A method, device and equipment for coordinating control of height and heading of a UAV

Abstract

The application discloses a kind of unmanned aerial vehicle height and heading coordination control method, device and equipment, the flight state parameter of unmanned aerial vehicle is preprocessed and feature extraction, constructs height-heading state space mapping relationship;Through frequency domain conversion and resonance mode analysis identifies coupling resonance characteristics;Establish decoupling rule base and generate dynamic decoupling strategy, form independent control channel;Analysis interactive relationship between channel and design coordination mechanism, obtain coordination constraint parameter;Based on constraint parameter constructs fractal geometry structure, creates fractal recursive control architecture by recursive level design, generates guide matrix by analyzing coupling energy flow direction;Predict coupling effect using guide matrix and generate reverse compensation sequence, output coordination control instruction by multilevel optimization;Finally, signal modulation and power distribution are carried out, realize the accurate coordination control of height and heading.The application solves the dynamic coupling problem that traditional method is difficult to handle, improves the control precision and stability of complex maneuver of unmanned aerial vehicle.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) flight control technology, and in particular to a method, apparatus, and equipment for coordinated control of UAV altitude and heading. Background Technology

[0002] With the rapid development of UAV technology and the continuous expansion of its application fields, the requirements for the accuracy and stability of UAV flight control are increasing. In complex flight missions, altitude control and heading control are two core control dimensions of UAVs, and their coupling problem seriously affects flight performance. Traditional control methods usually design altitude and heading as independent control loops, ignoring the dynamic coupling relationship between them, resulting in adverse phenomena such as oscillation and overshoot when performing compound maneuvers such as climb-turn and spiral-descent.

[0003] Existing decoupling control techniques are mainly based on linearized models and fixed parameter designs, making it difficult to adapt to the changing coupling characteristics of UAVs under different flight states. Furthermore, traditional methods lack in-depth analysis of the coupling energy flow, failing to achieve multi-level coordinated control, and resulting in a significant decrease in control effectiveness when facing complex environmental disturbances. Therefore, there is an urgent need to develop an intelligent control method capable of deeply analyzing the coupling mechanism between altitude and heading, achieving dynamic decoupling and multi-level coordination. Summary of the Invention

[0004] This invention provides a method, apparatus, and device for coordinated control of altitude and heading of a UAV based on a fractal recursive architecture. It aims to construct an altitude-heading state space mapping relationship, deeply analyze the coupling and resonance characteristics of the two, establish a dynamic decoupling rule base and a fractal recursive control architecture, achieve accurate prediction and compensation of coupling energy, and ultimately achieve precise coordinated control of altitude and heading, significantly improving the control performance and robustness of UAVs in complex flight missions.

[0005] The first aspect of this invention proposes a method for coordinated control of altitude and heading of an unmanned aerial vehicle (UAV), comprising the following steps: preprocessing and extracting features from the flight state parameters of the UAV to generate standardized state parameters; performing topological transformation on the standardized state parameters to construct an altitude-heading state space mapping relationship.

[0006] The altitude-heading state space mapping relationship is transformed in the frequency domain to generate frequency domain feature data. Frequency domain resonance mode analysis is performed on the frequency domain feature data to identify the coupling resonance characteristics of altitude and heading control.

[0007] A decoupling rule base is established based on the coupling resonance characteristics, and a dynamic decoupling strategy is generated based on the decoupling rule base. Independent altitude control channels and heading control channels are formed through the dynamic decoupling strategy.

[0008] An interaction relationship analysis is performed on the altitude control channel and the heading control channel to determine the channel interaction characteristics. Based on the channel interaction characteristics, a coordination mechanism is designed to obtain the coordination constraint parameters between channels.

[0009] A fractal geometric structure is constructed based on the coordination constraint parameters. A recursive hierarchical design is performed based on the fractal geometric structure to create a fractal recursive control architecture with self-similar properties. The coupling energy flow direction of each level is analyzed through the fractal recursive control architecture to generate a coupling energy guidance matrix.

[0010] Based on the coupled energy steering matrix, the coupled prediction value at each control moment is obtained. A reverse compensation sequence is generated according to the coupled prediction value. The reverse compensation sequence is allocated to each level through the fractal recursive control architecture to generate compensation allocation results. Based on the compensation allocation results, multi-level control strategy optimization is performed to output coordinated control commands.

[0011] The system performs signal modulation and power allocation according to the coordinated control command, generates a modulation control signal, and outputs an altitude control signal and a heading control signal through the modulation control signal to achieve coordinated control of the UAV's altitude and heading.

[0012] A second aspect of the present invention provides a drone altitude and heading coordination control device, comprising:

[0013] The state processing module is used to preprocess and extract features from the flight state parameters of the UAV, generate standardized state parameters, perform topological transformation processing on the standardized state parameters, and construct an altitude-heading state space mapping relationship.

[0014] The resonance analysis module is used to perform frequency domain transformation on the altitude-heading state space mapping relationship, generate frequency domain feature data, perform frequency domain resonance mode analysis on the frequency domain feature data, and identify the coupling resonance characteristics of altitude and heading control.

[0015] The channel establishment module is used to establish a decoupling rule base based on the coupling resonance characteristics, generate a dynamic decoupling strategy based on the decoupling rule base, and form independent altitude control channels and heading control channels through the dynamic decoupling strategy.

[0016] The coordination design module is used to analyze the interaction relationship between the altitude control channel and the heading control channel, determine the channel interaction characteristics, design a coordination mechanism based on the channel interaction characteristics, and obtain the coordination constraint parameters between the channels.

[0017] The architecture construction module is used to construct a fractal geometric structure based on the coordination constraint parameters, perform recursive hierarchical design based on the fractal geometric structure, create a fractal recursive control architecture with self-similar properties, and analyze the coupling energy flow direction of each level through the fractal recursive control architecture to generate a coupling energy guidance matrix.

[0018] The coordination processing module is used to obtain the coupling prediction value at each control moment based on the coupling energy steering matrix, generate a reverse compensation sequence according to the coupling prediction value, allocate the reverse compensation sequence to each level through the fractal recursive control architecture to generate compensation allocation results, and perform multi-level control strategy optimization based on the compensation allocation results to output coordinated control commands.

[0019] The signal output module is used to perform signal modulation and power distribution according to the coordinated control command, generate a modulation control signal, and output an altitude control signal and a heading control signal through the modulation control signal to realize the coordinated control of the UAV's altitude and heading.

[0020] A third aspect of the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a UAV altitude and heading coordinated control method disclosed in the first aspect.

[0021] The beneficial effects of this invention are reflected in the following points: First, by constructing an altitude-heading state space mapping relationship and performing frequency domain resonance analysis, coupled resonance frequency pairs under different flight states are identified. Based on this, a targeted decoupling rule base is established, realizing dynamic decoupling control based on coupling characteristics. This solves the problem that fixed parameter decoupling methods cannot adapt to changes in flight states, enabling the UAV to maintain stability when performing complex maneuvers such as spiral climbs and S-shaped maneuvers. Second, a fractal recursive control architecture is adopted to decompose the control task into multiple levels according to the self-similarity principle. The top level is responsible for macro-strategy formulation, the middle level performs task decomposition and coordination, and the bottom level performs precise control. Each level realizes information transmission and energy allocation through recursive relationships and energy guidance matrices, forming a control system that is both clearly defined and coordinated, improving the system's ability to handle complex tasks. Third, a predictive compensation mechanism based on the coupled energy guidance matrix is ​​established. By analyzing the energy flow pattern between each level, future coupling effects are predicted, and reverse compensation sequences are generated in advance and allocated to each level, realizing the active cancellation of coupling effects and avoiding the lag and oscillation problems caused by passive response. Finally, the coordination constraint parameters established through interactive characteristic analysis ensured that the altitude and heading control channels maintained independence while achieving necessary information exchange and coordination. This solved the problems of poor coordination caused by completely independent control and mutual interference caused by completely coupled control, providing a complete technical solution for the precise control of UAVs in complex environments.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0023] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0024] Unless otherwise specified or defined, the same reference numerals in different figures represent the same or similar technical features, and different reference numerals may be used to represent the same or similar technical features.

[0025] Figure 1 This is a flowchart illustrating a method for coordinated control of altitude and heading of an unmanned aerial vehicle (UAV) according to the present invention.

[0026] Figure 2 This is a structural block diagram of a drone altitude and heading coordination control device according to the present invention.

[0027] Figure 3 This is a schematic diagram of the structure of a computer device according to the present invention. Detailed Implementation

[0028] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0029] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0030] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0031] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0032] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0033] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0034] The technical solutions of the embodiments of this application will be described below.

[0035] like Figure 1As shown, this embodiment of the invention provides a method for coordinated control of altitude and heading of an unmanned aerial vehicle (UAV), including the following steps S110-S170:

[0036] Step S110: Perform data preprocessing and feature extraction on the flight state parameters of the UAV to generate standardized state parameters. Perform topological transformation on the standardized state parameters to construct an altitude-heading state space mapping relationship.

[0037] Specifically, the UAV collects raw flight state parameters through a multi-sensor system. The GPS navigation module provides three-dimensional position coordinates, the inertial measurement unit (IMU) outputs attitude angle and angular velocity data, the barometric altimeter measures barometric altitude, the magnetic compass provides magnetic heading angle, the pitot tube detects relative airspeed, and control command parameters from the flight control system are collected. All sensor data are collected synchronously at a sampling frequency of 100Hz and transmitted to the flight control computing unit via CAN bus for centralized processing. Multi-level data preprocessing is performed on the collected raw flight state parameters. First, data cleaning is performed, using the 3-sigma criterion to identify and remove outlier data points. The anomaly detection condition is |x_i - μ|>3σ, where x_i is the data at the i-th sampling point, μ is the data mean, and σ is the standard deviation. For example, if a UAV's GPS suddenly reports an obvious error with an altitude of -500 meters during normal flight, such outliers are automatically detected and removed. Parameters whose deviation from the mean by more than three times the standard deviation for three consecutive sampling points are corrected by interpolation. Time synchronization is implemented. Since the GPS update frequency is 10Hz and the IMU update frequency is 100Hz, a timestamp interpolation algorithm is used to align sensor data from different sampling frequencies to a unified time reference. The interpolation formula is x(t) = x_1 + (x_2 - x_1) × (t - t_1) / (t_2 - t_1), where x_1 and x_2 are data values ​​at adjacent sampling times, t_1 and t_2 are corresponding times, and t is the interpolation time, ensuring time consistency of flight status parameters at the same moment. Digital filtering is performed. Kalman filtering is used for position coordinates to reduce multipath interference noise of GPS signals in urban environments. Complementary filtering algorithms are used to fuse IMU and magnetic compass data for attitude angles and angular velocities to reduce the impact of sensor drift. Low-pass filters are used for acceleration signals to eliminate high-frequency interference such as propeller vibration. Data integrity checks are performed. When a sensor malfunctions and data is missing, linear interpolation or prediction algorithms based on historical flight trajectories are used to complete the data, ensuring the continuity and integrity of flight status parameters. Multidimensional feature extraction is performed based on the preprocessed flight state parameters. In the time domain, the statistical characteristics of each parameter, such as mean, variance, maximum and minimum values, are calculated within a 1-second sliding window. In the frequency domain, fast Fourier transform is performed on the attitude angle and angular velocity signals to extract the main frequency components and power spectral density features. The correlation coefficients between different flight parameters, such as the correlation between altitude and vertical velocity, and the correlation between heading angle and yaw rate, are analyzed. The derived features of velocity and acceleration are obtained by calculating the first and second derivatives of the position coordinates through numerical differentiation.The extracted multidimensional features are standardized using the Z-score standardization method to unify parameters with different dimensions and numerical ranges into a standard normal distribution space. The standardization formula is z = (x - μ) / σ, where x is the original data, μ is the mean of the parameter, σ is the standard deviation, and z is the standardized value. For example, the original altitude data of 500 meters is converted to a standardized value of 1.2, and the original pitch angle of 15 degrees is converted to a standardized value of 0.8. This eliminates the weight bias caused by the difference in dimensions between parameters. Position parameters are standardized using meters as the base unit, angle parameters are uniformly converted to radians, and velocity and acceleration are standardized using meters per second and meters per second squared, respectively. At the same time, a quality assessment label is established for each standardized parameter to mark the data's credibility level and processing status, and finally, standardized status parameters are generated.

[0038] In some embodiments, the step of performing topological transformation processing on the standardized state parameters to construct an altitude-heading state space mapping relationship includes: extracting the altitude dimension from the standardized state parameters to form an altitude state vector; extracting the heading dimension from the standardized state parameters to form a heading state vector; performing coordinate system transformation on the altitude state vector and the heading state vector to generate a unified coordinate system state vector; performing topological space projection on the unified coordinate system state vector to generate topological space coordinates; and establishing an altitude-heading state space mapping relationship based on the topological space coordinates.

[0039] First, the altitude dimension is extracted from the generated standardized state parameters. From these parameters, altitude-related components are selected, primarily including standardized barometric altitude, GPS altitude, vertical velocity, vertical acceleration, pitch angle, pitch rate, and throttle output. These altitude-related standardized parameters are arranged in a time series; for example, at a certain moment, the standardized barometric altitude is 1.2, the standardized vertical velocity is -0.3, the standardized pitch angle is 0.8, and the standardized throttle output is 0.5. These values ​​are combined to form a high-dimensional vector, creating the altitude state vector.

[0040] Then, the heading dimension is extracted from the standardized state parameters, and the parameter components related to heading control are selected, mainly including the standardized magnetic heading angle, GPS heading angle, yaw rate, lateral acceleration, rudder deflection angle, and roll angle. These heading-related standardized parameters are combined. For example, at a certain moment, the standardized magnetic heading angle is 0.6, the standardized yaw rate is -0.2, the standardized rudder deflection angle is 0.4, and the standardized roll angle is -0.1. These values ​​are arranged and combined according to their physical meaning to form the heading state vector.

[0041] Next, coordinate system transformation is performed on the altitude and heading state vectors. Since the two vectors may have different reference coordinate systems and dimensional bases, the transformation is performed using a rotation matrix R, where R = [cosθ -sinθ 0; sinθcosθ 0; 0 0 1] is the transformation matrix for rotating around the z-axis by an angle θ. The altitude state vector is transformed from the body coordinate system to the geographic coordinate system, and the heading state vector is transformed from the magnetic coordinate system to the true north coordinate system. Then, the transformed altitude and heading state vectors are fused according to a unified coordinate system base. For example, the transformed altitude and heading vectors are concatenated in the order of altitude first and then heading to generate a unified coordinate system state vector.

[0042] Subsequently, a topological space projection is performed on the unified coordinate system state vector. A nonlinear mapping method is used to project the high-dimensional unified coordinate system state vector into a low-dimensional topological space. Principal component analysis is used to identify the main directions of change in the state vector. The projection formula is Y = X × W, where X is the original state vector matrix, W is the principal component projection matrix, and Y is the projected low-dimensional representation. The original multidimensional state vector is compressed into a two-dimensional or three-dimensional topological space representation. For example, an 8-dimensional unified coordinate system state vector is mapped to a point in 3-dimensional space through a topological transformation. The coordinates of this point reflect the current altitude-heading integrated state of the UAV, generating topological space coordinates.

[0043] Finally, a state-space mapping relationship between altitude and heading is established based on the topological space coordinates. By analyzing the distribution pattern of topological space coordinates under different flight states, the coupling relationship and mutual influence mode between altitude control state and heading control state are identified. For example, when the UAV performs a climb maneuver, the topological space coordinates show a specific distribution pattern in the first quadrant, while when performing a turn maneuver, the coordinates show another distribution feature in the second quadrant. Through this mapping relationship, the dynamic correlation between altitude and heading control can be observed and analyzed intuitively, and finally, an altitude-heading state-space mapping relationship is constructed.

[0044] Step S120: Perform frequency domain transformation on the altitude-heading state space mapping relationship to generate frequency domain feature data, perform frequency domain resonance mode analysis on the frequency domain feature data, and identify the coupling resonance characteristics of altitude and heading control.

[0045] Specifically, a Fast Fourier Transform (FFT) is first applied to the altitude-heading state-space mapping relationship. The transformation formula is X(k) = Σ[n=0 to N-1] x(n)e^(-j2πkn / N), where x(n) is the time-domain signal, X(k) is the frequency-domain representation, and N is the number of sampling points. This converts the time-domain representation of the mapping relationship into a frequency-domain representation, identifying the response characteristics and oscillation modes of the mapping relationship at different frequency points. For example, when the UAV performs a standard spiral climb maneuver, the altitude-heading state-space mapping relationship exhibits a significant amplitude response reaching a peak value of 2.8 at the 0.3Hz frequency point, a significant phase shift of 45 degrees at the 1.2Hz frequency point, and strong harmonic components accompanied by an amplitude attenuation of 0.6 at the 2.5Hz frequency point. Then, power spectral density analysis was performed on the altitude-heading state space mapping relationship. The power spectral density calculation formula is PSD(f) = |X(f)|² / T, where X(f) is the frequency domain signal and T is the signal duration. The main frequency component distribution and energy concentration areas of the mapping relationship were identified. It was found that the mapping relationship in the low frequency band of 0.1-0.5Hz showed a smooth change and the power density remained at a low level. In the mid frequency band of 0.5-2.0Hz, the mapping relationship showed periodic fluctuations and the power density increased significantly to the peak range. In the high frequency band of 2.0-5.0Hz, the mapping relationship showed rapid oscillation characteristics, but the power density gradually decreased. Next, a short-time Fourier transform (STFT) was used to perform a time-frequency joint analysis of the altitude-heading state-space mapping relationship. The STFT formula is X(t,f) = ∫x(τ)w(τ-t)e^(-j2πfτ)dτ, where w(τ-t) is a window function. The evolution of frequency characteristics and transient changes of the mapping relationship within different time windows were observed. For example, in the initial 5 seconds of complex maneuvering flight, the dominant frequency of the mapping relationship was concentrated around 1.0Hz and the amplitude steadily increased. In the middle 10 seconds of maneuvering flight, the dominant frequency gradually migrated to 1.8Hz with amplitude fluctuations. In the last 5 seconds of maneuvering flight, the dominant frequency dropped back to 0.8Hz while the amplitude tended to stabilize. Subsequently, wavelet transform was used to perform multi-scale frequency domain decomposition of the altitude-heading state-space mapping relationship to obtain fine local features and multi-resolution representations of the mapping relationship at different frequency scales. It was identified that the transient frequency components in the mapping relationship were mainly distributed in the high-frequency band, reflecting rapid control response, while the steady-state frequency components were mainly concentrated in the low-frequency band, reflecting the basic dynamic characteristics of the system. Further adaptive frequency domain filtering is applied to the altitude-heading state space mapping relationship to extract mapping features and key frequency information within a specific frequency band. For example, a low-pass filter with a cutoff frequency of 0.5 Hz is designed to extract the low-frequency trend features of the mapping relationship, a band-pass filter with a center frequency of 1.5 Hz and a bandwidth of 2.0 Hz is applied to extract the mid-frequency periodic features of the mapping relationship, and a high-pass filter with a cutoff frequency of 3.0 Hz is used to extract the high-frequency oscillation features of the mapping relationship.Simultaneously, the nonlinear characteristics and harmonic distortion phenomena of the altitude-heading state space mapping relationship in the frequency domain are analyzed, and the total harmonic distortion rate is calculated. , where V_n is the amplitude of the nth harmonic, V_1 is the amplitude of the fundamental wave, and the nonlinearity metric NL = Σ|Y_actual - Y_linear| / Σ|Y_linear|, where Y_actual is the actual output and Y_linear is the predicted output of the linear model, to evaluate the linearity and distortion of the mapping relationship in different frequency bands. Finally, frequency domain correlation analysis was performed on the altitude-heading state space mapping relationship. The coherence function γ²(f) = |G_xy(f)|² / (G_xx(f)×G_yy(f)) between different frequency components was calculated, where G_xy(f) is the cross power spectral density, G_xx(f) and G_yy(f) are the self power spectral densities, the coherence function value range is [0,1], and the phase relationship φ(f) = arg(G_xy(f)) was also calculated. Frequency coupling phenomena and resonance characteristics in the mapping relationship were identified. All frequency domain analysis results were comprehensively processed and feature fused to form a complete dataset describing the frequency domain characteristics of the altitude-heading state space mapping relationship, generating frequency domain feature data.

[0046] In some embodiments, the step of performing frequency domain resonance mode analysis on the frequency domain feature data to identify the coupled resonance features of altitude and heading control includes: performing spectral analysis on the frequency domain feature data to extract the altitude control spectrum and the heading control spectrum; performing cross-correlation analysis on the altitude control spectrum and the heading control spectrum to obtain frequency domain correlation coefficients; identifying resonance frequency points based on the frequency domain correlation coefficients to determine multiple candidate resonance frequencies; performing amplitude and phase analysis on the multiple candidate resonance frequencies to obtain resonance mode parameters; and generating coupled resonance features of altitude and heading control based on the resonance mode parameters.

[0047] First, spectral decomposition analysis was performed on the generated frequency domain feature data to extract spectral components related to altitude control and heading control. Based on the feature information in the low-frequency band of 0.1-0.5Hz in the frequency domain feature data, the spectral components mainly reflecting the response of the altitude control system were extracted. This spectrum shows that altitude control has obvious resonance peaks at 0.2Hz and 0.4Hz, with amplitudes reaching 3.2 and 2.8, respectively, forming the altitude control spectrum. Simultaneously, from the feature information in the mid-frequency band of 0.5-2.0Hz in the frequency domain feature data, the spectral components mainly reflecting the characteristics of the heading control system were extracted. This spectrum shows significant energy concentration at 0.8Hz, 1.2Hz, and 1.6Hz, with amplitudes of 2.5, 3.1, and 2.3, respectively, forming the heading control spectrum. Then, cross-correlation analysis was performed on the extracted altitude control spectrum and heading control spectrum. The cross-correlation function was R_xy(τ) = E[x(t)y(t+τ)], where x(t) is the altitude control spectrum, y(t) is the heading control spectrum, and τ is the time delay. The correlation and mutual influence strength of the two spectra at different frequency points were calculated. The cross-correlation function calculation revealed a strong correlation between the peak value of the altitude control spectrum at 0.2 Hz and the peak value of the heading control spectrum at 0.8 Hz, with a correlation coefficient of 0.76. The correlation coefficient between the component of the altitude control spectrum at 0.4 Hz and the component of the heading control spectrum at 1.2 Hz was 0.68, yielding the frequency domain correlation coefficients. Next, the resonant frequency points of the system were identified based on the calculated frequency domain correlation coefficients. By setting a correlation coefficient threshold of 0.6, frequency combinations exceeding this threshold were selected as potential resonant frequencies. Analysis revealed that the heading control resonance frequency was 0.8Hz when the altitude control frequency was 0.2Hz, and 1.2Hz when the altitude control frequency was 0.4Hz. A significant coupling resonance phenomenon occurred when the frequency ratio of the two systems was 1:4. The ranges of 0.2Hz-0.8Hz, 0.4Hz-1.2Hz, and 0.3Hz-1.0Hz were identified as multiple candidate resonance frequencies. Detailed amplitude and phase analyses were then performed on these candidate resonance frequencies, calculating the oscillation amplitude, attenuation characteristics, and phase relationship at each frequency point. In the candidate resonant frequency pair of 0.2Hz-0.8Hz, the oscillation amplitude of the altitude control component is 3.2 and the phase φ_h = 15 degrees, while the oscillation amplitude of the heading control component is 2.5 and the phase φ_a = -30 degrees. The phase difference Δφ = φ_h - φ_a = 45 degrees indicates significant coupled oscillation. In the candidate resonant frequency pair of 0.4Hz-1.2Hz, the amplitude of the altitude control component is 2.8 and the phase is -10 degrees, while the amplitude of the heading control component is 3.1 and the phase is 25 degrees. The phase difference of 35 degrees also shows strong coupling characteristics. Resonance mode parameters were obtained.Finally, based on the obtained resonance mode parameters, the coupling mechanism and oscillation modes between the altitude and heading control systems were analyzed. It was found that when the system operates at the 0.2Hz-0.8Hz resonance frequency pair, changes in altitude control trigger a response in heading control after 0.125 seconds, resulting in periodic mutual excitation oscillations. When the system operates at the 0.4Hz-1.2Hz resonance frequency pair, adjustments in heading control affect the stability of altitude control after 0.083 seconds, producing a coupled resonance amplification effect. By identifying these coupled resonance characteristics, undesirable oscillations in the control system at specific frequencies can be effectively avoided, achieving decoupling optimization of altitude and heading control. Through parameter analysis and pattern recognition of all candidate resonance frequencies, a complete feature description was established, describing the coupled oscillation law, resonance frequency distribution, oscillation amplitude characteristics, and phase relationship of the altitude and heading control system, identifying the coupled resonance characteristics of altitude and heading control.

[0048] Step S130: Establish a decoupling rule base based on the coupling resonance characteristics, generate a dynamic decoupling strategy based on the decoupling rule base, and form independent altitude control channel and heading control channel through the dynamic decoupling strategy.

[0049] Specifically, firstly, low-frequency decoupling rules are established for the 0.2Hz-0.8Hz resonant frequency pair in the coupled resonance feature. Utilizing the characteristic that altitude control changes trigger a heading control response after 0.125 seconds with a 45-degree phase difference, a corresponding compensation decoupling strategy is formulated: when the altitude control command frequency is detected to be close to 0.2Hz in the coupled resonance feature, a compensation signal with an amplitude 0.8 times the original signal and a phase lead of 45 degrees is automatically introduced into the heading control loop to counteract the coupling effect described in the coupled resonance feature. Then, mid-frequency decoupling rules are established for the 0.4Hz-1.2Hz resonant frequency pair in the coupled resonance feature. Considering that heading control adjustments in this feature affect altitude control stability after 0.083 seconds with a 35-degree phase difference, a fast-response decoupling mechanism is designed: when the heading control command frequency approaches 1.2Hz in the coupled resonance feature, a feedforward compensation signal with an amplitude 0.9 times the original signal and a phase lag of 35 degrees is pre-injected into the altitude control loop to eliminate interference effects in the coupled resonance feature in advance. Next, a mid-band decoupling rule was established for the 0.3Hz-1.0Hz resonant frequency pair in the coupled resonance characteristics. The coupling strength exhibited by this frequency pair in the coupled resonance characteristics is between the previous two. The coupling strength C_strength = |R_xy(f)|×A_ratio, where R_xy(f) is the frequency domain correlation coefficient and A_ratio is the amplitude ratio. An adaptive decoupling strategy was adopted: the compensation signal was dynamically adjusted according to the oscillation amplitude information in the coupled resonance characteristics. When C_strength < 0.5, the compensation amplitude was set to 0.6 times; when the coupled resonance characteristics showed strong coupling strength, the compensation amplitude was increased to 1.1 times. Subsequently, a frequency ratio decoupling rule was established. Based on the 1:4 frequency ratio pattern found in the coupled resonance characteristics, a frequency avoidance strategy was formulated: when the ratio of the altitude control frequency to the heading control frequency is close to 1:4 in the coupled resonance characteristics, the operating frequency of one of the control loops is automatically adjusted to make the ratio deviate from the dangerous area in the coupled resonance characteristics. Further, a phase difference compensation rule is established. Based on the phase relationship of different frequency pairs in the coupled resonance characteristics, a phase decoupling algorithm is designed: by introducing an adjustable phaser into the control signal, the phase adjustment amount Δφ_adj = 90° - (φ_h - φ_a), where φ_h is the altitude control phase and φ_a is the heading control phase, the phase difference is adjusted in real time to maintain an orthogonal state of approximately 90 degrees to minimize the mutual influence in the coupled resonance characteristics. Simultaneously, an amplitude limiting rule is established. Based on the oscillation amplitude information in the coupled resonance characteristics, an amplitude constraint is set for the control signal: when the operating frequency is close to the resonant frequency in the coupled resonance characteristics, the amplitude of the control signal is automatically limited to no more than 0.7 times the normal value to prevent instability caused by the amplification effect in the coupled resonance characteristics.Finally, dynamic monitoring rules were established, and a real-time spectrum monitoring mechanism was designed using the response time information in the coupled resonance characteristics: the operating frequency is detected every 0.05 seconds, and when a frequency drift to the resonance region in the coupled resonance characteristics is detected, the corresponding decoupling rule is immediately activated to ensure the timeliness and effectiveness of decoupling control. By integrating these decoupling strategies targeting different coupling modes and frequency characteristics in the coupled resonance characteristics, a complete decoupling rule base was established.

[0050] In some embodiments, generating a dynamic decoupling strategy based on the decoupling rule base includes: performing rule matching analysis on the decoupling rule base to determine applicable decoupling rules; configuring strategy parameters according to the applicable decoupling rules to obtain decoupling strategy parameters; and generating a dynamic decoupling strategy based on the decoupling strategy parameters.

[0051] First, rule matching analysis is performed on the established decoupling rule base. Based on the current flight status and control frequency characteristics of the UAV, the most suitable decoupling rules for the current operating conditions are selected from the decoupling rule base. By real-time monitoring of the UAV's altitude control frequency and heading control frequency, when the altitude control frequency is detected to be 0.21Hz and the heading control frequency to be 0.83Hz, a matching is performed with the 0.2Hz-0.8Hz resonant frequency pair in the decoupling rule base. The matching degree is... Where f_h and f_a are the altitude and heading frequencies, respectively, the subscript actual represents the actual value, and rule represents the rule value. The calculated matching degree is 95%, confirming the low-frequency decoupling rule as an applicable decoupling rule. When the heading control frequency reaches 1.18Hz and the altitude control frequency is 0.39Hz, it matches the 0.4Hz-1.2Hz resonant frequency pair in the decoupling rule library, achieving a matching accuracy of 97%, confirming the mid-frequency decoupling rule as an applicable decoupling rule. When the ratio of the two control frequencies r = f_h / f_a = 1 / 3.8 ≈ 0.263, it approximates the 1:4 frequency ratio pattern in the decoupling rule library (standard value r_ref = 0.25), achieving a matching degree of 95%, confirming the frequency ratio decoupling rule as an applicable decoupling rule.

[0052] Then, strategy parameters are configured according to the determined applicable decoupling rules, and specific parameter settings for the corresponding rules are extracted from the decoupling rule library. For applicable low-frequency decoupling rules, core parameters such as the compensation signal amplitude coefficient of 0.8, the phase lead angle of 45 degrees, and the response delay time of 0.125 seconds are extracted from the decoupling rule library. The amplitude coefficient of 0.8 is set based on statistical analysis of a large amount of flight data, which can effectively offset the coupling effect without causing overcompensation; the phase lead angle of 45 degrees is the optimal compensation angle determined according to the phase characteristics of low-frequency coupling; the response delay of 0.125 seconds corresponds to the typical time for altitude control changes to be transmitted to heading control. In addition to the core parameters, supplementary parameters such as the rise time of the compensation signal of 0.02 seconds, the steady-state error tolerance of 2%, and the compensation duration window of 0.5 seconds are also extracted. These parameters together constitute a complete low-frequency decoupling configuration. For example, in long-distance cruise flight, when the UAV needs to climb slowly, these parameters ensure that altitude adjustments do not cause periodic oscillations in heading. For applicable mid-frequency decoupling rules, configuration parameters such as a feedforward compensation amplitude coefficient of 0.9, a phase lag angle of 35 degrees, and a response time of 0.083 seconds are obtained from the decoupling rule base. The feedforward compensation coefficient of 0.9 is slightly higher than in the low-frequency case, reflecting the strength characteristics of mid-frequency coupling; the phase lag of 35 degrees matches the phase of the influence of heading control on altitude control; and the response time of 0.083 seconds reflects the rapid characteristics of mid-frequency dynamics. Supporting parameters also include an adaptive adjustment range of 0.7-1.1 for the feedforward gain, a dynamic range of ±10 degrees for phase correction, and a bandwidth limit of 0.5-2.0 Hz for the compensation signal. These parameters play a crucial role when the UAV performs complex maneuvers such as S-shaped maneuvers or spiral descents, ensuring coordinated changes in altitude and heading. For applicable frequency ratio decoupling rules, parameter settings such as a frequency adjustment amplitude of ±0.1 Hz, an avoidance threshold of 1:4 ±0.2, and an adjustment response time of 0.05 seconds are read from the decoupling rule base. A frequency adjustment range of ±0.1Hz provides sufficient avoidance space, allowing the drone to escape the resonance region without deviating excessively from the normal operating frequency. An avoidance threshold of 1:4±0.2 defines the dangerous range requiring frequency adjustment. A fast response time of 0.05 seconds ensures frequency adjustment is completed before entering the resonance region. Additional parameters include a frequency scan range of 0.1-5Hz, resonance detection sensitivity of 0.05, and a frequency adjustment smoothing factor of 0.3. When flying in complex electromagnetic interference environments like urban areas, these parameters help the drone automatically adjust its control frequency to avoid frequency combinations that could cause resonance. Simultaneously, based on amplitude limiting rules in the decoupling rule base, decoupling strategy parameters such as an amplitude constraint coefficient of 0.7, a trigger threshold, and a dynamic adjustment range are obtained.An amplitude constraint coefficient of 0.7 represents a balance between safety margin and control effectiveness, ensuring that the system will not become unstable due to excessively strong control signals when approaching the resonant frequency. Trigger thresholds are limited by multiple criteria, including a frequency proximity threshold of 90%, a coupling strength threshold of 0.15, and an oscillation amplitude threshold of ±5 degrees. The dynamic adjustment range is set to 0.5-0.8, allowing for optimization of constraint strength within a safe range based on real-time conditions. Auxiliary parameters also include a constraint activation transition time of 0.1 seconds, a constraint release hysteresis time of 0.2 seconds, and a limit constraint coefficient of 0.5 in emergency situations. These parameters constitute a multi-layered safety protection mechanism, automatically limiting control output when encountering strong gusts or performing extreme maneuvers to prevent loss of control due to coupling resonance. Through systematic parameter extraction and configuration, complete decoupling strategy parameters were obtained.

[0053] Finally, based on the acquired decoupling strategy parameters, a specific dynamic decoupling strategy is generated, transforming the parameter configuration into an executable control algorithm and real-time adjustment mechanism. Utilizing the amplitude coefficient of 0.8 and phase lead of 45 degrees in the decoupling strategy parameters, a low-frequency decoupling strategy is generated: when the altitude control frequency is detected to be close to 0.2Hz, the compensation signal for the heading control loop is automatically calculated as the original altitude control signal × 0.8 × cos(ωt + 45°), and injected into the heading control loop after a delay of 0.125 seconds. Based on the feedforward compensation coefficient of 0.9 and phase lag of 35 degrees in the decoupling strategy parameters, a mid-frequency decoupling strategy is generated: when the heading control frequency approaches 1.2Hz, a compensation amount of the original heading control signal × 0.9 × cos(ωt - 35°) is pre-added to the altitude control loop, and the compensation action is executed 0.083 seconds in advance. By utilizing the frequency adjustment margin of ±0.1Hz in the decoupling strategy parameters, a frequency avoidance strategy is generated: when the frequency ratio approaches 1:4, the altitude control frequency is automatically fine-tuned by ±0.1Hz or the heading control frequency is reversed by ±0.1Hz to deviate the ratio from the danger zone. Simultaneously, based on the amplitude constraint coefficient of 0.7 in the decoupling strategy parameters, an amplitude limiting strategy is generated: when operating near the resonant frequency, the amplitude of the control signal is dynamically limited to no more than 70% of the normal value to prevent resonance amplification. By integrating all control algorithms and adjustment mechanisms based on the decoupling strategy parameters, a dynamic decoupling strategy capable of real-time adjustment and adaptive execution according to flight conditions is formed.

[0054] An independent channel structure for altitude and heading control is established using a dynamically decoupling strategy. Decoupling compensation and interference isolation are implemented to achieve mutual independence between the two control loops. First, the low-frequency decoupling strategy within the dynamic decoupling strategy is applied. The control frequency is monitored in real-time in the altitude control channel. When the altitude control signal frequency approaches 0.2Hz, the compensation amount is calculated according to the compensation algorithm in the dynamic decoupling strategy: original altitude control signal × 0.8 × cos(ωt + 45°). After a delay of 0.125 seconds, this compensation signal is injected into the opposite direction of the heading control loop to counteract the coupling effect of altitude control on heading control. This allows the altitude control channel to independently adjust altitude parameters without interfering with the heading response. Then, the mid-frequency decoupling strategy in the dynamic decoupling strategy is implemented. The control command frequency is monitored in the heading control channel. When the heading control frequency approaches 1.2Hz, the compensation value is pre-calculated according to the feedforward compensation formula in the dynamic decoupling strategy: original heading control signal × 0.9 × cos(ωt-35°). The compensation signal is added to the altitude control loop 0.083 seconds in advance to prevent heading adjustments from interfering with altitude stability and ensure that the heading control channel can independently perform heading maneuvers without affecting altitude maintenance. Next, the frequency avoidance strategy in the dynamic decoupling strategy is implemented. When the ratio of the altitude control frequency to the heading control frequency is detected to be close to 1:4, the altitude control frequency is finely adjusted by +0.1Hz or the heading control frequency is adjusted by -0.1Hz according to the frequency adjustment mechanism in the dynamic decoupling strategy. This causes the operating frequencies of the two control channels to leave the coupling resonance region, thereby achieving channel separation in the frequency domain. Simultaneously, the amplitude limiting strategy within the dynamic decoupling strategy is applied. When the operating frequency of any control channel approaches the resonant frequency, the control signal amplitude of that channel is limited to within 70% of its normal value according to the constraint mechanism in the dynamic decoupling strategy. This prevents the resonant amplification effect from propagating between channels and ensures the independent and stable operation of each channel. Furthermore, through the real-time monitoring mechanism in the dynamic decoupling strategy, the operating status of the two control channels is detected every 0.05 seconds. Based on the rule matching algorithm in the dynamic decoupling strategy, the decoupling mode is automatically selected and switched to achieve dynamic isolation and adaptive independent control between channels. By systematically applying various decoupling measures in the dynamic decoupling strategy, physically and logically independent altitude control and heading control channels are established. The altitude control channel is specifically responsible for tracking altitude commands and eliminating altitude errors, while the heading control channel is specifically responsible for executing heading commands and correcting heading deviations. Under the action of the dynamic decoupling strategy, the two channels operate completely independently, eliminating the problem in traditional coupled control where adjustments to one channel affect the performance of the other, significantly improving the accuracy and stability of UAV altitude and heading control.

[0055] Step S140: Perform an interaction relationship analysis on the altitude control channel and the heading control channel to determine the channel interaction characteristics, design a coordination mechanism based on the channel interaction characteristics, and obtain the coordination constraint parameters between channels.

[0056] Specifically, based on the established independent altitude control channel and heading control channel, this study analyzes in depth the interaction relationships and mutual influence patterns that still exist between the two channels during independent operation. First, residual coupling analysis is performed on the independent altitude control channel and heading control channel. Although channel independence is achieved through a dynamic decoupling strategy, a slight mutual influence still exists between the two channels under specific flight conditions. By monitoring the impact of the altitude control channel on the heading control channel when executing climb commands, it is found that when the altitude control channel outputs a large adjustment signal, the heading control channel will produce a small deviation of 0.02 degrees, which automatically converges within 2 seconds. Similarly, when the heading control channel performs a large-angle turn maneuver, the altitude control channel will experience a transient altitude fluctuation of 0.5 meters, lasting approximately 1.5 seconds. Then, the information exchange requirements between the independent altitude control channel and heading control channel are analyzed. Although the two channels achieve independence at the control level, information sharing is still required at the mission level to achieve coordinated flight. The altitude control channel needs to acquire the turning radius information from the heading control channel to predict the climb trajectory, while the heading control channel needs to know the climb rate information from the altitude control channel to adjust its turning strategy. The two channels exchange status information every 0.1 seconds to maintain consistency. Next, the necessity of coordinated control between the independent channels is analyzed. When the UAV performs complex maneuvers such as spiral climbs, although the altitude control and heading control channels operate independently, they need to maintain temporal synchronization. The climb maneuvers of the altitude control channel and the turning maneuvers of the heading control channel need to be coordinated to achieve the expected flight trajectory. Further quantitative analysis of the interaction strength between the independent channels is conducted. Where Δh represents altitude deviation, Δψ represents heading deviation, and h_ref and ψ_ref are reference values. By measuring the mutual influence between the two channels under different flight conditions, it was found that the interaction strength is 0.05 in level flight, increases to 0.12 during climb and turn, and reaches its maximum value of 0.18 during dive and turn, establishing a correspondence between interaction strength and flight condition. Subsequently, the interaction timing characteristics between independent channels were analyzed, studying the time delay and duration of the impact of altitude control channel actions on heading control channel. It was found that the adjustment signal from altitude control channel to heading control channel requires a delay of 0.08 seconds, with an average duration of 1.2 seconds; the impact delay of the reverse heading control channel on altitude control channel is 0.06 seconds, with an impact duration of approximately 0.9 seconds. Simultaneously, the interaction characteristics of the independent channels at different frequencies were analyzed. It was found that the two channels exhibited weak interaction in the low-frequency band (0.1-0.5Hz), moderate interaction in the mid-frequency band (0.5-2.0Hz), and negligible interaction in the high-frequency band (2.0-5.0Hz). Finally, by combining interaction intensity quantification and timing delay characteristics, a complete description of the relationship between the independent altitude control channel and heading control channel was formed, thus determining the channel interaction characteristics.

[0057] In some embodiments, the step of designing a coordination mechanism based on the channel interaction characteristics and obtaining inter-channel coordination constraint parameters includes: performing intensity quantification analysis on the channel interaction characteristics to obtain an interaction intensity coefficient; designing a coordination weight allocation strategy based on the interaction intensity coefficient and establishing a weight allocation matrix; setting constraint conditions based on the weight allocation matrix to form a constraint condition set; and parameterizing the constraint condition set to obtain inter-channel coordination constraint parameters.

[0058] First, the intensity of the determined channel interaction characteristics is quantitatively analyzed, converting the degree of interaction impact under different flight states described in the channel interaction characteristics into specific numerical coefficients. By establishing a mapping relationship between flight states and interaction intensity, and based on the measured data in the channel interaction characteristics, the interaction intensity coefficients for level flight, climb-turn, and dive-turn conditions are set to 0.05, 0.12, and 0.18, respectively. Simultaneously, based on the timing delay characteristics in the channel interaction characteristics, the time weighting coefficient formula W_t = 1 / (1 + τ_d / τ_r) is adopted, where τ_d is the delay time and τ_r is the duration. This formula reflects the design concept that the shorter the delay time and the longer the duration, the greater the weight. Substituting the timing parameters of the altitude control channel and the heading control channel into the formula, the time weighting coefficients for the two directions are calculated respectively, forming a quantitative index system that comprehensively considers interaction intensity and timing characteristics, obtaining the interaction intensity coefficient that includes both intensity coefficient and time weight.

[0059] Then, based on the obtained interaction strength coefficient, a coordination weight allocation strategy is designed, and the weight ratio of the two channels in coordinated control is determined by utilizing the numerical distribution law of the interaction strength coefficient. A three-level weight allocation strategy is designed based on the segmented control concept: when the interaction strength coefficient I < 0.08, the system is in a weakly coupled state, with a weight allocation of (w_h, w_a) = (0.6, 0.4), where altitude control dominates; when 0.08 ≤ I ≤ 0.15, the system is in a moderately coupled state, with a weight allocation of (0.5, 0.5), achieving balanced control; when I > 0.15, the system is in a strongly coupled state, with a weight allocation of (0.4, 0.6), where heading control takes priority. A time factor correction is introduced into the weight allocation, and the dynamic weight calculation formula is w'_i = w_i × W_i / Σ(w_j × W_j), where w_i is the original weight and W_i is the time weight coefficient. Normalization ensures that the sum of the weights is 1. This strategy enables adaptive adjustment of control weights based on the intensity of real-time interaction, establishing a weight allocation matrix that includes static weights and dynamic time corrections.

[0060] Next, constraints are set based on the established weight allocation matrix, utilizing the weight ratios in the matrix to define coordination constraints between channels. A weight mapping method is used to directly convert the allocated weights into control amplitude constraints, meaning the maximum adjustment amplitude of a channel is proportional to its weight, ensuring u_max_i = w_i × U_total, where u_max_i is the maximum adjustment of channel i, and U_total is the total control capacity. Corresponding constraint modes are set according to the three-level weight allocation strategy: altitude control is prioritized in weak coupling mode, dual-channel balance is achieved in medium coupling mode, and heading control is prioritized in strong coupling mode. A timing compensation mechanism is also introduced, calculating the start-up timing offset Δt = K × (W_h - W_a) between channels based on the time weight difference, where K is the timing compensation coefficient. Dynamic synchronization is achieved by starting the slower-responding channel earlier. Amplitude constraints, priority constraints, and timing constraints are integrated to form a multi-dimensional set of constraints.

[0061] Finally, the resulting set of constraints is parameterized, converting each constraint into specific control parameters and execution boundaries. Amplitude constraints are parameterized using percentage limiting, setting upper limits for the adjustment rate and rate of change for each channel. Balance constraints are quantified using error tolerance and response time windows to ensure consistency in the coordinated action of the two channels. Priority constraints are converted into priority coefficients P_i = w_i / w_ref, where w_ref is a reference weight used to dynamically adjust control authority. Timing constraints include startup time compensation, maximum allowable delay, and synchronization accuracy requirements. Through systematic parameter mapping and boundary setting, abstract constraints are transformed into numerical parameters directly applicable to the controller, forming a complete inter-channel coordination constraint parameter system, ultimately yielding the inter-channel coordination constraint parameters.

[0062] Step S150: Construct a fractal geometric structure based on the coordination constraint parameters, perform recursive hierarchical design based on the fractal geometric structure, create a fractal recursive control architecture with self-similar properties, analyze the coupling energy flow direction of each level through the fractal recursive control architecture, and generate a coupling energy guidance matrix.

[0063] Specifically, a fractal geometric control structure with self-similar properties is established using the acquired inter-channel coordination constraint parameters, and the numerical relationships in the constraint parameters are converted into a geometric topological description. First, based on the proportional relationship between the maximum adjustment rate of 60% for the altitude channel and the maximum adjustment rate of 40% for the heading channel in the inter-channel coordination constraint parameters, the geometric ratio of the basic fractal unit is established, and the weight ratio is converted into a geometric length ratio. The conversion formula is L_h / L_a = w_h / w_a = 0.6 / 0.4 = 1.5, where L_h and L_a are the geometric lengths corresponding to altitude and heading, respectively, and w_h and w_a are the corresponding weights. An elliptical basic fractal unit with 1.5 as the major axis and 1 as the minor axis is constructed. The major axis of this ellipse represents the dominant dimension of altitude control, and the minor axis represents the subordinate dimension of heading control. Then, using the dual-channel equal weight coefficient of 0.5 in the inter-channel coordination constraint parameters, the symmetry rule of the fractal unit is established. The basic elliptical fractal unit is mirrored along the centerline to generate a symmetrical double-elliptical structure. The upper ellipse represents the geometric space of the altitude control channel, and the lower ellipse represents the geometric space of the heading control channel. The two ellipses meet at the centerline to reflect the equal weight relationship of 0.5. Next, based on the heading priority coefficient of 1.5 and the altitude subordination coefficient of 0.67 in the inter-channel coordination constraint parameters, the double-elliptical structure is geometrically transformed. The area of ​​the heading control ellipse is enlarged by a factor of 1.5, and the area of ​​the altitude control ellipse is reduced by a factor of 0.67, forming an asymmetrical fractal basic structure to reflect the priority difference. Subsequently, using the height advance start time of 0.02 seconds, maximum timing deviation of 0.1 seconds, and synchronization accuracy requirement of 0.01 seconds from the inter-channel coordination constraint parameters, a time mapping relationship for the fractal structure is established, namely, the time-space mapping relationship T_s = k_t × T_r, where T_s is the spatial offset, T_r is the actual time, and k_t=1 is the time-space mapping coefficient. The 0.02-second advance start time is converted into a 0.02-unit offset of the basic fractal unit along the time axis, the 0.1-second maximum timing deviation is converted into a maximum geometric twist angle of 10 degrees for the fractal unit, and the 0.01-second synchronization accuracy is converted into a smoothness parameter of 1% for the fractal boundary. Further based on the adjustment rate limit of no more than 10% per second and the synchronization adjustment error of no more than 5% in the inter-channel coordination constraint parameters, a recursive scaling rule S_n = S_0 × r^n is established, where S_n is the nth level size, S_0 is the base size, and r=0.9 is the shrinkage factor (determined by the 10% rate limit). The size of each level of recursive structure is 90% of the previous level. The 5% synchronization error is used as the tolerance range of the fractal boundary, allowing the recursive structure to produce small deformations within the range of ±5%.Simultaneously, based on the response time difference of less than 0.05 seconds and the priority switching threshold of 0.15 in the inter-channel coordination constraint parameters, connection rules for the fractal structure are set. When the time deviation between adjacent fractal units exceeds 0.05 seconds, the connection is automatically disconnected; when the priority difference exceeds 0.15, the reorganization transformation of the fractal structure is triggered. By converting all inter-channel coordination constraint parameters into geometric attributes and recursive rules, a three-dimensional fractal geometric structure with ellipse as the basic unit, asymmetry, time mapping, and recursive self-similarity characteristics is established. This structure spatially reflects the coordination relationship between altitude and heading control channels, temporally reflects the synchronization constraints between channels, and scale-wise exhibits multi-level recursive control logic, thus constructing the fractal geometric structure.

[0064] In some embodiments, the step of creating a fractal recursive control architecture with self-similar properties by performing recursive hierarchical design based on the fractal geometry includes: defining a basic pattern for the fractal geometry to determine a basic control pattern; recursively replicating the basic control pattern to construct multiple hierarchical control patterns; constructing hierarchical relationships among the multiple hierarchical control patterns to form a recursive hierarchical structure; and forming a fractal recursive control architecture with self-similar properties based on the recursive hierarchical structure.

[0065] A basic pattern is defined for the constructed fractal geometry, and the core control pattern is extracted from the fractal geometry as the basic unit for recursive design. Based on the elliptical basic unit constructed with a geometric length ratio of 1.5:1 in the fractal geometry, the basic control pattern is defined as a double-elliptical control structure containing a dominant dimension of altitude control and a subordinate dimension of heading control. This basic pattern inherits the 60%:40% channel weight allocation, the 0.02-second time offset characteristic, and the 10-degree maximum torsional angle constraint from the fractal geometry. For example, when the UAV performs a standard reconnaissance flight mission, the upper ellipse in the basic control pattern maintains a dominant altitude of 500 meters, occupying 60% of the control resources, while the lower ellipse is responsible for heading adjustment to cover the target area, occupying 40% of the control resources. The two ellipses coordinate with each other through the central connection point. Utilizing the symmetry rules and asymmetric transformation characteristics in the fractal geometry, the basic control pattern is defined as a double-elliptical geometry with vertical symmetry but adjustable area. The upper ellipse carries the altitude control logic, the lower ellipse carries the heading control logic, and the connection center of the two ellipses represents the coordination and interaction point between the channels, thus determining the basic control pattern.

[0066] For example, the recursive replication of the basic control pattern to construct multiple hierarchical control patterns includes: extracting pattern features from the basic control pattern to generate basic pattern feature parameters; establishing a self-similarity transformation analysis based on the basic pattern feature parameters to form a set of transformation rules; performing a first-level recursive replication of the basic control pattern based on the set of transformation rules to construct a first-level control pattern; continuing the recursive replication operation on the first-level control pattern to construct a second-level control pattern; repeating the recursive replication operation until the recursive hierarchy design requirements are met to construct multiple hierarchical control patterns.

[0067] The established control baseline pattern is recursively replicated to construct multiple hierarchical control patterns. Pattern features are extracted from the control baseline pattern, including geometric parameters such as major axis length of 1.5, minor axis length of 1.0, ellipse area ratio of 1.5:0.67, center offset of 0.02, torsion angle constraint of 10 degrees, and boundary smoothness of 1%, generating baseline pattern feature parameters. Based on the generated baseline pattern feature parameters, a self-similarity transformation analysis is established. Scaling transformation rules are established using the major axis length of 1.5 as the scaling reference, rotation transformation rules are established using the minor axis length of 1.0 as the rotation radius, area adjustment rules are established based on the ellipse area ratio of 1.5:0.67, translation transformation rules are established based on the center offset of 0.02, and angle restriction rules are established through the torsion angle constraint of 10 degrees, forming a set of transformation rules including scaling, rotation, area adjustment, translation, and angle restriction. Based on the established set of transformation rules, the basic control pattern is recursively copied at the first level. Parameters are adjusted according to the recursive formula P_n = P_0 × r^n, where P represents each parameter (size, time, etc.). After scaling, the size S_1 = S_0 × 0.9, and the time offset t_1 = t_0 × 0.9 = 0.02 × 0.9 = 0.018 seconds. Maintaining the 1.5:1 major-to-minor axis ratio, the torsional angle constraint is adjusted to 9 degrees, constructing a first-level control pattern with similar geometric features but a smaller size. The recursive copying operation continues on the constructed first-level control pattern, again applying a 90% scaling factor and various rules from the transformation rule set, generating a control structure 0.81 times the size of the original basic pattern. The time offset is further adjusted to 0.0162 seconds, and the torsional angle constraint is 8.1 degrees, constructing the second-level control pattern. Repeat the above recursive copying operation to generate the third-level control pattern (0.729 times the size), the fourth-level control pattern (0.656 times the size), and the fifth-level control pattern (0.590 times the size) in sequence, until the size of the level control pattern is less than 0.6 times the original base pattern. Stop recursion when the size of the level control pattern is less than 0.6 times the original base pattern. This satisfies the recursive level design requirements and constructs multiple level control patterns that include the base pattern and five recursive levels.

[0068] A hierarchical relationship was constructed for the multiple control modes, establishing control logic connections and information transmission relationships between each level. The basic control mode was set as the top-level control node, responsible for receiving external control commands and formulating the overall control strategy; the first-level control mode was set as the main control layer, responsible for decomposing the top-level strategy into specific altitude and heading control tasks; the second to fifth-level control modes were set as execution control layers, responsible for refining control commands, supervising execution, providing feedback adjustments, and correcting errors in sequence. For example, in a low-altitude penetration mission in complex terrain, the top-level basic mode receives the overall command "maintain 200 meters altitude, heading 030 degrees," the first level decomposes it into specific tasks of "altitude control ±5 meters, heading control ±2 degrees," the second level further refines it into execution commands of "altitude adjustment not exceeding 1 meter per second, heading adjustment not exceeding 0.5 degrees per second," and the third to fifth levels are responsible for the fine-tuning operations of control surface control, sensor feedback, and error compensation in sequence. Establish information flow rules between levels: higher levels transmit control commands and constraint parameters to lower levels, while lower levels report execution status and performance indicators back to higher levels. Adjacent levels maintain a 0.01-second information transmission delay to ensure timing synchronization. Simultaneously, establish a coordination mechanism between levels: when a level experiences an anomaly or performance degradation, higher levels automatically adjust their control strategies and reallocate tasks to other levels, forming a recursive hierarchical structure with self-healing capabilities.

[0069] Based on the established recursive hierarchical structure, a unified fractal recursive control architecture is established by integrating the geometric features and control logic of all hierarchical control modes. This architecture inherits the self-similarity of fractal geometry; each level possesses the same double-elliptical geometry and control logic, but exhibits a recursive decreasing pattern in size, temporal characteristics, and control precision. Through geometric nesting and logical recursion, each level in the architecture achieves complete coverage from macroscopic control to microscopic execution. The top level handles large-scale coordination control tasks, the bottom level handles small-scale precision execution tasks, and the intermediate levels are responsible for the transition and connection between different scales. For example, during formation flight, the top-level architecture maintains the overall formation and flight path, the first level controls the relative position of individual aircraft, the second level handles the attitude fine-tuning of individual aircraft, and the bottom level handles the precise control of control surfaces and power. The levels maintain coordination and consistency through fractal recursion, ensuring the stability and precision of formation flight. Through this fractal recursive design, the control architecture achieves multi-scale control capabilities while maintaining overall consistency, forming a fractal recursive control architecture with self-similarity.

[0070] A fractal recursive control architecture is established to systematically analyze the coupling energy flow between different levels, quantifying the energy transfer relationships and distribution patterns between different levels. Coupling energy includes three types: control signal energy, information transfer energy, and execution action energy, exhibiting specific flow patterns between different levels of the fractal architecture. Based on the hierarchical structure of the fractal architecture, the energy flow function E_flow(i,j) = α × S_i × (P_i - P_j) × K_ij is defined, where i and j are level numbers, S_i is the geometric dimension of the i-th level, P_i and P_j are the control power of each level, K_ij is the coupling coefficient between levels, and α is the energy transfer efficiency (typically 0.85-0.95). Due to the fractal recursive characteristics, the coupling coefficient between adjacent levels satisfies K_ij = K_0 × r^|ij|, where K_0=0.8 is the basic coupling coefficient, and r=0.9 is the recursion factor, reflecting the characteristic that energy transfer attenuates with increasing level distance. By calculating the energy flow layer by layer, it was found that the energy flow from the top-level basic model to the first level, E_flow(0,1), accounts for 45% of the total energy, mainly used for strategy decomposition and instruction generation; the energy flow from the first level to the second level, E_flow(1,2), accounts for 30%, used for control refinement and parameter adjustment; the energy flow from the second level to the third level, E_flow(2,3), accounts for 15%, and the energy flow from the third level to the fourth level, E_flow(3,4), accounts for 7%, with the bottom layer accounting for only 3%, forming a clear cascading energy distribution. Analysis of the energy aggregation characteristics of each level revealed that the second and third levels are the key nodes for energy aggregation. These two levels undertake the core task of transforming macroscopic control into specific execution, with energy densities reaching 0.35 and 0.28 units / layer, respectively. Simultaneously considering energy loss between levels, a loss function L_ij = β × E_flow(i,j) × d_ij is defined, where β = 0.05 is the loss coefficient, and d_ij is the logical distance between levels. Calculations show that the energy loss between adjacent levels is approximately 5% of the transferred energy, increasing to 8-10% when transferring across levels. A coupled energy guidance matrix M_energy is constructed, where the matrix element M_ij = E_flow(i,j) / E_total represents the proportion of energy flowing from level i to level j. This matrix has an upper triangular property, reflecting the unidirectional flow of energy from higher to lower levels. The trace tr(M) = 0 indicates that there is no energy self-circulation within a level. For example, when performing complex maneuvering missions, the energy orientation matrix is ​​as follows: M_01=0.45 (top level to first level), M_12=0.30 (first level to second level), M_23=0.15 (second level to third level), M_02=0.05 (top level to second level, cross-level transfer), M_13=0.03 (first level to third level, cross-level transfer), and the remaining cross-level elements are less than 0.02.The maximum eigenvalue λ_max = 0.92 of the matrix reflects the energy convergence characteristics of the system, ensuring control stability. Through quantitative analysis, loss calculation, and matrix representation of the coupled energy flow, a coupled energy steering matrix that comprehensively describes the energy distribution, transmission laws, and loss characteristics in the fractal recursive control architecture is generated.

[0071] Step S160: Obtain the coupling prediction value at each control moment based on the coupling energy steering matrix, generate the reverse compensation sequence based on the coupling prediction value, allocate the reverse compensation sequence to each level through the fractal recursive control architecture to generate compensation allocation results, and optimize the multi-level control strategy based on the compensation allocation results to output coordinated control commands.

[0072] Specifically, the energy coupling trend at future control moments is analyzed using the coupled energy steering matrix to predict the coupling effect between different levels. The element M_ij in the coupled energy steering matrix M_energy reflects the proportion of energy flow from level i to level j. Based on this matrix, a coupled prediction model is established: E_pred(t+Δt) = M_energy × E_current(t) + ΔE_trend, where E_current(t) is the energy distribution vector at the current moment, and ΔE_trend is the energy change trend term. The energy distribution after k control cycles is predicted through matrix exponentiation M_energy^k, where k = Δt / T_c, and T_c is the control cycle (typically 0.05 seconds). The prediction process employs a sliding window mechanism, using historical data from the past 5 cycles to calculate the trend term ΔE_trend, and obtaining the energy change rate through least-squares fitting. For multi-step prediction, the recursive formula E_pred(t+kT_c) = (M_energy × L)^k × E_current(t) is used, where L is the loss matrix, the diagonal element L_ii=0.95 represents 5% intra-layer loss, and the off-diagonal element L_ij=0 considers inter-layer transfer loss already included in M_energy. Prediction accuracy is evaluated through residual analysis, and an adaptive correction factor is introduced when the prediction residual exceeds a threshold. The time window is set to 5 control cycles, i.e., predicting coupling changes within the next 0.25 seconds, ensuring real-time performance while covering the main dynamic processes. Through dynamic analysis and multi-step recursive calculation of the coupling energy steering matrix, the predicted coupling values ​​at each control moment are obtained.

[0073] Based on the obtained coupling prediction values, a reverse compensation mechanism is designed to generate a compensation sequence that can counteract adverse coupling effects. The core idea of ​​reverse compensation is to pre-inject a control quantity opposite to the predicted coupling to achieve active cancellation of coupling effects. The compensation quantity is calculated as C(t) = -K_comp × [E_pred(t) - E_desired(t)], where K_comp is the compensation gain matrix, and E_desired(t) is the desired energy distribution, which is usually set to a uniform distribution across all levels or a specific distribution according to task requirements. The compensation gain matrix is ​​designed by solving the algebraic Riccati equation A^T×P + P×A - P×B×R^(-1)×B^T×P + Q = 0, where A is the system matrix, B is the control matrix, and Q and R are the state and control weight matrices, respectively. The obtained P is used to calculate K_comp = R^(-1)×B^T×P. Differentiated compensation strategies are designed for different types of coupling effects: high-gain fast compensation (K_fast=1.2) is used for fast coupling to ensure rapid response; low-gain smooth compensation (K_slow=0.6) is used for slow coupling to avoid over-adjustment; and phase compensation (C_phase(t) = A_comp × sin(ωt + φ_comp)) is introduced for oscillation suppression through phase cancellation. The timing arrangement of the compensation sequence is gradient-designed based on the response characteristics of each level, forming a compensation sequence structure that is staggered in time and decreases in amplitude. The sequence length is set to 10 sampling points, with each sampling point spaced 0.05 seconds apart, covering a 0.5-second control period to ensure the continuity of the compensation effect. A reverse compensation sequence is generated through the design of a reverse compensation algorithm and a multi-type coupling processing strategy.

[0074] In some embodiments, the step of allocating the reverse compensation sequence to each level and generating compensation allocation results through the fractal recursive control architecture includes: performing hierarchical adaptation analysis based on the reverse compensation sequence to determine the compensation requirements of each level; decomposing the compensation amount according to the compensation requirements of each level to generate hierarchical compensation vectors; recursively allocating the hierarchical compensation vectors using the fractal recursive control architecture; and performing constraint verification on the recursive allocation results to generate compensation allocation results.

[0075] The compensation capacity of each level is analyzed based on the generated reverse compensation sequence to determine the appropriate compensation requirement for each level. The reverse compensation sequence contains compensation data at 10 time points C(t_k) = [C_0(t_k), C_1(t_k), ...,C_5(t_k)], where k = 1 to 10, and each C_i(t_k) represents the proposed compensation amount for the i-th level at time t_k. The level adaptation analysis first evaluates the remaining control capacity of each level R_capacity(i) = C_max(i) - C_current(i), where C_max(i) is the maximum control capacity of the i-th level, following the decreasing law of fractal recursive architecture C_max(i) = C_max(0) ×r^i, and C_current(i) is the current control occupancy. By comparing the demand of the compensation sequence with the level capacity, situations of insufficient capacity are identified and adjustments are made. When the compensation demand of a certain level exceeds the remaining capacity, the excess is recursively pushed to adjacent levels according to the principle of energy conservation to ensure that the total compensation amount remains unchanged. The adaptation process also considers the differences in dynamic characteristics of each level. Frequency components of the compensation sequence are extracted through spectral analysis, prioritizing the allocation of low-frequency, large-amplitude components to the top level, mid-frequency components to the middle levels, and high-frequency, small-amplitude components to the bottom level, achieving the optimal match between compensation requirements and level characteristics. Compensation requirements for each level are determined through capacity assessment, overflow handling, and frequency matching.

[0076] The determined compensation requirements for each level are quantitatively decomposed, and the total compensation amount is precisely allocated according to the level's capacity and characteristics. A weighted allocation algorithm is used for the compensation allocation, with the compensation allocation ratio for level i being w_i = (R_capacity(i) × P_i) / Σ(R_capacity(j) × P_j), where R_capacity(i) is the remaining capacity of that level, and P_i is the performance weight, reflecting the importance of that level in the overall control. Based on the self-similarity characteristics of the fractal recursive architecture, a compensation transfer relationship is established between levels: V_comp(i+1) = β × V_comp(i) + ΔV_local(i+1), where β is the transfer coefficient, and ΔV_local is the local compensation adjustment amount. When constructing the compensation vector, two dimensions are distinguished: altitude and heading, forming a two-dimensional vector V_comp(i) = [V_h(i), V_a(i)], ensuring that the compensation requirements of both control channels are met. The time-dimensional decomposition divides the 0.5-second compensation sequence into 10 time slices, with each level having a clearly defined compensation task within each time slice. The decomposition process maintains energy conservation, i.e., Σ(V_comp(i)) = V_total, ensuring the total compensation amount remains consistent before and after decomposition. Through weighted allocation, recursive propagation, and multi-dimensional decomposition, hierarchical compensation vectors are generated.

[0077] Utilizing the hierarchical transfer mechanism of the fractal recursive control architecture, hierarchical compensation vectors are distributed to each execution unit according to recursive rules. The recursive allocation follows the top-down propagation path of the fractal architecture, with the top-level compensation vector V_comp(0) being passed to lower levels via fractal transformation. The fractal transformation comprises two parts: geometric transformation and time transformation. The geometric transformation T_geo = S × R(θ) implements scaling S = 0.9 and angular rotation R(θ), while the time transformation T_time introduces a time offset τ to ensure ordered signal transmission. The compensation received by the (i+1)th layer is V_recv(i+1) = T_geo × V_comp(i) × T_time + V_local(i+1), where V_local is the supplementary compensation generated by this layer based on the local state. The recursive transfer process maintains the self-similarity of the fractal structure; each layer executes the same transfer rules, but the parameters are scaled proportionally. The parallel processing mechanism allows multiple control units within the same layer to simultaneously receive and process compensation instructions, ensuring coordination through time synchronization. The redundant design of the transmission path provides fault tolerance, automatically switching to the backup path when the primary path fails. Hierarchical allocation of compensation vectors is achieved through fractal transformation and recursive transmission mechanisms.

[0078] A comprehensive constraint check is performed on the recursive allocation results to ensure the feasibility and security of the compensation allocation. The constraint check system includes four core constraints: capacity constraints ensure that the compensation amount at each level does not exceed the remaining capacity |V_comp(i)| ≤ R_capacity(i), and triggers a compensation reduction mechanism when violated; stability constraints verify stability by calculating the eigenvalues ​​of the compensated system, requiring that the real part of all eigenvalues ​​be negative Re(λ_i) < 0; timing constraints check the time consistency of the compensation signal, and the propagation delay of adjacent levels must satisfy |δt(i+1) - δt(i) - δt_base| < ε, where δt_base is the standard delay and ε is the tolerance; coordination constraints ensure that altitude and heading compensation maintain a reasonable proportional relationship to avoid overcompensation of a single channel. The check process adopts an iterative optimization method. When a constraint is not satisfied, it is corrected by adjusting the allocation weight w_i or the propagation coefficient β, with a maximum of 10 iterations. After each iteration, the satisfaction level of each constraint is recalculated, forming a constraint margin vector: margin = [m_capacity, m_stability, m_timing, m_coord]. Verification passes when all margins are greater than zero. The verification result includes the final compensation allocation scheme, execution parameters at each level, and constraint margin information. Through systematic constraint verification and iterative optimization, the compensation allocation result is generated.

[0079] Based on the generated compensation allocation results, the multi-level control strategy is comprehensively optimized to generate the final coordinated control command. The compensation allocation results provide the specific compensation amount V_comp_final(i,t) and execution timing δt(i) of each level at each time step. These data are directly integrated into the control law to form a composite control strategy. The comprehensive control law is designed as u_i(t) = K_fb(i) × e_i(t) + V_comp_final(i,t) + u_ff(i,t), where K_fb(i) is the feedback gain matrix of that level, e_i(t) is the tracking error vector, V_comp_final(i,t) is the compensation term in the compensation allocation results, and u_ff(i,t) is the feedforward control term. For example, in a mountainous terrain tracking flight scenario, when an altitude deviation caused by terrain undulations is detected, the top level quickly adjusts the flight altitude according to the compensation allocation, the middle level coordinates the pitch angle and thrust, and the bottom level precisely controls the control surface deflection. The timing and amplitude of the actions of each level are strictly executed according to the compensation allocation results, achieving smooth tracking of complex terrain. Control commands are encoded in a structured format [Layer_ID, Control_Type, Time_Stamp, Control_Value, Priority] to ensure accurate transmission and execution. Command priorities are dynamically set based on the constraint margins in the compensation allocation results; layers with smaller margins receive higher priority to ensure critical constraints are met. An exception handling mechanism is embedded in the command flow; when execution deviations exceed the preset tolerance range of the compensation allocation results, a safety mode is automatically activated and an exception is reported. By fully integrating the compensation allocation results into the control strategy, coordinated control commands are output.

[0080] Step S170: Modulate the signal and distribute the power according to the coordinated control command to generate a modulation control signal. Output the altitude control signal and heading control signal through the modulation control signal to realize the coordinated control of the UAV's altitude and heading.

[0081] Specifically, signal modulation processing is performed on the output coordinated control commands to convert the digital control information in the commands into an analog signal format suitable for the actuators. Pulse Width Modulation (PWM) processing is applied to the altitude control component of the coordinated control commands, converting the digital command "elevator up 3 degrees, maintain for 0.8 seconds" into a PWM signal with a 70% duty cycle and a frequency of 50Hz to drive the elevator servo motor. Amplitude Modulation (AM) processing is applied to the heading control component of the coordinated control commands, converting the command "rudder right 2 degrees, maintain for 1.2 seconds" into a sinusoidal modulated signal with an amplitude of 1.5V and a frequency of 20Hz to control the rudder actuator. Simultaneously, frequency modulation is applied to the coordination synchronization signal in the coordinated control commands, converting the 0.1-second timing difference requirement into a carrier signal with a phase difference of 18 degrees to ensure that each actuator operates in a coordinated manner according to the predetermined timing.

[0082] Power allocation is performed based on the execution intensity requirements of the coordinated control commands, analyzing the relative importance of altitude control and heading control within the commands, as well as the power requirements of the actuators. Based on the urgency of the altitude climb task in the coordinated control commands, 80% of the available power is allocated to the elevator control, and 60% to the throttle control. Considering the accuracy requirements for heading avoidance in the coordinated control commands, 70% of the power is allocated to the rudder control, and 50% to the aileron control. Taking into account the 0.1-second timing difference synchronization requirement in the coordinated control commands, an additional 10% power margin is allocated to the synchronization control circuit to ensure accurate execution of the coordinated control. The modulated signal and power allocation results are then integrated to generate a modulated control signal containing complete information such as amplitude, frequency, phase, and power. The altitude control modulation signal integrates the timing characteristics of the PWM modulation signal and the amplitude information of 80% power allocation to form a composite control signal that drives the elevator and throttle; the heading control modulation signal integrates the frequency characteristics of the amplitude modulation signal and the intensity information of 70% power allocation to form a joint control signal that drives the rudder and aileron; the coordination and synchronization modulation control signal combines the phase characteristics of the frequency modulation and the guarantee mechanism of 10% power margin to form a coordination signal that ensures the timing synchronization of each control channel.

[0083] The generated modulation control signals output specific execution control signals. After signal separation and amplification, the modulation control signals output dedicated altitude control and heading control signals. The altitude control signal includes an elevator control signal (PWM duty cycle 70%, duration 0.8 seconds, phase reference 0 degrees) and a throttle control signal (voltage output 2.4V, gradient change 0.3V / second), directly driving the altitude control actuator. The heading control signal includes a rudder control signal (amplitude 1.5V, frequency 20Hz, phase lag 18 degrees) and an aileron control signal (differential output ±1.2V, response time 0.05 seconds), precisely controlling the heading adjustment actuator. Through these precisely output altitude and heading control signals, the UAV's elevator adjusts altitude according to a predetermined angle and timing, while the rudder and ailerons execute heading control according to calculated deflection and coordination timing. The two control systems work perfectly together under the unified coordination of the modulation control signals, ultimately achieving coordinated control of the UAV's altitude and heading.

[0084] To implement the UAV altitude and heading coordination control method corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects. See also Figure 2 , Figure 2This diagram illustrates a structural block diagram of a drone altitude and heading coordination control device 200 according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The drone altitude and heading coordination control device 200 provided in this embodiment includes:

[0085] The state processing module 201 is used to preprocess and extract features from the flight state parameters of the UAV, generate standardized state parameters, perform topological transformation processing on the standardized state parameters, and construct an altitude-heading state space mapping relationship.

[0086] The resonance analysis module 202 is used to perform frequency domain transformation on the altitude-heading state space mapping relationship, generate frequency domain feature data, perform frequency domain resonance mode analysis on the frequency domain feature data, and identify the coupling resonance characteristics of altitude and heading control.

[0087] The channel establishment module 203 is used to establish a decoupling rule base based on the coupling resonance characteristics, generate a dynamic decoupling strategy based on the decoupling rule base, and form an independent altitude control channel and heading control channel through the dynamic decoupling strategy.

[0088] The coordination design module 204 is used to perform interaction relationship analysis on the altitude control channel and the heading control channel, determine the channel interaction characteristics, design a coordination mechanism based on the channel interaction characteristics, and obtain coordination constraint parameters between channels.

[0089] Architecture building module 205 is used to construct a fractal geometric structure according to the coordination constraint parameters, perform recursive hierarchical design based on the fractal geometric structure, create a fractal recursive control architecture with self-similar characteristics, analyze the coupling energy flow direction of each level through the fractal recursive control architecture, and generate a coupling energy guidance matrix.

[0090] The coordination processing module 206 is used to obtain the coupling prediction value at each control moment based on the coupling energy steering matrix, generate a reverse compensation sequence according to the coupling prediction value, allocate the reverse compensation sequence to each level through the fractal recursive control architecture to generate compensation allocation results, and perform multi-level control strategy optimization based on the compensation allocation results to output coordinated control commands.

[0091] The signal output module 207 is used to perform signal modulation and power distribution according to the coordinated control command, generate a modulation control signal, and output an altitude control signal and a heading control signal through the modulation control signal to realize the coordinated control of the UAV's altitude and heading.

[0092] The aforementioned UAV altitude and heading coordination control device 200 can implement the UAV altitude and heading coordination control method of the above-described method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining content of this application embodiment can be referred to the content of the above method embodiments, and will not be repeated in this embodiment.

[0093] like Figure 3 As shown, the third embodiment of the present invention also provides a computer device, including a memory 301, a processor 302, and a computer program stored in the memory 301 and executable on the processor 302, characterized in that the processor 302 executes the program to implement the steps of the UAV altitude and heading coordinated control method described in the first embodiment of the present invention.

[0094] The purpose of the above embodiments is to reproduce and derive the technical solution of the present invention by way of example, and to fully describe the technical solution, purpose and effect of the present invention. The purpose is to enable the public to have a more thorough and comprehensive understanding of the disclosure of the present invention, and not to limit the scope of protection of the present invention.

[0095] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for coordinated control of altitude and heading of an unmanned aerial vehicle (UAV), characterized in that, include: The flight state parameters of the UAV are preprocessed and feature extracted to generate standardized state parameters. A topological transformation is then performed on these standardized state parameters to construct an altitude-heading state space mapping relationship. This includes: extracting the altitude dimension from the standardized state parameters to form an altitude state vector; extracting the heading dimension from the standardized state parameters to form a heading state vector; performing coordinate system transformation on the altitude state vector and the heading state vector to generate a unified coordinate system state vector; projecting the unified coordinate system state vector into topological space to generate topological space coordinates; and establishing an altitude-heading state space mapping relationship based on the topological space coordinates. The altitude-heading state space mapping relationship is transformed in the frequency domain to generate frequency domain feature data. Frequency domain resonance mode analysis is then performed on the frequency domain feature data to identify the coupled resonance characteristics of altitude and heading control. This includes: performing spectral analysis on the frequency domain feature data to extract the altitude control spectrum and heading control spectrum; performing cross-correlation analysis on the altitude control spectrum and the heading control spectrum to obtain frequency domain correlation coefficients; identifying resonance frequency points based on the frequency domain correlation coefficients to determine multiple candidate resonance frequencies; performing amplitude and phase analysis on the multiple candidate resonance frequencies to obtain resonance mode parameters; and generating coupled resonance characteristics of altitude and heading control based on the resonance mode parameters. A decoupling rule base is established based on the coupling resonance characteristics, and a dynamic decoupling strategy is generated based on the decoupling rule base. Independent altitude control channels and heading control channels are formed through the dynamic decoupling strategy. An interaction relationship analysis is performed on the altitude control channel and the heading control channel to determine the channel interaction characteristics. Based on the channel interaction characteristics, a coordination mechanism is designed to obtain the coordination constraint parameters between channels. A fractal geometric structure is constructed based on the coordination constraint parameters. A recursive hierarchical design is performed based on the fractal geometric structure to create a fractal recursive control architecture with self-similar properties. The coupling energy flow direction of each level is analyzed through the fractal recursive control architecture to generate a coupling energy guidance matrix. Based on the coupled energy steering matrix, the coupled prediction value at each control moment is obtained. A reverse compensation sequence is generated according to the coupled prediction value. The reverse compensation sequence is allocated to each level through the fractal recursive control architecture to generate compensation allocation results. Based on the compensation allocation results, multi-level control strategy optimization is performed to output coordinated control commands. The system performs signal modulation and power allocation according to the coordinated control command, generates a modulation control signal, and outputs an altitude control signal and a heading control signal through the modulation control signal to achieve coordinated control of the UAV's altitude and heading.

2. The method for coordinated control of UAV altitude and heading according to claim 1, characterized in that, The generation of dynamic decoupling strategies based on the decoupling rule base includes: Perform rule matching analysis on the decoupling rule base to determine the applicable decoupling rules; Configure the strategy parameters according to the applicable decoupling rules, and obtain the decoupling strategy parameters; A dynamic decoupling strategy is generated based on the decoupling strategy parameters.

3. The method for coordinated control of UAV altitude and heading according to claim 1, characterized in that, The design of the coordination mechanism based on the channel interaction characteristics, and the acquisition of inter-channel coordination constraint parameters, include: The interaction characteristics of the channels are subjected to intensity quantification analysis to obtain the interaction intensity coefficient; Based on the interaction strength coefficient, a coordination weight allocation strategy is designed, and a weight allocation matrix is ​​established. Constraints are set based on the weight allocation matrix to form a set of constraints. The set of constraints is parameterized to obtain the coordination constraint parameters between channels.

4. The method for coordinated control of UAV altitude and heading according to claim 1, characterized in that, The recursive hierarchical design based on the fractal geometry, creating a fractal recursive control architecture with self-similar properties, includes: Define the basic pattern of the fractal geometry and determine the control basic pattern; The basic control pattern is recursively copied to construct multiple hierarchical control patterns; The hierarchical relationship of the multiple hierarchical control modes is constructed to form a recursive hierarchical structure; Based on the aforementioned recursive hierarchical structure, a fractal recursive control architecture with self-similar properties is formed.

5. The method for coordinated control of UAV altitude and heading according to claim 1, characterized in that, The step of distributing the reverse compensation sequence to each level and generating compensation distribution results through the fractal recursive control architecture includes: Based on the reverse compensation sequence, hierarchical adaptation analysis is performed to determine the compensation requirements at each level. Based on the compensation requirements of each level, the compensation amount is decomposed to generate a hierarchical compensation vector; The hierarchical compensation vector is recursively allocated using the fractal recursive control architecture. The recursive allocation result is constrained and validated to generate a compensated allocation result.

6. The UAV altitude and heading coordinated control method according to claim 4, characterized in that, The recursive replication of the basic control pattern to construct multiple hierarchical control patterns includes: The basic control mode is subjected to mode feature extraction to generate basic mode feature parameters; A self-similarity transformation analysis is established based on the aforementioned basic pattern feature parameters to form a set of transformation rules; Based on the set of transformation rules, the control basic mode is recursively copied at the first level to construct the first level control mode; The first-level control pattern is recursively copied to construct the second-level control pattern. Repeat the recursive copying operation until the recursive hierarchy design requirements are met, and construct multiple hierarchical control patterns.

7. A drone altitude and heading coordinated control device, characterized in that, include: The state processing module is used to preprocess and extract features from the flight state parameters of the UAV to generate standardized state parameters. It then performs topological transformation on the standardized state parameters to construct an altitude-heading state space mapping relationship. This includes: extracting the altitude dimension from the standardized state parameters to form an altitude state vector; extracting the heading dimension from the standardized state parameters to form a heading state vector; performing coordinate system transformation on the altitude state vector and the heading state vector to generate a unified coordinate system state vector; projecting the unified coordinate system state vector into topological space to generate topological space coordinates; and establishing an altitude-heading state space mapping relationship based on the topological space coordinates. The resonance analysis module is used to perform frequency domain transformation on the altitude-heading state space mapping relationship to generate frequency domain feature data, and to perform frequency domain resonance mode analysis on the frequency domain feature data to identify the coupled resonance characteristics of altitude and heading control. This includes: performing spectral analysis on the frequency domain feature data to extract the altitude control spectrum and heading control spectrum; performing cross-correlation analysis on the altitude control spectrum and the heading control spectrum to obtain frequency domain correlation coefficients; identifying resonance frequency points based on the frequency domain correlation coefficients to determine multiple candidate resonance frequencies; performing amplitude and phase analysis on the multiple candidate resonance frequencies to obtain resonance mode parameters; and generating coupled resonance characteristics of altitude and heading control based on the resonance mode parameters. The channel establishment module is used to establish a decoupling rule base based on the coupling resonance characteristics, generate a dynamic decoupling strategy based on the decoupling rule base, and form independent altitude control channels and heading control channels through the dynamic decoupling strategy. The coordination design module is used to analyze the interaction relationship between the altitude control channel and the heading control channel, determine the channel interaction characteristics, design a coordination mechanism based on the channel interaction characteristics, and obtain the coordination constraint parameters between the channels. The architecture construction module is used to construct a fractal geometric structure based on the coordination constraint parameters, perform recursive hierarchical design based on the fractal geometric structure, create a fractal recursive control architecture with self-similar properties, and analyze the coupling energy flow direction of each level through the fractal recursive control architecture to generate a coupling energy guidance matrix. The coordination processing module is used to obtain the coupling prediction value at each control moment based on the coupling energy steering matrix, generate a reverse compensation sequence according to the coupling prediction value, allocate the reverse compensation sequence to each level through the fractal recursive control architecture to generate compensation allocation results, and perform multi-level control strategy optimization based on the compensation allocation results to output coordinated control commands. The signal output module is used to perform signal modulation and power distribution according to the coordinated control command, generate a modulation control signal, and output an altitude control signal and a heading control signal through the modulation control signal to realize the coordinated control of the UAV's altitude and heading.

8. A computer device, characterized in that, It includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and, when executing the computer program, implement the method as described in any one of claims 1 to 6.

Images