A method, system and computer device for flexible formation control of a UAV swarm

Abstract

The application discloses a kind of unmanned aerial vehicle cluster flexible formation control method, system and computer equipment, by embedding local auction mechanism in each unmanned aerial vehicle, calculate safe neighbor set, take maximum consensus mechanism to update bid, realize the task allocation of decentralization between unmanned aerial vehicles, only need local communication between neighbors in each round of auction, without global coordination, greatly reduce communication overhead.On the basis of maintaining communication topology, introduce artificial potential field method based on Leader-Follower formation theory, a few leaders guide the whole formation movement by predetermined trajectory or task target, the rest followers realize dynamic maintenance and adjustment to formation by using attractive force and repulsive force in artificial potential field.It is especially suitable for unmanned aerial vehicle cluster formation performance, high-altitude flight task and other scenes.The lightweight distributed algorithm developed by using the characteristics of unmanned cluster itself provides technical support for subsequent unmanned cluster formation method.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of unmanned aerial vehicle cluster formation control, and particularly relates to an unmanned aerial vehicle cluster flexible formation control method and system and a computer device. BACKGROUND

[0002] An unmanned aerial vehicle cluster refers to an autonomous cooperative system composed of multiple unmanned aerial vehicles, which can perform complex tasks through distributed control, intelligent cooperation, network communication and other technologies. Compared with a single unmanned aerial vehicle, an unmanned aerial vehicle cluster has stronger adaptability, flexibility and robustness, and can be widely applied to many fields such as military, agriculture, emergency rescue, logistics and smart city.

[0003] As a large-scale system, how to maintain communication connectivity is also a very important problem. At present, the methods for maintaining the formation connectivity of the unmanned cluster system mainly include a topology control method and an optimization method. The topology control method, such as the Delaunay triangulation algorithm, constructs a triangular network with the given set of plane unmanned vehicle points as the vertices, so as to meet the Delaunay condition. The Delaunay triangular network usually has the optimal maximum minimum angle, which reduces the instability of the communication link. However, in the face of high-speed motion of the unmanned cluster system, node joining / leaving and other actual problems, the Delaunay triangulation needs to be frequently reconstructed, the calculation complexity in the three-dimensional space sharply rises, the algorithm efficiency is low, the storage cost is large, and it is not suitable for real-time calculation. The optimization algorithm, such as the optimization method based on the Laplacian matrix, maximizes the second smallest eigenvalue of the Laplacian matrix to ensure the communication topology of the unmanned cluster system. However, when the large-scale unmanned aerial vehicle formation (100+ nodes) is used, the overall communication network eigenvalue calculation cost is large, and an approximate method needs to be used. The Laplacian matrix only reflects the network topology at the current time, and cannot directly predict the future topology change. Maximizing the second smallest eigenvalue is not equivalent to the shortest path optimization. When the unmanned aerial vehicle moves at high speed or the large-scale topology changes, the network topology optimization needs to frequently recalculate the Laplacian matrix, which affects the real-time performance. SUMMARY

[0004] In view of the above technical problems, the application provides an unmanned aerial vehicle cluster flexible formation control method, system and computer device.

[0005] The technical solution adopted by the application to solve the technical problems is as follows:

[0006] An unmanned aerial vehicle cluster flexible formation control method, the method comprising the following steps:

[0007] S100: calculating a safe neighbor set of each agent, and initializing the request of each unmanned aerial vehicle in the auction algorithm, the maximum bid set and the token vector;

[0008] S200: Each agent updates the maximum bid set through the token collection mechanism;

[0009] S300: Determine the winner according to the maximum bid set, and delete the edge corresponding to the highest bid, repeat S200 and S300, after multiple rounds of auction, the final overall network connectivity graph that meets the pre-set k-connectivity is obtained;

[0010] S400: Define the UAV cluster center, determine the leader according to the Euclidean distance between the UAV and the UAV cluster center, and define the rest of the UAV as the follower, initialize the key parameters in the artificial potential field algorithm for the target area, including the initial position of the leader, the coordinate range of the dangerous area, and the position of the final target point;

[0011] S500: Calculate the specific position information of the UAV cluster flight area and the size range of its structure, when the UAV cluster reaches the position of the final target point, the formation flight ends, and when it does not reach the position of the final target point, the next step is performed;

[0012] S600: According to the key parameters in the artificial potential field algorithm in S400, calculate the repulsive force and attractive force received by the UAV cluster system, obtain the total force, and according to the artificial potential field method, calculate the formation flight information of the leader at the next moment, including position information, speed information and yaw angle information;

[0013] S700: Each follower follows the leader to form and maintain the desired formation shape, returns to S500 until the formation flight ends.

[0014] Preferably, S100 includes:

[0015] S110: Each UAV i first checks its directly connected neighbor , and calculates the second smallest eigenvalue of the Laplacian matrix of the network in a local range, by analyzing the change of the second smallest eigenvalue, it is determined whether deleting an edge (i, j) will damage the connectivity of the overall network, finally, all nodes j corresponding to the edges that can be safely deleted form a set, i.e. the safe neighbor set ;

[0016] S120: Initialize the bid request of each agent, the request format is in vector form , where i is the number of the current UAV, represents the edge that the UAV hopes to delete, is the corresponding bid amount; the UAV performs different operations according to the situation of its safe neighbor set: if the UAV i still has a safe neighbor, i.e. If the UAV i has no safe neighbors, i.e. , it will not actually participate in the bidding, but only as an information transmission node, at this time, the request it submits is , indicating that it does not propose to delete any edge, nor set the bidding amount;

[0017] S130: Initialize the maximum bidding set of each agent and the token vector , is mainly used to record the bidding information of each UAV i itself, and to ensure that these information can be exchanged and shared with its neighbors If the UAV i still has safe neighbors, i.e. , the bidding information of the UAV i itself is taken as the initial value of the maximum bidding set, i.e. , which indicates that the current maximum bidding set only contains the bidding data of the UAV itself; if the UAV i has no safe neighbors, i.e. , it will not participate in the bidding, in which case the maximum bidding set is an empty set, i.e. ; the token vector is in the form of , and the vector length is equal to the total number of UAVs in the cluster, only the value at index i is 1, indicating that the UAV numbered i has participated in the bidding, while the other positions are 0, indicating that other UAVs have not submitted a bid.

[0018] Preferably, S200 comprises:

[0019] The initial state of the token vector of the UAV numbered i is , and the token will gradually spread with the exchange of information among neighbors. In this process, each agent i will collect token information from its neighbor nodes and constantly update its token vector:

[0020] ;

[0021] wherein represents a logical OR operation, meaning that agent i will constantly synchronize the token state of the neighbors to ensure that it obtains all visible bidding information;

[0022] When the token vector of a node is 0, i.e. , it indicates that the UAV is still bidding, and the auction continues; when the token vector of all nodes is 1, i.e. , all UAVs have completed the bidding, and the system enters a termination state, and the auction ends; wherein, during the bidding process, each agent i applies a maximum consensus algorithm to update its maximum bidding set Specifically, agent i compares its own bid value with that of its neighbors and selects the one with the largest third element as the new bid set. ,in, Let m be the set of bids submitted by agent m at time t. Let i represent the set of neighbors of node i at time t. The third element in the three-dimensional vector [initiator, edge to be deleted, bid] represents the bid value. Agent i collects the bid sets of its neighbors, merges all visible bid information, finds the bid information corresponding to the highest bid value, and uses it as the current maximum bid set. By iteratively propagating the bid information, all agents gradually synchronize their bid results.

[0023] Preferably, S300 includes:

[0024] Each agent determines the winner based on the maximum bid set. The expression is Once the highest bid is accepted, the edge is removed from the network topology. The determination of the winning bid follows a decentralized approach, with each agent making decisions based solely on its own and its neighbors' information. A winner is determined in each round of auctions. After multiple rounds of auctions, when each drone's request format is r=[i, None, 0], the final connected graph is determined, where i is the number of each drone, None indicates that there is no edge to be deleted, and 0 indicates that the bid is zero, i.e., no bid is made. The network structure is dynamically adjusted while maintaining global connectivity, and the relative positions between each drone are also determined accordingly.

[0025] Preferably, S400 includes:

[0026] For a small drone swarm of 4 or fewer drones, the drone closest to the target point in Euclidean distance is selected as the leader drone.

[0027] With 4 or more drones, the drone swarm center is defined as... , To determine the position of each drone in the final overall network connectivity graph, the radius r of the drone cluster center is defined according to the actual scenario requirements. When the Euclidean distance between the drone's position and the center point is less than r, all drones that meet this condition are classified as leaders, while the remaining drones are classified as followers. The roles of leaders and followers will not change thereafter.

[0028] The leader plays a core guiding role in the execution of the task, determining the overall direction of movement of the formation. Followers adjust their own trajectories according to the leader's movement status and the location information of the surrounding neighbors in order to maintain the stability of the formation and ensure network connectivity.

[0029] The key parameters of the artificial potential field algorithm are initialized, including the initial position of the leader, the coordinate range of the dangerous area, and the position of the final target point. At the same time, the core weight parameters in the potential field calculation are set, including the strength of the attractive force field, the action radius of the repulsive force field, and the avoidance sensitivity of the unmanned aerial vehicle to obstacles. Reasonable configuration of these parameters will directly affect the movement path of the unmanned aerial vehicle, so that it can smoothly advance to the target area.

[0030] Preferably, S600 comprises:

[0031] S610: Calculate the repulsive force and attractive force of the unmanned aerial vehicle based on the repulsive force potential function and the attractive force potential function to analyze the influence of the environment on its movement, wherein the repulsive force comes from obstacles, dangerous areas and the interaction between unmanned aerial vehicles, prevents collision and maintains a safe distance, and increases with the decrease of distance, and the attractive force is determined by the target point or the formation requirement, guides the unmanned aerial vehicle to move to the target area, ensures the continuity of the cluster, and the total force is synthesized by combining the two, determines the movement trend of the unmanned aerial vehicle, and affects the trajectory adjustment;

[0032] S620: Based on the artificial potential field method, the resultant force is calculated according to the gradient change of the potential function, so that the unmanned aerial vehicle moves along the direction of the resultant force, and then the position information, speed and yaw angle of the leader at the next moment are calculated.

[0033] Preferably, in S610, the repulsive force and attractive force of the unmanned aerial vehicle are calculated based on the repulsive force potential function and the attractive force potential function, and in S620, the resultant force is calculated according to the gradient change of the potential function, specifically:

[0034] The attractive force potential function is The repulsive force potential function is The negative gradient of the potential function is obtained, that is, the attractive force And the repulsive force The sum of the two is the size of the resultant force in the artificial potential field Wherein, is the attractive force gain coefficient, is the repulsive force gain coefficient, wherein, is the distance between the unmanned aerial vehicle and the obstacle, is the radius size of the influence of the obstacle on the unmanned aerial vehicle.

[0035] Preferably, S700 comprises:

[0036] Based on the leader information, and relying on the leader-follower formation theory, the follower keeps a fixed distance from the leader at a predetermined yaw angle, and when the formation needs to be changed, only the following distance and yaw angle need to be adjusted; each follower controls the trajectory according to the rules to realize synchronous formation flight; the leader does not need to pay attention to the state of the follower, when the leader moves, the follower adjusts under the action of the potential field and the auction algorithm, when the actual distance and the expected value error is within the allowable range or close to zero, it is considered that the formation is achieved, and finally returns to S500 to judge whether to end the formation flight.

[0037] An unmanned aerial vehicle cluster flexible formation control system, comprising an auction algorithm initialization module, a maximum bid set updating module, a network connectivity graph determination module, a role division module, a formation flight end judgment module, a formation flight information determination module and a formation flight module;

[0038] The auction algorithm initialization module is used to calculate the safe neighbor set of each agent, and initialize the request of each unmanned aerial vehicle in the auction algorithm, the maximum bid set and the token vector;

[0039] The maximum bid set updating module is used to update the maximum bid set of each agent through the token collection mechanism;

[0040] The network connectivity graph determination module is used to determine the winner according to the maximum bid set, and delete the edge corresponding to the highest bid, repeatedly execute the auction algorithm initialization module and the maximum bid set updating module, and after multiple rounds of auction, obtain the final overall network connectivity graph satisfying the pre-set k-connectivity;

[0041] The role division module is used to define the center of the unmanned aerial vehicle cluster, determine the leader according to the Euclidean distance between the unmanned aerial vehicle and the center of the unmanned aerial vehicle cluster, define the remaining unmanned aerial vehicles as followers, initialize the key parameters of the artificial potential field algorithm for the target area, including the initial position of the leader, the coordinate range of the dangerous area and the position of the final target point;

[0042] The formation flight end judgment module is used to calculate the specific position information of the unmanned aerial vehicle cluster flight area and the size range of its structure, when the unmanned aerial vehicle cluster reaches the position of the final target point, the formation flight ends, and when it does not reach the position of the final target point, the next step is performed;

[0043] The formation flight information determination module is used to calculate the repulsive force and attractive force of the unmanned aerial vehicle cluster system according to the key parameters of the artificial potential field algorithm in the role division module, obtain the total force, and calculate the formation flight information of the leader at the next moment according to the artificial potential field method, including position information, speed information and yaw angle information;

[0044] The formation flight module is used for each follower to follow a leader to form and maintain a desired formation shape, and returns to the end of formation flight judgment module until the end of formation flight.

[0045] The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the UAV cluster elastic formation control method when executing the computer program.

[0046] The UAV cluster elastic formation control method, system and computer device, by embedding a local auction mechanism in each UAV, the group can achieve target allocation, formation reconstruction and link optimization without a central controller, each round of auction only needs local communication between neighbors, which greatly reduces the communication overhead. On this basis, the artificial potential field method based on the Leader-Follower formation theory is introduced, a few leaders guide the overall movement of the formation, and the rest of the followers use the attractive force and repulsive force in the artificial potential field to achieve dynamic maintenance and adjustment of the formation. It is especially suitable for UAV cluster formation performance, high-altitude flight tasks and other scenes. The lightweight distributed algorithm developed by the UAV cluster itself provides technical support for subsequent UAV cluster formation methods. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 The flowchart of a UAV cluster elastic formation control method in an embodiment of the present application;

[0048] Figure 2 The flowchart of a UAV cluster elastic formation control method in another embodiment of the present application;

[0049] Figure 3 The example of processing a fixed UAV cluster network in an embodiment of the present application, wherein (a) is the original formation network graph of 50 UAVs, and (b) is the final network connectivity graph after auction;

[0050] Figure 4 The principle diagram of the distributed auction algorithm in an embodiment of the present application;

[0051] Figure 5 The principle diagram of the artificial potential field algorithm based on the Leader-Follower formation theory in an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the person in the art better understand the technical scheme of the present application, the present application will be further described in detail below with reference to the drawings.

[0053] In one embodiment, as Figure 1 and Figure 2The unmanned aerial vehicle cluster flexible formation control method comprises the following steps:

[0054] S100: Calculate the safe neighbor set of each agent, and initialize the request of each unmanned aerial vehicle in the auction algorithm, the maximum bid set and the token vector;

[0055] S200: Each agent updates the maximum bid set through the token collection mechanism;

[0056] S300: Determine the winner according to the maximum bid set, delete the corresponding edge of the highest bid, and repeat S200 and S300, After multiple rounds of auction, the final overall network connected graph satisfying the pre-set k-connectivity is obtained;

[0057] S400: Define the unmanned aerial vehicle cluster center, determine the leader according to the Euclidean distance between the unmanned aerial vehicle and the unmanned aerial vehicle cluster center, and define the rest unmanned aerial vehicles as followers, initialize the key parameters in the artificial potential field algorithm for the target area, including the initial position of the leader, the coordinate range of the dangerous area and the position of the final target point;

[0058] S500: Calculate the specific position information of the unmanned aerial vehicle cluster flight area and the size range of its structure, when the unmanned aerial vehicle cluster reaches the position of the final target point, the formation flight ends, and when it does not reach the position of the final target point, the next step is performed;

[0059] S600: According to the key parameters in the artificial potential field algorithm in S400, calculate the repulsive force and attractive force received by the unmanned cluster system, obtain the total force, and according to the artificial potential field method, calculate the formation flight information of the leader at the next moment, including position information, speed information and yaw angle information;

[0060] S700: Each follower follows the leader to form and maintain the expected formation shape, and returns to S500 until the formation flight ends.

[0061] The unmanned aerial vehicle cluster flexible formation control method combines distributed control and maximum bid principle to ensure network connectivity, and introduces the artificial potential field method based on Leader-Follower formation theory. The combination of the two makes the formation process more flexible and the cooperation efficiency higher, which is very suitable for multi-agent formation cooperation tasks.

[0062] The purpose of S100 is to find out which edges can be deleted without affecting the overall network connectivity of the unmanned cluster, and let each unmanned aerial vehicle in the cluster submit a request to delete the edge with a bid. According to graph theory, the overall system can be regarded as an undirected graph, and each unmanned aerial vehicle in the system is numbered.

[0063] In one embodiment, S100 comprises:

[0064] S110: Each UAV i first checks its directly connected neighbors , and calculates the second smallest eigenvalue of the Laplacian matrix of the network in the local range. By analyzing the change of the second smallest eigenvalue, it is determined whether deleting an edge (i, j) will destroy the connectivity of the whole network. Finally, the nodes j corresponding to all edges that can be safely deleted form a set, i.e., the safe neighbor set ;

[0065] S120: Initialize the bidding request of each agent, and the request format is in vector form , where i is the number of the current UAV, represents the edge that the UAV wants to delete, is the corresponding bid amount; the UAV performs different operations according to the situation of its safe neighbor set: if the UAV i still has a safe neighbor, i.e., , then select a target neighbor from the safe neighbor as the edge to be deleted and submit a deletion request; at the same time, the bid amount is set to the maximum connection distance in the current system; if the UAV i has no safe neighbor, i.e., , then it will not actually participate in bidding, but only act as an information transmission node. At this time, the request it submits is , indicating that it does not propose to delete any edge and does not set the bid amount;

[0066] S130: Initialize the maximum bidding set of each agent and the token vector , , which is mainly used to record the bidding information of each UAV i itself and ensure that these information can be exchanged and shared with its neighbors . If the UAV i still has a safe neighbor, i.e., , then its bidding information is taken as the initial value of the maximum bidding set, i.e., , which indicates that the current maximum bidding set only contains the bidding data of the UAV itself. If the UAV i has no safe neighbor, i.e., , then it will not participate in bidding. In this case, the maximum bidding set is an empty set, i.e., ; the token vector is used to mark whether the UAV has submitted a bid, and is used for subsequent information synchronization. The format is , where the vector length equals the total number of UAVs in the cluster, and only the value at index i is 1, indicating that the UAV with the current number i has participated in the bidding, while other positions are 0, indicating that other UAVs have not submitted the bid. When the token vector is initialized, only the position of agent i itself is 1, and other positions are 0.

[0067] As shown in Figure 3 , S200 formally enters the auction cycle. In an embodiment, S200 includes:

[0068] The initial state of the token vector of the UAV numbered i is , and the token will gradually spread with the exchange of information between neighbors. In this process, each agent i will collect token information from its neighbor nodes, and constantly update its own token vector:

[0069] ;

[0070] Wherein, denotes the logical or operation, meaning that agent i will constantly synchronize the token state of the neighbor, ensuring that it obtains all visible bid information;

[0071] When the token vector of the node is 0, that is, , it means that the UAV is still bidding, and the auction continues. When the token vector of all nodes is 1, that is, , all UAVs have completed the bidding, and the system enters the termination state, and the auction ends; wherein, during the bidding process, each agent i will apply the maximum consensus algorithm to update its own maximum bid set . Specifically, agent i will compare its own and the neighbor's bid value, and select the third element of the bid value that is the largest as the new bid set, , wherein, is the bid set of agent m at time t; denotes the neighbor set of node i at time t, denotes the third element in the three-dimensional vector [initiator, edge to be deleted, bid], that is, the bid value; agent i will collect the bid set of the neighbor, combine all visible bid information, find the bid information corresponding to the highest bid value, and take it as the current maximum bid set; through iterative propagation of bid information, all agents are gradually synchronized with the bid result.

[0072] As shown in Figure 4 , the goal of S300 is to delete the actual winning edge and update the network topology. In an embodiment, S300 includes:

[0073] Each agent determines the winner according to the maximum bid set, and the expression is Once the request with the highest bid is accepted, the edge will be removed from the network topology; wherein the determination of the winning bid follows a decentralized manner, each agent only makes decisions based on its own and neighbor's information, and the winner is determined in each round of auction, after multiple rounds of auction, when the request format of each UAV is r=[i, None, 0], the final connected graph is determined, wherein i is the number of each UAV, None indicates that there is no edge to be deleted, and 0 indicates that the bid is zero, i.e. no bid, the network structure is dynamically adjusted while maintaining global connectivity, and the relative position between each UAV is also determined.

[0074] In the control strategy of S400, in order to ensure the cooperative motion and task execution efficiency of the whole system, the UAVs are first divided into roles. In an embodiment, S400 includes:

[0075] For a small UAV cluster of 4 or less, the one closest to the target point in Euclidean distance is selected as the leader;

[0076] For more than 4, the UAV cluster center is defined as , For the position of each UAV in the final overall network connected graph, the radius r of the UAV cluster center is defined according to the actual scene requirements, when the Euclidean distance between the UAV and the center point is less than r, all UAVs that meet this condition are classified as leaders, and the remaining UAVs are classified as followers, and thereafter the roles of the leaders and followers do not change;

[0077] For a given target area, the key parameters of the artificial potential field algorithm need to be initialized, including the initial position of the leader, the coordinate range of the danger area, the position of the final target point, and the core weight parameters in the potential field calculation, including the strength of the attractive force field, the action radius of the repulsive force field, and the obstacle avoidance sensitivity of the UAV. Reasonable configuration of these parameters will directly affect the motion path of the UAV, enabling it to smoothly advance towards the target area.

[0078] S500 is the determination flag of formation flight end. When the entire UAV cluster system reaches the target position, the formation flight task is completed, and if it does not reach the target position, the next action is performed.

[0079] As shown in FIG. 6, in an embodiment, S600 includes: Figure 5

[0080] ​S610: Calculate repulsion force and attraction force of the UAV based on repulsion potential function and attraction potential function to analyze the influence of the environment on its movement, wherein the repulsion force comes from obstacles, dangerous areas and interaction between UAVs to prevent collision and maintain a safe distance, which increases as the distance decreases (usually, the size of the repulsion force is inversely proportional to the distance between the UAV and the obstacle or other UAVs, that is, the closer the distance, the greater the repulsion force, thereby forcing the UAV to keep a reasonable distance from the dangerous area or each other), and the attraction force is determined by the target point or the formation requirement to guide the UAV to move to the target area and ensure the continuity of the cluster; the resultant force is determined by the combination of the two, which determines the movement trend of the UAV and affects the trajectory adjustment.

[0081] S620: Calculate the resultant force according to the gradient change of the potential function to make the UAV move in the direction of the resultant force, and then calculate the position information, speed and yaw angle of the leader at the next time.

[0082] In one embodiment, the calculation of the repulsion force and attraction force of the UAV based on the repulsion potential function and the attraction potential function in S610, and the calculation of the resultant force according to the gradient change of the potential function in S620, are as follows:

[0083] The attraction potential function is The repulsion potential function is The negative gradient of the potential function is obtained, that is, the attraction force And the repulsion force The sum of the two is the size of the resultant force in the artificial potential field Wherein, is the attraction gain coefficient, is the repulsion gain coefficient, wherein, is the distance between the UAV and the obstacle, is the radius of the influence of the obstacle on the UAV. The reasonable configuration of these parameters will directly affect the movement trajectory of the UAV, so that it can not only smoothly advance towards the target area, but also effectively avoid obstacles and dangerous areas during flight, avoid getting into local minimum value problem, and finally realize safe and efficient cooperative flight of the UAV cluster.

[0084] In one embodiment, S700 includes:

[0085] Based on the leader information and relying on the leader-follower formation theory, the follower maintains a fixed distance from the leader at a predetermined yaw angle, and when the formation needs to be changed, only the following distance and yaw angle need to be adjusted; each follower controls the trajectory according to the rules to realize synchronous formation flight; the leader does not need to pay attention to the state of the follower, when the leader moves, the follower adjusts accordingly under the action of the potential field and the auction algorithm, when the actual distance and the expected value error is within the allowable range or close to zero, it is considered that the formation is achieved, and finally returns to S500 to judge whether to end the formation flight.

[0086] The unmanned aerial vehicle cluster flexible formation control method fuses a distributed auction mechanism and an artificial potential field method based on a Leader-Follower formation theory in a flexible formation control framework. The auction algorithm mainly solves a communication topology problem, updates a bid by adopting a maximum consensus mechanism through calculation of a safe neighbor set, and realizes decentralized task allocation among unmanned aerial vehicles. In a large-scale unmanned aerial vehicle system, through the auction mechanism, each unmanned aerial vehicle only needs to exchange information with local neighbors, can autonomously select and maintain a key communication link meeting a k-connectivity requirement, does not need global coordination, and thus effectively reduces a communication burden. On the basis of maintaining the communication topology, a Leader-Follower formation control is introduced, a few leaders (Leaders) guide overall formation movement through a predetermined trajectory or a task target, and most followers (Followers) are dynamically adjusted according to attractive force (approaching the leaders or the target) and repulsive force (avoiding collision or moving away from an obstacle) generated by the artificial potential field method, so that the k-connectivity is ensured while a reasonable formation structure is maintained. Among them, a network composed of unmanned aerial vehicle nodes can be regarded as an undirected graph, and the k-connectivity means that at least k edges need to be deleted, and the overall network is not connected.

[0087] The combination of the distributed auction algorithm and the artificial potential field method based on the Leader-Follower theory can realize efficient cooperation and flexible reconstruction of the unmanned aerial vehicle cluster in a dynamic environment while maintaining network connectivity and stability of the formation structure, and provides a reliable solution for autonomous control of a large-scale intelligent unmanned system.

[0088] In one embodiment, a flexible formation control system of an unmanned aerial vehicle cluster includes an auction algorithm initialization module, a maximum bid set updating module, a network connectivity graph determination module, a role division module, a formation flight end judgment module, a formation flight information determination module, and a formation flight module.

[0089] The auction algorithm initialization module is configured to calculate a safe neighbor set of each agent, and initialize a request of each unmanned aerial vehicle, a maximum bid set, and a token vector in the auction algorithm.

[0090] The maximum bid set updating module is configured to update the maximum bid set by each agent through a token collection mechanism.

[0091] The network connectivity graph determination module is configured to determine a winner according to the maximum bid set, delete an edge corresponding to the highest bid, repeatedly execute the auction algorithm initialization module and the maximum bid set updating module, and obtain a final overall network connectivity graph meeting a pre-set k-connectivity after multiple rounds of auction.

[0092] The role division module is configured to define a UAV cluster center, determine a leader according to the Euclidean distance between the UAV and the UAV cluster center point, define the rest of the UAVs as followers, initialize key parameters in the artificial potential field algorithm for a target area, including the initial position of the leader, the coordinate range of the dangerous area, and the position of the final target point.

[0093] The formation flight end judgment module is configured to calculate the specific position information of the UAV cluster flight area and the size range of the structure, and when the UAV cluster reaches the position of the final target point, the formation flight ends, and when the UAV cluster does not reach the position of the final target point, the next step is performed.

[0094] The formation flight information determination module is configured to calculate the repulsive force and attractive force received by the UAV cluster system according to the information in the role division module, obtain the resultant force, and calculate the formation flight information of the leader at the next moment according to the artificial potential field method, including the position information, the speed information, and the yaw angle information.

[0095] The formation flight module is configured to make the followers follow the leader to perform formation flight, form and maintain the expected formation shape, and return to the formation flight end judgment module until the formation flight ends.

[0096] The specific limitations of the UAV cluster flexible formation control system can be seen in the limitations of the UAV cluster flexible formation control method described above, and will not be repeated here. Each module in the UAV cluster flexible formation control system described above can be realized by software, hardware, or a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0097] A computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the UAV cluster flexible formation control method when executing the computer program.

[0098] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, storage, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. The non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory. The volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, the RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0099] The above provides a kind of unmanned aerial vehicle cluster flexibility formation control method, system and computer equipment provided by the present application in detail. The principle and implementation of the present application are described in this paper by applying specific examples. The above embodiment is only used to help understand the core idea of the present application. It should be pointed out that for those skilled in the art, without departing from the principles of the present application, the present application can be improved and modified in several ways, and these improvements and modifications also fall within the scope of the claims of the present application.

Claims

1. A method for flexible formation control of unmanned aerial vehicle (UAV) swarms, characterized in that, The method includes the following steps: S100: Calculate the safe neighbor set for each agent and initialize the request, maximum bid set, and token vector for each drone in the auction algorithm; S200: Each agent updates the maximum bid set through a token collection mechanism; S300: Determine the winner based on the set of highest bids and delete the edge corresponding to the highest bid. Repeat S200 and S300. After multiple rounds of auction, the final overall network connectivity graph that satisfies the pre-defined k-connectivity is obtained. S400: Define the drone swarm center, determine the leader based on the Euclidean distance between the drone and the drone swarm center point, and define the remaining drones as followers. For the target area, initialize the key parameters in the artificial potential field algorithm, including the initial position of the leader, the coordinate range of the danger zone, and the position of the final target point. S500: Calculates the specific location information and size range of the drone swarm's flight area. When the drone swarm reaches the final target point, the formation flight ends; if it does not reach the final target point, it proceeds to the next step. S600: Based on the key parameters in the artificial potential field algorithm in S400, calculate the repulsive and attractive forces on the unmanned swarm system to obtain the total resultant force. Then, based on the artificial potential field method, calculate the formation flight information of the leader at the next moment, including position information, speed information and yaw angle information. S700: Each follower follows the leader in formation flight, forming and maintaining the desired formation, and returns to S500 until the formation flight ends; S100 includes: S110: Each drone first checks its directly connected neighbors. The second smallest eigenvalue of the Laplacian matrix of the network is calculated locally. By analyzing the changes in the second smallest eigenvalue, it is determined whether deleting an edge (i,j) will destroy the overall connectivity of the network. Finally, a set is formed of the nodes j corresponding to all edges that can be safely deleted, namely the safe neighbor set. ; S120: Initialize the bidding request for each agent, with the request format being a vector. Where i is the current drone number, This indicates the edge that the drone wants to delete. The corresponding bidding amount; the drone performs different operations based on its own safe neighbor set: if drone i still has safe neighbors, i.e. If drone i has no safe neighbors, then select a target neighbor from the safe neighbors as the edge to be deleted and submit a deletion request; if drone i has no safe neighbors, then select a target neighbor from the safe neighbors as the edge to be deleted and submit a deletion request. In this case, it will not actually participate in the bidding, but will only act as an information transmission node. The request it submits will then be... This means it does not request the deletion of any edges, nor does it set a bid amount; S130: Initialize the maximum bid set for each agent. and token vector , It is mainly used to record each drone's own bidding information and ensure that this information can be shared with its neighbors. Exchange and share can be performed if drone i still has a secure neighbor, i.e. Then, its own bidding information will be used as the initial value of the maximum bidding set, that is... This means that the current maximum bid set only contains the bid data of the drone itself; if drone i has no safe neighbors, i.e. If the maximum bid is not reached, then the bidder will not participate in the bidding. In this case, the maximum bid set is an empty set. Token vector The format is The vector length is equal to the total number of drones in the cluster. Only the value at index i is 1, indicating that the drone with the current number i has participated in the bidding, while the values ​​at other positions are 0, indicating that other drones have not yet submitted bids. S200 includes: The initial state of the token vector for drone number i is The tokens gradually spread as information is exchanged between neighbors. During this process, each agent i collects token information from its neighboring nodes and continuously updates its own token vector. ； in, This indicates a logical OR operation, meaning that agent i will continuously synchronize the token status of its neighbors to ensure that it obtains all visible bidding information; When the token vector of a node is 0, that is This indicates that the drone is still bidding, and the auction continues until all nodes' token vectors are all 1. At this point, all drones have completed their bidding, the system enters a terminated state, and the auction ends. During the bidding process, each agent i applies a maximum consensus algorithm to update its maximum bid set. Specifically, agent i compares its own bid value with that of its neighbors and selects the one with the largest third element as the new bid set. ,in, Let m be the set of bids submitted by agent m at time t. Let i represent the set of neighbors of node i at time t. The third element in the 3D vector [initiator, edge to be deleted, bid] represents the bid value. Agent i collects the bid sets of its neighbors, merges all visible bid information, finds the bid information corresponding to the highest bid value, and uses it as the current maximum bid set. By iteratively propagating the bid information, all agents gradually synchronize their bid results. The S300 includes: Each agent determines the winner based on the maximum bid set. The expression is Once the highest bid is accepted, the edge is removed from the network topology. The determination of the winning bid follows a decentralized approach, with each agent making decisions based solely on its own and its neighbors' information. A winner is determined in each round of auctions. After multiple rounds of auctions, when each drone's request format is r=[i, None, 0], the final connected graph is determined, where i is the number of each drone, None indicates that there is no edge to be deleted, and 0 indicates that the bid is zero, i.e., no bid is made. The network structure is dynamically adjusted while maintaining global connectivity, and the relative positions between each drone are also determined accordingly.

2. The method according to claim 1, characterized in that, The S400 includes: For a small drone swarm of four or fewer drones, the drone closest to the target point in Euclidean distance is selected as the leader. With 4 or more drones, the drone swarm center is defined as... , To determine the position of each drone in the final overall network connectivity graph, the radius r of the drone cluster center is defined according to the actual scenario requirements. When the Euclidean distance between the drone's position and the center point is less than r, all drones that meet this condition are classified as leaders, while the remaining drones are classified as followers. The roles of leaders and followers will not change thereafter. The leader plays a core guiding role in the execution of the task, determining the overall direction of movement of the formation. Followers adjust their own trajectories according to the leader's movement status and the location information of the surrounding neighbors in order to maintain the stability of the formation and ensure network connectivity. The key parameters for initializing the artificial potential field algorithm include the leader's initial position, the coordinate range of the danger zone, and the position of the final target point. At the same time, the core weight parameters in the potential field calculation are set, including the strength of the attractive potential field, the radius of action of the repulsive potential field, and the drone's sensitivity to obstacle avoidance. The proper configuration of these parameters will directly affect the drone's movement path, enabling it to move smoothly toward the target area.

3. The method according to claim 2, characterized in that, The S600 includes: S610: Based on the repulsive potential function and the attractive potential function, the repulsive force and attractive force on the UAV are calculated to analyze the influence of the environment on its motion. The repulsive force comes from obstacles, dangerous areas and the interaction between UAVs to prevent collisions and maintain a safe distance. It increases as the distance decreases. The attractive force is determined by the target point or formation requirements to guide the UAV to move towards the target area and ensure the coherence of the swarm. The total resultant force is obtained by combining the two forces, which determines the UAV's motion trend and affects the trajectory adjustment. S620: Based on the artificial potential field method, the resultant force is calculated according to the gradient change of the potential function, so that the UAV moves in the direction of the resultant force, and then the position information, speed and yaw angle of the leader are calculated at the next moment.

4. The method according to claim 3, characterized in that, In S610, the repulsive and attractive forces acting on the UAV are calculated based on the repulsive and attractive potential functions, and in S620, the resultant force is calculated based on the gradient change of the potential function. Specifically: The attraction potential function is The repulsive potential function is Finding the negative gradient with respect to the potential function yields the attractive force. and repulsive force The sum of the two is the magnitude of the net force in the artificial potential field. ,in, For attraction gain coefficient, Let be the repulsive force gain coefficient, where The distance between the drone and the obstacle. The radius of the drone is affected by obstacles.

5. The method according to claim 4, characterized in that, The S700 includes: Based on leader information and relying on leader-follower formation theory, followers maintain a fixed distance from the leader at a predetermined yaw angle. When the formation needs to be changed, only the following distance and yaw angle need to be adjusted. Each follower controls its trajectory according to the rules to achieve synchronized formation flight. The leader does not need to pay attention to the follower status. When the leader moves, the followers adjust accordingly under the influence of the potential field and auction algorithm. When the error between the actual distance and the expected value is within the allowable range or close to zero, the formation is considered to be achieved. Finally, the system returns to S500 to determine whether to end the formation flight.

6. A flexible formation control system for unmanned aerial vehicle (UAV) swarms based on the method described in any one of claims 1 to 5, characterized in that, It includes an auction algorithm initialization module, a maximum bid set update module, a network connectivity graph determination module, a role partitioning module, a formation flight end judgment module, a formation flight information determination module, and a formation flight module; The auction algorithm initialization module is used to calculate the safe neighbor set for each agent and initialize the request, maximum bid set, and token vector for each drone in the auction algorithm. The maximum bid set update module is used by each agent to update the maximum bid set through a token collection mechanism. The network connectivity graph determination module is used to determine the winner based on the maximum bid set and delete the edge corresponding to the highest bid. The auction algorithm initialization module and the maximum bid set update module are repeatedly executed. After multiple rounds of auction, the final overall network connectivity graph that satisfies the pre-set k-connectivity is obtained. The role division module is used to define the center of the drone swarm. The leader is determined based on the Euclidean distance between the drone and the center point of the drone swarm, and the remaining drones are defined as followers. For the target area, the key parameters in the artificial potential field algorithm are initialized, including the initial position of the leader, the coordinate range of the danger zone, and the position of the final target point. The formation flight end judgment module is used to calculate the specific location information of the drone swarm flight area and the size range of its structure. When the drone swarm reaches the final target point, the formation flight ends; when it does not reach the final target point, the next step is performed. The formation flight information determination module is used to calculate the repulsive and attractive forces on the unmanned swarm system based on the information in the role division module, obtain the total resultant force, and then calculate the formation flight information of the leader at the next moment, including position information, speed information and yaw angle information, based on the artificial potential field method. The formation flight module is used by each follower to follow the leader in formation flight, form and maintain the desired formation, and return to the formation flight end judgment module until the formation flight ends.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

Images