Unmanned aerial vehicle high-precision simulation training method and system based on dynamic wind field modeling

Abstract

This invention discloses a high-precision simulation training method and system for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling, relating to the field of UAV simulation training. The method includes: activating a remote control acquisition engine to continuously acquire four-channel rod data at a preset frequency to establish raw control data; establishing a current control state flag based on the raw control data and sending it to a dynamic wind field controller to perform scene simulation initialization; executing two-dimensional drive control with the initialized dynamic wind field controller to establish a wind field dataset; inputting the data into a wind resistance calculation model and outputting wind resistance load data; after receiving the raw control data, wind field dataset, and wind resistance load data, the closed-loop responder performs data calculation through a UAV dynamics model, dynamically outputting the UAV's position and attitude for UAV simulation training management. This solves the technical problem of low simulation accuracy leading to poor training results in existing UAV simulation training, achieving the technical effect of improving simulation accuracy and enhancing training practicality.

Description

Technical Field

[0001] This application relates to the field of UAV simulation training, and in particular to a high-precision UAV simulation training method and system based on dynamic wind field modeling. Background Technology

[0002] In the field of UAV simulation training, achieving high-precision simulation to improve training effectiveness and ensure actual flight safety is crucial. This relates to the development of UAV operator skills and the stable operation of UAVs in complex environments. Currently, the main methods for solving the accuracy problems of environmental simulation and dynamic response in UAV simulation training are to use relatively basic static wind field simulation and simplified dynamic models. Static wind field simulation usually presets fixed wind speed and direction, and the dynamic model only performs rough calculations on some key motion parameters of the UAV. However, current methods cannot realistically reflect the dynamically changing wind field environment during actual flight, such as gusts and wind shear. At the same time, simplified dynamic models cannot accurately simulate the six-degree-of-freedom motion state of UAVs under complex wind fields and control commands. This leads to a large deviation between simulation training and actual flight conditions, significantly reducing training effectiveness and making it difficult to effectively improve operators' ability to cope with complex environments.

[0003] Currently, drone simulation training suffers from low simulation accuracy, leading to poor training results. Summary of the Invention

[0004] This application provides a high-precision simulation training method and system for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling. It employs a remote control acquisition engine to collect four-channel stick data at a preset frequency, generating raw control data which is then input into a closed-loop responder. Based on this raw control data, a current control state marker is established and sent to a dynamic wind field controller, completing the scene simulation initialization. The initialized dynamic wind field controller performs dual-dimensional drive control to generate a wind field dataset. This dataset is then input into a wind resistance calculation model to obtain wind resistance load data. The closed-loop responder receives the raw control data, the wind field dataset, and the wind resistance load data, and uses a pre-loaded UAV dynamics model to solve the problem, dynamically outputting the UAV's position and attitude for simulation training management. These techniques solve the technical problem of low simulation accuracy leading to poor training results in existing UAV simulation training, achieving the technical effect of improving simulation accuracy and enhancing training practicality.

[0005] This application provides a high-precision simulation training method for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling, comprising: activating a remote control acquisition engine to continuously acquire four-channel rod data at a preset frequency, establishing raw control data, and inputting the raw control data into a closed-loop responder; establishing a current control state flag based on the raw control data, sending the current control state flag to a dynamic wind field controller, and performing scene simulation initialization of the dynamic wind field controller; performing two-dimensional drive control with the initialized dynamic wind field controller to establish a wind field dataset; inputting the wind field dataset into a wind resistance calculation model and outputting wind resistance load data; after the closed-loop responder receives the raw control data, wind field dataset, and wind resistance load data, performing data calculation through a pre-loaded UAV dynamics model, and dynamically outputting the position and attitude of the UAV for UAV simulation training management.

[0006] In a possible implementation, the wind field dataset is input into the wind resistance calculation model, which outputs wind resistance load data and performs the following processing: After receiving the wind field dataset, the preprocessing layer in the wind resistance calculation model synthesizes the horizontal and vertical wind speeds to establish a real-time wind speed; before the real-time wind speed is input into the calculation layer of the wind resistance calculation model, an equivalence judgment for scene adaptation is performed; if the equivalence judgment for scene adaptation fails, the initial calculation unit in the calculation layer is activated to perform wind resistance calculation based on the real-time wind speed and outputs wind resistance load data. The calculation formula for the initial calculation unit is as follows: ;in, For wind resistance load, The air drag coefficient, air density, For the windward area of ​​the drone, The value is the real-time wind speed, and 0.5 is the inherent coefficient.

[0007] In a possible implementation, the equivalent discrimination for scene adaptation is performed, and the following processing is also performed: if the equivalent discrimination for scene adaptation passes, the equivalent calculation unit in the computing layer is activated, and the equivalent calculation unit is used to calculate wind resistance based on the real-time wind speed, and output wind resistance load data. The calculation formula of the equivalent calculation unit is as follows: Wherein, 0.04 is the equivalence coefficient, which is constructed by fitting actual flight data.

[0008] In a possible implementation, a dual-dimensional drive control is executed using the initialized dynamic wind field controller to establish a wind field dataset and perform the following processing: The dual-dimensional drive control includes horizontal wind field control and vertical wind field control, wherein: Horizontal wind field control includes autonomously executing random wind direction switching for 15 to 25 seconds and increasing the wind speed from the initial wind speed level to the target wind speed level in a second-by-second manner, generating dynamic parameters of wind speed and wind direction in the horizontal direction; Vertical wind field control includes executing vertical oscillation control when the vertical wind speed is greater than 0, wherein the vertical oscillation control is constructed based on a transition progress parameter, which represents the normalized progress within the oscillation period. ,in, For transition progress parameters, This represents the duration of the current vertical oscillation. The duration of the complete oscillation cycle is determined by the transition progress parameter. When the transition progress parameter reaches 1, it is automatically reset to 0, and the next oscillation cycle begins.

[0009] In a possible implementation, the following processing is performed: During the vertical wind field control process, the vertical oscillation control also includes an amplitude parameter, which is calculated as follows: Amplitude = sin(transition progress parameter × 2π) × 0.1 × vertical wind force level; where 0.1 is the basic amplitude coefficient and the vertical wind force level is a quantitative classification of airflow intensity.

[0010] In a possible implementation, the following processing is performed: after receiving the original control data, wind field dataset, and wind resistance load data, the closed-loop responder performs calculations through a six-degree-of-freedom dynamic model and outputs the UAV's real-time displacement, attitude adjustment amount, and wind resistance correction strategy.

[0011] In a possible implementation, the following processing is performed: the preset frequency is 50Hz, the four-channel stick quantities include stick quantities corresponding to throttle, pitch, roll, and yaw, and the sampling accuracy of the original control data is ±1%. The original control data is mapped to the pitch angle, roll angle, and thrust parameters of the UAV in real time. The remote control acquisition engine acquires the four-channel stick quantities through a USB interface, and the USB interface is compatible with the ET16S remote controller communication protocol.

[0012] In a possible implementation, a wind field dataset is established, and the following processes are performed: the wind field dataset is synchronized to a safety monitor; the safety monitor is used to analyze the wind speed level within the wind field dataset; if the wind speed level meets a preset threshold, a safety protection strategy is triggered, and wind speed limiting, data anomaly discarding, and simulation pause operations are performed according to the safety protection strategy.

[0013] In a possible implementation, the position and attitude of the UAV are dynamically output, and the following processing is performed: a visualization display component is configured, and the UAV simulation scene is rendered using the visualization display component based on the dynamically output position and attitude of the UAV, and the scene is displayed in real time through a display unit.

[0014] This application also provides a high-precision simulation training system for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling, comprising: a raw control data acquisition module, used to activate the remote control acquisition engine to continuously acquire four-channel rod quantities at a preset frequency, establish raw control data, and input the raw control data to a closed-loop responder; a scene simulation initialization module, used to establish a current control state flag based on the raw control data, send the current control state flag to the dynamic wind field controller, and execute the scene simulation initialization of the dynamic wind field controller; a dual-dimensional drive control module, used to execute dual-dimensional drive control with the initialized dynamic wind field controller to establish a wind field dataset; a wind resistance calculation module, used to input the wind field dataset into a wind resistance calculation model and output wind resistance load data; and a UAV simulation training module, used to perform data calculation through a pre-loaded UAV dynamics model after the closed-loop responder receives the raw control data, wind field dataset, and wind resistance load data, and dynamically output the position and attitude of the UAV for UAV simulation training management.

[0015] The proposed high-precision UAV simulation training method and system based on dynamic wind field modeling, as described in this application, first activates the remote control acquisition engine to continuously acquire four-channel rod data at a preset frequency, establishing raw control data. This raw control data is then input into a closed-loop responder. Next, a current control state flag is established based on the raw control data and sent to a dynamic wind field controller. This executes the scene simulation initialization of the dynamic wind field controller. Then, the initialized dynamic wind field controller performs two-dimensional drive control to establish a wind field dataset. This wind field dataset is then input into a wind resistance calculation model, outputting wind resistance load data. Finally, after receiving the raw control data, wind field dataset, and wind resistance load data, the closed-loop responder performs data calculation using a pre-loaded UAV dynamics model, dynamically outputting the UAV's position and attitude for UAV simulation training management. This achieves the technical effect of improving simulation accuracy and enhancing training practicality. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating the high-precision simulation training method for UAVs based on dynamic wind field modeling, provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the structure of a high-precision simulation training system for unmanned aerial vehicles based on dynamic wind field modeling, provided in an embodiment of this application.

[0019] Figure labeling: 10 raw control data acquisition module, 20 scene simulation initialization module, 30 dual-dimensional drive control module, 40 wind resistance calculation module, 50 UAV simulation training module. Detailed Implementation

[0020] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments; however, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0023] This application provides a high-precision simulation training method for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling, such as... Figure 1 As shown, the method includes:

[0024] Step S100: Activate the remote control acquisition engine to continuously acquire four-channel stick quantities at a preset frequency to establish raw control data. Input the raw control data into the closed-loop responder. The preset frequency is 50Hz. The four-channel stick quantities include stick quantities corresponding to throttle, pitch, roll, and yaw. The sampling accuracy of the raw control data is ±1%. The raw control data is mapped to the pitch angle, roll angle, and thrust parameters of the UAV in real time. The remote control acquisition engine acquires the four-channel stick quantities through a USB interface, and the USB interface is compatible with the ET16S remote controller communication protocol.

[0025] Specifically, the remote control acquisition engine is a hardware module used to acquire stick input signals from the remote controller. It is compatible with the ET16S remote controller communication protocol and can acquire signals at a preset frequency. The preset frequency is 50Hz, meaning it acquires data 50 times per second, ensuring the real-time nature and continuity of the data to meet the data update requirements of simulation training. The four stick input channels correspond to the stick input signals of the four control dimensions: throttle, pitch, roll, and yaw, reflecting the pilot's control commands to the drone.

[0026] By adapting the remote control acquisition engine to the ET16S remote controller communication protocol, the four-channel stick data is continuously acquired via USB interface at a fixed frequency of 50Hz. The acquired data is processed to establish raw control data, which includes the pilot's operation information. The sampling accuracy of this data is controlled within ±1%. It is then mapped in real time to the UAV's pitch angle, roll angle, and thrust parameters, and then input to the closed-loop response unit. The closed-loop response unit is a module used to receive the raw control data and provide basic data input for data processing and simulation training.

[0027] Specifically, by selecting high-precision sensors and optimizing signal conditioning circuits in the hardware, combined with precise sampling algorithms and data calibration and compensation in the software, the sampling accuracy is controlled to ±1%. High-resolution potentiometers or absolute encoders can be selected as high-precision sensors; low-noise operational amplifiers and low-pass filters can be used to optimize the signal conditioning circuits; high-precision timers can be used to trigger sampling interrupts and perform digital filtering for sampling; calibration curves can be established to eliminate zero-point drift and gain errors, and temperature compensation can be performed based on temperature characteristics for data calibration and compensation.

[0028] The real-time mapping of raw control data to the UAV's pitch, roll, and thrust parameters can be achieved by establishing a mapping model. Specifically, through extensive experimental data and theoretical analysis, a mathematical mapping model is established between the raw control data and the UAV's pitch, roll, and thrust parameters. For example, for pitch, data on the actual pitch angles achieved by the UAV under different pitch stick inputs are collected, and regression analysis and other methods are used to fit the functional relationship between the stick input and the pitch angle. Similarly, corresponding mapping functions are established for the roll and thrust parameters. During the simulation training, the remote control acquisition engine continuously collects four-channel stick inputs and generates raw control data, which the system reads in real time. For pitch stick data, it is substituted into the pre-established pitch angle mapping function to calculate the pitch angle parameter value at the current moment; for roll stick data, it is substituted into the roll angle mapping function to calculate the roll angle parameter value; and for throttle stick data, it is substituted into the thrust parameter mapping function to calculate the thrust parameter value. The calculated pitch angle, roll angle, and thrust parameters are used to update the current state parameters of the UAV to achieve real-time synchronization with the pilot's control.

[0029] Step S200: Establish a current control state flag based on the original control data, send the current control state flag to the dynamic wind field controller, and perform scene simulation initialization of the dynamic wind field controller.

[0030] Specifically, based on the received raw control data, a current control state marker is established through logical rules. This marker identifies the pilot's current control state, including characteristics such as control amplitude and direction, and is used by the dynamic wind field controller to understand the current control situation. The current control state marker is then sent to the dynamic wind field controller via a communication link, triggering the controller to perform scenario simulation initialization, preparing for subsequent wind field simulations. The dynamic wind field controller is the module responsible for controlling and managing the dynamic wind field simulation; it initializes the scenario simulation based on received information and generates realistic wind field data.

[0031] For example, when raw control data shows that the pilot is pushing the pitch stick significantly, the system generates a current control state flag based on this data, indicating that the UAV may be in a rapid climb or dive control state. The flag is then sent to the dynamic wind field controller, which initializes the scenario simulation based on this flag to simulate a wind field environment that matches the control state.

[0032] Step S300: Execute dual-dimensional drive control with the initialized dynamic wind field controller to establish a wind field dataset.

[0033] Specifically, the initialized dynamic wind field controller performs dual-dimensional drive control, which refers to simultaneously controlling and adjusting the wind field in both the horizontal and vertical directions to simulate a real and complex natural wind field environment. By controlling wind field parameters such as wind speed and wind direction in the horizontal and vertical directions respectively, a dataset containing rich wind field information is established, namely the wind field dataset.

[0034] In one possible implementation, a dual-dimensional drive control is executed using the initialized dynamic wind field controller to establish a wind field dataset. Step S300 further includes: the dual-dimensional drive control includes horizontal wind field control and vertical wind field control, wherein: the horizontal wind field control includes autonomously executing random wind direction switching for 15 to 25 seconds and increasing the wind speed from the initial wind speed level to the target wind speed level in a second-by-second manner, generating dynamic parameters of wind speed and wind direction in the horizontal direction; the vertical wind field control includes executing vertical oscillation control when the vertical wind speed is greater than 0, wherein the vertical oscillation control is constructed based on a transition progress parameter, the transition progress parameter representing the normalized progress within the oscillation period. ,in, For transition progress parameters, This represents the duration of the current vertical oscillation. The duration of the complete oscillation cycle is determined by the transition progress parameter. When the transition progress parameter reaches 1, it is automatically reset to 0, and the next oscillation cycle begins.

[0035] Specifically, horizontal wind field control includes wind speed and direction control along the positive and negative X and Z axes. The dynamic wind field controller incorporates a random number generation algorithm that generates a random angle value every 1-3 seconds within a time range of 15-25 seconds. This angle value represents the direction of wind change. For example, if the initial wind direction is due east with an angle of 90°, and the first randomly generated angle is 30°, then the wind direction changes to east-northeast with an angle of 90°-30°=60°. This method achieves random switching of wind direction, simulating the variable characteristics of wind direction in nature.

[0036] The initial and target wind speed levels are determined. The initial wind speed level is the baseline wind speed value at the start of horizontal wind field control, marking the beginning of wind speed increases. The target wind speed level is the desired wind speed value achieved during horizontal wind field control, representing the end point of wind speed increases. For example, the initial wind speed level is 2 m / s, and the target wind speed level is 10 m / s. The dynamic wind field controller adjusts the wind speed by increasing it second by second; that is, the wind speed increases by a fixed increment every second. The increment is calculated as: (target wind speed level - initial wind speed level) / total increment time. Assuming a total increment time of 20 seconds, the increment is (10-2) / 20 = 0.4 m / s. Therefore, in the first second, the wind speed is 2 + 0.4 = 2.4 m / s; in the second second, the wind speed is 2.4 + 0.4 = 2.8 m / s, and so on, until the target wind speed level of 10 m / s is reached.

[0037] By controlling the horizontal wind field, dynamic parameters of wind speed and direction in the horizontal direction are generated. These parameters record the wind speed and direction at different time points, providing horizontal information for the wind field dataset.

[0038] When the vertical wind speed is greater than 0, vertical oscillation control is implemented. Vertical oscillation control is based on a transition progress parameter. The normalized progress, representing the oscillation period, ranges from [0,1] and increases linearly with time. Its calculation formula is as follows: Where t is the current duration of the vertical oscillation, starting from the moment the oscillation is triggered, and T is the duration of the complete oscillation cycle, i.e., the time required to complete one complete up-and-down oscillation in the vertical oscillation control, which determines the frequency of vertical wind speed changes. For example, the default T = 3 seconds can be configured, which adapts to the common vertical airflow influence cycle. When the transition progress parameter... When the value reaches 1, it automatically resets to 0 and enters the next oscillation cycle, ensuring the periodic dynamic changes of vertical oscillation. Based on the transition progress parameter, the vertical wind speed is adjusted using a specific functional relationship to generate dynamic wind speed parameters in the vertical direction. The changes in wind speed in the vertical wind field at different time points are recorded, and together with the dynamic parameters of wind speed and direction in the horizontal direction, they constitute a complete wind field dataset, providing a realistic wind field environment for UAV simulation training.

[0039] In one possible implementation, step S300 further includes: during the vertical wind field control process, the vertical oscillation control also includes an amplitude parameter, which is calculated as follows: amplitude = sin(transition progress parameter × 2π) × 0.1 × vertical wind force level; where 0.1 is the basic amplitude coefficient and the vertical wind force level is a quantitative classification of airflow intensity.

[0040] Specifically, the amplitude parameter is calculated using a formula, where sin(transition progress parameter × 2π) is a sine function with a period of 1, ranging from -1 to 1. The transition progress parameter controls the phase change of the sine function, and 2π maps the transition progress to the interval [0, 2π], allowing the sine function to complete one full cycle. The base amplitude coefficient is set to 0.1, a value fitted from actual flight data, ensuring that the amplitude matches the vertical wind force level and that the simulation deviation is ≤5%. The base amplitude coefficient acts as a scaling factor to adjust the baseline amplitude. The vertical wind force level is a quantitative classification of airflow intensity, with different levels representing different airflow intensities. For example, the vertical wind force level can be divided into 1-5 levels, with higher levels indicating stronger vertical airflow.

[0041] The calculated sine function value, the basic amplitude coefficient, and the vertical wind force level are multiplied to obtain the final amplitude parameter. Based on this calculated amplitude parameter, the vertical wind speed is dynamically adjusted. Within one oscillation cycle, the vertical wind speed varies according to a sine curve, ranging from -amplitude to +amplitude (the negative sign indicates opposite wind direction). This amplitude parameter-based vertical oscillation control is applied to the vertical wind field, combined with horizontal wind field control, to construct a two-dimensional dynamic wind field, thereby establishing a rich and realistic wind field dataset.

[0042] In one possible implementation, a wind field dataset is established, and the method further includes: synchronizing the wind field dataset to a safety monitor; using the safety monitor to analyze the wind speed levels within the wind field dataset; if the wind speed levels meet a preset threshold, a safety protection strategy is triggered, and wind speed limiting, data anomaly discarding, and simulation pause operations are performed according to the safety protection strategy.

[0043] Specifically, a real-time data transmission protocol, such as a custom communication protocol based on TCP / IP, is used to synchronize the established wind farm dataset from the dynamic wind farm controller to the safety monitor. During data transmission, a data verification mechanism, such as CRC checksum, is employed to ensure data integrity and accuracy. When sending the wind farm dataset, the dynamic wind farm controller calculates the CRC checksum for that dataset and appends it to the data packet. Upon receiving the data packet, the safety monitor performs the same CRC calculation on the data in the dataset and compares the calculated checksum with the received checksum. If the two checksums match, it indicates that no errors occurred during data transmission and the data can be received and processed normally; if they do not match, it indicates a potential error in the data, and the safety monitor sends a retransmission request to the dynamic wind farm controller, requesting a retransmission of the dataset.

[0044] To achieve real-time synchronization, a fixed data update frequency is set, such as synchronizing the wind field dataset every 100ms, to ensure that the safety monitor can obtain the latest wind field data in a timely manner.

[0045] The safety monitor pre-stores wind speed classification standards, which can be set according to the actual application scenario and the performance parameters of the drone. For example, wind speed can be divided into four levels: low, medium, high, and extremely high. After receiving the wind field dataset, the safety monitor extracts the wind speed data for each location. Then, based on the pre-set wind speed classification standards, it determines the wind speed level for each location. Different wind speed level thresholds are set according to the drone's safe flight requirements and the simulation training objectives. For example, when the wind speed level reaches high, it is set as a level one warning threshold; when the wind speed level reaches extremely high, it is set as a level two danger threshold. If the wind speed level meets the level one warning threshold, the safety monitor sends a wind speed limiting command to the dynamic wind field controller. Upon receiving the command, the dynamic wind field controller adjusts the wind speed in the wind field dataset according to the preset limiting value. During wind speed level analysis, if significant anomalies are detected in wind speed data at certain locations—for example, a sudden increase in wind speed to an unreasonable value—it may be due to data acquisition errors or transmission interference. The safety monitor will mark these abnormal data as invalid and discard them from the wind field dataset. Simultaneously, it will record the location and time information of the abnormal data for subsequent troubleshooting and data analysis. When the wind speed level meets the level 2 hazard threshold, the safety monitor will immediately send a simulation pause command to the simulation training system. Upon receiving the command, the simulation training system will pause the drone's simulated flight and maintain the drone's current position. At the same time, corresponding warning messages will be displayed on the operating interface to alert the operator to the abnormal wind field situation. The operator can then take appropriate action, such as checking whether the wind field simulation equipment is functioning properly. After the problem is resolved, the operator can manually resume the simulation training.

[0046] Step S400: Input the wind field dataset into the wind resistance calculation model and output the wind resistance load data.

[0047] Specifically, the wind field dataset is taken as input and fed into a pre-built wind resistance calculation model. This model is a mathematical model based on principles of fluid mechanics and other fields, used to calculate the wind resistance load experienced by the UAV under specific wind field conditions. The wind resistance calculation model performs calculations based on the input wind field data and outputs corresponding wind resistance load data, representing information such as the magnitude of wind resistance experienced by the UAV under specific wind field conditions.

[0048] In one possible implementation, the wind field dataset is input into the wind resistance calculation model, which outputs wind resistance load data. Step S400 further includes step S410, whereby the preprocessing layer in the wind resistance calculation model, after receiving the wind field dataset, synthesizes the horizontal and vertical wind speeds to establish real-time wind speeds. Specifically, the preprocessing layer of the wind resistance calculation model receives the wind field dataset, which contains multi-dimensional information such as horizontal and vertical wind speeds. The preprocessing layer uses a vector synthesis method to synthesize the horizontal and vertical wind speeds. Simultaneously, the direction of the real-time wind speed can be calculated, and its angle in three-dimensional space can be determined using the arctangent function.

[0049] Step S420: Before inputting the real-time wind speed into the calculation layer of the wind resistance calculation model, an equivalence judgment for scene adaptation is performed. Specifically, before inputting the real-time wind speed into the calculation layer of the wind resistance calculation model, an equivalence judgment for scene adaptation is performed to determine whether the current simulation scenario is suitable for using the simplified formula for wind resistance calculation. Using the simplified formula can improve computational efficiency, reduce computation time during simulation, thereby reducing the delay in wind resistance response and enabling the simulation results to more promptly reflect the actual force situation of the UAV in the wind field. The simplified formula is derived and fitted under specific conditions and is not applicable to all scenarios. Therefore, the equivalence judgment for scene adaptation is used to determine whether the current scenario meets the conditions for using the simplified formula.

[0050] Specifically, the equivalence determination of scene adaptation is performed by pre-setting a series of scene feature parameters and determination rules. These scene feature parameters include the drone's flight altitude range, flight speed range, and wind speed change frequency. If the parameters of the current scene are all within the preset reasonable range, the equivalence determination of scene adaptation is considered to pass, meaning that the current scene is suitable for using the simplified formula to calculate wind resistance; conversely, if any parameter exceeds the preset range, the determination fails, and the standard formula needs to be used for calculation.

[0051] Step S430: If the scene adaptation equivalence judgment fails, the initial calculation unit in the calculation layer is activated to perform wind resistance calculation based on the real-time wind speed and output wind resistance load data. The calculation formula of the initial calculation unit is as follows: ;in, For wind resistance load, The air drag coefficient, air density, For the windward area of ​​the drone, The value represents the real-time wind speed, and 0.5 is an inherent coefficient. Specifically, when the equivalence judgment for scene adaptation fails, it indicates that the current simulated scene does not match the scene applicable to the equivalent calculation unit. In this case, the initial calculation unit within the calculation layer is activated to calculate wind resistance. The initial calculation unit adopts a relatively universal and accurate calculation method, which can adapt to various different scene conditions, ensuring accurate calculation of wind resistance load data in various complex environments.

[0052] The wind resistance calculation formula used in the initial calculation unit is: ,in, The wind resistance load is the magnitude of the resistance experienced by the UAV in the wind field. It reflects the degree of influence of wind on the flight of the UAV and is the core result of wind resistance calculation. The drag coefficient is a dimensionless coefficient related to factors such as the shape and surface roughness of the drone. Different drone models have different drag coefficients. For example, streamlined drones have lower drag coefficients, reducing air resistance during flight; drones with more complex shapes have higher drag coefficients, resulting in greater wind resistance. This coefficient can be determined through wind tunnel experiments or numerical simulations. Air density varies with factors such as altitude, temperature, and humidity. Generally, the higher the altitude, the lower the air density; the higher the temperature, the lower the air density; and the higher the humidity, the lower the air density. In calculations, the air density value needs to be determined based on the actual conditions of the current simulation scenario. The windward area of ​​a drone is its projected area in the direction of wind speed. The calculation method for windward area varies depending on the drone's shape. For example, for a cuboid-shaped drone, the windward area can be calculated based on its length, width, and height. A larger windward area results in greater wind resistance for the drone. The real-time wind speed is obtained through wind speed synthesis in step S410. The magnitude and direction of the real-time wind speed directly affect the magnitude of the wind resistance load; the higher the wind speed, the greater the wind resistance load. 0.5 is the natural coefficient, a fixed value derived from relevant theories and derivations in fluid mechanics.

[0053] Step S440: If the scene adaptation equivalence judgment passes, the equivalent calculation unit in the calculation layer is activated. The equivalent calculation unit performs wind resistance calculation based on the real-time wind speed and outputs wind resistance load data. The calculation formula of the equivalent calculation unit is as follows: Wherein, 0.04 is the equivalence coefficient, which is constructed by fitting actual flight data.

[0054] Specifically, if the equivalence judgment for scene adaptation passes, it means that the current simulation scene meets the conditions applicable to the equivalent calculation unit. At this time, the equivalent calculation unit in the calculation layer is activated to perform wind resistance calculation. Using the equivalent calculation unit can quickly calculate wind resistance load data using simplified formulas, improve calculation efficiency, reduce calculation time in the simulation process, thereby reducing the delay in wind resistance response and making the simulation results more timely and accurate.

[0055] The formula for calculating the equivalent computational unit is: ,in, Similarly, the wind resistance load has the same physical meaning as the wind resistance load calculated by the initial calculation unit, both representing the magnitude of drag experienced by the UAV in a wind field. 0.04 is an equivalence coefficient, constructed through fitting real-flight data. During the development process, a large amount of data on UAVs in actual flight was collected, including wind resistance load data under different flight altitudes, speeds, and wind speeds. This equivalence coefficient was obtained by analyzing and processing this real-flight data using a mathematical fitting method. The simplified formula constructed using this equivalence coefficient can accurately approximate the wind resistance load in specific scenarios, while simplifying the calculation process and improving computational efficiency. The real-time wind speed is consistent with the real-time wind speed in the initial calculation unit and is obtained through wind speed synthesis in step S410.

[0056] Step S500: After the closed-loop responder receives the original control data, wind field dataset, and wind resistance load data, it performs data calculation through the pre-loaded UAV dynamics model and dynamically outputs the position and attitude of the UAV for UAV simulation training management.

[0057] Specifically, the closed-loop responder receives multi-source data, including raw control data, wind field datasets, and wind resistance load data. It then uses a pre-loaded UAV dynamics model to comprehensively solve these data, dynamically outputting the UAV's position and attitude information through mathematical and physical calculations, thereby enabling the management and display of UAV simulation training. The UAV dynamics model is a mathematical model based on the physical characteristics and motion laws of the UAV, used to describe the UAV's motion state under different input conditions. Position refers to the UAV's coordinate information in space, while attitude includes information such as the UAV's pitch angle, roll angle, and yaw angle, reflecting the UAV's specific state in the air.

[0058] In one possible implementation, step S500 further includes step S510, whereby the closed-loop responder receives the original control data, wind field dataset, and wind resistance load data, and then performs calculations using a six-degree-of-freedom dynamic model to output the UAV's real-time displacement, attitude adjustment amount, and wind resistance correction strategy.

[0059] Specifically, the six-degree-of-freedom (DOF) dynamics model describes the motion of a UAV in three-dimensional space, including translational motion (forward, backward, ascent, descent, leftward, rightward) along three coordinate axes and rotational motion (pitch, roll, yaw) around these axes. Based on classical mechanics principles such as the Newton-Euler equations, this model integrates the influence of the UAV's mass, moment of inertia, external forces, and external torques on its motion. By establishing a system of mathematical equations, the six-DOF dynamics model can simulate the dynamic behavior of the UAV under various conditions.

[0060] The closed-loop responder takes the received raw control data, wind field dataset, and wind resistance load data as input parameters and substitutes them into the six-degree-of-freedom dynamic model for calculation. During the calculation process, the model calculates the resultant force and resultant torque acting on the UAV in real time based on these input parameters. Then, according to Newton's second law and Euler's equations, it further calculates the UAV's acceleration and angular acceleration, and then obtains the UAV's velocity and angular velocity through integration. Finally, it integrates again to obtain the UAV's displacement and attitude angles. For example, when receiving raw control data from the pilot to increase the throttle, the model calculates the increase in engine thrust, combines wind resistance load data and external forces such as gravity, calculates the resultant force on the UAV in the forward direction, and then obtains the UAV's acceleration and velocity changes, ultimately determining the UAV's real-time displacement after a period of time. At the same time, based on control data such as the control surface deflection angle, the model calculates the torque of the UAV around each coordinate axis, thereby obtaining the UAV's angular acceleration, angular velocity, and attitude angle changes, and determining the UAV's attitude adjustment amount.

[0061] Real-time displacement represents the UAV's position change in three-dimensional space, expressed using coordinates (x, z, y). The x-axis represents the east-west direction, the z-axis the north-south direction, and the y-axis the vertical direction (altitude). By monitoring the UAV's displacement in real time, it's possible to understand whether the UAV is flying along the predetermined route and how much it deviates due to wind conditions. Attitude adjustment reflects the UAV's rotation angles around three coordinate axes: pitch, roll, and yaw. Pitch angle represents the UAV's rotation angle around the y-axis, affecting its ascent and descent; roll angle represents the UAV's rotation angle around the x-axis, affecting its left and right tilt; and yaw angle represents the UAV's rotation angle around the z-axis, affecting its flight direction. Attitude adjustment is a key parameter for maintaining stable flight and performing various flight maneuvers. For example, when the UAV encounters crosswinds, it needs to adjust the roll angle to counteract the crosswind's effects and maintain flight direction stability. The wind resistance correction strategy is a set of specific operational suggestions provided to the pilot by a closed-loop responder based on real-time displacement and attitude adjustments, combined with wind field data and the performance characteristics of the UAV. These suggestions include adjusting throttle, changing control surface deflection angles, and adjusting flight altitude. For example, when a strong updraft is detected affecting the UAV and its altitude continues to rise, the wind resistance correction strategy will suggest appropriately reducing the throttle and adjusting the pitch angle to maintain a stable altitude for the UAV.

[0062] In one possible implementation, the position and attitude of the UAV are dynamically output. Step S500 further includes step S520, configuring a visualization display component, using the visualization display component to render a UAV simulation scene based on the dynamically output position and attitude of the UAV, and displaying the scene in real time through a display unit.

[0063] Specifically, the visualization component is the core of the visualization system, used to convert the UAV's position and attitude data into visualized image information. This component includes a series of algorithms and graphics processing modules capable of performing coordinate transformations, model rendering, and lighting calculations based on the input data, thereby generating realistic UAV simulation scenes. When configuring the visualization component, settings need to be tailored to the actual UAV model and the requirements of the simulation scene. For example, a suitable 3D UAV model should be selected, accurately reflecting the UAV's appearance and structural features; the background environment of the scene, such as the sky, ground, and buildings, should be set to enhance the realism of the simulation; and lighting parameters should be adjusted to ensure the UAV model has appropriate brightness and shadow effects within the scene. Furthermore, the component's interface needs to be configured to communicate with the module that dynamically outputs the UAV's position and attitude data, enabling real-time data acquisition and processing.

[0064] Rendering refers to the process of converting a 3D model into a 2D image. In drone simulation scene rendering, the visualization component performs coordinate transformation on the 3D drone model based on dynamically output drone position and attitude data, placing it in the correct position within the scene and rotating it according to its current attitude. Then, based on the lighting conditions in the scene, it calculates the light intensity and color values ​​of each point on the drone model's surface, combining this with optical effects such as shadows, reflections, and refractions to ultimately generate a realistic image. To improve rendering efficiency and image quality, the visualization component employs optimization techniques. For example, it uses layered detail technology, selecting different levels of detail for rendering based on the distance between the drone model and the observer—using a simplified model for greater distances and a refined model for closer distances—to reduce unnecessary computation. Texture mapping technology is used to apply pre-made texture images to the drone model's surface, increasing detail and realism. Multi-threaded rendering technology is utilized, distributing rendering tasks across multiple processor cores for parallel processing, thus improving rendering speed.

[0065] The display unit can be any type of display device, such as a computer monitor, virtual reality headset, or augmented reality glasses. The appropriate display unit is selected based on different application scenarios and requirements. The visualization display component transmits the rendered image data to the display unit in real time, and the display unit displays the image based on the received data.

[0066] The following is the comparative test verification data, including the test basic conditions, key indicator test data, and economic benefit verification data.

[0067] The basic testing conditions included the test object and the test environment. The test object was a Zhongke Haodian multi-rotor drone used for drone pilot training (body weight 14.5kg, takeoff weight 25kg, equipped with an 8220 100KV motor and FOC / 80A ESC, price: 49,800 RMB); real flight data and simulated data were collected simultaneously. The test environment was a standard outdoor airspace (covering stable wind speed areas and turbulent areas); the simulation system ran on a standard industrial computer (CPU: i7-12700, memory: 32GB).

[0068] The test data results for key indicators are shown in Table 1.

[0069] Table 1: Key Indicator Test Data Results

[0070] Technical indicators Test scenarios / methods Sample data Results Explanation Vertical wind simulation Simulate a level 4 vertical wind (wind speed 3.4 m / s) and compare the vertical displacement fluctuation curves of the UAV with and without the "sine oscillation algorithm". When the algorithm is enabled, the displacement fluctuation period fit is ≥95% (compared to the theoretical sine curve); when disabled, there are irregular fluctuations. Verify the effectiveness of the sinusoidal oscillation algorithm in simulating vertical wind fields. Smoothness of wind speed switching The wind speed was increased second by second, from level 2 (0.4 m / s) to level 5 (10.7 m / s), and the rate of change in wind speed was collected. Wind speed abrupt change rate ≤3% (theoretical gradual change vs. actual switching data); existing technology abrupt change rate >20%. Prove the smoothness advantage of stepwise gradient logic Wind resistance response delay Injecting a step wind speed (0→6 level wind, 1.2m / s→13.8m / s), the response time is calculated by recording wind resistance using a high-speed data acquisition card. Average latency 18ms (minimum 15ms, maximum 22ms); existing technology averages 120ms. Formula simplification and hardware adaptation achieve low latency Controlling the acquisition precision A standard lever signal (such as a throttle channel step change from 0% to 100%) is sent at a frequency of 50Hz, and the deviation between the values ​​collected by the simulation system and the actual values ​​is compared. The absolute value of the deviation is ≤0.8% (mean of 1000 samples); the average deviation of existing technology is 4.2%. High-frequency acquisition + algorithm calibration improves accuracy Actual flight fit Five pilots (including two instructors and three trainees) were selected to simultaneously record 12 key indicators (position deviation, attitude response, etc.) in three types of tasks: "hovering against wind," "headwind flight," and "vertical airflow response," and the consistency between actual flight and simulation was calculated. The average agreement rate for Task 1 was 93%, for Task 2 it was 91%, and for Task 3 it was 92%; existing technologies averaged 68%. The high degree of fit between the multi-scenario verification simulation system and actual flight tests

[0071] The economic benefit verification is based on the core scenario, taking the Zhongke Haodian multi-rotor UAV used for training (single unit cost of 49,800 yuan) as the object, and according to the loss rule of "10 sessions per year, 40 people per session, and 1 UAV damaged per session when there is no simulation", the total cost difference of "no simulation and full actual flight" and "40% simulation training replacement" over 5 years is compared to verify the long-term benefits of this application.

[0072] The 5-year benefit calculation results are as follows:

[0073] (a) Full live flight mode without simulation (benchmark group)

[0074] Annual drone damage cost: 1 drone damaged per period × 10 periods per year = 10 drones / year, annual damage cost = 10 drones × 49,800 yuan / drone = 498,000 yuan; total damage cost over 5 years: 498,000 yuan / year × 5 years = 2,490,000 yuan; other loss costs: calculated based on "annual loss per drone during full flight including 16,600 yuan depreciation + 20,000 yuan maintenance", 5-year depreciation + maintenance cost of 40 drones = (16,600 + 20,000) / drone / year × 40 drones × 5 years = 7,320,000 yuan; total cost over 5 years: 2,490,000 yuan + 7,320,000 yuan = 9,810,000 yuan.

[0075] (ii) 40% simulation training alternative mode (experimental group)

[0076] Annual drone damage cost: With a 40% reduction in actual flight frequency, the probability of damage decreases by 40%. 0.6 drones damaged per period × 10 periods per year = 6 drones / year. Annual damage cost = 6 drones × 49,800 RMB / drone = 298,800 RMB. Total damage cost over 5 years: 298,800 RMB / year × 5 years = 1,494,000 RMB. Other depreciation costs: Maintenance costs are reduced to 0 due to reduced actual flight frequency; only depreciation is retained. Depreciation cost for 40 drones over 5 years = 16,600 RMB / drone / year × 40 drones × 5 years = 3,320,000 RMB. Total cost over 5 years: 1,494,000 RMB + 3,320,000 RMB = 4,814,000 RMB.

[0077] (III) Cost Reduction Effect over 5 Years

[0078] Comparing the two models, after replacing 40% with simulation training: the total cost over 5 years decreased from 9.81 million yuan to 4.814 million yuan, with a cumulative cost reduction of approximately 4.996 million yuan over 5 years, and an average annual cost reduction of approximately 999,200 yuan.

[0079] This application embodiment employs a remote control acquisition engine to acquire four-channel rod data at a preset frequency to generate raw control data, which is then input into a closed-loop responder. Based on the raw control data, a current control state flag is established and sent to a dynamic wind field controller to complete its scene simulation initialization. The initialized dynamic wind field controller performs dual-dimensional drive control to generate a wind field dataset. The wind field dataset is input into a wind resistance calculation model to obtain wind resistance load data. The closed-loop responder receives the raw control data, the wind field dataset, and the wind resistance load data, and uses a pre-loaded UAV dynamics model to calculate and dynamically output the UAV position and attitude for simulation training management. These technical means solve the technical problem of low simulation accuracy leading to poor training results in existing UAV simulation training, achieving the technical effect of improving simulation accuracy and enhancing training practicality.

[0080] In the above text, refer to Figure 1 This paper describes in detail a high-precision simulation training method for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling according to embodiments of the present invention. Next, we will refer to... Figure 2 This invention describes a high-precision simulation training system for unmanned aerial vehicles (UAVs) based on dynamic wind field modeling, according to an embodiment of the present invention.

[0081] The high-precision UAV simulation training system based on dynamic wind field modeling according to embodiments of the present invention addresses the technical problem of low simulation accuracy leading to poor training results in existing UAV simulation training, thereby improving simulation accuracy and enhancing training practicality. The high-precision UAV simulation training system based on dynamic wind field modeling includes: a raw control data acquisition module 10, a scene simulation initialization module 20, a dual-dimensional drive control module 30, a wind resistance calculation module 40, and a UAV simulation training module 50.

[0082] The raw control data acquisition module 10 is used to activate the remote control acquisition engine to continuously acquire four-channel rod quantities at a preset frequency, establish raw control data, and input the raw control data to the closed-loop responder; the scene simulation initialization module 20 is used to establish a current control state flag based on the raw control data, send the current control state flag to the dynamic wind field controller, and execute the scene simulation initialization of the dynamic wind field controller; the dual-dimensional drive control module 30 is used to execute dual-dimensional drive control with the initialized dynamic wind field controller to establish a wind field dataset; the wind resistance calculation module 40 is used to input the wind field dataset into the wind resistance calculation model and output wind resistance load data; the UAV simulation training module 50 is used to perform data calculation through a pre-loaded UAV dynamics model after the closed-loop responder receives the raw control data, wind field dataset, and wind resistance load data, and dynamically output the position and attitude of the UAV for UAV simulation training management.

[0083] The specific configuration of the wind resistance calculation module 40 will be described in detail below. As mentioned above, the wind field dataset is input into the wind resistance calculation model, and wind resistance load data is output. The wind resistance calculation module 40 may further include: a wind speed synthesis unit, which uses the preprocessing layer in the wind resistance calculation model to synthesize the horizontal and vertical wind speeds after receiving the wind field dataset to establish a real-time wind speed; an equivalence discrimination unit, which performs scene adaptation equivalence discrimination before the real-time wind speed is input into the calculation layer of the wind resistance calculation model; and an initial calculation unit, which, if the scene adaptation equivalence discrimination fails, activates the initial calculation unit in the calculation layer to perform wind resistance calculation based on the real-time wind speed and outputs wind resistance load data. The calculation formula of the initial calculation unit is as follows: ;in, For wind resistance load, The air drag coefficient, air density, For the windward area of ​​the drone, The value is the real-time wind speed, and 0.5 is the inherent coefficient.

[0084] The wind resistance calculation module 40, which performs equivalent discrimination for scene adaptation, may further include: an equivalent calculation unit that, if the equivalent discrimination for scene adaptation passes, activates the equivalent calculation unit within the calculation layer, uses the equivalent calculation unit to perform wind resistance calculation based on the real-time wind speed, and outputs wind resistance load data. The calculation formula of the equivalent calculation unit is as follows: Wherein, 0.04 is the equivalence coefficient, which is constructed by fitting actual flight data.

[0085] The specific configuration of the dual-dimensional drive control module 30 will be described in detail below. As mentioned above, the dual-dimensional drive control is executed using the initialized dynamic wind field controller to establish a wind field dataset. The dual-dimensional drive control module 30 may further include: the dual-dimensional drive control includes horizontal wind field control and vertical wind field control, wherein: the horizontal wind field control includes autonomously executing random wind direction switching for 15 to 25 seconds and increasing the wind speed from the initial wind speed level to the target wind speed level in a second-by-second manner, generating dynamic parameters of wind speed and wind direction in the horizontal direction; the vertical wind field control includes executing vertical oscillation control when the vertical wind speed is greater than 0, wherein the vertical oscillation control is constructed based on a transition progress parameter, and the transition progress parameter represents the normalized progress within the oscillation period. ,in, For transition progress parameters, This represents the duration of the current vertical oscillation. The duration of the complete oscillation cycle is determined by the transition progress parameter. When the transition progress parameter reaches 1, it is automatically reset to 0, and the next oscillation cycle begins.

[0086] The dual-dimensional drive control module 30 may further include: during the vertical wind field control process, the vertical oscillation control also includes an amplitude parameter, which is calculated as follows: amplitude = sin(transition progress parameter × 2π) × 0.1 × vertical wind force level; where 0.1 is the basic amplitude coefficient and the vertical wind force level is the quantitative classification of airflow intensity.

[0087] The specific configuration of the UAV simulation training module 50 will be described in detail below. As mentioned above, the UAV simulation training module 50 may further include: a calculation unit for the closed-loop responder to receive the original control data, wind field dataset, and wind resistance load data, and then calculates them using a six-degree-of-freedom dynamic model to output the UAV's real-time displacement, attitude adjustment amount, and wind resistance correction strategy.

[0088] The specific configuration of the raw control data acquisition module 10 will be described in detail below. As mentioned above, the raw control data acquisition module 10 may further include: a preset frequency of 50Hz; four-channel stick quantities including stick quantities corresponding to throttle, pitch, roll, and yaw; and a sampling accuracy of ±1% for the raw control data. The raw control data is mapped in real time to the pitch angle, roll angle, and thrust parameters of the UAV. The remote control acquisition engine acquires the four-channel stick quantities through a USB interface, and the USB interface is compatible with the ET16S remote controller communication protocol.

[0089] The system, which establishes a wind field dataset, may further include: a data synchronization module for synchronizing the wind field dataset to a safety monitor; a wind speed level analysis module for performing wind speed level analysis within the wind field dataset using the safety monitor; and a safety protection module for triggering a safety protection strategy if the wind speed level meets a preset threshold, and performing wind speed limiting, data anomaly discarding, and simulation pause operations according to the safety protection strategy.

[0090] The drone simulation training module 50, which dynamically outputs the position and attitude of the drone, may further include: a simulation scene rendering unit for configuring a visualization display component, which uses the visualization display component to render a drone simulation scene based on the dynamically output position and attitude of the drone, and displays and outputs it in real time through the display unit.

[0091] The UAV high-precision simulation training system based on dynamic wind field modeling provided in this embodiment of the invention can execute the UAV high-precision simulation training method based on dynamic wind field modeling provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0092] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0093] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A high-precision simulation training method for UAVs based on dynamic wind field modeling, characterized in that, The method includes: The remote control acquisition engine is activated to continuously acquire four-channel lever data at a preset frequency, establishing raw control data, which is then input into the closed-loop responder. A current control state flag is established based on the original control data, and the current control state flag is sent to the dynamic wind field controller to perform scene simulation initialization of the dynamic wind field controller. The initialized dynamic wind field controller is used to perform two-dimensional drive control to establish a wind field dataset. Input the wind field dataset into the wind resistance calculation model and output the wind resistance load data; After receiving the raw control data, wind field dataset, and wind resistance load data, the closed-loop responder performs data calculations through a pre-loaded UAV dynamics model and dynamically outputs the UAV's position and attitude for UAV simulation training management. The wind field dataset is input into the wind resistance calculation model, and the wind resistance load data is output, including: After receiving the wind field dataset, the preprocessing layer in the wind resistance calculation model synthesizes the horizontal and vertical wind speeds to establish real-time wind speeds. Before the real-time wind speed is input into the calculation layer of the wind resistance calculation model, an equivalent judgment for scenario adaptation is performed. If the scene adaptation equivalence judgment fails, the initial calculation unit in the calculation layer is activated to perform wind resistance calculation based on the real-time wind speed and output wind resistance load data. The calculation formula of the initial calculation unit is as follows: ； in, For wind resistance load, The air drag coefficient, air density, For the windward area of ​​the drone, The value represents the real-time wind speed, and 0.5 is an inherent coefficient. The equivalence determination for execution scenario adaptation also includes: If the scene adaptation equivalence judgment passes, the equivalent calculation unit in the calculation layer is activated. The equivalent calculation unit performs wind resistance calculation based on the real-time wind speed and outputs wind resistance load data. The calculation formula of the equivalent calculation unit is as follows: ； Wherein, 0.04 is the equivalence coefficient, which is constructed by fitting actual flight data; The initialized dynamic wind field controller performs two-dimensional drive control to establish a wind field dataset, including: The dual-dimensional drive control includes horizontal wind field control and vertical wind field control, wherein: Horizontal wind field control includes autonomously executing random wind direction switching for 15 to 25 seconds and increasing the wind speed from the initial wind speed level to the target wind speed level in a second-by-second manner, generating dynamic parameters of wind speed and wind direction in the horizontal direction; Vertical wind field control includes executing vertical oscillation control when the vertical wind speed is greater than 0. This vertical oscillation control is constructed based on a transition progress parameter, which characterizes the normalized progress within the oscillation period. ,in, For transition progress parameters, This represents the duration of the current vertical oscillation. The duration of the complete oscillation cycle is determined by the transition progress parameter. When the transition progress parameter reaches 1, it is automatically reset to 0, and the next oscillation cycle begins. In the process of vertical wind field control, vertical oscillation control also includes an amplitude parameter, which is calculated as follows: Amplitude = sin(transition progress parameter × 2π) × 0.1 × vertical wind force level; Here, 0.1 is the basic amplitude coefficient, and the vertical wind force level is a quantitative classification of airflow intensity.

2. The high-precision simulation training method for UAVs based on dynamic wind field modeling as described in claim 1, characterized in that, After receiving the original control data, wind field dataset, and wind resistance load data, the closed-loop responder performs calculations using a six-degree-of-freedom dynamic model, and outputs the UAV's real-time displacement, attitude adjustment, and wind resistance correction strategy.

3. The high-precision simulation training method for UAVs based on dynamic wind field modeling as described in claim 1, characterized in that, The preset frequency is 50Hz, the four-channel stick quantities include stick quantities corresponding to throttle, pitch, roll, and yaw, and the sampling accuracy of the raw control data is ±1%. The raw control data is mapped to the pitch angle, roll angle, and thrust parameters of the UAV in real time. The remote control acquisition engine acquires the four-channel stick quantities through a USB interface, and the USB interface is compatible with the ET16S remote controller communication protocol.

4. The high-precision simulation training method for UAVs based on dynamic wind field modeling as described in claim 1, characterized in that, Establish a wind field dataset, including: Synchronize the wind field dataset to the safety monitor; The safety monitor was used to perform wind speed level analysis within the wind field dataset. If the wind speed level meets the preset threshold, a safety protection strategy is triggered, and wind speed limiting, data anomaly discarding, and simulation pause operations are performed according to the safety protection strategy.

5. The high-precision simulation training method for UAVs based on dynamic wind field modeling as described in claim 1, characterized in that, Dynamically output the drone's position and attitude, including: Configure a visualization display component, use the visualization display component to render the drone simulation scene based on the dynamically output drone position and attitude, and display the output in real time through the display unit.

6. A high-precision simulation training system for unmanned aerial vehicles based on dynamic wind field modeling, characterized in that, The system is used to implement the high-precision simulation training method for unmanned aerial vehicles based on dynamic wind field modeling as described in any one of claims 1-5, and the system comprises: The raw control data acquisition module is used to activate the remote control acquisition engine to continuously acquire four-channel lever quantities at a preset frequency, establish raw control data, and input the raw control data to the closed-loop responder. The scenario simulation initialization module is used to establish a current control state flag based on the original control data, send the current control state flag to the dynamic wind field controller, and perform scenario simulation initialization of the dynamic wind field controller. The dual-dimensional drive control module is used to execute dual-dimensional drive control with the initialized dynamic wind field controller and establish a wind field dataset. The wind resistance calculation module is used to input the wind field dataset into the wind resistance calculation model and output wind resistance load data. The UAV simulation training module is used to perform data calculations through a pre-loaded UAV dynamics model after the closed-loop responder receives the raw control data, wind field dataset, and wind resistance load data, and dynamically output the position and attitude of the UAV for UAV simulation training management.

Images