Low-altitude navigation test flight condition forecasting method based on Bayesian network algorithm

Abstract

The invention discloses a low-altitude navigation test flight condition forecasting method based on a Bayesian network algorithm, and the method is characterized in that the method comprises the following steps: S1, collecting meteorological data related to low-altitude navigation; s2, preprocessing the meteorological data; s3, a Bayesian network model composed of a group of nodes and edges is constructed, the nodes represent meteorological elements, and the edges represent condition dependence relations among the meteorological elements; s4, parameter learning is carried out, and probabilities under different conditions are optimized and calculated; s5, performing Bayesian network reasoning, namely, calculating posterior probability distribution of other meteorological elements by inputting known meteorological data; s6, carrying out flight adaptability prediction and evaluation; and S7, outputting a prediction result. And S8, result feedback and optimization are carried out. According to the method, the prediction accuracy, the real-time performance and the automation level are improved, and the capability of dealing with complex meteorological conditions is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of low-altitude general aviation flight, and in particular to a method for predicting low-altitude general aviation test flight conditions based on a Bayesian network algorithm. Background Art

[0002] With the increasing complexity of low-altitude general aviation missions and the increasing demand for related mission planning, meteorological support for low-altitude general aviation faces increasingly severe challenges. Meteorological factors play a vital role in the stability, safety, and mission effectiveness of low-altitude flights. Traditional low-altitude general aviation flight test evaluation methods mainly rely on manual judgment or simple threshold judgment models for analysis. However, due to the large uncertainty of medium- and long-term vertical forecasts and the difficulty in directly evaluating the impact of low-altitude general aviation flight paths and array positions, existing methods have significant limitations in terms of real-time and accuracy.

[0003] The technical problems to be solved by the present invention are: (1) how to construct a scientific and reasonable automation model to meet the needs of low-altitude general aviation mission planning and other automated command and management systems for low-altitude general aviation flight fitness assessment products; (2) how to design an accurate and explainable intelligent prediction model; (3) how to improve the parameterization level of the model and optimize the performance of the model in sparse sample prediction in response to the training problem of the intelligent model in the case of sparse samples. Summary of the Invention

[0004] In order to solve the existing problems, the present invention provides a method for predicting low-altitude general aviation test flight conditions based on a Bayesian network algorithm. The specific scheme is as follows:

[0005] A method for predicting low-altitude general aviation test flight conditions based on a Bayesian network algorithm comprises the following steps:

[0006] S1, collects meteorological data related to low-altitude aviation;

[0007] S2, preprocessing of meteorological data;

[0008] S3, constructing a Bayesian network model consisting of a set of nodes and edges, where the nodes represent meteorological elements and the edges represent the conditional dependencies between meteorological elements;

[0009] S4, performs parameter learning and optimizes the probability under different conditions;

[0010] S5, perform Bayesian network reasoning, that is, calculate the posterior probability distribution of other meteorological elements by inputting known meteorological data;

[0011] S6, conduct flight suitability prediction and assessment;

[0012] S7, output prediction results;

[0013] S8, provide feedback and optimize the results.

[0014] Preferably, the meteorological data in step S1 includes real-time meteorological forecast data and historical meteorological data. The real-time meteorological forecast data is provided by meteorological satellites, meteorological station observations, and numerical forecast systems, and covers key meteorological elements such as wind speed, visibility, temperature, cloud cover, and ice accumulation.

[0015] Preferably, the data preprocessing in step S2 specifically includes the following steps:

[0016] S21, data cleaning, removal of abnormal data; use statistical methods to detect abnormal data: for each meteorological element If its value exceeds the specified upper and lower limits, it is considered abnormal data;

[0017] S22, data standardization, use mean-standard deviation method Z-Score standardization to perform data standardization processing, the formula is: ;in, is the original data, and Meteorological elements The mean and standard deviation of It is the standardized data;

[0018] S23, use linear interpolation or KNN interpolation method to fill missing data, for meteorological elements A missing value , and its interpolation calculation method is: .

[0019] Preferably, the conditional probability of each node in step S4 is represented by a conditional probability table CPT, which is the probability distribution of each meteorological element under different conditions in the Bayesian network; the conditional probability of the node is calculated by training the historical meteorological data. , that is, under the condition of a given parent node - meteorological element, meteorological element The probability of occurrence; optimize the calculation of conditional probability through maximum likelihood estimation or Bayesian estimation method.

[0020] Preferably, the core of the reasoning in step S5 is to use Bayes' theorem to calculate the joint probability distribution and conditional probability;

[0021] For meteorological feature collections , the joint probability distribution of the Bayesian network can be expressed as:

[0022]

[0023] Preferably, the probability distribution of each meteorological element is obtained according to Bayesian network reasoning , using a weighted sum method to calculate the flyability score ,

[0024]

[0025] in, It is a meteorological element The weight of It is a meteorological element The influence function of , which indicates the influence of this factor on the flyability;

[0026] The airworthiness is judged by scoring Compare with the preset threshold to determine whether it is suitable for flying.

[0027] Preferably, step S7 specifically includes: generating an airworthiness assessment report through the reasoning results of the Bayesian network, the report including a conclusion on whether the aircraft is airworthy, and listing in detail the predicted values ​​of various meteorological elements, the corresponding credibility, and the impact of these meteorological elements on the airworthiness assessment; the prediction results are output in the form of a report for the flight mission command center and meteorological support personnel to make decisions, helping them to evaluate the feasibility and safety of the flight mission.

[0028] Preferably, step S8 specifically includes: After the flight mission is executed, collecting weather data and flyability assessment results encountered during the actual flight. By comparing the predicted results with the actual situation, the model's accuracy and prediction bias are analyzed. Based on this feedback, the structure and parameters of the Bayesian network model are adjusted and optimized to improve the model's predictive capabilities under complex weather conditions.

[0029] A computer-readable storage medium stores a computer program, and when the computer program is run, the method described in any one of the above items is executed.

[0030] A computer system includes a processor and a storage medium, wherein a computer program is stored on the storage medium. The processor reads and runs the computer program from the storage medium to execute any of the methods described above.

[0031] The beneficial effects of the present invention are:

[0032] (1) Improved prediction accuracy: This invention uses a Bayesian network to model the complex dependencies between meteorological elements, comprehensively considering the influence of multiple meteorological factors, thereby providing more accurate flight suitability predictions. Compared with traditional threshold-based judgment methods, Bayesian networks can avoid misjudgments and improve prediction accuracy, especially under complex and dynamically changing meteorological conditions.

[0033] (2) Improved real-time performance and automation: Bayesian networks can automatically process real-time meteorological data and conduct flyability assessments, avoiding the inefficiency of traditional manual assessment methods. By updating prediction results in real time, this invention can provide timely and accurate meteorological support for low-altitude general aviation missions, significantly improving the efficiency of mission planning.

[0034] (3) Improved model interpretability: The present invention uses a Bayesian network to make the airworthiness prediction process more transparent. The impact of each meteorological factor on the airworthiness prediction can be clearly explained, helping flight crews understand and trust the prediction results, and increasing their reliance on the prediction model.

[0035] (4) Enhanced ability to cope with complex meteorological conditions: Bayesian networks can handle the conditional dependencies between complex meteorological elements, so that when faced with sudden meteorological events or extreme weather conditions, the system can make rapid adjustments and provide more accurate airworthiness assessments.

[0036] (5) Optimizing the model’s continuous learning capability: The Bayesian network model can make adaptive adjustments based on feedback from actual flight missions, continuously optimizing prediction accuracy and thus improving the stability and accuracy of long-term predictions. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 Flowchart of the present invention;

[0039] Figure 2 This is the Bayesian network model diagram of the present invention. DETAILED DESCRIPTION

[0040] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0041] The existing low-altitude general aviation airworthiness prediction methods mainly include the following technical solutions:

[0042] 1. Manual Airworthiness Prediction Methods: Currently, most low-altitude general aviation meteorological support systems rely on manual analysis by meteorologists. These analyses typically assess an airport's airworthiness or the suitability of a specific aircraft type for low-altitude general aviation flights 12 or 24 hours in advance, based on a comprehensive assessment of past meteorological data, numerical forecast products, and on-site observations. Manual analysis can leverage the experience of meteorologists to provide relatively accurate assessments of complex meteorological conditions. However, this method has significant limitations. First, manual analysis relies on a large number of professional meteorologists, resulting in significant human and material costs. Second, manual assessments typically only provide short-term forecasts (primarily 12- to 24-hour airworthiness assessments) and lack effective coverage for mission forecasts over longer timeframes. Therefore, they cannot meet the long-term airworthiness assessment requirements of low-altitude general aviation missions. Furthermore, manual methods rely entirely on human experience, making them susceptible to human bias, resistant to automation and standardization, and susceptible to workload inconsistency.

[0043] Second, a threshold-based automated approach: This approach assesses whether an airport meets the appropriateness criteria by setting thresholds for key meteorological elements (such as wind speed, visibility, and air pressure) within numerical forecast products. If the value of a particular meteorological element does not meet the preset threshold, the airport is deemed unsuitable for low-altitude operations. This approach is simple and easy to implement, and can improve forecasting efficiency. However, threshold-based judgments have significant limitations. First, there is often a certain degree of deviation between the forecast product and the actual observations, especially in conditions of rapid weather changes or incomplete data, which can lead to inaccurate judgments. Second, threshold-based methods are overly simplistic and fail to account for the complex relationships between meteorological elements and the coupling effects of multiple factors. This makes it difficult to effectively capture complex weather conditions, particularly those between suitability and unsuitability for low-altitude operations. Threshold-based methods struggle to accurately identify and address certain edge cases, resulting in low forecast accuracy. In particular, when suitability is difficult to determine, threshold-based methods often fail to accurately identify and address complex weather conditions.

[0044] The following is the technical solution process of the automated method based on meteorological data threshold judgment:

[0045] (1) Data collection

[0046] Use the output data of aircraft report data and T1279 data to obtain key meteorological elements (such as wind speed, visibility, air pressure, temperature, etc.).

[0047] (2) Data preprocessing

[0048] 1) Data cleaning: remove missing data and outliers.

[0049] 2) Data standardization: unify the formats and units of different data sources.

[0050] 3) Data interpolation: interpolation to fill in missing data or data with irregular time intervals.

[0051] (3) Setting threshold conditions

[0052] 1) Wind Speed ​​Threshold: A wind speed threshold is set, for example, exceeding which a certain wind speed is deemed unsuitable for low-altitude operations. The upper wind speed limit for low-altitude operations is set at no more than 15 m / s (or the equivalent of 34 mph). This threshold is based on an analysis of the aircraft's maneuverability and stability in strong winds and is intended to ensure flight safety in a wide range of wind speeds.

[0053] 2) Visibility Threshold: A visibility threshold is set below which low-altitude air navigation is deemed unsuitable. To ensure clear vision for pilots and accurate flight paths, visibility is set at no less than 2 kilometers (or equivalently 1.24 miles) as the minimum visibility limit for low-altitude air navigation. In extreme or adverse weather conditions, this visibility requirement may be increased as needed to ensure adequate flight safety margins.

[0054] 3) Thresholds for other meteorological factors (such as temperature and air pressure) are set based on specific application requirements. A cloud base height of no less than 200 meters is set as the lower limit for low-altitude air navigation. This threshold takes into account the aircraft's safety margin during low-altitude flight and the potential impact of clouds on the flight path. Furthermore, this threshold can be flexibly adjusted based on the geographic characteristics of a specific region and aircraft performance to meet safety requirements in different scenarios.

[0055] (4) Threshold judgment

[0056] Each meteorological element is compared to see if it meets the set threshold. If any meteorological element does not meet the set threshold, the current conditions are considered unsuitable for low-altitude air traffic. If all meteorological elements meet the set threshold, the current conditions are considered suitable for low-altitude air traffic.

[0057] (5) Automated airworthiness assessment

[0058] 1) Automated processing: Automatically updates the airworthiness assessment based on real-time changes in meteorological data and threshold judgment results.

[0059] 2) Airworthiness Forecast Output: Generates an airworthiness forecast report, including the airworthiness assessment results and detailed information on various meteorological factors. If necessary, short-term or medium-term airworthiness forecasts are generated to assist in mission planning.

[0060] (6) Result output

[0061] The airworthiness assessment results are output through visualization tools, such as charts, reports, or warning messages, and provided to the flight control center and meteorological support personnel for further analysis and decision-making.

[0062] The shortcomings of the aforementioned threshold-based automated approach include: 1. Simplified judgment logic: This approach relies solely on comparing a single meteorological factor against a set threshold, ignoring the complex interrelationships between meteorological factors and the coupled effects of multiple factors. Therefore, it cannot fully and accurately reflect the suitability of low-altitude general aviation, resulting in a lack of precision in the prediction results. 2. Threshold setting issues: Thresholds are often based on historical experience or standards and may not accurately match the actual needs under varying meteorological conditions. Due to the complex and rapid changes in meteorological conditions, fixed thresholds may not adapt to real-time changes, leading to misjudgments. 3. Inability to cope with complex meteorological conditions: Low-altitude general aviation suitability involves the combined influence of multiple meteorological factors. A single threshold cannot capture the complex interactions of these factors, and is particularly prone to misjudgments in highly variable meteorological environments. 4. Inaccurate predictions: Due to the oversimplification of the threshold model, it often fails to recognize subtle changes in meteorological conditions, especially at the critical boundary between suitability and unsuitability for low-altitude flight. This can easily lead to misjudgments or omissions, resulting in low prediction accuracy. 5. Lack of dynamic adjustment capability: This method has poor response capabilities to sudden weather changes and cannot dynamically adjust threshold conditions according to real-time meteorological changes, resulting in the inability to make accurate and timely judgments on flight suitability.

[0063] III. Airworthiness Prediction Models Based on Artificial Intelligence Algorithms: With the rapid development of artificial intelligence technology, airworthiness prediction models based on machine learning and deep learning are gradually entering the field of low-altitude general aviation meteorological support. These models automatically identify the potential relationships between meteorological factors and airworthiness by analyzing historical meteorological data, flight mission data, and real-time observation data, and then make predictions through model training. Artificial intelligence algorithms have strong self-learning capabilities and can continuously optimize predictions using large amounts of data, theoretically providing high accuracy. However, in practical applications, these intelligent models still face several challenges. For example, the quality and quantity of training data directly affect the model's prediction performance. If historical data is incomplete or has significant deviations, the model training effect will be significantly reduced. Furthermore, the model training process typically requires a large amount of computing resources and time, and when dealing with sudden meteorological events or extreme weather conditions, the model's responsiveness is weak and may not be able to accurately respond to the impact of sudden weather events. More importantly, because the meteorological conditions for low-altitude general aviation involve multiple complex factors, the prediction accuracy of existing intelligent models still has significant room for improvement, especially when dealing with sparse samples and atypical weather conditions, where accuracy is low.

[0064] The present invention discloses a low-altitude general aviation flight suitability condition forecasting technology based on a Bayesian network algorithm. The technology utilizes the Bayesian network to model the dependency relationship between meteorological elements, and provides a more accurate low-altitude general aviation flight suitability prediction through inference calculation and real-time data input, thereby improving the safety and feasibility of flight missions.

[0065] like Figure 1 As shown, a method for predicting low-altitude general aviation test flight conditions based on a Bayesian network algorithm includes the following steps:

[0066] S1. Collect meteorological data related to low-altitude air traffic. Collect meteorological data related to low-altitude air traffic from multiple meteorological data sources.

[0067] This data includes both real-time forecasts and historical meteorological data. The real-time forecasts are provided by meteorological satellites, weather station observations, numerical forecast systems, and other systems. They cover key meteorological elements such as wind speed, visibility, temperature, cloud cover, and ice accumulation. By combining historical and real-time data, they provide a training dataset for the Bayesian network model. The collected data contains complex relationships between different meteorological elements, and by integrating data from different sources, we ensure data integrity and timeliness.

[0068] S2. Preprocess the meteorological data. The collected raw meteorological data needs to be cleaned and standardized. Data cleaning involves removing missing or outlier observations to ensure data accuracy and consistency. Data standardization unifies the format and units of data from different sources, eliminating incompatibilities caused by format differences. Furthermore, interpolation methods are used to fill in missing or irregular data to ensure data integrity and continuity, thereby providing a high-quality dataset for the Bayesian network model input.

[0069] Data preprocessing specifically includes the following steps:

[0070] S21, data cleaning, remove abnormal data, ensure data quality. Use statistical methods to detect abnormal data: For each meteorological element , if its value exceeds the specified upper and lower limits, for example, wind speed If the data is greater than 100m / s or less than -10m / s, it is considered abnormal data.

[0071] S22, Data Standardization - Since data for different meteorological elements may come from different sensors or sources with different units and scales, they need to be standardized. The goal of standardization is to unify the dimensions of all meteorological elements so that they can be processed in the same model.

[0072] Data standardization is performed using the mean-standard deviation method Z-Score standardization, the formula is: ;in, is the original data, and Meteorological elements The mean and standard deviation of It is the standardized data;

[0073] S23, use linear interpolation or KNN interpolation to fill missing data, for meteorological elements A missing value , and its interpolation calculation method is: .

[0074] S3, by analyzing historical meteorological data and combining expert knowledge, determine the dependencies between various meteorological elements. Construct a Bayesian network model consisting of a set of nodes and edges, where nodes represent meteorological elements and edges represent conditional dependencies between meteorological elements. Based on these dependencies, construct a Bayesian network structure, such as Figure 2 shown.

[0075] S4, performs parameter learning and optimizes the probability calculation under different conditions.

[0076] The conditional probability of each node is represented by the conditional probability table CPT, which is the probability distribution of each meteorological element under different conditions in the Bayesian network. The conditional probability of the node is calculated by training the historical meteorological data. , that is, under the condition of a given parent node - meteorological element, meteorological element Probability of occurrence. Assuming there are meteorological factors (wind speed) and (temperature) has a dependency, then the conditional probability This represents the probability distribution of wind speed given a known temperature. The conditional probability is calculated using maximum likelihood estimation or Bayesian estimation.

[0077] S5 performs Bayesian network inference. Bayesian network inference involves inputting known meteorological data such as wind speed, temperature, and cloud cover to calculate the posterior probability distribution of other meteorological factors. The core of the inference is to use Bayes' theorem to calculate joint probability distributions and conditional probabilities.

[0078] For meteorological feature collections , the joint probability distribution of the Bayesian network can be expressed as:

[0079]

[0080] During the inference process, the Bayesian network will update the conditional probability of each meteorological factor based on real-time data, and comprehensively consider the impact of each meteorological factor on flight suitability to conduct a flyability assessment.

[0081] S6, conduct airworthiness prediction and assessment.

[0082] Once the Bayesian network model is complete, real-time meteorological data can be input and used for inference calculations. Based on the input data, the model infers the probability distribution of each meteorological factor and, based on these probabilistic results, determines whether the aircraft meets the low-altitude airworthiness standards. Bayesian networks can handle the dependencies between multiple meteorological factors, comprehensively assessing their impact on flight safety, and thus providing more accurate airworthiness predictions.

[0083] Probability distribution of various meteorological elements obtained by Bayesian network reasoning , the flyability score can be calculated using a weighted sum method .

[0084]

[0085] in, It is a meteorological element The weight of It is a meteorological element The influence function of represents the influence of this factor on the flyability.

[0086] The flyability will be judged by scoring Compare with preset thresholds to determine if it is suitable for flight:

[0087]

[0088] S7: Output prediction results. Step S7 specifically involves generating an airworthiness assessment report based on the Bayesian network's inference results. The report includes a conclusion on whether the aircraft is airworthy and details the predicted values ​​of various meteorological factors, their corresponding reliability, and their impact on the airworthiness assessment, as shown in Table 1. The prediction results are output as a report for decision-making by the mission control center and meteorological support personnel, helping them assess the feasibility and safety of the mission.

[0089] surface Forecast results of meteorological elements Meteorological elements Predicted value Confidence (range: 0-100%) Impact on flyability Weather conditions 1 85% No severe dangerous weather wind speed 1 90% Wind speed does not exceed 12m / s visibility 0.95 70% Does not affect visual navigation safety

[0090] S8, feedback and optimization of the results. Specifically:

[0091] After the mission is completed, weather data and flyability assessment results from the actual flight are collected. By comparing the predicted results with the actual conditions, the model's accuracy and prediction bias are analyzed. Based on this feedback, the Bayesian network model's structure and parameters are adjusted and optimized to improve its predictive capabilities under complex weather conditions. Through this continuous optimization process, the Bayesian network model will continuously improve its accuracy in flyability predictions and its ability to respond to sudden weather events. Table 2 shows the weather data and flyability assessment results.

[0092] surface Meteorological data and airworthiness assessment results Meteorological elements Accuracy Accuracy Recall F1 score AUC Weather conditions 0.90 0.88 0.92 0.90 0.85 wind speed 0.85 0.82 0.88 0.85 0.80 visibility 0.82 0.80 0.84 0.82 0.78 Airworthiness prediction 0.93 0.92 0.94 0.93 0.90

[0093] The advantages of the low-altitude general aviation flyability prediction method based on Bayesian network adopted in the present invention are mainly reflected in:

[0094] (1) Improved prediction accuracy: This invention uses a Bayesian network to model the complex dependencies between meteorological elements, comprehensively considering the influence of multiple meteorological factors, thereby providing more accurate flight suitability predictions. Compared with traditional threshold-based judgment methods, Bayesian networks can avoid misjudgments and improve prediction accuracy, especially under complex and dynamically changing meteorological conditions.

[0095] (2) Improved real-time performance and automation: Bayesian networks can automatically process real-time meteorological data and conduct flyability assessments, avoiding the inefficiency of traditional manual assessment methods. By updating prediction results in real time, this invention can provide timely and accurate meteorological support for low-altitude general aviation missions, significantly improving the efficiency of mission planning.

[0096] (3) Improved model interpretability: The present invention uses a Bayesian network to make the airworthiness prediction process more transparent. The impact of each meteorological factor on the airworthiness prediction can be clearly explained, helping flight crews understand and trust the prediction results, and increasing their reliance on the prediction model.

[0097] (4) Enhanced ability to cope with complex meteorological conditions: Bayesian networks can handle the conditional dependencies between complex meteorological elements, so that when faced with sudden meteorological events or extreme weather conditions, the system can make rapid adjustments and provide more accurate airworthiness assessments.

[0098] (5) Optimizing the model’s continuous learning capability: The Bayesian network model can make adaptive adjustments based on feedback from actual flight missions, continuously optimizing prediction accuracy and thus improving the stability and accuracy of long-term predictions.

[0099] The present invention also discloses a computer-readable storage medium and a computer system, wherein the computer storage medium stores a computer program, and when the computer program is executed, it performs any of the methods described above. A computer system includes a processor and a storage medium, wherein the storage medium stores the computer program, and the processor reads and executes the computer program from the storage medium to perform any of the methods described above.

[0100] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithmic steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. A skilled person may implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention.

[0101] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor so that the processor can read and write information from / to the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and storage medium may reside in a user terminal as discrete components.

[0102] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0103] Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein; 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 method for predicting low-altitude general aviation test flight conditions based on a Bayesian network algorithm, characterized in that: The following steps are involved: S1, collects meteorological data related to low-altitude aviation; S2, preprocessing of meteorological data; S3, constructing a Bayesian network model consisting of a set of nodes and edges, where the nodes represent meteorological elements and the edges represent the conditional dependencies between meteorological elements; S4, performs parameter learning and optimizes the probability under different conditions; S5, perform Bayesian network reasoning, that is, calculate the posterior probability distribution of other meteorological elements by inputting known meteorological data; S6, conduct flight suitability prediction and assessment; specifically, the various meteorological elements obtained by Bayesian network reasoning The probability distribution of , using a weighted sum method to calculate the flyability score , ； in, It is a meteorological element The weight of It is a meteorological element The influence function of , which indicates the influence of this factor on the flyability; The flyability will be judged by scoring Compare with preset thresholds to determine whether it is suitable for flight; S7, output prediction results; S8, provide feedback and optimize the results.

2. The method according to claim 1, wherein: The meteorological data in step S1 includes real-time meteorological forecast data and historical meteorological data. The real-time meteorological forecast data is provided by meteorological satellites, meteorological station observations, and numerical forecast systems, and covers key meteorological elements such as wind speed, visibility, temperature, cloud cover, and ice accumulation.

3. The method according to claim 1, wherein: The data preprocessing in step S2 specifically includes the following steps: S21, data cleaning, removal of abnormal data; use statistical methods to detect abnormal data: for each meteorological element If its value exceeds the specified upper and lower limits, it is considered abnormal data; S22, data standardization, use mean-standard deviation method Z-Score standardization to perform data standardization processing, the formula is: ;in, is the original data, and Meteorological elements The mean and standard deviation of It is the standardized data; S23, use linear interpolation or KNN interpolation to fill missing data, for meteorological elements A missing value , and its interpolation calculation method is: , where j is the spatial location index, representing different observation stations or geographic grid points.

4. The method according to claim 1, wherein: The conditional probability of each node in step S4 is represented by the conditional probability table CPT, which is the probability distribution of each meteorological element under different conditions in the Bayesian network; the conditional probability of the node is calculated by training the historical meteorological data. , that is, under the condition of a given parent node - meteorological element, meteorological element The probability of occurrence is obtained by optimizing the calculation of conditional probability through maximum likelihood estimation or Bayesian estimation method.

5. The method according to claim 1, wherein: The core of the reasoning in step S5 is to use Bayes’ theorem to calculate the joint probability distribution and conditional probability; For meteorological feature collections , the joint probability distribution of the Bayesian network is expressed as: 。 6. The method according to claim 1, characterized in that Step S7 specifically includes: generating an airworthiness assessment report based on the inference results of the Bayesian network, the report including a conclusion on whether the aircraft is airworthy, and listing in detail the predicted values ​​of various meteorological elements, the corresponding credibility, and the impact of these meteorological elements on the airworthiness assessment; the prediction results are output in the form of a report for the flight mission command center and meteorological support personnel to make decisions, helping them to evaluate the feasibility and safety of the flight mission.

7. The method according to claim 1, characterized in that Step S8 specifically involves collecting weather data and flight suitability assessment results from the actual flight after the mission is completed. By comparing the predicted results with the actual situation, the model's accuracy and prediction bias are analyzed. Based on this feedback, the Bayesian network model's structure and parameters are adjusted and optimized to improve the model's predictive capabilities under complex weather conditions.

8. A computer-readable storage medium, characterized in that: The medium stores a computer program, and after the computer program is run, the method according to any one of claims 1 to 7 is executed.

9. A computer system, characterized in that: The method comprises a processor and a storage medium, wherein the storage medium stores a computer program, and the processor reads and runs the computer program from the storage medium to execute the method according to any one of claims 1 to 7.

Images