Drone airport battery replacement control method and system

Abstract

This invention discloses a battery swapping control method and system for unmanned aerial vehicles (UAVs), relating to the field of UAV control technology. The method includes: estimating the pose deviation between a panoramic image from a global camera and a local image from a robotic arm's end effector camera; performing visual servo fusion using camera confidence weights to obtain the desired pose of the end effector relative to a target position; controlling the robotic arm to move to an alignment area at a set distance from the target position, and obtaining the precise pose of the actuator relative to the target position based on depth information from the target position; when the robotic arm reaches the target position and performs battery insertion / removal actions, estimating the contact force of the actuator based on torque sensing data from each joint of the robotic arm; when the contact force exceeds a set threshold, adjusting the attitude based on the impedance relationship between the actuator position and the contact force until the battery is fully inserted. This achieves battery swapping control through multi-source data fusion.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and in particular to a method and system for controlling the battery swapping of UAVs at airports. Background Technology

[0002] The core function of a battery-swapping drone airport is to provide rapid and automated battery replacement services for drones on missions. Its implementation relies primarily on a high-precision positioning and control system, as well as automated mechanical equipment. After the drone automatically returns to the airport, a robotic arm automatically removes the old battery and installs a new one. The reliability and accuracy of the battery-swapping control system directly impact the drone's operational efficiency and safety. The core of the battery-swapping control system lies in achieving high-precision, high-reliability battery insertion and removal operations. Therefore, achieving high-precision, high-robust control of the drone battery-swapping process through comprehensive evaluation is particularly important.

[0003] Traditional solutions often rely on monocular or binocular vision guidance. When there are drastic changes in lighting (such as backlight or low light) or when the battery surface features are blurred or there is oil or dirt obstructing the view, the visual positioning is prone to failure or a sharp drop in accuracy.

[0004] Secondly, there may be slight assembly tolerances in the installation of the battery inside the drone body, and the actual position of the battery connector cannot be accurately reflected by the positioning of the body alone.

[0005] In addition, if the battery swapping robotic arm relies solely on closed-loop position control when inserting or removing batteries, it is prone to generating a huge impact force due to positional errors when encountering obstruction, which can damage the battery interface or the robotic arm itself. Summary of the Invention

[0006] To address the aforementioned issues, this invention proposes a battery swapping control method and system for unmanned aerial vehicles (UAVs). The system utilizes a global camera for coarse guidance, a terminal camera for precise alignment, a laser rangefinder for depth constraints, and a torque sensor for safe interaction. This multi-source data fusion enables battery swapping control.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for controlling battery swapping at an unmanned aerial vehicle (UAV) airport, comprising: The pose estimation discrepancy between the panoramic image acquired by the global camera and the local image acquired by the end-effector camera of the robotic arm is addressed by using camera confidence weights for visual servo fusion, thereby obtaining the desired pose of the end-effector relative to the target position. After the robotic arm is moved to the alignment area at a set distance from the target position according to the desired pose, the precise pose of the actuator relative to the target position is obtained based on the depth information of the distance from the target position. When the robotic arm reaches the target position and performs battery insertion / removal actions based on the precise pose control, the contact force of the actuator is estimated based on the torque sensor data of each joint of the robotic arm. When the contact force is greater than the set threshold, the posture is adjusted according to the impedance relationship between the actuator position and the contact force until the battery is fully inserted.

[0008] As an alternative implementation method, the visual servo fusion error obtained after visual servo fusion... for: ; in, The pose estimation deviation between the current preliminary pose calculated from the panoramic image of the global camera and the target pose; The pose estimation deviation between the current preliminary pose calculated from the local image of the end camera and the target pose; For robust kernel functions; and These are the global camera confidence weights and the terminal camera confidence weights, respectively, satisfying... .

[0009] As an alternative implementation method, the confidence weight is calculated as follows: ; Where i represents the camera, g is the global camera, and e is the end camera; The ratio of feature-matched inlier points between the current frame image acquired by the i-th camera and the target image; The variance of the pose estimation deviation between the current preliminary pose calculated by the i-th camera and the target pose; Let be the exponential decay factor of the i-th camera.

[0010] As an alternative implementation method, the ICP error function for calculating the precise pose based on the depth information from the target location is: ; Where R is the rotation matrix; t is the translation vector; These are image feature points in a local image captured by the end-point camera; These are the corresponding image feature points in the preset model library; yes The normal vector; It is the distance to the battery port calculated by the laser rangefinder in the m-th calculation; It is the theoretical distance calculated based on the current pose (R,t); It is the laser data weighting factor.

[0011] As an alternative implementation method, contact force for: ; Where J(q) is the Jacobian matrix of the robotic arm; These are measured values ​​from torque sensors; It is the compensation value of the joint friction torque model; It is the compensation value of the gravity torque model.

[0012] As an alternative implementation, the impedance relationship between the actuator position and the contact force is: ; in, This is the actual position of the end effector; It is the expected position of the end effector; These are actual velocity and actual acceleration; These are the expected velocity and the expected acceleration; These are the desired inertia matrix, desired damping matrix, and desired stiffness matrix, respectively.

[0013] Secondly, the present invention provides a UAV airport battery swapping control system, comprising: The fusion module is configured to perform visual servo fusion using camera confidence weights on the pose estimation deviation between the panoramic image acquired by the global camera and the local image acquired by the end-effector of the robotic arm, thereby obtaining the desired pose of the end-effector of the robotic arm relative to the target position. The depth constraint module is configured to control the robotic arm to move to an alignment area at a set distance from the target position according to the desired pose, and then obtain the precise pose of the actuator relative to the target position based on the depth information of the distance from the target position. The insertion and removal control module is configured to estimate the contact force of the actuator based on the torque sensing data of each joint of the robotic arm when the robotic arm reaches the target position and performs the battery insertion and removal action according to the precise pose control. When the contact force is greater than the set threshold, the attitude is adjusted according to the impedance relationship between the actuator position and the contact force until the battery is fully inserted.

[0014] Thirdly, the present invention provides an electronic device including a memory and a processor, and computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the method described in the first aspect.

[0015] Fourthly, the present invention provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the method described in the first aspect.

[0016] Fifthly, the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the method described in the first aspect.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention innovatively proposes a multi-source data fusion architecture encompassing global coarse positioning, local fine positioning, depth constraints, and contact force sensing. It designs a multi-source data fusion positioning method for UAV airport battery swapping control, constructs a visual positioning model fused with dynamic confidence weights and depth constraints, and develops a battery swapping control system combining visual guidance and force control compliance. In the visual guidance stage, dynamic confidence weights are introduced to dynamically fuse pose calculation results from the global camera and the end-effector camera, improving anti-interference capabilities. Simultaneously, depth constraints provided by the end-effector laser rangefinder accelerate and correct the convergence of the visual iterative nearest-point algorithm, improving control accuracy. In the battery insertion / removal stage, joint torque sensing data is introduced, switching the battery swapping robotic arm from pure position control to impedance control. Position adjustments are made based on impedance relationships to avoid rigid impacts, improving the compliance and safety of the insertion / removal process. This addresses the problems of visual positioning being susceptible to environmental interference and the large rigid impacts during battery insertion / removal, achieving precise positioning and smooth, safe insertion / removal, reducing the risk of operational errors and equipment damage, and improving environmental anti-interference capabilities, pose calculation accuracy, and battery swapping control accuracy.

[0018] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart of the UAV airport battery swapping control method provided in Embodiment 1 of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0023] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form as well. Furthermore, it should be understood that the terms “comprising” and “including”, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0025] Existing drone airport battery swapping technologies are mainly divided into two categories: The first category is a battery swapping method based on mechanical positioning. This method uses mechanical structures such as centering mechanisms and limiting devices to fix the drone in a predetermined position, and then a robotic arm performs the battery swapping operation. After long-term use, mechanical wear and tear can lead to a decrease in positioning accuracy, and it also has poor adaptability to different drone models.

[0026] The second type is a battery swapping method based on visual servoing. Image data is acquired through a fixed camera and an end-effector camera. After extracting feature points, the pose is estimated based on a perspective n-point algorithm. The confidence weights of each camera are calculated and then weighted averaged and fused to generate control commands for the robotic arm.

[0027] The existing methods described above still have the following problems in practical applications: (1) The sensor type is limited, relying solely on vision sensors, which are prone to failure under complex lighting conditions.

[0028] (2) Vibration or slight displacement of the battery swapping platform may occur during transportation or operation, causing the preset calibration parameters to fail and the battery swapping success rate to decrease.

[0029] (3) The lack of a dynamic correction mechanism for the relative position of the end effector of the robotic arm and the battery compartment makes it easy for the battery to get stuck or damaged under the influence of cumulative errors.

[0030] Therefore, this invention proposes a UAV airport battery swapping control method based on multi-source data fusion. The global camera provides coarse guidance, the terminal camera provides precise alignment, the laser rangefinder provides depth constraint, and the torque sensor provides safe interaction. Thus, the final command is generated through multi-source data fusion to achieve battery swapping control.

[0031] Example 1 like Figure 1 As shown in the figure, this embodiment proposes a battery swapping control method for unmanned aerial vehicles (UAVs) at airports, including: The pose estimation discrepancy between the panoramic image acquired by the global camera and the local image acquired by the end-effector camera of the robotic arm is addressed by using camera confidence weights for visual servo fusion, thereby obtaining the desired pose of the end-effector relative to the target position. After the robotic arm is moved to the alignment area at a set distance from the target position according to the desired pose, the precise pose of the actuator relative to the target position is obtained based on the depth information of the distance from the target position. When the robotic arm reaches the target position and performs the battery insertion / removal action based on the precise pose control, the contact force of the actuator is estimated based on the torque sensor data of each joint of the robotic arm. When the contact force exceeds a set threshold, the posture is adjusted according to the impedance relationship between the actuator position and the contact force until the battery is fully inserted. The method of this embodiment will be described in detail below.

[0032] S1: After the drone lands on the battery swapping platform, it first undergoes preliminary physical calibration via the centering mechanism on the platform. Then, a panoramic image of the entire view is acquired using a global camera. Visual labels for the battery compartment area are identified based on the panoramic image, and the initial pose is solved using the Perspective N-Point Problem (PnP) algorithm. Finally, the initial pose will be... Perform coordinate transformation to convert it into a target point in the robot arm's base coordinate system, so as to guide the robot arm to move to the pre-observed position (a set distance from the battery port, such as 10cm).

[0033] In this step, considering the recognition jitter caused by uneven illumination, time-series filtering is introduced when calculating the pose: ; in, These are the filter coefficients; The pose matrix calculated for the current frame; This is the optimal estimate from the previous frame. This is the optimal estimate for the current frame.

[0034] In this step, after the drone autonomously lands and docks in the working area of ​​the battery swapping platform, the platform's integrated centering and correction mechanism first corrects the drone's physical attitude and position. As an example, the centering mechanism can consist of elastic limit baffles, guide ramps, and electric push-pull components. Based on the mechanical limit adaptive retraction principle, it limits and pushes the drone's body laterally and longitudinally, gradually correcting the drone to the center of the preset reference workstation on the battery swapping platform. This eliminates problems such as drone landing deviation, body tilt, and positional offset, ensuring the drone's overall posture is regular and providing a stable physical reference for subsequent visual data acquisition, accurate recognition, and robotic arm operations.

[0035] In this step, a global camera is installed on the top of the battery swapping platform using a fixed hoisting method, covering the path of the drone entering the warehouse. After the drone lands on the battery swapping platform and completes physical centering correction, the global camera collects a coarse panoramic image of the overall perspective to identify the battery compartment area.

[0036] As an example, high-contrast, standardized visual QR code labels are affixed to the outer edge of the battery compartment as location markers for the battery compartment area. The original panoramic image is optimized using machine vision preprocessing algorithms such as grayscale preprocessing, image noise reduction, and threshold segmentation to remove environmental noise. Then, combined with template matching and contour detection algorithms, the visual label area of ​​the battery compartment within the panoramic image is quickly located, and the target range of the battery compartment is accurately selected, thus completing the precise extraction of the target area.

[0037] After completing the visual label recognition and positioning, the PnP perspective N-point solution algorithm is introduced to perform preliminary pose calculation. Then, the preliminary pose is transferred to the corresponding target point of the battery compartment in the robot arm's base coordinate system. Based on this, the safe movement path of the robot arm is planned, and the robot arm is controlled to move smoothly to the pre-observation position, which is a preliminary preparation for subsequent automated operation.

[0038] S2: When the robotic arm moves to the pre-observation position, the end-effector camera of the robotic arm acquires a local image. The local image acquisition range focuses on the battery compartment area, which complements the panoramic image acquired by the global camera, forming a dual-view data complement, laying the data foundation for subsequent pose fusion.

[0039] For local images, the PnP perspective N-point solution algorithm can also be used to solve the initial pose under the end-effector camera. At the same time, in order to overcome the feature loss or blurring that may occur in a single viewpoint, a weighted fusion pose estimation algorithm based on camera confidence is introduced. At this time, the visual servo fusion stage is entered to obtain the desired pose of the end effector of the robotic arm relative to the target position.

[0040] Specifically: the visual servo fusion error is: ; in, For visual servo fusion error; The pose estimation deviation between the current preliminary pose calculated from the panoramic image of the global camera and the target pose; The pose estimation deviation between the current preliminary pose calculated from the local image of the end camera and the target pose; This is a robust kernel function used to suppress outliers; and These are the global camera confidence weights and the terminal camera confidence weights, respectively, satisfying... .

[0041] The confidence weight is calculated as follows: ; Where i represents the camera, g is the global camera, and e is the end camera; The ratio of feature-matched inliers between the current frame image acquired by the i-th camera and the target image reflects the richness of features. The variance of the pose estimation deviation between the current preliminary pose calculated by the i-th camera and the target pose reflects the uncertainty; Let be the exponential decay factor for the i-th camera; This is the normalized denominator.

[0042] The feature matching inlier ratio is as follows: image features, including QR code label corner points and battery compartment corner points, are extracted from the current frame panoramic image acquired by the global camera and the current frame local image acquired by the end camera; the image features of the current frame image under each camera are matched with the corresponding image features in the target location image to obtain several candidate matching pairs; each candidate matching pair is then verified, and the candidate matching pairs that are considered correct after verification are regarded as inliers; the ratio of the number of inliers to the total number of candidate matching pairs is the inlier ratio.

[0043] This formula allows cameras with rich features and low uncertainty to receive higher fusion weights. After obtaining the visual servoing fusion error, the desired pose of the current end effector relative to the target position is obtained by combining it with the current actual pose. .

[0044] S3: Under visual servo fusion control, based on the desired pose The robotic arm is controlled to move to an alignment area at a set distance from the target position (e.g., 2-3 cm from the battery port). At this point, the positioning accuracy may decrease due to visual defocusing or a lack of feature points, so a laser rangefinder sensor installed at the end of the robotic arm is used in conjunction with this.

[0045] Laser rangefinders provide high-precision depth information. It is used to construct local point cloud constraints in the coordinate system of the end camera, utilizing depth information. Accelerate the convergence of the visual ICP (Iterative Closest Point) algorithm to obtain the precise pose of the end effector relative to the battery port. Local point cloud constraints refer to the geometric consistency of all point cloud data during the solution process.

[0046] The laser-assisted ICP error function is as follows: ; Where R is the rotation matrix; t is the translation vector; These are image feature points (QR code corner points and battery compartment corner points) of a local image of the battery port captured by the end camera. These are the corresponding image feature points in the preset model library; yes The normal vector; It is the distance to the battery port calculated by the laser rangefinder in the m-th calculation; It is the theoretical distance calculated based on the current pose (R,t); It is the laser data weighting factor.

[0047] S4: Based on the precise pose of the actuator relative to the target position, guide the robotic arm to the battery port and perform the battery insertion / removal action; at this time, due to the friction and possible structural obstruction at the battery port, compliant control is required.

[0048] Specifically: (1) Real-time acquisition of torque sensing data of each joint of the robotic arm, and estimation of the contact force on the end effector in Cartesian space through dynamic model. .

[0049] The mapping from joint torque sensing data to end effector contact force is as follows: ; Where J(q) is the Jacobian matrix of the robotic arm; These are measured values ​​from torque sensors; It is the compensation value of the joint friction torque model; These are the compensation values ​​for the gravity torque model, and the model compensation values ​​are preset.

[0050] (2) Based on contact force Switch the robotic arm to impedance control mode. The goal of impedance control is to establish the end effector position correction. (The difference between the actual end effector position and the desired end effector position) and contact force The dynamic relationship between these elements enables the robotic arm to exhibit the desired mass-spring-damping characteristics.

[0051] The impedance control model is as follows: ; in, This is the actual position of the end effector; It is the expected position of the end effector; These are actual velocity and actual acceleration; These are the expected velocity and the expected acceleration; These are the desired inertia matrix, desired damping matrix, and desired stiffness matrix, respectively.

[0052] (3) When contact force is detected When the set threshold is exceeded (e.g., excessive insertion resistance), the position is adjusted according to the impedance control model to avoid rigid impact.

[0053] (4) When the force feedback shows that the battery has been fully inserted (such as when the buckle is in place and a characteristic force signal is generated), the robotic arm stops moving and releases the battery gripper.

[0054] This method utilizes the aforementioned fusion mechanism, with a global camera providing coarse guidance, a terminal camera providing precise alignment, a laser rangefinder providing depth constraints, and a torque sensor providing safe interaction. Ultimately, it achieves battery swapping control through the fusion of multi-source data.

[0055] The advantages of the above-mentioned UAV airport battery swapping control method based on multi-source data fusion are: (1) Construct a multi-source data fusion architecture of global coarse positioning, local fine positioning and contact force sensing, deploy different sensors in working stages that are suitable for their physical characteristics, introduce dynamic confidence weights in the visual guidance stage, dynamically fuse the pose calculation results of the global camera and the end camera, and improve the environmental anti-interference capability.

[0056] (2) By utilizing the depth constraints provided by the end laser rangefinder, the convergence of the visual iterative closest point (ICP) algorithm is accelerated and corrected, thereby further improving the control accuracy.

[0057] (3) During the battery insertion and removal stage, joint torque sensing data is introduced to switch the battery swapping robot arm from pure position control to impedance control. The robot arm is adjusted according to the impedance relationship to avoid rigid impact and improve the smoothness and safety of the insertion and removal process.

[0058] Example 2 This embodiment provides a UAV airport battery swapping control system, including: The fusion module is configured to perform visual servo fusion using camera confidence weights on the pose estimation deviation between the panoramic image acquired by the global camera and the local image acquired by the end-effector of the robotic arm, thereby obtaining the desired pose of the end-effector of the robotic arm relative to the target position. The depth constraint module is configured to control the robotic arm to move to an alignment area at a set distance from the target position according to the desired pose, and then obtain the precise pose of the actuator relative to the target position based on the depth information of the distance from the target position. The insertion and removal control module is configured to estimate the contact force of the actuator based on the torque sensing data of each joint of the robotic arm when the robotic arm reaches the target position and performs the battery insertion and removal action according to the precise pose control. When the contact force is greater than the set threshold, the attitude is adjusted according to the impedance relationship between the actuator position and the contact force until the battery is fully inserted.

[0059] It should be noted that the above modules correspond to the steps described in Embodiment 1, and the examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1. It should also be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0060] In further embodiments, the following is also provided: An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the method described in Embodiment 1. For brevity, further details are omitted here.

[0061] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0062] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.

[0063] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the method described in Embodiment 1.

[0064] The method in Example 1 can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, a detailed description is not provided here.

[0065] A computer program product includes a computer program that, when executed by a processor, implements the method described in Embodiment 1.

[0066] The present invention also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as instructions included in program modules, which execute in a device on a target real or virtual processor to perform the processes / methods described above. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc., that perform specific tasks or implement specific abstract data types. In various embodiments, the functionality of program modules can be combined or divided among program modules as needed. The machine-executable instructions for the program modules can execute within a local or distributed device. In a distributed device, the program modules can reside in both local and remote storage media.

[0067] The computer program code used to implement the methods of the present invention may be written in one or more programming languages. This computer program code may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the computer or other programmable data processing device, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a computer, partially on a computer, as a stand-alone software package, partially on a computer and partially on a remote computer, or entirely on a remote computer or server.

[0068] In the context of this invention, computer program code or related data may be carried by any suitable carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, and the like. Examples of signals may include electrical, optical, radio, sound, or other forms of propagation signals, such as carrier waves, infrared signals, etc.

[0069] Those skilled in the art will recognize that the units and algorithm steps described in connection with the various examples of this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0070] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for controlling battery swapping at an unmanned aerial vehicle (UAV) airport, characterized in that, include: The pose estimation discrepancy between the panoramic image acquired by the global camera and the local image acquired by the end-effector camera of the robotic arm is addressed by using camera confidence weights for visual servo fusion, thereby obtaining the desired pose of the end-effector relative to the target position. After the robotic arm is moved to the alignment area at a set distance from the target position according to the desired pose, the precise pose of the actuator relative to the target position is obtained based on the depth information of the distance from the target position. When the robotic arm reaches the target position and performs battery insertion / removal actions based on the precise pose control, the contact force of the actuator is estimated based on the torque sensor data of each joint of the robotic arm. When the contact force is greater than the set threshold, the posture is adjusted according to the impedance relationship between the actuator position and the contact force until the battery is fully inserted.

2. The UAV airport battery swapping control method as described in claim 1, characterized in that, Visual servo fusion error obtained after visual servo fusion for: ； in, The pose estimation deviation between the current preliminary pose calculated from the panoramic image of the global camera and the target pose; The pose estimation deviation between the current preliminary pose calculated from the local image of the end camera and the target pose; For robust kernel functions; and These are the global camera confidence weights and the terminal camera confidence weights, respectively, satisfying... .

3. The UAV airport battery swapping control method as described in claim 2, characterized in that, The confidence weight is calculated as follows: ； Where i represents the camera, g is the global camera, and e is the end camera; The ratio of feature-matched inlier points between the current frame image acquired by the i-th camera and the target image; The variance of the pose estimation deviation between the current preliminary pose calculated by the i-th camera and the target pose; Let be the exponential decay factor of the i-th camera.

4. The UAV airport battery swapping control method as described in claim 1, characterized in that, The ICP error function for calculating the precise pose based on the depth information from the target location is: ； Where R is the rotation matrix; t is the translation vector; These are image feature points in a local image captured by the end-point camera; These are the corresponding image feature points in the preset model library; yes The normal vector; It is the distance to the battery port calculated by the laser rangefinder in the m-th calculation; It is the theoretical distance calculated based on the current pose (R,t); It is the laser data weighting factor.

5. The UAV airport battery swapping control method as described in claim 1, characterized in that, Contact force for: ； Where J(q) is the Jacobian matrix of the robotic arm; These are measured values ​​from torque sensors; It is the compensation value of the joint friction torque model; It is the compensation value of the gravity torque model.

6. The UAV airport battery swapping control method as described in claim 1, characterized in that, The impedance relationship between the actuator position and the contact force is as follows: ； in, This is the actual position of the end effector; It is the expected position of the end effector; These are actual velocity and actual acceleration; These are the expected velocity and the expected acceleration; These are the desired inertia matrix, desired damping matrix, and desired stiffness matrix, respectively.

7. A UAV airport battery swapping control system, characterized in that, include: The fusion module is configured to perform visual servo fusion using camera confidence weights on the pose estimation deviation between the panoramic image acquired by the global camera and the local image acquired by the end-effector of the robotic arm, thereby obtaining the desired pose of the end-effector of the robotic arm relative to the target position. The depth constraint module is configured to control the robotic arm to move to an alignment area at a set distance from the target position according to the desired pose, and then obtain the precise pose of the actuator relative to the target position based on the depth information of the distance from the target position. The insertion and removal control module is configured to estimate the contact force of the actuator based on the torque sensing data of each joint of the robotic arm when the robotic arm reaches the target position and performs the battery insertion and removal action according to the precise pose control. When the contact force is greater than the set threshold, the attitude is adjusted according to the impedance relationship between the actuator position and the contact force until the battery is fully inserted.

8. An electronic device, characterized in that, It includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, Used to store computer instructions, which, when executed by a processor, perform the method described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, implements the method described in any one of claims 1-6.

Images