Aircraft formation cooperative control method based on self-learning

Abstract

The invention provides an aircraft formation cooperative control method based on self-learning, and the method comprises the steps: carrying out the prediction of the flight data of each aircraft at the current moment based on an observer, and obtaining the leader flight state and system matrix of a leader aircraft at the next moment; and outputting a cooperative control strategy based on the controller application leader flight state and the system matrix. According to the method provided by the invention, the estimation of the leader flight state of the leader and the system matrix of the aircraft formation is realized through the observer constructed based on the neural network, the generalization ability of unknown data is realized through the internal law and characteristics between the sample flight data obtained through deep learning, the limitation of traditional physical model construction is avoided, and the method has the advantages of high reliability and high reliability. And the accuracy and generalization of the estimated leader flight state and system matrix are greatly improved. In addition, a controller constructed based on a neural network is adopted, an optimal cooperative control strategy is output, and control strategy optimization of the whole aircraft formation is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of aircraft control, and particularly to a self-learning based cooperative control method for aircraft formation. Background Technique

[0002] The cooperative control of aircraft formation is an important topic in the aviation field, which is crucial for the formation maintenance, mission execution of aircraft, and avoiding collisions with each other. With the continuous progress of aviation technology, the flight missions of aircraft are becoming more and more complex, and the requirements for the accuracy and real-time performance of formation cooperative control are also getting higher and higher. Traditional aircraft formation cooperative control methods mainly rely on physical models, such as numerical simulation methods based on dynamic equations.

[0003] However, during the actual flight of an aircraft, it will be affected by various uncertain factors, such as changes in wind speed and direction, and the uncertainty of the aircraft's own state, etc. These all make it very difficult to establish an accurate physical model. Therefore, the accuracy and generalization ability of aircraft formation cooperative control relying on physical models need to be improved, resulting in some limitations in practical applications. Summary of the Invention

[0004] The present invention provides a self-learning based cooperative control method for aircraft formation to solve the deficiencies in the accuracy and generalization ability of aircraft formation cooperative control relying on physical models in the prior art.

[0005] The present invention provides a self-learning based cooperative control method for aircraft formation, including: Obtain the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; Based on an observer, use the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; Based on a controller, use the leading flight state and the system matrix to output a cooperative control strategy; The system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0006] According to a self-learning based cooperative control method for aircraft formation provided by the present invention, the step of using the leading flight state and the system matrix based on the controller to output a cooperative control strategy includes: Based on the evaluation network in the controller, evaluate the flight state data at the current moment to obtain the current policy reward; Based on the action network in the controller, apply the leading flight state, the system matrix, and the current policy reward to output the cooperative control policy.

[0007] According to a self-learning based cooperative control method for an aircraft formation provided by the present invention, the aircraft formation includes the leader aircraft and a plurality of follower aircraft; The observers are respectively arranged on the leader aircraft and each follower aircraft; The flight data of each aircraft at the current moment is shared among the leader aircraft and the plurality of follower aircraft.

[0008] According to a self-learning based cooperative control method for an aircraft formation provided by the present invention, the sample flight data includes sample flight state data and sample state influence data; The sample state influence data includes control instructions and external influence data; The external influence data includes at least one of aerodynamic parameters, aircraft structure parameters, and environmental parameters.

[0009] According to a self-learning based cooperative control method for an aircraft formation provided by the present invention, the determination step of the sample flight data includes: It is obtained by combining the sample flight state data, control instructions, and at least one randomly selected external influence data at any historical moment.

[0010] According to a self-learning based cooperative control method for an aircraft formation provided by the present invention, based on the observer, applying the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation, includes: Based on the observer, applying the flight data and input weights at the current moment for prediction, and outputting the leading flight state and the system matrix; The input weights are respectively used to control the importance degrees of the flight state data and the state influence data; The input weights include two symmetric matrices.

[0011] The present invention also provides a self-learning based cooperative control device for an aircraft formation, including: An acquisition unit, which acquires the flight data of each aircraft in the aircraft formation at the current moment, and the flight data includes flight state data and state influence data; An observation unit, based on the observer, applying the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; The control unit, based on the controller, applies the leadership flight state and the system matrix, and outputs a cooperative control strategy; The system model includes the observer and the controller. The system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0012] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the self-learning-based cooperative control method for aircraft formations as described in any one of the above.

[0013] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the self-learning-based cooperative control method for aircraft formations as described in any one of the above.

[0014] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the self-learning-based cooperative control method for aircraft formations as described in any one of the above.

[0015] A self-learning-based cooperative control method for aircraft formations provided by the present invention estimates the leadership flight state of the leader and the system matrix of the aircraft formation through an observer constructed based on a neural network. By learning the internal laws and characteristics between the sample flight data, it has the generalization ability for unknown data, avoids the limitations of traditional physical model construction, and greatly improves the accuracy and generalization of the estimated leadership flight state and system matrix. In addition, a controller constructed based on a neural network is used to output an optimal cooperative control strategy, realizing the optimization of the overall control strategy of the aircraft formation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is a schematic flowchart of the self-learning-based cooperative control method for aircraft formations provided by the present invention; Figure 2 is a schematic structural diagram of the self-learning-based cooperative control device for aircraft formations provided by the present invention; Figure 3 is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] It should be noted that considering that the calculation of physical models is usually relatively complex and requires a large amount of computing resources and time, this is an important limiting factor in real-time formation cooperative control. In addition, with the improvement of aircraft performance and the diversification of flight missions, traditional formation cooperative control methods are difficult to meet new requirements. Especially when facing a complex and changeable flight environment, their prediction accuracy and adaptability are insufficient, resulting in some limitations in the actual application of the aircraft formation cooperative control method that relies on physical models.

[0020] In view of the above problems, the present invention provides a self-learning based aircraft formation cooperative control method to achieve accurate and highly generalized formation cooperative control. Figure 1 is a schematic flowchart of the self-learning based aircraft formation cooperative control method provided by the present invention, as Figure 1 shown, the method includes: Step 110, obtaining the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; Here, the flight state data refers to the motion state data of the aircraft, such as data on position, speed, acceleration, etc. The state influence data here refers to the data that affects the flight state of the aircraft, which may include internal influence data and external influence data. Among them, the internal influence data is, for example, the control instruction of the aircraft. The external influence data may include at least one of aerodynamic parameters, aircraft structure parameters, and environmental parameters.

[0021] Specifically, for obtaining the flight data of a single aircraft, the real-time acquisition and monitoring data can be obtained through sensors pre-set on each aircraft in the aircraft formation, so as to obtain the flight data of each aircraft at the current moment. For example, through sensors such as GPS, gyroscopes, and accelerometers, the flight state data such as position, speed, acceleration, and attitude angle can be collected in real time. In addition, the state influence data can be obtained by acquiring the control instructions received by the aircraft and by obtaining the external factor data that may affect the aircraft trajectory, such as wind speed, wind direction, air pressure, temperature, etc., from a weather station, radar or other external data sources.

[0022] Thus, the aircraft can communicate and connect with each other through a wireless communication network to obtain the flight data of other aircraft except itself.

[0023] Step 120, based on the observer, predict using the flight data of each aircraft at the current moment to obtain the leading flight state of the leading aircraft at the next moment and the system matrix of the aircraft formation. Here, the leading aircraft refers to the aircraft that leads the flight direction of the entire aircraft formation. Thus, the leading flight state refers to the flight state data of the leading aircraft at the next moment, such as position, speed, acceleration, and other data.

[0024] In addition, the system matrix here refers to the mathematical model of the aircraft dynamic characteristics of each aircraft in the aircraft formation, which is used to analyze and design the system control strategy.

[0025] Specifically, all the flight data of each aircraft at the current moment can be input into the observer. Based on all the flight data, the observer predicts the leading flight state of the leading aircraft at the next moment and the system matrix of the aircraft formation.

[0026] Step 130, based on the controller, apply the leading flight state and the system matrix to output a cooperative control strategy. The system model includes the observer and the controller. The system model is obtained by training the initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0027] Specifically, in the application stage of the system model, the leading flight state and the system matrix can be input into the controller, and the controller outputs the cooperative control strategy of the aircraft formation. Then, the cooperative control strategy can be sent to each aircraft in the aircraft formation simultaneously to control the flight trajectories of each aircraft, update the flight states of each aircraft, and achieve the optimal output clustering synchronization control of heterogeneous multi-aircraft.

[0028] Here, the observer and the controller can jointly form the system model. The aircraft formation is constructed based on a neural network and is self-learned and trained in a data-driven manner through sample flight data. It should be noted that in the training stage of the system model, the observer and the controller can be trained together as the system model. In the application stage of the system model, the observer can be deployed on each aircraft in the aircraft formation, and the controller can be deployed on the central computing device that establishes a communication connection with each aircraft, or on the leading aircraft.

[0029] The training steps of the system model include: First, determine the initial system model and sample flight data. Specifically, initial parameters can be selected within an appropriate range, and these parameters may include state variables such as the initial position, velocity, and acceleration of the aircraft, as well as control parameters such as control gains and observer gains. An error bound and a stopping criterion can also be selected. Symmetric matrices Q and R are selected. Among them, the error bound is used to determine the acceptable range of control errors, and the stopping criterion is used to judge the convergence of the algorithm; the symmetric matrices Q and R are used to control the importance of flight state data and state influence data in the performance index respectively. It should be noted that the above selected model parameters ensure that the algorithm can obtain the optimal control law, thus ensuring the effectiveness and reliability of the control strategy.

[0030] Then, the sample flight data can be input into the action network in the initial system model, and the sample cooperative control strategy can be output through the action network. Then, the updated flight state corresponding to the sample cooperative control strategy is generated. Then, based on the evaluation network in the initial system model, the updated flight state is evaluated to obtain the training strategy reward; based on the training strategy reward, the model parameters of the initial system model are iteratively parameterized to obtain the final system model.

[0031] It should be noted that the system model trained in the embodiment of the present invention optimizes the control strategy through the data-driven policy iteration method, reduces the computational complexity, improves the computational efficiency, and thus meets the requirements of real-time control and has a wide range of application prospects. Therefore, the trained system model can adapt to the improvement of the aircraft performance and the diversification of flight missions. Especially when facing a complex and changeable flight environment, it can maintain a high prediction accuracy and adaptability. And, on the premise of ensuring flight safety, it can also optimize the energy consumption and flight performance of the aircraft, improve the efficiency and economy of flight missions, and thus provide a new solution for the cooperative control of aircraft formations in the fields of unmanned aerial vehicle formations and satellite formations.

[0032] The method provided by the embodiment of the present invention estimates the leading flight state of the leader and the system matrix of the aircraft formation through an observer constructed based on a neural network. By learning the internal laws and characteristics between the sample flight data through deep learning, it has the generalization ability for unknown data, avoids the limitations of traditional physical model construction, and greatly improves the accuracy and generalization of the estimated leading flight state and system matrix. In addition, a controller constructed based on a neural network is used to output the optimal cooperative control strategy, realizing the optimization of the overall control strategy of the aircraft formation.

[0033] Based on any of the above embodiments, step 130 includes: Based on the evaluation network in the controller, evaluate the flight state data at the current moment to obtain the current policy reward; Based on the action network in the controller, apply the leading flight state, the system matrix, and the current policy reward to output the cooperative control strategy.

[0034] Here, the controller may include an evaluation network and an action network, and both the evaluation network and the action network here are constructed based on neural networks. Among them, the evaluation network is used to approximate the performance index function, and the action network is used to approximate the system control law.

[0035] Specifically, through the evaluation network, evaluate the flight state data at the current moment to obtain the current policy reward. It can be understood that the current policy reward here can reflect the control reward of the cooperative control strategy at the previous moment.

[0036] Then, through the action network, apply the leading flight state, the system matrix, and the current policy reward to find the optimal control law to satisfy the performance index function, and then output the cooperative control strategy at the next moment.

[0037] The method provided by the embodiments of the present invention evaluates the flight state data at the current moment through the evaluation network in the controller to obtain the current policy reward; based on the action network in the controller, the leading flight state, the system matrix, and the current policy reward, output the cooperative control strategy, realize adaptive dynamic programming, and use the input flight state data to learn the optimal control law of each aircraft, thereby realizing effective suppression of disturbances. During the control process, input constraints and unknown disturbances during the flight process are fully considered, making the control law more robust.

[0038] Based on any of the above embodiments, the aircraft formation includes the leader aircraft and multiple follower aircraft; The observers are respectively arranged on the leader aircraft and each follower aircraft; The flight data of each aircraft at the current moment is shared between the leader aircraft and multiple follower aircraft.

[0039] Specifically, in the actual application stage, the trained observers can be respectively arranged in the leader aircraft and each follower aircraft, and each aircraft realizes the sharing of flight data during the flight. For example, for any aircraft, after obtaining its own flight data, it can send its own flight data to other aircraft through wireless communication. At the same time, it also receives the flight data sent by other aircraft. In one embodiment, each follower aircraft can send its own flight data to the leader aircraft, and then the leader aircraft distributes the flight data of each follower aircraft to each follower aircraft respectively.

[0040] It should be noted that during flight, when individual aircraft malfunction, the leader flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation can be estimated through the observers set in other normal aircraft, thereby improving the robustness of the cooperative control of the aircraft formation.

[0041] The method provided by the embodiment of the present invention can accurately estimate the state and system matrix of the aircraft through a distributed adaptive observer and a data-driven optimal control strategy, thereby improving the accuracy of the cooperative control of the aircraft formation.

[0042] Based on any of the above embodiments, the sample flight data includes sample flight state data and sample state influence data; The sample state influence data includes control commands and external influence data; The external influence data includes at least one of aerodynamic parameters, aircraft structure parameters, and environmental parameters.

[0043] Specifically, in the construction stage of the sample flight data, the flight data at historical moments can be obtained as the sample flight data.

[0044] Among them, the sample flight data includes sample flight state data and sample state influence data. The sample state influence data includes control commands and external influence data. Among them, the external influence data includes at least one of aerodynamic parameters, aircraft structure parameters, and environmental parameters.

[0045] In detail, the aerodynamic parameters are used to reflect the interaction between the aircraft and the surrounding airflows, including the lift, drag, stability, and maneuverability of the aircraft, etc.; the aircraft structure parameters reflect information on aspects such as the structural design and material properties of the aircraft, including the weight, center of gravity position, structural strength, stiffness, and material properties of the aircraft, etc. The aircraft structure parameters have important impacts on the stability, maneuverability, and safety of the aircraft; the environmental parameters are used to reflect the atmospheric conditions in the flight environment where the aircraft is located, including air temperature, air pressure, humidity, wind speed, and wind direction, etc.

[0046] In addition, it should also be noted that after the system model training is completed, the external influence data in the sample flight data can be randomly combined to obtain simulation data. The trained system model is simulated using the simulation data to verify the effectiveness of the cooperative control method, realizing the cooperative control of the aircraft formation.

[0047] Based on any of the above embodiments, step 120 includes: Based on the observer, use the flight data and input weights at the current moment for prediction, and output the leader flight state and the system matrix; The input weights are respectively used to control the importance degrees of the flight state data and the state influence data; The input weights include two confrontation matrices.

[0048] Specifically, the importance degrees of the flight state data and the state influence data can be controlled by setting two confrontation matrices respectively. It should be noted that the confrontation matrices here are optimized based on the data-driven training process. In the training stage, the performance indexes for evaluating the cooperative control strategy include indexes such as stability, synchronization accuracy, and robustness.

[0049] In the actual application stage, the flight data at the current moment can be input into the observer, and the flight state data and the state influence data are weighted respectively through the input weights, so as to output more accurate leading flight states and system matrices through the observer.

[0050] Based on any of the above embodiments, Figure 2 is a schematic structural diagram of a formation cooperative control device for an aircraft based on self-learning, as Figure 2 shown, the device includes: An acquisition unit 210, which acquires the flight data of each aircraft in the aircraft formation at the current moment, and the flight data includes flight state data and state influence data; An observation unit 220, based on an observer, applies the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leading aircraft at the next moment and the system matrix of the aircraft formation; A control unit 230, based on a controller, applies the leading flight state and the system matrix to output a cooperative control strategy; The system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0051] The device provided by the embodiment of the present invention realizes the estimation of the leading flight state of the leader and the system matrix of the aircraft formation through an observer constructed based on a neural network, and has the generalization ability for unknown data by learning the internal laws and characteristics between the sample flight data, avoiding the limitations of traditional physical model construction, and greatly improving the accuracy and generalization of the estimated leading flight state and system matrix. In addition, a controller constructed based on a neural network is adopted to output an optimal cooperative control strategy, realizing the optimization of the overall control strategy of the aircraft formation.

[0052] Based on any of the above embodiments, the control unit is specifically used for: Evaluate the flight state data at the current moment based on the evaluation network in the controller to obtain the current policy reward; Based on the action network in the controller, apply the leading flight state, the system matrix, and the current policy reward to output the cooperative control policy.

[0053] Based on any of the above embodiments, the aircraft formation includes the leader aircraft and multiple follower aircraft; The observers are respectively arranged on the leader aircraft and each follower aircraft; The flight data of each aircraft at the current moment is shared between the leader aircraft and multiple follower aircraft.

[0054] Based on any of the above embodiments, the sample flight data includes sample flight state data and sample state influence data; The sample state influence data includes control commands and external influence data; The external influence data includes at least one of aerodynamic parameters, aircraft structure parameters, and environmental parameters.

[0055] Based on any of the above embodiments, the device further includes a training unit, and the training unit is specifically used for: Obtained based on the sample flight state data, control commands, and at least one randomly selected external influence data combination at any historical moment.

[0056] Based on any of the above embodiments, the observation unit is specifically used for: Based on the observer, apply the flight data and input weights at the current moment for prediction, and output the leading flight state and the system matrix; The input weights are respectively used to control the importance of the flight state data and the state influence data; The input weights include two symmetric matrices.

[0057] Figure 3 Illustrates a schematic physical structure diagram of an electronic device, such as Figure 3As shown in the figure, the electronic device may include: a processor 310, a communications interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communications interface 320, and the memory 330 complete communication with each other through the communication bus 340. The processor 310 may call the logical instructions in the memory 330 to execute a collaborative control method for aircraft formation based on self-learning. The method includes: obtaining the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; based on an observer, applying the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; based on a controller, applying the leading flight state and the system matrix to output a collaborative control strategy; the system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0058] In addition, when the logical instructions in the above-mentioned memory 330 can be implemented in the form of software functional units and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk, or an optical disc that can store program codes.

[0059] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the self-learning-based cooperative control method for an aircraft formation provided by each of the above methods. The method includes: obtaining the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; based on an observer, applying the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; based on a controller, applying the leading flight state and the system matrix to output a cooperative control strategy; the system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0060] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute the self-learning-based cooperative control method for an aircraft formation provided by each of the above methods. The method includes: obtaining the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; based on an observer, applying the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; based on a controller, applying the leading flight state and the system matrix to output a cooperative control strategy; the system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

[0061] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0062] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A formation cooperative control method for aircraft based on self-learning, characterized in that, Including: Obtain the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; Based on an observer, apply the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; Based on a controller, apply the leading flight state and the system matrix to output a cooperative control strategy; The system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

2. The method for collaborative control of aircraft formation based on self-learning according to claim 1, wherein The step of applying the leading flight state and the system matrix based on the controller to output a cooperative control strategy includes: Based on the evaluation network in the controller, evaluate the flight state data at the current moment to obtain the current policy reward; Based on the action network in the controller, apply the leading flight state, the system matrix, and the current policy reward to output the cooperative control strategy.

3. The collaborative control method for aircraft formation based on self-learning according to claim 1, wherein The aircraft formation includes the leader aircraft and multiple follower aircraft; The observer is respectively arranged on the leader aircraft and each follower aircraft; The flight data of each aircraft at the current moment is shared between the leader aircraft and multiple follower aircraft.

4. The self-learning-based formation cooperative control method for aircraft according to any one of claims 1 to 3, characterized in that The sample flight data includes sample flight state data and sample state influence data; The sample state influence data includes control instructions and external influence data; The external influence data includes at least one of aerodynamic parameters, aircraft structure parameters, and environmental parameters.

5. The method for cooperative control of aircraft formation based on self-learning according to claim 4, characterized in that The determination step of the sample flight data includes: It is obtained by combining the sample flight state data, control instructions, and at least one randomly selected external influence data at any historical moment.

6. The self-learning-based formation cooperative control method for aircraft according to any one of claims 1 to 3, characterized in that The step of applying the flight data of each aircraft at the current moment based on the observer for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation includes: Based on the observer, apply the flight data at the current moment and input weights for prediction, and output the leading flight state and the system matrix; The input weights are respectively used to control the importance degrees of the flight state data and the state influence data; The input weights include two symmetric matrices.

7. An aircraft formation cooperative control device based on self-learning, characterized in that Including: An acquisition unit that acquires the flight data of each aircraft in the aircraft formation at the current moment, where the flight data includes flight state data and state influence data; An observation unit that, based on an observer, applies the flight data of each aircraft at the current moment for prediction to obtain the leading flight state of the leader aircraft at the next moment and the system matrix of the aircraft formation; A control unit that, based on a controller, applies the leading flight state and the system matrix to output a cooperative control strategy; The system model includes the observer and the controller, and the system model is obtained by training an initial system model based on sample flight data, and the initial system model is constructed based on a neural network.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the self-learning-based cooperative control method for aircraft formations as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the self-learning-based cooperative control method for aircraft formations as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the self-learning-based cooperative control method for aircraft formations as described in any one of claims 1 to 6.

Images