Method for Generating Aircraft Motion Model Parameters of Radar Control Simulation Training System

Abstract

The present invention discloses a method for generating aircraft motion model parameters of a radar control simulation training system, and the steps are as follows: extracting historical flight trajectory data of flights; performing three-level data refinement classification on the historical flight trajectory data of flights; obtaining aircraft motion model parameters under corresponding altitude layers; obtaining the frequency statistics result of aircraft motion model parameter items of an altitude profile; obtaining a generation model of aircraft motion model parameters of an altitude profile; obtaining aircraft motion model parameters. By mining historical flight trajectories, the present invention improves the simulation degree of a civil aviation radar control simulation training system, thereby better helping radar controllers to conduct control skill training, improving the professional level, and ensuring the safety of civil aviation air traffic control.

Description

Technical Field

[0001] The present invention belongs to the field of civil aviation aircraft simulation systems, and particularly relates to a method for generating aircraft motion model parameters of a civil aviation radar control simulation training system. Background Art

[0002] The civil aviation radar control simulation training system is used for tasks such as training, on-the-job training, and professional assessment of air traffic radar controllers, aiming to quickly cultivate and improve the control skills of radar controllers. It is the main means and standard equipment for training approach and area controllers. The civil aviation radar control simulation training system plays an irreplaceable role in improving the control level and handling skills of controllers, helping controllers cope with the pressure of flight growth and a substantial increase in control load, thus ensuring the safety of civil aviation air traffic control.

[0003] To achieve the above goals, the civil aviation radar control simulation training system emphasizes maximizing the simulation degree.

[0004] In the existing civil aviation radar control simulation training system, the aircraft motion model parameters include the flight speed at a specified altitude output externally, the climb rate at a specified altitude, etc. In the current system, these parameters are obtained by calculating through the BADA model (base of Aircraft data) of the aircraft. The BADA model is an aircraft flight dynamics analysis model developed by the European Organization for Aviation Safety. This model is used to determine the performance parameters of the aircraft during the climb, cruise, and descent phases, as well as the fuel quantity estimation in each flight phase, and is mainly applied to four aspects: flight simulation, trajectory prediction, emission assessment, and fuel consumption calculation.

[0005] Since the BADA model is an ideal model, through the analysis of the real historical flight data of domestic flights, it can be found that there are many unreasonable aspects in the flight parameters provided by the BADA model.

[0006] In the era of big data, data mining technology has been more widely applied in the actual generation process.

[0007] In the existing published research literature, the trajectory analysis research based on data mining mainly focuses on the 4D flight trajectory prediction task. However, flight trajectory prediction and the flight simulation of the aircraft in the simulation system are two seemingly similar but different tasks. 4D flight trajectory prediction focuses on the accurate prediction of the next moment during a single flight. 4D flight trajectory prediction is a time series problem, and the state parameters of the previous few moments are tightly coupled with the state parameters of the next moment, and it is weakly correlated with other repeated trajectories in history. While the task of flight simulation training is to be able to approximate the distribution characteristics of the real trajectory through the sampling of each simulation exercise, so that the controllers can fully rehearse and train various situations that occur in real air traffic.

[0008] The deficiencies existing in the prior art are summarized as follows:

[0009] 1. For the calculation of the aircraft motion model parameters in the civil aviation radar control simulation training system based on the BADA model, it is an ideal model that does not consider the influence of different air routes on aircraft parameters, does not consider the influence of human factors on flight parameters during flight, and does not consider the randomness of flight parameters in the specific flight process.

[0010] 2. The existing 4D flight trajectory analysis work based on historical data mining is not applicable to the extraction of aircraft motion model parameters in the civil aviation radar control simulation training system. Summary of the Invention

[0011] Aiming at the deficiencies of the above prior art, the purpose of the present invention is to provide a method for generating aircraft motion model parameters of a radar control simulation training system, so as to solve the problem that the aircraft motion model in the prior art lacks considerations of factors such as heading, route, altitude, etc., and lacks considerations of the influence of flight randomness on model parameters, resulting in the deviation of the aircraft motion model parameter settings from the real flight trajectory.

[0012] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0013] A method for generating aircraft motion model parameters of a radar control simulation training system of the present invention comprises the following steps:

[0014] 1) Extract the historical flight trajectory data of the flight;

[0015] 2) Conduct three-level data refinement classification on the historical flight trajectory data of the flight;

[0016] 3) Take the third-level classification result in step 2) as the input data, extract the altitude profile data, and preprocess the altitude profile data to obtain the aircraft motion model parameters corresponding to the altitude layer, including speed, acceleration, climb rate / descent rate;

[0017] 4) Conduct frequency statistics on the speed, acceleration, climb rate, and descent rate of the altitude profile in step 3) to obtain the frequency statistics result of the aircraft motion model parameter items of the altitude profile;

[0018] 5) Fit the distribution function to the frequency statistics result of the aircraft motion model parameter items of the altitude profile in step 4) to obtain the generation model of the aircraft motion model parameters of the altitude profile;

[0019] 6) Input the current flight altitude of the aircraft into the generation model in step 5), calculate the speed, acceleration, climb rate / descent rate, and obtain the aircraft motion model parameters: [v h ,a h ,aup , a down .

[0020] Furthermore, step 1) specifically includes: parsing the track data from the civil aviation air traffic control automation system, saving the flight data of one flight as an independent file, and this file contains aircraft type, departure airport code, arrival airport code, flight altitude, flight speed, heading route, latitude and longitude coordinates, climb flag bit, and descent flag bit.

[0021] Furthermore, step 2) specifically includes:

[0022] First-level classification: classifying all the obtained historical flight track data according to the aircraft type;

[0023] Second-level classification: re-classifying the results of the first-level classification according to the flight route;

[0024] Third-level classification: re-classifying the results of the second-level classification according to the departure and arrival airports.

[0025] Furthermore, the altitude profile data in step 3) refers to the altitude from 1200 meters to 12000 meters, starting from 1200 meters, with each 300 meters as a height layer, and the corresponding aircraft motion model parameters.

[0026] Furthermore, the preprocessing in step 3) is specifically: calculating the derivative of the speed to obtain the acceleration; dividing the input data into climb segment data / descent segment data, and calculating the derivative of the climb segment data / descent segment data to obtain the climb rate / descent rate.

[0027] Furthermore, step 5) specifically includes: using 11 univariate distributions (norm distribution, lognorm distribution, t distribution, pareto distribution, expon distribution, dweibull distribution, genextreme distribution, gamma distribution, beta distribution, uniform distribution, loggamma distribution) according to the empirical distribution to fit the frequency statistics data of the aircraft motion model parameter items statistically obtained in step 4); scoring the 11 fitting distributions through the sum of squared residuals and hypothesis testing methods, and returning the best scoring distribution, that is, obtaining the sampling generation model p H＝i (v), the sampling generation model p H＝i (a), the sampling generation model p H＝i (a up ) and the sampling generation model p H＝i (a down ) of the descent rate, where v represents speed, a represents acceleration, a up represents the climb rate, a downrepresents the descent rate, and H is the height.

[0028] Further, step 6) specifically includes:

[0029] q(x) represents a truncated normal distribution;

[0030] q0 ∼ q(x) means sampling q(x) to obtain the sampling value q0;

[0031] is a constant;

[0032] p H＝i (v) represents the velocity fitting distribution function of the i-th altitude layer;

[0033] p H＝i (a) represents the acceleration fitting distribution function of the i-th altitude layer;

[0034] p H＝i (a up ) represents the climb rate fitting distribution function of the i-th altitude layer;

[0035] p H＝i (a down ) represents the descent rate fitting distribution function of the i-th altitude layer;

[0036] u0 ∼ Uniform[0,1] means sampling the uniform distribution on [0,1] to obtain the value u0;

[0037] If then accept the sampling q0 of q(x) and assign it to the velocity v h , otherwise reject this sampling and resample;

[0038] If then accept the sampling q0 of q(x) and assign it to the horizontal acceleration a h , otherwise reject this sampling and resample;

[0039] If then accept the sampling q0 of q(x) and assign it to the climb rate a up , otherwise reject this sampling and resample;

[0040] If then accept the sampling q0 of q(x) and assign it to the descent rate a down , otherwise reject this sampling and resample.

[0041] The beneficial effects of the present invention:

[0042] 1. The present invention improves the simulation degree of the civil aviation radar control simulation training system by mining historical flight trajectories, thereby better assisting radar controllers in conducting control skill training, improving their professional levels, and ensuring the safety of civil aviation traffic control.

[0043] 2. The present invention provides a statistical basis for setting aircraft parameters in simulated flight by calculating the statistical distribution characteristic values of real flight trajectory data.

[0044] 3. The present invention fully considers the differences between air routes and sets different parameters for simulated aircraft according to different air routes.

[0045] 4. The present invention fully considers the randomness during flight, provides a parameter generation model including data distribution for setting aircraft parameters in simulated flight, and gives a method to obtain the final output parameters through simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0047] For the convenience of those skilled in the art, the present invention will be further described below in conjunction with embodiments and drawings. The content mentioned in the embodiments does not limit the present invention.

[0048] Refer to Figure 1 As shown, a method for generating parameters of an aircraft motion model in a civil aviation radar control simulation training system of the present invention is as follows:

[0049] 1) Extract historical flight trajectory data of flights; specifically include: parsing track data from the civil aviation air traffic control automation system, and saving the flight data of one flight as an independent file, which contains aircraft type, departure airport code, arrival airport code, flight altitude, flight speed, heading route, latitude and longitude coordinates, climb flag bit, and descent flag bit.

[0050] 2) Conduct three-level data refinement classification for the historical flight trajectory data of flights; specifically include:

[0051] The first-level classification: Classify all the obtained historical flight track data according to aircraft type.

[0052] The second-level classification: Classify the results of the first-level classification again according to flight routes.

[0053] The third-level classification: Classify the results of the second-level classification again according to departure and arrival airports.

[0054] 3) Taking the third-level classification result in step 2) as input data, extracting altitude profile data, and preprocessing the altitude profile data to obtain the aircraft motion model parameters at the corresponding altitude layer, which include speed, acceleration, climb rate / descent rate;

[0055] The altitude profile data refers to the altitudes from 1,200 meters to 12,000 meters, with 1,200 meters as the starting point and every 300 meters as an altitude layer, and the corresponding aircraft motion model parameters.

[0056] The preprocessing is specifically as follows: calculating the derivative of the speed to obtain the acceleration; dividing the input data into climbing segment data / descending segment data, calculating the derivative of the climbing segment data / descending segment data, and obtaining the climbing rate / descending rate.

[0057] 4) Perform frequency statistics on the speed, acceleration, climb rate and descent rate of the altitude profile in step 3) to obtain frequency statistics results of the aircraft motion model parameter items of the altitude profile.

[0058] 5) fitting the distribution function of the frequency statistics of the aircraft motion model parameter items of the altitude profile in step 4) to obtain the generation model of the aircraft motion model parameters of the altitude profile; specifically comprising: fitting the frequency statistics of the aircraft motion model parameter items obtained in step 4) using 11 univariate distributions (norm distribution, lognorm distribution, t distribution, pareto distribution, expon distribution, dweibull distribution, genextreme distribution, gamma distribution, beta distribution, uniform distribution, loggamma distribution) according to the empirical distribution; scoring the 11 fitting distributions by the residual sum of squares and hypothesis testing method, returning the best score distribution, that is, obtaining the sampling generation model p of the velocity on the corresponding altitude layer profile H＝i (v) Acceleration sampling generation model p H＝i (a) Sampling generation model of climbing rate p H＝i (a up ) and the sampling generation model of the drop rate p H＝i (a down ), where v represents velocity, a represents acceleration, and a up represents the climb rate, a down represents the descent rate and H is the height.

[0059] 6) Input the current flight altitude of the aircraft into the parameter generation model of step 5), calculate the speed, acceleration, climb rate / descent rate, and obtain the aircraft motion model parameters: [v h ,a h ,a up ,a down ]; specifically including:

[0060] q(x) represents a truncated normal distribution;

[0061] q0 ∼ q(x) means sampling q(x) to obtain the sampled value q0;

[0062] is a constant;

[0063] p H＝i (v) represents the velocity fitting distribution function of the i-th altitude layer;

[0064] p H＝i (a) represents the acceleration fitting distribution function of the i-th altitude layer;

[0065] p H＝i (a up ) represents the climb rate fitting distribution function of the i-th altitude layer;

[0066] p H＝i (a down ) represents the descent rate fitting distribution function of the i-th altitude layer;

[0067] u0 ∼ Uniform[0,1] means sampling the uniform distribution on [0,1] to obtain the value u0;

[0068] If then accept the sampled value q0 of q(x) and assign it to the velocity v h , otherwise reject this sampling and resample;

[0069] If then accept the sampled value q0 of q(x) and assign it to the horizontal acceleration a h , otherwise reject this sampling and resample;

[0070] If then accept the sampled value q0 of q(x) and assign it to the climb rate a up , otherwise reject this sampling and resample;

[0071] If then accept the sampled value q0 of q(x) and assign it to the descent rate a down , otherwise reject this sampling and resample.

[0072] There are many specific application ways of the present invention. The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements can be made, and these improvements should also be regarded as the protection scope of the present invention.

Claims

1. A method for generating aircraft motion model parameters of a radar control simulation training system, characterized in that, The steps are as follows: 1) Extract the historical flight trajectory data of flights; 2) Conduct a three-level data refinement classification for the historical flight trajectory data of flights; 3) Using the third-level classification result in step 2) as input data, extract altitude profile data, and preprocess the altitude profile data to obtain the aircraft motion model parameters corresponding to each altitude layer, which include speed, acceleration, climb rate / descent rate; 4) Conduct a frequency statistics on the speed, acceleration, climb rate, and descent rate of the altitude profile in step 3) to obtain the frequency statistics result of the aircraft motion model parameter items of the altitude profile; 5) Conduct a distribution function fitting on the frequency statistics result of the aircraft motion model parameter items of the altitude profile in step 4) to obtain the generation model of the aircraft motion model parameters of the altitude profile; 6) Input the current flight altitude of the aircraft into the generation model in step 5), calculate the speed, acceleration, climb rate / descent rate, and obtain the aircraft motion model parameters; The specific content of step 2) includes: First-level classification: Classify all the obtained historical flight track data according to the aircraft model; Second-level classification: Classify the result of the first-level classification again according to the flight route; Third-level classification: Classify the result of the second-level classification again according to the departure and arrival airports.

2. The method for generating the aircraft motion model parameters of the radar control simulation training system according to claim 1, wherein, The specific content of step 1) includes: Parse the track data from the civil aviation air traffic control automation system, and save the flight data of one flight as an independent file, which contains the aircraft model, departure airport code, arrival airport code, flight altitude, flight speed, course route, longitude and latitude coordinates, climb flag bit, descent flag bit.

3. The method for generating the aircraft motion model parameters of the radar control simulation training system according to claim 1, characterized in that, The altitude profile data in step 3) refers to the altitude from 1,200 meters to 12,000 meters, with 1,200 meters as the starting point, and each 300 meters as an altitude layer, and the corresponding aircraft motion model parameters.

4. The method for generating aircraft motion model parameters of the radar control simulation training system according to claim 1, characterized in that, The specific preprocessing in step 3) is: Calculate the derivative of the speed to obtain the acceleration; Cut the input data into climb segment data / descent segment data, and calculate the derivative of the climb segment data / descent segment data to obtain the climb rate / descent rate.

5. The method for generating the aircraft motion model parameters of the radar control simulation training system according to claim 1, characterized in that, Step 5) specifically includes: fitting the frequency statistical data of the aircraft motion model parameter items obtained in step 4) with 11 univariate distributions according to the empirical distribution; scoring the 11 fitting distributions through the sum of squared residuals and hypothesis testing methods, and returning the best scoring distribution, that is, obtaining the sampling generation model p of the speed on the cross-section of the corresponding altitude layer H＝i (v), the sampling generation model p of the acceleration H＝i (a), the sampling generation model p of the climb rate H＝i (a up ) and the sampling generation model p of the descent rate H＝i (a down ), where v represents speed, a represents acceleration, a up represents the climb rate, a down represents the descent rate, and H is the altitude.

6. The method for generating the aircraft motion model parameters of the radar control simulation training system according to claim 5, wherein The 11 univariate distributions are norm distribution, lognorm distribution, t distribution, pareto distribution, expon distribution, dweibull distribution, genextreme distribution, gamma distribution, beta distribution, uniform distribution, loggamma distribution.

Images