A fixed-wing unmanned aerial vehicle dense formation management and control method

Abstract

The application discloses a fixed-wing unmanned aerial vehicle dense formation management and control method, which is realized by adopting a hierarchical distributed control architecture and belongs to the technical field of unmanned aerial vehicle cluster control. In the hierarchical distributed control architecture, the method steps comprise the following steps: a fixed-wing unmanned aerial vehicle dense formation model is designed, and initialization of a flight control state, a mesh topology communication structure and a cruising route is completed; a top-layer architecture initializes a unmanned aerial vehicle formation model and a flight task configuration, and manages and maintains the fixed-wing unmanned aerial vehicle dense formation model based on a formation management strategy of a variable spanning tree structure; and a bottom-layer architecture realizes fixed-wing unmanned aerial vehicle dense formation flight based on a dense formation speed tracking controller of a virtual waypoint. By adopting the method, the autonomy, survivability and dynamic reconstruction efficiency of the dense formation are improved, the control complexity and collision risk are reduced, and stable dense formation flight of the fixed-wing unmanned aerial vehicle under complex scenes such as member increase and decrease and failure of a long machine is realized.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) swarm control technology, and particularly relates to a method for managing and controlling dense formations of fixed-wing UAVs. Background Technology

[0002] Unmanned aerial vehicle (UAV) swarms hold significant application potential, research importance, and strategic value in areas such as target reconnaissance and area coverage. Among these, dense swarms of fixed-wing UAVs can flexibly combine tactics such as deception and low-altitude penetration to form a multi-layered, three-dimensional penetration posture, meeting the needs of coordinated defense and concentrated attack within combat clusters. Dense swarms of fixed-wing UAVs require that, when flying in a pre-defined formation, the lateral distance between adjacent members must not exceed twice the wingspan, placing high demands on the real-time performance and accuracy of information exchange within the swarm.

[0003] To ensure the flight effectiveness of dense formations of fixed-wing UAVs, it is necessary to establish formation management and control models that can adapt to dynamic conditions. Currently, most existing research on the management and control of dense formations of fixed-wing UAVs tends to adopt a centralized control architecture, and rarely considers the formation reconstruction and maintenance issues after some members of the swarm become disabled. Furthermore, the increasingly complex future combat scenarios place higher demands on the stability and adaptability of UAV swarms, necessitating further development of management and autonomous control methods for dense formations of fixed-wing UAVs. Summary of the Invention

[0004] The purpose of this invention is to provide a method for managing and controlling dense formations of fixed-wing UAVs, which can dynamically manage formation members and optimize the maneuver control effect of formation flight, so as to solve the problem of formation reconstruction and maintenance after some members of a dense formation of fixed-wing UAVs become disabled.

[0005] To achieve the above objectives, this invention provides a method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles (UAVs), based on a hierarchical distributed control architecture. The hierarchical distributed control architecture includes a top-level architecture and a bottom-level architecture. The method includes the following steps:

[0006] S1. Design a model of a dense formation of fixed-wing UAVs, and complete the initialization of flight control status, mesh topology communication structure and loitering route before performing dense formation flight missions;

[0007] S2. The top-level architecture controls the dense formation model of fixed-wing UAVs, including initializing the UAV formation model and flight mission configuration, and managing and maintaining the dense formation model of fixed-wing UAVs based on the formation management strategy of variable spanning tree structure.

[0008] S3, the underlying architecture is based on a dense formation speed tracking controller with virtual waypoints, which realizes dense formation flight of fixed-wing UAVs by having the lead aircraft lead independently and the wingmen follow the dynamic virtual target waypoints.

[0009] Preferably, the design of a fixed-wing UAV dense formation model that meets the requirements of dense formation in step S1 includes:

[0010] Based on the single-aircraft motion model, considering the relative motion relationships in a dense formation of fixed-wing UAVs, the instantaneous relative positional relationship between two adjacent UAVs in the ground coordinate system is obtained. Here, the two UAVs in the formation are pre-defined as the i-th and i+1-th UAVs, and their only difference in altitude is the instantaneous velocity ζ. The three-dimensional spatial relative motion equations between adjacent UAVs in the formation are then established:

[0011]

[0012] In the formula, V i With V i+1 ψ represents the instantaneous velocity scalars of the two machines, respectively. i With ψ i+1 y, r represent the instantaneous yaw angles of the two aircraft, x, y, r represent the longitudinal, lateral, and center-to-center distances between the two aircraft, respectively, and z is the altitude difference between the two aircraft;

[0013] Among them, the instantaneous yaw angle error ψ between two adjacent aircraft in the formation e =ψ i -ψ i+1 .

[0014] Preferably, the fixed-wing UAV dense formation model is set under ideal conditions. ... e =0;

[0015] Close formation flying of fixed-wing UAVs requires that, when flying in a pre-defined formation, the lateral distance between adjacent members must not exceed twice the wingspan.

[0016] Preferably, the formation management strategy based on the variable spanning tree structure in step S2 is divided into two stages: initialization and dynamic management.

[0017] Initialization phase: Construct an undirected graph based on the dense formation model of fixed-wing UAVs. Based on the undirected graph, specify the root node of the tree structure and use the Dynamic-Prim algorithm with dynamically adjusted edge weights to construct the minimum spanning tree.

[0018] Dynamic management phase: The minimum directed spanning tree structure T generated in the initialization phase. t Maintenance is performed, specifically by generating a data table D with a variable spanning tree structure. v The model of a dense formation of fixed-wing UAVs is maintained. The maintenance includes node number, node survival status, and parent node information of all nodes except the root node.

[0019] Preferably, the construction of the undirected graph based on the dense formation model of fixed-wing UAVs includes: assuming that the number of UAVs in the dense formation cluster of fixed-wing UAVs is n and the positions of the UAVs are randomly distributed, treating all UAVs in the cluster as numbered vertices, treating the communication connections between UAVs as edges, and setting weights for the edges according to the relative distance between vertices, thereby establishing a connected undirected graph of the dense formation cluster of fixed-wing UAVs.

[0020] Preferably, the process of maintaining the dense formation model of fixed-wing UAVs includes:

[0021] When a new member joins a dense formation of fixed-wing UAVs, it will be added to data table D by default. v The node information is appended to the tail, and a leaf node in the formation is randomly assigned as the parent node of the appended node;

[0022] If a node fails or exits, there are three possibilities:

[0023] If the disabled node is a leaf node, then in data table D v Delete the corresponding information and report the situation;

[0024] If the disabled node is the root node, then one of its child nodes needs to be designated as the new root node, inheriting the decision information of the original root node, in data table D. v Delete the corresponding information and report the situation;

[0025] If the disabled node is an intermediate node, then its child nodes inherit from its parent node and replace the disabled node's original position in the formation, while updating data table D. v .

[0026] Preferably, the control strategy of the dense formation speed tracking controller based on virtual waypoints in step S3 is as follows: when the flight state of the lead aircraft in the dense formation of fixed-wing UAVs changes, the error between the controlled UAV and the target UAV is eliminated by controlling the speed, heading and altitude of each wingman, and the dense following of the virtual target waypoint is maintained. The heading is controlled through the lateral channel and the speed is controlled through the longitudinal channel.

[0027] Preferably, the longitudinal channel controls the projection Δx of the relative distance between the controlled UAV and the corresponding virtual target waypoint in the x-axis direction, and performs speed compensation through the relative speed ΔV between the controlled UAV and the corresponding virtual target waypoint, specifically:

[0028] The error control amount ΔX for the longitudinal channel is a linear combination of the longitudinal spacing error Δx and the relative speed error ΔV, and the formula is:

[0029] ΔX=w x Δx+w V ΔV;

[0030] In the formula, Δx is the projection of the relative distance between the controlled UAV and the corresponding virtual target waypoint on the x-axis, and w x with w V Gain weights are assigned to the spacing error and the velocity error, respectively, satisfying w x +w V =1;

[0031] Then, the speed setpoint at time t is calculated using the following formula. The formula is:

[0032]

[0033] In the formula, V ref The reference velocity for the virtual target point. and These are the control coefficients for the proportional, integral, and derivative terms, respectively, and the attenuation factor γ. x <1, ΔX n Let ΔX be the error control quantity of the longitudinal channel at time n. t Let ΔX be the error control value of the longitudinal channel at time t. t-1 Let ΔX be the error control quantity of the longitudinal channel at time t-1.

[0034] Preferably, the lateral channel controls the projection Δy of the relative distance between the controlled UAV and the corresponding virtual target waypoint in the y-axis direction, and performs yaw compensation by the relative heading angle Δψ between the controlled UAV and the corresponding virtual target waypoint.

[0035] The error control amount ΔY for the lateral channel is a linear combination of the lateral spacing error Δy and the relative yaw angle error Δψ, and the formula is:

[0036] ΔY=w y Δy+w ψ Δψ;

[0037] In the formula, w y with w ψ Gain weights are assigned to the spacing error and yaw angle error, respectively, which satisfy wy +w ψ =1;

[0038] Then, the yaw angle setpoint at time t is calculated using the following formula. The formula is:

[0039]

[0040] In the formula, and These are the control coefficients for the proportional, integral, and derivative terms, respectively, γ. Y Let be the attenuation factor introduced into the integral term, which satisfies γ Y <1, ΔY n Let ΔY be the error control quantity of the lateral channel at time n. t Let ΔY be the error control quantity of the lateral channel at time t. t-1 Let ΔY be the error control quantity of the transverse channel at time t-1;

[0041] Considering the maneuvering and turning during close formation flight of fixed-wing UAVs, the minimum turning radius R of the lead aircraft is constrained by the formation width B. fly To ensure the speed command of the UAVs on the inner side of the formation during maneuvering and turning. Not lower than the minimum loitering speed V of a fixed-wing aircraft L ;

[0042] Construct a roll maneuver adjuster to compensate for the positional offset of UAVs in dense formations through dynamic amplitude limiting. The resulting lateral control error, among which the roll maneuver adjustment formula for close formation maneuvers and turns of fixed-wing UAVs is:

[0043]

[0044] In the formula, clip represents the clipping function. and These are the lower and upper limits for cropping, respectively, K. inner With K outter K represents the inner and outer correction factors during formation maneuvers and turns. a With K b These represent the adjustment factors for the lower and upper limits after medial and lateral correction, respectively. For the cropped Set the roll angle value for the drone numbered i.

[0045] Preferably, the altitude channel controls the projection Δz of the relative distance between the controlled UAV and the corresponding virtual target waypoint along the z-axis. The error control amount ΔZ involved is only affected by the elevation spacing error Δz, as shown in the formula:

[0046] ΔZ=w z Δz;

[0047] In the formula, w z Assign gain weights to elevation errors;

[0048] The height setting value at time t is calculated using the following formula. The formula is:

[0049]

[0050] In the formula, and These are the control coefficients for the proportional, integral, and derivative terms, respectively, γ. Z Let be the attenuation factor introduced into the integral term, which satisfies γ Z <1, ΔZ n Let ΔZ be the height error control quantity at time n. t Let ΔZ be the height error control quantity at time t. t-1 Let ΔZ be the height error control quantity at time t-1.

[0051] Therefore, the present invention employs the above-described method for managing and controlling dense formations of fixed-wing UAVs, which has the following beneficial effects:

[0052] (1) Significantly improved survivability of the formation: Through the hierarchical distributed control architecture and the dynamic management mechanism of the variable spanning tree, the formation collapse problem caused by the single point failure of the long machine in the centralized architecture is effectively solved; when some members are disabled, the system can autonomously reconstruct the spanning tree topology to achieve seamless connection of the formation function and ensure the continuity of the task.

[0053] (2) Design a dense formation speed tracking controller based on virtual waypoints to enable dense formations to flexibly respond to maneuvering tasks and reduce control difficulty and collision risk.

[0054] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0055] Figure 1 A flowchart for the management and control of dense formations of fixed-wing UAVs;

[0056] Figure 2 This is a diagram showing the planar relative motion between the i-th and i+1-th UAVs in the formation.

[0057] Figure 3 A simplified model diagram of a dense formation of fixed-wing UAVs;

[0058] Figure 4 A mesh communication topology for dense formations of fixed-wing UAVs;

[0059] Figure 5 An example of a tree-like management structure for a dense formation system of fixed-wing UAVs;

[0060] Figure 6 An example of the formation process of a five-aircraft close formation system of fixed-wing UAVs;

[0061] Figure 7 The diagram shows the dynamic changes in the relationships between nodes after a node fails in a dense formation structure. (a) is a schematic diagram of the formation initialization phase; (b) is a schematic diagram of the formation dynamic management phase.

[0062] Figure 8 The diagram illustrates the disabling and reconfiguration process of a five-unit close formation of fixed-wing UAVs, where (a) is a diagram of random disabling of formation members and (b) is a diagram of dynamic reconfiguration after disabling of formation members.

[0063] Figure 9 This is a schematic diagram of the motion of a wingman approaching a virtual target waypoint in a close formation, where (a) is the horizontal plane motion diagram and (b) is the vertical plane motion diagram.

[0064] Figure 10 Flowchart of a speed tracking controller for dense formation of fixed-wing UAVs. Detailed Implementation

[0065] Example

[0066] This invention provides a method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles (UAVs), based on a hierarchical distributed control architecture, including a top-level architecture and a bottom-level architecture. The top-level architecture is responsible for the construction and maintenance of the overall formation model, while the bottom-level architecture controls the movement of each individual aircraft in the formation in a distributed manner to maintain the stability and consistency of the dense formation.

[0067] As attached Figure 1 The diagram shown is a flowchart for the management and control of dense formations of fixed-wing UAVs. In the dense formation model of fixed-wing UAVs, all UAVs are isomorphic, possessing the same functional modules and motion models. The specific steps of the method include:

[0068] S1. Design a fixed-wing UAV dense formation model that meets the requirements of dense formation, and complete the initialization of flight control status, mesh topology communication structure and loitering route before performing dense formation flight mission.

[0069] S2, the top-level architecture controls the dense formation model of fixed-wing UAVs, including initializing the UAV formation model and flight mission configuration, and managing and maintaining the dense formation model of fixed-wing UAVs based on the formation management strategy of variable spanning tree structure.

[0070] S3, the underlying architecture is based on a dense formation speed tracking controller with virtual waypoints, which realizes dense formation flight of fixed-wing UAVs by having the lead aircraft lead independently and the wingmen follow the dynamic virtual target waypoints.

[0071] In this embodiment, designing a fixed-wing UAV dense formation model that meets the requirements of dense formation specifically includes:

[0072] Based on the established single-aircraft motion model, and considering the relative motion relationships within a formation, the instantaneous relative positional relationships between adjacent aircraft in the ground coordinate system are shown in the attached figure. Figure 2 As shown. Let the two drones in the queue be the i-th and i+1-th drones, then V i With V i+1 ψ represents the instantaneous velocity scalars of the two machines, respectively. i With ψ i+1 y, r represent the instantaneous yaw angles of the two aircraft, x, y, r represent the longitudinal, lateral, and center-to-center distances between the two aircraft, respectively, and z represents the altitude difference between the two aircraft.

[0073] Assuming that the two aircraft have only an instantaneous velocity difference ζ in the height direction, the three-dimensional spatial relative motion equations of adjacent aircraft in the formation can be established:

[0074]

[0075] Among them, the instantaneous yaw angle error ψ between two adjacent aircraft in the formation e =ψ i -ψ i+1 The above formula shows the instantaneous relative distance between the (i+1)th UAV and the i-th UAV in three-dimensional space. Therefore, given the instantaneous state vector and instantaneous three-dimensional position of the i-th UAV, the instantaneous position of the (i+1)th UAV in three-dimensional space can be calculated using the above formula by combining the instantaneous state vector of the (i+1)th UAV itself.

[0076] Consider designing dense formations for fixed-wing UAVs. Under ideal conditions, ... e =0.

[0077] Close formation flying of fixed-wing UAVs requires that, when flying in a pre-defined formation, the lateral distance between adjacent members must not exceed twice the wingspan. In most application scenarios, the formation model of dense formations of fixed-wing UAVs can be simplified as shown in the attached figure. Figure 3 As shown. Where, offset x ,offsety These refer to the preset longitudinal and lateral spacing between machines in a dense formation queue. According to the aforementioned requirements of the definition of dense formation, the offset should be satisfied. y ≤2b.

[0078] Based on the aforementioned dense formation model, and considering the potential for communication link disruptions, aircraft malfunctions, or system failures in complex missions, a mesh communication topology is established for each data node within the dense formation, as shown in the attached diagram. Figure 4 As shown in the diagram. This communication topology establishes channels between adjacent UAVs in the cluster, enabling any pair of publishers and subscribers in the cluster to communicate directly or via relay through multiple links. In this case, the optimal channel is calculated and activated using graph search algorithms such as DFS, while the remaining channels that are not yet activated are called redundant channels. Simultaneously, if individual communication nodes or existing channels fail, the redundant channels need to be recalculated and activated to ensure the effectiveness and real-time nature of inter-UAV information exchange. (See attached diagram) Figure 4 As shown, assuming the cost of direct communication between the two machines in this communication topology is the same, if machine 5 receives a message published by machine 1, since there is no direct channel between the two machines, communication requires at least one relay node. The DFS algorithm calculates two shortest channels, 1-2-5 and 1-4-5, which are relayed through machines 2 and 4 respectively. Therefore, even if machine 2 suddenly loses connection during the task, machines 1 and 5 can still communicate via relay node 4.

[0079] In this embodiment, the cluster member management strategy based on the variable spanning tree structure is divided into two stages when applied during dense formation flight: the initialization stage and the dynamic management stage.

[0080] During the initialization phase, a spanning tree-based control structure is optimally established for the cluster. Given a cluster of n drones with randomly distributed initial positions, all drones can be considered as numbered vertices, and the communication connections between drones as edges. Weights are assigned to the edges based on the relative distances between vertices, thus constructing a connected undirected graph G = (u, v) for the drone cluster. Based on this undirected graph G, the root node of the tree structure is specified, and the minimum spanning tree of the graph is constructed using the Dynamic-Prim algorithm, which dynamically adjusts edge weights.

[0081] The basic idea of ​​the classic Prim algorithm is to start from a specified root node in the graph, select the vertex with the smallest edge weight to expand the tree structure, and recursively construct the minimum spanning tree in the undirected graph G by updating the leaf nodes and using them as the starting point for a new round of expansion. The Dynamic-Prim algorithm improves upon the classic Prim algorithm by adaptively changing the weights of the edges in the graph as the tree structure expands, each time the minimum spanning tree structure is updated in the undirected graph G.

[0082] Taking a close formation of five fixed-wing aircraft as an example, this illustrates how the Dynamic-Prim algorithm builds a variable tree structure for the cluster by optimizing distance costs. Figure 5 As shown.

[0083] First, a variable spanning tree structure T is established to optimally create virtual formation target waypoints planned by the lead aircraft for the wingmen in the preset formation. v And in establishing a tree structure T for the cluster t Traversing T during the process v Each node in the graph updates its weights by recalculating the distance between the target point and each vertex in the graph.

[0084] At this point, based on the minimum spanning tree structure calculated by Prim's algorithm, the undirected graph G can be transformed into a directed graph G according to the connection hierarchy between vertices. d The direction of each edge points from shallow vertices to deep vertices. From this, we can derive the directed graph relationships formed by the cluster members during the initialization phase and the minimum directed spanning tree structure T used to manage the cluster members. t . Figure 4 This paper demonstrates a tree-like member management structure for a 5-machine dense formation based on a minimum directed spanning tree. The lead machine, UAV 0, is the root node of the directed spanning tree structure and also manages UAV 1 and UAV 2 as its parent node. In the subtree with UAV 2 as its parent node, UAV 4 is a child node managed by UAV 2. UAV 3 is located at the deepest level of the tree structure, i.e., the edge layer, and therefore serves as a leaf node in the entire formation's tree-like management structure, possessing the lowest control priority. Similarly, the formation member management structure during the initialization phase can be represented by the relational mapping form shown in Table 1.

[0085] Table 1. Example of the "parent node-child node" relationship mapping in the tree-structured management of a dense formation system.

[0086]

[0087] During the dynamic management phase, the fixed-wing UAV dense formation system will analyze the minimum directed spanning tree structure T generated during the initialization phase. t Maintenance is performed, i.e., based on the variable spanning tree structure T. vDynamic management of formation members is implemented. Specifically, it involves determining whether the actual drone ID matches the formation ID. If they match, low-level control is performed; otherwise, the variable tree structure is updated, and the process of maintaining the variable tree structure based on the real-time status of the drones is repeated.

[0088] This dynamic management phase will maintain a data table D with a variable spanning tree structure for dense formations of fixed-wing UAVs. v The maintained information includes node ID, node liveness status, and parent node information for all nodes except the root node. (Appendix) Figure 6 The study demonstrates the state changes of a 5-machine cluster with randomly distributed initial positions, based on a variable spanning tree structure, after optimizing the formation system management strategy, from the initialization phase to the dynamic management phase.

[0089] By maintaining this data table D v The dynamic management module in the cluster can handle changes in cluster members relatively stably. Specifically, if a new member is added to a densely clustered cluster, it will be added to data table D by default. v The node information is appended to the tail, and a leaf node in the formation is randomly assigned as the parent node of the appended node. If a node fails or exits, three cases need to be considered: if the disabled node is a leaf node, then in data table D... v Delete the corresponding information and report the situation; if the disabled node is the root node, then one of its child nodes needs to be assigned as the new root node, inheriting the decision information of the original root node, and in data table D. v The corresponding information is deleted and the situation is reported; if the disabled node is an intermediate node, its child nodes inherit from its parent node and replace the disabled node's original position in the formation, while updating data table D. v This is to maintain the stability of the dense formation.

[0090] Managing dense formations of fixed-wing UAVs using a variable spanning tree structure enables each data node to trace upstream nodes based on a clear understanding of its local "leader-follower" relationship. Furthermore, the relationships between nodes can be dynamically changed when necessary, as shown in the attached diagram. Figure 7 As shown. This example illustrates formation restructuring caused by changes in members within a dense formation. Figure 8 Demonstrates based on Figure 6 The dynamic changes during the reconfiguration of a dense formation of fixed-wing UAVs. When UAV 5 malfunctions and leaves the formation, its child node UAV 3 inherits from its parent node UAV 1 and approaches the new virtual target waypoint.

[0091] It is important to note that in scenarios where the lead drone fails, the remaining intact drones in the formation initially maintain their original flight status. Then, the formation management model reassigns the wingman closest to the lead drone in the original variable spanning tree structure as the new lead drone, placing it under the lead drone's control strategy. Simultaneously, a new variable spanning tree structure is generated based on the updated drone distance matrix, forming a new formation structure.

[0092] In this embodiment, the control strategy of the dense formation speed tracking controller based on virtual waypoints is as follows:

[0093] Unlike rotary-wing UAVs, fixed-wing UAVs in flight need to maintain a certain speed, therefore they can only form and maintain close formations during forward movement. Based on a simplified model of close cooperative formation, the motion of fixed-wing UAVs during formation flight can be decomposed into two directions: lateral and longitudinal. Lateral motion refers to the movement of fixed-wing UAVs in the horizontal heading direction, involving roll and yaw angle control; longitudinal motion encompasses both altitude and speed, involving pitch and throttle control.

[0094] Under the dynamic conditions of fixed-wing UAV swarm flight, a dense formation speed tracking controller based on virtual waypoints is proposed. This controller achieves dense formation flight of fixed-wing UAVs by having a lead aircraft independently navigate and wingmen follow corresponding virtual target waypoints. In this scenario, each wingman in the formation generates and adjusts its corresponding virtual target waypoint position based on independent observation information, according to a preset formation and its own number, by observing the position of the lead aircraft. It then approaches this virtual target waypoint, thereby achieving collaborative maintenance of the dense formation. Taking a two-aircraft dense following mission as an example, the horizontal and vertical motions of the wingman approaching the virtual target waypoint are shown in the attached figures. Figure 9 (a) and appendix Figure 9 As shown in (b).

[0095] For fixed-wing UAVs in close formation flight missions, when the flight status of the lead aircraft changes, the virtual target waypoints of each wingman will also move in unison. At this time, by controlling the speed, heading, and altitude of each wingman, the spacing error between the controlled UAV and the target—that is, the close formation consistency error—can be effectively eliminated, ensuring that the UAV maintains close following of the virtual target waypoint. The heading is controlled via the lateral channel, and the speed is controlled via the longitudinal channel. (See attached...) Figure 10 As shown, the designed dense formation speed tracking controller takes dense formation consistency information as input and assigns yaw compensation and speed compensation to the virtual target waypoint position errors in the horizontal and vertical channels, respectively. Then, the desired attitude and throttle settings are generated by the roll maneuver adjuster and the total energy controller. The attitude and throttle control quantities after being limited are further output to the PX4 attitude loop to realize the speed tracking control of the fixed-wing UAV.

[0096] Based on the model of dense formation of fixed-wing UAVs, this invention designs a speed tracking controller for dense formation based on virtual waypoints. Through mechanisms such as yaw compensation, speed compensation, and roll maneuver adjustment, it enables distributed dense formations to adapt to maneuvering turns, reducing control difficulty and collision risk.

[0097] In this embodiment, the control strategies for the longitudinal, lateral, and height channels are as follows:

[0098] (a) Longitudinal Channel: The longitudinal channel controls the state variables on the XOZ plane, specifically the projection Δx of the relative distance between the controlled UAV and the corresponding virtual target waypoint along the x-axis, i.e., the longitudinal spacing error, and compensates for the velocity using the relative velocity ΔV between them. The error control variable ΔX of the longitudinal channel is a linear combination of the longitudinal spacing error Δx and the relative velocity error ΔV, as shown in the formula:

[0099] ΔX=w x Δx+w V ΔV;

[0100] In the formula, w represents the gain weights assigned to the spacing error and the velocity error, respectively. x with w V It satisfies w x +w V =1;

[0101] The speed setpoint at time t can then be calculated using the following formula. The formula is:

[0102]

[0103] In the formula, V ref The reference velocity for the virtual target point. and These are the control coefficients for the proportional, integral, and derivative terms, respectively, and the attenuation factor γ. x <1, ΔX n Let ΔX be the error control quantity of the longitudinal channel at time n. t Let ΔX be the error control value of the longitudinal channel at time t. t-1 Let ΔX be the error control quantity of the longitudinal channel at time t-1;

[0104] (b) Lateral Channel: The lateral channel enables control of state variables on the YOZ plane in the body coordinate system, specifically controlling the projection Δy of the relative distance between the controlled UAV and the corresponding virtual target waypoint along the y-axis, i.e., the lateral spacing error, and compensating for yaw using the relative heading angle Δψ between them. The error control quantity ΔY of the lateral channel is a linear combination of the lateral spacing error Δy and the relative yaw angle error Δψ, as shown in the formula:

[0105] ΔY=w y Δy+w ψ Δψ;

[0106] In the formula, w represents the gain weights assigned to the spacing error and the yaw angle error, respectively. y with w ψ It satisfies w y +w ψ =1;

[0107] The yaw angle setpoint at time t can then be calculated using the following formula. The formula is:

[0108]

[0109] In the formula, and These are the control coefficients for the proportional, integral, and derivative terms, respectively, γ. Y Let be the attenuation factor introduced into the integral term, which satisfies γ Y <1, ΔY n Let ΔY be the error control quantity of the lateral channel at time n. t Let ΔY be the error control quantity of the lateral channel at time t. t-1 Let ΔY be the error control quantity of the transverse channel at time t-1;

[0110] The minimum turning radius R of the lead aircraft is constrained by the formation width B. fly Ensure the speed command of the UAV i inside the formation during maneuvering and turning. Not lower than the minimum loitering speed V of a fixed-wing aircraft L ;

[0111] Construct a roll maneuver adjuster to compensate for the positional offset of UAVs in dense formations through dynamic amplitude limiting. The resulting lateral control error is adjusted for roll maneuvering during close formation maneuvers and turns of fixed-wing UAVs using the following formula:

[0112]

[0113] In the formula, clip represents the clipping function. and These are the lower and upper limits for cropping, respectively, K. inner With K outter K represents the inner and outer correction factors during formation maneuvers and turns. a With K b These represent the adjustment factors for the lower and upper limits after medial and lateral correction, respectively. For the cropped Set the roll angle value for the drone numbered i;

[0114] (c) Altitude Channel: In the altitude channel, the projection Δz of the relative distance between the controlled UAV and the corresponding virtual target waypoint on the z-axis is mainly controlled, i.e., the elevation distance error. The error control amount ΔZ involved in this channel is only affected by the elevation distance error Δz, and the formula is:

[0115] ΔZ=w z Δz;

[0116] In the formula, w z Assign gain weights to elevation errors;

[0117] The height setting value at time t is calculated using the following formula. The formula is:

[0118]

[0119] In the formula, and These are the control coefficients for the proportional, integral, and derivative terms, respectively, γ. Z Let be the attenuation factor introduced into the integral term, which satisfies γ Z <1, ΔZ n Let ΔZ be the height error control quantity at time n. t Let ΔZ be the height error control quantity at time t. t-1 Let ΔZ be the height error control quantity at time t-1.

[0120] Finally, the speed tracking control command is calculated, the flight control execution command is output and the movement is initiated, thus concluding the formation control process.

[0121] Therefore, the present invention adopts the above-mentioned method for managing and controlling dense formations of fixed-wing UAVs. Through the collaborative design of hierarchical distributed control architecture, variable spanning tree structure dynamic management, and virtual waypoint speed tracking controller, it significantly improves the autonomy, survivability, and dynamic reconfiguration efficiency of dense formations, while reducing control complexity and collision risk. It enables stable dense formation flight of fixed-wing UAVs in complex scenarios such as crew additions and subtractions and lead aircraft failure.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles (UAVs), characterized in that, This is implemented based on a layered distributed control architecture, which includes a top-level architecture and a bottom-level architecture. The method and steps include: S1. Design a model of a dense formation of fixed-wing UAVs, and complete the initialization of flight control status, mesh topology communication structure and loitering route before performing dense formation flight missions; S2. The top-level architecture controls the dense formation model of fixed-wing UAVs, including initializing the UAV formation model and flight mission configuration, and managing and maintaining the dense formation model of fixed-wing UAVs based on the formation nodes of the variable spanning tree structure. S3, the underlying architecture is a dense formation speed tracking controller based on virtual waypoints, which realizes dense formation flight of fixed-wing UAVs by the method of independent lead aircraft and wingmen following dynamic virtual target waypoints; Considering the maneuvering and turning during close formation flight of fixed-wing UAVs, based on the formation width Minimum turning radius for constraining the flight of the lead aircraft To ensure the speed command of the drones on the inner side of the formation during maneuvering and turning. Not lower than the minimum loitering speed of a fixed-wing aircraft ; Construct a roll maneuver adjuster to compensate for the positional offset of UAVs in dense formations through dynamic amplitude limiting. The resulting lateral control error, among which the roll maneuver adjustment formula for close formation maneuvers and turns of fixed-wing UAVs is: ； In the formula, This represents the clipping function. and These are the lower and upper limits for cropping, respectively. and These represent the inner and outer correction factors during formation maneuvers and turns. and These represent the adjustment factors for the lower and upper limits after medial and lateral correction, respectively. For the cropped , For the number The roll angle setting of the drone.

2. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 1, characterized in that, Step S1 involves designing a close formation model of fixed-wing UAVs that meets the requirements of close formation, including: Based on a single-aircraft motion model, considering the relative motion relationships in a dense formation of fixed-wing UAVs, the instantaneous relative positional relationship between two adjacent UAVs in the ground coordinate system is obtained. Here, the two UAVs pre-defined in the formation are the [missing information - likely a specific aircraft or group]. i frame and the first i +1 drone, with only an instantaneous speed difference between the two drones in the altitude direction. Then, establish the three-dimensional spatial relative motion equations between two adjacent machines in the formation: ； In the formula, and These are the instantaneous velocity scalars of the two machines, respectively. and These are the instantaneous yaw angles of the two aircraft, x,y These represent the longitudinal and lateral spacing between the two machines, respectively. z The height difference between the two machines; Among them, the instantaneous yaw angle error between two adjacent aircraft in the formation .

3. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 2, characterized in that: The fixed-wing UAV dense formation model is assumed to be under ideal conditions. These represent the ideal longitudinal and lateral spacing and height difference in a close formation, respectively. Let be the wingspan of the fixed-wing UAV. At this point, the velocity vector directions of adjacent UAVs should be consistent, and the yaw angle difference should satisfy... ; Close formation flying of fixed-wing UAVs requires that, when flying in a pre-defined formation, the lateral distance between adjacent members must not exceed twice the wingspan. .

4. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 1, characterized in that: The formation management strategy based on the variable spanning tree structure in step S2 is divided into two phases: initialization and dynamic management. Initialization phase: Construct an undirected graph based on the dense formation model of fixed-wing UAVs. Based on the undirected graph, specify the root node of the tree structure and use the Dynamic-Prim algorithm with dynamically adjusted edge weights to construct the minimum spanning tree. Dynamic management phase: The minimum directed spanning tree structure generated during the initialization phase. Maintenance is performed by generating a data table with a variable spanning tree structure. The model of a dense formation of fixed-wing UAVs is maintained, including node number, node liveness status, and parent node information of all nodes except the root node.

5. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 4, characterized in that, The construction of the undirected graph based on the dense formation model of fixed-wing UAVs includes: assuming the number of UAVs in the dense formation of fixed-wing UAVs is... n Furthermore, the drones are randomly distributed. All drones in the cluster are considered as numbered vertices, and the communication connections between drones are considered as edges. The weights of the edges are set according to the relative distance between the vertices, thus establishing a connected undirected graph of the dense formation of fixed-wing drones.

6. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 4, characterized in that: The process of maintaining the dense formation model of fixed-wing UAVs includes: When a new member joins a dense formation of fixed-wing UAVs, it will be added to the data table by default. The node information is appended to the tail, and a leaf node in the formation is randomly assigned as the parent node of the appended node; If a node fails or exits, there are three possibilities: If the disabled node is a leaf node, then in the data table Delete the corresponding information and report the situation; If the disabled node is the root node, then one of its child nodes needs to be designated as the new root node, inheriting the decision information of the original root node, and this information needs to be added to the data table. Delete the corresponding information and report the situation; If the disabled node is an intermediate node, then its child nodes inherit from its parent node and replace the disabled node's original position in the formation, while simultaneously updating the data table. .

7. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 1, characterized in that: The control strategy of the dense formation speed tracking controller based on virtual waypoints in step S3 is as follows: when the flight state of the lead aircraft in the dense formation of fixed-wing UAVs changes, the error between the controlled UAV and the target UAV is eliminated by controlling the speed, heading and altitude of each wingman, and the dense following of the virtual target waypoint is maintained. The heading is controlled through the lateral channel and the speed is controlled through the longitudinal channel.

8. The method for managing and controlling dense formations of fixed-wing unmanned aerial vehicles according to claim 7, characterized in that: The projection of the relative distance between the controlled UAV and the corresponding virtual target waypoint on the z-axis of the altitude channel. To perform control, the error control quantity involved Based solely on elevation spacing error The effect, the formula is: ； In the formula, Assign gain weights to elevation errors; The height setting value at time t is calculated using the following formula. The formula is: ； In the formula, , and These are the control coefficients for the proportional, integral, and derivative terms, respectively. Let be the attenuation factor introduced into the integral term, which satisfies , for Time-based height error control quantity , for Time-based height error control quantity , for Time-based height error control quantity .

Images