteration number , initial simulation beamforming , initial digital beamforming , Initial User Merger ;

[0103] Typically, drones, acting as transmitters, use beamforming (also known as beamforming or spatial filtering), a signal processing technique that uses an array antenna to transmit and receive signals in a directional manner. Specifically, it focuses the beam by adjusting the phase and amplitude of each antenna element. Users, acting as receivers, use combining technology (also known as a combiner) for signal processing. Similarly, they focus the beam by adjusting the phase and amplitude of the transmitted signal.

[0104] The specific development of the fully distributed UAV cluster communication model (referred to as the solution model) also includes:

[0105] (1) Constraints:

[0106] ;

[0107] in, The matrix defined , for The identity matrix of dimension , for A complex matrix of dimension , for A complex matrix of dimension and satisfying the condition , for A complex vector of dimension , is a combiner of all users calculated locally by the b-th drone; it can be understood that the constraint conditions in this embodiment are the specific mathematical expressions of the constraint conditions in S13.

[0108] (2) Alternating update calculation conditions set for solving the model:

[0109] ;

[0110] ;

[0111] ;

[0112] ;

[0113] ;

[0114] ;

[0115] in, is the tth iteration, is the transpose operation, is the conjugate transpose operation, To simulate beamforming, For digital beamforming, is a complex space, and are the analog beamforming and digital beamforming matrices of the b-th UAV, respectively, and , is a matrix The Kth column vector of is the combined vector of all users, The combiner for all users computed locally for the b-th drone, is the (t+1)th iteration , The combiner for all users computed locally for non-b-th drone, is the tth iteration gather, and are the first auxiliary variable and the second auxiliary variable generated in the fractional programming transformation process, is the dual variable at the t-th iteration, is the hybrid beamforming matrix of the b-th UAV, is the beamforming matrix set at the tth iteration, is the beamforming matrix set at the (t+1)th iteration, is the beamforming matrix set of the b-th UAV, is the (t+1)th iteration , is the beamforming matrix set of non-b-th UAV, is the tth iteration , is the penalty coefficient, is the communication rate weight of the kth user, To take the real part of a complex number, The second auxiliary variable The conjugate transpose of the kth element of is the conjugate transpose of the combined vector of the kth user, is the millimeter wave channel between the b-th UAV and the k-th user, and , is the digital beamforming vector of the b-th UAV for the k-th user, is the digital beamforming vector of the b-th UAV for the j-th user, is the noise variance, is the augmented Lagrangian function, is the objective function; is the Lagrange multiplier used by the b-th drone, is a custom parameter, and , is a matrix The Frobenius norm of ; and represents an indicator function, for example, if , Equal to 0, otherwise .

[0116] S104: After the solution process of S103 converges, the current calculated result is output as the optimal condition. The optimal condition includes: 、 、 、 ;in, is the optimal simulated beamforming matrix for the b-th UAV, is the optimal digital beamforming matrix for the b-th UAV, Calculate the optimal merge vector of all users locally for the b-th drone, is the optimal sum rate for the communication system.

[0117] This embodiment actually proposes a beamforming design method in a fully distributed UAV cluster millimeter wave MIMO communication system, which improves the throughput of the communication system by designing the hybrid beamformer of the UAV and the synthesizer of the user.

[0118] For example, consider a fully distributed downlink UAV swarm mmWave MIMO communication scenario, where a group of UAVs With a group of users Each UAV is equipped with N array antennas and N RF RF chains, and each user is equipped with M array antennas and 1 RF chain. Since the communication system is fully distributed, there is no need for a ground station to perform coordination and computing tasks between drones. The signaling information shared between drones is transmitted through the backhaul link, and all baseband signal processing is completed locally on the drone. Taking into account the constraints of hardware cost and power consumption, each drone adopts a hybrid beamforming structure to meet and At this time, the signal received by the kth user is: ;

[0119] in, Indicates that the kth user obeys The distributed additive white Gaussian noise, represents the millimeter wave channel between the b-th drone and the k-th user, represents the symbol of the jth user, represents the symbol vector of K users, represents the simulated beamforming of the b-th UAV, represents the digital beamforming of the b-th UAV, represents the k-th user utilizing the merged vector.

[0120] Furthermore, assume that each UAV sends K data streams to all users, and the transmitted signal from the b-th UAV can be expressed as: ;in represents a symbolic vector and satisfies , for The identity matrix of dimension . and They represent the analog beamforming matrix and digital beamforming matrix of the b-th UAV respectively. The transmission power of the b-th UAV satisfies .

[0121] According to the UAV cluster millimeter wave MIMO communication model, the received signal of the kth user is the superposition of the signals transmitted by B UAVs, which can be expressed as ;in, Indicates that the kth user obeys The distributed additive white Gaussian noise, represents the millimeter wave channel between the bth UAV and the kth user. Due to the sparse characteristics of the millimeter wave channel, the present invention adopts the Saleh-Valenzuela channel model with a small number of scattering paths. In order to simplify the notation, a uniform linear array is used to represent Defined as: ;in, is the number of propagation paths, is the complex gain of the lth path between the bth UAV and the kth user. In particular, In the high frequency band, it is mainly determined by propagation attenuation and molecular absorption. In addition, and represents the arrival azimuth and departure azimuth of the lth path. and represent the antenna array response vectors of the b-th UAV and the k-th user respectively.

[0122] Then, the kth user uses the merged vector To further process the received signal, we can get ;

[0123] The final received signal of the kth user is It is divided into three parts: desired signal, interference signal and noise.

[0124] Based on the above formula, the signal-to-interference-and-noise ratio of the kth user can be expressed as: ;

[0125] in represents the effective noise variance, is the original noise variance. Given the signal-to-interference-noise ratio, the sum rate of the UAV cluster millimeter wave MIMO communication system can be written as: ;

[0126] in represents the WSR weight and satisfies the constraint .

[0127] Furthermore, for the fully distributed UAV cluster millimeter wave MIMO communication system, to ensure that all UAVs reach a consensus in the combiner of K users. Specifically, for the consistency constraint, set it to satisfy the condition ,in represents the combiner of K users calculated locally by the b-th drone. Therefore, the optimization problem (i.e., the solution model) is set as:

[0128]

[0129] in and .

[0130] It is worth noting that in the centralized UAV cluster millimeter wave MIMO communication system, the UAV cluster and the ground station need to conduct a large amount of information interaction, especially with the increase in the number of UAVs and the increase in the MIMO channel dimension, which will lead to a sharp increase in the backhaul signaling overhead and computational complexity.

[0131] To solve this problem, the present invention proposes a fully distributed UAV cluster communication system. The alternating direction multiplier method is used to transform the centralized communication system optimization problem into a distributed consistency optimization problem. Then, the block gradient descent algorithm and the popular optimization algorithm are used to optimize the UAV's beamforming matrix and the user's merging vector respectively, ultimately maximizing the sum rate of the communication system.

[0132] Furthermore, in order to illustrate the solution process of obtaining the solution information by calculating according to the solution model, the solution process of step S102 will be described in detail below:

[0133] For example, in some embodiments, S102 includes the following steps: S201, obtaining initial parameters in the current communication state, the initial parameters including: the channel between the UAV and the user; , the undirected graph formed by the backhaul links between the drones , the incidence matrix corresponding to the undirected graph , initial iteration number , initial simulation beamforming , initial simulation beamforming , Initial User Merger ;

[0134] S202, when the number of iterations When , steps S203-S216 are executed in a loop;

[0135] S203, set the relationship between drone index and iteration number: ;in, is the drone index, mod represents the modulo operator;

[0136] S204: The bth UAV receives the first post-transmission signaling information and the second post-command information ,in, is the tth iteration Dimensional signaling matrix , is the tth iteration Dimensional signaling vector ;

[0137] S205: Update the first auxiliary variable: ;

[0138] S206, updating the second auxiliary variable: ;

[0139] S207, the bth drone transmits signaling information after updating: ,in, is the complex number at the (t+1)th iteration and satisfy , is a complex number at the tth iteration and satisfies , is the conjugate transpose of the combiner of the b-th drone for the j-th user at the t-th iteration, is the simulated beamforming at the tth iteration, For the The conjugate transpose of the combiner of the UAV for the jth user at the tth iteration, is the digital beamforming vector of the b-th UAV targeting the j-th user at the t-th iteration;

[0140] S208, the bth drone transmits signaling information after updating: , is the second subsequent command information that does not include the b-th UAV in the t+1th iteration, is the combiner of all users of the b-th drone at the t-th iteration;

[0141] S209, the b-th UAV updates the hybrid beamforming matrix: , is the hybrid beamforming matrix at the t+1th iteration;

[0142] S210, the b-th UAV updates the simulated beamforming matrix: , is the simulated beamforming matrix at the t+1th iteration;

[0143] S211, according to the optimized variable set Update the digital beamforming matrix for the b-th UAV: , is the digital beamforming matrix at the t+1th iteration;

[0144] S212: The b-th UAV updates the combiners of all users based on the updated hybrid beamforming matrix, analog beamforming matrix, and digital beamforming matrix: , is the user merger of the b-th UAV at the t+1th iteration;

[0145] S213, update auxiliary variables: ;

[0146] S214, the bth drone transmits signaling information after updating: ;

[0147] S215, the bth drone transmits signaling information after updating: ;

[0148] S216, sending updated post-transmission signaling information and to adjacent drones;

[0149] S217, calculating the sum rate of the current communication system The solution is considered converged if the difference between the current sum rate and the sum rate of the previous iteration is less than the set difference degree. The sum rate is the total amount of data that the communication system can transmit per unit time. The sum rate refers to the sum of the communication rates between all drones and all users, which is the communication capacity of the entire system.

[0150] It is worth noting that the distributed technology proposed in this application can not only compress the data dimension and data volume of information transmission, but also maintain excellent speed performance during the transmission of limited information.

[0151] Specifically, this application employs a fully distributed transmission scheme to achieve decentralization, eliminating the need for communication between drones and ground stations. Furthermore, through distributed consistency optimization, this application enables adjacent drones to complete local updates and calculations of beamforming with minimal backhaul signaling, maximizing communication capacity.

[0152] In some embodiments, the communication system meets the following requirements: , .

[0153] In some embodiments, the first auxiliary variable and the second auxiliary variable are calculated using the following model: ; ;

[0154] ; ;

[0155] and ;

[0156] in, After optimization for the kth user , is the signal-to-interference-and-noise ratio of the kth user, After optimization for the kth user , is the conjugate transpose of the combiner of the b-th drone for the k-th user, First post-transmission signaling information The k-th row and k-th column element of First post-transmission signaling information The k-th row and j-th column element of .

[0157] To further solve the fully distributed hybrid beamforming problem, the bth UAV can also optimize multiple beamforming matrices locally (such as 、 and ), the present invention uses the BCD algorithm to design analog beamforming and digital beamforming for the UAV, and the specific implementation process is as follows.

[0158] 1) Optimization : Given a fixed set of variables , so that about The augmented Lagrangian function is defined as ;in, is the hybrid beamforming matrix of the b-th UAV, is the Lagrange multiplier used by the b-th drone, is the transmission power of the b-th UAV, is the penalty coefficient; its Karush-Kuhn-Tucker (KKT) condition can be expressed as:

[0159] , , , ;

[0160] in, is a matrix The Frobenius norm of ;

[0161] After algebraic operations, The closed-form solution can be expressed as: ;

[0162] 2) Optimization : Considering and is fixed, and the minimization problem of the simulated beamforming design for the b-th UAV is defined as: ;

[0163] Among them about The new objective function is expressed as:

[0164] ;

[0165] in, for .

[0166] For non-convex problems under constant modulus constraints, popular optimization algorithms can be used to solve them. The constraints are mapped onto a smooth Riemannian manifold and the following equations are introduced: , the above optimization problem can be rewritten as:

[0167] ;in,

[0168] ;

[0169]

[0170]

[0171] ;

[0172] in, is a matrix Column vectorization of For vector The objective function, is a vector The conjugate transpose of is a matrix The conjugate transpose of is a matrix The conjugate transpose of is a vector The conjugate of is a vector The transpose of for , is a matrix The transpose of for The identity matrix of dimension , is a matrix Column vectorization of ;

[0173] Furthermore, It is defined as a complex circular Riemann manifold, which can be written as

[0174] ;in, is a vector The first element of is a vector No. elements;

[0175] For a given point , the Riemann gradient The Euclidean gradient On a Riemannian manifold Upper corresponding point The tangent space Therefore, It can be expressed as

[0176] ;in Indicates The tangent vector at .

[0177] At the point Euclidean gradient at It can be expressed as ;

[0178] In Euclidean space, the Riemann gradient is a tangent vector and represents the fastest descent direction of the function. Then, It can be expressed as: ;

[0179] For the Riemann gradient, the following step is the contraction operation, which ensures that the required tangent vector Points considered above Projected to the specified Riemann manifold. Based on the above analysis, the contraction operation It can be expressed as: ; where superscript i is the iteration index, is the step size of the MO algorithm. In addition, the step size selection belongs to the category of Riemannian gradient descent (RGD) algorithm, and the classic RGD step size selection strategy is Armijo backtracking line search. More specifically, the Armijo backtracking line search method uses controllable parameters (usually 0.5 or 0.8) Iteratively reduce the tentative step size until the preset condition is met. For a specific iteration, the famous Armijo-Goldstein condition can be expressed as:

[0180] ;

[0181] in, represents a constant determined according to practical applications, and the above conditions can achieve rapid descent on Riemannian manifolds. It is an empirical parameter between 0 and 1 and needs to be debugged according to the actual application when used; It is the iteration step of the MO algorithm, which needs to be debugged according to the actual application; is the vector of the i-th iteration in the MO algorithm ;

[0182] 3) Optimization : Fixed optimized variable set , then The digital beamforming design is defined as an unconstrained minimization problem, so that the objective function can be written as: ;in, By introducing ,and The question was further rephrased as: ;

[0183] in, is the first auxiliary parameter,

[0184] is the second auxiliary parameter, is the third auxiliary parameter,

[0185] is the fourth auxiliary parameter, is the fifth auxiliary parameter,

[0186] is the sixth auxiliary parameter,

[0187] is the seventh auxiliary parameter, It is the eighth auxiliary parameter.

[0188] It is worth noting that the above problem is defined as an unconstrained convex problem and contains a quadratic objective function, so by Setting the derivative of to zero can obtain the optimal digital beamforming as follows:

[0189] ;

[0190] about The simulated combiner design for K users can be written as:

[0191] in,

[0192] ;

[0193] Furthermore, in this embodiment, the above design is simplified into a more compact form as follows:

[0194]

[0195] in, is the ninth auxiliary parameter,

[0196] is the tenth auxiliary parameter, is the eleventh auxiliary parameter,

[0197] is the twelfth auxiliary parameter, is the thirteenth auxiliary parameter,

[0198] is the fourteenth auxiliary parameter, is the fifteenth auxiliary parameter,

[0199] is the sixteenth auxiliary parameter, is the seventeenth auxiliary parameter,

[0200] is the eighteenth auxiliary parameter, It is the nineteenth auxiliary parameter.

[0201] In particular, the above optimization problem is a non-convex quadratic constrained quadratic programming problem with a quadratic objective function and constant modulus constraints, so it can be well solved using existing popular optimization algorithms.

[0202] Further, see Figure 3-Figure 4 As shown, the present invention also conducts a simulation test on the reliability of the communication system. Specifically, Figure 2 The figure shows the centralized communication relationship diagram of the UAV cluster millimeter wave MIMO communication system. Figure 3 The figure is a schematic diagram of the distributed communication relationship of the UAV cluster millimeter wave MIMO communication system. Figure 3 For example, this embodiment simulates an actual UAV cluster millimeter wave MIMO communication system, where all UAVs and users are randomly distributed within an area of ​​400×400 square meters, and the heights of UAVs and users are set to 6 meters and 2 meters (wherein the link between UAVs is set as a backhaul link, and the link between UAVs and users is set as a wireless link). Since millimeter wave communication links are easily blocked by obstacles, it is assumed that there is always one line-of-sight path and three non-line-of-sight paths between each UAV and each user. The complex gain of the line-of-sight path is mainly determined by the free space loss, while the complex gain of the non-line-of-sight path is calculated by the multipath loss. The departure azimuth and arrival azimuth of the millimeter wave channel are calculated from The number of array antennas equipped on each UAV and each user is N=64 and M=8 respectively. In addition, the operating frequency and noise power of the centralized and distributed UAV cluster millimeter wave MIMO communication systems are configured as GHz, dBm. Figure 3 It can be seen that the sum rate performance gap between centralized and fully distributed UAV swarm millimeter-wave MIMO communication systems is small and negligible in actual communication scenarios. The above simulation process shows that the computational complexity and backhaul signaling overhead of the fully distributed communication system in this application are not affected by the number of UAV clusters and the MIMO channel dimension, making it suitable for large-scale UAV swarm application scenarios.

[0203] Moreover, compared with the centralized UAV cluster millimeter wave MIMO communication system, the method protected by this application can significantly reduce the network's computational complexity and backhaul signaling overhead, and ensure that the network's sum and rate performance remain basically unchanged, which will make the fully distributed communication system easier to expand to the application scenarios of large-scale UAV clusters.

[0204] See also Figure 4As shown, the pairwise communication mechanism in the present application can not only avoid the interference of obstacles on communication efficiency, but also under the same UAV transmission power, the sum rate performance of the present application is almost consistent with the sum rate performance of the traditional centralized UAV cluster millimeter wave MIMO communication system, that is, the avoidance of obstacles in the present application does not reduce the communication speed of the UAV cluster.

[0205] Example 2: See Figure 5 As shown, the present invention also provides a fully distributed UAV cluster communication system, including: a decentralized module 01 for obtaining initial conditions and using an undirected graph to calculate the bth UAV and the Transmission information between drones; wherein the decentralized module includes: a graph definition unit 011, used to define an undirected graph for describing the communication relationship between multiple drones, wherein the nodes of the undirected graph represent the drones, and one node is connected to another node through at least one edge; a parameter definition unit 012, used to define initial parameters according to the undirected graph, wherein the initial parameters include: channel, undirected graph, association matrix, signed association matrix, analog beamforming, digital beamforming, and user combiner; a constraint definition unit 013, used to define constraint conditions according to the initial parameters: ;in, 、 are user mergers for the b-th and non-b-th UAVs, The drone cluster; a solution definition unit 014, used to generate a solution model according to the initial parameters and the constraint conditions; a solution unit 015, used to calculate the solution information according to the solution model, and the solution information includes: the optimal analog beamforming, digital beamforming, optimal merging vector of all users, and optimal sum rate of the b-th drone; a result unit 016, used to determine the transmission information according to the solution information; an update module 02, used to obtain a new b-th drone from the drone cluster, and define the transmission information of the previous b-th drone as the initial condition, and enter the decentralized solution module; a traversal module 03, returning to the decentralized module to traverse the drones in the drone cluster to complete the current round of communication.

[0206] In some embodiments, the communication system meets the following requirements: , . Furthermore, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the method described in any one of the embodiments is implemented. Furthermore, the present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a computer, the method described in any one of the embodiments is implemented. In some embodiments, the communication system is a millimeter wave MIMO communication system. In some embodiments, the communication link between at least one of the users and at least one of the drones is blocked by an obstacle.

[0207] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0208] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for enabling a computer terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0209] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are protected by the present invention.

Claims

1. A fully distributed UAV cluster communication method, characterized in that: Including steps: S1, obtain the initial conditions and use the undirected graph to calculate the bth UAV and the Transmitting information between drones; wherein S1 includes the steps of: S11, defining an undirected graph for describing communication relationships between drones, wherein nodes of the undirected graph represent the drones, and a node is connected to another node via at least one edge; S12, defining initial parameters according to the undirected graph, the initial parameters including: a channel, an undirected graph, an incidence matrix, a signed incidence matrix, an analog beamforming, a digital beamforming, and a user combiner; S13, defining constraints based on the initial parameters: ;in, 、 are user mergers for the b-th and non-b-th UAVs, The drone cluster; S14, generating a solution model according to the initial parameters and the constraint conditions; S15, calculating solution information according to the solution model, wherein the solution information includes: optimal analog beamforming, digital beamforming, optimal merging vectors of all users, and optimal sum rate of the b-th UAV; S16, determining the transfer information according to the solution information; S2, obtaining a new b-th drone from the drone cluster, and defining the transmission information of the previous b-th drone as the initial condition; S3, return to S1 until the drones in the drone cluster are traversed to complete the current round of communication.

2. The method according to claim 1, characterized in that The S15 includes: Set the relationship between drone index b and iteration number t: , mod represents the modulo operator, B is the number of drones; The bth UAV receives the Signaling information sent by drones and ; Update auxiliary variables generated during fractional transformation and auxiliary variables ,in, Represents a user collection; The bth drone transmits signaling information after updating: and ;in, is the complex number at the t+1th iteration , is the complex number at the t-th iteration, 、 is the conjugate transpose and digital beamforming vector of the combiner of the b-th UAV for the j-th user at the t-th iteration, For the channel, is the simulated beamforming at the tth iteration; is the second subsequent command information that does not include the b-th UAV in the t+1th iteration, is the combiner of all users of the b-th drone at the t-th iteration, is a complex matrix; Optimize the hybrid beamforming matrix of the b-th UAV using the block gradient descent algorithm , simulated beamforming matrix , digital beamforming matrix ; The b-th drone updates the combiner vector of K users ; Update the dual variables generated during the alternating direction multiplication method ; The bth drone transmits signaling information after updating: and , and send it to adjacent drones; After the solution process converges, the output solution information includes: the optimal analog beamforming and digital beamforming of the b-th UAV, the optimal merging vector of all users, and the optimal sum rate.

3. The method according to claim 1, characterized in that The S11 includes: Construct an undirected graph based on the connection relationship between drones , wherein the undirected graph has B nodes, and the nodes represent the drones, and the edges represents the backhaul link between the UAVs; The incidence matrix Defined as The incidence matrix of The rows and columns correspond to The nodes and edges of Convert to a signed incidence matrix , and for the e-th connection, given edge , to Defined as: ; Wherein, b is the index of the drone, is a complex space.

4. The method according to any one of claims 1 to 3, characterized in that The S11 also includes: Calculating a communication barrier degree of at least one communication partition, wherein a communication partition includes: at least one node; selecting the number of edges connected to the node according to the degree of communication barrier; The undirected graph is generated or updated according to the number of the edges.

5. The method according to claim 4, characterized in that The S11 also includes: When the communication barrier degree of the communication partition is greater than a preset first level, the first nodes in the communication partition are connected to the second number of second nodes using a second number of edges; and the second nodes are selected from the communication partitions whose communication barrier degree is less than the preset second level.

6. The method according to claim 5, characterized in that The communication partition includes multiple nodes, and correspondingly, S11 also includes: Setting a first subset of edges connecting a first node located in a central area of ​​the communication partition; A second subnumber of edges is set for a first node located in an edge area of ​​the communication zone, and the first subnumber is greater than or equal to the second subnumber.

7. The method according to claim 4, characterized in that The solution model includes: ； in, is the objective function, 、 For the analog beamforming and digital beamforming of the b-th UAV, is the transmission power of the b-th UAV, is the element in row i1 and column j1 of the simulated beamforming matrix of the b-th UAV, i1 and j1 are the row and column numbers of the elements in the matrix respectively. The i2th element in the user merge vector calculated locally for the bth drone, where i2 is the sequence number of the element in the vector. , , , 、 These are the two types of auxiliary variables introduced accordingly.

8. A fully distributed UAV cluster communication system, characterized by: Including steps: Decentralized module, used to obtain initial conditions and calculate the bth drone and the bth drone using an undirected graph Transmission of information between drones; wherein the decentralized module includes: A graph definition unit, configured to define an undirected graph for describing the communication relationship between a plurality of drones, wherein the nodes of the undirected graph represent the drones, and a node is connected to another node via at least one edge; A parameter definition unit, configured to define initial parameters according to the undirected graph, wherein the initial parameters include: a channel, an undirected graph, an incidence matrix, a signed incidence matrix, an analog beamforming, a digital beamforming, and a user combiner; A constraint definition unit is used to define constraint conditions according to the initial parameters: ;in, 、 are user mergers for the b-th and non-b-th UAVs, The drone cluster; A solution definition unit, configured to generate a solution model according to the initial parameters and the constraint conditions; a solution unit, configured to calculate solution information according to the solution model, wherein the solution information includes: optimal analog beamforming and digital beamforming for the b-th UAV, optimal combining vectors for all users, and optimal sum rate; A result unit, configured to determine the transfer information according to the solution information; An update module is used to obtain a new b-th drone from the drone cluster, define the transmission information of the previous b-th drone as the initial condition, and enter the decentralized solution module. The traversal module returns to the decentralization module to traverse the drones in the drone cluster to complete the current round of communication.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when the computer program is executed by a computer.

Images