Vehicle-mounted unmanned aerial vehicle recovery control method, device and equipment in mobile environment

Abstract

The present application relates to a mobile environment vehicle-mounted unmanned aerial vehicle recovery control method, device and equipment. The method comprises the following steps: acquiring the relative positioning information between the unmanned aerial vehicle and the vehicle-mounted recovery platform in real time, and sensing the aerodynamic flow field information around the vehicle-mounted recovery platform in real time; determining the disturbance influence vector of the aerodynamic disturbance on the unmanned aerial vehicle according to the aerodynamic flow field information, and generating an active flow field compensation control quantity; judging whether the relative positioning information of the unmanned aerial vehicle meets the preset recovery condition; if the preset recovery condition is met, generating a cooperative control instruction according to the relative positioning information and the active flow field compensation control quantity in combination with the vehicle motion state, dynamically regulating and controlling the flight parameters of the unmanned aerial vehicle, so as to guide the unmanned aerial vehicle to recover with the vehicle-mounted recovery platform. The method realizes real-time sensing and active compensation of the vehicle-mounted aerodynamic flow field, and comprehensively considers the vehicle motion state and the relative positioning information, so that the unmanned aerial vehicle can stably approach and accurately align the recovery platform, and the recovery success rate is improved.

Description

Technical Field

[0001] This invention relates to the field of drone recovery technology, and in particular to a method, apparatus and equipment for controlling the recovery of vehicle-mounted drones in a mobile environment. Background Technology

[0002] With the rapid development of drone technology, drones are increasingly being used in logistics delivery, reconnaissance patrol, and emergency rescue. Among these applications, autonomous take-off, landing, and recovery technology for drones on mobile platforms (such as vehicles and ships) has become a key element in expanding the operational range of drones and improving mission continuity. However, in high-speed vehicle-mounted environments, complex unsteady airflow disturbances occur around the vehicle. Existing recovery technologies mostly employ passive control methods, failing to provide real-time perception and active compensation for these disturbances. This can lead to safety hazards such as attitude instability, docking failure, or even collisions with the vehicle when the drone approaches the recovery device. Summary of the Invention

[0003] This invention provides a method, apparatus, and equipment for controlling the recovery of vehicle-mounted unmanned aerial vehicles in a mobile environment, in order to overcome at least one of the above-mentioned technical problems existing in the prior art.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions: In a first aspect, the present invention provides a method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment, comprising: The relative positioning information between the drone and the vehicle-mounted recycling platform is acquired in real time, and the aerodynamic flow field information around the vehicle-mounted recycling platform is perceived in real time. Based on the aerodynamic flow field information, the disturbance influence vector of aerodynamic disturbance on the UAV is determined, and an active flow field compensation control quantity is generated; Determine whether the relative positioning information of the drone meets the preset recovery conditions; If the relative positioning information of the UAV meets the preset recovery conditions, a coordinated control command is generated based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, to dynamically adjust the flight parameters of the UAV, so as to guide the UAV to be recovered by the vehicle-mounted recovery platform.

[0005] In one possible implementation of the first aspect, the relative positioning information includes a relative pose error vector, and the real-time acquisition of the relative positioning information between the UAV and the vehicle-mounted recovery platform includes: The image of the cooperation mark on the vehicle-mounted recovery platform is acquired by an airborne vision sensor, and the relative position vector and relative yaw angle of the UAV relative to the vehicle-mounted recovery platform are calculated by the PnP algorithm. The three-axis angular velocity and three-axis acceleration of the UAV are obtained by the airborne inertial measurement unit, and the standard deviation of the angular velocity is calculated within a preset sliding time window to obtain the attitude jitter. The absolute distance between the drone and the recycling inlet is obtained by a distance sensor installed at the recycling inlet of the vehicle-mounted recycling platform. The noise covariance matrix of visual observations in the extended Kalman filter is adaptively adjusted according to the attitude jitter amount, and the relative position vector, the relative yaw angle and the absolute distance are used as observation values ​​to be input into the extended Kalman filter for fusion to obtain the relative pose error vector.

[0006] In one possible implementation of the first aspect, determining the disturbance influence vector of the aerodynamic disturbance on the UAV based on the aerodynamic flow field information includes: The vehicle-mounted recovery platform's driving speed and acceleration are acquired in real time, the drone's flight speed vector and the drone's three-dimensional position coordinates relative to the vehicle-mounted recovery platform are acquired in real time, and the local airflow rate, airflow direction and pressure change data in the aerodynamic flow field information are acquired in real time to construct an input feature vector. The input feature vector is input into a pre-trained unsteady aerodynamic flow field model, and the unsteady aerodynamic flow field model is used to predict the disturbance acceleration vector and disturbance torque vector of the aerodynamic disturbance around the vehicle-mounted recovery platform on the UAV during future sampling periods.

[0007] In one possible implementation of the first aspect, determining whether the relative positioning information of the UAV meets the preset recovery conditions includes: Determine whether the relative distance and relative speed errors in the relative positioning information meet the first determination condition; If the relative distance and relative velocity errors in the relative positioning information meet the first determination condition, then based on the relative distance and relative velocity errors, combined with the relative pose error in the relative positioning information and the disturbance influence vector, the weighted fusion algorithm is used to calculate the recovery alignment confidence. Determine whether the recovery alignment confidence level exceeds a preset recovery alignment confidence threshold; If the recovery alignment confidence level exceeds a preset recovery alignment confidence level threshold, it indicates that the UAV meets the preset recovery alignment conditions.

[0008] In one possible implementation of the first aspect, generating a cooperative control command based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the onboard recycling platform, includes: Extract the vehicle's yaw angle and steering angular velocity from the vehicle's motion state; Using the vehicle yaw angle and the steering angular velocity as coordinate transformation parameters, coordinate system alignment is performed on the active flow field compensation control quantity to obtain the aligned compensation control quantity. The aligned compensation control quantity is used as the feedforward control quantity and superimposed and fused with the feedback control quantity generated based on the relative positioning information; Extract the vehicle speed from the vehicle's motion state and the relative distance from the relative positioning information. Based on the vehicle speed and the relative distance, determine the fusion weight coefficient in real time. The fusion weight coefficient is positively correlated with the vehicle speed and negatively correlated with the relative distance. The cooperative control command is generated based on the feedforward control quantity, the feedback control quantity, and the fusion weight coefficient.

[0009] In one possible implementation of the first aspect, determining the fusion weight coefficients in real time based on the vehicle speed and the relative distance includes: When the vehicle speed is less than or equal to a first preset speed threshold and the relative distance is greater than or equal to a first preset distance threshold, the fusion weight coefficient is set to a first weight value. When the vehicle speed is greater than or equal to a second preset speed threshold and the relative distance is less than or equal to a second preset distance threshold, the fusion weight coefficient is set to a second weight value, and the second weight value is greater than the first weight value. When the vehicle speed is between the first preset speed threshold and the second preset speed threshold, and the relative distance is between the first preset distance threshold and the second preset distance threshold, the fusion weight coefficient increases non-linearly with the vehicle speed and the relative distance.

[0010] In one possible implementation of the first aspect, the preset recovery condition is a phased triggering condition, and the generation of cooperative control commands to dynamically adjust the flight parameters of the UAV includes: When the recovery alignment confidence is greater than the first preset alignment confidence threshold, and the relative distance between the UAV and the vehicle-mounted recovery platform is greater than the third preset distance threshold or the relative speed error between the UAV and the vehicle-mounted recovery platform is greater than the first preset relative speed error threshold, the cooperative control command is generated based on the feedback control quantity. When the recovery alignment confidence is greater than or equal to the second preset alignment confidence threshold, the relative distance is less than or equal to the third preset distance threshold and greater than the fourth preset distance threshold, and the relative speed error is less than or equal to the second preset relative speed error threshold, the cooperative control command is generated according to the feedforward control quantity, the feedback control quantity and the fusion weight coefficient. When the recovery alignment confidence is greater than or equal to the third preset alignment confidence threshold, the relative distance is less than or equal to the fourth preset distance threshold, and the relative speed error is less than or equal to the third preset relative speed error threshold, the fusion weight coefficient is set to 1, and the differential control gain is increased to enable the UAV to dock smoothly with the vehicle-mounted recovery platform.

[0011] In one possible implementation of the first aspect, the dynamic adjustment of the flight parameters of the UAV further includes: Real-time monitoring of the relative positioning information and the disturbance influence vector; Determine whether the relative positioning information or the disturbance influence vector exceeds a preset safety threshold; If the relative positioning information or the disturbance influence vector exceeds a preset safety threshold, the recovery process is interrupted and the drone is controlled to enter a hovering or detached state.

[0012] Compared with the prior art, the present invention has at least the following beneficial effects: The present invention provides a vehicle-mounted UAV recovery control method in a mobile environment. By real-time sensing and active compensation of the vehicle's aerodynamic flow field, and by comprehensively considering the vehicle's motion state and relative positioning information, the UAV can stably approach and accurately align with the recovery platform in complex aerodynamic environments, thereby improving the recovery success rate in high-speed mobile environments.

[0013] Secondly, the present invention provides a vehicle-mounted drone recovery control device for mobile environments, comprising: The data acquisition module is used to acquire the relative positioning information between the drone and the vehicle-mounted recycling platform in real time, and to perceive the aerodynamic flow field information around the vehicle-mounted recycling platform in real time. The active flow field compensation module is used to determine the disturbance influence vector of aerodynamic disturbance on the UAV based on the aerodynamic flow field information, and generate active flow field compensation control quantity; The recovery condition determination module is used to determine whether the relative positioning information of the UAV meets the preset recovery conditions; The collaborative control module is used to generate collaborative control commands based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, when the relative positioning information of the UAV meets the preset recovery conditions, so as to dynamically adjust the flight parameters of the UAV and guide the UAV to be recovered by the vehicle-mounted recovery platform.

[0014] Thirdly, the present invention provides an electronic device comprising: at least one processor and at least one memory, wherein the memory stores computer-readable instructions; the computer-readable instructions are executed by one or more of the processors, causing the electronic device to implement the vehicle-mounted drone recovery control method in a mobile environment as described in any implementation of the first aspect.

[0015] Fourthly, the present invention provides a storage medium having a computer-executable program stored thereon, the computer-executable program being used to cause a computer to execute a vehicle-mounted unmanned aerial vehicle recovery control method in a mobile environment as described in any implementation of the first aspect.

[0016] Understandably, the beneficial effects achieved by the apparatus of the second aspect, the electronic device of the third aspect, and the storage medium of the fourth aspect provided above can be referred to the beneficial effects of the first aspect and any of its possible design embodiments, which will not be repeated here. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention; Figure 2 A flowchart illustrating a method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment, as provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of a vehicle-mounted drone recovery control device in a mobile environment, provided as an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be described below with reference to the accompanying drawings. In the description of the present invention, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. The "or" in the present invention is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A or B can represent: A alone, A and B simultaneously, and B alone. A and B can be singular or plural. Furthermore, in the description of the present invention, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items.

[0020] Furthermore, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0021] In this embodiment of the invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in this embodiment of the invention should not be construed as superior or more advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0022] With the rapid development of drone technology, drones are increasingly being used in logistics delivery, reconnaissance patrols, and emergency rescue. Among these applications, the autonomous take-off, landing, and recovery technology of drones on mobile platforms (such as vehicles and ships) has become a key element in expanding the operational range of drones and improving mission continuity. However, achieving precise recovery of drones in high-speed vehicle-mounted environments faces numerous technical challenges, such as: (1) Poor environmental adaptability: When a vehicle travels at high speed, complex unsteady airflow disturbances will be generated around the vehicle body, including wake, eddies and turbulence. Existing recovery technologies mostly adopt passive control methods, which cannot perceive and actively compensate for flow field disturbances in real time, resulting in the UAV losing attitude, failing to dock, or even crashing when approaching the recovery device.

[0023] (2) Insufficient recovery accuracy under high-speed relative motion: When there is a large relative speed between the UAV and the vehicle, the traditional control method based on position or visual error has a response delay, making it difficult to achieve accurate alignment and speed matching. Multiple recovery failures will seriously affect the system's operational efficiency.

[0024] (3) Weak adaptability to complex working conditions: Under conditions such as crosswinds, gusts, rapid vehicle acceleration or sharp turns, the existing system lacks comprehensive consideration of the vehicle's motion state and aerodynamic influence, resulting in a significant decrease in control performance.

[0025] (4) Safety and reliability need to be improved: Due to the lack of prediction and active compensation mechanism for flow field disturbances, the system cannot adjust in time when the airflow changes suddenly, which poses a major safety hazard of collision between the UAV and the vehicle.

[0026] In view of this, on the one hand, embodiments of the present invention provide a vehicle-mounted unmanned aerial vehicle (UAV) recovery control method in a mobile environment, comprising: acquiring relative positioning information between the UAV and the vehicle-mounted recovery platform in real time, and sensing aerodynamic flow field information around the vehicle-mounted recovery platform in real time; determining the disturbance influence vector of aerodynamic disturbance on the UAV based on the aerodynamic flow field information, and generating an active flow field compensation control quantity; determining whether the relative positioning information of the UAV meets preset recovery conditions; if the relative positioning information of the UAV meets the preset recovery conditions, generating a cooperative control command based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, and dynamically adjusting the flight parameters of the UAV to guide the UAV to be recovered from the vehicle-mounted recovery platform.

[0027] The present invention provides a vehicle-mounted UAV recovery control method in a mobile environment. By real-time sensing and active compensation of the vehicle's aerodynamic flow field, and by comprehensively considering the vehicle's motion state and relative positioning information, the UAV can stably approach and accurately align with the recovery platform in a complex aerodynamic environment, thereby improving the recovery success rate in high-speed mobile environments.

[0028] In some embodiments, the vehicle-mounted drone recovery control method in a mobile environment provided by the present invention can be executed by any electronic device 20 with data processing capabilities, such as a general-purpose computer, personal computer, laptop computer, switch, or tablet computer, etc. The specific implementation of the electronic device 20 is not limited here.

[0029] Figure 1 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention is shown. The electronic device 20 includes a processor 210, a memory 220, and a communication interface 230.

[0030] Processor 210 may include one or more processing cores. Processor 210 connects to various parts within electronic device 20 using various interfaces and lines, and performs various functions and processes data of electronic device 20 by running or executing instructions, programs, code sets, or instruction sets stored in memory 220, and by calling data stored in memory 220. Optionally, processor 210 may be implemented using at least one of the following hardware forms: Central Processing Unit (CPU), Graphics Processing Unit (GPU), Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA).

[0031] The memory 220 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 220 may include a non-transitory computer-readable storage medium. The memory 220 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 220 may include a program storage area. This program storage area may store instructions for implementing an operating system, instructions for implementing at least one function, instructions for implementing the various method embodiments described above, etc.

[0032] Communication interface 230 is used to communicate with other devices, equipment or communication networks, such as data storage devices, image processing devices or Ethernet, wireless access network (RAN), wireless local area network (WLAN), etc.

[0033] In terms of physical implementation, the aforementioned devices (such as processor 210, memory 220, and communication interface 230) can each be devices within the same device (such as a laptop computer). Alternatively, at least two of these devices can be located within the same device, i.e., as different devices within the same device, similar to the deployment of devices or components in a distributed system.

[0034] It is understood that the structure illustrated in this embodiment does not constitute a specific limitation on the electronic device 20. In other embodiments of the present invention, the electronic device 20 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0035] The following description, in conjunction with the accompanying drawings, illustrates a method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle (UAV) in a mobile environment, as provided by an embodiment of the present invention.

[0036] like Figure 2 As shown, this embodiment of the invention provides a method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment, which may include, but is not limited to: S1: Real-time acquisition of relative positioning information between the drone and the vehicle-mounted recycling platform, and real-time perception of aerodynamic flow field information around the vehicle-mounted recycling platform.

[0037] In one feasible implementation, the relative positioning information in this embodiment of the invention includes a relative pose error vector, and the real-time acquisition of the relative positioning information between the UAV and the vehicle-mounted recovery platform may include, but is not limited to: The image of the cooperation mark on the vehicle-mounted recovery platform is acquired by an airborne vision sensor, and the relative position vector and relative yaw angle of the UAV relative to the vehicle-mounted recovery platform are calculated by the PnP algorithm. The three-axis angular velocity and three-axis acceleration of the UAV are obtained by the airborne inertial measurement unit, and the standard deviation of the angular velocity is calculated within a preset sliding time window to obtain the attitude jitter. The absolute distance between the drone and the recycling inlet is obtained by a distance sensor installed at the recycling inlet of the vehicle-mounted recycling platform. The noise covariance matrix of visual observations in the extended Kalman filter is adaptively adjusted according to the attitude jitter amount, and the relative position vector, the relative yaw angle and the absolute distance are used as observation values ​​to be input into the extended Kalman filter for fusion to obtain the relative pose error vector.

[0038] In specific implementation, this embodiment of the invention adopts the direct relative measurement method (i.e., without relying on global positioning systems such as GPS, directly measuring the relative pose relationship between the UAV and the recovery platform through airborne sensors). It estimates the relative pose error vector in real time through multi-source information fusion from visual sensors, inertial measurement units (IMUs), and distance sensors. The specific steps are as follows: (i) Relative pose observation based on visual sensors In this embodiment of the invention, a visual cooperation marker is set at a predetermined location (e.g., the geometric center of the platform) on the vehicle-mounted recycling platform. The cooperation marker has known geometric dimensions and spatial layout (e.g., Apriltag, QR code, or infrared LED array). A downward-looking or forward-looking camera mounted on the drone acquires images of the cooperation marker in real time. The six-degree-of-freedom pose of the drone relative to the cooperation marker is calculated using the Perspective-n-Point (PnP) algorithm, yielding: Relative position observations: [x v y v , z v ] T , where x v y is the longitudinal (front-to-back) offset of the UAV's center of mass relative to the center of the vehicle-mounted recovery platform; v z is the lateral offset of the center of mass of the vehicle-mounted UAV relative to the center of the vehicle-mounted recovery platform. v The height difference between the center of mass of the UAV and the center of the recovery platform; Relative attitude observation: yaw angle ψ v This indicates the yaw angle deviation between the direction the drone's nose is pointing and the direction the vehicle is traveling.

[0039] In this embodiment of the invention, the visual observation values ​​obtained through a visual sensor provide an absolute reference at the time of image acquisition, and the sampling frequency can be 30~60Hz. The noise covariance matrix R of the visual observation... vision Based on the sharpness of the marker and the dynamic estimation of lighting conditions (e.g., through the variance calculation of reprojection error), no restrictions are imposed here.

[0040] (ii) IMU-based state prediction The inertial measurement unit (IMU) onboard the UAV outputs the three-axis angular velocity ω in real time. u =[ω x ω y ω z ] T With triaxial acceleration a u =[a x a y a z ] T The sampling frequency can be 200~1000Hz. IMU data is used for the prediction step of the extended Kalman filter (EKF), that is, between two visual observations, the relative pose error vector is updated over time based on the UAV kinematic model (such as a uniformly accelerated kinematic model).

[0041] Specifically, define the state vector: ; Where, xe This refers to the longitudinal (front-to-back) positional deviation; y e This refers to the lateral (left-right) positional deviation; z e This refers to the vertical (height) positional deviation. The longitudinal relative velocity (the velocity component of the drone relative to the vehicle-mounted recovery platform in the direction of travel). The lateral relative velocity (the velocity component of the drone relative to the vehicle-mounted recovery platform in the lateral direction). The vertical relative velocity (the velocity component of the UAV relative to the recovery platform in the height direction, positive for upward movement); φ e For relative roll angle error; θ e For relative pitch angle error; ψ e For relative yaw angle error; b ω b α Zero bias for IMU.

[0042] The position and attitude components in the aforementioned state vector constitute the relative pose error vector E=[x e y e , z e φ e θ e , ψ e ] T .

[0043] By integrating angular velocity and acceleration, the state and its covariance are recursively predicted. IMU prediction ensures high-frequency continuity of the state output (up to 200Hz or higher), but relying solely on the IMU can lead to integral drift, thus requiring periodic correction by visual and distance observations.

[0044] (III) Absolute distance observation based on distance sensor A laser rangefinder or ultrasonic sensor is installed at the recovery inlet of the vehicle-mounted recovery platform to directly detect the absolute distance L between the drone and the recovery inlet. raw This distance sensor is unaffected by ambient light and achieves centimeter-level accuracy at close range (0.2~10 m).

[0045] The absolute distance is used as another observation value of EKF, and the observation equation is: ; Among them, v L To measure noise, its variance is set based on the sensor calibration value. This observation is used to correct for visual distortion in the depth direction (z). e The estimation error of ) provides a stable constraint, especially when the vision is occluded or the texture is insufficient.

[0046] The observation matrix (Jacobi matrix) corresponding to this observation equation is: H L =[x e / L,y e / L, z e [L, 0, ..., 0] (a 1×12 matrix, with columns 4 through 12 all being 0).

[0047] in, In the EKF update step, the H L The matrix is ​​used to calculate the Kalman gain and assign the range observation residuals to x in the state vector. e y e z e The spatial position constraints are transmitted across the three components.

[0048] (iv) Calculation and adaptive adjustment of attitude jitter Although IMU data is used for state prediction, the fluctuation of its output angular velocity can reflect the flight stability of the UAV. This embodiment of the invention calculates the standard deviation of the angular velocity within a sliding time window (e.g., 0.5 seconds) to obtain the attitude jitter. ; The attitude jitter amount E imu Instead of being directly input into the EKF as the observation value, it is used to adaptively adjust the noise covariance matrix R of the visual observations. vision : when E imu When it is large (e.g., greater than the threshold E) th E th The value is set to 5° / s (this value is pre-calibrated through ground tests or simulations, and can be adjusted within the range of 3~10° / s in practical applications depending on the UAV model). This indicates that the UAV is severely shaking due to airflow disturbances, at which point the reliability of the visual solution decreases, and the system automatically increases R. vision Reduce the weight of visual observations in EKF updates; when E imu A smaller value indicates that the UAV's attitude is stable, and the system reduces R. vision It trusts visual observation. This mechanism makes the fusion algorithm robust to environmental changes.

[0049] (v) Extended Kalman filter fusion This embodiment uses an extended Kalman filter (EKF) to achieve the fusion of the above-mentioned multi-source information. The specific process is as follows: Initialization: Set the initial state vector And the covariance matrix P0.

[0050] Prediction Step: At each IMU sampling time (high frequency), the state is integrally recursively derived using acceleration and angular velocity, while simultaneously updating the covariance. Specifically, the prediction step employs the following discretized recursive form: ; .

[0051] in, The kinematic state transition function of the UAV is (which can be approximated by a uniform kinematic model or a uniformly accelerated kinematic model, and the specific model selection is determined in advance according to the dynamic characteristics of the UAV), and u(k-1) is the control input (including acceleration and angular velocity) at the previous moment. Let be the state transition matrix (Jacobi matrix), and Q be the process noise covariance matrix.

[0052] Update step: When a new visual observation or distance observation arrives, perform an EKF update. The observation vector for EKF is Z=[x v y v , z v , ψ v L raw ] T The observation equation is expressed as: Z = h(X) + v.

[0053] Where h(X) is the prediction function for the above observations based on the current state vector X, v is the observation noise vector, and the covariance matrix R of the observation noise vector v is a 5×5 diagonal matrix, the first four diagonal elements of which are the noise variances of each component of the visual observation, and the fifth diagonal element is the measurement noise variance σ of the distance sensor. L 2 .

[0054] The complete observation matrix (Jacobi matrix) is H = [H vision T H L T ] T H vision H is the 4×12 Jacobian matrix corresponding to the visual observations (position and yaw angle). L Let K be the 1×12 Jacobian matrix corresponding to the range observation. In the update step, the Kalman gain K = P·H T ·(H·P·H T + R) (-1) The gain is used to correct the state estimate and its covariance.

[0055] Adaptive adjustment: based on the current attitude jitter amount E imu Dynamically adjust the noise covariance matrix R of visual observations vision (e.g. R) vision =R vision0 ×(1+k×E imu ), where k is an empirical coefficient, ranging from 0.5 to 2.0; R vision0This is the observation noise covariance matrix pre-calibrated by the vision sensor under standard operating conditions (no significant airflow disturbance, good lighting).

[0056] After the above EKF fusion, the system outputs the optimal estimate at each time step. The relative pose error vector E is extracted from it. This estimate has the advantages of high frequency (guaranteed by IMU prediction), no drift (corrected by visual / distance observation), and anti-interference (adaptive adjustment), meeting the high real-time and high-precision requirements of high-speed vehicle-mounted recovery.

[0057] (vi) Calculation of relative distance Based on the above fusion estimation results, the relative distance L between the UAV and the vehicle-mounted recovery platform is further calculated: ; This relative distance is different from the absolute distance L directly measured by the distance sensor. raw While physically representing the same spatial distance, the merged L exhibits higher stability and reliability.

[0058] In the specific implementation process, in order to quantitatively describe the alignment state between the UAV and the vehicle-mounted recovery platform, and to provide a unified mathematical benchmark for subsequent control algorithms, this embodiment of the invention further provides a theoretical definition of the relative pose error vector, as follows: Define the vehicle coordinate system {B}: the origin is located at the center of the vehicle-mounted recycling platform, the x-axis points in the direction of vehicle travel, the y-axis points to the left side of the vehicle (according to the right-hand rule), and the z-axis is vertically upward.

[0059] Define the UAV coordinate system {U}: the origin is located at the center of mass of the UAV, the x-axis points towards the nose, the y-axis points to the right side of the wing, and the z-axis is perpendicular to the fuselage and pointing upwards.

[0060] At the same time, a world coordinate system {W} (such as the northeast-sky coordinate system) is defined as the reference inertial frame.

[0061] Then, the relative pose error vector is defined as: E=[x e y e , z e φ e θ e , ψ e ] T .

[0062] Wherein, the position error vector P e =[x e y e , z e ] T =R B W (P)u -P b R B W P is the rotation transformation matrix from the world coordinate system to the vehicle coordinate system; u P is the position vector of the UAV in the world coordinate system. b x is the position vector of the vehicle-mounted recycling platform in the world coordinate system; e This refers to the longitudinal (front-to-back) positional deviation; y e This refers to the lateral (left-right) positional deviation; z e This refers to the vertical (height) positional deviation.

[0063] Attitude error vector Φ e =[φ e θ e , ψ e ] T =Φ u -Φ b Φ u Φ is the attitude angle vector of the UAV; b φ is the attitude angle vector of the vehicle-mounted recycling platform. e The relative roll angle error characterizes the difference in lateral tilt between the UAV and the vehicle-mounted recovery platform; θ e The relative pitch angle error characterizes the difference in the degree of pitching up / down between the UAV and the vehicle-mounted recovery platform; ψ e The relative yaw angle error characterizes the alignment deviation between the UAV's nose and the vehicle's direction of travel.

[0064] In a high-speed recovery environment, this embodiment of the invention introduces a predictive derivative to address unsteady flow field disturbances, and its state-space expression is as follows: ; in, Let ΔV be the rate of change of relative pose error, Δω be the relative velocity vector between the UAV and the vehicle-mounted recovery platform, and D be the relative angular velocity vector. flow The disturbance effect vector, Let be a function determined by the relative kinematic relationship between the UAV and the vehicle-mounted recovery platform, used to describe the rate of change of the current pose error E, relative velocity ΔV, and relative angular velocity Δω with respect to the error. The physical effects.

[0065] It should be noted that the definition of relative pose error based on world coordinate difference described above is physically equivalent to the relative pose error vector obtained by the aforementioned direct relative measurement method under ideal measurement conditions, but it is not the only implementation of this invention. In practical engineering, the relative pose error vector E described in this embodiment of the invention is preferably obtained in real time using the aforementioned direct relative measurement method. In open areas with good GPS signals, the formula based on world coordinate difference can also be used as an auxiliary or backup method.

[0066] S2: Based on the aerodynamic flow field information, determine the disturbance influence vector of aerodynamic disturbance on the UAV, and generate an active flow field compensation control quantity.

[0067] In one feasible implementation, the determination of the aerodynamic disturbance influence vector on the UAV based on the aerodynamic flow field information in this embodiment of the invention may include, but is not limited to: The speed and acceleration of the vehicle-mounted recovery platform are acquired in real time, the flight speed vector of the UAV and the three-dimensional position coordinates of the UAV relative to the vehicle-mounted recovery platform are acquired in real time, and the local airflow rate, airflow direction and pressure change data in the aerodynamic flow field information are acquired in real time to construct an input feature vector; The input feature vector is input into a pre-trained unsteady aerodynamic flow field model, and the unsteady aerodynamic flow field model is used to predict the disturbance acceleration vector and disturbance torque vector of the aerodynamic disturbance around the vehicle-mounted recovery platform on the UAV during future sampling periods.

[0068] In specific implementation, embodiments of the present invention arrange pneumatic sensors around the vehicle-mounted recycling platform, which may include, but are not limited to: Wind speed sensor: detects the relative velocity v of local airflow. wind ; Wind direction sensor: acquires the angle d of airflow relative to the vehicle's centerline. wind ; Pressure sensor array: Arranged at different positions on the surface of the recovery platform to collect pressure change values ​​ΔP i (i=1, ..., n).

[0069] Meanwhile, embodiments of the present invention can, but are not limited to, obtain the real-time driving speed V of the on-board recycling platform via the vehicle's CAN bus. car and acceleration a car The flight speed vector V of the UAV is obtained through the UAV communication link. u And the three-dimensional position coordinates P of the drone relative to the center of the vehicle-mounted recovery platform. rel (That is, the relative position vector obtained through fusion).

[0070] At each sampling time, the above data is combined into an input feature vector: X=[V car a car v wind d wind ΔP i P rel V u ] T .

[0071] Then, the input feature vector is input into a pre-trained unsteady aerodynamic flow field model to predict the disturbance acceleration vector and disturbance torque vector of the aerodynamic disturbance around the vehicle-mounted recovery platform on the UAV within future sampling periods.

[0072] In specific implementation, the unsteady aerodynamic flow field model in this embodiment of the invention may, but is not limited to, employ a lightweight one-dimensional convolutional neural network (1D-CNN) model. The lightweight 1D-CNN model may, but is not limited to, include: Input layer: Receives feature sequences from T historical time steps (e.g., T=10), with tensor dimension T×(7+n). First convolutional layer: kernel size 3, number of filters 64, ReLU activation; Max pooling layer: pooling window size 2; Second convolutional layer: kernel size 3, number of filters 128, ReLU activation; Global average pooling layer; Fully connected layer: 32 neurons, ReLU activation, Dropout=0.2; Output layer: Fully connected layer, 6 linear neurons, output vector Y=[a dist , τ dist ] T Among them, a dist For the predicted perturbation acceleration vector (three-dimensional, including a) dist_x a dist_y a dist_z ), τ dist The predicted disturbance moment vector (three-dimensional, including roll moment, pitch moment, and yaw moment).

[0073] The model is trained using a large amount of pre-collected real-vehicle data or CFD simulation data, with mean squared error (MSE) as the loss function. During the training phase, the input feature vectors of the training samples are collected through real-vehicle road tests at different speeds (10–120 km / h), different crosswind conditions (0–15 m / s), and different acceleration / deceleration conditions, and expanded using CFD simulation data. The output labels (i.e., the true values ​​of the disturbance acceleration vector and the disturbance torque vector) are obtained through joint calibration using a high-precision inertial navigation system and wind tunnel test data. A time-sliding window strategy is employed during training to predict the disturbance value for the next sampling period using the feature sequence of T historical time steps.

[0074] In specific implementation, after obtaining the predicted disturbance acceleration vector and disturbance torque vector, the embodiments of the present invention generate active flow field compensation control quantities based on the predicted disturbance acceleration vector and disturbance torque vector.

[0075] In specific implementation, the active flow field compensation control quantity in this embodiment of the invention may include, but is not limited to, thrust compensation correction quantity and attitude angle compensation correction quantity. The thrust compensation correction quantity is calculated using the following formula: ; Where, Δ F This indicates the thrust compensation correction amount. m This indicates the overall mass of the drone. Indicates the vertical disturbance acceleration. ρ Indicates air density, C d Indicates the drag coefficient. A Indicates the effective windward area of ​​the drone. This indicates the speed of the drone relative to the local airflow. The attitude angle compensation correction amount is calculated using the following formula: ; in, This indicates the attitude angle compensation correction amount. Denotes the disturbance torque vector, Δ t This indicates the control sampling period. I This represents the moment of inertia of the drone.

[0076] The thrust compensation correction amount Δ F and attitude angle compensation correction amount By combining these parameters, we obtain the active flow field compensation control parameters: U comp =[ΔF, ΔΦ] T .

[0077] In specific implementation, embodiments of the present invention can also utilize the inverse dynamics model G. -1 To perform a more precise conversion, i.e., U comp =G -1 ([a dist , τ dist ] T (to adapt to the dynamic characteristics of different drones).

[0078] S3: Determine whether the relative positioning information of the UAV meets the preset recovery conditions.

[0079] In one feasible implementation, the determination of whether the relative positioning information of the UAV meets the preset recovery conditions in this embodiment of the invention may include, but is not limited to: Determine whether the relative distance and relative speed errors in the relative positioning information meet the first determination condition; If the relative distance and relative velocity errors in the relative positioning information meet the first determination condition, then based on the relative distance and relative velocity errors, combined with the relative pose error in the relative positioning information and the disturbance influence vector, the weighted fusion algorithm is used to calculate the recovery alignment confidence. Determine whether the recovery alignment confidence level exceeds a preset recovery alignment confidence threshold; If the recovery alignment confidence level exceeds a preset recovery alignment confidence level threshold, it indicates that the UAV meets the preset recovery alignment conditions.

[0080] In practical implementation, to avoid false triggering caused by a single threshold determination, this embodiment of the invention introduces a multi-source information fusion-based recovery alignment confidence determination mechanism. The recovery alignment confidence determination mechanism may include, but is not limited to, the following: Basic condition judgment: First, determine the relative distance L in the relative positioning information (i.e., , where x e y e , z e Whether the fusion result and the relative velocity error ΔV (i.e., the resultant velocity difference between the UAV and the vehicle platform) meet the first judgment condition. For example: preset relative distance threshold L th1 =50m, relative velocity error threshold ΔV th1 =5m / s. If L <L th1 And ΔV < ΔV th1 If the confidence level is positive, proceed to the next step of confidence calculation; otherwise, continue guiding.

[0081] Multi-dimensional confidence fusion: To more accurately assess whether recycling is currently suitable, this embodiment of the invention constructs three sub-functions: Space vector constraint function Where ΔP is the relative distance residual (Euclidean distance), and ΔV is the relative velocity residual. The output value of this function is between 0 and 1; the smaller the distance and velocity errors, the larger the output value.

[0082] Pose Consistency Evaluation Function :in The relative pose error is calculated using a visual sensor. This represents the attitude jitter based on IMU feedback. This function outputs the confidence level of the pose alignment, with an output value between 0 and 1. The smaller the visual pose error and the more stable the IMU pose, the larger the output value.

[0083] Flow field environment suitability evaluation function D flow The disturbance influence vector (i.e., [a]) dist , τ dist ] T ), σ flow The function outputs the stability of the current aerodynamic environment, which is the flow field fluctuation rate (e.g., the variance of pressure sensor data). The output value is between 0 and 1. The smaller the disturbance and the lower the flow field fluctuation rate, the larger the output value.

[0084] Then, an adaptive weighted fusion algorithm is used to calculate the overall recovery alignment confidence: ; Where ω1, ω2, and ω3 are weighting coefficients, satisfying ω1 + ω2 + ω3 = 1. These weighting coefficients are dynamically adjusted using the inverse variance formula: ; in, This corresponds to the measurement noise variance of the sensor / model within the current sampling period. For example, when vision is affected by illumination interference, causing E... vision When noise increases, As the velocity increases, ω2 automatically decreases, and the system becomes more dependent on distance / velocity conditions and the flow field model.

[0085] Threshold determination: In this embodiment of the invention, multiple recovery alignment confidence thresholds are preset. When C is greater than the corresponding recovery alignment confidence threshold, the corresponding collaborative control and recovery actions are executed.

[0086] S4: If the relative positioning information of the UAV meets the preset recovery conditions, a coordinated control command is generated based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, to dynamically adjust the flight parameters of the UAV in order to guide the UAV to be recovered by the vehicle-mounted recovery platform.

[0087] In one feasible implementation, the generation of cooperative control commands based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the on-board recycling platform, in this embodiment of the invention may include, but is not limited to: Extract the vehicle's yaw angle and steering angular velocity from the vehicle's motion state; Using the vehicle yaw angle and the steering angular velocity as coordinate transformation parameters, coordinate system alignment is performed on the active flow field compensation control quantity to obtain the aligned compensation control quantity. The aligned compensation control quantity is used as the feedforward control quantity and superimposed and fused with the feedback control quantity generated based on the relative positioning information; Extract the vehicle speed from the vehicle's motion state and the relative distance from the relative positioning information. Based on the vehicle speed and the relative distance, determine the fusion weight coefficient in real time. The fusion weight coefficient is positively correlated with the vehicle speed and negatively correlated with the relative distance. The cooperative control command is generated based on the feedforward control quantity, the feedback control quantity, and the fusion weight coefficient.

[0088] In the specific implementation process, due to the influence of the disturbance vector D flow Typically, calculations are based on the vehicle coordinate system, while UAV control requires a UAV body coordinate system, thus necessitating coordinate transformation. This embodiment of the invention uses the vehicle yaw angle and the steering angular velocity as coordinate transformation parameters to transform the active flow field compensation control quantity U... comp The aligned compensation control quantity is obtained by rotating from the vehicle coordinate system to the UAV heading coordinate system. .

[0089] Aligned compensation control quantity As the feedforward control variable U ff Simultaneously, feedback control quantities are generated based on relative positioning information (i.e., relative pose error vector E). K p K d The preset proportional-derivative control gain matrix, It can be obtained through differential control. The feedforward control quantity and the feedback control quantity are superimposed and fused to form a preliminary joint control command.

[0090] After generating the initial joint control command, this embodiment of the invention extracts the vehicle speed V from the vehicle's motion state. car The relative distance L in the relative positioning information is used to determine the fusion weight coefficient λ (V) in real time. carThe coefficient is positively correlated with vehicle speed (the higher the vehicle speed, the stronger the airflow disturbance, and the more feedforward compensation is needed), and negatively correlated with relative distance (the closer the distance, the more feedforward is needed to ensure accuracy), with a value range of [0, 1].

[0091] In one feasible implementation, the real-time determination of the fusion weighting coefficient based on the vehicle speed and the relative distance in this embodiment of the invention may include, but is not limited to: When the vehicle speed is less than or equal to a first preset speed threshold and the relative distance is greater than or equal to a first preset distance threshold, the fusion weight coefficient is set to a first weight value. When the vehicle speed is greater than or equal to a second preset speed threshold and the relative distance is less than or equal to a second preset distance threshold, the fusion weight coefficient is set to a second weight value, and the second weight value is greater than the first weight value. When the vehicle speed is between the first preset speed threshold and the second preset speed threshold, and the relative distance is between the first preset distance threshold and the second preset distance threshold, the fusion weight coefficient increases non-linearly with the vehicle speed and the relative distance.

[0092] In specific implementation, the weighting coefficients of this invention are designed according to actual needs, for example: When V car When the speed is ≤20km / h and L≥50m, the airflow disturbance has little impact or no fine compensation is required, so λ=0 (first weight value) is set. When V car When the speed is ≥80km / h and L≤5m, feedforward compensation is the most critical under high-speed close-range conditions, and λ=1 (second weight value) is set. Nonlinear interpolation (such as sigmoid function or linear interpolation) is used in the intermediate region.

[0093] Finally, based on the feedforward control quantity, the feedback control quantity, and the fusion weighting coefficient, the required cooperative control command (cooperative control law) is generated: .

[0094] The command is sent to the UAV's flight control system, where it is converted into motor speed and control surface deflection, enabling dynamic control of the UAV's attitude, speed, and trajectory.

[0095] In one feasible implementation, the preset recovery condition in this embodiment of the invention is a phased triggering condition, and the generation of cooperative control commands to dynamically adjust the flight parameters of the UAV may include, but is not limited to: When the recovery alignment confidence is greater than a first preset alignment confidence threshold, and the relative distance between the drone and the vehicle-mounted recovery platform is greater than a third preset distance threshold or the relative velocity error between the drone and the vehicle-mounted recovery platform is greater than a first preset relative velocity error threshold, a cooperative control instruction is generated according to the feedback control quantity to perform long-distance guidance; When the recovery alignment confidence is greater than or equal to a second preset alignment confidence threshold, the relative distance is less than or equal to the third preset distance threshold and greater than a fourth preset distance threshold, and the relative velocity error is less than or equal to a second preset relative velocity error threshold, a cooperative control instruction is generated according to the feedforward control quantity, the feedback control quantity and the fusion weight coefficient to perform short-distance high-precision pose synchronization; When the recovery alignment confidence is greater than or equal to a third preset alignment confidence threshold, the relative distance is less than or equal to the fourth preset distance threshold, and the relative velocity error is less than or equal to a third preset relative velocity error threshold, the fusion weight coefficient is set to 1, and the differential control gain is increased to n times that in the long-distance guidance stage, so that the drone is smoothly docked with the vehicle-mounted recovery platform.

[0096] In the embodiment of the present invention, the recovery process is divided into three consecutive stages, and different control parameters and trigger conditions are adopted in each stage to adapt to the changes of distance from far to near and precision requirement from low to high. The specific control parameters and trigger conditions can be set as follows: Stage 1: Long-distance guidance stage Trigger conditions: recovery alignment confidence C>0.6, and relative distance L>50m or relative velocity error ΔV>5m / s.

[0097] Control strategy: Only feedback control is used (i.e., λ=0), with the goal of fast track tracking, and the drone focuses on reducing the distance and velocity difference with the recovery platform. The flow field compensation is only monitored and not actually applied.

[0098] Stage 2: Short-distance stabilization stage Trigger conditions: C≥0.8, 5m<L≤50m and ΔV≤2m / s.

[0099] Control strategy: Activate flow field feedforward compensation, and adopt a complete cooperative control law , where λ is dynamically adjusted according to the vehicle speed and distance. The goal is to achieve high-precision pose synchronization and counteract the basic wake disturbance.

[0100] Stage 3: Final recovery stage Trigger conditions: C≥0.9, L≤5m and ΔV≤0.3m / s.

[0101] Control strategy: Control the feedforward control quantity to full power output (i.e., set λ to 1), while increasing the derivative control gain K. d The distance is increased to n times the distance of the long-distance guidance phase (n ranges from 1.2 to 2) to enhance system damping, suppress speed overshoot and attitude oscillation during docking, and ensure that the UAV enters the recovery device smoothly and accurately.

[0102] In one feasible implementation, the dynamic control of the flight parameters of the UAV in this embodiment of the invention further includes: Real-time monitoring of the relative positioning information and the disturbance influence vector; Determine whether the relative positioning information or the disturbance influence vector exceeds a preset safety threshold; If the relative positioning information or the disturbance influence vector exceeds a preset safety threshold, the recovery process is interrupted and the drone is controlled to enter a hovering or detached state.

[0103] In this embodiment of the invention, during the entire recovery process (not just before entry), relative positioning information and disturbance impact vectors are monitored in real time, and it is determined whether they exceed a preset safety threshold. For example: If |y e |>0.2m or |ψ e If a position is more than 5° for more than 0.5 seconds, it is considered an abnormal pose. If ||a dist ||>2m / s 2 Or pressure fluctuation rate σ flow >Threshold value indicates that the flow field disturbance exceeds the limit; If the communication delay is greater than 200ms or the sensor fault flag is set, the system is considered to be in an abnormal state.

[0104] If any of the above exceeds the safety threshold, the system will immediately interrupt the recovery process and use control commands to make the drone hover safely or move away from the vehicle to prevent a collision.

[0105] After recovery, the system uses contact sensors, position lock signals, and attitude lock signals to determine that the drone has been successfully secured in the recovery device. Then, the drone's flight control output is turned off, and the system state machine is reset to standby mode, ready for the next mission.

[0106] The above-mentioned vehicle-mounted UAV recovery control method in a mobile environment provided by the present invention effectively reduces the impact of unsteady airflow on the flight attitude and trajectory of the UAV by real-time sensing and active compensation of the aerodynamic flow field, enabling the UAV to stably approach and accurately align with the recovery platform in a complex aerodynamic environment, and significantly improving the recovery success rate on the mobile platform.

[0107] Furthermore, by comprehensively considering the vehicle's motion state, relative positioning information, and aerodynamic disturbance factors, this embodiment of the invention achieves coordinated control between the UAV and the vehicle, enabling the system to maintain stable operation even under complex conditions such as crosswinds, gusts, vehicle acceleration, or turning.

[0108] Furthermore, this embodiment of the invention uses a phased recovery control and multiple safety verification mechanisms to monitor relative positioning information and disturbance impact vectors in real time. When an anomaly is detected, the recovery can be interrupted in a timely manner and a safety control strategy can be executed, effectively reducing the risk of collision.

[0109] Based on the above-described method for recovering and controlling a vehicle-mounted drone in a mobile environment provided in the first aspect, embodiments of the present invention provide a device for recovering and controlling a vehicle-mounted drone in a mobile environment, such as... Figure 3 As shown, the vehicle-mounted drone recovery control device in the mobile environment includes: The data acquisition module 110 is used to acquire the relative positioning information between the drone and the vehicle-mounted recycling platform in real time, and to perceive the aerodynamic flow field information around the vehicle-mounted recycling platform in real time. The active flow field compensation module 120 is used to determine the disturbance influence vector of aerodynamic disturbance on the UAV based on the aerodynamic flow field information, and generate an active flow field compensation control quantity. The recovery condition determination module 130 is used to determine whether the relative positioning information of the UAV meets the preset recovery conditions; The collaborative control module 140 is used to generate collaborative control commands based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, when the relative positioning information of the UAV meets the preset recovery conditions, so as to dynamically adjust the flight parameters of the UAV to guide the UAV to be recovered with the vehicle-mounted recovery platform.

[0110] In one feasible implementation, the data acquisition module 110 in this embodiment of the invention has the function of: The image of the cooperation mark on the vehicle-mounted recovery platform is acquired by an airborne vision sensor, and the relative position vector and relative yaw angle of the UAV relative to the vehicle-mounted recovery platform are calculated by the PnP algorithm. The three-axis angular velocity and three-axis acceleration of the UAV are obtained by the airborne inertial measurement unit, and the standard deviation of the angular velocity is calculated within a preset sliding time window to obtain the attitude jitter. The absolute distance between the drone and the recycling inlet is obtained by a distance sensor installed at the recycling inlet of the vehicle-mounted recycling platform. The noise covariance matrix of visual observations in the extended Kalman filter is adaptively adjusted according to the attitude jitter amount, and the relative position vector, the relative yaw angle and the absolute distance are used as observation values ​​to be input into the extended Kalman filter for fusion to obtain the relative pose error vector.

[0111] In one feasible implementation, the active flow field compensation module 120 in this embodiment of the invention may include, but is not limited to: The disturbance impact prediction submodule is used to acquire the driving speed and acceleration of the vehicle-mounted recovery platform in real time, the flight speed vector of the UAV and the three-dimensional position coordinates of the UAV relative to the vehicle-mounted recovery platform in real time, and the local airflow rate, airflow direction and pressure change data in the aerodynamic flow field information in real time, and construct an input feature vector; input the input feature vector into a pre-trained unsteady aerodynamic flow field model, and use the unsteady aerodynamic flow field model to predict the disturbance acceleration vector and disturbance torque vector of the UAV caused by the aerodynamic disturbance around the vehicle-mounted recovery platform in future sampling periods.

[0112] The active flow field compensation control quantity generation submodule is used to generate active flow field compensation control quantities based on the predicted disturbance acceleration vector and the disturbance torque vector.

[0113] In one feasible implementation, the recycling condition discrimination module 130 in this embodiment of the invention may include, but is not limited to: The basic condition discrimination submodule is used to determine whether the relative distance and relative speed errors in the relative positioning information meet the first judgment condition. The recovery confidence calculation submodule is used to calculate the recovery alignment confidence based on the relative distance and relative velocity errors in the relative positioning information, combined with the relative pose error in the relative positioning information and the disturbance influence vector, using a weighted fusion algorithm when the relative distance and relative velocity errors in the relative positioning information meet the first determination condition. The recovery confidence determination submodule is used to determine whether the recovery alignment confidence exceeds a preset recovery alignment confidence threshold. If the recovery alignment confidence exceeds the preset recovery alignment confidence threshold, it indicates that the UAV meets the preset recovery alignment conditions.

[0114] In one feasible implementation, the collaborative control module 140 in this embodiment of the invention may include, but is not limited to: The compensation control quantity alignment submodule is used to extract the vehicle yaw angle and steering angular velocity in the vehicle motion state; and to perform coordinate system alignment processing on the active flow field compensation control quantity using the vehicle yaw angle and steering angular velocity as coordinate transformation parameters to obtain the aligned compensation control quantity. The preliminary joint control command generation submodule is used to superimpose and fuse the aligned compensation control quantity as a feedforward control quantity with the feedback control quantity generated based on the relative positioning information to form a preliminary joint control command. The dynamic gain adjustment submodule is used to extract the vehicle speed in the vehicle's motion state and the relative distance in the relative positioning information, and to determine the fusion weight coefficient in real time based on the vehicle speed and the relative distance. The fusion weight coefficient is positively correlated with the vehicle speed and negatively correlated with the relative distance. The collaborative control instruction generation submodule is used to generate the collaborative control instruction based on the feedforward control quantity, the feedback control quantity, and the fusion weight coefficient.

[0115] In one feasible implementation, the dynamic gain adjustment submodule in this embodiment of the invention is specifically used for: When the vehicle speed is less than or equal to a first preset speed threshold and the relative distance is greater than or equal to a first preset distance threshold, the fusion weight coefficient is set to a first weight value. When the vehicle speed is greater than or equal to a second preset speed threshold and the relative distance is less than or equal to a second preset distance threshold, the fusion weight coefficient is set to a second weight value, and the second weight value is greater than the first weight value. When the vehicle speed is between the first preset speed threshold and the second preset speed threshold, and the relative distance is between the first preset distance threshold and the second preset distance threshold, the fusion weight coefficient increases non-linearly with the vehicle speed and the relative distance.

[0116] In one feasible implementation, the preset recycling condition in this embodiment of the invention is a phased triggering condition, and the collaborative control module 140 further includes: The collaborative control command execution submodule is used to generate collaborative control commands for long-distance guidance based on the feedback control quantity when the recovery alignment confidence is greater than a first preset alignment confidence threshold, and the relative distance between the UAV and the vehicle-mounted recovery platform is greater than a third preset distance threshold or the relative speed error between the UAV and the vehicle-mounted recovery platform is greater than a first preset relative speed error threshold; when the recovery alignment confidence is greater than or equal to a second preset alignment confidence threshold, the relative distance is less than or equal to the third preset distance threshold and greater than a fourth preset distance threshold, and the relative speed error is less than or equal to the second preset relative speed error threshold, it generates collaborative control commands for close-range high-precision pose synchronization based on the feedforward control quantity, the feedback control quantity, and the fusion weight coefficient; when the recovery alignment confidence is greater than or equal to a third preset alignment confidence threshold, the relative distance is less than or equal to the fourth preset distance threshold, and the relative speed error is less than or equal to the third preset relative speed error threshold, it sets the fusion weight coefficient to 1 and increases the differential control gain to n times that of the long-distance guidance stage, so that the UAV and the vehicle-mounted recovery platform can dock smoothly.

[0117] In one feasible implementation, the preset recycling condition in this embodiment of the invention is a phased triggering condition, and the collaborative control module 140 further includes: The safety verification and anomaly handling submodule is used to monitor the relative positioning information and the disturbance influence vector in real time; determine whether the relative positioning information or the disturbance influence vector exceeds a preset safety threshold; if the relative positioning information or the disturbance influence vector exceeds the preset safety threshold, interrupt the recovery process and control the drone to enter a hovering or detached state.

[0118] Based on the vehicle-mounted drone recovery control method in a mobile environment provided in the first aspect, this embodiment of the invention also provides a storage medium storing a computer-executable program. The computer-executable program is used to cause a computer to execute the vehicle-mounted drone recovery control method in a mobile environment as described in any implementation of the first aspect. Explanations of the relevant content and descriptions of the beneficial effects of any of the computer-readable storage media provided above can be found in the corresponding embodiments described above, and will not be repeated here.

[0119] Those skilled in the art will understand that the program for implementing all or part of the steps of the above embodiments, which can be executed by a program instructing related hardware, can be stored in a computer-readable storage medium. The storage medium mentioned above can be a read-only memory, a random access memory, etc. The processing unit or processor mentioned above can be a central processing unit, a general-purpose processor, an application-specific integrated circuit (ASIC), a microprocessor (DSP), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof.

[0120] This invention also provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform any of the methods described in the above embodiments. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., SSD), etc.

[0121] It should be noted that the devices for storing computer instructions or computer programs provided in the embodiments of the present invention, such as, but not limited to, the aforementioned memory, computer-readable storage medium, and communication chip, are all non-transitory. Those skilled in the art should recognize that the functions described in the embodiments of the present invention in one or more of the above examples can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium accessible to general-purpose or special-purpose computers.

[0122] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle (UAV) in a mobile environment, characterized in that, include: The relative positioning information between the drone and the vehicle-mounted recycling platform is acquired in real time, and the aerodynamic flow field information around the vehicle-mounted recycling platform is perceived in real time. Based on the aerodynamic flow field information, the disturbance influence vector of aerodynamic disturbance on the UAV is determined, and an active flow field compensation control quantity is generated; Determine whether the relative positioning information of the drone meets the preset recovery conditions; If the relative positioning information of the UAV meets the preset recovery conditions, a coordinated control command is generated based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, to dynamically adjust the flight parameters of the UAV in order to guide the UAV to be recovered by the vehicle-mounted recovery platform. The step of generating cooperative control commands based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the on-board recycling platform, includes: Extract the vehicle's yaw angle and steering angular velocity from the vehicle's motion state; Using the vehicle yaw angle and the steering angular velocity as coordinate transformation parameters, coordinate system alignment is performed on the active flow field compensation control quantity to obtain the aligned compensation control quantity. The aligned compensation control quantity is used as the feedforward control quantity and superimposed and fused with the feedback control quantity generated based on the relative positioning information; Extract the vehicle speed from the vehicle's motion state and the relative distance from the relative positioning information. Based on the vehicle speed and the relative distance, determine the fusion weight coefficient in real time. The fusion weight coefficient is positively correlated with the vehicle speed and negatively correlated with the relative distance. The cooperative control command is generated based on the feedforward control quantity, the feedback control quantity, and the fusion weight coefficient.

2. The method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment according to claim 1, characterized in that, The relative positioning information includes a relative pose error vector, and the real-time acquisition of the relative positioning information between the UAV and the vehicle-mounted recovery platform includes: The image of the cooperation mark on the vehicle-mounted recovery platform is acquired by an airborne vision sensor, and the relative position vector and relative yaw angle of the UAV relative to the vehicle-mounted recovery platform are calculated by the PnP algorithm. The three-axis angular velocity and three-axis acceleration of the UAV are obtained by the airborne inertial measurement unit, and the standard deviation of the angular velocity is calculated within a preset sliding time window to obtain the attitude jitter. The absolute distance between the drone and the recycling inlet is obtained by a distance sensor installed at the recycling inlet of the vehicle-mounted recycling platform. The noise covariance matrix of visual observations in the extended Kalman filter is adaptively adjusted according to the attitude jitter amount, and the relative position vector, the relative yaw angle and the absolute distance are used as observation values ​​to be input into the extended Kalman filter for fusion to obtain the relative pose error vector.

3. The method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment according to claim 1, characterized in that, The step of determining the disturbance influence vector of aerodynamic disturbance on the UAV based on the aerodynamic flow field information includes: The vehicle-mounted recovery platform's driving speed and acceleration are acquired in real time, the drone's flight speed vector and the drone's three-dimensional position coordinates relative to the vehicle-mounted recovery platform are acquired in real time, and the local airflow rate, airflow direction and pressure change data in the aerodynamic flow field information are acquired in real time to construct an input feature vector. The input feature vector is input into a pre-trained unsteady aerodynamic flow field model, and the unsteady aerodynamic flow field model is used to predict the disturbance acceleration vector and disturbance torque vector of the aerodynamic disturbance around the vehicle-mounted recovery platform on the UAV during future sampling periods.

4. The method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment according to claim 1, characterized in that, The step of determining whether the relative positioning information of the drone meets the preset recovery conditions includes: Determine whether the relative distance and relative speed errors in the relative positioning information meet the first determination condition; If the relative distance and relative velocity errors in the relative positioning information meet the first determination condition, then based on the relative distance and relative velocity errors, combined with the relative pose error in the relative positioning information and the disturbance influence vector, the weighted fusion algorithm is used to calculate the recovery alignment confidence. Determine whether the recovery alignment confidence level exceeds a preset recovery alignment confidence threshold; If the recovery alignment confidence level exceeds a preset recovery alignment confidence level threshold, it indicates that the UAV meets the preset recovery alignment conditions.

5. The method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment according to claim 1, characterized in that, The step of determining the fusion weighting coefficient in real time based on the vehicle speed and the relative distance includes: When the vehicle speed is less than or equal to a first preset speed threshold and the relative distance is greater than or equal to a first preset distance threshold, the fusion weight coefficient is set to a first weight value. When the vehicle speed is greater than or equal to a second preset speed threshold and the relative distance is less than or equal to a second preset distance threshold, the fusion weight coefficient is set to a second weight value, and the second weight value is greater than the first weight value. When the vehicle speed is between the first preset speed threshold and the second preset speed threshold, and the relative distance is between the first preset distance threshold and the second preset distance threshold, the fusion weight coefficient increases non-linearly with the vehicle speed and the relative distance.

6. The method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment according to claim 4, characterized in that, The preset recovery conditions are phased triggering conditions, and the generation of collaborative control commands to dynamically adjust the flight parameters of the UAV includes: When the recovery alignment confidence is greater than the first preset alignment confidence threshold, and the relative distance between the UAV and the vehicle-mounted recovery platform is greater than the third preset distance threshold or the relative speed error between the UAV and the vehicle-mounted recovery platform is greater than the first preset relative speed error threshold, a collaborative control command is generated based on the feedback control quantity for long-distance guidance. When the recovery alignment confidence is greater than or equal to the second preset alignment confidence threshold, the relative distance is less than or equal to the third preset distance threshold and greater than the fourth preset distance threshold, and the relative speed error is less than or equal to the second preset relative speed error threshold, a cooperative control command is generated to perform close-range high-precision pose synchronization based on the feedforward control quantity, the feedback control quantity and the fusion weight coefficient. When the recovery alignment confidence is greater than or equal to the third preset alignment confidence threshold, the relative distance is less than or equal to the fourth preset distance threshold, and the relative speed error is less than or equal to the third preset relative speed error threshold, the fusion weight coefficient is set to 1, and the differential control gain is increased to n times that of the long-distance guidance stage, so that the UAV and the vehicle-mounted recovery platform can dock smoothly.

7. The method for controlling the recovery of a vehicle-mounted unmanned aerial vehicle in a mobile environment according to claim 1, characterized in that, The dynamic control of the flight parameters of the UAV also includes: Real-time monitoring of the relative positioning information and the disturbance influence vector; Determine whether the relative positioning information or the disturbance influence vector exceeds a preset safety threshold; If the relative positioning information or the disturbance influence vector exceeds a preset safety threshold, the recovery process is interrupted and the drone is controlled to enter a hovering or detached state.

8. A vehicle-mounted unmanned aerial vehicle (UAV) recovery control device for mobile environments, characterized in that, include: The data acquisition module is used to acquire the relative positioning information between the drone and the vehicle-mounted recycling platform in real time, and to perceive the aerodynamic flow field information around the vehicle-mounted recycling platform in real time. The active flow field compensation module is used to determine the disturbance influence vector of aerodynamic disturbance on the UAV based on the aerodynamic flow field information, and generate active flow field compensation control quantity; The recovery condition determination module is used to determine whether the relative positioning information of the UAV meets the preset recovery conditions; The collaborative control module is used to generate collaborative control commands based on the relative positioning information and the active flow field compensation control quantity, combined with the vehicle motion state of the vehicle-mounted recovery platform, when the relative positioning information of the UAV meets the preset recovery conditions, so as to dynamically adjust the flight parameters of the UAV and guide the UAV to be recovered by the vehicle-mounted recovery platform. The collaborative control module includes: The compensation control quantity alignment submodule is used to extract the vehicle yaw angle and steering angular velocity in the vehicle motion state; and to perform coordinate system alignment processing on the active flow field compensation control quantity using the vehicle yaw angle and steering angular velocity as coordinate transformation parameters to obtain the aligned compensation control quantity. The preliminary joint control command generation submodule is used to superimpose and fuse the aligned compensation control quantity as a feedforward control quantity with the feedback control quantity generated based on the relative positioning information to form a preliminary joint control command. The dynamic gain adjustment submodule is used to extract the vehicle speed in the vehicle's motion state and the relative distance in the relative positioning information, and to determine the fusion weight coefficient in real time based on the vehicle speed and the relative distance. The fusion weight coefficient is positively correlated with the vehicle speed and negatively correlated with the relative distance. The collaborative control instruction generation submodule is used to generate the collaborative control instruction based on the feedforward control quantity, the feedback control quantity, and the fusion weight coefficient.

9. An electronic device, characterized in that, include: A memory, one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, which, when executed by the processor, cause the electronic device to perform the vehicle-mounted unmanned aerial vehicle recovery control method in a mobile environment as described in any one of claims 1 to 7.

Images