Unmanned aerial vehicle cluster formation cooperative control method, system, device and medium

Abstract

The invention relates to the technical field of unmanned aerial vehicle control, and discloses an unmanned aerial vehicle cluster formation cooperative control method, system and device and a medium. The method comprises the following steps: establishing a dynamic model of each unmanned aerial vehicle in an unmanned aerial vehicle cluster and a dynamic model of an obstacle; establishing an information exchange topology model for describing a data interaction path between the unmanned aerial vehicles in the unmanned aerial vehicle cluster; according to each established model, designing a time-varying formation tracking protocol and a collision avoidance controller based on an artificial potential field method; all the unmanned aerial vehicles in the unmanned aerial vehicle cluster are controlled to form a preset formation which dynamically changes according to task requirements in the flight process through a time-varying formation tracking protocol, and a preset flight path which dynamically changes according to the task requirements is tracked; all the unmanned aerial vehicles are attracted to move towards a preset target point through a gravitational field of the collision avoidance controller, and all the unmanned aerial vehicles are repelled to move towards obstacles and each unmanned aerial vehicle is repelled to move towards other unmanned aerial vehicles through a repulsive force field of the collision avoidance controller, so that control over the unmanned aerial vehicle cluster is achieved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and in particular to a method, system, device, and medium for collaborative control of UAV swarm formations. Background Technology

[0002] As a cutting-edge technology, the collaborative control of unmanned aerial vehicle (UAV) swarms exhibits significant innovation in its technical characteristics and application value. Unlike single-aircraft operation modes, the distributed architecture of multi-aircraft collaboration offers multidimensional advantages at the task execution level. From a spatial coverage perspective, multi-aircraft swarms can form a wide-area monitoring network, enabling simultaneous processing of multiple target tasks through parallel operation mechanisms, effectively improving operational accuracy and task stability. The distributed fault-tolerance mechanism introduced at the system architecture level can trigger a task redistribution protocol when a unit node fails, maintaining the integrity of system functionality.

[0003] Current collaborative control methods for UAV swarms treat the entire swarm as a rigid virtual structure using a virtual structure-based control scheme, achieving overall motion through centralized planning of the trajectory of the virtual leader. However, this method has significant limitations. Its swarm configuration is typically preset and fixed, lacking flexibility to adapt to dynamic mission environments. If the virtual leader, as the core control point, fails (single point of failure), it can easily trigger the collapse of the entire swarm system. Furthermore, such methods usually lack integrated dynamic obstacle avoidance capabilities, leading to a sharp increase in flight risks in complex environments. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, device and medium for collaborative control of unmanned aerial vehicle (UAV) swarm formations, which can solve the problems of existing technologies lacking flexibility in dealing with dynamic mission environments, being prone to causing the entire swarm system to crash and lacking integrated dynamic obstacle avoidance capabilities.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide a method for cooperative control of unmanned aerial vehicle (UAV) swarm formations, comprising the following steps: Establish a dynamic model for each drone in the drone swarm, and establish a dynamic model for obstacles existing in the pre-defined flight path of the drone swarm; Establish an information exchange topology model to describe the data interaction paths between drones in a drone swarm; Based on the information exchange topology model and the dynamic model of each UAV, a time-varying formation tracking protocol is designed to control all UAVs in the UAV swarm to form a preset formation that dynamically changes according to mission requirements during flight and to track a preset flight trajectory that dynamically changes according to mission requirements. Based on the artificial potential field method, and according to the information exchange topology model, the dynamic model of each UAV and the dynamic model of the obstacle, a collision avoidance controller containing a gravitational field and a repulsive field is designed. The gravitational field attracts all UAVs to move towards the preset target point, and the repulsive field repels all UAVs from moving towards the obstacle and repels each UAV from moving towards other UAVs. By combining a time-varying formation tracking protocol and a collision avoidance controller, the drone swarm can be controlled.

[0006] Furthermore, the time-varying formation tracking protocol is as follows: ; In the formula, It is a weighted adjacency matrix, and the desired time-varying formation is composed of vectors. Specify, Followers A continuously differentiable piecewise formation vector. and These represent the agent's desired positional offset and velocity offset relative to the leader within the formation, respectively. The extended state is , Let be a constant gain matrix, denoted as , .

[0007] Furthermore, the collision avoidance controller is: ; in, For obstacle avoidance, and , , For the set of all obstacles, To improve obstacle avoidance, For obstacles speed; It is a group coordination term, and ; For drones The neighborhood group, For the location of the neighbor, These are the weighting coefficients. , It is Gaussian white noise. These are elements of the adjacency matrix; For collision avoidance between drones, , This represents the collision avoidance gain coefficient between unmanned aerial vehicles (UAVs).

[0008] Furthermore, the control of the drone swarm by combining the time-varying formation tracking protocol and the collision avoidance controller includes: The weights of the time-varying formation tracking protocol and the collision avoidance controller are determined based on the distance between each drone and the obstacle. Each drone in the drone swarm is controlled according to the time-varying formation tracking protocol and the weights of the collision avoidance controller.

[0009] Furthermore, each drone in the drone swarm is controlled using the following formula: ; ; ; In the formula, , , , This is an indicator function; its value is 1 if the condition is true, and 0 otherwise. For drones The neighborhood group, For the neighbor's central location, These are the weighting coefficients for velocity matching, centroid projection, and noise. For standard normally distributed random noise, Let V be the velocity vector of the obstacle. These are the elements of the adjacency matrix.

[0010] Embodiments of the present invention also provide a drone swarm formation cooperative control system, comprising: The first model building module is used to build a dynamic model of each drone in the drone swarm, and to build a dynamic model of obstacles existing in the preset drone swarm flight trajectory; The second model building module is used to build an information exchange topology model that describes the data interaction paths between drones in a drone cluster. The formation protocol design module is used to design a time-varying formation tracking protocol based on the information exchange topology model and the dynamic model of each UAV, so as to control all UAVs in the UAV cluster to form a preset formation that dynamically changes according to mission requirements and track a preset flight trajectory that dynamically changes according to mission requirements during flight. The collision avoidance control design module is used to design a collision avoidance controller that includes a gravitational field and a repulsive field, based on the artificial potential field method, the information exchange topology model, the dynamic model of each UAV, and the dynamic model of the obstacle. The gravitational field attracts all UAVs to move towards the preset target point, and the repulsive field repels all UAVs from moving towards the obstacle and repels each UAV from moving towards other UAVs. The drone control module is used to control the drone swarm by combining a time-varying formation tracking protocol and a collision avoidance controller.

[0011] Embodiments of the present invention also provide a computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described UAV swarm formation cooperative control method.

[0012] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described UAV swarm formation cooperative control method.

[0013] The UAV swarm formation cooperative control method provided by this invention has at least the following beneficial effects: First, we establish the dynamic model of each UAV in the UAV swarm, the dynamic model of obstacles existing in the preset UAV swarm flight trajectory, and the information exchange topology model to describe the data interaction path between the UAVs in the UAV swarm. Based on this, we design a time-varying formation tracking protocol to enable the UAV swarm to form a formation that is dynamically adjusted according to mission requirements and to track the flight trajectory that is dynamically changed according to mission requirements (i.e., time-varying formation and dynamic tracking). At the same time, we design a collision avoidance controller based on the artificial potential field method to enable each UAV to avoid collisions with obstacles and other UAVs during flight.

[0014] As can be seen, this invention considers the flight state of each UAV and the interaction between them by defining the dynamic model of each UAV and the information exchange path between them. This time-varying formation tracking protocol, designed in this way, does not rely on a centralized control node but coordinates the behavior of a group of independent agents (i.e., UAVs) to achieve a common goal. Compared with centralized control, this distributed control scheme can effectively cope with the flexibility of dynamic task environments and avoid the collapse of the entire cluster system due to the failure of the virtual leader at the core control point. Simultaneously, a collision avoidance controller based on an artificial potential field is designed, which, combined with the aforementioned time-varying formation tracking protocol, can improve the safety of the UAV cluster during flight. Attached Figure Description

[0015] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0016] Figure 1 A flowchart illustrating a method for collaborative control of unmanned aerial vehicle (UAV) swarm formations provided by the present invention; Figure 2 This is a schematic diagram of two types of local minimum points provided by the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0019] One embodiment of the present invention relates to a method for cooperative control of unmanned aerial vehicle (UAV) swarm formations. The specific process of the UAV swarm formation cooperative control method in this embodiment can be as follows: Figure 1 As shown, it includes: Step 101: Establish a dynamic model for each drone in the drone swarm, and establish a dynamic model for obstacles existing in the preset drone swarm flight trajectory.

[0020] Step 102: Establish an information exchange topology model to describe the data interaction paths between drones in the drone cluster.

[0021] Step 103: Based on the information exchange topology model and the dynamic model of each UAV, design a time-varying formation tracking protocol to control all UAVs in the UAV cluster to form a preset formation that dynamically changes according to mission requirements and to track a preset flight trajectory that dynamically changes according to mission requirements during flight.

[0022] Step 104: Based on the artificial potential field method, according to the information exchange topology model, the dynamic model of each UAV and the dynamic model of the obstacle, design a collision avoidance controller that includes a gravitational field and a repulsive field. The gravitational field attracts all UAVs to move towards the preset target point, and the repulsive field repels all UAVs from moving towards the obstacle and repels each UAV from moving towards other UAVs.

[0023] Step 105: Combine time-varying formation tracking protocol and collision avoidance controller to control the drone swarm.

[0024] This embodiment considers the analysis and design problem of time-varying formation tracking in a second-order multi-agent system with a switching interaction topology, where the states of followers form a predefined time-varying formation while tracking the state of the leader. A formation tracking protocol is constructed based on the relative information of neighboring agents. Necessary and sufficient conditions for achieving time-varying formation tracking in a multi-agent system with a switching interaction topology are proposed, and feasibility constraints for formation tracking are proposed using graph theory. A method for designing a formation tracking protocol is proposed by solving the algebraic Riccati equation, and the stability of the method is proved using the common Lyapunov stability theory. The results are applied to solve the target encirclement problem of a multi-rotor UAV system consisting of a leader (target) quadrotor UAV and ten follower quadrotor UAVs.

[0025] The following is a detailed description of the implementation details of the UAV swarm formation cooperative control method in this embodiment. The following content is only for the convenience of understanding and is not necessary for implementing this solution.

[0026] First, we introduce graph theory and establish a mathematical model of UAVs. Then, we introduce related basic knowledge such as formation consensus protocol.

[0027] set up It is a A weighted directed graph with vertices, where , , respectively representing the vertex set, edge set, and connection set. The relevant weighted adjacency matrix. Let... for An edge in the middle, where vertex Called vertex The neighbors. For any ,if ,but ,otherwise, .like and Then it is called The edge is undirected. The neighborhood collection is recorded as .vertex The in-degree is defined as .set up for The degree matrix. The Laplace matrix is ​​defined as .from arrive The path is in the form of an ordered series of edges, that is... A graph is called a graph if there exists a path from at least one vertex to all other vertices. It has a spanning tree.

[0028] The graph relating the multi-agent system interaction topology in this embodiment can be switched. Let... Represents all possible graphs, with its index set as follows: ,in Let represent the set of natural numbers. It is a switching signal with a value of Index of the timeline. Definition From the vertex arrive The weights of the edges. (Using...) and express The graph at time t and its corresponding Laplacian matrix. Let... as an agent exist The neighbor set at that time. Assume that the allowed switching signals have a dwell time. .

[0029] Laplace matrix It has the following properties: (1) The sum of the rows is zero, and zero is A feature root of , and the elements in its corresponding feature vector are all 1; (2) If the figure If it is either strongly connected or undirected connected, then zero is its... A single characteristic root; (3) If the figure If it contains only one set of directed spanning trees, then the negative value is... There is only one zero eigenvalue, and the real parts of all other non-zero eigenvalues ​​are negative.

[0030] Definition 1: Given a matrix , For matrix If the elements are in the matrix, then the matrix... sum matrix The Kronecker product is: .

[0031] Mathematical model of drones: Consider a containing A multi-agent system for unmanned aerial vehicles. The switching interaction topology between drones can be represented by a diagram. To describe, each of the drones is One of the vertices, drones To drones The interaction is by the edge This is used to define a drone as follows: If a drone has no neighbors, it is called a leader; if it has at least one neighbor, it is called a follower. Assume there exists a leader and... One follower. Let... For the index set of followers. The control objective is to make all Each follower forms a predefined, time-varying formation while tracking the leader's trajectory. The leader's dynamic characteristics are described by the following equation:

[0032] ; in and These are the position vector and velocity vector of leader 1, respectively. Represents the dimension of space. and It is a known damping constant. The dynamics of a follower can be modeled as follows:

[0033] ; in , and They are followers Position, velocity, and control input vectors.

[0034] In drone formation control, consistency is the core objective to ensure that multiple drones achieve overall behavioral coordination through local information exchange. For example, in drone formation control... A formation system consisting of drones is assumed to have a dynamic model in the form of a second-order integrator:

[0035] ; in The first The position and speed of the drone To control the input, a distributed control protocol needs to be designed to achieve time-varying formation tracking. This enables drones to meet the following requirements:

[0036] ; in This represents the offset of the time-varying formation reference trajectory.

[0037] Consistency control protocols are typically based on adjacency communication topology design. This involves defining an adjacency matrix. ,in Indicates drone and drones A communication connection exists; otherwise, it is... Combining the Laplace matrix Combined with formation error, a typical consistency control law can be expressed as:

[0038] ; in For coupling gain, This represents the desired relative formation offset.

[0039] Next, to achieve time-varying formation control under a switching topology based on a leader-follower model, the necessary and sufficient conditions for realizing time-varying formation tracking in a multi-agent system with a switching interaction topology are proposed using the leader-follower model. Feasibility constraints for formation tracking are also proposed using graph theory. A method for designing formation tracking protocols is proposed by solving the algebraic Riccati equation, and the stability of this method is proven using the common Lyapunov stability theory.

[0040] The time-varying formation tracking protocol is designed as follows: The time-varying formation matrix that followers need to form is composed of Specify, where Followers Piecewise continuously differentiable formation vectors and They are The components corresponding to position and velocity. Let... .

[0041] Definition 2: If for any given bounded initial state: ; This means that the multi-agent system has achieved time-varying formation tracking.

[0042] For multi-agent systems with switching topologies, the following time-varying formation tracking protocol is designed: ; in Let be a constant gain matrix. Let be... , Considering the leader-follower topology, the Laplace matrix... It has the following block format:

[0043] ; in and Let represent the interaction relationships from leader to follower and among followers, respectively. Under the influence of the protocol, the dynamics of the closed-loop system can be compactly represented as:

[0044] ; The second-order model includes position and velocity feedback terms (i.e., damping terms) for each agent. The role of the damping terms is to... The eigenvalues ​​are configured to the desired location in the complex plane to specify the leader's motion pattern. Because... Controllable; a suitable damping constant can always be found for any given motion mode. It is worth noting that leader dynamics does not contain external control inputs, therefore, given... Once the initial state is reached, the leader's trajectory can be pre-calculated. However, the leader's trajectory is unknown for each follower.

[0045] Necessary and sufficient conditions for time-varying formation tracking and their proof: Assumption 1 (Topological condition): ① For each possible interaction topology any follower There exists at least one path originating from leader 1; ② The interaction topology among the followers is an undirected graph.

[0046] Lemma 1 (Property of the Laplace Matrix): like If it contains a spanning tree, then its Laplace matrix is... satisfy: ; in It is a vector of all 1s. Sort by real part in ascending order.

[0047] Lemma 2 (Hierarchical Decomposition): Under assumption 1, the Laplace matrix can be decomposed as follows: ; in For a symmetric positive definite matrix, its eigenvalues ​​satisfy: ; Theorem 3 (Necessary and sufficient condition for time-varying formation tracking): A multi-agent system can achieve time-varying formation tracking under a time-varying formation tracking protocol if and only if the following condition is met: ① Formation feasibility constraints: ; ② Switching system stability: system ; Global asymptotic stability.

[0048] prove: make The multi-agent system can be rewritten as: ; Define the global error vector ; in The state of all intelligent agents. This is the formation vector.

[0049] Define the coordinate transformation matrix: ; Its inverse matrix is: ; The original system is decoupled using a transformation matrix: ; in For the relative error of the follower: ; The original dynamic equations can be expanded as follows: ; Among them, the disturbance term is: ; After applying coordinate transformation, the block diagonal form is obtained: ; The error system is decoupled as follows: ① Leader Subsystem (Uncontrolled): ; Leader dynamics are determined by the damping coefficient. The trajectory is determined solely by the initial state and inherent dynamics, without any external control input. Controllable, can be selected Configure the leader's motion mode (e.g., stable / oscillating).

[0050] ② Tracking error subsystem: ; Disturbance term Time-varying formation The non-ideal properties of cause it to tend to 0, and the necessary and sufficient condition for it to approach 0 is: ; Formation speed component The derivative of the position component must be asymptotically tracked. This ensures that time-varying formations are compatible with system dynamics.

[0051] Considering the error dynamic system, since Symmetric, with orthogonal matrices. Diagonalize it: ; in for The eigenvalues. Define new coordinates. ,

[0052] The system dynamic equations become: ; Define the common Lyapunov quadratic form function: ; Construction of Algebraic Riccati Equations: Choosing a Weight Matrix Please provide a solution: ; in Controllability Guarantee Solution exist.

[0053] pass Positive definite matrices provide a unified measure of error energy under different topologies, offering a common stability criterion for switching systems.

[0054] Calculate the derivative along the orbit: for Differentiate and substitute into the error dynamic equation: ; Will After substituting the expression and expanding it: ; Control gain design: Select the control gain. ; in , To adjust the robustness parameters, using The smallest non-zero eigenvalue of all possible topologies is given a unified upper bound, which guarantees that the energy function monotonically decreases during the switching process.

[0055] Will Substituting, we get: ; Combining terms in the algebraic Riccati equation: ; Substitution have to: ; Analyze the sign of the derivative: ; Symbol analysis of item 1: due to and ,have: ; Therefore, term 1 is a negative definite matrix.

[0056] Contribution: due to The global derivative satisfies: ; Theorem 4: Lassalle Invariance Principle when It must meet the following requirements: ① For all Established (by) (semi-negative qualitative) ②The system status at this time .

[0057] According to Lassalle's invariance principle, the trajectory of the system will converge to the maximally invariant set. That is, the system is globally asymptotically stable.

[0058] This embodiment first designs a time-varying formation tracking controller with switching topologies based on the leader-follower model. Then, it proposes the necessary and sufficient conditions for the stability of time-varying formation tracking and proves the necessary and sufficient conditions for the stability of time-varying formation tracking through steps such as coordinate transformation, system decoupling, solving the algebraic Riccati equation, constructing Lyapunov functions, and verifying stability conditions.

[0059] Meanwhile, inspired by flocks of birds and similar creatures, this invention proposes a distributed collision avoidance controller based on the artificial potential field method. By constructing a potential energy function, the target and the obstacle are repelled, guiding the drone away from the obstacle.

[0060] To design a collision avoidance strategy based on the artificial potential field method, a collision avoidance controller is constructed based on an improved artificial potential field method. During flight, each UAV is subjected not only to the virtual repulsive force of other UAVs but also to the repulsive force of obstacles. At the same time, a group coordination component consisting of velocity matching, centripetal aggregation, and random perturbation is added to avoid falling into the minimum trap of the traditional artificial potential field method.

[0061] The principle of obstacle avoidance using the artificial potential field method is as follows: The artificial potential field method is a local path planning method based on a virtual force field. Its core idea is to enable autonomous obstacle avoidance of mobile robots by constructing an abstract potential field environment. This method treats the target point as an attractive source and obstacles as repulsive sources, guiding the robot to move along the descent direction of the potential field gradient through the virtual resultant force generated by the superposition of potential fields.

[0062] The gravitational potential field generated at the target point is positively correlated with the current position of the UAV, and its function is defined as follows: ; In the formula, For drones The current position coordinates, The coordinates of the target point, Let be the gravitational gain function. The corresponding gravitational vector can be expressed as:

[0063] ; The repulsive potential field generated by an obstacle is limited to its radius of influence. Inside, the function is defined as ; in, For drones Euclidean distance to the obstacle The repulsion gain coefficient is given by the corresponding repulsion vector expression. ; drones The net potential field is a linear superposition of the gravitational field and all repulsive fields. ; The corresponding motion control direction is determined by the negative gradient direction of the potential field. ; drones will along Motion decisions are made based on vector direction, and a potential field equilibrium state is reached when the net external force is zero.

[0064] Design of a distributed collision avoidance controller for a single UAV: The fundamental reason limiting the application of traditional artificial potential field methods is that obstacle distribution easily leads to local minima in the potential field, causing UAVs to get trapped in potential field traps and fail in path planning. The main types of local minima include... Figure 2 As shown, local minima in both single-obstacle and multi-obstacle scenarios can make it difficult for drones to reach the target location, thus requiring optimization in obstacle avoidance problems.

[0065] There are two common solutions. The first is to use search strategies such as optimal priority or simulated annealing to try to "jump out" of a local minimum after getting stuck in it. Some scholars have tried to increase the kinetic energy of the moving object to break through the potential field barrier. However, this undoubtedly places more stringent demands on the performance of the moving object, and the resulting trajectory is not globally optimal. Furthermore, the increased kinetic energy must be between the barrier of the local minimum and the potential energy of the local maximum. The second solution is to use mathematical methods to eliminate local minima outside the target area in advance. Before planning, all local minima in the space are solved in advance. By adding a local repulsive potential field at the local minima, the total potential field function is adjusted to "fill" the concave potential field, allowing the moving object to pass through the trap smoothly, thereby eliminating the local minima.

[0066] This embodiment designs a distributed collision avoidance controller that replaces the gravitational term in the traditional artificial potential field method with a group coordination term. By adding velocity matching, centripetal aggregation and random perturbation, it avoids the UAV from getting trapped in the potential field trap, thus enabling the UAV to flexibly avoid unknown and dynamic threats or obstacles during flight.

[0067] Inspired by the behavior of flocks of birds avoiding predators, the improved method involves birds perceiving their neighbors by detecting the edges of light and shadow cast by those neighbors. The birds then use these perceived shadows to adjust their flight speed. Because each bird adjusts its speed based on the behavior of its neighbors, the entire process exhibits typical group characteristics.

[0068] To simulate this mechanism, we designed the following collision avoidance controller for each drone, taking into account the distance and relative speed to obstacles, the relative speed and safe distance between itself and neighboring drones, and noise.

[0069] The distributed collision avoidance controller takes the following form: ; Collision avoidance controller input It consists of the following three parts: ① Obstacle avoidance items : When drones and obstacles The distance is less than the safety threshold At that time, an acceleration opposite to the direction of the obstacle's motion is applied: ; in , For the set of all obstacles, To improve obstacle avoidance gain, the intensity of the reverse acceleration is controlled. For obstacles The speed.

[0070] ② Group Coordination Items : This design ensures the aggregation and collision avoidance of the drone swarm. The core concept is to guide the drones towards areas with lower formation density. Group consistency is maintained through speed matching, centripetal aggregation, and random perturbation. Speed ​​matching ensures that each drone maintains the same speed as its neighbors, guaranteeing coordinated group movement and preventing collisions or dispersal due to speed differences, similar to individual birds adjusting their speed to synchronize their flight direction. Centripetal aggregation, by sensing the distribution of surrounding drones, guides them towards areas of lower density, avoiding excessive local aggregation. For example, integrating the direction of surrounding shadows encourages dispersal, reducing collision risk, similar to the natural dispersal behavior of birds avoiding obstacles. Random perturbation introduces randomness, preventing the group from falling into a mechanized movement pattern and enhancing the system's adaptability to dynamic environments. Noise simulates the subtle random behaviors of individuals in natural flocks (such as slight turns in bird flight), improving the robustness and flexibility of the obstacle avoidance strategy.

[0071] ; in For drones The set of neighbors (defined by the communication topology). For the location of the neighbor, These are weighting coefficients that balance the contributions of each component, and , Gaussian white noise is used to enhance collision avoidance robustness. These are the elements of the adjacency matrix.

[0072] ③ Collision avoidance between drones : When drones with neighbors The distance is less than the internal safety distance. At this time, a repulsive force is generated: ; Direction by drone Pointing to drones To avoid collisions This represents the collision avoidance gain coefficient between unmanned aerial vehicles (UAVs).

[0073] Divided into two modes: ① Obstacle Avoidance Mode (Obstacle detected): Activated Combining group coordination and drone collision; ② Normal Mode: Enabled only To maintain the safety of group activities.

[0074] In practical implementation, the design of a multi-UAV controller considering formation obstacle avoidance is as follows: The design of a multi-UAV controller that considers formation obstacle avoidance combines formation control and collision avoidance control, and performs fusion control based on weighted smooth transition.

[0075] ; in .

[0076] (1) Obstacle avoidance mode control input: ; (2) Formation mode control input ; in This is an indicator function; its value is 1 if the condition is true, and 0 otherwise. For drones The set of neighbors (defined by a fixed topology). For the neighbor's central location, These are the weighting coefficients for velocity matching, centroid projection, and noise. For standard normally distributed random noise, This is the velocity vector of the obstacle (this term disappears when the velocity is zero). These are the elements of the adjacency matrix.

[0077] This embodiment first designs a collision avoidance controller for a single UAV based on the artificial potential field method, and then designs a comprehensive collision avoidance controller for UAV formations.

[0078] In summary, considering the limitations of a single UAV in terms of operational range and fault tolerance, and the research value and application prospects of multi-UAV systems in both civilian and military fields, this invention achieves obstacle avoidance control for UAV formations, mainly including: (1) Considering the multi-UAV formation control problem, this invention simplifies the UAV flight control model into a second-order integral system and uses graph theory to describe the information exchange topology between UAVs. For the UAV formation control problem, this invention designs a time-varying formation controller based on a leader-follower model with a switching topology.

[0079] (2) Considering the collision problem of UAV formation, this invention introduces the obstacle avoidance principle of artificial potential field method, and considers the speed matching, centripetal aggregation and random disturbance between UAVs in the collision avoidance process to maintain the consistency of the group, based on the collision avoidance of UAVs to obstacles. It also considers the collision avoidance between UAVs.

[0080] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this invention. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, without changing the core design of the algorithm and process, are also within the protection scope of this invention.

[0081] Another embodiment of the present invention relates to a drone swarm formation cooperative control system. The implementation details of this drone swarm formation cooperative control system are described below. The following details are provided for ease of understanding and are not essential for implementing this solution. The drone swarm formation cooperative control system of this embodiment includes: The first model building module is used to build a dynamic model of each drone in the drone swarm, and to build a dynamic model of obstacles existing in the preset drone swarm flight trajectory; The second model building module is used to build an information exchange topology model that describes the data interaction paths between drones in a drone cluster. The formation protocol design module is used to design a time-varying formation tracking protocol based on the information exchange topology model and the dynamic model of each UAV, so as to control all UAVs in the UAV cluster to form a preset formation that dynamically changes according to mission requirements and track a preset flight trajectory that dynamically changes according to mission requirements during flight. The collision avoidance control design module is used to design a collision avoidance controller that includes a gravitational field and a repulsive field, based on the artificial potential field method, the information exchange topology model, the dynamic model of each UAV, and the dynamic model of the obstacle. The gravitational field attracts all UAVs to move towards the preset target point, and the repulsive field repels all UAVs from moving towards the obstacle and repels each UAV from moving towards other UAVs. The drone control module is used to control the drone swarm by combining a time-varying formation tracking protocol and a collision avoidance controller.

[0082] It is not difficult to see that this embodiment is a system embodiment corresponding to the above method embodiments, and this embodiment can be implemented in conjunction with the above method embodiments. The relevant technical details and technical effects mentioned in the above embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above embodiments.

[0083] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0084] Another embodiment of the present invention relates to a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the UAV swarm formation cooperative control method of the above embodiments.

[0085] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0086] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0087] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0088] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0089] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A method for cooperative control of unmanned aerial vehicle (UAV) swarm formations, characterized in that, The method includes: Establish a dynamic model for each drone in the drone swarm, and establish a dynamic model for obstacles existing in the pre-defined flight path of the drone swarm; Establish an information exchange topology model to describe the data interaction paths between drones in a drone swarm; Based on the information exchange topology model and the dynamic model of each UAV, a time-varying formation tracking protocol is designed to control all UAVs in the UAV swarm to form a preset formation that dynamically changes according to mission requirements during flight and to track a preset flight trajectory that dynamically changes according to mission requirements. Based on the artificial potential field method, and according to the information exchange topology model, the dynamic model of each UAV and the dynamic model of the obstacle, a collision avoidance controller containing a gravitational field and a repulsive field is designed. The gravitational field attracts all UAVs to move towards the preset target point, and the repulsive field repels all UAVs from moving towards the obstacle and repels each UAV from moving towards other UAVs. By combining a time-varying formation tracking protocol and a collision avoidance controller, the drone swarm can be controlled.

2. The UAV swarm formation cooperative control method according to claim 1, characterized in that, The time-varying formation tracking protocol is as follows: ； In the formula, It is a weighted adjacency matrix, and the desired time-varying formation is composed of vectors. Specify, Followers A continuously differentiable piecewise formation vector. and These represent the agent's desired positional offset and velocity offset relative to the leader within the formation, respectively. The extended state is , Let be a constant gain matrix, denoted as , .

3. The UAV swarm formation cooperative control method according to claim 1, characterized in that, The collision avoidance controller is: ； in, For obstacle avoidance, and , , For the set of all obstacles, To improve obstacle avoidance, Obstacles speed; It is a group coordination term, and ; For drones The neighborhood group, For the location of the neighbor, These are the weighting coefficients. , It is Gaussian white noise. These are elements of the adjacency matrix; For collision avoidance between drones, , This represents the collision avoidance gain coefficient between unmanned aerial vehicles (UAVs).

4. The UAV swarm formation cooperative control method according to claim 1, characterized in that, The method of controlling a drone swarm by combining a time-varying formation tracking protocol and a collision avoidance controller includes: The weights of the time-varying formation tracking protocol and the collision avoidance controller are determined based on the distance between each drone and the obstacle. Each drone in the drone swarm is controlled according to the time-varying formation tracking protocol and the weights of the collision avoidance controller.

5. The UAV swarm formation cooperative control method according to claim 4, characterized in that, The following formula is used to control each drone in the drone swarm: ； ； ； In the formula, , , , This is an indicator function; its value is 1 if the condition is true, and 0 otherwise. For drones The neighborhood group, For the neighbor's central location, These are the weighting coefficients for velocity matching, centroid projection, and noise. For standard normally distributed random noise, Let V be the velocity vector of the obstacle. These are the elements of the adjacency matrix.

6. A collaborative control system for unmanned aerial vehicle (UAV) swarm formations, characterized in that, The system includes: The first model building module is used to build a dynamic model of each drone in the drone swarm, and to build a dynamic model of obstacles existing in the preset drone swarm flight trajectory; The second model building module is used to build an information exchange topology model that describes the data interaction paths between drones in a drone cluster. The formation protocol design module is used to design a time-varying formation tracking protocol based on the information exchange topology model and the dynamic model of each UAV, so as to control all UAVs in the UAV cluster to form a preset formation that dynamically changes according to mission requirements and track a preset flight trajectory that dynamically changes according to mission requirements during flight. The collision avoidance control design module is used to design a collision avoidance controller that includes a gravitational field and a repulsive field, based on the artificial potential field method, the information exchange topology model, the dynamic model of each UAV, and the dynamic model of the obstacle. The gravitational field attracts all UAVs to move towards the preset target point, and the repulsive field repels all UAVs from moving towards the obstacle and repels each UAV from moving towards other UAVs. The drone control module is used to control the drone swarm by combining a time-varying formation tracking protocol and a collision avoidance controller.

7. A computer device, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the UAV swarm formation cooperative control method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the UAV swarm formation cooperative control method as described in any one of claims 1 to 5.

Images