Unmanned aerial vehicle cluster time-varying formation tracking control method and device based on RBF neural network, and storage medium

Abstract

The invention provides an unmanned aerial vehicle cluster time-varying formation tracking control method and device based on an RBF neural network, and a storage medium, so as to effectively fit the external input of a leader unmanned aerial vehicle, thereby achieving better time-varying formation tracking control precision. The method comprises the following steps: constructing an expected time-varying formation configuration instruction of a follower unmanned aerial vehicle in an unmanned aerial vehicle cluster; dynamic models of the leader unmanned aerial vehicle and the follower unmanned aerial vehicle are established; constructing a communication topological relation between the leader unmanned aerial vehicle and the follower unmanned aerial vehicle, and calculating a characteristic value of a communication topological matrix of the follower unmanned aerial vehicle; controller parameters are obtained according to the characteristic values of the communication topology matrix and the dynamical model; periodically iteratively calculating a control instruction of the follower unmanned aerial vehicle based on an RBF neural network, the follower unmanned aerial vehicle expected time-varying formation configuration instruction and controller parameters; and performing formation tracking control based on the control instruction.

Description

Technical Field

[0001] This application relates to the field of multi-UAV formation flight control, specifically to a UAV swarm time-varying formation tracking control method, device, and storage medium based on RBF neural network. Background Technology

[0002] In recent years, multi-aircraft formation control has become an important direction in the field of aircraft control. Compared with single-aircraft missions, multi-aircraft formation missions have significant advantages in terms of time consumption, scheme selection, and mission completion rate. In practical applications, it is often necessary to form a formation while tracking the leader drone. However, traditional control methods are limited to situations where the leader drone has no external input, which is still somewhat lacking in practical applications. In reality, whether it is tracking, encircling, and striking enemy targets or following friendly formations, the leader drone usually has some external input.

[0003] Therefore, in order to address the above issues, how to effectively fit the external inputs of the leader drone to achieve better time-varying formation tracking control accuracy is a problem that needs to be solved. Summary of the Invention

[0004] This application provides a device and storage medium for a time-varying formation tracking and control method for UAV swarms based on RBF neural networks, which can effectively fit the external input of the leader UAV, thereby achieving better time-varying formation tracking and control accuracy.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] In a first aspect, this application provides a time-varying formation tracking and control method for UAV swarms based on RBF neural networks, the method comprising:

[0007] Construct time-varying formation configuration commands for follower drones in a drone swarm; and develop dynamic models for the leader and follower drones;

[0008] Construct the communication topology between the leader drone and the follower drone, and calculate the eigenvalues ​​of the follower drone communication topology matrix;

[0009] The controller parameters are obtained based on the eigenvalues ​​of the communication topology matrix and the dynamic model;

[0010] Based on the RBF neural network, the expected time-varying formation configuration command of the follower UAV and the controller parameters, the control command of the follower UAV is calculated periodically and iteratively.

[0011] Formation tracking control is performed based on control commands.

[0012] One possible design scheme, the first aspect of the method, also includes a drone swarm comprising N drones, of which at least one leader drone and at least one follower drone, where N is a positive integer greater than or equal to two.

[0013] One possible design approach, the first aspect of which also includes the following dynamic models for the leader drone and follower drone:

[0014] Where i is the number of the N drones, t represents time, and A and B are constant matrices. U represents the state of the i-th drone at time t. i (t) represents the control input of the i-th UAV at time t, where i is a positive integer greater than zero and less than or equal to N.

[0015] One possible design approach, the first aspect of which further includes constructing the communication topology between the leader drone and the follower drones, and calculating the eigenvalues ​​of the follower drone communication topology matrix, including:

[0016] The communication topology between the leader drone and the follower drones can be represented by a Laplace matrix as follows:

[0017] Where L1 is the communication topology matrix between follower drones in the drone swarm, and L2 is the communication topology matrix between follower drones and leader drones in the drone swarm.

[0018] By solving the characteristic equation det(L1-λI)=0, the eigenvalues ​​λ of the communication topology matrix between follower drones in the drone swarm are obtained, where I is the identity matrix.

[0019] One possible design approach, the first aspect of which further includes obtaining controller parameters based on the eigenvalues ​​of the communication topology matrix and the dynamic model, including:

[0020] Based on the eigenvalue matrix λ of the communication topology matrix between follower drones in a drone swarm, the maximum value term λ of the eigenvalue matrix λ is obtained. max and the minimum term λ min ;

[0021] Based on the dynamic models of the leader drone and the follower drone, constant matrices A and B are obtained;

[0022] The controller parameters must meet the following conditions: Matrix M is The solution, where,

[0023] One possible design approach, the first aspect of the method, further includes periodically iteratively calculating the control commands of the follower drone based on an RBF neural network, including: obtaining the control commands of the follower drone at any given time based on the RBF neural network.

[0024]

[0025] Where K is the controller parameter, n is the ID of the leader drone in the drone swarm, i is a positive integer greater than zero and less than or equal to N, and i is not equal to n. i Let j represent the set of drone numbers among N drones that have communication costs with the i-th drone, where j belongs to N. i And j is not equal to n; w ij (t) represents the communication cost between the i-th drone and the j-th drone; h represents the state of the i-th UAV at time t; i (t) represents the expected time-varying formation configuration vector of the i-th UAV at time t; h represents the state of the j-th UAV at time t; j (t) represents the expected time-varying formation configuration vector of the j-th UAV at time t; w in (t) represents the communication cost between the i-th drone and the n-th drone; This represents the state of the nth drone at time t; Let represent the first derivative of the expected time-varying formation velocity vector of the i-th UAV at time t; and s i For RBF neural network elements.

[0026] One possible design approach, the first aspect of the method, also includes, The following conditions must be met:

[0027] Where, γ i σ represents the learning rate of the neural network. i These are the parameters of the neural network.

[0028] In a second aspect, a time-varying formation tracking and control device for UAV swarms based on an RBF neural network is provided, the device including a module for performing the method described in the first aspect.

[0029] In one possible design, the RBF neural network-based UAV swarm time-varying formation tracking and control device of the second aspect may further include a transceiver. This transceiver can be a transceiver circuit or an interface circuit. The transceiver can be used for communication between the RBF neural network-based UAV swarm time-varying formation tracking and control device of the second aspect and other devices.

[0030] In one possible design, the second aspect of the RBF neural network-based UAV swarm time-varying formation tracking and control device may further include a memory. This memory may be integrated with the processor or configured separately. The memory may be used to store instructions related to the method of the first aspect.

[0031] Thirdly, a time-varying formation tracking and control device for UAV swarms based on an RBF neural network is provided. This RBF neural network-based time-varying formation tracking and control device includes: a processor coupled to a memory, the processor executing instructions stored in the memory to cause the RBF neural network-based time-varying formation tracking and control device to perform the method described in the first aspect.

[0032] In one possible design, the RBF neural network-based UAV swarm time-varying formation tracking and control device described in the third aspect may further include a transceiver. This transceiver can be a transceiver circuit or an interface circuit. The transceiver can be used for communication between the RBF neural network-based UAV swarm time-varying formation tracking and control device and other devices.

[0033] Fourthly, a time-varying formation tracking and control device for UAV swarms based on an RBF neural network is provided, comprising: a processor and a memory; the memory is used to store instructions, and when the processor executes the instructions, the time-varying formation tracking and control device for UAV swarms based on an RBF neural network performs the method described in the first aspect.

[0034] In one possible design, the RBF neural network-based UAV swarm time-varying formation tracking and control device described in the fourth aspect may further include a transceiver. This transceiver can be a transceiver circuit or an interface circuit. The transceiver can be used for communication between the RBF neural network-based UAV swarm time-varying formation tracking and control device and other devices.

[0035] Fifthly, a computer-readable storage medium is provided, the computer-readable storage medium including storage of a computer program or instructions, which, when executed, cause the time-varying formation tracking and control method for UAV swarms based on the RBF neural network described in the first aspect to be executed.

[0036] In this embodiment, when the leader drone has external input, the external input of the leader drone is estimated by means of neural networks, state information of follower drones, etc., so as to compensate for it, reduce the interference of the external input of the leader drone on the follower drone, and realize the tracking of the leader drone by the follower drone.

[0037] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 A flowchart illustrating the time-varying formation tracking and control method for UAV swarms based on RBF neural networks provided in this application embodiment;

[0040] Figure 2 A directed interaction topology diagram of each drone in a drone swarm provided in an embodiment of this application;

[0041] Figure 3 The drone swarm movement trajectory provided in the embodiments of this application;

[0042] Figure 4 Error curves of the tracker drone provided in the embodiments of this application;

[0043] Figure 5 Schematic diagram of the structure of the UAV swarm time-varying formation tracking and control device based on RBF neural network provided in the embodiments of this application. Figure 1 ;

[0044] Figure 6 Schematic diagram of the structure of the UAV swarm time-varying formation tracking and control device based on RBF neural network provided in the embodiments of this application. Figure 2 . Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. At the same time, in the description of the embodiments of this application, the terms "first," "second," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0046] For ease of understanding, the technical terminology used in this application will be introduced below: RBF neural network theory.

[0047] RBF (Radial Basis Function) neural networks consist of an input layer, hidden layers, and an output layer. The relationship between the input and hidden layers is a nonlinear mapping formed by radial basis functions, exhibiting strong nonlinear fitting capabilities and capable of mapping arbitrarily complex nonlinear relationships. The most commonly used radial basis function is the Gaussian kernel function.

[0048]

[0049] m is the number of hidden layer nodes in the neural network, b i c is the width of the basis functions. i =[c i1 ,c i2 ,…,c im ] T Let z be the center vector, and z be the input to the neural network.

[0050] set up If f(z) is a continuous function, then f(z) can be fitted using an RBF neural network:

[0051] f(z) = W *T s(z)+η(z)

[0052] in Let s(z) represent the optimal weights of the neural network, where s(z) = [s1(z), s2(z), ..., s...]. m (z)] T η(z) is the estimation error. The ideal optimal weight W * The definition is as follows:

[0053]

[0054] In this embodiment of the application, the ability of RBF neural networks to approximate arbitrary nonlinear functions is utilized to design a time-varying formation tracking control method for UAV swarms based on RBF neural networks.

[0055] Figure 1 A flowchart illustrating the time-varying formation tracking and control method for UAV swarms based on RBF neural networks provided in this application embodiment. Figure 1 This time-varying formation tracking control method for UAV swarms based on RBF neural networks is applicable when the leader UAV in a UAV swarm has external inputs. It fits the external inputs to enable follower UAVs to track the time-varying formation.

[0056] The process of the UAV swarm time-varying formation tracking control method based on RBF neural network is as follows:

[0057] Step S101: Construct the desired time-varying formation configuration instructions for follower drones in the drone swarm; and construct the dynamic models of the leader drone and follower drones.

[0058] It should be noted that the drone swarm consists of N drones, and the N drones include at least one leader drone and at least one follower drone, where N is a positive integer greater than or equal to two.

[0059] To facilitate understanding of this application, the following description will use the drone numbered N in the drone swarm as the leader drone and the drones in the drone swarm with the number set F = {1,2,…,N-1} as the follower drones.

[0060] The instructions for constructing the desired time-varying formation configuration of follower drones in a drone swarm are as follows: Where h i (t)=[h ix (t),h iv (t)] T (i∈F), h ix (t) is the desired time-varying formation position vector, h iv (t) is the velocity vector of the desired time-varying formation configuration, and the desired time-varying formation configuration satisfies It is h ix The first derivative of (t).

[0061] The dynamic model of the leader drone can be represented as: in, and These are the position and velocity vectors of the leader drone. This represents the leader's own control vector, which is the content that the leader drone can input externally.

[0062] The follower dynamics model can be expressed as: in, and The position and velocity vectors of the follower drone are F = {1, 2, ..., N-1}, u i (t) represents its corresponding control input.

[0063] The status of each drone in the drone swarm is as follows Subsequently, the dynamics models of the leader drone and the follower drone can be unified and converted to... Where t represents time, and A and B are constant matrices, with B being [0,1]. T B1 can be [1,0] T A is B1B T .

[0064] Step S102: Construct the communication topology relationship between the leader drone and the follower drone, and calculate the eigenvalues ​​of the follower drone communication topology matrix.

[0065] It should be noted that the interactions between each drone in a drone swarm can be described using a graph structure, G = {V, S, W}. Each drone in the swarm can be described as a vertex set V = {v1, v2, ..., v...}. N}, where the i-th drone corresponds to node v in the graph. i The influence of the i-th drone on the j-th drone corresponds to an edge e in the graph. ij =(v i ,v j ), where node v i It can be called the parent node, node v j These can be called child nodes. The interactions among all drones in a drone swarm can be represented by the edge set of the graph. This indicates that if the i-th drone and the j-th drone interact, i.e., if e ij For all ∈S, e ji If any two edges are connected to the same node (S), then graph G is called an undirected graph; otherwise, it is called a directed graph. The weight set W assigns a weight to each edge, representing the strength of the influence between drones, communication latency, data transmission rate, etc., and can also be called the communication cost between the two nodes corresponding to each edge.

[0066] For example, suppose there are four drones U1, U2, U3, and U4. Their interactions can be described by the following topological graph: Vertex set V = {U1, U2, U3, U4}; Edge set S = {(U1, U2), (U2, U3), (U1, U3), (U3, U4)}, representing direct interactions between U1 and U2, U2 and U3, U1 and U3, and U3 and U4. The weight of edge (U1, U2) is w. 12 The weight of edge (U2, U3) is w. 23 The weight of edge (U1, U3) is w. 13 The weight of edge (U3, U4) is w. 34 .

[0067] The communication topology between the leader drone and the follower drones can be represented by a Laplace matrix as follows: Wherein, L1 is the communication topology matrix between follower drones in the drone swarm, and L2 is the communication topology matrix between follower drones and leader drones in the drone swarm.

[0068] In other words, L1 represents the communication topology between follower drones in a drone swarm, and L2 represents the communication topology between follower drones and the leader drone in the same swarm. Specifically, it can be represented as: (i,j)∈F, i∈F.

[0069] By solving the characteristic equation det(L1-λI)=0, the eigenvalues ​​λ of the communication topology matrix between follower drones in the drone swarm are obtained, where I is the identity matrix.

[0070] Step S103: Obtain the controller parameters based on the eigenvalues ​​of the communication topology matrix and the dynamic model.

[0071] In step S102, the obtained eigenvalue λ matrix has N-1 elements, namely λ1, λ2, ..., λ N-1 And each element satisfies the following relationship: λ1<λ2<…<λ N-1 Furthermore, based on the eigenvalue matrix λ of the communication topology matrix between follower drones in the drone swarm, the maximum value term λ of the eigenvalue matrix λ can be obtained. max (i.e., λ) N-1 ) and minimum term λ min (i.e., λ1);

[0072] Furthermore, in step S101, the dynamic models of the leader drone and the follower drone are as follows: We can obtain constant matrices A and B. In this application, B is defined as [0,1]. T B1 is [1,0] T A is B1B T For example, the value can also be selected according to the specific situation in the drone swarm, which will not be elaborated here.

[0073] The controller parameters can be determined based on the maximum value of the eigenvalue matrix λ. max (i.e., λ) N-1 ) and minimum term λ min (i.e., λ1) and constant matrices A and B are determined. Where, matrix M is The solution here

[0074] Step S104: Based on the RBF neural network, the expected time-varying formation configuration command of the follower UAV, and the controller parameters, the control command of the follower UAV is calculated periodically and iteratively.

[0075] In this application, since the RBF neural network has the ability to approximate any nonlinear function, adaptive time-varying formation tracking control is performed through the RBF neural network.

[0076] Furthermore, the specific details of the RBF neural network can be found in the preceding description and will not be repeated here.

[0077] Based on the RBF neural network, the control command for the follower drone at any given moment is obtained using the following formula:

[0078]

[0079] Where K is the controller parameter, n is the ID of the leader drone in the drone swarm, i is a positive integer greater than zero and less than or equal to N, and i is not equal to n. i Let j represent the set of drone IDs that have communication costs with the i-th drone out of N drones, where j belongs to N. i And j is not equal to n; w ij (t) represents the communication cost between the i-th drone and the j-th drone; h represents the state of the i-th UAV at time t; i (t) represents the expected time-varying formation configuration vector of the i-th UAV at time t; h represents the state of the j-th UAV at time t; j (t) represents the expected time-varying formation configuration vector of the j-th UAV at time t; w in(t) represents the communication cost between the i-th drone and the n-th drone (the leader drone); This represents the state of the nth drone at time t; Let represent the first derivative of the expected time-varying formation velocity vector of the i-th UAV at time t; and s i For RBF neural network elements.

[0080] The following conditions must be met:

[0081] Where, γ i σ represents the learning rate of the neural network. i These are the parameters of the neural network.

[0082] Step S105: Perform formation tracking control based on control commands.

[0083] The control commands u corresponding to each follower drone obtained in step S104 are used to... i (t) Input the controller of each follower drone to control the position and speed of each follower drone.

[0084] In summary, this application embodiment estimates the external input of the leader drone by using neural networks and the state information of the follower drone when the leader drone has external input, thereby compensating for it, reducing the interference of the leader drone's external input on the follower drone, and enabling the follower drone to track the leader drone.

[0085] The above combination Figure 1 This application provides a detailed description of the UAV swarm time-varying formation tracking and control method based on RBF neural networks, as illustrated below. Figures 2-4 This paper introduces a time-varying formation tracking control method for UAV swarms based on RBF neural networks and its specific application scenarios in UAV swarms.

[0086] Taking a drone swarm of four drones as an example: one is the leader drone (numbered 4), and the other three are follower drones (numbered 1, 2, and 3). The three follower drones are required to perform time-varying formation tracking of the leader drone in a two-dimensional plane, i.e., the XY plane. The directed action topology of the drones in the swarm is as follows: Figure 2 As shown, and it is assumed that the weights in the graph are all between 0 and 1.

[0087] The time-varying grouping instruction for followers can be represented as follows:

[0088]

[0089] Where r = 3m, ω = 1rad / s, the follower drones will maintain a time-varying circular formation around the leader drone.

[0090] The initial states of the leader drone and follower drone are [0, 0, 0, 3], [1.45, 1.41, 0.05, 1.2], [-1, 2.3, -1, 0.6], [5.91, 3.6, 0, 1.6].

[0091] Assume the external inputs to the leader drone are:

[0092] Where r N =10m,ω N =0.3rad / s, meaning the leader also moves in a time-varying circular trajectory.

[0093] Based on the above conditions and initial state, the neural network learning rate γ is selected. i =10, neural network parameter σ i =0.005. And according to steps S102 and S103, the controller parameter K = [-6.9655, -13.9309] is obtained.

[0094] Based on the calculation method in step S104, a simulation can be performed to obtain the following results: Figure 3 The drone swarm trajectory shown is as follows: the blue trajectory represents the leader drone (UAV4), and the red, green, and black trajectories correspond to the three follower drones (UAV1, UAV2, and UAV3).

[0095] As can be seen, the initial states of the four drones are represented by small circles, and their final states are represented by squares, plus signs, stars, and triangles, respectively. The diagram shows that the three follower drones formed an equilateral triangle configuration and moved in circles, simultaneously tracking the leader drone's position, achieving the desired time-varying formation configuration and tracking.

[0096] In addition, according to Figure 4 The error curves shown demonstrate that, according to the method described in this application, the three follower UAVs (UAV1, UAV2, and UAV3) achieved time-varying formation tracking with minimal error.

[0097] The above combination Figures 1-4 This application provides a detailed description of the UAV swarm time-varying formation tracking and control method based on RBF neural networks, as illustrated in the embodiments below. Figures 5-6 This document describes in detail the time-varying formation tracking and control device for UAV swarms based on RBF neural networks provided in the embodiments of this application.

[0098] Figure 5This is a schematic diagram of the structure of the UAV swarm time-varying formation tracking and control device based on RBF neural network provided in the embodiments of this application. Figure 1 For example, such as Figure 5 As shown, the UAV swarm time-varying formation tracking and control device 500 based on RBF neural network includes: a transceiver module 501 and a processing module 502. For ease of explanation, Figure 5 Only the main components of this time-varying formation tracking and control device for UAV swarms based on the RBF neural network are shown.

[0099] The transceiver module 501 is used to perform the transceiver function of the above-mentioned UAV swarm time-varying formation tracking control method based on RBF neural network, and the processing module 502 is used to perform other functions of the above-mentioned UAV swarm time-varying formation tracking control method based on RBF neural network besides the transceiver function.

[0100] Optionally, the transceiver module 501 may include a transmitting module ( Figure 5 (not shown in the image) and receiving module ( Figure 5 (Not shown in the image). The transmitting module is used to implement the transmitting function of the RBF neural network-based UAV swarm time-varying formation tracking and control device 500, and the receiving module is used to implement the receiving function of the RBF neural network-based UAV swarm time-varying formation tracking and control device 500.

[0101] Optionally, the RBF neural network-based UAV swarm time-varying formation tracking and control device 500 may also include a storage module ( Figure 5 (Not shown in the image), the storage module stores programs or instructions. When the processing module 502 executes the program or instructions, the RBF neural network-based UAV swarm time-varying formation tracking and control device 500 can execute the RBF neural network-based UAV swarm time-varying formation tracking and control method in the embodiments of this application.

[0102] The following is combined with Figure 6 A detailed description of each component of the 600 UAV swarm time-varying formation tracking and control device based on RBF neural network is provided below:

[0103] The processor 601 is the control center of the UAV swarm time-varying formation tracking and control device 600 based on the RBF neural network. It can be a single processor or a collective term for multiple processing elements. For example, the processor 601 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement the embodiments of this application, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0104] Optionally, the processor 601 can execute various functions of the RBF neural network-based UAV swarm time-varying formation tracking control device 600 by running or executing software programs stored in the memory 602 and calling data stored in the memory 602, such as executing the RBF neural network-based UAV swarm time-varying formation tracking control method in the embodiments of this application.

[0105] In a specific implementation, as one example, processor 601 may include one or more CPUs, for example... Figure 6 CPU0 and CPU1 are shown in the diagram.

[0106] In a specific implementation, as one example, the UAV swarm time-varying formation tracking and control device 600 based on the RBF neural network may also include multiple processors, for example... Figure 6 The processors 601 and 604 are shown in the diagram. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, "processor" can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions). The memory 602 is used to store the software program executing the scheme of this application, and its execution is controlled by the processor 601. Specific implementation methods can be found in the above method embodiments, and will not be repeated here.

[0107] Optionally, the memory 602 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 602 may be integrated with the processor 601 or exist independently, and may be connected via the interface circuit of the RBF neural network-based UAV swarm time-varying formation tracking control device 600. Figure 6 (Not shown in the image) is coupled to the processor 601, but this embodiment does not specifically limit this.

[0108] Transceiver 603 is used for communication with other communication devices. For example, in the case of a UAV swarm time-varying formation tracking and control device 600 based on an RBF neural network, which is the first device, transceiver 603 can be used to communicate with a second device or a third device.

[0109] Alternatively, transceiver 603 may include a receiver and a transmitter. Figure 6 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.

[0110] Optionally, the transceiver 603 can be integrated with the processor 601 or exist independently, and can be connected to the interface circuit of the RBF neural network-based UAV swarm time-varying formation tracking and control device 600. Figure 6 (Not shown in the image) is coupled to the processor 601, but this embodiment does not specifically limit this.

[0111] Understandable, Figure 6 The structure of the RBF neural network-based UAV swarm time-varying formation tracking control device 600 shown does not constitute a limitation on the RBF neural network-based UAV swarm time-varying formation tracking control device. The actual RBF neural network-based UAV swarm time-varying formation tracking control device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0112] Furthermore, the technical effects of the UAV swarm time-varying formation tracking and control device 600 based on RBF neural network can be referred to the technical effects of the method described in the above method embodiments, and will not be repeated here.

[0113] It should be understood that the processor in the embodiments of this application can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0114] It should also be understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0115] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

Claims

1. A time-varying formation tracking control method for UAV swarms based on RBF neural networks, characterized in that, The method includes: Construct the desired time-varying formation configuration instructions for follower drones in a drone swarm; and the dynamic model of the leader drone and the follower drones; Construct the communication topology between the leader drone and the follower drone, and calculate the eigenvalues ​​of the communication topology matrix of the follower drone; The controller parameters are obtained based on the eigenvalues ​​of the communication topology matrix and the dynamic model. Based on the RBF neural network, the desired time-varying formation configuration command of the follower UAV, and the controller parameters, the control command of the follower UAV is calculated periodically and iteratively. Formation tracking control is performed based on the control commands.

2. The UAV swarm time-varying formation tracking control method based on RBF neural network according to claim 1, characterized in that, The drone swarm comprises N drones, including at least one leader drone and at least one follower drone, where N is a positive integer greater than or equal to two.

3. The UAV swarm time-varying formation tracking and control method based on RBF neural network according to claim 2, characterized in that, The dynamic model of the leader drone and the follower drone is as follows: Where i is the number of the N drones, t represents time, and A and B are constant matrices. U represents the state of the i-th drone at time t. i (t) represents the control input of the i-th UAV at time t, where i is a positive integer greater than zero and less than or equal to N.

4. The UAV swarm time-varying formation tracking control method based on RBF neural network according to claim 3, characterized in that, The process of constructing the communication topology between the leader drone and the follower drones, and calculating the eigenvalues ​​of the follower drone communication topology matrix, includes: The communication topology between the leader drone and the follower drone is represented by a Laplace matrix as follows: Wherein, L1 is the communication topology matrix between follower drones in the drone cluster, and L2 is the communication topology matrix between follower drones and leader drones in the drone cluster; By solving the characteristic equation det(L1-λI)=0, the eigenvalue λ of the communication topology matrix between the follower drones in the drone cluster is obtained, where I is the identity matrix.

5. The UAV swarm time-varying formation tracking control method based on RBF neural network according to claim 4, characterized in that, The process of obtaining controller parameters based on the eigenvalues ​​of the communication topology matrix and the dynamic model includes: Based on the eigenvalue matrix λ of the communication topology matrix between follower drones in the drone swarm, the maximum value term λ of the eigenvalue matrix λ is obtained. max and the minimum term λ min ; Based on the dynamic models of the leader drone and the follower drone, the constant matrix A and the constant matrix B are obtained; The controller parameters satisfy the following conditions: Matrix M is The solution, where, 6. The UAV swarm time-varying formation tracking control method based on RBF neural network according to claim 5, characterized in that, The control commands for the follower UAV, based on the RBF neural network, the desired time-varying formation configuration command of the follower UAV, and the controller parameters, are periodically iteratively calculated, including: Based on the RBF neural network, the control commands for the follower drone at any given time are obtained. Wherein, K is the controller parameter, n is the ID of the leader drone in the drone cluster, i is a positive integer greater than zero and less than or equal to N, and i is not equal to n, and N... i This represents the set of drone IDs among the N drones that have communication costs with the i-th drone, where j belongs to N. i And j is not equal to n; the w ij (t) represents the communication cost between the i-th drone and the j-th drone; the The h represents the state of the i-th UAV at time t; i (t) represents the expected time-varying formation configuration vector of the i-th UAV at time t; the The h represents the state of the j-th UAV at time t; j (t) represents the expected time-varying formation configuration vector of the j-th UAV at time t; the w in (t) represents the communication cost between the i-th drone and the n-th drone; the... This represents the state of the nth UAV at time t; the... The first derivative of the expected time-varying formation velocity vector of the i-th UAV at time t; and s i For RBF neural network elements.

7. The UAV swarm time-varying formation tracking control method based on RBF neural network according to claim 6, characterized in that, The The following conditions must be met: Where, γ i σ represents the learning rate of the neural network. i These are the parameters of the neural network.

8. A time-varying formation tracking and control device for UAV swarms based on RBF neural networks, characterized in that, The apparatus includes a module for performing the method as described in any one of claims 1-7.

9. A time-varying formation tracking and control device for UAV swarms based on RBF neural networks, characterized in that, The RBF neural network-based UAV swarm time-varying formation tracking and control device includes a processor and a memory; the memory is used to store computer instructions, and when the processor executes the instructions, it causes the RBF neural network-based UAV swarm time-varying formation tracking and control device to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a computer program or instructions that, when executed, cause the method as described in any one of claims 1-7 to be performed.

Images