System and method for realizing air end tool switching

Abstract

The invention provides a system and a method for realizing air end tool switching, and belongs to the technical field of air docking of unmanned aerial vehicles. The system provided by the invention comprises an operation unmanned aerial vehicle, the operation unmanned aerial vehicle is provided with a terminal tool, the operation unmanned aerial vehicle is configured to obtain positioning information of a toolbox unmanned aerial vehicle below and size information of a plurality of replacement tools on the toolbox unmanned aerial vehicle in the air, and a target replacement tool is determined based on the size information, realizing end tool switching based on contact between the positioning information and the target replacement tool; a plurality of replacement tools are arranged on the top plane of the tool box unmanned aerial vehicle, and when the operation unmanned aerial vehicle makes contact with the tool box unmanned aerial vehicle to replace the tail end tools, the tool box unmanned aerial vehicle body and the operation unmanned aerial vehicle body are in flexible butt joint. According to the system and the method for realizing air terminal tool switching provided by the invention, the air flexible docking of the unmanned aerial vehicle can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle air docking, and in particular to a system and method for realizing air end tool switching. BACKGROUND

[0002] With the deep integration of unmanned aerial vehicle technology and mechanical arm technology, aerial operation robots have become the core equipment for replacing manual work to complete high-risk, high-altitude and remote area operations due to their combined advantages of rapid movement and high-precision control, and are widely used in power inspection, wind turbine maintenance, high-altitude rescue, infrastructure detection and other fields.

[0003] The existing technology mainly falls into two categories: one is that the aerial operation robot is provided with a single end execution mechanism, and can only complete a specific task, and if multiple-step operation is required, the aerial operation robot needs to return to the ground base station multiple times; the other is a ground fixed tool box scheme proposed in the prior art, that is, a ground tool box is arranged near the operation starting point of the aerial operation robot, and the aerial operation robot returns to the ground tool box to realize the replacement of the end tool. This scheme has been tried in some low-altitude and short-distance operation scenarios, and to some extent, it solves the functional limitations of a single tool.

[0004] However, the existing technology still has significant defects and cannot adapt to complex operation scenarios at high altitudes and long distances. On the one hand, the single end execution mechanism scheme cannot meet the needs of multi-step collaborative operation, for example, wind turbine blade cleaning needs to be performed by first spraying a cleaning agent with a spraying tool and then removing stains with a brushing tool. This scheme requires the deployment of multiple aerial operation robots with different functions to work collaboratively, which not only increases equipment costs but also poses safety risks in multi-robot collaborative scheduling. On the other hand, the ground fixed tool box scheme has low switching efficiency. Since the operation site of the aerial operation robot is often far away from the ground tool box and at a high altitude, the unmanned aerial vehicle spends a large proportion of time returning to the ground to replace the tool, which greatly reduces the operation efficiency. Moreover, the unmanned aerial vehicle is easily affected by air flow and obstacles during the return process, which increases the risk of equipment damage and operation delay and cannot meet the needs of efficient and continuous high-altitude operation. SUMMARY

[0005] Therefore, the present application provides a system and method for realizing air end tool switching to realize flexible docking of unmanned aerial vehicles in the air.

[0006] Specifically, the present application is realized by the following technical solutions:

[0007] The first aspect of the present application provides a system for realizing air end tool switching, which comprises:

[0008] The operational drone is equipped with an end-effector tool. The operational drone is configured to acquire, while in the air, the positioning information of a toolbox drone below and the size information of multiple replacement tools on the toolbox drone, determine a target replacement tool based on the size information, and make contact with the target replacement tool based on the positioning information to achieve end-effector tool switching.

[0009] The toolbox drone has multiple replacement tools on its top plane, and each replacement tool has a guide positioning post on its top. The working drone has a docking interface on its bottom, which is shaped and connected to the guide positioning post. When the working drone and the toolbox drone come into contact to replace the end tool, the body of the toolbox drone and the body of the working drone flexibly dock.

[0010] A second aspect of this application provides a method for implementing over-the-air terminal tool handover, the method comprising:

[0011] Obtain the positioning information of the toolbox drone and the size information of multiple replacement tools;

[0012] The target replacement tool is determined based on the size information;

[0013] The relative positions of the operation drone and the toolbox drone are determined based on the positioning information.

[0014] Based on the relative position, the operation drone and the toolbox drone are flexibly docked to the target replacement tool.

[0015] The system and method for aerial end-tool switching provided in this application improve the efficiency and docking stability of tool switching in aerial operation scenarios through the collaborative design of the operational UAV and the toolbox UAV. On the one hand, the operational UAV directly acquires the positioning information and replacement tool size information of the toolbox UAV in the air, and can determine the target replacement tool and complete the switching without returning to the ground. Compared with the traditional ground-based fixed toolbox solution, it shortens the tool replacement time, avoids the efficiency loss caused by long-distance round trips, and is suitable for high-altitude and long-distance operation scenarios, meeting the needs of multi-tool collaboration in complex tasks. On the other hand, the guide positioning post of the replacement tool and the docking interface of the operational UAV form a shape-fit connection. Combined with the flexible docking design when the toolbox UAV and the operational UAV come into contact, it effectively compensates for possible attitude deviations and position errors during docking, reduces the risk of equipment collisions caused by hard contact, and ensures the connection reliability during end-tool switching. This solves the problems of limited single-tool functionality and low switching efficiency of traditional aerial operation robots, and provides strong support for continuous and efficient operations in high-altitude and high-risk fields. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of a system embodiment 1 for implementing over-the-air end-tool switching provided in this application;

[0017] Figure 2 The flowchart is a second embodiment of the method for implementing over-the-air terminal tool switching provided in this application.

[0018] Explanation of reference numerals in the attached figures:

[0019] 1: Operational drones;

[0020] 2: Toolbox Drone. Detailed Implementation

[0021] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0022] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0023] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0024] The following specific embodiments are given to illustrate the technical solution of this application in detail.

[0025] Figure 1 This is a schematic diagram of a system embodiment for implementing over-the-air end-point tool handover provided in this application. Please refer to... Figure 1 The system provided in this embodiment includes:

[0026] The operation drone 1 is equipped with an end-effector tool. The operation drone 1 is configured to acquire the positioning information of the toolbox drone 2 below and the size information of multiple replacement tools on the toolbox drone 2 when it is in the air, determine the target replacement tool based on the size information, and make contact with the target replacement tool based on the positioning information to realize the end-effector tool switching.

[0027] The toolbox drone 2 has multiple replacement tools on its top plane, and a guide positioning post is provided on the top of each replacement tool. The bottom of the work drone 1 has a docking interface, which is shaped and connected to the guide positioning post. When the work drone 1 and the toolbox drone 2 come into contact to replace the end tool, the body of the toolbox drone 2 and the body of the work drone 1 are flexibly docked.

[0028] Optionally, the drone is equipped with a robotic arm at its bottom, and a docking interface is installed at the bottom of the robotic arm. The replacement tool is equipped with a guide positioning post, the top of which is a conical boss, and the docking interface is a downward-opening conical groove. The shape of the guide positioning post matches the docking interface. An electromagnet is provided at the top of the guide positioning post, and a magnetizing part is provided at the bottom of the docking interface. The magnetizing part and the electromagnet are magnetically attracted and connected.

[0029] For details, please refer to Figure 2 The work drone is positioned above the toolbox drone. A robotic arm is mounted on the bottom of the work drone, with a downward-opening docking interface at its base. The inner wall of the docking interface has a conical groove structure. This interface is used to dock with replacement tools on the toolbox drone to form an end effector. Multiple replacement tools are arranged in an array on the top of the toolbox drone. Each replacement tool has an integrally formed guide positioning post at its top. The guide positioning post has a conical boss structure, with the top diameter smaller than the bottom diameter. The outer taper of the conical boss matches the inner taper of the docking interface on the work drone, creating a shape-fitting structure between the conical boss and the conical groove. This allows for correction of slight positional deviations during docking via conical guidance. Furthermore, the toolbox drone integrates a compliant control module to adjust its attitude during docking.

[0030] Furthermore, it can be understood that the upper parts of the multiple replacement tools on the toolbox drone all share the same docking structure with guide positioning posts, while the lower parts are each equipped with the corresponding tool. Each replacement tool has an independent power supply, controller, and WiFi module, enabling real-time communication with a host computer via WiFi. It can also remotely control the electromagnet and tool via WiFi. Moreover, the size, model, and material of each replacement tool on the toolbox drone are pre-determined based on information provided by the manufacturer and stored within the toolbox drone for immediate use.

[0031] Optionally, the operational drone adopts a quadcopter design with a wheelbase of 0.65 meters, offering enhanced maneuverability and flexibility. Equipped with a high-precision Delta robotic arm, the operational drone achieves translational motion of the end effector in three-dimensional space through precise control of three servos. To ensure accurate positioning during collaboration, the operational drone is also equipped with a high-resolution camera system, enabling millimeter-level relative positioning accuracy by recognizing a QR code mounted on the toolbox drone. For the installation and release of end-effector tools, a circular metal plate is included at the end of the robotic arm, which can combine with an electromagnet to generate attraction. The toolbox drone, serving as the system's mobile tool library, employs a hexacopter design with a wheelbase of 1.02 meters, ensuring flight stability and load capacity. Its core component is a specially designed modular toolbox equipped with four circular tool mounting holes, each 0.12 meters in diameter, accommodating a variety of end-effector tools.

[0032] Furthermore, during system operation, the toolbox drone flies to the designated location according to the task planning results and hovers. The operation drone acquires real-time images of the toolbox drone through a visual sensor, first identifying the positions of each replacement tool on the top plane and calculating the toolbox drone's positioning information. Then, it extracts the contours of each guide positioning post through image edge detection, calculates the actual size of the replacement tool, and determines the corresponding target replacement tool based on the task requirements. Subsequently, the operation drone moves towards the target replacement tool based on the positioning information. As the interface gradually approaches the guide positioning post, the two achieve initial alignment through the shape matching of the conical surfaces. After initial alignment, the toolbox drone's compliant control module is activated, compensating for docking deviations by fine-tuning the body attitude, achieving flexible docking between the operation drone and the toolbox drone, and finally completing the end-effector tool switching.

[0033] Furthermore, the operational drone first fuses the toolbox drone positioning information collected by the visual sensor with the pre-stored relative position of the target replacement tool in the toolbox drone's body coordinate system. Combining this with the rotation matrix from the toolbox drone's body coordinate system to the world coordinate system, the docking point coordinates at the top of the target replacement tool's guide positioning post are calculated using coordinate transformation formulas, thus clarifying the path endpoint. A smooth path combining straight lines and circular transitions is adopted. A speed segmentation threshold is set based on the initial distance between the operational drone and the docking point. The operational drone, considering its current position relative to the docking point, generates a path using the A* algorithm (A-Star algorithm). The path contains multiple path nodes, and a collision detection algorithm verifies whether each node meets obstacle avoidance constraints. If a conflict exists, the node coordinates are adjusted (e.g., offset by 0.3m along a direction perpendicular to the obstacle). Finally, an executable preliminary docking path is output, ensuring that the operational drone smoothly approaches the target tool along the preset trajectory.

[0034] Furthermore, the compliant mode is an active attitude adjustment mode activated by the toolbox drone to compensate for positional deviations and attitude tilts during docking and to reduce hard contact impacts. Through closed-loop control of dynamic attitude correction, the toolbox drone has the ability to flexibly follow the operating drone, ensuring that the guide positioning post and the docking interface always maintain a close fit, avoiding docking failure or equipment damage caused by rigid collisions. When the work drone approaches the target replacement tool along the initial path and meets any of the following conditions (1. the relative distance between the two is less than a threshold; 2. the work drone enters a smooth docking and maintains a low and stable speed; 3. the binocular vision sensor detects that the relative attitude tilt angle between the work drone and the toolbox drone is greater than a threshold), the toolbox drone automatically triggers the compliant mode. After the compliant mode is activated, the toolbox drone's UWB positioning sensor and binocular vision sensor collect the relative position deviation and attitude deviation in real time. Based on the size information of the target replacement tool, the matching compliant control parameters are retrieved from the pre-stored size-parameter mapping table. Based on the collected deviation value and the matching compliant control parameters, the vertical compensation force is calculated and converted into the attitude adjustment command of the toolbox drone. Based on the attitude adjustment command, the toolbox drone is controlled to flexibly dock.

[0035] Furthermore, the docking speed of the operation drone is determined based on the relative distance between the operation drone and the toolbox drone, wherein the smaller the relative distance, the slower the docking speed.

[0036] Specifically, when the docking interface of the operational drone is not in contact with the guide positioning post of the toolbox drone, the operational drone uses a visual sensor to monitor the relative vertical distance to the guide positioning post on the target replacement tool in real time. A segmented speed control strategy is employed: when the relative distance is greater than a first distance, the operational drone docks at a first speed; when the relative distance is greater than a second distance but less than or equal to the first distance, the operational drone docks at a second speed; and when the relative distance is less than or equal to the second distance, the operational drone docks at a third speed. It can be understood that the first distance > the second distance, the first speed > the second speed > the third speed, and each value is specifically set according to actual needs. For example, in one embodiment, when the relative distance is greater than two meters, the horizontal speed Δy is corrected at 50-80 cm / s, and the vertical speed is maintained at a uniform descent of 30-50 cm / s. The distance is quickly reduced through coarse adjustment. When the relative distance is greater than 50 cm and less than or equal to two meters, the vertical speed is reduced to 10-20 cm / s, and the attitude fine adjustment is initiated to ensure that the deviation between the centerline of the interface and the centerline of the guide positioning post is ≤5°. When the relative distance is less than or equal to 50 cm, the vertical speed is reduced to 3-5 cm / s to prepare for the force buffer at the moment of contact. At this time, the compliant control algorithm has been preloaded with parameters and is waiting to be triggered.

[0037] Furthermore, at the instant of contact between the interface and the guide positioning post, an initial collision occurs between the conical surfaces of the interface and the guide positioning post. The impact force is calculated based on the contact stiffness and relative velocity. When the acceleration of the operating UAV is 0, the docking force is calculated based on flexible docking control. The impact is suppressed by a damping term, and the position is corrected by an elastic term. After the interface and the guide positioning post are fully matched and docked, there is no relative motion between them, and the docking force stabilizes as a static adhesion force, which can be calculated based on the electromagnet attraction force and the elastic preload. The docking force is ultimately converted into the attitude angle adjustment of the toolbox UAV and the motor output tool. By dividing the docking force into horizontal and vertical components, the horizontal component is converted into roll and pitch angles using the force-attitude mapping formula, and the vertical component is converted into lift compensation. Lift compensation is achieved by adjusting the total throttle output of the toolbox UAV.

[0038] Furthermore, in addition to the aforementioned operational drone and toolbox drone, the system also includes a ground control station. The ground control station is communicatively connected to the operational drone and the toolbox drone. The ground control station receives the positioning information and the size information of multiple replacement tools from the toolbox drone, generates a docking path instruction based on the positioning information, generates a target replacement tool docking instruction based on the size information, and sends the target replacement tool docking instruction to the operational drone.

[0039] Specifically, the ground control station, as the system's command center and information interaction core, establishes a two-way data connection with the operational drone and toolbox drone via wireless communication links (such as WiFi, 4G / 5G, or dedicated drone communication frequencies). This connection integrates data from multiple devices, generates precise control commands, and ensures the orderly and reliable tool switching process. The ground control station receives two types of key data in real time: first, the toolbox drone's own positioning information (collected by the toolbox drone's UWB positioning sensor, RTK module, or visual positioning system, including parameters such as latitude, longitude, altitude, and attitude angle, with centimeter-level accuracy), and the dimensional information of multiple replacement tools (pre-stored through offline measurement and real-time image calibration by the toolbox drone, covering key dimensional parameters such as the diameter of the guide positioning post, the length and width of the tool base, and the overall height); second, the operational drone's own status data (such as current location, remaining battery power, end-effector working status, and real-time images collected by the visual sensor), used to determine whether the operational drone is ready for docking.

[0040] Furthermore, for the docking path command, the positioning information of the toolbox UAV is first converted into coordinate values ​​in the world coordinate system. Then, combined with the current position of the operation UAV, the optimal flight path from the operation UAV to directly above the target replacement tool on the toolbox UAV is generated through the path planning algorithm. The path will avoid the preset obstacle area and set multiple speed control points (such as the speed thresholds for the long-range approach segment and the short-range fine-tuning segment). For the target replacement tool docking command, the received target replacement tool size information is compared with the preset task requirement size range (such as the diameter range of the guide positioning post for the end tool to the interface adaptation, and the tool base size threshold corresponding to the work load). The target replacement tool that meets the requirements is selected, and the command content containing the target replacement tool number, the relative position coordinates on the toolbox UAV, and the attitude adjustment reference value during docking is generated. The ground control station sends the generated docking path instructions and target replacement tool docking instructions to the operation drone. After receiving the instructions, the operation drone adjusts its flight trajectory according to the docking path instructions, and at the same time, drives the robotic arm to approach the target replacement tool according to the relative position coordinates in the target replacement tool docking instructions. During this process, the operation drone and the toolbox drone will continuously transmit docking progress data (such as relative distance, attitude deviation, contact status, etc.) back to the ground control station. The ground control station displays these data in real time through a visual interface. If problems such as path deviation or size mismatch occur, emergency correction instructions can be generated and sent to the corresponding drone for adjustment.

[0041] Furthermore, the steps for flexibly docking the toolbox drone fuselage and the operational drone fuselage include:

[0042] (1) When the end tool is not in contact with the target replacement tool, the compliant docking mode is activated according to the real-time flight speed of the operation drone;

[0043] Specifically, when the end-effector tool is not in contact with the target replacement tool, the flight speed of the operation drone is acquired in real time. It can be understood that the flight speed of the operation drone during docking is preset according to the task requirements. When the real-time flight speed reaches the preset speed, it indicates that the distance between the end-effector tool and the target replacement tool is relatively close. At this time, the compliant docking mode is activated.

[0044] (2) The toolbox UAV determines compliance control parameters based on the size information of the target replacement tool, and the compliance control parameters include the elastic coefficient and damping coefficient of the target replacement tool;

[0045] Specifically, the toolbox drone has a pre-stored mapping table of the size information of each replacement tool and the compliance control parameters. The elastic coefficient is positively correlated with the diameter of the guide positioning post of the target replacement tool, and the damping coefficient is positively correlated with the tool base area. Based on the diameter and base area in the size information, the corresponding elastic coefficient and damping coefficient can be directly looked up from the mapping table.

[0046] (3) The toolbox UAV acquires the relative position deviation and attitude tilt angle of the operation UAV, and calculates the compensation force based on the elastic coefficient and damping coefficient;

[0047] Specifically, the relative position deviation is collected by the UWB positioning module on the toolbox drone, where Δx and Δy are the horizontal deviations (left and right, forward and backward), and Δz is the vertical deviation. The attitude tilt angle is collected by the binocular vision sensor of the toolbox drone, where Δα is the roll direction deviation (left and right tilt angle of the fuselage), and Δβ is the pitch direction deviation (forward / backward tilt angle of the fuselage).

[0048] Furthermore, since the distance between the docking points of the two machines is fixed, the attitude tilt angle can be converted into a vertical position deviation compensation amount through trigonometric function conversion, thus obtaining the total vertical deviation. The vertical compensation force is then calculated based on the elastic coefficient, damping coefficient, and total vertical deviation, and used as the compensation force.

[0049] Furthermore, the vertical deviation compensation amount can be calculated using the following formula:

[0050] ;

[0051] Where L is the vertical projection deviation caused by attitude tilt;

[0052] This refers to the deviation in the roll direction;

[0053] This refers to the pitch deviation.

[0054] Furthermore, the compensating force can be calculated using the following formula:

[0055] ;

[0056] in, The elastic coefficient;

[0057] The damping coefficient;

[0058] This refers to the relative vertical deviation;

[0059] This is the amount of compensation for deviation in the vertical direction;

[0060] The vertical relative speed between the two machines.

[0061] (4) Adjust the hovering attitude of the toolbox UAV according to the compensation force, wherein the direction of the compensation force is opposite to the flight direction of the operation UAV.

[0062] Specifically, the compensation force is converted into an altitude adjustment amount using a force-altitude conversion formula. The toolbox UAV flight control system then adds the altitude adjustment amount to the current altitude command and adjusts the total throttle output to achieve altitude adjustment.

[0063] By monitoring the flight speed of the operational drone in real time and initiating a compliant docking mode, buffer control is activated in advance during the safe phase when the end-effector tool approaches but does not make contact with the target replacement tool, avoiding equipment collision damage caused by high-speed impacts. The design of the vertical compensation force being opposite to the flight direction of the operational drone can offset the approaching impact force through the reverse force. Based on the size of the target replacement tool, the appropriate elastic coefficient and damping coefficient are retrieved from a pre-stored mapping table to solve the problem of insufficient adaptability of a single parameter. At the same time, UWB positioning and binocular vision are used to collect relative position deviations and attitude tilt angles, and the attitude deviations are converted into vertical compensation amounts through trigonometric functions to ensure docking coaxiality accuracy. The toolbox drone actively converts the compensation force into throttle adjustment commands through force-altitude conversion to adjust the hovering attitude, rather than relying on the operational drone for unilateral adjustment. This reduces the control load of the operational drone and enables two-drone collaboration, ultimately significantly improving the success rate of aerial tool switching in high-altitude environments, reducing equipment wear and tear, and providing reliable support for drone collaborative operations.

[0064] Furthermore, the system also includes:

[0065] (1) Determine the first position vector of the operation UAV and the second position vector of the target replacement tool in the world coordinate system;

[0066] Specifically, the world coordinate system adopts a right-handed Cartesian coordinate system, with the starting point of the task as the origin. The X-axis points horizontally to the task area, the Y-axis is perpendicular to the X-axis and parallel to the ground, and the Z-axis is perpendicular to the ground and pointing upwards. The task drone is equipped with an RTK (Real-Time Kinematic) module and an IMU (Inertial Measurement Unit). The RTK module acquires the drone's coordinates in the world coordinate system in real time through satellite signals, while the IMU collects the drone's acceleration and angular velocity data. The system fuses the two types of data using a Kalman filter algorithm to output the drone's first position vector.

[0067] Furthermore, the toolbox drone obtains its own center of gravity position vector in the world coordinate system through the UWB positioning sensor installed on the outer casing; secondly, since the target replacement tool is pre-fixed in the modular toolbox on the top of the toolbox drone, its relative position in the toolbox drone's body coordinate system has been pre-stored in the system through offline measurement; finally, based on the rotation matrix from the toolbox drone's body coordinate system to the world coordinate system (calculated from the attitude angles collected by the toolbox drone's gyroscope and magnetometer), the second position vector of the target replacement tool in the world coordinate system is calculated through the coordinate transformation formula.

[0068] (2) Calculate the relative position deviation vector based on the first position vector and the second position vector;

[0069] Specifically, the difference between the first position vector and the second position vector is used as the relative position deviation vector.

[0070] (3) Obtain the rate of change of the relative position deviation vector in real time, and calculate the compliant enhancement control law of the toolbox UAV based on the elastic coefficient, the damping coefficient and the rate of change;

[0071] Specifically, the rate of change of the relative position deviation vector, i.e., the derivative of the deviation vector with respect to time, reflects the dynamic trend of the deviation. The following formula can be used to calculate the compliant enhancement control law for the UAV:

[0072] ;

[0073] in, The desired location for the toolbox drone.

[0074] This is the relative position deviation vector;

[0075] This represents the rate of change of the relative position deviation vector;

[0076] , , The mass, damping, and spring constant of the ideal mass-damped-spring system are respectively.

[0077] This is the acceleration vector at the end of the robotic arm of the operational drone.

[0078] The above formula yields the result. The desired speed of the drone below can be obtained by integration. and desired location .

[0079] Furthermore, after obtaining the compliant augmentation control law of the toolbox UAV, the desired position of the toolbox UAV output by the compliant augmentation control law is used as the target position, and the toolbox UAV is controlled to fly towards the target position to achieve compliant docking.

[0080] Furthermore, obtain the location information of the drone in the toolbox below, including:

[0081] (1) Acquire real-time images of the toolbox drone;

[0082] (2) Identify the positional features in the real-time image;

[0083] Specifically, the operational drone is equipped with a camera that captures real-time images of the toolbox drone. A circular visual marker is placed at the end of the operational drone's robotic arm, and a square positioning frame is placed on the top of the replacement tool. Digital templates of these two types of features are pre-stored in the recognition algorithm library.

[0084] Furthermore, a target detection algorithm is used for real-time target detection. The anchor frame size is optimized for aerial scenes (to adapt the imaging size of the marker at different distances). Combined with the SIFT feature point matching algorithm, the pixel coordinates of the circular marker and the square positioning box are located in the image as position features.

[0085] (3) Calculate the relative position and relative attitude angle between the operating UAV and the replacement tool based on the positional characteristics;

[0086] Specifically, based on the triangulation principle of binocular vision, the system utilizes the parallax of feature points collected by the left and right cameras, combined with the camera baseline distance and focal length to calculate the three-dimensional relative position. By fitting the rotation state in three-dimensional space through the pixel coordinates of the four vertices of the square positioning box in the image, the PnP algorithm is used to calculate the roll angle, pitch angle, and yaw angle of the operating UAV relative to the replacement tool, which are used as the relative attitude angles.

[0087] (4) Generate motion control commands for the operation drone based on the relative position, and generate adjustment commands for the replacement tool based on the relative attitude angle.

[0088] Specifically, the position of the replacement tool within the body coordinate system is represented as follows: (Measured by the camera), the position of the operating drone in the inertial coordinate system is represented as follows: (Indoor measurements were taken by Vicon, and outdoor measurements were taken by RTK). Let be the rotation matrix from the drone's body coordinate system to the world coordinate system. For the positions of the drone's end effector and the replacement tool in the world coordinate system, we have the following expression:

[0089] ;

[0090] ;

[0091] From the expression above, we can obtain:

[0092] ;

[0093] ;

[0094] in, To replace the position of the tool within the machine coordinate system;

[0095] The position of the terminal of the operational drone in the world coordinate system;

[0096] To replace the tool's position in the world coordinate system;

[0097] It is a rotation matrix;

[0098] The position of the operating drone in the inertial coordinate system;

[0099] The position of the end effector of the robotic arm of the operational drone in the body coordinate system.

[0100] make .make Therefore, the automatic docking control law is set to... , Under the aforementioned control law, the system can achieve stability, and the error eventually converges. The desired location for the operation drone. The position of the robotic arm end effector of the operational drone in the body coordinate system. The position of the target in the machine coordinate system, It is a constant greater than 0, used to control the gain. To determine the position of the operational drone in the inertial coordinate system, The desired location for the operation drone. It is a constant greater than 0, used to control the gain.

[0101] Furthermore, the system provided in this embodiment also includes:

[0102] (1) Acquire real-time images of the toolbox drone;

[0103] Specifically, based on the description above, it will not be repeated here.

[0104] (2) Extract the edge contour of the replacement tool in the real-time image based on the edge detection algorithm, and generate a dynamic tracking box;

[0105] Specifically, the real-time image is converted to grayscale (to reduce computational load), Gaussian filtered to remove noise, and histogram equalization is used to enhance edge contrast. The Canny operator is used to extract image edges, and combined with the prior shape features of the replacement tool (such as a square base or cylindrical pillar), Hough transform is used to filter out straight / curve segments that conform to the tool's contour. The filtered edge contours are then fitted with a minimum bounding rectangle to generate a dynamic tracking box containing the complete contour of the replacement tool, and a unique ID is assigned to the tracking box for continuous frame matching. This process eliminates background interference (such as the sky, clouds, and ground objects), retaining only the effective contour of the replacement tool, providing a clear tracking target for subsequent displacement calculations.

[0106] (3) Determine the displacement change of the dynamic tracking box in the continuous frame real-time image, and calculate the relative linear velocity between the operation drone and the replacement tool based on the displacement change;

[0107] Specifically, for the dynamic tracking box in two consecutive frames, the pixel displacement of its center coordinates is calculated. Combined with the camera intrinsic parameters and the distance from the replacement tool to the camera, the pixel displacement is converted into the actual physical displacement. The relative linear velocity is then calculated based on the ratio of the actual physical displacement to time.

[0108] Furthermore, if, while tracking the displacement changes of real-time images, the original tracking box ID is not detected for N consecutive frames, or the matching degree between the tracking box and the prior shape of the replacement tool is less than a threshold, it is determined that the dynamic tracking box is lost. Based on the displacement changes of the previous M frames, a motion trajectory is fitted (using quadratic polynomial prediction) to generate a predicted tracking box position. The template matching algorithm is then activated to search for areas within the predicted region that match the edge template (pre-stored high-definition outline) of the replacement tool. If the matching degree is greater than the threshold, the tracking box is regenerated and the calculation continues. In addition, if the dynamic tracking box is lost for too long, a wide-angle camera is triggered to assist in shooting, expanding the search range. Combining the GPS position of the operating drone with the relative position of the toolbox drone, the search area is narrowed. If tracking is still not restored, docking is stopped.

[0109] (4) Generate a speed closed-loop control command based on the relative linear velocity, and control the operation drone based on the speed closed-loop control command.

[0110] Specifically, the deviation between the target relative linear velocity and the actual relative linear velocity is used as input, and a PID controller is used to calculate the speed adjustment of the operation drone as a speed closed-loop control command; this ensures that the operation drone approaches the replacement tool at a stable speed, avoiding collisions or docking deviations caused by excessive relative speed.

[0111] Furthermore, the implementation steps for determining the target replacement tool include:

[0112] (1) Acquire real-time images of the toolbox drone and identify the tool area of ​​each replacement tool in the real-time image based on the target detection algorithm;

[0113] Specifically, the acquired real-time images are first preprocessed, then input into the target detection algorithm, which outputs all detected tool regions. Low-confidence targets are filtered out using a confidence threshold, retaining only valid tool regions. This solves the occlusion problem caused by densely packed tools, ensuring that all replacement tools are fully identified, and providing accurate region localization for subsequent contour extraction and size calculation.

[0114] (2) Perform edge detection on each tool region and extract the contour features of each tool region;

[0115] Specifically, the images of each identified tool region are cropped to obtain images containing only a single tool region; the cropped tool region images are then converted to grayscale; a suitable edge detection algorithm is selected to perform edge detection on the grayscale tool region images to obtain the edge information of the tools; the detected edge information is then denoised and connected to eliminate edge breaks and false edges; the tool contour features formed by the processed edge information are extracted, including parameters such as the perimeter, area, number of inflection points, and curvature changes of the contour; the extracted contour features are then standardized to eliminate the influence of factors such as image size and scaling ratio on the contour feature parameters.

[0116] Furthermore, the original real-time image is cropped, generating an independent tool region image for each tool region. During the cropping process, it is ensured that the tool region is completely contained within the cropped image, and the distance between the image edge and the tool edge is not less than a preset number of pixels, avoiding the cropping of critical edge information of the tool. Image preprocessing is performed on the image of a single tool region, and contour feature extraction is further performed on the preprocessed image. For the steps of contour feature extraction, please refer to the description in relevant technologies, which will not be repeated here.

[0117] (3) Determine the actual size of each replacement tool based on the contour features, and determine the target replacement tool according to the actual size.

[0118] Specifically, a mapping relationship between pixel size and actual size is established using camera calibration parameters. The pixel size of each contour feature is calculated based on the contour features and the pixel size, and the actual size of that contour feature is obtained further according to the mapping relationship. By comparing the parameter information in the replacement tools in the task, the replacement tool whose actual size matches the parameter information is identified as the target replacement tool.

[0119] In addition, if there are multiple contour fusions, the gradient direction of each edge pixel is calculated within the fusion region, and K-means clustering is performed on the gradient direction to divide the pixels into two groups of candidate edges. For each group of candidate edges, the gradient direction difference and Euclidean distance between adjacent pixels are calculated. Edge segments with a continuity score greater than the score threshold are retained, and discrete noise pixels are removed. At the fusion point of the two groups of edges, the inflection point is detected by the second derivative of the pixel gray value, and the fused contour is divided into two independent sub-contours.

[0120] For example, when two cylindrical tools are placed side by side, causing edge merging, gradient direction clustering can distinguish the tangent directions of the two arcs (which differ by 180°). Combined with inflection point detection, the merging point can be accurately segmented, ensuring that the segmented sub-contour matches the edge of the single tool with ≥90%. In this way, the gradient features of adjacent pixels are used to achieve local separation of the merging contour, solving the edge adhesion problem caused by physical contact or imaging interference, and providing preliminary segmentation results for subsequent trajectory prediction.

[0121] Furthermore, for the tool region in M ​​consecutive frames of images, the contour features of each tool are extracted and stored, the motion trajectory and shape change trend of each tool are established, and the matching degree between the center coordinates of the sub-contours obtained by segmentation in the current frame and the trajectory of each tool in the historical database is calculated. If the matching error between a sub-contour and the historical tool trajectory is less than the error threshold, it is assigned to that tool. The shape similarity is calculated by comparing the sub-contour with the historical shape features (such as aspect ratio, convex hull defects) of the matching tool (≥80% is considered consistent). If they are inconsistent, re-segmentation is triggered.

[0122] Furthermore, weights are assigned to pixel feature analysis and contour trajectory analysis respectively. The final confidence score is calculated based on the weighted sum of the two, and the final segmented sub-contour is output, updating the tool contour library for the current frame. The segmentation error (compared to the contours of subsequent unfused frames) is fed back to the pixel feature analysis module to dynamically optimize the clustering threshold and inflection point detection parameters. By combining pixel features in the spatial domain with trajectory patterns in the temporal domain, a complementary verification decision mechanism is formed, effectively solving the contour fusion problem in complex scenes and ensuring the reliability of tool recognition and size calculation.

[0123] Furthermore, the coordinate positions of the operation drone and the toolbox drone are determined separately. The toolbox drone retrieves the installation position parameters of the target tool in the toolbox drone's body coordinate system, as well as the attitude deflection angle during tool installation. The coordinates of the docking point in the local coordinate system of the replacement tool are calculated using the coordinate transformation formula to obtain the coordinates of the docking point in the toolbox drone's body coordinate system. The coordinates of the docking point in the toolbox drone's body coordinate system are substituted into the world coordinate transformation formula to calculate the coordinates of the docking point in the world coordinate system, which serves as the endpoint of the path planning.

[0124] In addition, the system provided in this embodiment also includes:

[0125] When the operating drone docks with the target replacement tool, the target replacement tool in the toolbox drone is powered on;

[0126] After the operating drone completes its previous operation and before docking with the target replacement tool for the next operation, a power-off command is sent to the original end-effector to disconnect the original end-effector from the operating drone.

[0127] Specifically, when the operating UAV visually locates the target replacement tool (e.g., by recognizing the tool's unique ID code) and enters the docking preparation phase, the ground control station or the toolbox UAV flight control system sends a power-on command to the independent power supply module of the target replacement tool, activating only the electromagnet power supply of that tool. The modular power supply unit of the toolbox UAV adopts a one-tool-one-circuit design, with each replacement tool corresponding to an independent relay switch. Simultaneously with the target tool's power-on, the system sends a power-off command to the relays of all non-target tools, cutting off their electromagnets and controller power, ensuring that non-target tools have no magnetic attraction capability. Furthermore, current sensors can collect the current values ​​of each tool's power supply circuit in real time. If the current in a non-target tool's circuit is >0, an alarm is immediately triggered and the power supply module is restarted, forcibly cutting off the abnormal circuit and preventing accidental power supply due to relay failure. This completely eliminates the magnetic attraction capability of the non-target tool's electromagnet, avoiding additional attraction caused by other charged tools during docking, and preventing problems such as misalignment, collision damage, etc.

[0128] Optionally, if the relative vertical distance between the interface and the guide positioning post of the target replacement tool is less than a threshold and there is a deviation in the centerline, the operating UAV generates and sends a docking signal to the toolbox UAV and the ground control station. The ground control station sends a target tool identity verification command to the toolbox UAV. The toolbox UAV verifies the target tool by comparing the tool hardware ID. After successful verification, it sends an identity verification pass signal back to the ground control station. The ground control station generates and issues a target tool power-on command. The toolbox UAV receives the target tool power-on command, controls the target tool relay to close and power on, and simultaneously sends a disconnect command to the non-target tool relay.

[0129] Furthermore, when the operational scenario requires replacement tools to remain on standby and all replacement tools to maintain power supply, a magnetic shield is installed on the outside of the electromagnet of each replacement tool to avoid magnetic field interference between adjacent tools. An isolation transformer and a common-mode inductor are used to achieve electrical isolation of the power supply circuit, and double-layer shielding is used for communication cables to reduce electromagnetic interference. On the software side, authentication commands ensure that only the target tool responds and connects. The current of the target tool's electromagnet is dynamically adjusted based on the relative distance, and a Hall sensor monitors the magnetic field in real time. In case of an anomaly, the current of non-target tools is reduced, and the current of the target tool is compensated. This approach ensures continuous power supply to all tools while reducing the false engagement rate, shortening tool switching time, and meeting the requirements for rapid response.

[0130] Furthermore, the system automatically or manually selects the power supply path based on the type of task. For routine tasks, the default path is path one, which aims for low power consumption and high safety. For emergency tasks, the system switches to path two to enable rapid tool response. When switching paths, the system first cuts off the power supply to all tools within 100ms, and then reconfigures the power supply status according to the target path to avoid power supply conflicts or interference during the switching process, ensuring the adaptability and stability of tool power supply control in different scenarios.

[0131] Optionally, after the drone detects that the operation is completed and hovers directly above the original tool placement hole, it generates and sends an original tool retrieval ready signal to the ground control station; the ground control station generates an original tool power-off command and sends it to the drone; the drone receives the original tool power-off command and controls the robotic arm to cut off the power supply to the original tool, while simultaneously detecting the load force of the robotic arm, and determines whether separation is complete based on the load force of the robotic arm.

[0132] Furthermore, before the drone completes its previous task and docks with the next target tool, it flies to the original tool's recovery area, hovers, and sends a request to recover the original tool. After confirming the location, the system sends a power-off command to the original tool's power supply interface, cutting off the power to the magnetic or mechanical latch. The original tool is recovered under the action of gravity or mechanical unlocking. The load force is then detected by the force sensor at the end of the robotic arm. Once the load force is confirmed to have dropped to the no-load threshold, the original tool is disconnected. If it gets stuck, a second power-off is performed, and the device is shaken to assist in separation, ensuring that the original tool is completely detached and making room for the next docking.

[0133] The system for aerial end-tool switching provided in this embodiment, on the one hand, allows the operational drone to directly acquire the toolbox drone's positioning information and the replacement tool's size information in the air. This enables target tool identification and switching without returning to the ground, significantly shortening tool replacement time compared to ground-based fixed toolbox solutions. It avoids efficiency losses and safety risks associated with long-distance round trips, adapting to high-altitude, long-distance operation scenarios and meeting the multi-tool collaboration needs in complex tasks such as wind turbine maintenance and power line inspection, reducing the cost and scheduling risks of multi-device collaborative operations. On the other hand, the conical boss guide positioning post of the replacement tool and the conical groove of the operational drone form a shape fit at the interface. Combined with the compliant docking control of the toolbox drone, and based on the operational drone's speed start mode, the system matches the elasticity and damping coefficients according to the tool size. By acquiring the deviation and calculating the reverse compensation force to adjust the attitude, it effectively compensates for docking deviations and reduces hard contact losses. Simultaneously, a power supply scheme combining directional power supply to the target tool with power-off for non-target tools or full tool power supply with magnetic shielding and dynamic current adjustment prevents accidental adsorption and ensures reliable docking. This improves the success rate and efficiency of aerial tool switching, providing support for continuous and efficient operations in high-altitude, high-risk areas.

[0134] Corresponding to the aforementioned embodiment of a system for implementing in-flight end-tool handover, this application also provides an embodiment of a method for implementing in-flight end-tool handover.

[0135] Figure 2 This is a flowchart of Embodiment 2 of the method for implementing over-the-air terminal tool handover provided in this application. Please refer to... Figure 2 The method provided in this embodiment includes:

[0136] S201. Obtain the positioning information of the toolbox drone and the size information of multiple replacement tools;

[0137] S202. Determine the target replacement tool based on the size information;

[0138] S203. Determine the relative positions between the operation drone and the toolbox drone based on the positioning information;

[0139] S204. Based on the relative position, control the operation drone and the toolbox drone to flexibly dock the target replacement tool.

[0140] The method in this embodiment can be used to execute Figure 1 The steps of the system embodiment shown are similar in principle and process, and will not be repeated here.

[0141] The specific implementation process of the functions and roles of each step in the above method can be found in the implementation process of the corresponding steps in the above system, and will not be repeated here.

[0142] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A system for enabling over-the-air end tool switching, the system comprising: The system comprises: A work unmanned aerial vehicle, which is provided with an end tool, and is configured to acquire positioning information of an underlying tool box unmanned aerial vehicle and size information of a plurality of replacement tools on the tool box unmanned aerial vehicle when in the air, determine a target replacement tool based on the size information, and contact the target replacement tool based on the positioning information, so as to realize end tool switching; A tool box unmanned aerial vehicle, which is provided with a plurality of replacement tools on a top plane, and the replacement tools are provided with guiding positioning columns on the top, and the bottom of the work unmanned aerial vehicle is provided with a docking interface, which is connected with the guiding positioning columns in a shape fitting mode, and the tool box unmanned aerial vehicle body and the work unmanned aerial vehicle body are flexibly docked when the work unmanned aerial vehicle and the tool box unmanned aerial vehicle contact to replace the end tool.

2. The system of claim 1, wherein, The system further comprises: A ground control station, which is in communication connection with the work unmanned aerial vehicle and the tool box unmanned aerial vehicle, receives positioning information of the tool box unmanned aerial vehicle and size information of a plurality of replacement tools, generates a docking path instruction according to the positioning information, generates a target replacement tool docking instruction according to the size information, and sends the target replacement tool docking instruction to the work unmanned aerial vehicle.

3. The system of claim 1, wherein, The flexible docking of the tool box unmanned aerial vehicle body and the work unmanned aerial vehicle body comprises: When the end tool and the target replacement tool are not in contact, a flexible docking mode is started according to the real-time flight speed of the work unmanned aerial vehicle; The tool box unmanned aerial vehicle determines flexible control parameters according to the size information of the target replacement tool, and the flexible control parameters comprise an elastic coefficient and a damping coefficient of the target replacement tool; The tool box unmanned aerial vehicle acquires a relative position deviation and an attitude inclination angle of the work unmanned aerial vehicle, calculates a compensation force based on the elastic coefficient and the damping coefficient; The hovering attitude of the tool box unmanned aerial vehicle is adjusted according to the compensation force, and the direction of the compensation force is opposite to the flight direction of the work unmanned aerial vehicle.

4. The system of claim 3, wherein, The system further comprises: A first position vector of the work unmanned aerial vehicle and a second position vector of the target replacement tool in a world coordinate system are determined; A relative position deviation vector is calculated according to the first position vector and the second position vector; The change rate of the relative position deviation vector is acquired in real time, and a flexible enhanced control law of the tool box unmanned aerial vehicle is calculated based on the elastic coefficient, the damping coefficient and the change rate.

5. The system of claim 1, wherein, The acquisition of the positioning information of the underlying tool box unmanned aerial vehicle comprises: Real-time images of the tool box unmanned aerial vehicle are acquired; Position features in the real-time images are recognized; A relative position and a relative attitude angle between the work unmanned aerial vehicle and the replacement tool are calculated according to the position features; Motion control instructions of the work unmanned aerial vehicle are generated according to the relative position, and adjustment instructions of the replacement tool are generated according to the relative attitude angle.

6. The system of claim 1, wherein, The system further comprises: Real-time images of the tool box unmanned aerial vehicle are acquired; Edge contours of the replacement tool in the real-time images are extracted based on an edge detection algorithm, and a dynamic tracking frame is generated; Determine displacement change of the dynamic tracking frame in the continuous frame real-time image, calculate relative linear velocity between the work unmanned aerial vehicle and the replacement tool based on the displacement change; Generate speed closed-loop control instruction according to the relative linear velocity, and control the work unmanned aerial vehicle based on the speed closed-loop control instruction.

7. The system of claim 1, wherein, The target replacement tool is determined based on the size information, comprising: Collect real-time image of the tool box unmanned aerial vehicle, and identify tool area of each replacement tool in the real-time image based on target detection algorithm; Edge detection is performed on each tool area, and contour features of each tool area are extracted; The actual size of each replacement tool is determined based on the contour features, and the target replacement tool is determined according to the actual size.

8. The system of claim 1, wherein, The system further comprises: When the work unmanned aerial vehicle is docked with the target replacement tool, power on the target replacement tool in the tool box unmanned aerial vehicle; After the work unmanned aerial vehicle completes the last work, before docking the target replacement tool of the next work, send power-off instruction to the original end tool, and disconnect the original end tool from the work unmanned aerial vehicle.

9. The system of claim 1, wherein, The work unmanned aerial vehicle is installed with a mechanical arm at the bottom, the mechanical arm is installed with a docking port at the bottom, the replacement tool is installed with a guide positioning column, the top end of the guide positioning column is a conical boss, the docking port is a downward opening conical groove, the shape of the guide positioning column matches the docking port, the top of the guide positioning column is provided with an electromagnet, the bottom of the docking port is provided with a magnetic induction part, and the magnetic induction part and the electromagnet are connected by magnetic force.

10. A method of enabling over-the-air end tool switching, the method being implemented based on the system of claims 1-9, characterized in that, The method comprises: Obtain positioning information of the tool box unmanned aerial vehicle and size information of a plurality of replacement tools; Determine the target replacement tool according to the size information; Determine the relative position between the work unmanned aerial vehicle and the tool box unmanned aerial vehicle according to the positioning information; Control the work unmanned aerial vehicle and the tool box unmanned aerial vehicle to flexibly dock the target replacement tool based on the relative position.

Images