Machine learning-oriented aircraft moving trajectory accurate segmentation evaluation method and device

Abstract

The invention discloses a machine learning-oriented aircraft moving trajectory precision segmentation evaluation method, device and equipment and a storage medium. The method comprises the following steps: collecting flight track data from a flight log or a sensor, dividing a track into four stages of taking off, cruising, combat and landing according to a flight task, and respectively calculating evaluation indexes of each stage; by comprehensively considering the performance of the aircraft in different flight stages, especially the influence of key indexes such as speed, acceleration, energy consumption and flight stability in each stage, the quality of the flight path can be accurately evaluated, the high-quality flight path can be extracted from the flight path to serve as a training sample of a machine learning model, and the training efficiency is improved. Therefore, the prediction and optimization capabilities of the model in an actual flight task are improved.

Description

Technical Field

[0001] The present invention relates to the field of aerospace technology, and in particular to a method, device, equipment and storage medium for accurately evaluating the trajectory of an aircraft for machine learning. Background Art

[0002] The trajectory of an unmanned aerial vehicle during its mission contains a large amount of dynamic information. The evaluation of the aircraft's trajectory can be used to find the best performing flight trajectory and use it as a training sample for the machine learning model, thereby optimizing the model performance. At the same time, with the rapid development of unmanned aerial vehicles and automatic flight systems, it is particularly important to accurately evaluate and analyze the trajectory of the aircraft.

[0003] The Chinese patent "A system and method for testing the flight performance of fixed-wing UAVs" (publication number: CN108045596A, publication date: May 18, 2018) records and judges the flight trajectory of fixed-wing UAVs through a fixed-wing UAV flight trajectory evaluation system, and evaluates the performance of the UAV based on the detection parameters; the Chinese patent "A method for distributed multi-UAV relay network coverage" (publication number: CN116017479A, publication date: April 25, 2023) establishes a multi-UAV communication channel model, energy consumption model, and collaborative trajectory optimization mathematical model, builds a multi-agent deep reinforcement learning algorithm with centralized training and distributed execution, and equips each UAV with an actor and an evaluator. The actor outputs the flight trajectory according to the UAV status information, and the evaluator outputs the corresponding action value to train the actor, and constructs an actor with a dual data stream structure to improve the dynamics of the UAV.

[0004] However, traditional flight trajectory evaluation methods often rely on simple statistical analysis, lack detailed consideration of the characteristics of different flight phases, and cannot make full use of data to optimize the training process of machine learning models.

[0005] Therefore, there is an urgent need for a method for accurately evaluating the segmented operation trajectory of an aircraft, which can conduct a detailed evaluation of the operation phases with different characteristics of the aircraft and provide support for flight trajectory optimization and machine learning model training. Summary of the invention

[0006] Purpose of the invention: The present invention aims to provide a method and device for accurately evaluating the segmented flight trajectory of an aircraft for machine learning, which can perform detailed evaluation of different flight phases and select high-quality trajectories as machine learning training samples.

[0007] Technical solution: To solve the above technical problems, the technical solution adopted by the present invention is: In a first aspect, a method for accurately evaluating the trajectory of an aircraft by segmentation based on machine learning is provided, comprising the following steps: (1) Obtain flight trajectory data; (2) Segment the flight trajectory; (3) Calculate the evaluation index of each segment and complete the segment evaluation of the flight trajectory.

[0008] As a further solution of the present invention, the flight trajectory is divided into four sections according to the flight mission: take-off section, cruise section, combat section, and landing section.

[0009] As a further solution of the present invention, the calculation method of the evaluation index of the take-off stage is: S takeoff = k a • A takeoff + k θ •SD( θ takeoff ) in, S takeoff is the evaluation index of the take-off phase, A takeoff is the takeoff acceleration, SD( θ takeoff ) is the standard deviation of the pitch angle during takeoff, k a and k θ They are A takeoff and SD( θ takeoff )’s weight coefficient; The calculation formulas for the standard deviation of takeoff acceleration and takeoff pitch angle are: , Where ΔV is the velocity difference between the start and end of the takeoff segment, Δt is the time difference between the start and end of the takeoff segment, It is the take-off phase i The pitch angle of a track point, is the average pitch angle during takeoff, n 1 is the number of trajectory points for the takeoff segment.

[0010] As a further solution of the present invention, the calculation method of the evaluation index of the cruise segment is: S cruise = k f •F +k h •SD( H ) in, S cruise is the evaluation index of the cruise segment, F is the energy efficiency, SD( H ) is the standard deviation of the flight altitude during the cruise phase, k f and k h They are F and SD( H )’s weight coefficient; The calculation formulas for energy efficiency and standard deviation of cruise altitude are: ,

[0011] Among them, F T is the total energy consumption in the cruise phase, F D is the cruising flight distance, It is the cruise section i The aircraft altitude at 2 track points, is the average flight altitude during the cruise phase, n 2 is the number of track points in the cruise segment.

[0012] As a further solution of the present invention, the evaluation index calculation method of the combat phase is: S combat = k m • M + k r • R in, S combat is the evaluation index of the combat phase, M For mobility, R is the reaction speed, k m and k r They are M and the weight coefficient of R; The calculation formulas for mobility and reaction speed are: , in, It is the battle section i 3 The maximum acceleration or deceleration during the maneuver, It is the battle sectioni 3 The maneuvering frequency of the maneuver, k acc and k freq are respectively and The weight coefficient of n 3 is the total number of maneuvers during the battle phase, T react,j is the time of the jth reaction in the combat phase, m It is the total number of reactions in the combat phase.

[0013] As a further solution of the present invention, the evaluation index calculation method of the landing stage is: S landing = k d • D + k v •SD(V landing ) in, S landing is the evaluation index of the landing phase, D is the landing accuracy, SD(V landing ) is the standard deviation of the falling velocity, k d and k v They are D and SD(V landing )’s weight coefficient; The calculation formulas for the standard deviation of landing accuracy and landing speed are: ,

[0014] in, x act , y act The actual landing points x , y coordinate, x pla , y pla The scheduled landing points x , y coordinate, It is the landing stage i The landing speed of the 4 trajectory points, is the average descent velocity during the descent phase, n 4 is the number of trajectory points for the landing phase.

[0015] In the second aspect, a device for accurately evaluating the trajectory of an aircraft by segmentation based on machine learning is also provided, comprising: A data acquisition unit, used for acquiring flight trajectory data; A segmentation unit, used for segmenting the flight trajectory; An index calculation unit, used to calculate the evaluation index of each segment; The evaluation unit is used to complete the segmented evaluation of the flight trajectory according to the evaluation index of each segment.

[0016] In a third aspect, a method for processing an aircraft trajectory data set is also provided, wherein the aircraft trajectory data set is used for model training for aircraft trajectory planning based on machine learning, and the specific steps of the method are as follows: 1) Using the segmented evaluation method described above, each flight trajectory data in the aircraft operation trajectory data set is evaluated segment by segment; 2) The trajectory point data corresponding to the evaluation indicators that meet the set requirements in each segment are selected to form a new data set as training samples for the model of aircraft trajectory planning based on machine learning.

[0017] In a fourth aspect, an electronic device is also provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned segmented evaluation method when executing the computer program.

[0018] In a fifth aspect, a computer-readable storage medium is also provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned segmented evaluation method are implemented.

[0019] Beneficial effects: The proposed method and device for accurate segmented evaluation of aircraft trajectory for machine learning can provide detailed evaluation indicators for flight trajectories at different stages, which are helpful for screening out high-quality flight data and providing more accurate and reliable training samples for flight path prediction and optimization based on machine learning. In addition, the method of the present invention can also help flight control systems better understand the performance characteristics of different flight stages and optimize the design and operation of aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the content of the present invention clearer, the present invention is described in detail below in conjunction with specific embodiments.

[0022] In this example, it is assumed that the UAV undergoes standard operations in four stages: take-off, cruising, combat, and landing in a test mission. Figure 1 As shown, the accurate segmented evaluation method of aircraft trajectory for machine learning includes the following steps: (1) Read flight trajectory data from flight logs or sensors.

[0023] (2) According to the flight mission, the flight trajectory is divided into four stages: take-off, cruise, combat and landing.

[0024] The main flight trajectory data of the four stages of this example are as follows: Take-off phase: Speed ​​change: 0m / s to 30m / s, 10s Pitch Angle: Increased from 0° to 15°.

[0025] Cruising stage: Flight distance: 60km Height stability: Fluctuation ±2m Energy consumption: 10 kWh.

[0026] Battle Phase: Maneuverability: Perform three rapid maneuvers with a maximum acceleration of 20 m / s² Reaction speed: Detects threats and responds, with an average reaction time of 2 seconds Landing phase: Landing accuracy: Deviation from the planned landing point is 3m Landing speed fluctuation: from 20m / s to 0m / s, fluctuation ±1m / s.

[0027] The weight coefficient parameters of this instance are set as follows: k a =0.5, k θ =0.5, k f =0.3, k h =0.7, k m =0.6, kr =0.4, k d =0.8, k v =0.2.

[0028] (3) Calculation of evaluation indicators for the takeoff phase S takeoff : S takeoff = k a • Atakeoff + k θ •SD( θ takeoff ); Takeoff acceleration A takeoff : ; Take-off stability SD( θ takeoff ): Assume that it is simplified to the variable SD( θ takeoff )=15°-0°=15°; Evaluation indicators S takeoff : S takeoff = 0.5•3+0.5•15=1.5+7.5=9.

[0029] (4) Calculation of evaluation indicators during the cruise phase S cruise S cruise = k f •F + k h •SD( H ); Energy efficiency F : ; High stability SD( H ): SD( H )=2m; Evaluation indicators S cruise calculate: S cruise =0.3•0.167+0.7•2=0.05+1.4=1.45.

[0030] (5) Calculation of evaluation indicators for the combat phase S combat : S combat = k m • M + k r • R ; flexibilityM : M=20m / s; Reaction speed R : R = 2s; Evaluation indicators S combat : S combat =0.6•20+0.4•2=12+0.8=12.8.

[0031] (6) Calculation of evaluation indicators for the landing phase S landing : S landing = k d • D + k v •SD(V landing ) Landing accuracy D : D = 3m; Landing stability SD(V landing ): Assume that it is simplified to the variable SD(V landing )=1m / s; Evaluation indicators S landing : S landing =0.8•3+0.2•1=2.4+0.2=2.6.

[0032] Through calculation, the evaluation indicators of each stage are obtained. These data can be used to select the best performing flight trajectories as training samples for machine learning models, thereby optimizing model performance.

[0033] The present invention can accurately evaluate the quality of the flight trajectory by comprehensively considering the performance of the aircraft in different flight phases, especially the influence of key indicators such as speed, acceleration, energy consumption and flight stability in each phase, which is conducive to extracting high-quality flight trajectories as training samples for machine learning models, thereby improving the prediction and optimization capabilities of the model in actual flight missions.

[0034] Furthermore, the present invention also proposes a device for accurately segmenting and evaluating aircraft trajectory for machine learning, comprising: A data acquisition unit, used for acquiring flight trajectory data; A segmentation unit, used for segmenting the flight trajectory; An index calculation unit, used to calculate the evaluation index of each segment; The evaluation unit is used to complete the segmented evaluation of the flight trajectory according to the evaluation index of each segment.

[0035] The technical solution of the above-mentioned segmented extraction device is consistent with the technical solution of the aforementioned segmented extraction method, and will not be repeated here.

[0036] Furthermore, the present invention also proposes a method for processing an aircraft trajectory data set, wherein the aircraft trajectory data set is used for model training for aircraft trajectory planning based on machine learning, and the specific steps of the method are as follows: 1) Using the segmented evaluation method described above, each flight trajectory data in the aircraft operation trajectory data set is evaluated segment by segment; 2) The trajectory point data corresponding to the evaluation indicators that meet the set requirements in each segment are selected to form a new data set as the training samples of the model for aircraft trajectory planning based on machine learning.

[0037] Based on the same technical solution, the present invention also proposes a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device executes the above-mentioned segmented evaluation method.

[0038] Based on the same technical solution, the present invention also proposes a computing device, comprising one or more processors, one or more memories and one or more programs, wherein the one or more programs are stored in the one or more memories and are configured to be executed by the one or more processors, and the one or more programs include instructions for executing the above-mentioned segmented evaluation method.

[0039] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0040] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0041] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0042] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0043] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A precise segmented evaluation method for aircraft trajectory based on machine learning, characterized in that: The steps include: (1) Obtain flight trajectory data; (2) Segment the flight trajectory; (3) Calculate the evaluation index of each segment and complete the segment evaluation of the flight trajectory.

2. The segmented evaluation method according to claim 1, characterized in that: According to the flight mission, the flight trajectory is divided into four sections: take-off section, cruise section, combat section, and landing section.

3. The segmented evaluation method according to claim 2, characterized in that: The calculation method of the evaluation index of the take-off phase is: S takeoff = k a • A takeoff + k θ •SD( θ takeoff ) in, S takeoff is the evaluation index of the take-off phase, A takeoff is the takeoff acceleration, SD( θ takeoff ) is the standard deviation of the pitch angle during takeoff, k a and k θ They are A takeoff and SD( θ takeoff )’s weight coefficient; The calculation formulas for the standard deviation of takeoff acceleration and takeoff pitch angle are: , , where ΔV is the velocity difference between the start and end trajectory points of the takeoff segment, Δt is the time difference between the start and end trajectory points of the takeoff segment, It is the take-off phase i The pitch angle of a track point, is the average pitch angle during takeoff, n 1 is the number of trajectory points for the takeoff segment.

4. The segmented evaluation method according to claim 2, characterized in that: The calculation method of the evaluation index of the cruise segment is: S cruise = k f •F + k h •SD( H ) in, S cruise is the evaluation index of the cruise segment, F is the energy efficiency, SD( H ) is the standard deviation of the flight altitude during the cruise phase, k f and k h They are F and SD( H )’s weight coefficient; The calculation formulas for energy efficiency and standard deviation of cruise altitude are: , , where F T is the total energy consumption in the cruise phase, F D is the cruising flight distance, It is the cruise section i The aircraft altitude at 2 track points, is the average flight altitude during the cruise phase, n 2 is the number of track points in the cruise segment.

5. The segmented evaluation method according to claim 2, characterized in that: The calculation method of the evaluation index of the combat phase is: S combat = k m • M + k r • R in, S combat is the evaluation index of the combat phase, M For mobility, R is the reaction speed, k m and k r They are M and the weight coefficient of R; The calculation formulas for mobility and reaction speed are: , ,in, It is the battle section i 3 The maximum acceleration or deceleration during the maneuver, It is the battle section i 3 The maneuvering frequency of the maneuver, k acc and k freq are respectively and The weight coefficient of n 3 is the total number of maneuvers during the battle phase, T react,j is the time of the jth reaction in the combat phase, m It is the total number of reactions in the combat phase.

6. The segmented evaluation method according to claim 2, characterized in that: The calculation method of the evaluation index of the landing phase is: S landing = k d • D + k v •SD(V landing ) in, S landing is the evaluation index of the landing phase, D is the landing accuracy, SD(V landing ) is the standard deviation of the falling velocity, k d and k v They are D and SD(V landing )’s weight coefficient; The calculation formulas for the standard deviation of landing accuracy and landing speed are: , ,in, x act , y act The actual landing points x , y coordinate, x pla , y pla The scheduled landing points x , y coordinate, It is the landing stage i Landing speed of 4 trajectory points, is the average descent velocity during the descent, n 4 is the number of trajectory points for the landing phase.

7. The device for accurately segmenting and evaluating the trajectory of an aircraft for machine learning is characterized by: include: A data acquisition unit, used for acquiring flight trajectory data; A segmentation unit, used for segmenting the flight trajectory; An index calculation unit, used to calculate the evaluation index of each segment; The evaluation unit is used to complete the segmented evaluation of the flight trajectory according to the evaluation index of each segment.

8. A method for processing an aircraft trajectory data set, wherein the aircraft trajectory data set is used for model training for aircraft trajectory planning based on machine learning, characterized in that: The specific steps of this method are as follows: 1) using the segmented evaluation method as described in any one of claims 1 to 6 to perform segmented evaluation on each flight trajectory data in the aircraft operation trajectory data set; 2) The trajectory point data corresponding to the evaluation indicators that meet the set requirements in each segment are selected to form a new data set as training samples for the model of aircraft trajectory planning based on machine learning.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the segmented evaluation method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the segmented evaluation method according to any one of claims 1 to 6 are implemented.

Images