A vehicle-mounted modular drone intelligent airport system and control method

Abstract

The present invention relates to the technical field of intelligent control of unmanned aerial vehicle (UAV) airports, and discloses a vehicle-mounted modularized UAV intelligent airport system and a control method. Based on ranging data of platform tags and UAV tags, the ranging data are processed by weighted least squares fitting to obtain a rough position solution of the UAV; the rough position solution and the ranging data are combined to obtain the attitude quaternion of the UAV through multi-layer coupled attitude closure analysis; the rough position solution, the attitude quaternion and the ranging data are fused by extended Kalman filtering to determine a fine position solution of the UAV; based on the fine position solution, a preset landing point and preset trajectory parameters, a desired spiral landing trajectory is generated; and according to the desired spiral landing trajectory and the fine position solution, a control signal of the UAV motor is output through layered adjustment of inner loop control and outer loop control.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control of unmanned aerial vehicle airports, and more specifically, to a vehicle-mounted modular unmanned aerial vehicle intelligent airport system and a control method. Background Art

[0002] As a mobile drone deployment platform, the vehicle-mounted modular drone smart airport integrates UWB positioning base stations, autonomous take-off and landing guidance systems, and energy supply modules to achieve rapid response and all-weather operation capabilities for drones.

[0003] However, the stability of autonomous landing of drones in complex environments at such airports still faces the following problems:

[0004] The traditional UWB base station layout cannot provide an attitude reference, resulting in large-angle flip ambiguity in attitude calculation. Especially in the presence of multipath interference or tag occlusion, the positioning error is large, leading to the risk of unstable landing or even falling. Summary of the Invention

[0005] The present invention provides a vehicle-mounted modularized unmanned aerial vehicle intelligent airport system and a control method, which solve the technical problems existing in the background technology.

[0006] In a first aspect, the present invention provides a vehicle-mounted modular UAV intelligent airport control method, comprising:

[0007] Labels are placed on the vehicle-mounted airport platform and drone in a preset structure, including platform labels and drone labels;

[0008] Initialize and generate a platform coordinate system with the center of the vehicle-mounted airport platform as the origin;

[0009] In response to the distance parameter between any platform tag and any drone tag being less than the preset guidance distance, the landing guidance process is triggered, including:

[0010] Step 1: Based on the ranging data of the platform tag and the drone tag, the ranging data is processed by weighted least squares fitting to obtain a rough solution of the drone's position;

[0011] Step 2: Combining the rough position solution with the ranging data, the attitude quaternion of the UAV is calculated through multi-layer coupled attitude closure analysis.

[0012] Step 3: The rough position solution, attitude quaternion and ranging data are fused through extended Kalman filtering to determine the precise position solution of the UAV;

[0013] Step 4: Generate the desired spiral landing trajectory based on the refined position solution, the preset landing point, and the preset trajectory parameters;

[0014] Step 5: Based on the desired spiral landing trajectory and position, the control signal of the drone motor is output through layered adjustment of the inner and outer loop controls.

[0015] Furthermore, the preset structure includes:

[0016] The center platform label of the vehicle-mounted airport platform center preset height layout;

[0017] Peripheral platform labels are distributed in a regular pentagonal pattern with the central platform label as the center;

[0018] Drone tags are distributed in a non-coplanar triangular pyramid on the drone.

[0019] Further, the weighted least squares fitting process includes:

[0020] Get the position coordinates of the platform tag in the platform coordinate system;

[0021] Based on the position coordinates of the platform tag, the ranging data corresponding to the drone tag of each platform tag is subjected to minimum cost fitting processing to form the ranging geometric equation;

[0022] The ranging geometric equation is solved by least squares to obtain the rough solution of the UAV's position.

[0023] Further, multi-layer coupled posture closure analysis includes:

[0024] Eliminate outlier observations in ranging data;

[0025] Based on the distance matching and directional coupling between platform tags and the rough position solution, a weighted covariance matrix is constructed.

[0026] By performing eigendecomposition and iterative refinement on the weighted covariance matrix, and using the rough position solution as the initial condition to assist in the solution, the attitude quaternion of the UAV is obtained.

[0027] Furthermore, the Kalman filter fusion process is extended to include:

[0028] The rough position solution, attitude quaternion and ranging data are used as observation quantities, and the refined position solution of the UAV is output through the extended Kalman filter fusion algorithm.

[0029] Furthermore, based on the refined position, the preset landing point, and the preset trajectory parameters, a desired spiral landing trajectory is generated, including:

[0030] The preset trajectory parameters include: radial distance attenuation model based on altitude change and rotation angle change model;

[0031] With the preset landing point as the center, the expected spiral landing trajectory is generated by combining the position solution, radial distance attenuation model and rotation angle change model; the expected spiral landing trajectory indicates the expected coordinates of the position solution corresponding to the platform coordinate system.

[0032] Furthermore, layered adjustments are made through inner-loop control and outer-loop control, including:

[0033] The outer loop controller generates the desired posture based on the error between the refined position solution and the expected coordinates;

[0034] The inner loop controller generates a motor speed regulation signal based on the error between the attitude quaternion and the desired attitude.

[0035] In a second aspect, a vehicle-mounted modular unmanned aerial vehicle intelligent airport control system is applied to any of the vehicle-mounted modular unmanned aerial vehicle intelligent airport control methods described above, comprising:

[0036] UWB tag module, used to deploy tags on vehicle-mounted airport platforms and drones;

[0037] Coordinate initialization module, used to establish the platform coordinate system;

[0038] Landing guidance module, including:

[0039] The data acquisition unit is used to process the ranging data based on the platform tag and the drone tag through weighted least squares fitting to obtain a rough solution of the drone's position;

[0040] The data processing unit is used to combine the rough position solution and the ranging data, and calculate the attitude quaternion of the UAV through multi-layer coupled attitude closure analysis;

[0041] The data fusion unit is used to fuse the rough position solution, attitude quaternion and ranging data through the extended Kalman filter to determine the precise position solution of the UAV;

[0042] A trajectory prediction unit, configured to generate a desired spiral landing trajectory based on the refined position, a preset landing point, and preset trajectory parameters;

[0043] The motor control unit is used to output the control signal of the drone motor through layered adjustment of inner and outer loop control according to the desired spiral landing trajectory and position.

[0044] The beneficial effects of the present invention are as follows: a geometric coupling system is constructed by deploying a regular pentagonal coplanar UWB tag array on a vehicle-mounted airport platform and a triangular pyramid non-coplanar tag array on the drone end; combined with weighted least squares position rough solution, multi-layer coupled attitude closure analysis and extended Kalman filter fusion, the key problem of autonomous landing of drones on a vehicle-mounted dynamic platform is solved, high-precision positioning and unique attitude solution are achieved, and the attitude flip ambiguity and multipath interference of traditional solutions are eliminated, so that drones can land stably in complex scenarios such as vehicle movement and bumps. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This invention provides a vehicle-mounted modular UAV intelligent airport control method;

[0046] Figure 2 The invention relates to a vehicle-mounted modularized unmanned aerial vehicle (UAV) intelligent airport control system. DETAILED DESCRIPTION

[0047] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0048] like Figures 1 and 2 As shown, a vehicle-mounted modular UAV intelligent airport control method includes:

[0049] Labels are placed on the vehicle-mounted airport platform and drone in a preset structure, including platform labels and drone labels;

[0050] In one embodiment of the present invention, the preset structure includes:

[0051] The center platform label of the vehicle-mounted airport platform center preset height layout;

[0052] Peripheral platform labels are distributed in a regular pentagonal pattern with the central platform label as the center;

[0053] Drone tags are distributed in a non-coplanar triangular pyramid on the drone.

[0054] Specifically, platform tags include: 、 、 、 、 and ; Drone tags include: 、 and ; The platform label at the center of the vehicle-mounted airport platform at height z ; Distributed on the vehicle-mounted airport platform, Platform labels for the vertices of the regular pentagon built as the center 、 、 、 and ; Drone tag located on the drone 、 and , 、 and They are distributed in the form of non-coplanar triangular pyramids extending upward.

[0055] It should be noted that the platform tag Located high in the center, it can reduce obstructions (such as propeller interference) and ensure signal stability; platform label 、 、 、 and It forms a regular pentagonal coplanar array, provides an attitude reference plane, and multi-vertex redundant observation improves positioning reliability; UAV tag 、 and The non-coplanar triangular pyramid arrangement allows for uniquely calculating the drone's six degrees of freedom (DOF) using any three-point ranging, eliminating mirror image ambiguity. This layout leverages complementary geometric properties to achieve high-precision, unambiguous positioning and attitude calculation, ensuring the accuracy and stability of vehicle-mounted drone landings.

[0056] Initialize and generate a platform coordinate system with the center of the vehicle-mounted airport platform as the origin;

[0057] In response to the distance parameter between any platform tag and any drone tag being less than the preset guidance distance, the landing guidance process is triggered, including:

[0058] Step 1: Based on the ranging data of the platform tag and the drone tag, the ranging data is processed by weighted least squares fitting to obtain a rough solution of the drone's position;

[0059] In one embodiment of the present invention, the weighted least squares fitting process includes:

[0060] Get the position coordinates of the platform tag in the platform coordinate system;

[0061] Based on the position coordinates of the platform tag, the ranging data corresponding to the drone tag of each platform tag is subjected to minimum cost fitting processing to form the ranging geometric equation;

[0062] The ranging geometric equation is solved by least squares to obtain the rough solution of the UAV's position.

[0063] In one embodiment of the present invention, the UWB distance set is processed by weighted least squares fitting to obtain a rough solution of the UAV's position, including:

[0064] Determine the position of each platform label in the platform coordinate system , , , 、 and Indicates the kth platform label in the platform coordinate system The horizontal coordinate, vertical coordinate and height coordinate of Represents the transpose of a vector, which facilitates subsequent matrix operations;

[0065] Determine the UWB distance between the drone tag and the platform tag (ranging data) , ;

[0066] For each UWB distance, construct the ranging geometry equation, , , represents the rough solution of the position to be solved, 、 and Indicates the rough solution of the position to be solved with respect to the platform coordinate system axis, Axis and The position of the axis, Indicates ranging error;

[0067] The ranging geometric equation is linearized as follows: ;in, 、 and Indicates the Platform tags With the center platform label In the platform coordinate system axis, Axis and The coordinate difference of the axis, , Indicates the position of the center platform label in the platform coordinate system;

[0068] Constructing the observation matrix , where the elements ;

[0069] Constructing the observation vector , ;in, Indicates the position of the cth platform label in the platform coordinate system, represents the UWB distance between the c-th platform tag and the i-th drone tag;

[0070] Constructing a diagonal weighted matrix based on distance weighting , , , express function;

[0071] The objective function is constructed as: ;

[0072] The position rough solution is obtained by solving the objective function using the least squares method , .

[0073] It should be noted that the platform label position is determined , establish the platform label coordinate benchmark. Respectively represent the horizontal, vertical, and height coordinates of the kth platform tag in the platform coordinate system P, and clarify the tag's spatial position. Determine the UWB distance (including 15 distance parameters) to obtain the ranging observation value between the drone and the platform tag as the positioning input. Indicates platform label With drone label The measured distance is the key observation data for positioning.

[0074] Constructing geometric equations , in order to establish a positioning model based on geometric relationships and describe the relationship between measured distance and coordinate difference. is the theoretical distance between the drone and the platform tag, The geometric equation is the basic model of measurement error, and the measured value = theoretical value + error. Linearization processing converts the nonlinear geometric equation into a linear form, which is convenient for least squares solution. By squaring and sorting both sides of the geometric equation, and linearizing it using coordinate difference, the observation matrix is obtained. and vector b, observation matrix and vector b, facilitating the transformation of the geometric equations and adapting them to the least squares algorithm. A weighting matrix is constructed to weight the distance observations, giving observations with smaller errors (typically closer distances) greater weight in the solution, improving accuracy. The objective function is the standard form of weighted least squares, minimizing the weighted sum of squares of the observation residuals to obtain a rough position solution.

[0075] Step 2: Combining the rough position solution with the ranging data, the attitude quaternion of the UAV is calculated through multi-layer coupled attitude closure analysis.

[0076] In one embodiment of the present invention, multi-layer coupled posture closure analysis includes:

[0077] Eliminate outlier observations in ranging data;

[0078] Based on the distance matching and directional coupling between platform tags and the rough position solution, a weighted covariance matrix is constructed.

[0079] By performing eigendecomposition and iterative refinement on the weighted covariance matrix, and using the rough position solution as the initial condition to assist in the solution, the attitude quaternion of the UAV is obtained.

[0080] In one embodiment of the present invention, a multi-layer coupled attitude closure analysis is performed on the rough position solution of the UAV and the UWB distance set to obtain the attitude quaternion of the UAV, including:

[0081] Calculate each platform label With drone label Distance measurement between Rough analysis of the drone's position The distance residuals of , in order to remove outlier observations based on the distance residuals, are as follows:

[0082] Distance residuals: ;

[0083] Remove outlier observations: ;

[0084] in, Indicates platform label With drone label The distance residual between represents the set of valid observation indices, Indicates the standard deviation of the UWB distance in the UWB distance set;

[0085] Tectonic coupling factor ,as follows:

[0086]

[0087]

[0088] in, Indicates platform label With drone label The coupling factor, represents the nominal baseline distance, , Indicates the platform label spacing, Indicates the weight of the platform label spacing, , represents the coupling adjustment constant, Indicates platform label With drone label The axial coupling degree, Indicates The drone coordinate system constructed as the origin Download drone tag location;

[0089] Based on the coupling factor , based on the effective observation index set , construct the weighted covariance matrix without mean, as follows:

[0090]

[0091]

[0092]

[0093] in, Based on the effective observation index set and weights The calculated drone label is in the platform coordinate system The weighted average position under Represents the drone coordinate system The weighted average position of the drone labels, represents the mean-removed weighted covariance matrix, Indicates the drone label in drone coordinate system B The coordinates in the platform coordinate system P, Represents the drone coordinate system Download drone tag coordinates of

[0094] Based on the weighted covariance matrix , construct the feature matrix ,as follows:

[0095] , ;

[0096] in, represents the weighted covariance matrix traces, represents the identity matrix, the weighted covariance matrix is a 3×3 matrix, 、 、 、 、 and Both are weighted covariance matrices The elements in is a weighted covariance matrix The column vector is composed of the difference of specific elements of By extracting the weighted covariance matrix The difference information between the elements in the image is used to participate in operations such as feature decomposition, helping to solve the UAV's attitude quaternion; express The transpose of express The transpose of

[0097] For the feature matrix Perform eigenvalue decomposition to obtain the eigenvector corresponding to the maximum eigenvalue, and normalize the eigenvector to obtain the initial quaternion;

[0098] The initial quaternion is iteratively processed through a 6D attitude estimation coupled iterative refinement algorithm to obtain the attitude quaternion.

[0099] It should be noted that the distance residual is calculated and outlier observations are eliminated to avoid data interference with large errors by eliminating outliers, thereby screening valid observations and improving the reliability of attitude solution. Used to measure the deviation between the observed value and the theoretical value. The valid observation index set is used to retain the residual Observations of . Among them, is the standard deviation of UWB distance, reflecting the degree of data dispersion.

[0100] The coupling factor is constructed to quantify the contribution weight of different observations to the attitude solution, highlighting the role of effective observations. It comprehensively reflects the coupling strength between the platform label and the drone label. is the nominal baseline distance, which is used to measure the degree of match between the observed distance and the theoretical design baseline; Characterizes the axial coupling degree (i.e. directional consistency) of the two tags in their respective coordinate systems, is the adjustment constant.

[0101] Matching distance Consistency with direction Combined, the closer the distance is to the nominal value, the higher the directional coupling degree is. The larger the distance, the higher the weight of the observation in the attitude solution. This design breaks through the limitations of traditional single distance weighting and distinguishes the observation quality based on the essence of geometric relationships, thus achieving high-precision attitude solution in an innovative way.

[0102] A weighted covariance matrix without mean value is constructed to extract the statistical correlation between coordinates from valid observations and establish a mathematical model for attitude solution. and Based on the effective observation, the platform coordinate system P and the drone coordinate system B are respectively and The weighted average position of , removing the mean to highlight the coordinate deviation characteristics. This is a weighted covariance matrix with no mean, capturing the coordinated variation of label coordinate deviations between the two coordinate systems. It is the core matrix connecting the two coordinate systems and solving pose transformations. Position offsets are eliminated through weighted averaging, and then the covariance matrix is calculated to meet the statistical requirements for extracting correlations between variables. This matrix transforms discrete label coordinates into structured statistical features.

[0103] The feature matrix K is constructed and decomposed to transform the geometric features of the covariance matrix into an algebraic form and efficiently extract the posture information. ,in, is the trace of the matrix S (the sum of the diagonal elements, representing the matrix energy), It is composed of the antisymmetric elements of the matrix S, where I is the identity matrix. The construction of the characteristic matrix K is based on matrix theory, and its eigenvalue decomposition can extract the main geometric features of the covariance matrix S. The eigenvector corresponding to the largest eigenvalue contains the main transformation direction between the two coordinate systems. After normalization, this eigenvector is used as the initial quaternion. Essentially, this transforms the complex three-dimensional rotation problem into an algebraic feature extraction, improving the efficiency and accuracy of attitude solution.

[0104] By iteratively refining the initial quaternion, errors caused by data noise or simplified models are eliminated, improving attitude solution accuracy. Although the initial quaternion is based on the main eigenvectors, it may contain local optima or noise. Using the 6D attitude estimation coupled iterative refinement algorithm, the quaternion is gradually corrected to better match the true attitude. It should be noted that the 6D attitude estimation coupled iterative refinement algorithm is existing technology and will not be described in detail here.

[0105] Step 3: The rough position solution, attitude quaternion and ranging data are fused through extended Kalman filtering to determine the precise position solution of the UAV;

[0106] In one embodiment of the present invention, the extended Kalman filter fusion process includes:

[0107] The rough position solution, attitude quaternion and ranging data are used as observation quantities, and the refined position solution of the UAV is output through the extended Kalman filter fusion algorithm.

[0108] In one embodiment of the present invention, the rough position solution, attitude quaternion and UWB distance set are fused through an extended Kalman filter to determine the precise position solution of the UAV, including:

[0109] The rough position solution, attitude quaternion and UWB distance set are used as observation quantities to obtain the precise position solution of the UAV based on the extended Kalman filter fusion algorithm. .

[0110] It should be noted that three types of data are used as observation input: the rough solution of the drone's position obtained earlier, the attitude quaternion representing the drone's attitude, and the set of UWB distances between the drone and the platform's tags. These observations are processed using the extended Kalman filter fusion algorithm, which effectively addresses the uncertainty and noise in nonlinear systems. By fusing and analyzing multi-source data and optimizing estimation, the accuracy of position estimation is further improved, ultimately obtaining a more accurate drone position solution, providing more reliable position information support for precise drone landing.

[0111] It should be noted that the extended Kalman fusion filter algorithm is a commonly used technical means in the field of multi-source data fusion, so it will not be described in detail.

[0112] Step 4: Generate the desired spiral landing trajectory based on the refined position solution, the preset landing point, and the preset trajectory parameters;

[0113] In one embodiment of the present invention, generating a desired spiral landing trajectory based on the refined position solution, a preset landing point, and preset trajectory parameters includes:

[0114] The preset trajectory parameters include: radial distance attenuation model based on altitude change and rotation angle change model;

[0115] With the preset landing point as the center, the expected spiral landing trajectory is generated by combining the position solution, radial distance attenuation model and rotation angle change model; the expected spiral landing trajectory indicates the expected coordinates of the position solution corresponding to the platform coordinate system.

[0116] In one embodiment of the present invention, generating a desired spiral landing trajectory based on the precise location of the drone, a preset landing point, and preset landing trajectory parameters includes:

[0117] Location Analysis Expressed as: , 、 and Indicates the UAV is in the platform coordinate system The horizontal coordinate, vertical coordinate and height coordinate of

[0118] The preset landing point coordinates are any point on the vehicle-mounted airport platform: ;in, Indicates the horizontal coordinate of the preset landing point, Indicates the vertical coordinate of the preset landing point;

[0119] Preset landing trajectory parameters, including: initial radius , attenuation coefficient , initial angle and the initial angular velocity ;

[0120] The expected spiral landing trajectory includes: radial distance, rotation angle, spiral trajectory abscissa and spiral trajectory ordinate, as follows:

[0121] Radial distance of the expected spiral landing trajectory for: , is the natural base;

[0122] The rotation angle of the desired spiral landing trajectory for: ;

[0123] The horizontal coordinate of the spiral trajectory of the expected spiral landing trajectory for: ;

[0124] The spiral trajectory ordinate of the expected spiral landing trajectory for: .

[0125] It should be noted that the position is refined: the position is refined Expressed as This clarifies the precise horizontal, vertical, and altitude coordinates of the UAV in the platform coordinate system P, providing a real-time position reference for trajectory generation.

[0126] The default landing point coordinates are user-defined and are set to The preset landing point coordinates provide the endpoint guidance for the trajectory, ensuring that the drone can eventually reach the designated location of the vehicle-mounted airport platform.

[0127] Preset landing trajectory parameters: including initial radius , attenuation coefficient , initial angle and the initial angular velocity These parameters are the control factors of the spiral trajectory, the initial radius Determine the trajectory starting scale and attenuation coefficient Control the radial distance with the height of the shrink rate, the initial angle and the initial angular velocity Determine the starting state and speed of rotation to jointly realize the shape and dynamic control of the spiral trajectory.

[0128] Radial distance formula: With the height reduce( Reduced), radial distance The exponential decay satisfies the logic of the drone gradually approaching the landing point as it descends, ensuring that the horizontal distance shrinks synchronously during the altitude reduction process.

[0129] Rotation angle formula Initial Angle Determine the starting direction, Make the rotation angle gradually increase as the height decreases, and the rate of change is the same as 、 Matching to form a smooth spiral rotation effect, ensuring the continuity and regularity of the trajectory.

[0130] Horizontal and vertical coordinate formulas: and . To preset landing point As the center, through trigonometric functions and The radial distance Decompose it into horizontal and vertical directions to form a spiral trajectory around the landing point, strictly follow the geometric definition of the spiral line, and ensure that the drone approaches the landing point along a reasonable path.

[0131] Step 5: Based on the desired spiral landing trajectory and position, the control signal of the drone motor is output through layered adjustment of the inner and outer loop controls.

[0132] In one embodiment of the present invention, hierarchical regulation is performed through inner-loop control and outer-loop control, including:

[0133] The outer loop controller generates the desired posture based on the error between the refined position solution and the expected coordinates;

[0134] The inner loop controller generates a motor speed regulation signal based on the error between the attitude quaternion and the desired attitude.

[0135] In one embodiment of the present invention, inner loop control and outer loop control are performed respectively according to the desired spiral landing trajectory and position refinement to determine the control signal of the motor of the target UAV, including:

[0136] Get location details Corresponding predicted position solution of expected spiral landing trajectory ;

[0137] Location-based analysis and predicted position analysis Position error , to determine the outer loop control output through the outer loop controller, as follows:

[0138]

[0139] in, Indicates the output of the outer loop control, 、 and are the proportional coefficient, integral coefficient and differential coefficient of the outer loop controller respectively, represents the integral term of the outer loop controller, represents the derivative term of the outer loop controller;

[0140] Based on the UAV dynamics model, the outer loop control output Convert to desired attitude quaternion ;

[0141] Calculate attitude quaternion and expected attitude quaternion The attitude error ;

[0142] Based on attitude quaternion and expected attitude quaternion The attitude error , to determine the inner loop control output through the inner loop controller, as follows:

[0143]

[0144] in, Indicates the output of the inner loop control, 、 and are the proportional coefficient, integral coefficient and differential coefficient of the inner loop controller respectively, represents the integral term of the inner loop controller, represents the derivative term of the inner loop controller;

[0145] Based on the UAV motor model, the inner loop control output Converted into control signals for the motors of the target drone.

[0146] It should be noted that the precise predicted position solution for the desired spiral landing trajectory, corresponding to the precise solution of the drone's current position, is determined to clearly define the target position the drone should reach. Then, based on the position error between the precise solution of the drone's actual position and the predicted solution, an outer-loop controller (using proportional, integral, and differential control mechanisms) calculates the outer-loop control output to correct the position error and bring the drone closer to the desired trajectory. Next, the attitude error between the drone's current attitude quaternion and the desired attitude quaternion is calculated. The inner-loop controller (also using proportional, integral, and differential control mechanisms) then determines the inner-loop control output to adjust the drone's attitude to the desired attitude. Finally, based on the drone's motor model, the inner-loop control output is converted into actual motor control signals, driving the drone along the desired spiral landing trajectory. This ensures that both its position and attitude accurately track the set trajectory, achieving a precise landing. This entire process, through layered control of the inner and outer loops, precisely adjusts both position and attitude, ensuring stable and accurate flight.

[0147] It's important to note that the output of the outer-loop control reflects the position error adjustment requirements and, based on the dynamics model, must be converted into the attitude required to achieve this position control. For example, the required total lift and torque are calculated from the output of the outer-loop control. The desired attitude (including roll and pitch angles) that satisfies these mechanical conditions is then inferred through the dynamics equations. Ultimately, this is represented as a quaternion, providing the target for the inner-loop attitude control.

[0148] It's important to note that the inner-loop control output adjusts for attitude error, and the motor model converts it into specific motor control signals. For example, the inner-loop control output determines the required torque or speed. By understanding the relationship between speed and PWM signals, as well as the relationship between lift and speed in the motor model, a corresponding PWM control signal is generated to drive the motor to adjust speed, generating the required lift and torque, achieving precise attitude adjustment and ensuring the drone flies along the desired trajectory.

[0149] A vehicle-mounted modular unmanned aerial vehicle intelligent airport control system, applied to any of the vehicle-mounted modular unmanned aerial vehicle intelligent airport control methods described above, comprising:

[0150] UWB tag module, used to deploy tags on vehicle-mounted airport platforms and drones;

[0151] Coordinate initialization module, used to establish the platform coordinate system;

[0152] Landing guidance module, including:

[0153] The data acquisition unit is used to process the ranging data based on the platform tag and the drone tag through weighted least squares fitting to obtain a rough solution of the drone's position;

[0154] The data processing unit is used to combine the rough position solution and the ranging data, and calculate the attitude quaternion of the UAV through multi-layer coupled attitude closure analysis;

[0155] The data fusion unit is used to fuse the rough position solution, attitude quaternion and ranging data through the extended Kalman filter to determine the precise position solution of the UAV;

[0156] A trajectory prediction unit, configured to generate a desired spiral landing trajectory based on the refined position, a preset landing point, and preset trajectory parameters;

[0157] The motor control unit is used to output the control signal of the drone motor through layered adjustment of inner and outer loop control according to the desired spiral landing trajectory and position.

[0158] The above describes the embodiments of this embodiment, but this embodiment is not limited to the above specific implementation methods. The above specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make many forms based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A vehicle-mounted modular UAV intelligent airport control method, characterized in that: include: Labels are placed on the vehicle-mounted airport platform and drone in a preset structure, including platform labels and drone labels; Initialize and generate a platform coordinate system with the center of the vehicle-mounted airport platform as the origin; In response to the distance parameter between any platform tag and any drone tag being less than the preset guidance distance, the landing guidance process is triggered, including: Step 1: Based on the ranging data of the platform tag and the drone tag, the ranging data is processed by weighted least squares fitting to obtain a rough solution of the drone's position; Step 2: Combining the rough position solution with the ranging data, the attitude quaternion of the UAV is calculated through multi-layer coupled attitude closure analysis. Step 3: The rough position solution, attitude quaternion and ranging data are fused through extended Kalman filtering to determine the precise position solution of the UAV; Step 4: Generate the desired spiral landing trajectory based on the refined position solution, the preset landing point, and the preset trajectory parameters; Step 5: Based on the desired spiral landing trajectory and position, the control signal of the drone motor is output through layered adjustment of the inner and outer loop controls.

2. A vehicle-mounted modular UAV intelligent airport control method according to claim 1, characterized in that: Preset structure, including: The center platform label of the vehicle-mounted airport platform center preset height layout; Peripheral platform labels are distributed in a regular pentagonal pattern with the central platform label as the center; Drone tags are distributed in a non-coplanar triangular pyramid on the drone.

3. The vehicle-mounted modular UAV intelligent airport control method according to claim 2, characterized in that: Weighted least squares fitting processing, including: Get the position coordinates of the platform tag in the platform coordinate system; Based on the position coordinates of the platform tag, the ranging data corresponding to the drone tag of each platform tag is subjected to minimum cost fitting processing to form the ranging geometric equation; The ranging geometric equation is solved by least squares to obtain the rough solution of the UAV's position.

4. The vehicle-mounted modular UAV intelligent airport control method according to claim 3 is characterized in that: Multi-layer coupled posture closure analysis, including: Eliminate outlier observations in ranging data; Based on the distance matching and directional coupling between platform tags and the rough position solution, a weighted covariance matrix is constructed. By performing eigendecomposition and iterative refinement on the weighted covariance matrix, and using the rough position solution as the initial condition to assist in the solution, the attitude quaternion of the UAV is obtained.

5. The vehicle-mounted modular UAV intelligent airport control method according to claim 4 is characterized in that: Extended Kalman filter fusion processing, including: The rough position solution, attitude quaternion and ranging data are used as observation quantities, and the refined position solution of the UAV is output through the extended Kalman filter fusion algorithm.

6. The vehicle-mounted modular UAV intelligent airport control method according to claim 5, characterized in that: Generate the desired spiral landing trajectory based on the position solution, the preset landing point, and the preset trajectory parameters, including: The preset trajectory parameters include: radial distance attenuation model based on altitude change and rotation angle change model; With the preset landing point as the center, the expected spiral landing trajectory is generated by combining the position solution, radial distance attenuation model and rotation angle change model; the expected spiral landing trajectory indicates the expected coordinates of the position solution corresponding to the platform coordinate system.

7. The vehicle-mounted modular UAV intelligent airport control method according to claim 6, characterized in that: Through the inner and outer loop control layered regulation, including: The outer loop controller generates the desired posture based on the error between the refined position solution and the expected coordinates; The inner loop controller generates a motor speed regulation signal based on the error between the attitude quaternion and the desired attitude.

8. A vehicle-mounted modular UAV intelligent airport control system, applied to a vehicle-mounted modular UAV intelligent airport control method according to any one of claims 1 to 7, characterized in that: include: UWB tag module, used to deploy tags on vehicle-mounted airport platforms and drones; Coordinate initialization module, used to establish the platform coordinate system; Landing guidance module, including: The data acquisition unit is used to process the ranging data based on the platform tag and the drone tag through weighted least squares fitting to obtain a rough solution of the drone's position; The data processing unit is used to combine the rough position solution and the ranging data, and calculate the attitude quaternion of the UAV through multi-layer coupled attitude closure analysis; The data fusion unit is used to fuse the rough position solution, attitude quaternion and ranging data through the extended Kalman filter to determine the precise position solution of the UAV; A trajectory prediction unit, configured to generate a desired spiral landing trajectory based on the refined position, a preset landing point, and preset trajectory parameters; The motor control unit is used to output the control signal of the drone motor through layered adjustment of inner and outer loop control according to the desired spiral landing trajectory and position.

Images