Intelligent cooperative control method and system based on unmanned aerial vehicle group

Abstract

The invention discloses an intelligent cooperative control method and system based on an unmanned aerial vehicle group, and relates to the technical field of unmanned aerial vehicle group intelligent control, and the method comprises the steps: collecting multi-source data, constructing a sparse tensor, defining unmanned aerial vehicles as nodes, and constructing a dynamic category graph; nodes of the dynamic category graph are mapped to manifold points of a Riemannian manifold, coordinates of the unmanned aerial vehicle are converted into grid indexes, an objective function is defined, the Riemannian gradient of the objective function is calculated, manifold points are updated through index mapping, and updated space coordinates and task vectors are obtained; and generating a control instruction vector by using the weighted linear combination, and converting the control instruction vector into a control instruction and executing the control instruction. By constructing a sparse tensor and a dynamic category graph and combining Riemannian manifold optimization to balance task priorities and collaborative constraints, the task coverage rate and collaborative robustness are improved, Riemannian gradient optimization is combined with index mapping update, and the autonomous collaborative ability is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control of unmanned aerial vehicle swarm, in particular to an intelligent collaborative control method and system based on unmanned aerial vehicle swarm. BACKGROUND

[0002] With the progress of sensor technology, embedded computing power and communication network, unmanned aerial vehicle swarm can perform complex collaborative tasks such as search and rescue, environmental monitoring, and precision management in agriculture. Information fusion based on multi-source sensors (such as vision, radar, and inertial measurement unit), collaborative path planning based on distributed algorithms, and behavior control methods based on swarm intelligence have become the mainstream technology route for collaborative control of unmanned aerial vehicle swarm. However, in the face of high dynamic, nonlinear, and complex and variable task environment, existing collaborative control methods still face many challenges.

[0003] The existing unmanned aerial vehicle swarm collaborative control technology still has some deficiencies. The existing technology is based on a navigation method using Euclidean geometry, which is difficult to achieve effective task allocation and path planning under the driving of high-dimensional information flow, and the system flexibility is low. Existing optimization methods are mostly based on Euclidean space, ignoring the nonlinear geometric characteristics of the task space, and cannot effectively balance the task priority and collaborative constraints, thereby affecting the global optimality of path planning. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides an intelligent collaborative control method and system based on unmanned aerial vehicle swarm, which solves the problem that the existing technology is based on a navigation method using Euclidean geometry, which is difficult to achieve effective task allocation and path planning under the driving of high-dimensional information flow, and the system flexibility is low. Existing optimization methods are mostly based on Euclidean space, ignoring the nonlinear geometric characteristics of the task space, and cannot effectively balance the task priority and collaborative constraints, thereby affecting the global optimality of path planning.

[0006] To solve the above technical problems, the present application provides the following technical solutions: In a first aspect, the present application provides an intelligent collaborative control method based on unmanned aerial vehicle swarm, which comprises, Collecting multi-source data and preprocessing, constructing a sparse tensor, using finite difference method to calculate the partial derivative of each dimension in the sparse tensor, using vector synthesis method to combine the partial derivative as the gradient vector field of information entropy, defining the unmanned aerial vehicle as a node, and constructing a dynamic category graph. Map the nodes of the dynamic category graph to the popular points of the Riemann manifold, convert the UAV coordinates into grid indexes, calculate the task intensity influence of the UAV, calculate the metric tensor, add a regularization term to the metric tensor, define the objective function, calculate the Riemann gradient of the objective function, update the manifold points using the exponential mapping, and obtain the updated spatial coordinates and task vectors; Generate a control instruction vector using a weighted linear combination, convert the control instruction vector into control instructions, and execute them; Construct a visual interface to display the control instructions.

[0007] As a preferred scheme of the intelligent collaborative control method based on a UAV group, the multi-source data is collected, a sparse tensor is constructed, the UAV is defined as a node, and a dynamic category graph is constructed, including: Collect multi-source data of the UAV group using intelligent sensors and perform denoising and normalization processing; The multi-source data includes images, three-dimensional point clouds, spatial coordinates, velocities, and acceleration data; Fuse the spatial coordinates, velocities, and acceleration data using an extended Kalman filter, cluster the fused data using a K-means clustering algorithm, identify the images using a pre-trained Swin-Transformer model, output the corresponding state variables, stack the spatial coordinates, three-dimensional point clouds, velocities, accelerations, clustering results, and state labels using tensor stacking to construct a sparse tensor, sum the non-zero elements of the sparse tensor, calculate the task intensity probability, calculate the information entropy based on the task intensity probability, calculate the partial derivatives of each dimension in the sparse tensor using the finite difference method based on the information entropy, and combine the partial derivatives using vector synthesis as the gradient vector field of the information entropy; Extract the target grid coordinates from the sparse tensor, convert the grid coordinates into actual geographic coordinates, calculate the direction vector of the UAV to the target location, normalize the direction vector, and obtain the task vector; Calculate the control mapping function between the UAVs, set a screening threshold using the fixed threshold method, and define the screened control mapping function as a control morphism; Define the UAV as a node, define the control morphism greater than 0 as an edge, construct an initial category graph, set a change threshold using statistical analysis, calculate the Euclidean distance between the spatial position of each UAV and the previous time, screen the Euclidean distances greater than the change threshold, mark them as affected control morphisms, and recalculate the control morphisms, otherwise retain the control morphisms of the previous time, and generate a dynamic category graph.

[0008] As a preferred scheme of the intelligent cooperative control method based on the UAV group, wherein: the nodes of the dynamic category graph are mapped to the popular points of the Riemann manifold, the UAV coordinates are converted into grid indexes, a target function is defined, and the target function comprises: Based on the UAV, a Riemann manifold is constructed, the Euclidean distance between the manifold points is calculated using the Euclidean distance formula, and the Euclidean metric is defined. The initial metric tensor is set using the Euclidean metric, and the nodes of the dynamic category graph are mapped to the popular points of the Riemann manifold. The edges of the dynamic category graph are used as adjacency constraints on the manifold to construct an adjacency matrix on the manifold. The UAV coordinates are converted into grid indexes using a grid mapping table, the task intensity influence of the UAV is calculated, the metric tensor is calculated, and a regularization term is added to the metric tensor by positive definite regularization. The target function is defined using a weighted target construction.

[0009] As a preferred scheme of the intelligent cooperative control method based on the UAV group, wherein: the Riemann gradient of the target function is calculated, the manifold points are updated using exponential mapping, and the updated spatial coordinates and task vectors are obtained, comprising: The Riemann gradient of the target function is calculated by combining the gradient vector and the metric tensor, and the manifold points are updated based on the Riemann gradient using exponential mapping. The updated spatial coordinates and task vectors are decomposed from the updated manifold points using vector decomposition.

[0010] As a preferred scheme of the intelligent cooperative control method based on the UAV group, wherein: the control instruction vector is generated using weighted linear combination, comprising: The position increment is calculated using vector subtraction, the control instruction vector is generated using weighted linear combination, and normalization processing is performed.

[0011] As a preferred scheme of the intelligent cooperative control method based on the UAV group, wherein: the control instruction vector is converted into control instructions and executed, comprising: The normalized control instruction vector is converted into control instructions using vector set encapsulation; The adjacency matrix of the dynamic category graph is extracted, the communication channels of the UAVs are allocated using a wireless communication protocol, the control instructions are sent to the corresponding UAVs, and the UAVs receive the control instructions and execute them.

[0012] As a preferred scheme of the intelligent cooperative control method based on the UAV group, wherein: the control instructions are displayed on a visual interface, comprising: A visual interface is constructed using the front-end framework React.js, and the control instructions obtained by analysis and the multi-source data collected are visualized and displayed. Allow the user after real-name verification to consult.

[0013] In a second aspect, the present application provides an intelligent collaborative control system based on a UAV group, comprising, A collection composition module is configured to collect multi-source data, pre-process the data, construct a sparse tensor, calculate partial derivatives of each dimension in the sparse tensor using a finite difference method, combine the partial derivatives using a vector synthesis method, use the combined partial derivatives as a gradient vector field of information entropy, define a UAV as a node, and construct a dynamic category graph. A target updating module is configured to map nodes of the dynamic category graph to popular points of a Riemann manifold, convert UAV coordinates into grid indexes, calculate a task intensity influence of the UAV, calculate a metric tensor, add a regularization term to the metric tensor through positive definite regularization, define an objective function, calculate a Riemann gradient of the objective function, update the manifold points using an exponential mapping, and obtain updated spatial coordinates and a task vector. An instruction execution module is configured to generate a control instruction vector using a weighted linear combination, convert the control instruction vector into control instructions, and execute the control instructions. A visualization module is configured to construct a visualization interface to display the control instructions.

[0014] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program, when executed by the processor, implements any step of the intelligent collaborative control method based on a UAV group according to the first aspect of the present application.

[0015] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step of the intelligent collaborative control method based on a UAV group according to the first aspect of the present application.

[0016] The present application has the following advantages: the present application collects multi-source data, pre-processes the data, constructs a sparse tensor, calculates partial derivatives of each dimension in the sparse tensor using a finite difference method, combines the partial derivatives using a vector synthesis method, uses the combined partial derivatives as a gradient vector field of information entropy, defines a UAV as a node, and constructs a dynamic category graph; the present application maps nodes of the dynamic category graph to popular points of a Riemann manifold, converts UAV coordinates into grid indexes, calculates a task intensity influence of the UAV, calculates a metric tensor, adds a regularization term to the metric tensor through positive definite regularization, defines an objective function, calculates a Riemann gradient of the objective function, updates the manifold points using an exponential mapping, and obtains updated spatial coordinates and a task vector; the present application generates a control instruction vector using a weighted linear combination, converts the control instruction vector into control instructions, and executes the control instructions; the present application solves the deficiencies of traditional methods in complex dynamic environments, significantly improves task coverage and collaborative robustness, and enhances autonomous collaboration capabilities. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0018] Fig. 1 Flow chart of the intelligent cooperative control method based on the UAV group in embodiment 1.

[0019] Fig. 2 Schematic diagram of the intelligent cooperative control system based on the UAV group in embodiment 1.

[0020] Fig. 3 Schematic diagram of the cooperative control in the intelligent cooperative control method based on the UAV group in embodiment 1. DETAILED DESCRIPTION

[0021] In order to make the above objectives, features and advantages of the present application more apparent, the specific embodiments of the present application will be described in detail in conjunction with the drawings in the specification.

[0022] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be practiced without the specific details, other than those described herein, and it is understood that the present application will encompass a variety of implementations beyond those described herein. Accordingly, the present application is not limited to the embodiments described herein.

[0023] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. The "in one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0024] Embodiment 1, refer to Figs. 1 to 3 , the first embodiment of the present application, the embodiment provides an intelligent cooperative control method based on the UAV group, including the following steps: S1, collect multi-source data and pre-process, construct sparse tensor, use finite difference method to calculate the partial derivative of each dimension in the sparse tensor, use vector synthesis method to combine the partial derivative as the gradient vector field of information entropy, define the UAV as a node, and construct a dynamic category graph; Specifically, multi-source data is collected and preprocessed, a sparse tensor is constructed, the partial derivatives of each dimension in the sparse tensor are calculated using the finite difference method, the partial derivatives are combined using the vector composition method as the gradient vector field of information entropy, a dynamic category graph is constructed by defining the unmanned aerial vehicle as a node, including: Collecting multi-source data of the unmanned aerial vehicle group using intelligent sensors and performing denoising and normalization processing; The multi-source data includes images, three-dimensional point clouds, spatial coordinates, velocity, and acceleration data; The intelligent sensors include cameras, lidar, GPS, and IMU (inertial measurement unit) sensors; The spatial coordinates, velocity, and acceleration data are fused using an extended Kalman filter, the fused data is clustered using a K-means clustering algorithm, the number of clusters is set using the contour coefficient method, a pre-trained Swin-Transformer model is used to recognize the images, outputting corresponding state variables, the spatial coordinates, three-dimensional point clouds, velocity, acceleration, clustering results, and state labels are stacked using tensor stacking to construct a sparse tensor, the non-zero elements of the sparse tensor are summed to calculate the task intensity probability, the information entropy is calculated based on the task intensity probability, the partial derivatives of each dimension in the sparse tensor are calculated using the finite difference method based on the information entropy, and the partial derivatives are combined using the vector composition method as the gradient vector field of information entropy; The target grid coordinates corresponding to the maximum intensity unit in the tensor are extracted from the sparse tensor, the grid coordinates are converted to actual geographic coordinates (converted through a grid mapping table), the direction vector of the unmanned aerial vehicle to the target location is calculated, and the direction vector is normalized to obtain the task vector; The control mapping function between the unmanned aerial vehicles is calculated, a fixed threshold method is used to set a screening threshold, and the screened control mapping function is defined as a control morphism; The unmanned aerial vehicle is defined as a node, the control morphism greater than 0 is defined as an edge, an initial category graph is constructed, a statistical analysis method is used to set a change threshold, the Euclidean distance between the spatial position of each unmanned aerial vehicle and the previous time is calculated, the Euclidean distances greater than the change threshold are screened, marked as affected control morphisms, and the control morphisms are recalculated, otherwise the control morphisms of the previous time are retained, and a dynamic category graph is generated; The task intensity probability is calculated using a normalized probability distribution based on the sparse task tensor, and the formula is: , Wherein is the task intensity probability of the sparse tensor at time t, is the sparse tensor, is the intensity sum, indicating the sum of the non-zero elements of the sparse tensor, i is the index of the task type dimension, j is the index of the spatial coordinates, and k is the index of the state variable; The information entropy of the sparse tensor is calculated based on the task intensity probability using Shannon entropy, and the formula is: , Wherein is the information entropy of the sparse tensor at time t; The target grid coordinates are extracted, and the formula is: , Wherein is the target grid coordinates of the unmanned aerial vehicle at time t; The direction vector of the unmanned aerial vehicle to the target position is calculated, and the formula is: , Wherein is the direction vector of the lth unmanned aerial vehicle, is the actual geographic coordinates of the unmanned aerial vehicle at time t, is the spatial coordinates of the Jth unmanned aerial vehicle; The direction vector is normalized using the vector normalization method to obtain the task vector. If the task vector is 0, it indicates that the unmanned aerial vehicle has reached the target. Let , and the formula is: , , Wherein is the task vector of the lth unmanned aerial vehicle; The control mapping function between the unmanned aerial vehicles is calculated, and the formula is: , , Wherein is the control mapping function, indicating the control dependence intensity of the lth unmanned aerial vehicle to the uth unmanned aerial vehicle , is the Euclidean distance between the lth and uth unmanned aerial vehicles (calculated using the Euclidean distance formula), is a minimum constant, is the task vector of the uth unmanned aerial vehicle, is the control mapping of the lth unmanned aerial vehicle to the uth unmanned aerial vehicle at time t, indicating the value of the control dependence relationship, is a screening threshold.

[0025] The present application integrates multi-source heterogeneous data into a sparse tensor, combines information entropy and gradient vector field to guide the task allocation of the UAV group, and the existing technology usually processes data (such as images or point clouds) separately or uses high-dimensional dense tensors, resulting in high computational complexity. The present application reduces storage requirements through a sparse tensor, quantifies task uncertainty through information entropy, provides a global driving direction through a gradient vector field, and controls the morphism through real-time updating to adapt to the topological changes of the UAV group. Vector synthesis ensures the collaborative expression of gradient information in each dimension, avoiding path shock or execution ambiguity caused by single-dimensional optimization. The addition of Transformer enhances the context modeling capability of information, especially suitable for identifying target types under complex conditions such as weak light and occlusion. The control morphism can avoid control shock caused by dense graph structure through fixed threshold method to filter effective edges, while retaining the main control path. The dynamic updating mechanism ensures that the structure responds to task evolution or position disturbance, thereby realizing self-organizing update of control relationship in the region and ensuring the coordination and stability of large-scale UAV group under dynamic task driving.

[0026] S2, mapping the nodes of the dynamic category graph to the popular points of the Riemann manifold, converting the UAV coordinates into grid indexes, calculating the task intensity influence of the UAV, calculating the metric tensor, adding a regularization term to the metric tensor through positive definite regularization, defining the objective function, calculating the Riemann gradient of the objective function, updating the manifold points using exponential mapping, and obtaining the updated spatial coordinates and task vectors; Specifically, the nodes of the dynamic category graph are mapped to the popular points of the Riemann manifold, the UAV coordinates are converted into grid indexes, the task intensity influence of the UAV is calculated, the metric tensor is calculated, a regularization term is added to the metric tensor through positive definite regularization, and the objective function is defined, including: Based on the UAV, the manifold dimension is defined, and the formula is: , Wherein is the manifold dimension, and 3 represents the dimension of the spatial coordinates; The popular points are constructed, and the formula is: , Wherein is the spatial coordinates of the lth popular point at time t, is the spatial coordinates of the lth UAV at time t, is the local coordinate space; The Euclidean distance between the manifold points is calculated using the Euclidean distance formula, and the Euclidean metric is defined. The initial metric tensor is set using the Euclidean metric; The nodes of the dynamic category graph are mapped to the popular points of the Riemann manifold, and the formula is: , where is the popular point of the lth UAV on the manifold at time t, containing spatial coordinates and task vector, is a Riemannian manifold; The edges of the dynamic category graph are taken as the adjacency constraints on the manifold, and the adjacency matrix on the manifold is constructed; The task intensity influence of the UAV is calculated by using the grid mapping table to convert the UAV coordinates into grid indices, and the formula is: , where is the task intensity influence of the lth UAV at time t, and the indicator function is 1 only when the grid index of the jth UAV matches the position; The metric tensor is calculated, and the formula is: , where is the metric tensor component at the popular point , Q and R are the indices of the row and column of the manifold dimension respectively, is the Euclidean metric, is the information entropy of the sparse tensor at time t The gradient vector of the UAV position is calculated (using the finite difference method); The positive definite regularization adds a regularization term to the metric tensor, and the formula is: , where I is the identity matrix; The objective function is defined by using the weighted objective construction, and the formula is: , where is the objective function, representing the total cost of the UAV swarm cooperative optimization, is the geodesic distance between the lth and uth popular points on the manifold (calculated based on the metric using the Dijkstra algorithm).

[0027] By embedding the unmanned aerial vehicle and its relationship into the manifold, the system can naturally integrate the curvature and local geometric characteristics, making the cooperative decision more robust in complex task environment, not limited by traditional Euclidean modeling, solving the problem of static modeling in the prior art that cannot capture the dynamics of the group, improving the adaptability to dynamic tasks and uncertainty, through the matrix constraint expression, it is convenient to introduce linear algebra and optimization tools, so that the model has higher calculation efficiency in numerical solution, overcoming the defects of weak dynamic relationship processing ability and lack of unified optimization interface in the existing system, and solving the problem of large continuous optimization calculation amount and insufficient real-time performance in the prior art, at the same time, the task intensity significantly improves the accuracy of task allocation, and the positive definiteness constraint introduced by the metric tensor solves the problem of singular matrix and numerical divergence in high-dimensional optimization in the prior art, ensuring the convergence of the optimization path.

[0028] Further, the Riemannian gradient of the objective function is calculated, and the manifold point is updated using exponential mapping to obtain the updated spatial coordinates and task vector, including: The gradient vector and the metric tensor are combined to calculate the Riemannian gradient of the objective function, and the formula is: , Wherein is the Riemannian gradient of the objective function J at the manifold point at time t, and is the manifold point at time t. The manifold point is updated based on the Riemannian gradient using exponential mapping, and the formula is: , Wherein is the updated manifold point, is the step size (set using line search method); The updated spatial coordinates and task vector are decomposed from the updated manifold point using vector decomposition, and the formula is: , Wherein and are the updated coordinate space and task vector, respectively.

[0029] The definition of the gradient on the manifold makes the optimization step have geometric consistency, avoids invalid or even destructive update direction, solves the problem that the optimization direction in the prior art is not adaptive to the curved surface space, leading to convergence failure, the exponential mapping ensures that the updated point is still on the manifold, without the need for forced projection back to the manifold, the optimization process is more natural and efficient, avoiding the high-cost operation of frequent constraint projection in the existing Euclidean optimization, improving the calculation efficiency and accuracy, ensuring seamless connection from the abstract optimization space to the specific execution space, solving the problem that the optimization result in the prior art is difficult to explain and directly applied to control instructions, improving the practicality of the system.

[0030] S3, generating a control instruction vector using weighted linear combination, converting the control instruction vector into control instructions and performing execution; Specifically, the control instruction vector is generated using weighted linear combination, comprising: The position increment is calculated using vector subtraction (current position of the unmanned aerial vehicle minus the updated coordinate space), the control instruction vector is generated using weighted linear combination, and normalization processing is performed, and the formula is: , Wherein is the instruction vector of the unmanned aerial vehicle at the lth position at time t, is the task vector weight (set using the empirical rule); If , only the position increment is used, if , only the task vector is used .

[0031] Vector subtraction provides a simple and low-error increment acquisition method, which helps to improve real-time performance and accuracy, and normalization is often used in single-objective optimization in the prior art, which is difficult to simultaneously meet the needs of multiple tasks. Weighted linear combination provides a natural multi-task fusion mechanism, and normalization solves the problem of scale imbalance of control instructions, ensuring the overall stability of the multi-unmanned aerial vehicle system. Traditional multi-unmanned aerial vehicle communication schemes are usually based on static network topology, which is difficult to meet the communication reconstruction needs under dynamic tasks.

[0032] Further, the control instruction vector is converted into control instructions and executed, comprising: The normalized control instruction vector is converted into control instructions using vector set encapsulation; The adjacency matrix of the dynamic category graph is extracted, the communication channel of the unmanned aerial vehicle is allocated using the wireless communication protocol, the control instruction is sent to the corresponding unmanned aerial vehicle, and the unmanned aerial vehicle receives the control instruction and performs execution.

[0033] The scheme solves the problems of conflict, delay, packet loss, etc. in group communication through dynamic adjustment of the adjacency matrix and protocol allocation, significantly improving the communication efficiency and robustness.

[0034] S4, constructing a visual interface display control instruction; Specifically, the constructing a visual interface display control instruction comprises: The visual interface is constructed using a front-end framework React.js, and the control instruction obtained through analysis and the collected multi-source data are visualized and displayed. The user after real-name verification is allowed to consult.

[0035] Data visualization often uses static pages or re-rendering, which is slow in response and poor in interactivity. The present application uses React.js to enable real-time and low-latency rendering of changes in the instruction and multi-source data on the front end, greatly improving operation efficiency and user experience. In existing unmanned aerial vehicle group systems, most of them use open interfaces or access control based on simple passwords, which has a large security risk. Real-name verification not only improves the security level of the system, but also provides mechanism support for task tracing and responsibility division.

[0036] The embodiment also provides an intelligent collaborative control system based on an unmanned aerial vehicle group, comprising: A collection composition module is configured to collect multi-source data and perform preprocessing, construct a sparse tensor, calculate the partial derivative of each dimension in the sparse tensor using a finite difference method, combine the partial derivatives using a vector synthesis method, use the combined partial derivatives as a gradient vector field of information entropy, define an unmanned aerial vehicle as a node, and construct a dynamic category graph. A target updating module is configured to map the nodes of the dynamic category graph to popular points of a Riemann manifold, convert the coordinates of the unmanned aerial vehicles into grid indexes, calculate the task intensity influence of the unmanned aerial vehicles, calculate a metric tensor, add a regularization term to the metric tensor through positive definite regularization, define an objective function, calculate the Riemann gradient of the objective function, update the manifold points using exponential mapping, and obtain updated spatial coordinates and task vectors. An instruction execution module is configured to generate a control instruction vector using weighted linear combination, convert the control instruction vector into a control instruction, and execute the control instruction. A visualization module is configured to construct a visual interface to display the control instruction.

[0037] The embodiment also provides a computer device suitable for the intelligent collaborative control method based on an unmanned aerial vehicle group, comprising a memory and a processor. The memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions to implement the intelligent collaborative control method based on an unmanned aerial vehicle group proposed in the above embodiment.

[0038] The computer device can be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected by a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved by WIFI, a carrier network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, a trackball or a touchpad arranged on the shell of the computer device, or an external keyboard, a touchpad or a mouse, etc.

[0039] The embodiment also provides a storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the method for implementing intelligent cooperative control based on a UAV group. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0040] To sum up, the present application collects multi-source data and pre-processes, constructs a sparse tensor, uses finite difference method to calculate the partial derivative of each dimension in the sparse tensor, uses vector synthesis method to combine the partial derivative as the gradient vector field of information entropy, defines the unmanned aerial vehicle as a node, and constructs a dynamic category graph; maps the nodes of the dynamic category graph to the popular points of the Riemann manifold, converts the unmanned aerial vehicle coordinates into grid indexes, calculates the task intensity influence of the unmanned aerial vehicle, calculates the metric tensor, adds a regularization term to the metric tensor by positive definite regularization, defines the objective function, calculates the Riemann gradient of the objective function, updates the manifold points using exponential mapping, and obtains the updated spatial coordinates and task vector; uses weighted linear combination to generate a control instruction vector, converts the control instruction vector into control instructions and executes, solves the shortcomings of the traditional method in the complex dynamic environment, significantly improves the task coverage and collaborative robustness, and improves the autonomous collaboration ability.

[0041] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. An intelligent collaborative control method based on drone swarms, characterized by: include, Collect and preprocess multi-source data to construct a sparse tensor. Use the finite difference method to calculate the partial derivatives of each dimension in the sparse tensor. Use the vector synthesis method to combine the partial derivatives as the gradient vector field of information entropy. Define drones as nodes and construct a dynamic category graph. Map the nodes of the dynamic category graph to the popular points of the Riemann manifold, convert the drone coordinates to grid indices, calculate the mission intensity impact of the drone, calculate the metric tensor, add a regularization term to the metric tensor using positive definite regularization, define the objective function, calculate the Riemannian gradient of the objective function, and use the exponential mapping to update the manifold points to obtain the updated spatial coordinates and mission vector. generating a control instruction vector using a weighted linear combination, converting the control instruction vector into a control instruction and executing the control instruction; Build a visual interface to display control instructions.

2. The intelligent collaborative control method based on a drone swarm according to claim 1, characterized in that: The process of collecting multi-source data, constructing a sparse tensor, defining drones as nodes, and building a dynamic category graph includes: Use smart sensors to collect multi-source data from drone swarms and perform denoising and normalization processing; The multi-source data includes images, three-dimensional point clouds, spatial coordinates, velocity and acceleration data; Use the extended Kalman filter to fuse the spatial coordinates, velocity, and acceleration data, cluster the fused data using the K-means clustering algorithm, recognize the image using the pre-trained Swin-Transformer model, output the corresponding state variables, use tensor stacking to stack the spatial coordinates, three-dimensional point cloud, velocity, acceleration, clustering results, and state labels, construct a sparse tensor, traverse the non-zero elements of the sparse tensor and sum them, calculate the task intensity probability, calculate the information entropy based on the task intensity probability, use the finite difference method to calculate the partial derivatives of each dimension in the sparse tensor based on the information entropy, and use the vector synthesis method to combine the partial derivatives as the gradient vector field of the information entropy; Extract the target grid coordinates from the sparse tensor, convert the grid coordinates into actual geographic coordinates, calculate the direction vector from the drone to the target position, normalize the direction vector, and obtain the mission vector; Calculate the control mapping function between UAVs, use the fixed threshold method to set the screening threshold, and define the filtered control mapping function as the control morphism; Drones are defined as nodes, and control morphisms greater than 0 are defined as edges. An initial category graph is constructed. A statistical analysis method is used to set a change threshold. The Euclidean distance between the spatial position of each drone and the previous time is calculated. Control morphisms with Euclidean distances greater than the change threshold are screened and marked as affected. The control morphisms are then recalculated. Otherwise, the control morphisms at the previous time are retained to generate a dynamic category graph.

3. The intelligent collaborative control method based on a swarm of drones according to claim 2, characterized in that: The nodes of the dynamic category graph are mapped to the popular points of the Riemann manifold, the drone coordinates are converted into grid indices, and the objective function is defined, including: Based on drones, a Riemannian manifold is constructed. The Euclidean distance formula is used to calculate the Euclidean distance between manifold points, which is defined as the Euclidean metric. The Euclidean metric is used to set the initial metric tensor and map the nodes of the dynamic category graph to the popular points of the Riemannian manifold. The edges of the dynamic category graph are used as adjacency constraints on the manifold to construct the adjacency matrix on the manifold; Use the grid mapping table to convert the drone coordinates into grid indices, calculate the drone's mission intensity impact, calculate the metric tensor, and add a regularization term to the metric tensor using positive definite regularization; Using the weighted objective construction, define the objective function.

4. The intelligent collaborative control method based on a swarm of drones according to claim 3, characterized in that: The Riemann gradient of the objective function is calculated, and the manifold points are updated using the exponential mapping to obtain the updated spatial coordinates and task vector, including: Combine the gradient vector and the metric tensor to calculate the Riemann gradient of the objective function, and update the manifold points using the exponential mapping based on the Riemann gradient; Decompose the updated spatial coordinates and task vector from the update manifold points using vector decomposition.

5. The intelligent collaborative control method based on drone swarms according to claim 4, characterized in that: The method of generating a control instruction vector by using a weighted linear combination includes: The position increment is calculated using vector subtraction, and the control command vector is generated using weighted linear combination and normalized.

6. The intelligent collaborative control method based on a swarm of drones according to claim 5, characterized in that: The converting the control instruction vector into a control instruction and executing the control instruction includes: Use vector set encapsulation to convert the normalized control instruction vector into a control instruction; The adjacency matrix of the dynamic category graph is extracted, and the communication channels of the drones are allocated using the wireless communication protocol. The control instructions are sent to the corresponding drones, and the drones receive and execute the control instructions.

7. The intelligent collaborative control method based on a drone swarm according to claim 6, characterized in that: The construction of a visual interface to display control instructions includes: Use the front-end framework React.js to build a visual interface to visualize the analyzed control instructions and collected multi-source data; Users who have passed real-name verification are allowed to access the information.

8. An intelligent collaborative control system based on a swarm of unmanned aerial vehicles, based on the intelligent collaborative control method based on a swarm of unmanned aerial vehicles according to any one of claims 1 to 7, characterized in that: include, The collection and composition module is used to collect and preprocess multi-source data, construct a sparse tensor, calculate the partial derivatives of each dimension in the sparse tensor using the finite difference method, combine the partial derivatives using the vector synthesis method, and use them as the gradient vector field of information entropy. The drones are defined as nodes to construct a dynamic category graph. The target update module is used to map the nodes of the dynamic category graph to the popular points of the Riemann manifold, convert the drone coordinates to grid indices, calculate the mission intensity impact of the drone, calculate the metric tensor, add a regularization term to the metric tensor using positive definite regularization, define the objective function, calculate the Riemann gradient of the objective function, and update the manifold points using the exponential map to obtain the updated spatial coordinates and mission vector; An instruction execution module, configured to generate a control instruction vector using a weighted linear combination, convert the control instruction vector into a control instruction, and execute the control instruction; Visualization module, used to build a visual interface to display control instructions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the intelligent collaborative control method based on a drone swarm are implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent collaborative control method based on a drone swarm are implemented as described in any one of claims 1 to 7.

Images