A low-altitude equipment adaptive control optimization method based on local wind interference resistance

Abstract

The application relates to the technical field of low-altitude equipment intelligent control, and discloses a low-altitude equipment adaptive control optimization method based on local wind interference resistance, which comprises the following steps: collecting the rotating speed signal and phase current signal of a rotor motor and an attitude error, calculating a tension state and constructing a virtual force receiving surface, extracting local geometric distortion features and carrying out denoising treatment to obtain a real aerodynamic disturbance component; constructing a rotor adjacency topology graph, tracking cross-node propagation delay, constructing an aerodynamic force propagation chain and generating a feedforward compensation amount; calculating a cross-influence matrix of lift and torque, constructing an orthogonal decoupling space; generating a basic attitude control instruction, superimposing the feedforward compensation amount and the basic attitude control instruction, and carrying out projection decomposition to obtain a decoupling compensation component; constructing a reachable polyhedron, truncating the decoupling compensation component and converting the decoupling compensation component into a rotor rotating speed control instruction. The application reduces attitude invalid oscillation of low-altitude equipment, and improves the adaptive anti-interference capability and operation stability of the low-altitude equipment.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for low-altitude equipment, specifically to an adaptive control optimization method for low-altitude equipment based on local wind interference resistance. Background Technology

[0002] When performing low-altitude operations close to the surface of large structures, the flight control of low-altitude equipment such as multi-rotors usually relies on proportional-integral-derivative control or conventional disturbance observers. The global attitude error is obtained through inertial measurement units, and the feedback compensation is calculated by combining a fixed control allocation matrix and numerical filtering algorithms. This drives the various actuators in order to maintain the attitude balance and trajectory stability of the equipment in a normal, uniform, open airflow.

[0003] However, the airflow within the boundary layer is not a uniform flow field, and low-altitude equipment inevitably suffers from highly spatially localized multi-vortex alternating interference. When the rotor cuts through the scattered vortices, it may trigger asymmetric high-frequency aerodynamic impacts. Existing frequency domain filtering technology is unable to separate these local abrupt changes from the structural resonance or electrical noise of the airframe itself, directly leading to high-frequency ineffective oscillations. Secondly, real vortices exhibit a spatial sweeping propagation phenomenon on the fuselage surface, similar to water waves. Existing error-triggered passive feedback mechanisms ignore the propagation delay of disturbances between rotor nodes and lack the ability to predict them in advance. Furthermore, when a single rotor is disturbed, it is inevitably accompanied by strong coupling between lift and attitude torque. Existing static hybrid control logic will cause cross-interference between the altitude loop and the attitude loop, and it is difficult to assess the aerodynamic margin of each rotor in real time during extreme transient compensation cycles, leading to physical saturation and thrust collapse of the single-side motor due to control command overflow. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for adaptive control optimization of low-altitude equipment based on local wind interference, comprising:

[0006] The rotational speed signal, phase current signal, and attitude error of the rotor motor in the low-altitude equipment are collected. The thrust state of the rotor is calculated based on the rotational speed signal and phase current signal, and a virtual force surface is constructed. The local geometric distortion features of the virtual force surface are extracted and denoised to obtain the real aerodynamic disturbance components.

[0007] Based on the physical connection relationship of the rotor, a rotor adjacency topology graph is constructed, the cross-node propagation delay of the real aerodynamic disturbance component is tracked, an aerodynamic propagation chain is constructed, and a feedforward compensation amount is generated.

[0008] The cross-influence matrix of lift and torque is calculated using the tension state to construct an orthogonal decoupling space; a basic attitude control command is generated based on the attitude error; the feedforward compensation amount and the basic attitude control command are superimposed and projected onto the orthogonal decoupling space to obtain the decoupling compensation component;

[0009] Based on the upper and lower limits of rotor thrust, an reachable polyhedron is constructed. The decoupling compensation component is truncated and converted into rotor speed control commands through the reachable polyhedron, thereby realizing the control optimization of low-altitude equipment based on local wind interference resistance.

[0010] As a preferred embodiment of the adaptive control optimization method for low-altitude equipment based on local wind interference described in this invention, the calculation of the rotor thrust state and the construction of a virtual force surface based on the rotation speed signal and phase current signal includes: extracting the electromagnetic torque components of each phase winding through the phase current signal; obtaining the rotor disk rotation speed of each rotor through the rotation speed signal; and calculating the real-time thrust of each rotor based on the electromagnetic torque components and rotor disk rotation speed to obtain the thrust state.

[0011] Using the normal offset of the thrust state of each rotor relative to the reference plane of the airframe as the ordinate, the normal offset is mapped to the corresponding node according to the spatial coordinates of each rotor node in the airframe coordinate system. Discrete interpolation is performed on the normal offset between adjacent nodes to obtain a virtual force surface covering the entire rotor distribution range of the low-altitude equipment. The virtual force surface is updated in each control cycle with the real-time thrust of each rotor, reflecting the current spatial distribution state of the aerodynamic load in real time.

[0012] As a preferred embodiment of the adaptive control optimization method for low-altitude equipment based on local wind interference described in this invention, the following steps are included: extracting the local geometric distortion features of the virtual force-bearing surface and performing denoising processing to obtain the real aerodynamic disturbance components; calculating the curvature of the normal offset of each node in the virtual force-bearing surface to obtain the local Gaussian curvature of each node; marking nodes whose local Gaussian curvature exceeds the global background curvature field mean as curvature distortion candidate points; and constructing local geometric distortion features using the spatial location of the curvature distortion candidate points and their corresponding local Gaussian curvature.

[0013] The spatial migration trajectory of the curvature distortion candidate points is tracked within consecutive time frames. Based on the migration direction of the curvature distortion candidate points in adjacent time frames and the direction of the line connecting adjacent rotor nodes, the curvature distortion candidate points with consistent directions are identified as real aerodynamic disturbances with spatial sweep trajectories and retained. The curvature distortion candidate points with inconsistent directions are identified as transient distortion points caused by structural coupling or motor commutation and eliminated. The retained curvature distortion candidate points and local Gaussian curvature are used as real aerodynamic disturbance components.

[0014] As a preferred embodiment of the adaptive control optimization method for low-altitude equipment based on local wind interference described in this invention, the following steps are included: tracking the cross-node propagation delay of the actual aerodynamic disturbance component, constructing an aerodynamic propagation chain, and generating feedforward compensation: on the rotor adjacency topology, taking the rotor node where the curvature distortion candidate point first appears in the actual aerodynamic disturbance component as the starting node, determining each downstream node sequentially according to the migration order of the curvature distortion candidate point between adjacent rotor nodes, and recording the time interval between the migration of the curvature distortion candidate point between adjacent nodes as the cross-node propagation delay; constructing an aerodynamic propagation chain based on the starting node, the downstream nodes, and the cross-node propagation delay; updating the aerodynamic propagation chain in each control cycle according to the latest state of the actual aerodynamic disturbance component.

[0015] Based on the cross-node propagation delay corresponding to each downstream node in the aerodynamic propagation chain, a cancellation command is applied to the control channel of each downstream node in advance at the current moment; the amplitude of the cancellation command is determined based on the local Gaussian curvature of the actual aerodynamic disturbance component at the corresponding node; the set of cancellation commands for all downstream nodes is statistically analyzed to obtain the feedforward compensation amount.

[0016] As a preferred embodiment of the adaptive control optimization method for low-altitude equipment based on local wind interference described in this invention, the method for constructing an orthogonal decoupling space by calculating the cross-influence matrix of lift and torque using the tension state includes using a rotor aerodynamic model to back-calculate the current rotational speed of each rotor using the real-time tension of each rotor in the tension state.

[0017] Based on the current rotational speed of each rotor, the partial derivatives of the unit rotational speed change of each rotor with respect to the total lift of the entire aircraft and with respect to the roll moment, pitch moment and yaw moment are calculated and used as row and column elements to construct a cross-influence matrix of the combined effect of single rotor output change on the lift and moment of the entire aircraft.

[0018] Singular value decomposition is performed on the cross-influence matrix to extract cross-coupling components that represent the degree of mutual interference between the lift channel and each torque channel. With the goal of minimizing the cross-coupling components, the attitude compensation direction vector that has the least impact on the total lift and the lift compensation direction vector that has the least impact on the torque of each axis are solved in the control command space. An orthogonal decoupling space is constructed based on the attitude compensation direction vector and the lift compensation direction vector.

[0019] As a preferred embodiment of the adaptive control optimization method for low-altitude equipment based on local wind interference described in this invention, the following steps are included: constructing an reachable polyhedron based on the upper and lower limits of the rotor's thrust, which involves determining the upper and lower limits of the thrust that each rotor can currently output based on the real-time thrust of each rotor in the thrust state, combined with the physical output boundaries of each rotor motor; and calculating the rotational speed based on the upper and lower limits of the thrust of each rotor to obtain the upper and lower limits of the rotational speed of each rotor.

[0020] By combining the upper and lower speed limits of each rotor with the cross-influence matrix, the contribution of each rotor to the total lift, roll moment, and pitch moment under the upper and lower speed limits is obtained. Based on the combined boundary of the contributions of all rotors, a convex hull region is formed in the three-dimensional control space with the total lift, roll moment, and pitch moment as coordinate axes, thus obtaining an reachable polyhedron.

[0021] As a preferred embodiment of the adaptive control optimization method for low-altitude equipment based on local wind interference described in this invention, the method of truncating the decoupling compensation component and converting it into rotor speed control command through the reachable polyhedron includes: constructing a directional ray with the geometric center of the reachable polyhedron as the starting point and the pointing direction of the decoupling compensation component in the three-dimensional control space as the ray direction; traversing each boundary surface of the reachable polyhedron, calculating the intersection points of the directional ray with each boundary surface, taking the intersection point closest to the geometric center as the truncation point, and using the distance from the geometric center to the truncation point as the physical reachability limit of the low-altitude equipment in the control direction corresponding to the directional ray;

[0022] Based on the comparison between the decoupling compensation component and the physical reach limit, the decoupling compensation component whose amplitude exceeds the physical reach limit is scaled to the cutoff point along the direction ray; the decoupling compensation component whose amplitude does not exceed the physical reach limit remains unchanged, thus obtaining the truncated decoupling compensation component.

[0023] The decoupling compensation components after truncation are distributed to each rotor through the inverse mapping of the cross-influence matrix to generate rotor speed control commands.

[0024] An adaptive control optimization system for low-altitude equipment based on local wind interference resistance, wherein:

[0025] The aerodynamic disturbance module collects the rotational speed signal, phase current signal, and attitude error of the rotor motor in the low-altitude equipment. Based on the rotational speed signal and phase current signal, it calculates the rotor's thrust state and constructs a virtual force surface. It extracts the local geometric distortion features of the virtual force surface and performs noise reduction processing to obtain the real aerodynamic disturbance components.

[0026] The feedforward compensation module constructs a rotor adjacency topology graph based on the physical connection relationship of the rotor, tracks the cross-node propagation delay of the real aerodynamic disturbance component, constructs an aerodynamic propagation chain, and generates a feedforward compensation amount.

[0027] The command decoupling module calculates the cross-influence matrix of lift and torque using the tension state to construct an orthogonal decoupling space; it generates basic attitude control commands based on the attitude error, superimposes the feedforward compensation amount and the basic attitude control commands, and projects and decomposes them onto the orthogonal decoupling space to obtain decoupling compensation components.

[0028] The command truncation module constructs an reachable polyhedron based on the upper and lower limits of the rotor's thrust. The reachable polyhedron is used to truncate the decoupling compensation component and convert it into rotor speed control commands, thereby achieving low-altitude equipment control optimization based on local wind interference resistance.

[0029] A computer device includes: a memory and a processor; the memory stores a computer program, wherein: when the processor executes the computer program, it implements the steps of the method described in any one of the present invention.

[0030] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described in any one of the present invention.

[0031] The beneficial effects of this invention are as follows: By constructing a virtual force surface and extracting local curvature distortion, combined with the continuity of spatial migration trajectory to eliminate transient noise, this invention achieves morphological separation between real wind disturbance and electrical noise, improving the accuracy of disturbance signal extraction; based on the rotor topology map, it tracks the cross-node propagation delay of disturbances and injects cancellation commands in advance before the wave peak reaches the downstream, improving the response speed of equipment in the face of sudden wind disturbances; by constructing an orthogonal decoupling space through real-time calculation of the cross-influence matrix, it isolates lift and attitude compensation, reduces multi-axis cross-interference under asymmetric wind disturbances, and enhances the stability of multi-channel coordinated adjustment; by mapping the rotor physical limit to a three-dimensional reachable polyhedron and using the intersection of directional rays to geometrically truncate commands, the safe compensation range shrinks as the rotor aerodynamic capability decays, effectively suppressing the overload of motor thrust in transient compensation. By overcoming the local interference caused by the airflow boundary layer in low-altitude operations, it reduces the ineffective attitude oscillation of low-altitude equipment, improving the adaptive anti-disturbance capability and operational stability of low-altitude equipment. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is an overall flowchart of an adaptive control optimization method for low-altitude equipment based on local wind interference, provided in an embodiment of the present invention.

[0034] Figure 2 A flowchart of truncation and instruction mapping based on reachable polyhedra provided for embodiments of the present invention.

[0035] Figure 3 This is a computer device diagram illustrating an adaptive control optimization method for low-altitude equipment based on local wind interference, provided in an embodiment of the present invention. Detailed Implementation

[0036] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0037] Example 1, referring to Figures 1-3 As an embodiment of the present invention, a method for adaptive control optimization of low-altitude equipment based on local wind interference is provided, comprising:

[0038] S1: Collect the rotational speed signal, phase current signal, and attitude error of the rotor motor in the low-altitude equipment. Calculate the rotor's thrust state based on the rotational speed signal and phase current signal and construct a virtual force surface. Extract the local geometric distortion features of the virtual force surface and perform noise reduction processing to obtain the real aerodynamic disturbance components.

[0039] In this embodiment, the acquisition of rotation speed signals, phase current signals, and attitude errors of the rotor motors in the low-altitude equipment includes: the control system synchronously acquires the rotation speed signals and phase current signals of each rotor motor from the drive circuit of each rotor motor in the low-altitude equipment, and obtains the current attitude error of the airframe from the airborne inertial measurement unit; wherein, the rotation speed signal reflects the current rotor disk angular velocity state of each rotor, the phase current signal carries the electromagnetic load information of each phase winding, and the attitude error serves as the feedback quantity for attitude control.

[0040] It should be noted that during the flight of low-altitude equipment, the aerodynamic disturbances experienced by each rotor are not uniformly distributed. Local vortex intrusion often only affects individual rotor areas, manifested as a local deviation of the real-time thrust of individual rotors relative to the rest of the rotors. If the thrust of each rotor is monitored independently, it is impossible to perceive the spatial distribution of the disturbance between the rotors, and thus it is difficult to determine the source direction and range of the disturbance. It is necessary to map the real-time thrust information of each rotor into a unified spatial geometric expression in order to perceive the local characteristics of aerodynamic disturbances from the perspective of the overall spatial form.

[0041] Specifically, calculating the rotor thrust state and constructing a virtual force surface based on the rotational speed signal and phase current signal includes the following steps: extracting the electromagnetic torque components of each phase winding in each rotor motor through the phase current signal; obtaining the rotor disk rotational speed through the rotational speed signal; and calculating the real-time thrust of each rotor based on the electromagnetic torque components and rotor disk rotational speed through rotor aerodynamic relationships, thereby obtaining the thrust state of each rotor. In this embodiment, the standard electromagnetic torque estimation method in the field of rotor motor drive is adopted. The instantaneous electromagnetic torque components of each phase winding are calculated through the phase current signal, and combined with the rotor disk rotational speed, the real-time thrust of each rotor is obtained according to the correspondence between thrust, motor output power, and blade aerodynamic coefficient.

[0042] Taking a six-rotor low-altitude equipment as an example, the six rotor nodes are numbered as follows: to Then the first The real-time thrust of each rotor is denoted as The rated thrust reference value of each rotor when the aircraft is hovering is denoted as: The normal direction of the fuselage reference plane is the positive z-axis direction; rotor node normal offset The calculation formula is as follows:

[0043]

[0044] in, This represents the total number of rotors; For rotor node The normal offset.

[0045] Furthermore, the normal offset of the thrust state of each rotor relative to the airframe reference plane is considered. The vertical axis represents the coordinates of each rotor node in the airframe coordinate system. The normal offset of each node Mapping to the corresponding node, the normal offset between adjacent nodes is discretized to obtain a virtual force surface covering the entire rotor distribution range of the low-altitude equipment; the virtual force surface is updated synchronously with the real-time thrust of each rotor in each control cycle, reflecting the current spatial distribution state of the aerodynamic load in real time.

[0046] It should be noted that by mapping the dispersed thrust information of each rotor into a unified spatial surface expression, the originally independent thrust changes of each rotor form a continuous spatial morphological description in the same geometric space. This overcomes the limitation that the spatial distribution of disturbances cannot be perceived when monitoring each rotor individually, and provides a unified data carrier for extracting local aerodynamic disturbance characteristics.

[0047] After obtaining the virtual force surface, it is necessary to further extract characteristic quantities that can reflect local aerodynamic disturbances. Among them, low-altitude equipment faces two types of disturbance sources during flight: one is the overall attitude change of the airframe (such as slow pitch or roll), at which time the overall thrust of each rotor changes in coordination, and the virtual force surface tends to be coplanar; the other is the intrusion of local vortices into a certain rotor area, causing a local sudden change in the regional thrust, and the corresponding node shows obvious non-coplanar warping. The two types of changes are fundamentally different in geometric shape. Therefore, by using the local Gaussian curvature as a characteristic quantity, the two types of changes can be separated at the geometric shape level. The Gaussian curvature corresponding to the global coplanar change approaches zero, while the Gaussian curvature corresponding to the local non-coplanar warping increases.

[0048] Specifically, the local geometric distortion features of the virtual force-bearing surface are extracted and denoised to obtain the real aerodynamic disturbance components. The specific process is as follows: the curvature of the normal offset of each node in the virtual force-bearing surface is calculated to obtain the local Gaussian curvature of each node. In this embodiment, the standard finite difference calculation method for local Gaussian curvature is adopted. The specific calculation process is as follows: the normal offset at each node in the virtual force-bearing surface is calculated by performing finite difference second-order partial derivatives along the longitudinal axis and transverse axis of the body, respectively, to obtain the second-order partial derivatives of the longitudinal axis and the transverse axis at each node, and then the mixed partial derivative at each node is calculated; the local Gaussian curvature of the node is obtained by subtracting the square of the mixed partial derivative from the product of the second-order partial derivative of the longitudinal axis and the second-order partial derivative of the transverse axis, as shown in the specific formula below:

[0049]

[0050] in, For nodes The second partial derivative of the normal offset along the vertical axis; For nodes The second partial derivative of the normal offset along the horizontal axis; For nodes Mixed partial derivatives at; For nodes The local Gaussian curvature at that point.

[0051] After obtaining the local Gaussian curvature of each node, the local Gaussian curvature of all nodes within the current control cycle is statistically analyzed to form a global background curvature field, and the mean value of the global background curvature field is calculated. Local Gaussian curvature Exceeding the global background curvature field mean The nodes are marked as candidate points for curvature distortion; local geometric distortion features are constructed using the spatial location of the candidate points for curvature distortion and the corresponding local Gaussian curvature.

[0052] Furthermore, since the local geometric distortion features of a single frame are insufficient to determine the nature of the disturbance corresponding to the curvature distortion candidate point, the coupled vibration of the airframe structure or the commutation of the motor can also produce an abnormal increase in the local Gaussian curvature on the virtual force surface. However, this type of noise is an in-situ excitation, and its curvature distortion candidate point disappears after appearing in situ, without forming a directional spatial migration between adjacent rotor nodes. In contrast, real aerodynamic disturbances (such as local vortices) must, in terms of physical mechanism, form a spatial sweep trajectory with a clear migration path between adjacent rotors along the airflow sweep direction. Therefore, the spatial migration attribute of the curvature distortion candidate point within a continuous time frame can be used as a classification criterion to distinguish between the two types of sources.

[0053] Specifically, the spatial migration trajectory of the curvature distortion candidate points is tracked within consecutive time frames. The judgment is made based on the migration direction of the curvature distortion candidate points and the direction of the line connecting adjacent rotor nodes within adjacent time frames: curvature distortion candidate points with consistent directions are judged as real aerodynamic disturbances with spatial sweep trajectories and retained; curvature distortion candidate points with inconsistent directions are judged as transient distortion points caused by structural coupling or motor commutation and eliminated; the retained curvature distortion candidate points and local Gaussian curvature are used as real aerodynamic disturbance components.

[0054] It should be noted that, in response to the problem that traditional frequency domain filtering is difficult to distinguish between real aerodynamic disturbances and structural electrical noise in the same frequency band, a virtual force surface is constructed to transform the real-time thrust state of each rotor into a spatial geometric shape. Local Gaussian curvature is used to separate global attitude changes and local aerodynamic intrusions in the geometric domain. Furthermore, the spatial propagation continuity criterion is introduced to identify candidate points of curvature distortion, thus accurately separating real aerodynamic disturbances with clear spatial sweep trajectories from transient noise of in-situ excitation. This preserves a complete time margin and ensures the purity and timeliness of real aerodynamic disturbance components.

[0055] S2: Based on the physical connection relationship of the rotor, construct the rotor adjacency topology graph, track the cross-node propagation delay of the real aerodynamic disturbance component, construct the aerodynamic propagation chain, and generate the feedforward compensation amount.

[0056] Furthermore, although the actual aerodynamic disturbance components are obtained through step S1, which include the spatial position sequence of each curvature distortion candidate point and the corresponding local Gaussian curvature amplitude, simply knowing which rotor node the disturbance has currently intruded into is insufficient to achieve advance suppression. Traditional wind resistance control intervenes to compensate only after the disturbance effect is sensed, which is essentially a post-event response and cannot eliminate the attitude deviation that has already occurred when the disturbance peak arrives. At the same time, since local vortices physically sweep adjacent rotors sequentially along the airflow direction, if a cancellation command can be applied in advance before the disturbance peak reaches the downstream rotor, the lag of post-event compensation can be fundamentally eliminated.

[0057] Specifically, constructing a rotor adjacency topology graph based on the physical connection relationships of the rotors includes: using each rotor node in the low-altitude equipment as the vertices of the graph, and the physical connection relationships of the arms between each rotor node as the edges of the graph; recording the spatial coordinates of each rotor node and the direction vectors of the lines connecting adjacent nodes in the body coordinate system to construct a discrete rotor adjacency topology graph. In this embodiment, the construction of the rotor adjacency topology graph adopts the standard modeling method of directed adjacency graphs in graph theory. The specific modeling process is as follows: storing the spatial coordinates of each rotor node in the body coordinate system as vertex attributes in the node table; traversing the physical arm connection relationships of each rotor node, storing the unit vectors of the line directions between two physically connected node pairs as edge attributes in the edge table; and finally forming a topology graph structure containing all rotor nodes and their adjacency relationships.

[0058] For example, taking a six-rotor low-altitude equipment as an example, the six rotor nodes to The nodes are arranged in a ring shape, with physical connections between adjacent nodes via robotic arms. The adjacency topology contains six vertices and six edges, with each edge carrying a unit vector indicating the direction of the connection between the corresponding node pairs. :

[0059]

[0060] in, and Rotor nodes and rotor node Spatial coordinates in the body coordinate system; For rotor node Pointing to rotor node The direction vector of the connecting line; Let be the magnitude of the vector.

[0061] It should be noted that the rotor adjacency topology is constructed based on the actual physical connection relationship of the airframe, rather than using the assumption of continuous manifold or full connectivity. This ensures that the topology is consistent with the actual aerodynamic disturbance propagation path of the airframe, providing a spatial reference framework that conforms to physical reality for subsequent tracking of the migration trajectory of candidate points of curvature distortion on the topology.

[0062] After obtaining the rotor adjacency topology, it is necessary to further trace the propagation law of the real aerodynamic disturbance components on it. The real aerodynamic disturbance components output in step S1 contain the spatial position sequence of the curvature distortion candidate points in a continuous time frame. The spatial position sequence in the rotor adjacency topology is represented by the path of the distortion center starting from a certain starting node and migrating along the edge to each downstream node. By recording the time interval consumed by the migration of the distortion center between each adjacent node, the cross-node propagation delay of each segment can be obtained. Then, the combination of the starting node, the downstream node sequence and the cross-node propagation delay of each segment constitutes the aerodynamic propagation chain.

[0063] Specifically, tracking the cross-node propagation delay of the actual aerodynamic disturbance component, constructing the aerodynamic propagation chain, and generating feedforward compensation includes the following steps: On the rotor adjacency topology, taking the rotor node where the curvature distortion candidate point first appears in the actual aerodynamic disturbance component as the starting node, determining each downstream node sequentially according to the migration order of the curvature distortion candidate point between adjacent rotor nodes, and recording the time interval between the migration of the curvature distortion candidate point between adjacent nodes as the cross-node propagation delay; constructing the aerodynamic propagation chain based on the starting node, the downstream nodes, and the cross-node propagation delay.

[0064] Specifically, assume that the candidate point for curvature distortion first appears at node The process migrates sequentially through downstream nodes. The cross-node propagation delay between adjacent nodes is denoted as The specific formula for the aerodynamic propagation chain is as follows:

[0065]

[0066] in, The starting node; For the first One downstream node; For from the first downstream node to the first Cross-node propagation delay of downstream nodes; The total number of downstream nodes; the aerodynamic propagation chain is updated in each control cycle according to the latest migration status of the actual aerodynamic disturbance components to track changes in the eddy sweep path.

[0067] It should be noted that, compared to the method of calculating tangential velocity on a continuous manifold, using a discrete rotor adjacency topology graph as a carrier to track the node-by-node migration of candidate points for curvature distortion makes the extraction of propagation paths completely fit the actual discrete rotor distribution structure of the aircraft, avoiding the approximation error introduced by the continuous assumption in the sparse rotor node scenario, and ensuring that the aerodynamic propagation chain accurately depicts the real disturbance propagation law.

[0068] After obtaining the aerodynamic propagation chain, the cross-node propagation delay of each downstream node recorded therein can be used to apply a cancellation command to the control channel of the corresponding downstream node in advance at the current moment, thereby completing the compensation before the eddy wave peak arrives and realizing the transformation from reactive response to proactive suppression.

[0069] Specifically, based on the cross-node propagation delay corresponding to each downstream node in the aerodynamic propagation chain, a cancellation command is applied to the control channel of each downstream node in advance at the current moment; the amplitude of the cancellation command is determined based on the local Gaussian curvature of the actual aerodynamic disturbance component at the corresponding node; the set of cancellation commands for all downstream nodes is statistically analyzed to obtain the feedforward compensation amount. Specifically, in this embodiment, a linear mapping method is used to determine the cancellation command. The specific process is as follows: the local Gaussian curvature amplitude at the corresponding downstream node in the actual aerodynamic disturbance component is taken, and the amplitude of the cancellation command to be applied to the control channel of the corresponding node is calculated according to the pre-calibrated proportional relationship between the curvature amplitude and the rotor speed compensation amount. The pre-calibrated proportional relationship is obtained by aerodynamic bench tests before the low-altitude equipment leaves the factory.

[0070] Secondly, for downstream nodes, the timing of the cancellation command application is as follows:

[0071]

[0072] in, This is the current control moment (i.e., the moment when the candidate point of curvature distortion first appears at the starting node). The first in the aerodynamic propagation chain Cross-node propagation delay of the segment; The expected arrival time of the eddy current peak at the downstream node. At the critical moment, by applying the cancellation command in advance, we ensure that the compensation action arrives before the disturbance peak.

[0073] It should be noted that since wind resistance control can only intervene to compensate after the disturbance has occurred, there is an inherent response lag. By tracing the node-by-node migration path and cross-node propagation delay of the real aerodynamic disturbance components on the discrete rotor adjacency topology, an aerodynamic propagation chain is established. Utilizing the time window required for eddies to propagate between adjacent rotor nodes, a cancellation command corresponding to the amplitude and local Gaussian curvature is injected into the control channel in advance before the disturbance peak reaches the downstream rotor. This advances the intervention timing of wind resistance control from after the disturbance occurs to before the disturbance arrives, realizing the transformation from passive compensation to active prediction and suppression.

[0074] S3: Calculate the cross-influence matrix of lift and torque using the tension state, and construct an orthogonal decoupling space; generate basic attitude control commands based on the attitude error, superimpose the feedforward compensation amount and the basic attitude control commands, and project and decompose them onto the orthogonal decoupling space to obtain the decoupling compensation components.

[0075] When the feedforward compensation and the basic attitude control commands generated based on attitude errors are superimposed and directly issued to each rotor, an inherent physical conflict arises: each rotor of a multi-rotor low-altitude equipment simultaneously bears the dual responsibility of generating total lift and differential torque. Adjusting the rotational speed of any rotor changes both the total lift and the attitude torque, creating an unavoidable cross-coupling between the altitude control loop and the attitude control loop. Adjustment in one loop often triggers a response deviation in the other. Without decoupling, the superimposed composite commands will interfere with each rotor, weakening the anti-disturbance effect and even causing overall aircraft oscillation. Therefore, it is necessary to calculate the cross-influence relationship between each rotor output on lift and torque in real time based on the current transient aerodynamic state, construct a physically meaningful orthogonal decoupling space, and project the superimposed composite commands into this space to achieve decoupling and separation of the lift channel and the torque channel.

[0076] Specifically, the calculation of the cross-influence matrix of lift and torque using the aforementioned thrust state, and the construction of an orthogonal decoupling space, includes the following steps: First, using the rotor aerodynamic model, the current rotational speed of each rotor is inferred from the real-time thrust of each rotor in the aforementioned thrust state. The specific calculation process is as follows: Based on the aerodynamic relationship that the rotor thrust is proportional to the square of the rotor disk rotational speed, the entropy of the real-time thrust of each rotor and the corresponding rotor aerodynamic coefficient is calculated, and then the arithmetic square root of the result is taken to obtain the current rotational speed of each rotor; wherein the rotor aerodynamic coefficient is determined by the blade geometric parameters and air density, and is calibrated when the low-altitude equipment leaves the factory.

[0077] After obtaining the current rotational speed of each rotor, the partial derivatives of the unit rotational speed change of each rotor with respect to the total lift of the entire aircraft, and with respect to the roll moment, pitch moment, and yaw moment, are calculated based on the current rotational speed of each rotor. A cross-influence matrix is ​​then constructed using these partial derivatives as row and column elements. Specifically, in this embodiment, a Jacobian matrix is ​​used for construction. The specific modeling process is as follows: For the first... For each rotor, calculate the partial derivatives of the unit change in speed with respect to total lift, roll moment, pitch moment, and yaw moment at its current rotational speed. Using the output values ​​as rows and the rotor number as columns, fill the corresponding row and column positions with the partial derivatives of all rotors with respect to all output values, forming four rows. Column cross-influence matrix , of which Line number The elements of the column represent the first... The change in unit rotational speed of the first rotor relative to the first... The degree of influence of each output quantity is described, thereby describing the comprehensive cross-influence relationship between the output changes of a single rotor on the lift and torque of the entire aircraft.

[0078] It should be noted that the partial derivatives are calculated based on the current real-time rotational speed of each rotor, rather than using parameters such as the rated hovering speed. This ensures that the cross-influence matrix accurately reflects the real coupling relationship under the current transient aerodynamic state, avoiding model mismatch caused by the aerodynamic state deviating from the rated point.

[0079] After obtaining the cross-influence matrix, it is necessary to further extract the coupling structure between the lift channel and the torque channel, and construct an orthogonal decoupling space that makes the two channels independent of each other.

[0080] Specifically, singular value decomposition is performed on the cross-influence matrix to extract the cross-coupling components representing the degree of mutual interference between the lift channel and each torque channel. In this embodiment, singular value decomposition is used to calculate the cross-influence matrix. It is decomposed into the product of a left singular matrix, a diagonal matrix of singular values, and a right singular matrix. The distribution of singular values ​​reflects the coupling strength between each output channel and corresponds to the main direction of the cross-influence between the lift channel and each torque channel.

[0081] Furthermore, with the goal of minimizing the cross-coupling components, the attitude compensation direction vector that minimizes the impact on the total lift and the lift compensation direction vector that minimizes the impact on the moments of each axis are solved in the control command space. Based on these attitude compensation direction vectors and lift compensation direction vectors, the orthogonal decoupling space is constructed. Specifically, the solution process involves extracting the null space direction corresponding to the singular values ​​of the lift channels from the right singular matrix of the singular value decomposition, which is then used as the attitude compensation direction vector that minimizes the impact on the total lift. Extract the null space direction corresponding to the singular value of the torque channel, and use it as the lift compensation direction vector with the least impact on the torque of each axis. ; based on attitude compensation direction vector and lift compensation direction vector Constructing an orthogonal decoupling space The specific formula is as follows:

[0082]

[0083] in, The lift compensation direction vector that has the least impact on the torque of each axis; The attitude compensation direction vector that has the least impact on the total lift is zero, and the inner product of the two is zero to ensure the independence of the lift channel and the torque channel after projection. It is an orthogonal decoupling space.

[0084] It should be noted that the null space direction of the cross-influence matrix under the current transient aerodynamic state is used as the decoupling basis, so that the orthogonal decoupling space can be adjusted in real time with the change of aerodynamic state, ensuring that the decoupling accuracy does not decrease due to model mismatch under local aerodynamic degradation conditions, and reducing the failure of the decoupling method under disturbed conditions.

[0085] Furthermore, after completing the construction of the orthogonal decoupling space, it is necessary to merge the feedforward compensation amount output in step S2 with the generated basic attitude control command as a unified composite command input, and then project and decompose it onto the orthogonal decoupling space to obtain the decoupled compensation components.

[0086] Specifically, a basic attitude control command is generated based on the attitude error. The feedforward compensation amount and the basic attitude control command are superimposed and projected onto the orthogonal decoupling space to obtain the decoupling compensation component. The specific process is as follows: Based on the attitude error collected in step S1, the attitude controller calculates the roll, pitch, and yaw differential compensation amounts to be applied at the current moment to form the basic attitude control command. In this embodiment, the attitude controller adopts the proportional-integral-derivative control method to calculate the basic attitude control command based on the attitude error.

[0087] The feedforward compensation and the basic attitude control command are superimposed to obtain a composite control command. This composite control command is then directed to an orthogonal decoupling space. Projective decomposition is performed to obtain the decoupling compensation components. Specifically, the composite control commands are directed towards the attitude compensation direction vector. and lift compensation direction vector Orthogonally project the direction to obtain the lift compensation component and attitude compensation component, and combine the two to form the decoupling compensation component.

[0088] It should be noted that, in response to the inherent cross-coupling between the lift channel and the torque channel in multi-rotor low-altitude equipment due to rotor actuator sharing, where a single adjustment causes a deviation in the overall aircraft response, the cross-influence matrix is ​​calculated in real time based on the current transient thrust state. The zero-space direction is extracted from this matrix to construct an orthogonal decoupling space. The composite command resulting from the superposition of the feedforward compensation and the basic attitude control command is projected onto the orthogonal decoupling space, ensuring that the lift compensation component and the attitude compensation component are physically orthogonal and do not interfere with each other. This reduces the cross-coupling resonance of the control loop caused by actuator sharing, while ensuring that the feedforward compensation is not reduced by the coupling effect during the superposition process, thus preserving the complete effect of advance suppression.

[0089] S4: Construct an reachable polyhedron based on the upper and lower limits of the rotor's thrust, and use the reachable polyhedron to truncate the decoupling compensation component and convert it into rotor speed control commands to achieve low-altitude equipment control optimization based on local wind interference resistance.

[0090] As a directional vector in the control command, the decoupling compensation component may not be within the physical output capability range of each rotor. When a local aerodynamic disturbance causes a decrease in the aerodynamic efficiency of a rotor, the upper limit of the rotor's available thrust will also decrease. If a command that exceeds the physical boundary is directly issued, it will lead to motor saturation instability. It is necessary to adjust the decoupling compensation component by using the current real-time upper and lower limits of the thrust of each rotor to push the disturbance rejection command limit to the edge of the physical limit.

[0091] Specifically, such as Figure 2 The diagram shows the truncation and command mapping flowchart based on reachable polyhedra. Constructing reachable polyhedra based on the upper and lower limits of rotor thrust includes the following steps: Based on the real-time thrust of each rotor in the thrust state, and combined with the physical output boundary of each rotor motor, determine the upper and lower limits of the thrust that each rotor can currently output; wherein, the process of determining the upper and lower limits of the thrust of each rotor is as follows: taking the maximum electromagnetic torque corresponding to the rated maximum current of each rotor motor as the physical upper limit, and calculating the upper limit of the thrust of each rotor through the rotor aerodynamic relationship in combination with the current rotor disk speed; taking the aerodynamic output corresponding to the current minimum stable speed of each rotor as the lower limit of the thrust, when a local aerodynamic disturbance causes a decrease in the aerodynamic efficiency of a certain rotor region, the upper limit of the thrust of the corresponding rotor decreases accordingly.

[0092] After obtaining the upper and lower limits of the thrust of each rotor, the rotational speed is calculated based on the upper and lower limits of the thrust of each rotor to obtain the upper and lower limits of the rotational speed of each rotor. This includes converting the upper and lower limits of the thrust of each rotor into the corresponding upper and lower limits of the rotational speed, based on the aerodynamic relationship that the rotor thrust is proportional to the square of the rotor disk rotational speed.

[0093] By using the upper and lower speed limits of each rotor, combined with the cross-influence matrix, the contribution of each rotor to the total lift, roll moment, and pitch moment under the upper and lower speed limits is obtained. Based on the combined boundary of all rotor contributions, a convex hull region is formed in the three-dimensional control space with total lift, roll moment, and pitch moment as coordinate axes, thus obtaining an reachable polyhedron. Specifically, the construction process of the reachable polyhedron is as follows: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] Substitute the upper and lower speed limits of each rotor into the cross-influence matrix. The corresponding columns are used to obtain the contributions of the rotor to the total lift, roll moment, and pitch moment under two limiting speed conditions. All rotors are traversed, and all combinations of rotor speeds taking the upper or lower limit are enumerated. The three-dimensional control space coordinates obtained by mapping each combination through the cross-influence matrix are collected as a vertex set. A convex hull calculation is performed on the vertex set in the three-dimensional control space, and the interior of the convex hull region enclosed by all vertices is taken as the reachable polyhedron.

[0094] In each control cycle, the reachable polyhedron is reconstructed based on the latest state of the upper and lower limits of the thrust of each rotor. When a local aerodynamic disturbance causes the upper limit of the thrust of a certain rotor to decrease, the corresponding limit vertex of the rotor contracts inward, and the reachable polyhedron undergoes asymmetric deformation in the corresponding direction. This reflects the real reachable control boundary of the actuators of the entire machine in real time, ensuring that the control boundary is strictly consistent with the real combination space of the physical output capabilities of each rotor, reducing the fitting error between the artificial envelope and the real boundary. At the same time, the convex hull polyhedron naturally supports asymmetric deformation in each control direction, which can accurately reflect the directional influence of local aerodynamic degradation in the three-dimensional control space.

[0095] Furthermore, after obtaining the reachable polyhedron, the decoupling compensation component obtained in step S3 needs to be compared with the physical boundary of the reachable polyhedron. The components that exceed the boundary are precisely truncated, and the truncated result is converted into control commands for the rotational speed of each rotor.

[0096] Specifically, the process of truncating the decoupling compensation component and converting it into a rotor speed control command through the reachable polyhedron includes the following steps: constructing a directional ray with the geometric center of the reachable polyhedron as the starting point and the pointing direction of the decoupling compensation component in the three-dimensional control space as the ray direction; wherein, the calculation of the geometric center of the reachable polyhedron adopts the method of the centroid of the convex polyhedron, taking the arithmetic mean of the coordinates of all vertices.

[0097] Traverse each boundary surface of the reachable polyhedron, calculate the intersection point of the directional ray with each boundary surface, take the intersection point closest to the geometric center as the cutoff point, and take the distance from the geometric center to the cutoff point as the physical reach limit of the low-altitude equipment in the control direction corresponding to the directional ray.

[0098] In this embodiment, the specific calculation process for the intersection points of the directional ray and each boundary surface is as follows: For each boundary surface of the reachable polyhedron, it is expressed as a plane equation determined by the normal vector and the intercept; the parametric equation of the directional ray is substituted into each plane equation to solve for the ray parameters and obtain the intersection point coordinates; after traversing all boundary surfaces, the intersection point with the smallest positive ray parameter (i.e., the intersection point closest to the geometric center and located in the positive direction of the ray) is taken as the cutoff point, and the cutoff distance is... for:

[0099]

[0100] in, The coordinates of the cutoff point; The coordinates of the geometric center of the reachable polyhedron; The cutoff distance is the physical limit that low-altitude equipment can reach in the direction of the decoupling compensation component.

[0101] Based on the comparison between the decoupling compensation component and the physical reachability limit: the decoupling compensation component whose amplitude exceeds the physical reachability limit is scaled to the cutoff point along the directional ray direction; the decoupling compensation component whose amplitude does not exceed the physical reachability limit remains unchanged, thus obtaining the truncated decoupling compensation component.

[0102] The truncated decoupling compensation components are distributed to each rotor through the inverse mapping of the cross-influence matrix, generating rotor speed control commands. The specific inverse mapping process is as follows: a matrix pseudo-inverse method is used, with the cross-influence matrix... The pseudo-inverse is mapped to the decoupling compensation component after truncation processing to obtain the required speed adjustment amount for each rotor, and then superimposed with the current speed of each rotor to obtain the final rotor speed control command.

[0103] It should be noted that, in response to the inability to perceive the differences in aerodynamic margins in different control directions, which may lead to overfitting or insufficient truncation under local aerodynamic degradation conditions, the true upper and lower limits of each rotor's thrust are mapped to the three-dimensional control space after being converted by rotational speed and enclosed as an reachable polyhedron. The precise physical reachability limit in the direction of the decoupling compensation component is calculated in real time using vector ray projection. Components exceeding the boundary are scaled to the truncation point along the original direction. Under the premise of absolutely ensuring that each rotor does not exceed its physical output capability, the compression operation is pushed to the physically permissible limit edge. At the same time, the truncated three-dimensional control commands are precisely distributed to each rotor through the pseudo-inverse inverse mapping of the cross-influence matrix, completing the conversion from the control space to the actuator space.

[0104] On the other hand, this embodiment also provides a low-altitude equipment adaptive control optimization system based on local wind interference, which includes:

[0105] The aerodynamic disturbance module collects the rotational speed signal, phase current signal, and attitude error of the rotor motor in the low-altitude equipment. Based on the rotational speed signal and phase current signal, it calculates the rotor's thrust state and constructs a virtual force surface. It then extracts the local geometric distortion features of the virtual force surface and performs noise reduction processing to obtain the real aerodynamic disturbance components.

[0106] The feedforward compensation module constructs a rotor adjacency topology graph based on the physical connection relationship of the rotor, tracks the cross-node propagation delay of the real aerodynamic disturbance component, constructs an aerodynamic propagation chain, and generates a feedforward compensation amount.

[0107] The command decoupling module calculates the cross-influence matrix of lift and torque using the tension state to construct an orthogonal decoupling space; it generates basic attitude control commands based on the attitude error, superimposes the feedforward compensation amount and the basic attitude control commands, and projects and decomposes them onto the orthogonal decoupling space to obtain decoupling compensation components.

[0108] The command truncation module constructs an reachable polyhedron based on the upper and lower limits of the rotor's thrust. The reachable polyhedron is used to truncate the decoupling compensation component and convert it into rotor speed control commands, thereby achieving low-altitude equipment control optimization based on local wind interference resistance.

[0109] like Figure 3 As shown, if the above functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0111] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0112] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for adaptive control optimization of low-altitude equipment based on local wind interference, characterized in that, include: The rotational speed signal, phase current signal, and attitude error of the rotor motor in the low-altitude equipment are collected. The thrust state of the rotor is calculated based on the rotational speed signal and phase current signal, and a virtual force surface is constructed. The local geometric distortion features of the virtual force surface are extracted and denoised to obtain the real aerodynamic disturbance components. Based on the physical connection relationship of the rotor, a rotor adjacency topology graph is constructed, the cross-node propagation delay of the real aerodynamic disturbance component is tracked, an aerodynamic propagation chain is constructed, and a feedforward compensation amount is generated. The cross-influence matrix of lift and torque is calculated using the tensile state, and an orthogonal decoupling space is constructed. Based on the attitude error, a basic attitude control command is generated, the feedforward compensation amount and the basic attitude control command are superimposed, and then projected and decomposed onto the orthogonal decoupling space to obtain the decoupling compensation component. Based on the upper and lower limits of rotor thrust, an reachable polyhedron is constructed. The decoupling compensation component is truncated and converted into rotor speed control commands through the reachable polyhedron, thereby realizing the control optimization of low-altitude equipment based on local wind interference resistance.

2. The adaptive control optimization method for low-altitude equipment based on local wind interference as described in claim 1, characterized in that: Calculating the rotor thrust state and constructing a virtual force surface based on the rotation speed signal and phase current signal includes: extracting the electromagnetic torque component of each phase winding through the phase current signal; obtaining the rotor disk rotation speed through the rotation speed signal; and calculating the real-time thrust of each rotor based on the electromagnetic torque component and rotor disk rotation speed to obtain the thrust state. Using the normal offset of the thrust state of each rotor relative to the reference plane of the airframe as the ordinate, the normal offset is mapped to the corresponding node according to the spatial coordinates of each rotor node in the airframe coordinate system. Discrete interpolation is performed on the normal offset between adjacent nodes to obtain a virtual force surface covering the entire rotor distribution range of the low-altitude equipment. The virtual force surface is updated in each control cycle with the real-time thrust of each rotor, reflecting the current spatial distribution state of the aerodynamic load in real time.

3. The adaptive control optimization method for low-altitude equipment based on local wind interference as described in claim 2, characterized in that: The process of extracting local geometric distortion features of the virtual force-bearing surface and performing denoising processing to obtain the real aerodynamic disturbance components includes: calculating the curvature of the normal offset of each node in the virtual force-bearing surface to obtain the local Gaussian curvature of each node; marking nodes whose local Gaussian curvature exceeds the global background curvature field mean as curvature distortion candidate points; and constructing local geometric distortion features using the spatial location of the curvature distortion candidate points and their corresponding local Gaussian curvature. The spatial position migration trajectory of the curvature distortion candidate points is tracked within consecutive time frames. The judgment is made based on the position migration direction of the curvature distortion candidate points and the connection direction between adjacent rotor nodes within adjacent time frames. Curvature distortion candidate points with consistent directions are judged as real aerodynamic disturbances with spatial sweep trajectories and retained. Curvature distortion candidate points with inconsistent directions are judged as transient distortion points caused by structural coupling or motor commutation and eliminated. The retained curvature distortion candidate points and local Gaussian curvature are used as the actual aerodynamic disturbance components.

4. The adaptive control optimization method for low-altitude equipment based on local wind interference as described in claim 3, characterized in that: The process of tracking the cross-node propagation delay of the actual aerodynamic disturbance component, constructing an aerodynamic propagation chain, and generating feedforward compensation includes: on the rotor adjacency topology, taking the rotor node where the curvature distortion candidate point first appears in the actual aerodynamic disturbance component as the starting node, determining each downstream node sequentially according to the migration order of the curvature distortion candidate point between adjacent rotor nodes, and recording the time interval between the migration of the curvature distortion candidate point between adjacent nodes as the cross-node propagation delay; constructing an aerodynamic propagation chain based on the starting node, the downstream nodes, and the cross-node propagation delay; and updating the aerodynamic propagation chain in each control cycle according to the latest state of the actual aerodynamic disturbance component. Based on the cross-node propagation delay corresponding to each downstream node in the aerodynamic propagation chain, a cancellation command is applied to the control channel of each downstream node in advance at the current moment; the amplitude of the cancellation command is determined based on the local Gaussian curvature of the actual aerodynamic disturbance component at the corresponding node; the set of cancellation commands for all downstream nodes is statistically analyzed to obtain the feedforward compensation amount.

5. The adaptive control optimization method for low-altitude equipment based on local wind interference as described in claim 4, characterized in that: The cross-influence matrix of lift and torque is calculated using the aforementioned tension state, and an orthogonal decoupling space is constructed by using a rotor aerodynamic model to infer the current rotational speed of each rotor from the real-time tension of each rotor in the aforementioned tension state. Based on the current rotational speed of each rotor, the partial derivatives of the unit rotational speed change of each rotor with respect to the total lift of the entire aircraft and with respect to the roll moment, pitch moment and yaw moment are calculated and used as row and column elements to construct a cross-influence matrix of the combined effect of single rotor output change on the lift and moment of the entire aircraft. Singular value decomposition is performed on the cross-influence matrix to extract cross-coupling components that represent the degree of mutual interference between the lift channel and each torque channel. With the goal of minimizing the cross-coupling components, the attitude compensation direction vector that has the least impact on the total lift and the lift compensation direction vector that has the least impact on the torque of each axis are solved in the control command space. An orthogonal decoupling space is constructed based on the attitude compensation direction vector and the lift compensation direction vector.

6. The adaptive control optimization method for low-altitude equipment based on local wind interference as described in claim 5, characterized in that: Constructing an accessible polyhedron based on the upper and lower limits of rotor thrust includes determining the upper and lower limits of the thrust that each rotor can currently output based on the real-time thrust of each rotor in the thrust state, combined with the physical output boundary of each rotor motor; and calculating the rotational speed based on the upper and lower limits of the thrust of each rotor to obtain the upper and lower limits of the rotational speed of each rotor. By combining the upper and lower speed limits of each rotor with the cross-influence matrix, the contribution of each rotor to the total lift, roll moment, and pitch moment under the upper and lower speed limits is obtained. Based on the combined boundary of the contributions of all rotors, a convex hull region is formed in the three-dimensional control space with the total lift, roll moment, and pitch moment as coordinate axes, thus obtaining an reachable polyhedron.

7. The adaptive control optimization method for low-altitude equipment based on local wind interference as described in claim 6, characterized in that: The process of truncating the decoupling compensation component and converting it into rotor speed control commands through the reachable polyhedron includes: constructing a directional ray starting from the geometric center of the reachable polyhedron and using the pointing direction of the decoupling compensation component in the three-dimensional control space as the ray direction; traversing each boundary surface of the reachable polyhedron, calculating the intersection points of the directional ray with each boundary surface, taking the intersection point closest to the geometric center as the truncation point, and using the distance from the geometric center to the truncation point as the physical reachability limit of the low-altitude equipment in the control direction corresponding to the directional ray; Based on the comparison between the decoupling compensation component and the physical reach limit, the decoupling compensation component whose amplitude exceeds the physical reach limit is scaled to the cutoff point along the direction ray; the decoupling compensation component whose amplitude does not exceed the physical reach limit remains unchanged, thus obtaining the truncated decoupling compensation component. The decoupling compensation components after truncation are distributed to each rotor through the inverse mapping of the cross-influence matrix to generate rotor speed control commands.

8. A low-altitude equipment adaptive control optimization system based on local wind interference, employing the method described in any one of claims 1-7, characterized in that: The aerodynamic disturbance module collects the rotational speed signal, phase current signal, and attitude error of the rotor motor in the low-altitude equipment. Based on the rotational speed signal and phase current signal, it calculates the rotor's thrust state and constructs a virtual force surface. It extracts the local geometric distortion features of the virtual force surface and performs noise reduction processing to obtain the real aerodynamic disturbance components. The feedforward compensation module constructs a rotor adjacency topology graph based on the physical connection relationship of the rotor, tracks the cross-node propagation delay of the real aerodynamic disturbance component, constructs an aerodynamic propagation chain, and generates a feedforward compensation amount. The command decoupling module uses the tension state to calculate the cross-influence matrix of lift and torque, and constructs an orthogonal decoupling space; Based on the attitude error, a basic attitude control command is generated, the feedforward compensation amount and the basic attitude control command are superimposed, and then projected and decomposed onto the orthogonal decoupling space to obtain the decoupling compensation component. The command truncation module constructs an reachable polyhedron based on the upper and lower limits of the rotor's thrust. The reachable polyhedron is used to truncate the decoupling compensation component and convert it into rotor speed control commands, thereby achieving low-altitude equipment control optimization based on local wind interference resistance.

9. A computer device, comprising: A memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-7.

Images