Multi-rotor unmanned aerial vehicle adaptive feedforward control method based on mesoscopic wind field model

Abstract

The invention belongs to the technical field of multi-rotor unmanned aerial vehicle control, and particularly relates to a multi-rotor unmanned aerial vehicle self-adaptive feedforward control method based on a mesoscopic wind field model. Comprising the following steps: firstly, based on a relaxation time model of a lattice Boltzmann method, establishing a mapping relation between a Knudsen number and turbulence intensity and thermal noise intensity of a macroscopic wind field, and realizing parameterized representation of a mesoscopic wind field; then, constructing a mesoscopic wind field model by synthesizing a real-time wind field containing average wind, turbulent flow, gust and thermal noise components; then, calculating the wind resistance according to the predicted wind speed, designing a self-adaptive gain mechanism fusing a real-time tracking error, an error wind direction included angle and a historical error trend, and generating a dynamic feedforward control quantity; and finally, combining the feedforward control quantity with a feedback control quantity based on gravity compensation to form a comprehensive wind resistance control law. Starting from the mesoscale, the disturbance suppression capability and trajectory tracking precision of the unmanned aerial vehicle in complex environments such as strong wind and turbulent flow are enhanced.

Description

Technical Field

[0001] This invention belongs to the field of multi-rotor unmanned aerial vehicle (UAV) control technology, and particularly relates to an adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model. Background Technology

[0002] In recent years, with the rapid development of multi-rotor drone technology, its application in fields such as agricultural plant protection, power line inspection, emergency rescue, and logistics distribution has become increasingly widespread. However, wind field disturbance has become a key technical bottleneck limiting the performance improvement of multi-rotor drones, especially in complex terrain areas such as urban canyons, mountains, and coastal areas, where wind field disturbance has a significant impact on the flight control of multi-rotor drones.

[0003] Current wind-resistant control technology for multi-rotor UAVs still faces several key bottlenecks in practical applications. First, regarding control strategies, the commonly used feedback controllers suffer from significant phase lag, only able to compensate for existing trajectory deviations and unable to respond promptly to rapid dynamic changes in the wind field. Simultaneously, control parameters often lack online adaptive adjustment mechanisms, resulting in insufficient adaptability and decreased control performance in complex and variable wind environments. Second, in composite control architectures, existing feedforward controllers mostly employ fixed-gain designs, failing to effectively adapt to the spatiotemporal non-uniformity and multi-scale turbulence characteristics of the wind field, thus limiting wind resistance in practical applications. Furthermore, existing methods have limitations in wind field modeling. High-fidelity computational fluid dynamics models are computationally intensive and difficult to use for real-time control; while engineering empirical models (such as the Dryden model) lack physical mechanisms and have poor environmental adaptability. The Lattice Boltzmann Method (LBM), as a mesoscale fluid simulation tool, has unique advantages in maintaining physical mechanisms and computational efficiency, but how to effectively integrate it with online flight control of multi-rotor UAVs remains an unsolved problem. Summary of the Invention

[0004] In view of this, the present invention aims to provide a wind field modeling technique and adaptive feedforward control method based on mesoscopic dynamics theory to improve the control accuracy and robustness of multi-rotor UAVs under complex wind fields. Based on the Lattice Boltzmann model, this invention introduces the mesoscopic physical parameter Knudsen number and establishes its mapping relationship with macroscopic wind field parameters, thereby constructing an environment-adaptive mesoscopic-scale wind field model that better reflects real wind field characteristics and retains the physical mechanism of LBM, achieving low computational cost and applicability to real-time control systems. Furthermore, an adaptive gain adjustment mechanism coupling tracking error with the wind field is proposed, overcoming the limitations of linearization in traditional feedforward control, thereby improving the adaptability and robustness of multi-rotor UAVs under wind disturbances.

[0005] To achieve the above objectives, the technical solution created by this invention is implemented as follows: An adaptive feedforward control method for a multi-rotor UAV based on a mesoscopic wind field model includes the following steps: S1: Simulate the motion of virtual particles on discrete lattice points based on the lattice Boltzmann evolution equation, and construct the correlation between Knudsen number and relaxation time based on the BGK model; S2: Based on the mapping relationship between wind field characteristic parameters and Knudsen number, a parameterized characterization of mesoscopic wind field characteristics is achieved; S3: Based on the characteristics of the mesoscopic wind field, the average wind component, turbulence component, gust component and thermal noise component are linearly superimposed to establish the wind speed expression of the real-time wind field model; S4: Based on the wind speed expression of the real-time wind field model, a mesoscopic wind field model is constructed by combining linear prediction methods; S5: Calculate wind resistance based on the predicted wind speed output by the mesoscopic wind field model, obtain adaptive gain parameters through three different dimensions, and obtain feedforward control quantity through wind resistance and adaptive gain parameters. S6: Obtain feedback control quantity with gravity compensation based on proportional-integral-derivative controller; S7: Obtain the wind resistance control law of the multi-rotor UAV based on feedforward control and feedback control, and realize adaptive wind resistance control based on mesoscopic wind field model.

[0006] Furthermore, in step S1, the relationship between the Knudsen number and the relaxation time is as follows: ; ; in, As the baseline relaxation time, For adjustment coefficients, For Knudsen numbers, For relaxation time, Here, T is the Boltzmann constant, and T is the absolute temperature. The effective diameter of an air molecule. Let L be the gas pressure and L be the characteristic length of the multi-rotor UAV.

[0007] Furthermore, in step S2, the wind field characteristic parameters include turbulence intensity. Thermal noise intensity ,in: Turbulence intensity The expression is: ; in, This is the first linear scaling factor; Thermal noise intensity The expression is: ; in, This is the second linear scaling factor.

[0008] Furthermore, in step S3, the expression for the average wind component in the three-dimensional coordinate system is: ; in, It is a constant. For time, , , All are the amplitudes of wind fluctuations. , , All of these are the frequencies of wind fluctuations. Let x be the component of the average wind in the x-direction. Let be the component of the average wind in the y-direction. This represents the component of the average wind in the z-direction; The expression for the turbulent component is: ; in, As the reference turbulence intensity, and To construct turbulent components of different frequencies, For the x-component of turbulence, For the y-component of turbulence, For the z-component of turbulence; Gust weight The expression is: ; in, Indicates the amplitude of gusts. It is a unit vector in direction. For the duration of gusts, The initial time of the gust; Thermal noise component The expression is: ; in, This represents three-dimensional Gaussian white noise with a mean of zero and a variance of 1.

[0009] Furthermore, in step S3, the wind speed expression is: ; in, Let t be the wind speed at time t.

[0010] Furthermore, in step S4, the extrapolated wind speed is calculated based on the linear prediction method: ; ; ;

[0011] in, To predict the time span, It is a first-order difference. It is a second-order difference. It is a third-order difference. The sampling interval is... for Predicted wind speed at any time; The extrapolated wind speed is low-pass filtered to obtain a mesoscopic wind field model. : ; in, This represents the smoothing coefficient of the low-pass filter.

[0012] Furthermore, in step S5, the formula used to calculate wind resistance is as follows: ; in, air density, The drag coefficient, For the characteristic reference area of ​​a multi-rotor unmanned aerial vehicle, To predict wind speed and direction, For wind resistance, For mesoscopic wind field models.

[0013] Furthermore, in step S5, the three factors with different dimensions are the gain factor, the orientation alignment factor, and the learning fine-tuning factor, wherein: Gain Factor The expression is: ; in, For the position tracking error of multi-rotor UAVs, For speed tracking error, , , All are gain values, and , , , All are preset error thresholds; Orientation Alignment Factor The expression is: ; in, Minimum directional alignment factor, , Let be the unit direction vector of the position tracking error. The unit direction vector for predicting wind speed; Learning fine-tuning factor The expression is: ; in, This is the historical tracking error trend vector. This is the unit direction vector of the historical tracking error trend vector. The threshold for a significant trend. The threshold for the alignment between trend and wind direction. , For positive fine-tuning of gain; Adaptive gain parameters for: ; The feedforward control quantity is obtained by using wind resistance and adaptive gain parameters: ; in, Let be the feedforward compensation amount at time t. The sampling period is For the acceleration of the desired trajectory, These are the filter coefficients for the feedforward force channel. , where m is the mass of the drone.

[0014] Furthermore, in step S6, the feedback control quantity... The expression is:

[0015] in, , , These correspond to proportional parameters, differential parameters, and integral parameters. For position tracking error, For speed tracking error, For the z-direction Gravity compensation.

[0016] Furthermore, in step S7, the wind resistance control law of the multi-rotor UAV... The expression is: ; in, For feedback control, This is the feedforward compensation amount.

[0017] Compared with the prior art, the present invention can achieve the following beneficial effects: (1) The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model described in this invention is based on mesoscopic dynamics theory. It introduces the Knudsen number as a coupling criterion and designs a quantitative mapping relationship between mesoscopic physical quantities and key features of macroscopic wind fields, thereby constructing a mesoscopic wind field model that combines physical mechanisms and engineering real-time performance. On this basis, a wind field prediction module is embedded to realize the dynamic expression and prediction of the real environment.

[0018] (2) The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model described in this invention designs an adaptive enhanced feedforward mechanism based on the real-time state of the control system. Robustness is ensured through a gain factor, compensation efficiency is optimized through a direction alignment factor, and slow adaptation is achieved through a learning fine-tuning factor. Thus, it achieves better anti-interference performance and trajectory tracking accuracy than traditional feedforward methods in complex time-varying wind fields. All parameters (such as threshold, gain, and coefficients) are configurable variables and can be tuned according to the specific UAV platform and performance requirements. Attached Figure Description

[0019] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 A flowchart illustrating the adaptive feedforward control method for a multi-rotor UAV based on a mesoscopic wind field model, as described in an embodiment of the present invention. Figure 2 A schematic diagram of horizontal tracking of a six-rotor UAV under outdoor wind disturbance of force 4. Figure 3 This is a schematic diagram of altitude tracking for a six-rotor drone under outdoor wind conditions of level 4. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof.

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0022] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0023] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0024] The invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] like Figure 1 As shown, this invention proposes an adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model, which specifically includes the following steps: S1: Simulate the motion of virtual particles on discrete lattice points based on the lattice Boltzmann evolution equation, and construct the correlation between Knudsen number and relaxation time based on the BGK model; S2: Based on the mapping relationship between wind field characteristic parameters and Knudsen number, a parameterized characterization of mesoscopic wind field characteristics is achieved; S3: Based on the characteristics of the mesoscopic wind field, the average wind component, turbulence component, gust component and thermal noise component are linearly superimposed to establish the wind speed expression of the real-time wind field model; S4: Based on the wind speed expression of the real-time wind field model, a mesoscopic wind field model is constructed by combining linear prediction methods; S5: Calculate wind resistance based on the predicted wind speed output by the mesoscopic wind field model, obtain adaptive gain parameters through three different dimensions, and obtain feedforward control quantity through wind resistance and adaptive gain parameters. S6: Obtain feedback control quantity with gravity compensation based on proportional-integral-derivative controller; S7: Obtain the wind resistance control law of the multi-rotor UAV based on feedforward control and feedback control, and realize adaptive wind resistance control based on mesoscopic wind field model.

[0026] It should be noted that, firstly, the motion of virtual particles on discrete lattice points is simulated based on the lattice Boltzmann evolution equation. Based on the single-relaxation-time BGK model, the correlation between the Knudsen number (characterizing the rarefaction of gas flow) and the relaxation time related to fluid viscosity is designed. Through predefined parameterized correspondence rules, a mapping relationship between macroscopic wind field characteristic parameters and the Knudsen number is established, achieving parameterized characterization of mesoscopic wind field characteristics. The four physically meaningful wind field components—mean wind component, turbulence component, gust component, and thermal noise component—are linearly superimposed, and a modular method is used to establish the wind speed expression for the real-time wind field model (including turbulence component, mean wind component, gust component, and thermal noise component). Based on the wind speed expression of the established real-time wind field model, as well as the prediction time and the smoothing coefficient of the low-pass filter, a mesoscopic wind field model is constructed. Based on the mesoscopic wind field model, air density, drag coefficient, and the characteristic reference area of ​​the multi-rotor UAV, the predicted wind drag generated by the predicted wind speed on the multi-rotor UAV is obtained. A gain factor is designed based on the trajectory tracking position error (the difference between the expected trajectory and the actual trajectory of the UAV) and a preset threshold. A direction alignment factor is determined based on the angle between the tracking position error direction (the ratio of the UAV's tracking position error to its magnitude) and the predicted wind direction. A learning fine-tuning factor is designed based on the historical trajectory error (the average of accumulated trajectory tracking errors). Adaptive gain parameters are obtained based on the gain factor, direction alignment factor, and learning fine-tuning factor. Finally, a feedforward control quantity is obtained from the adaptive gain parameter and the predicted wind resistance to achieve adaptive enhanced feedforward control. A gravity-compensated feedback control quantity is designed based on the trajectory tracking position error, velocity error (the difference between the expected trajectory velocity and the actual trajectory velocity of the UAV), the mass of the UAV, and preset proportional, integral, and derivative coefficients to achieve feedback control of the UAV trajectory tracking. Based on the feedforward control quantity and the feedback control quantity, the wind resistance control law of the UAV is obtained, achieving adaptive wind resistance control based on a mesoscopic wind field model.

[0027] Furthermore, this invention first establishes a mapping relationship between the mesoscopic physical quantity Knudsen number and macroscopic wind field characteristic parameters, determines the mesoscopic wind field model, and then constructs the mesoscopic wind field model to calculate wind drag in real time. Subsequently, an adaptive feedforward mechanism is designed to embed wind drag into the control loop, ultimately achieving adaptive enhanced feedforward control. This significantly improves the dynamic control performance and disturbance rejection capability of multi-rotor UAVs in complex flight environments. The adaptive feedforward control method for multi-rotor UAVs based on the mesoscopic wind field model opens up a new path for refined wind field modeling and provides a new approach for high-precision trajectory tracking control of multi-rotor UAVs, possessing both theoretical reference value and practical application value.

[0028] In some embodiments, in step S1, the relaxation time With Knudsen Related: ; in, As the baseline relaxation time, The Knudsen number, a modulating factor, is the ratio of the molecular mean free path to the characteristic length, further expressed as: ; in, Here, T is the Boltzmann constant, and T is the absolute temperature. The effective diameter of an air molecule. Indicates gas pressure. The characteristic length of a multi-rotor drone.

[0029] In some embodiments, in step S2, the wind field characteristic parameters include turbulence intensity. Thermal noise intensity turbulence intensity Thermal noise intensity They are represented as follows: ; ; in, The first linear proportionality coefficient has the physical meaning that as the Knudsen number increases, the randomness of gas molecule motion increases, which macroscopically manifests as more significant turbulent fluctuations. The second linear scaling factor intuitively reflects the proportional relationship between the intensity of the random velocity of molecular thermal motion and the Knudsen number, thus enabling the wind field model to have a physics-based environmental adaptability.

[0030] Turbulence intensity A dimensionless quantity used to describe the random fluctuations in wind speed. The greater the turbulence intensity, the more violent the wind speed fluctuations, and the greater the disturbance to the drone. At the mesoscale, the random motion of air molecules generates random forces on the drone, which can be considered as a form of thermal noise. Thermal noise intensity. The strength of this random force is described.

[0031] In some embodiments, in step S3, the average wind component is designed as follows: ; in, It is a constant value. Indicates time, , , Indicates the amplitude of wind fluctuations. , , This represents the frequency of wind fluctuations. The average wind component is used to simulate large-scale, slowly changing background wind fields.

[0032] The turbulent components are designed as follows: ; in, As the reference turbulence intensity, and Used to construct turbulent components of different frequencies (designed according to actual needs) to simulate multi-scale turbulence caused by air pulsation.

[0033] Gust components are expressed as: ; in, Indicates the amplitude of gusts. It is a unit vector in direction. The duration of the gust is represented by the event-driven method to simulate sudden, short-duration strong wind events, describing the typical characteristics of rapid gust intensification followed by parabolic decay.

[0034] In addition, the thermal noise components are designed as follows: ; in, This represents a three-dimensional Gaussian white noise process with a mean of zero and a variance of 1, whose amplitude is determined by... This determines the microscopic random perturbations caused by molecular thermal motion.

[0035] In summary, the wind speed is obtained. The expression: ; The synthesized wind speed sequence is then subjected to physical plausibility testing followed by low-pass filtering before being output.

[0036] In some embodiments, in step S4, the extrapolated wind speed is obtained using a linear prediction method: ; in, Indicates the prediction time span, first difference Second-order difference Third-order difference .

[0037] Furthermore, the extrapolated wind speed is low-pass filtered to obtain the predicted wind speed module:

[0038] in, This represents the smoothing coefficient of the low-pass filter. The smaller the value, the smoother the predicted output, but the greater the lag.

[0039] In some embodiments, in step S5, the wind resistance is calculated based on the predicted wind speed output by the mesoscopic wind field model: ; in, Given air density, the drag coefficient is expressed as... It is related to the predicted wind speed. Indicates the initial drag coefficient. As a velocity-sensitive coefficient, it reflects the physical characteristics of the drag coefficient changing with relative velocity, ensuring the rationality of the mesoscopic wind field model at high and low speeds. This represents the characteristic reference area of ​​a multi-rotor unmanned aerial vehicle (UAV). This indicates the direction of the predicted wind speed. It can be seen that the predicted wind resistance is opposite to the predicted wind speed and is the disturbance force of the drone.

[0040] Gain Factor The design is as follows: ; in, For UAV position tracking error, For speed tracking error, , , It is the gain value, and satisfies , , , The preset error threshold is used. The gain factor switches between three gain levels—conservative, aggressive, and medium—based on the magnitude of the current tracking error, aiming to ensure the stability of the UAV system under large tracking errors and the high control accuracy (such as small overshoot and short settling time) under small tracking errors.

[0041] Directional alignment factor based on error direction optimization , means as follows: ; in, Minimum alignment factor Let be the unit direction vector of the position tracking error. This represents the unit direction vector for predicted wind speed.

[0042] By analyzing historical error trends, a learning fine-tuning factor is designed. The following measures are taken to overcome model mismatch and slow time-varying environmental perturbations: ; in, The historical tracking error trend vector is obtained by linearly fitting the slope of the position error over several control cycles. This is the unit direction vector of the historical tracking error trend vector. Indicates the threshold for a significant trend. For the trend and wind direction alignment threshold, the gain is positively fine-tuned. It is a small positive number. In the adaptive enhanced feedforward controller, when the magnitude of the historical tracking error trend vector is greater than the trend significance threshold, when a continuous position error trend is detected and the direction of the position error trend is highly consistent with the predicted wind direction, the learning fine-tuning factor will be slightly greater than 1, thereby slowly enhancing the feedforward compensation to suppress the steady-state deviation.

[0043] Therefore, to comprehensively reflect the real-time status of the UAV tracking and control system, the adaptive gain is designed as follows:

[0044] The adaptive gain is then applied to predict wind drag, and a desired acceleration compensation term is added. Furthermore, to provide the inertial force required for changes in the UAV's motion state and reduce phase lag in trajectory tracking, a desired acceleration compensation term is added to the adaptive enhanced feedforward controller. After smoothing through a first-order low-pass filter, the feedforward control quantity is obtained as follows:

[0045] in, This represents the feedforward compensation force at the current moment. The sampling period of the control system, For the acceleration of the desired trajectory, For the feedforward controller, the filter coefficients are designed to eliminate high-frequency computational noise, making the drone's flight smoother.

[0046] In some embodiments, in step S6, a feedback control quantity with gravity compensation is obtained based on a proportional-integral-derivative controller:

[0047] in, , , These are represented as proportional parameter, differential parameter, and integral parameter, respectively. Indicates position tracking error. Indicates speed tracking error. Indicates the z-direction Gravity compensation.

[0048] In some embodiments, in step S7, by combining the feedback control quantity and the feedforward control quantity, the wind resistance control law of the multi-rotor UAV is expressed as follows:

[0049] The following uses a six-rotor UAV as an example to illustrate the adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model provided by this invention: To verify the robustness of this invention, a hexacopter UAV was used as an example. A trajectory tracking flight experiment was conducted outdoors in level 4 wind conditions. The maximum instantaneous wind speed measured using an Omega anemometer was 6.8 m / s. Code for an adaptive feedforward control method for the multi-rotor UAV based on a mesoscopic wind field model was written and input into the hexacopter's onboard control chip to complete the trajectory tracking flight experiment. The parameters were set as follows: , , ,like Turbulence was added in all three directions of the trajectory: x, y, and z. Gaussian white noise is used to simulate the dramatic velocity fluctuations under extremely rarefied conditions. .

[0050] Figures 2-3 The trajectory tracking results of the hexacopter UAV are described. It can be seen that, based on the feedforward control method provided by this invention, the trajectory tracking error of the hexacopter UAV in all three directions is within ±0.9 meters under level 4 wind disturbance, enabling it to maintain good flight performance even in windy conditions. Experiments show that the method designed in this invention can guarantee excellent trajectory tracking control performance for the hexacopter UAV and resist external level 4 wind disturbance, exhibiting strong robustness.

[0051] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0052] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-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 invention should be included within the scope of protection of this invention.

Claims

1. An adaptive feedforward control method for a multi-rotor unmanned aerial vehicle based on a mesoscopic wind field model, characterized in that, Specifically, the steps include the following: S1: Simulate the motion of virtual particles on discrete lattice points based on the lattice Boltzmann evolution equation, and construct the correlation between Knudsen number and relaxation time based on the BGK model; S2: Based on the mapping relationship between wind field characteristic parameters and Knudsen number, a parameterized characterization of mesoscopic wind field characteristics is achieved; S3: Based on the characteristics of the mesoscopic wind field, the average wind component, turbulence component, gust component and thermal noise component are linearly superimposed to establish the wind speed expression of the real-time wind field model; S4: Based on the wind speed expression of the real-time wind field model, a mesoscopic wind field model is constructed by combining linear prediction methods; S5: Calculate wind resistance based on the predicted wind speed output by the mesoscopic wind field model, obtain adaptive gain parameters through three different dimensions, and obtain feedforward control quantity through wind resistance and adaptive gain parameters. S6: Obtain feedback control quantity with gravity compensation based on proportional-integral-derivative controller; S7: Obtain the wind resistance control law of the multi-rotor UAV based on feedforward control and feedback control, and realize adaptive wind resistance control based on the mesoscopic wind field model.

2. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S1, the relationship between the Knudsen number and the relaxation time is as follows: ； ； in, As the baseline relaxation time, For adjustment coefficients, For Knudsen numbers, For relaxation time, Here, T is the Boltzmann constant, and T is the absolute temperature. The effective diameter of an air molecule. Let L be the gas pressure and L be the characteristic length of the multi-rotor UAV.

3. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S2, the wind field characteristic parameters include turbulence intensity. Thermal noise intensity ,in: Turbulence intensity The expression is: ； in, This is the first linear scaling factor; Thermal noise intensity The expression is: ； in, This is the second linear scaling factor.

4. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S3, the expression for the average wind component in the three-dimensional coordinate system is: ； in, It is a constant. For time, , , All are the amplitudes of wind fluctuations. , , All of these are frequencies of wind fluctuations. Let x be the component of the average wind in the x-direction. Let be the average wind component in the y-direction. This represents the component of the average wind in the z-direction; The expression for the turbulent component is: ； in, As the reference turbulence intensity, and To construct turbulent components of different frequencies, For the x-component of turbulence, For the y-component of turbulence, For the z-component of turbulence; Gust weight The expression is: ； in, Indicates the amplitude of gusts. It is a unit vector in direction. For the duration of gusts, The initial time of the gust; Thermal noise component The expression is: ； in, This represents three-dimensional Gaussian white noise with a mean of zero and a variance of 1.

5. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 4, characterized in that: In step S3, the wind speed expression is: ； in, Let t be the wind speed at time t.

6. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 5, characterized in that: In step S4, the extrapolated wind speed is calculated based on the linear prediction method: ； ； ； in, To predict the time span, It is a first-order difference. It is a second-order difference. It is a third-order difference. The sampling interval is... for Predicted wind speed at any time; The extrapolated wind speed is low-pass filtered to obtain a mesoscopic wind field model. : ； in, This represents the smoothing coefficient of the low-pass filter.

7. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S5, the formula used to calculate wind resistance is: ； in, air density, The drag coefficient, For the characteristic reference area of ​​a multi-rotor unmanned aerial vehicle, To predict wind speed and direction, For wind resistance, For mesoscopic wind field models.

8. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S5, the three factors with different dimensions are the gain factor, the orientation alignment factor, and the learning fine-tuning factor, where: Gain Factor The expression is: ； in, For the position tracking error of multi-rotor UAVs, For speed tracking error, , , All are gain values, and , , , All are preset error thresholds; Orientation Alignment Factor The expression is: ； in, Minimum directional alignment factor, , Let be the unit direction vector of the position tracking error. The unit direction vector for predicting wind speed; Learning fine-tuning factor The expression is: ； in, This is the historical tracking error trend vector. This is the unit direction vector of the historical tracking error trend vector. The threshold for a significant trend. The threshold for the alignment between trend and wind direction. , For positive fine-tuning of gain; Adaptive gain parameters for: ； The feedforward control quantity is obtained by using wind resistance and adaptive gain parameters: ； in, Let be the feedforward compensation amount at time t. The sampling period is For the acceleration of the desired trajectory, These are the filter coefficients for the feedforward force channel. , where m is the mass of the drone.

9. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S6, the feedback control quantity The expression is: in, , , These correspond to proportional parameters, differential parameters, and integral parameters. For position tracking error, For speed tracking error, For the z-direction Gravity compensation.

10. The adaptive feedforward control method for multi-rotor UAVs based on a mesoscopic wind field model according to claim 1, characterized in that: In step S7, the wind resistance control law of the multi-rotor UAV... The expression is: ； in, For feedback control, This is the feedforward compensation amount.

Images