Uncertainty Evaluation Method for Intelligent Decision-making of UAV Swarms Facing Complex Tasks

Abstract

The present invention relates to an uncertain evaluation method for intelligent decision-making of an unmanned aerial vehicle (UAV) cluster facing complex tasks. The method includes: obtaining multi-objective operation tasks of an operation scenario, where the multi-objective operation tasks include: a plurality of operation objectives, external environmental factors, and an operation plan. Constructing an evaluation objective functional for the UAV cluster to execute the operation according to the multi-objective operation tasks. Decomposing the operation activities of the UAV cluster according to the operation scenario to obtain a plurality of sub-operation activities. Constructing an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, the sub-operation activities, the operation objectives, the external environmental factors, and the uncertainty parameters. After determining the intelligent decision-making model according to the index evaluation system, performing an uncertain evaluation on the operation plan. Using this method can improve the analysis efficiency of the UAV cluster operation system and reduce the overall test cost.

Description

Technical Field

[0001] The present invention relates to the technical field of UAV swarm decision-making evaluation, and particularly to an intelligent decision-making uncertainty evaluation method for UAV swarms facing complex tasks. Background Art

[0002] When a UAV swarm faces complex tasks, the system usually has a complex composition and high experimental costs. Therefore, it is necessary to build a surrogate model based on experimental data to improve the analysis efficiency and reduce the overall experimental costs. Due to the high experimental costs, the UAV swarm system often faces experimental data with a small sample size, which is suitable for building a surrogate model based on dynamic Gaussian process regression. The purpose of the UAV swarm is to complete related operations, which often requires the UAV swarm to cooperate with other operation parties for systematic application. Therefore, the operations of the swarm need to be carried out from a systematic perspective.

[0003] UAV swarms usually have a complex composition, high experimental costs, and often face experimental data with a small sample size. UAV swarms have high uncertainty during operations. Their prominent feature is the diverse application methods, and under different application methods, they are affected by various factors such as operation parties, objects to be operated on, and the environment. The high uncertainty of UAVs requires the comprehensive application of surrogate models. The actual UAV swarm may simultaneously have characteristics such as strong uncertainty, complex parameters, spatial instability, and spatio-temporal instability, and randomly convert the system representation form due to changes in the environment and working conditions. At this time, a single surrogate model cannot comprehensively build the system, and the comprehensive application of surrogate models is required. However, the problem of uncertainty representation of the UAV swarm system has not been fully solved, which affects the comprehensive application of surrogate models, resulting in low operation energy efficiency and high enterprise application costs. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide an intelligent decision-making uncertainty evaluation method for UAV swarms facing complex tasks that can improve the overall operation efficiency of UAV swarms.

[0005] An intelligent decision-making uncertainty evaluation method for UAV swarms facing complex tasks, the method comprising:

[0006] Obtain multi-objective operation tasks for the operation scenario, where the multi-objective operation tasks include: a plurality of operation objectives, external environmental factors, and operation plans.

[0007] Construct an evaluation objective functional for the UAV swarm to execute operations according to the multi-objective operation tasks.

[0008] Decompose the operation activities of the UAV swarm according to the operation scenario to obtain a number of sub-operation activities.

[0009] Construct an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, sub-operation activities, operation objectives, external environmental factors, and uncertainty parameters.

[0010] After determining the intelligent decision-making model according to the index evaluation system, conduct an uncertainty evaluation of the operation plan.

[0011] An uncertainty evaluation device for intelligent decision-making of an unmanned aerial vehicle (UAV) cluster for complex tasks, the device includes:

[0012] A multi-objective operation task acquisition module, used to acquire the multi-objective operation tasks of the operation scenario, where the multi-objective operation tasks include: several operation objectives, external environmental factors, and operation plans.

[0013] An objective functional construction module, used to construct an evaluation objective functional for the UAV cluster to execute operations according to the multi-objective operation tasks.

[0014] A task decomposition module, used to decompose the operation activities of the UAV cluster according to the operation scenario to obtain several sub-operation activities.

[0015] An index evaluation system construction module, used to construct an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, sub-operation activities, operation objectives, external environmental factors, and uncertainty parameters;

[0016] An uncertainty evaluation module, used to conduct an uncertainty evaluation of the operation plan after determining the intelligent decision-making model according to the index evaluation system.

[0017] A computer device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0018] Acquire the multi-objective operation tasks of the operation scenario, where the multi-objective operation tasks include: several operation objectives, external environmental factors, and operation plans.

[0019] Construct an evaluation objective functional for the UAV cluster to execute operations according to the multi-objective operation tasks.

[0020] Decompose the operation activities of the UAV cluster according to the operation scenario to obtain several sub-operation activities.

[0021] Construct an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, sub-operation activities, operation objectives, external environmental factors, and uncertainty parameters.

[0022] After determining the intelligent decision-making model according to the index evaluation system, conduct an uncertainty evaluation of the operation plan.

[0023] The above-mentioned method for uncertain evaluation of intelligent decision-making of UAV clusters for complex tasks fully considers various factors in multi-objective operation tasks, such as several operators, external environmental factors, and operation plans, when constructing the evaluation objective functional, to ensure that the functional can comprehensively reflect the actual situation of UAV cluster operations. After obtaining sub-operation activities by disassembling operation activities, an index evaluation system is constructed by combining the evaluation objective functional, multi-objective states, external environmental factors, and uncertainty parameters. In this process, more in-depth research and analysis should be carried out on the uncertainty parameters to accurately characterize the uncertainty of the UAV cluster system. Next, when determining the intelligent decision-making model, multiple surrogate models should be comprehensively used. For example, for different sub-operation activities and uncertainty factors, appropriate surrogate models are selected and then these models are organically integrated. The advantages of different surrogate models can be complemented to enable the comprehensive model to more comprehensively describe the UAV cluster system. At the same time, during the process of using surrogate models, the model parameters are continuously optimized to improve the model's adaptability to uncertainty. Finally, when the UAV cluster faces complex tasks, a more reasonable surrogate model can be constructed. This can not only improve the analysis efficiency of the UAV cluster operation system but also reduce the overall test cost, enabling the UAV cluster to better play its role in complex task environments and providing more reliable technical support for practical applications. Brief Description of the Drawings

[0024] Figure 1 It is a schematic flowchart of the method for uncertain evaluation of intelligent decision-making of UAV clusters for complex tasks in an embodiment;

[0025] Figure 2 It is a flowchart of the uncertainty evaluation index in an embodiment;

[0026] Figure 3 It is a flowchart of the analysis of the UAV cluster operation system in an embodiment;

[0027] Figure 4 It is a schematic diagram of the house of quality in an embodiment;

[0028] Figure 5 It is a mapping relationship diagram of the index system based on quality function deployment in an embodiment;

[0029] Figure 6 It is a structural block diagram of the device for uncertain evaluation of intelligent decision-making of UAV clusters for complex tasks in an embodiment;

[0030] Figure 7 It is an internal structure diagram of a computer device in an embodiment. Detailed Embodiments

[0031] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0032] In one embodiment, as Figure 1 shown, a method for uncertain evaluation of intelligent decision-making of an unmanned aerial vehicle (UAV) cluster for complex tasks is provided, including the following steps:

[0033] Step 102: Obtain multi-objective operation tasks of the operation scenario, where the multi-objective operation tasks include: several operation objectives, external environmental factors, and operation plans.

[0034] Step 104: Construct an evaluation objective functional for the UAV cluster to execute the operation according to the multi-objective operation tasks.

[0035] Step 106: Decompose the operation activities of the UAV cluster according to the operation scenario to obtain several sub-operation activities.

[0036] Step 108: Construct an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, sub-operation activities, operation objectives, external environmental factors, and uncertainty parameters.

[0037] Step 110: After determining the intelligent decision-making model according to the index evaluation system, perform uncertain evaluation on the operation plan.

[0038] In the above-mentioned method for uncertain evaluation of intelligent decision-making of UAV clusters for complex tasks, when constructing the evaluation objective functional, various factors in multi-objective operation tasks are fully considered, such as several operators, external environmental factors, and operation plans, etc., to ensure that the functional can comprehensively reflect the actual situation of UAV cluster operations. After decomposing the operation activities into sub-operation activities, an index evaluation system is constructed by combining the evaluation objective functional, multi-objective states, external environmental factors, and uncertainty parameters. In this process, more in-depth research and analysis should be carried out on the uncertainty parameters to accurately characterize the uncertainty of the UAV cluster system. Next, when determining the intelligent decision-making model, multiple surrogate models should be comprehensively used. For example, for different sub-operation activities and uncertainty factors, appropriate surrogate models are selected, and then these models are organically integrated. The advantages of different surrogate models can be complemented to enable the comprehensive model to more comprehensively describe the UAV cluster system. At the same time, during the process of using the surrogate model, the model parameters are continuously optimized to improve the adaptability of the model to uncertainty. Finally, when the UAV cluster faces complex tasks, a more reasonable surrogate model can be constructed. This can not only improve the analysis efficiency of the UAV cluster operation system but also reduce the overall test cost, enabling the UAV cluster to better play its role in complex task environments and providing more reliable technical support for practical applications.

[0039] In one embodiment, according to the multi-objective operation task, the evaluation objective functional is constructed for the influencing factors of each UAV subgroup in the UAV cluster by using the functional principle:

[0040] 。

[0041] Where N is the number of UAV clusters or the number of UAV subgroups, is the state of each UAV cluster or each UAV subgroup changing with time t, is the state change rate, is the external environmental factor, and M is the number of external environmental factors, is the performance weight of the operation activity corresponding to the current time t in all operation task profiles, and K is the total number of operation activities, is the random variable of uncertainty, is the performance function of the i-th UAV cluster or each UAV subgroup at time t, represents the starting point of the evaluation time, is the end point of the evaluation time, is the influence function of the j-th external environmental factor at time t, is the random term considering the uncertainty factor.

[0042] In one embodiment, the sub-job activities are arranged with both parties of the job in chronological order to obtain an ordered matrix of sub-jobs. A Gantt chart is used to formulate the job execution order and data flow relationship for the ordered matrix of sub-jobs, resulting in uncertain sub-activities.

[0043] In one embodiment, an index evaluation system for the corresponding job party of the uncertain sub-activity is constructed using a House of Quality according to the evaluation objective functional, sub-job activities, job objectives, external environmental factors, and uncertainty parameters. The index evaluation system includes: a job objective - job scenario House of Quality, a job scenario - job activity House of Quality, and a job activity - uncertainty parameter House of Quality.

[0044] In one embodiment, in the job party - job scenario House of Quality of the job objective - job scenario House of Quality, a set of job objectives corresponding to the current round of the UAV cluster is set, and the mapping relationships in the job scenarios corresponding to each job objective in the set of job objectives are sorted out respectively to determine the uncertainty of the job objectives.

[0045] In one embodiment, in the job scenario - job activity House of Quality, the relationship of the uncertainty ranges of different sub-job activities in different job scenarios is mapped. In the job activity - uncertainty parameter House of Quality, the relationship of the uncertainty ranges between the UAV cluster or UAV subgroup as the job party and the job object in different sub-job activities is mapped.

[0046] In one embodiment, according to the index evaluation system, when there is uncertainty in the process of the UAV subgroup executing the job activity, test sample points are selected according to the uncertainty parameters. After sequential testing of the test sample points through the constructed standard surrogate model, an uncertain evaluation of the job plan is carried out.

[0047] In one embodiment, a standard surrogate model for the test sample points is constructed using the Gaussian process regression algorithm:

[0048] ;

[0049] ;

[0050] where is the time predicted by the standard surrogate model and the output value of the test sample point in the uncertainty parameter, is the uncertainty parameter, is the function of the standard surrogate model, is the parameter set of the standard surrogate model, is the covariance matrix between the current moment test sample point and the next moment test sample point, is the Gaussian process regression function, is the covariance function, is the mean function. The Bayesian method is used to design the experiment for the test sample points through the constructed standard surrogate model, and the Sobol index method is used to conduct a global sensitivity analysis on the output of the test sample points. This index measures the contribution of a certain variable or a group of variables to the output variability after removing the influence of all other variables. The overall sensitivity index is obtained:

[0051]

[0052] where, is the total variance of the output, is the overall sensitivity index, is the variable being analyzed in the current experiment for all other variables in the current experiment is the variance after taking the average. In the evaluation of the sub-job activities of the UAV swarm, according to the overall sensitivity index, the parameters of the standard surrogate model are calibrated by the inverse problem, and the probability distribution of the uncertainty parameters is updated through Bayesian inversion to complete the uncertain evaluation of the operation plan:

[0053]

[0054] where, is the posterior distribution of the parameters of the standard surrogate model, is the observed data is the probability under the condition of the parameters of the standard surrogate model, is the prior distribution of the parameters of the standard surrogate model.

[0055] In one embodiment, as Figure 2 shown, a flowchart of the uncertainty evaluation index is provided. The four-layer index system constructed by the uncertainty analysis process is "mission - operation scenario - operation activity - uncertainty parameter". The first three layers of indexes correspond to the three layers of assessment objects of "operator ability - operator efficiency - operator performance". The bottom layer of the index system is the uncertainty parameter. There may be multiple sources of uncertainty parameters in the whole operation activity, such as UAV technical performance indicators, operation strategy parameters, environmental parameters, etc. In a single sub-job, the uncertainty parameters affect the progress of the sub-job activity and put forward requirements for the overall performance of the UAV.

[0056] In one embodiment, as Figure 3 shown, a flowchart of the analysis of the UAV swarm operation system is provided, which specifically includes the following steps:

[0057] The first step is to construct the evaluation objective functional.

[0058] In the evaluation of UAV swarm systems, a complex network composed of interdependent and interacting systems or subsystems is faced. Adopting the functional concept, the UAV system can be regarded as a whole, and its multi-dimensional performance can be regarded as a function of influencing factors. The constructed functional not only depends on the states of all systems or subsystems, but also on the rate of change of these states over time and external environmental factors, and is used to comprehensively evaluate the effectiveness of the equipment system. The UAV system's capabilities or its ability to complete tasks can be regarded as a functional in a higher-dimensional space. The evaluation objective functional of the complex equipment system can be expressed as:

[0059]

[0060] Among them, there are N systems or subsystems, and each system or subsystem has a state that changes over time t, is the rate of change of the state, and is affected by M external environmental factors 's influence, represents the performance weight of the equipment system in different mission profiles, K is the total number of operation activities, and the random variable represents the influence brought by uncertainty. is the performance function of the i-th UAV swarm or each UAV subgroup at time t, represents the starting point of the evaluation time, is the end point of the evaluation time, is the influence function of the j-th external environmental factor at time t, is a random term considering uncertainty factors.

[0061] Second step, design of the structured analysis method.

[0062] 2.1 Configure the operation scenario and draw the operation concept diagram.

[0063] Configure the operation scenario where the party to be operated is located, and describe the composition method, cooperation relationship, activity relationship, etc. between the UAV swarm and the party to be operated.

[0064] 2.2 Decompose the operation activities and draw the Gantt chart of the operation sub-activities.

[0065] Decompose the UAV swarm activities into operation sub-activities, arrange them according to the time sequence and the two parties of the operation, and use the Gantt chart to clarify the time relationship of the UAV swarm operation sub-activities, so as to formulate the activity execution sequence and data flow relationship in a specific operation scenario, and clarify the sub-activities with uncertainty.

[0066] 2.3 Construct an index system to characterize the uncertainty of the equipment under test.

[0067] According to the four-layer index system of "objective - scenario - operation task - uncertainty parameter", such asFigure 4 As shown, matrix decomposition is carried out through the quality house of quality function deployment, and it is expanded step by step to form an index system for the party being operated, which characterizes the uncertainty of the party being operated. The quality house consists of the following parts:

[0068] A Correlation Matrix: The roof of the quality house, which is the autocorrelation matrix of the lower-level indicators, indicating whether there is a correlation between the lower-level indicators. If there is no correlation, the roof can be removed.

[0069] B Lower-Level Indicators: The ceiling of the quality house, representing the lower-level indicators. The determination of the lower-level indicators depends on the upper-level indicators.

[0070] C Internal Relationship Matrix: The room of the quality house, representing the corresponding relationship between the upper-level indicators and the lower-level indicators.

[0071] D Additional Information of Lower-Level Indicators: The floor of the quality house, which generally can include elements such as the importance of the lower-level indicators and the measurement scheme. These can be used as the basis for subsequent test item tailoring and parameter construction.

[0072] E Upper-Level Indicators: The left wall of the quality house, which is the basis for determining the lower-level indicators.

[0073] F Additional Information of Upper-Level Indicators: The right wall of the quality house, which generally includes the importance of the upper-level indicators. It plays the same role as the area.

[0074] As Figure 5 shown, the index system is combined in sequence to form a mapping relationship. There are three quality houses in the index system based on quality function deployment, namely the "Goal - Operation Scenario" quality house, the "Operation Scenario - Operation Activity" quality house, and the "Operation Activity - Uncertainty Parameter" quality house. All three quality houses represent the mapping between interrelated relationships. For example, in the "Goal - Operation Scenario" quality house, first, the operation goals of this UAV cluster are set as "environmental detection, item transportation, and photogrammetry", and the mapping relationships of the three operation goals in the "traffic monitoring, agricultural plant protection, and building planning" operation scenarios are sorted out respectively to clarify whether there is uncertainty in the operation goals in specific scenarios, whether there is information about the party being operated, etc. The "Operation Scenario - Operation Activity" quality house sorts out whether there is uncertainty in different operation activities in different operation scenarios, while the "Operation Activity - Uncertainty Parameter" quality house sorts out the mapping relationship of the uncertainty ranges between the operator and the party being operated in different operation activities.

[0075] 2.4 Comprehensively apply digital models to analyze test data

[0076] When there is uncertainty in the process of a drone subgroup performing tasks, experimental design is carried out according to the parameters of the uncertainty, experimental sample points are selected, models are constructed and analyzed to obtain research conclusions. Supplementary experiments can also be carried out according to the actual situation, or the operation scenario can be adjusted for iterative experiments.

[0077] Step 3 Uncertainty quantification and evaluation

[0078] Uncertainty quantification requires constructing a suitable surrogate model, performing sensitivity analysis on the influencing factors to determine the important factors, further based on the inverse problem solving framework, designing a Bayesian inversion method for parameter inference to obtain the final quantification evaluation result. The present invention uses the self-developed nudt_UQtool toolbox, which covers experimental design methods, surrogate model construction methods, sensitivity analysis methods, etc., and proposes a series of uncertainty quantification methods with stronger adaptability, such as dynamic Gaussian process regression, multi-fidelity surrogate model construction, sequential adaptive experimental design, and sensitivity analysis, for the requirements of drone swarm operations.

[0079] 3.1 Surrogate model construction and sequential experimental design method based on surrogate model

[0080] Use Gaussian process regression to construct the functional relationship between input parameters and output responses. Given the training data set , the goal of Gaussian process regression is to learn a Gaussian process and predict the output corresponding to the new input point . The posterior distribution can be expressed as:

[0081]

[0082] where is the mean function, is the covariance matrix between training data points, is the covariance vector between training data points and new input points, and is the self-covariance of the new input point.

[0083] Then use the spatio-temporal Gaussian process regression method to dynamically discover the local space, mine diverse time-varying patterns to construct a spatio-temporal surrogate model, and capture system change points and coupling structures. Specifically, the spatio-temporal surrogate model can be expressed by the following formula:

[0084]

[0085] where is the output value of the surrogate model prediction at time and spatial position ; is in the form of a surrogate model, which can be a polynomial, neural network, Gaussian process, or other forms of functions; is a set of parameters of the surrogate model. In a Gaussian process, it can be represented by the following formula:

[0086]

[0087] where is the mean function, which describes the expected value at time and position ; is the covariance function, which measures the correlation between observations at different time points and spatial positions. The covariance function needs to be modified accordingly. For example, the spatio-temporal covariance function can adopt a separable form, that is, the time and space parts can be processed independently:

[0088]

[0089] where is the time covariance function; is the spatial covariance function.

[0090] Then Bayesian optimization is used for experimental design, and Gaussian process regression is used for digital model construction.

[0091] 3.2 Global Sensitivity Analysis

[0092] The global sensitivity analysis method used in the present invention is the Sobol index method. This index measures the contribution of a variable or a group of variables to the output variability after removing the influence of all other variables. The Sobol total sensitivity index can be expressed as:

[0093]

[0094] where is the total variance of the output, and is the variance after fixing the variable and averaging over other variables .

[0095] 3.3 Parameter Inference Based on Bayesian Inversion

[0096] In the evaluation of UAV swarm activities, inverse problems can be used to calibrate model parameters so as to match experimental data with actual operating conditions. The present invention updates the probability distribution of unknown parameters through Bayesian inversion:

[0097]

[0098] where is the parameter The posterior distribution, is the likelihood function, which describes the observed data given the parameter and is the prior distribution of the parameter.

[0099] It should be understood that although Figures 1 - 3 , Figure 5 the steps in the flowchart of Figures 1 - 3 , Figure 5 are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover,

[0100] In one embodiment, as Figure 6 shown, a device for uncertain evaluation of intelligent decision-making of an unmanned aerial vehicle (UAV) cluster for complex tasks is provided, including: a multi-objective operation task acquisition module 602, an objective functional construction module 604, a task decomposition module 606, an index evaluation system construction module 608, and an uncertainty evaluation module 610, where:

[0101] The multi-objective operation task acquisition module 602 is configured to acquire the multi-objective operation tasks of the operation scenario, where the multi-objective operation tasks include: several operation objectives, external environmental factors, and operation plans.

[0102] The objective functional construction module 604 is configured to construct an evaluation objective functional for the UAV cluster to execute the operation according to the multi-objective operation tasks.

[0103] The task decomposition module 606 is configured to decompose the operation activities of the UAV cluster according to the operation scenario to obtain several sub-operation activities.

[0104] The index evaluation system construction module 608 is configured to construct an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, sub-operation activities, operation objectives, external environmental factors, and uncertainty parameters;

[0105] The uncertainty evaluation module 610 is configured to perform an uncertain evaluation of the operation plan after determining the intelligent decision-making model according to the index evaluation system.

[0106] For the specific limitations of the UAV cluster intelligent decision-making uncertainty evaluation device for complex tasks, reference can be made to the limitations of the UAV cluster intelligent decision-making uncertainty evaluation method for complex tasks in the above text, which will not be elaborated here. Each module in the above UAV cluster intelligent decision-making uncertainty evaluation device for complex tasks can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0107] In one embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 7 shown. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used 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 the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a UAV cluster intelligent decision-making uncertainty evaluation method for complex tasks. 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 covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0108] Those skilled in the art can understand that Figures 6 - 7 the structure shown in

[0109] is only a block diagram of some structures related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0109] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0110] Obtain a multi-objective operation task for the operation scenario, where the multi-objective operation task includes: several operation objectives, external environmental factors, and an operation plan.

[0111] Construct an evaluation objective functional for the UAV cluster to execute the operation according to the multi-objective operation task.

[0112] Decompose the operation activities of the UAV cluster according to the operation scenario to obtain several sub-operation activities.

[0113] Construct an index evaluation system for the corresponding operation activities of each operator according to the evaluation objective functional, sub-operation activities, operation objectives, external environmental factors, and uncertainty parameters.

[0114] After determining the intelligent decision-making model according to the index evaluation system, perform an uncertainty evaluation on the operation plan.

[0115] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided by the present invention can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0116] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0117] The above-described embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the appended claims.

Claims

1. An uncertain evaluation method for intelligent decision-making of UAV clusters facing complex tasks, characterized in that The method includes: Obtaining a multi-objective operation task for an operation scenario, where the multi-objective operation task includes: a number of operation objectives, external environmental factors, and an operation plan; Constructing an evaluation objective functional for the operation of a UAV cluster according to the multi-objective operation task; Decomposing the operation activities of the UAV cluster according to the operation scenario to obtain a number of sub-operation activities; Arranging the sub-operation activities in chronological order with both parties of the operation to obtain an ordered matrix of sub-operations, and using a Gantt chart to formulate the operation execution order and data flow relationship for the ordered matrix of sub-operations to obtain uncertain sub-activities; Constructing an index evaluation system for the operation activities corresponding to each party to be operated according to the evaluation objective functional, the sub-operation activities, the operation objectives, the external environmental factors, and uncertain parameters; Constructing an index evaluation system for the uncertain sub-activities corresponding to the party to be operated using a quality function deployment (QFD) based on the evaluation objective functional, the sub-operation activities, the operation objectives, the external environmental factors, and uncertain parameters; The index evaluation system includes: an operation objective - operation scenario quality function deployment (QFD), an operation scenario - operation activity QFD, and an operation activity - uncertain parameter QFD; After determining an intelligent decision-making model according to the index evaluation system, performing an uncertain evaluation on the operation plan.

2. The method according to claim 1, wherein Constructing an evaluation objective functional for the operation of a UAV cluster according to the multi-objective operation task, including: Constructing an evaluation objective functional for the influencing factors of the operation of each UAV subgroup in the UAV cluster using the functional principle according to the multi-objective operation task; ； where N is the number of UAV subgroups, is the state of each UAV subgroup varying with time t, is the state change rate, is the external environmental factor, and M is the number of external environmental factors, is the performance weight of the operation activity corresponding to the current time t in all operation task profiles, and K is the total number of operation activities, is a random variable of uncertainty, is the performance function of each UAV subgroup at time t, represents the starting point of the evaluation time, is the ending point of the evaluation time, is the influence function of the j-th external environmental factor at time t, is a random term considering uncertainty factors.

3. The method according to claim 1, characterized in that In the quality function deployment (QFD) of the party to be operated - operation scenario, setting the set of operation objectives corresponding to the UAV cluster in the current round, and respectively sorting out the mapping relationships in the operation scenarios corresponding to each operation objective in the set of operation objectives to determine the uncertainty of the operation objectives.

4. The method according to claim 3, characterized in that, In the quality function deployment (QFD) of the operation scenario - operation activity, mapping the relationship of the uncertainty ranges of different sub-operation activities in different operation scenarios; In the quality function deployment (QFD) of the operation activity - uncertain parameter, mapping the relationship of the uncertainty ranges between the UAV cluster or UAV subgroup as the operating party and the party to be operated in different sub-operation activities.

5. The method according to claim 4, wherein After determining an intelligent decision-making model according to the index evaluation system, performing an uncertain evaluation on the operation plan, including: According to the index evaluation system, when there is uncertainty in the process of the UAV subgroup performing operation activities, selecting test sample points according to the uncertain parameters; After performing sequential tests on the test sample points through the constructed standard surrogate model, performing an uncertain evaluation on the operation plan.

6. The method according to claim 5, wherein Using a Gaussian process regression algorithm to construct the standard surrogate model for the test sample points; ； ； Among them, is the time predicted by the standard surrogate model and the output value of the test sample point in the uncertainty parameter, is the uncertainty parameter, is the function of the standard surrogate model, is the parameter set of the standard surrogate model, is the covariance matrix between the test sample point at the current moment and the test sample point at the next moment, is the Gaussian process regression function, is the covariance function, is the mean function; After performing sequential tests on the test sample points through the constructed standard surrogate model, performing an uncertain evaluation on the operation plan, including: Using the constructed standard surrogate model, Bayesian is adopted to conduct experimental design on the test sample points, and the Sobol index method is used to conduct global sensitivity analysis on the outputs of the test sample points to obtain the overall sensitivity index: ； wherein, is the total variance of the output, is the overall sensitivity index, is the variable being analyzed in the current experiment for all other variables in the current experiment is the variance after taking the average. In the evaluation of the sub-operation activities of the UAV swarm, according to the overall sensitivity index, the parameters of the standard surrogate model are calibrated by inverse problem, and the probability distribution of the uncertainty parameters is updated by Bayesian inversion to complete the uncertainty evaluation of the operation plan: ； wherein, is the posterior distribution of the parameters of the standard surrogate model, is the observed data is the probability under the condition of the parameters of the standard surrogate model, is the prior distribution of the parameters of the standard surrogate model.

Images