Application system and application method of flight simulation training data packet

Abstract

The invention provides an application system and an application method of a flight simulation training data packet, and belongs to the technical field of aviation simulation training. The system comprises a data acquisition module, an engineering data packet processing module, a training data packet generation module, an error verification module and an adaptive output module. According to the method, an engineering data packet is constructed by arranging aircraft design data and wind tunnel data, a training data packet meeting the standard of a CAAC full-motion D-level simulator is formed after flight experiment correction, and the error is controlled within 5%. The system realizes localization of flight simulation training data packets, breaks through foreign technical monopoly, guarantees national defense safety, improves training fidelity and economical efficiency, and can be widely applied to the field of civil aviation and military aircraft simulation training.

Description

Technical Field

[0001] This invention relates to the field of aviation simulation training technology, and in particular to an application system and method for flight simulation training data packages. Background Technology

[0002] As my country's first domestically developed large passenger aircraft, the C919 relies on a full-motion Level D simulator for flight simulation training. Currently, this simulator and its core aerodynamic data packages are monopolized by the Canadian CAE company. This poses a risk of leakage to critical performance data and pilot training data of the C919, threatening national defense security. Furthermore, the high cost of imported simulators and reliance on foreign technology severely restricts the independent development of China's domestic large aircraft industry.

[0003] In the existing technology, the development of domestic simulation training data packages has the following shortcomings:

[0004] The core technologies rely on imports, lack independent intellectual property rights, and cannot meet national defense and security needs;

[0005] The data packets have a large error compared to the real machine data, resulting in insufficient simulation realism and affecting the training effect;

[0006] Its poor adaptability and incompatibility with multiple simulator models limit the promotion and application of the technology.

[0007] Therefore, developing a domestically produced, high-precision, and highly adaptable C919 flight simulation training data package application system is of great practical significance. Summary of the Invention

[0008] The embodiments of the present invention provide an application system and method for flight simulation training data packages. The system enables the localization of flight simulation training data packages, improves simulation accuracy and adaptability, and safeguards national defense security.

[0009] To achieve the above objectives, the present invention adopts the following technical solution.

[0010] An application system for flight simulation training data packages includes:

[0011] The multi-source data acquisition module is used to: collect real aircraft flight test data, pilot simulation training data, and aircraft design data; and to preprocess the collected real aircraft flight test data, pilot simulation training data, and aircraft design data.

[0012] The engineering data package processing module is used to: construct engineering simulator data packages based on time-domain data obtained by the multi-source data acquisition module; the engineering simulator data package contains the aircraft's aerodynamic model, flight control model, engine model, and ground response model;

[0013] The training data package generation module is used to: correct the engineering simulator data package using the aircraft's flight test data; optimize the aerodynamic model parameters in the engineering simulator data package using the frequency domain response identification method; generate flight simulation training data packages; and the flight simulation training data packages are training data packages containing dynamic models, functional logic data, and QTG data.

[0014] The dynamic optimization module is used to: collect emergency response operation data of the flight simulator in real time, update the emergency response parameters of the flight simulation training data package through reinforcement learning algorithm; update the test flight data of the aircraft in real time, and iteratively optimize the flight simulation training data package through sliding window algorithm;

[0015] The error verification module is used to calculate the deviation between the flight simulation training data package and the actual data using the root mean square error method. If the deviation is greater than a preset threshold, the dynamic optimization module is triggered to correct the flight simulation training data package.

[0016] The adapter output module is used to output the flight simulation training data package optimized by the dynamic optimization module to the flight simulator.

[0017] Preferably, the multi-source data acquisition module includes:

[0018] High-precision sensors are used to synchronously collect flight test data of the aircraft;

[0019] The cockpit operation recorder is used to synchronously collect pilot simulation training data for the aircraft.

[0020] Aircraft design data includes wind tunnel test data and aircraft structural data.

[0021] Preferably, the process by which the training data packet generation module optimizes the aerodynamic model parameters in the engineering simulator data packet using the frequency domain response identification method specifically includes:

[0022] The flight test data of the aircraft is subjected to frequency sweep processing to obtain channel response data;

[0023] By fitting the aerodynamic equation coefficients of the aircraft's flight test data using the least squares method, a three-dimensional aerodynamic model including angle of attack and Mach number is established.

[0024] The parameters of the three-dimensional aerodynamic model were optimized using the unscented Kalman filtering method.

[0025] The channel response data and the three-dimensional aerodynamic model are subjected to multi-input processing, combined window processing and state space identification operations to establish the MIMO system coefficient matrix and generate frequency domain response data packets.

[0026] Preferably, the working process of the adapter output module specifically includes: converting the dynamic parameters in the flight simulation training data package into control signals of the flight simulator hardware platform, so that the flight simulator's six-degree-of-freedom motion system, visual system and avionics system can achieve synchronous response.

[0027] Preferably, the flight simulation training data package includes a dynamic model, functional logic data, and QTG data; the QTG data includes objective test results comparing simulation and test flights, covering response parameters for scenarios such as takeoff and landing route flight, airspace flight, and emergency handling.

[0028] Secondly, the present invention provides a method for applying flight simulation training data packets, including:

[0029] Collect real aircraft flight test data, pilot simulation training data, and aircraft design data; preprocess the collected real aircraft flight test data, pilot simulation training data, and aircraft design data to obtain time domain data;

[0030] Based on the MIMO system identification method, the time-domain data obtained by the multi-source data acquisition module is processed to obtain flight simulation training data packets; the flight simulation training data packets have the frequency domain response characteristics of time-domain data;

[0031] Real-time acquisition of emergency handling operation data from flight simulators; updating emergency response parameters in flight simulation training data packages using reinforcement learning algorithms; real-time updating of aircraft test flight data; and iterative optimization of flight simulation training data packages using a sliding window algorithm.

[0032] The deviation between the flight simulation training data package and the actual flight data is calculated using the root mean square error method. If the deviation is greater than a preset threshold, the dynamic optimization module is triggered to correct the flight simulation training data package.

[0033] The optimized flight simulation training data package is output to the flight simulator.

[0034] As can be seen from the technical solutions provided by the embodiments of the present invention above, the present invention provides an application system and method for flight simulation training data packages, belonging to the field of aviation simulation training technology. The system includes a data acquisition module, an engineering data package processing module, a training data package generation module, an error verification module, and an adaptation output module. It constructs engineering data packages by organizing aircraft design data and wind tunnel data, and after correction through flight experiments, forms training data packages that meet the CAAC full-motion Class D simulator standard, with errors controlled within 5%. This system realizes the localization of flight simulation training data packages, breaks through foreign technological monopolies, ensures national defense security, improves training realism and economy, and can be widely applied in the fields of civil aviation and military aircraft simulation training.

[0035] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of the invention. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A logic block diagram of an application system for flight simulation training data packets provided by the present invention;

[0038] Figure 2 A flowchart illustrating the construction process of a flight simulation training data package in an application system provided by this invention;

[0039] Figure 3 This is a schematic diagram illustrating the process of constructing aerodynamic parameter frequency domain identification for a flight simulation training data package application system provided by the present invention. Detailed Implementation

[0040] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0041] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or couplings. The term “and / or” as used herein includes any and all combinations of one or more of the associated listed items.

[0042] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0043] To facilitate understanding of the embodiments of the present invention, the following will provide further explanation and description with reference to the accompanying drawings and several specific embodiments. These embodiments do not constitute a limitation on the embodiments of the present invention.

[0044] See Figure 1 The present invention provides an application system for flight simulation training data packets, including a multi-source data acquisition module 101, an engineering data packet processing module 102, a training data packet generation module 103, a dynamic optimization module 104, an error verification module 105, and an adaptation output module 106.

[0045] The multi-source data acquisition module 101 is used to acquire aircraft design data, wind tunnel test data and flight test data. It includes a sensor unit and a data preprocessing submodule, which can filter and reduce noise in the raw data to ensure data quality.

[0046] Based on the collected data, the engineering data package processing module 102 constructs an engineering simulator data package containing an aerodynamic model, a flight control model, an engine model, and a ground response model. The aerodynamic model employs Newton's laws of motion and multibody dynamics, fitting the functional relationship between lift coefficient, drag coefficient, and angle of attack using wind tunnel data.

[0047] The training data package generation module 103 corrects the engineering data package using flight test data, optimizes the dynamic model parameters using the frequency domain response identification method, and generates a training data package containing the dynamic model, functional logic data, and QTG data. The QTG data covers test results for scenarios such as takeoff and landing routes, airspace flight, and emergency handling, ensuring that the CAAC full-motion Class D simulator certification requirements are met.

[0048] The error verification module 105 verifies the error by comparing the output results of the training data package with those of the engineering data and flight test data, and uses the root mean square error calculation method to ensure that the error is controlled within 5%.

[0049] The adapter output module 106 uses a parameter mapping algorithm to adapt the training data package to the aircraft's full-motion D-level simulator or other types of simulators, achieving interface compatibility with the six-degree-of-freedom motion system, visual system, and avionics system.

[0050] In a preferred embodiment provided by the present invention, the multi-source data acquisition module 101 includes:

[0051] High-precision sensors are used to synchronously collect flight test data of the aircraft;

[0052] The cockpit operation recorder is used to synchronously collect pilot simulation training data for the aircraft.

[0053] Aircraft design data includes wind tunnel test data and aircraft structural data.

[0054] The training data packet generation module 103 performs the above-mentioned optimization using the frequency domain response identification method, specifically including the following process for optimizing the aerodynamic model parameters in the engineering simulator data packet:

[0055] The flight test data of the aircraft is subjected to frequency sweep processing to obtain channel response data;

[0056] By fitting the aerodynamic equation coefficients of the aircraft's flight test data using the least squares method, a three-dimensional aerodynamic model including angle of attack and Mach number is established.

[0057] The parameters of the three-dimensional aerodynamic model were optimized using the unscented Kalman filtering method.

[0058] The channel response data and the three-dimensional aerodynamic model are subjected to multi-input processing, combined window processing and state space identification operations to establish the MIMO system coefficient matrix and generate frequency domain response data packets.

[0059] The specific working process of the adapter output module 106 includes: converting the dynamic parameters in the flight simulation training data package into control signals of the flight simulator hardware platform, so that the flight simulator's six-degree-of-freedom motion system, visual system and avionics system can achieve synchronous response.

[0060] The flight simulation training data package includes dynamics models, functional logic data, and QTG data; the QTG data includes objective test results comparing simulation and test flights, covering response parameters for scenarios such as takeoff and landing routes, airspace flight, and emergency handling.

[0061] The present invention also provides an embodiment, taking a simulator of the C919 passenger aircraft as an example, to illustrate the implementation of the present invention.

[0062] This embodiment provides an application system for flight simulation training data packages, including:

[0063] Multi-source data acquisition module 101: includes an accelerometer, angular velocity sensor, and atmospheric data sensor installed on the C919 test aircraft to acquire parameters such as flight attitude, speed, and altitude; the built-in data preprocessing submodule uses wavelet transform algorithm to reduce noise in the raw data, and the sampling frequency is set to 1kHz;

[0064] Engineering data package processing module 102: Models are built based on the MATLAB / Simulink platform. The aerodynamic model input parameters are angle of attack (-15°~15°) and Mach number (0.2~0.8), and the output parameters are lift coefficient and drag coefficient. The flight control model uses a PID control algorithm to simulate the control surface response.

[0065] Training data packet generation module 103: Flight experiments were conducted at the 13,000-square-meter R&D base in the Daxing International Airport Comprehensive Bonded Zone. Response data was collected via frequency sweep input (frequency range 0.1~10Hz), and after UKF Kalman filtering, a dynamic model was established using a multiple-input multiple-output (MIMO) system identification method to generate training data packets (such as...). Figure 2 (As shown).

[0066] MIMO (Multiple-Input Multiple-Output) system frequency domain identification technology is the core technology for constructing high-precision aerodynamic models in this invention, specifically reflected in the entire process of the frequency domain identification modeling module, such as... Figure 3 As shown, the key steps are as follows:

[0067] Multi-channel data acquisition and conversion is designed for the flight characteristics of the C919. Time-domain data from six key channels are acquired simultaneously (inputs include pilot stick and rudder operations, throttle position, etc., and outputs include pitch angle, roll angle, yaw angle, airspeed, altitude, overload, etc.). The time-domain data is converted into frequency-domain response characteristics (frequency range 0.1~10Hz) through Fourier transform, and a frequency-domain mapping relationship between multiple inputs and multiple outputs is established.

[0068] Frequency sweeping experiments and model fitting were conducted using real aircraft flight test data to obtain aerodynamic parameter responses at different frequencies. Based on MIMO system theory, the least squares method was used to fit the coefficients of the three-dimensional aerodynamic equations, which include angle of attack and Mach number, to clarify the quantitative relationship between multiple input parameters (such as control surface deflection and engine thrust) and multiple output aerodynamic characteristics (such as lift coefficient and drag coefficient).

[0069] Parameter optimization and error convergence

[0070] The Unscented Kalman Filter (UKF) algorithm is introduced to optimize the initial model parameters. By iteratively adjusting the transfer function matrix of the MIMO system, the deviation between the model output and the flight test data is reduced from the initial 6% to within 5%, ultimately achieving high-precision matching between the aerodynamic model and the dynamic characteristics of the actual aircraft.

[0071] Error verification module 105: Select 100 sets of test flight data as the test set, calculate the root mean square error between the training data package output and the actual test data, and the result shows that the mean error is 3.2%, which meets the requirement of ≤5%.

[0072] The adapter output module 106 uses a parameter mapping interface written in C++ to convert parameters such as attitude angles and velocities in the training data package into displacement commands (range ±1.5m) for the six-degree-of-freedom platform and the field of view (120° horizontally and 80° vertically) for the visual system, enabling real-time linkage.

[0073] Secondly, the present invention provides a method for applying a flight simulation training data package, which includes the following steps:

[0074] Data acquisition: Flight test data of C919 at different altitudes (0~12000m) and speeds (200~900km / h) were collected by sensors, and preprocessed using Butterworth filters with a cutoff frequency of 50Hz.

[0075] Engineering data package construction: An aerodynamic model is established in MATLAB, and the lift coefficient CL=0.5×α²+0.1×α+0.3 (α is the angle of attack, in rad) is obtained by fitting; the flight control model adopts lateral yaw damping control, and the damping coefficient is set to 0.8.

[0076] Training data packet generation: Four typical regions within the flight envelope (low Mach number / low altitude, high Mach number / low altitude, low Mach number / high altitude, and high Mach number / high altitude) were selected for frequency sweep experiments. The transfer function was identified using the LMS algorithm, the dynamic model parameters were corrected, and training data packets were generated.

[0077] The training data packet is the core output of this invention, and its generation and optimization depend on the following algorithm:

[0078] Data preprocessing: Kalman filtering algorithm is used to remove noise (such as sensor drift error) from the flight test data to ensure that the data signal-to-noise ratio is >50dB;

[0079] Model construction: The aerodynamic equations are fitted by the least squares method identified by MIMO frequency domain, and the parameters are optimized by combining the UKF algorithm to generate the initial training data package;

[0080] Dynamic updates: Based on the reinforcement learning PPO strategy, combined with the sliding window algorithm (window size 500 flights), the data packet parameters are iteratively optimized monthly, incorporating new aircraft test flight data and pilot operation data;

[0081] Error verification: Perform 1000 Monte Carlo simulations on the training data package and calculate the error with the test flight data. The error is ≤4.8% within the 95% confidence interval, which meets the requirements.

[0082] The error verification module uses the root mean square error (RMSE) as its core algorithm. The specific process is as follows:

[0083] The RMSE between the training data packet output and the actual flight test data (such as 10 parameters including altitude, airspeed, pitch angle, etc.) is calculated using the formula: RMSE = n1∑i=1n(yi−y^i)2 (where yi is the actual flight data, y^i is the simulated data from the data packet, and n is the sample size).

[0084] If RMSE > 5%, the dynamic optimization module is triggered to readjust the parameters; if RMSE ≤ 5%, a verification report is output (including indicators such as altitude error ≤ 2m and airspeed error ≤ 3km / h), meeting the CAAC D-level certification standard.

[0085] Adaptive output: Load the training data package into the C919 full-motion D-level simulator, communicate with the six-degree-of-freedom platform through the OPC server interface, and control the platform response latency ≤10ms; transmit data with the visual system via UDP protocol, with the frame rate set to 60fps.

[0086] The encrypted transmission method used is based on existing mature technology (Chinese national cryptographic algorithm), but it has been specifically designed to suit the characteristics of flight data. The specific process is as follows:

[0087] It adopts the SM4 national cryptographic algorithm (block cipher algorithm, national commercial cryptography standard), which is an open and widely used existing encryption technology. Its security has been verified in the fields of finance and communications, and it is suitable for encrypted transmission of flight data.

[0088] Data classification encryption: Sensitive data (such as aerodynamic parameters, pilot operating characteristics, and emergency handling models) are classified and encrypted. Core aerodynamic data is encrypted using the SM4 algorithm with a key length of 128 bits.

[0089] Transmission link protection: In the data transmission from the multi-source data acquisition module to the modeling module and the dynamic optimization module, a dual mechanism of "encryption + verification" is adopted - first, ciphertext is generated by SM4 encryption, and then a hash value based on SHA-256 is added to ensure that the data transmission is not tampered with during the transmission process;

[0090] Key management: Dynamic distribution and regular updates of keys are achieved through a domestically developed key management system (compliant with the GM / T 0028-2014 standard) to avoid the risk of key leakage.

[0091] Model Update: 500 hours of pilot training data are collected monthly, and the random forest algorithm is used to optimize the emergency response model, improving the accuracy of engine failure simulation to 98%.

[0092] Real-time operation data feedback and parameter adjustment

[0093] Pilot operational data on the simulator (especially emergency response actions, such as stick and rudder operations during engine failure and stall recovery) is collected and used as "environmental feedback" for reinforcement learning. A proximal policy optimization (PPO) strategy is employed, using the "accuracy of emergency response simulation" as the reward function to dynamically update the model's emergency response parameters (such as control surface efficiency correction coefficients and thrust decay curves after engine failure), ensuring that the simulator's emergency performance is synchronized with the characteristics of real aircraft operation.

[0094] Iterative optimization of flight test data

[0095] Each month, new aircraft test flight data provided by COMAC (with a window size of 500 flights) is incorporated. Through a reinforcement learning "experience playback" mechanism, historical data is combined with new data to continuously optimize the parameters of the aerodynamic and flight control models. For example, for the BeiDou navigation integration parameters added to the C919 after 2024, reinforcement learning algorithms are used to quickly adjust the coupling relationship between the avionics system and the aerodynamic model to ensure the timeliness of the model.

[0096] In summary, this invention provides an application system and method for flight simulation training data packages, belonging to the field of aviation simulation training technology. The system includes a data acquisition module, an engineering data package processing module, a training data package generation module, an error verification module, and an adaptation output module. It constructs engineering data packages by organizing aircraft design data and wind tunnel data, and after flight experiment correction, forms training data packages that meet the CAAC full-motion Class D simulator standard, with errors controlled within 5%. This system achieves the localization of flight simulation training data packages, breaks through foreign technological monopolies, ensures national defense security, improves training realism and economy, and can be widely applied in the fields of civil aviation and military aircraft simulation training.

[0097] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0098] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0099] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for apparatus or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The apparatus and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0100] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An application system of flight simulation training data package, characterized in that, The method comprises the following steps: a multi-source data acquisition module is used to acquire real machine test flight data, pilot simulation training data and aircraft design data of an aircraft; the acquired real machine test flight data, pilot simulation training data and aircraft design data of the aircraft are preprocessed; an engineering data packet processing module is used to construct an engineering simulator data packet based on time domain data obtained by the multi-source data acquisition module; the engineering simulator data packet has an aerodynamic model, a flight control model, an engine model and a ground response model of the aircraft; a training data packet generation module is used to correct the engineering simulator data packet by using flight test data of the aircraft; a frequency domain response identification method is used to optimize aerodynamic model parameters in the engineering simulator data packet; a flight simulation training data packet is generated; the flight simulation training data packet has a training data packet of a kinetic model, functional logic data and QTG data; a dynamic optimization module is used to acquire special situation handling operation data of a flight simulator in real time, and update special situation response parameters of the flight simulation training data packet by using a reinforcement learning algorithm; real-time update of test flight data of the aircraft is performed, and the flight simulation training data packet is iteratively optimized by using a sliding window algorithm; an error checking module is used to calculate deviation of the flight simulation training data packet from real machine data by using a root mean square error method; if the deviation is greater than a preset threshold, the dynamic optimization module is triggered to correct the flight simulation training data packet; an adaptive output module is used to output the flight simulation training data packet optimized by the dynamic optimization module to a flight simulator.

2. The application system according to claim 1, characterized in that, The multi-source data acquisition module comprises: a high-precision sensor is used to synchronously acquire test flight data of the aircraft; a cockpit operation recorder is used to synchronously acquire pilot simulation training data of the aircraft; the aircraft design data comprises wind tunnel test data and aircraft structure data of the aircraft.

3. The application system according to claim 1, wherein, The process of using the frequency domain response identification method to optimize the aerodynamic model parameters in the engineering simulator data packet performed by the training data packet generation module specifically comprises: sweep frequency processing is performed on the test flight data of the aircraft to obtain channel response data; least square fitting is performed on aerodynamic equation coefficients of the test flight data of the aircraft to establish a three-dimensional aerodynamic model including an angle of attack and a Mach number; parameters of the three-dimensional aerodynamic model are optimized by using an unscented Kalman filter method; MIMO system coefficient matrixes are established by performing multi-input processing, combined window processing and state space identification operations on the channel response data and the three-dimensional aerodynamic model, and a frequency domain response data packet is generated.

4. The application system according to claim 1, wherein, The working process of the adaptive output module specifically comprises: converting the kinetic parameters in the flight simulation training data packet into control signals of a hardware platform of the flight simulator, so that a driving six-degree-of-freedom motion system, a visual system and an avionics system of the flight simulator realize synchronous response.

5. The application system according to claim 1, wherein, The flight simulation training data packet comprises a kinetic model, functional logic data and QTG data; the QTG data comprises objective test results of simulation simulation and test flight comparison, and response parameters of response parameters in scenes such as take-off and landing flight, airspace flight and special situation handling.

6. A method of applying flight simulation training data packets, characterized by, The method comprises the following steps: acquiring real machine test flight data, pilot simulation training data and aircraft design data of an aircraft; The collected real machine test flight data of the aircraft, pilot simulation training data and aircraft design data are preprocessed to obtain time domain data; Based on the MIMO system identification method, the time domain data obtained by the multi-source data acquisition module is processed to obtain a flight simulation training data package; the flight simulation training data package has the frequency domain response characteristics of the time domain data; Real-time collection of special situation handling operation data of the flight simulator, updating of special situation response parameters of the flight simulation training data package through a reinforcement learning algorithm; Real-time updating of the test flight data of the aircraft, and iterative optimization of the flight simulation training data package through a sliding window algorithm; The deviation between the flight simulation training data package and the real machine data is calculated by a root mean square error method, and if the deviation is greater than a preset threshold, the dynamic optimization module is triggered to correct the flight simulation training data package; The optimized flight simulation training data package is output to the flight simulator.

Images