A system and method for enabling over-the-air end tool switching

Abstract

The application provides a system and method for realizing aerial end tool switching, and belongs to the technical field of aerial docking of unmanned aerial vehicles. The system provided by the application comprises: a work unmanned aerial vehicle, the work unmanned aerial vehicle is provided with an end tool, and the work unmanned aerial vehicle is configured to acquire positioning information of a tool box unmanned aerial vehicle below 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, contact the target replacement tool based on the positioning information, and realize end tool switching; a plurality of replacement tools are arranged on a top plane of the tool box unmanned aerial vehicle, 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. The system and method for realizing aerial end tool switching provided by the application can realize flexible docking of unmanned aerial vehicles in the air.

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 work unmanned aerial vehicle is provided with an end tool, and is configured to acquire positioning information of a tool box unmanned aerial vehicle below 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, contact the target replacement tool based on the positioning information, and realize end tool switching.

[0009] The tool box unmanned aerial vehicle is provided with a plurality of replacement tools on a top plane, the top of the replacement tools is provided with a guide positioning column, the bottom of the work unmanned aerial vehicle is provided with a docking interface, the docking interface is connected in shape fitting with the guide positioning column, 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 are in contact to replace the end tool.

[0010] The second aspect of the application provides a method for realizing air end tool switching, the method comprising:

[0011] Acquiring positioning information of the tool box unmanned aerial vehicle and size information of a plurality of replacement tools;

[0012] Determining a target replacement tool according to the size information;

[0013] Determining a relative position between a work unmanned aerial vehicle and a tool box unmanned aerial vehicle according to the positioning information;

[0014] Controlling the work unmanned aerial vehicle and the tool box unmanned aerial vehicle to flexibly dock the target replacement tool based on the relative position.

[0015] The system and method for realizing air end tool switching provided by the application improve the tool switching efficiency and docking stability of an air work scene through the collaborative design of a work unmanned aerial vehicle and a tool box unmanned aerial vehicle. On the one hand, the work unmanned aerial vehicle directly acquires positioning information of the tool box unmanned aerial vehicle and size information of replacement tools in the air, determines a target replacement tool and completes switching without returning to the ground, shortens the tool replacement time, avoids the efficiency loss caused by long-distance round trips, is suitable for high-altitude and long-distance work scenes, and meets the needs of multi-tool cooperation in complex tasks. On the other hand, the guide positioning column of the replacement tool and the docking interface of the work unmanned aerial vehicle form a shape fitting connection, and the flexible docking design of the tool box unmanned aerial vehicle and the work unmanned aerial vehicle when in contact effectively compensates for the attitude deviation and position error that may occur in the docking process, reduces the risk of equipment collision caused by hard contact, ensures the connection reliability during end tool switching, solves the problem of single tool function limitation and low switching efficiency of traditional air work robots, and provides strong support for continuous and efficient work in high-altitude and high-risk fields. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 FIG. 1 is a schematic diagram of a system for switching aerial end tools according to an embodiment of the present application;

[0017] Figure 2 FIG. 2 is a flowchart of a method for switching aerial end tools according to an embodiment of the present application.

[0018] Legend of Signs:

[0019] 1: work unmanned aerial vehicle;

[0020] 2: tool box unmanned aerial vehicle. DETAILED DESCRIPTION

[0021] The exemplary embodiments will be described in detail herein below with reference to the accompanying drawings. In the following description, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application.

[0022] The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in the present application, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0023] It is to be understood that the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0024] The specific embodiments are given below to describe the technical solutions of the present application in detail.

[0025] Figure 1 FIG. 1 is a schematic diagram of a system for switching aerial end tools according to an embodiment of the present application. Please refer to FIG. 1, the system provided by the embodiment of the present application comprises: Figure 1

[0026] ​The work unmanned aerial vehicle 1 is provided with an end tool, and is configured to acquire positioning information of a tool box unmanned aerial vehicle 2 below and size information of a plurality of replacement tools on the tool box unmanned aerial vehicle 2 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.

[0027] The tool box unmanned aerial vehicle 2 is provided with a plurality of replacement tools on a top plane, and the top of each replacement tool is provided with a guide positioning column. The bottom of the work unmanned aerial vehicle 1 is provided with a docking interface, and the docking interface is connected with the guide positioning column in a shape fitting mode. When the work unmanned aerial vehicle 1 and the tool box unmanned aerial vehicle 2 are in contact for replacement of the end tool, the body of the tool box unmanned aerial vehicle 2 and the body of the work unmanned aerial vehicle 1 are in flexible docking.

[0028] Optionally, the work unmanned aerial vehicle is provided with a mechanical arm at the bottom, and the mechanical arm is provided with a docking interface at the bottom. The replacement tool is provided with a guide positioning column, and the top end of the guide positioning column is a conical boss. The docking interface is a downwardly opening conical groove, and the shape of the guide positioning column matches the shape of the docking interface. The top of the guide positioning column is provided with an electromagnet, and the bottom of the docking interface is provided with a magnetic attraction part, which is connected with the electromagnet in a magnetic force adsorption mode.

[0029] Specifically, please refer to Figure 2 The work unmanned aerial vehicle is located above the tool box unmanned aerial vehicle. The work unmanned aerial vehicle is provided with a mechanical arm at the bottom, and the mechanical arm is provided with a downwardly opening docking interface at the bottom. The inner wall of the docking interface is in a conical groove structure, and the docking interface is used for docking with the replacement tool on the tool box unmanned aerial vehicle to form an end tool. The top of the tool box unmanned aerial vehicle is provided with a plurality of replacement tools in an array mode. The top end of each replacement tool is integrally formed with a guide positioning column, and the guide positioning column is in a conical boss structure. The diameter of the top end of the boss is smaller than the diameter of the bottom end, and the taper of the outer side surface of the conical boss matches the taper of the inner side surface of the docking interface of the work unmanned aerial vehicle, so as to form a shape fitting structure of the conical boss and the conical groove, and to correct slight position deviation through taper surface guidance during docking. In addition, the body of the tool box unmanned aerial vehicle is integrated with a compliant control module, which is used for adjusting the attitude of the body during docking.

[0030] Further, it can be understood that the upper half of the plurality of replacement tools on the tool box UAV is the docking structure of the same guide positioning column, and the lower half is respectively installed with the corresponding tool; each replacement tool has an independent power supply, a controller and a WiFi module, and can realize real-time communication with the upper computer through WiFi communication, and can also remotely control the electromagnet and the tool through WiFi. And the size, model and material of each replacement tool on the tool box UAV are determined in advance according to the information provided by the manufacturer, and are stored in the tool box UAV for standby at any time.

[0031] Optionally, the operation UAV adopts a four-rotor design, with an axle distance of 0.65 meters, and has higher mobility and flexibility. The operation UAV is equipped with a high-precision Delta mechanical arm, which can realize the translational motion of the mechanical arm end in three-dimensional space through the precise control of three steering engines. In order to ensure accurate positioning during cooperation, the operation UAV is also equipped with a high-resolution camera system, which can realize millimeter-level relative positioning accuracy by identifying the two-dimensional code installed on the tool box UAV. In order to realize the installation and release of the end tool, a circular iron sheet is contained at one end of the mechanical arm, which can generate suction force in combination with the electromagnet; the tool box UAV as the mobile tool library of the system adopts a six-rotor design, with an axle distance of 1.02 meters, which ensures the stability and load capacity of flight, and its core component is a specially designed modular tool box, which is equipped with four circular tool placing holes with a diameter of 0.12 meters, which can accommodate a variety of end tools.

[0032] Further, when the system is working, the tool box UAV flies to the specified position according to the task planning result and keeps hovering, and the operation UAV collects the real-time image of the tool box UAV through the visual sensor, identifies the position of each replacement tool on the top plane first, calculates the positioning information of the tool box UAV; then the outline of each guide positioning column is extracted through image edge detection, the actual size of the replacement tool is converted, and the corresponding target replacement tool is determined according to the operation task demand. Subsequently, the operation UAV moves to the target replacement tool based on the positioning information, and when the docking port gradually approaches the guide positioning column, the two are preliminarily aligned through the shape cooperation of the conical surface; after preliminary alignment, the compliant control module of the tool box UAV is started, and the docking deviation is compensated by fine-tuning the body attitude, realizing the flexible docking of the operation UAV and the tool box UAV, and finally completing the switching of the end tool.

[0033] Further, the work unmanned vehicle first fuses the tool box unmanned vehicle positioning information collected by the visual sensor with the relative position of the target replacement tool in the tool box unmanned vehicle body coordinate system, combines the rotation matrix from the tool box unmanned vehicle body coordinate system to the world coordinate system, calculates the docking point coordinates of the target replacement tool guide positioning column top through the coordinate transformation formula, and determines the path end point. A smooth path of straight line segments combined with circular arc transitions is adopted, the velocity segmentation threshold is set according to the initial distance between the work unmanned vehicle and the docking point, the work unmanned vehicle generates a path using the A* algorithm (A-Star algorithm) according to the relative position between the current position and the docking point, the path includes multiple path nodes, and each node is checked by a collision detection algorithm to determine whether it meets the obstacle avoidance constraint. If there is a conflict, the node coordinates are adjusted (such as shifting 0.3 m in a direction perpendicular to the obstacle), and finally an executable preliminary docking path is output to ensure that the work unmanned vehicle approaches the target tool along the preset trajectory smoothly.

[0034] Further, the compliant mode is an active attitude adjustment mode started by the tool box unmanned vehicle to compensate for position deviation and attitude tilt during docking and reduce hard contact impact. Through closed-loop control of attitude dynamic correction, the tool box unmanned vehicle has the ability to flexibly follow the work unmanned vehicle, ensuring that the guide positioning column and the docking port always maintain a sticking trend, avoiding docking failure or equipment damage due to rigid collision. When the work unmanned vehicle approaches the target replacement tool along the preliminary path, any of the following conditions is met (1, the relative distance between the two is less than a threshold; 2, the work unmanned vehicle enters stable docking, and the speed remains low and stable; 3, the binocular visual sensor detects that the relative attitude tilt angle between the work unmanned vehicle and the tool box unmanned vehicle is greater than a threshold), the tool box unmanned vehicle automatically triggers the compliant mode. After the compliant mode is started, the relative position deviation and attitude deviation are collected in real time by the UWB positioning sensor and the binocular visual sensor of the tool box unmanned vehicle, the matching compliant control parameters are retrieved from the pre-stored size-parameter mapping table according to the size information of the target replacement tool, and the compensation force in the vertical direction is calculated based on the collected deviation values and the matching compliant control parameters. The compensation force is converted into an attitude adjustment instruction for the tool box unmanned vehicle, and the tool box unmanned vehicle is controlled for flexible docking based on the attitude adjustment instruction.

[0035] Further, the docking speed of the work unmanned vehicle is determined according to the relative distance between the work unmanned vehicle and the tool box unmanned vehicle, wherein the smaller the relative distance, the slower the docking speed.

[0036] Specifically, when the docking interface of the work drone does not contact the guide positioning column of the tool box drone, the work drone monitors the relative vertical distance between the docking interface and the guide positioning column of the target replacement tool in real time through a visual sensor, adopts a segmented speed control strategy, and makes the work drone dock at a first speed when the relative distance is greater than a first distance, makes the work drone dock at a second speed when the relative distance is greater than a second distance and less than or equal to the first distance, and makes the work drone dock at a third speed when the relative distance is less than or equal to the second distance. 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 an embodiment, when the relative distance is greater than two meters, the horizontal speed correction Ay is 50-80 cm / s, the vertical direction maintains a uniform speed of 30-50 cm / s, the distance is quickly reduced through coarse adjustment, when the relative distance is greater than fifty centimeters and less than or equal to two meters, the vertical speed is reduced to 10-20 cm / s, and at the same time, the attitude fine adjustment is started to ensure that the deviation between the central axis of the docking interface and the central axis of the guide positioning column is ≤5°, and when the relative distance is less than or equal to fifty centimeters, 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 soft control algorithm has preloaded parameters, and is waiting for triggering.

[0037] Further, at the moment of contact between the docking interface and the guide positioning column, initial collision occurs between the conical surface of the docking interface and the conical surface of the guide positioning column, the impact force is calculated according to the contact stiffness and the relative speed, the docking force is calculated based on the flexible docking control when the acceleration of the work drone is 0, the impact is suppressed through the damping term, and the position correction is driven through the elastic term. After the docking interface and the guide positioning column are completely matched and docked, there is no relative motion between them, and the docking force is stable as a static fitting force, which can be calculated according to the electromagnetic attraction and the elastic pre-tightening force. The docking force is finally converted into the attitude angle adjustment amount of the tool box drone and the motor output tool. The horizontal component and the vertical component are obtained by resolving the docking force, the horizontal component is converted into the roll angle and the pitch angle through the force-attitude mapping formula, and the vertical component is converted into the lift compensation amount, and the lift compensation is realized by adjusting the total throttle output of the tool box drone.

[0038] Further, in addition to the work drone and the tool box drone described above, the system further comprises a ground control station, the ground control station is in communication connection with the work drone and the tool box drone, the ground control station receives the positioning information of the tool box drone and the 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 drone.

[0039] Specifically, the ground control station serves as the command center and information interaction core of the system. It establishes a two-way data connection with the work unmanned aerial vehicle and the tool box unmanned aerial vehicle through a wireless communication link (such as WiFi, 4G / 5G, or a dedicated unmanned aerial vehicle communication frequency band) to integrate multi-device data and generate precise control instructions, ensuring the orderliness and reliability of the tool switching process. The ground control station receives two types of key data in real time. The first type is the positioning information of the tool box unmanned aerial vehicle (collected by the UWB positioning sensor, RTK module, or visual positioning system carried by the tool box unmanned aerial vehicle, including latitude, longitude, height, attitude angle, etc., with an accuracy of centimeters) and the size information of multiple replacement tools (pre-stored by the tool box unmanned aerial vehicle through offline measurement and uploaded after real-time image calibration, covering key size parameters such as the diameter of the guide positioning column, the length and width of the tool base, and the overall height of the replacement tool). The second type is the state data of the work unmanned aerial vehicle (such as the current position, remaining power, end tool working state, and real-time images collected by the visual sensor), which is used to determine whether the work unmanned aerial vehicle meets the docking conditions.

[0040] Further, for the docking path instruction, the positioning information of the tool box unmanned aerial vehicle is first converted into coordinate values in the world coordinate system, and then combined with the current position of the work unmanned aerial vehicle to generate an optimal flight path for the work unmanned aerial vehicle from the current position to the target replacement tool above the tool box unmanned aerial vehicle through a path planning algorithm. The path will avoid pre-set obstacle areas and set multiple speed control points (such as speed thresholds for the remote approach segment and the near fine-tuning segment). For the target replacement tool docking instruction, the received target replacement tool size information is compared with the pre-set task requirement size range (such as the diameter interval of the guide positioning column compatible with the end tool docking interface and the tool base size threshold corresponding to the work load), and the target replacement tool that meets the requirements is selected to generate an instruction content containing the target replacement tool number, the relative position coordinates on the tool box unmanned aerial vehicle, and the attitude adjustment reference value during docking. The ground control station sends the generated docking path instruction and target replacement tool docking instruction to the work unmanned aerial vehicle. After receiving the instructions, the work unmanned aerial vehicle adjusts the flight trajectory according to the docking path instruction and drives the mechanical arm to approach the target replacement tool based on the relative position coordinates in the target replacement tool docking instruction. In this process, the work unmanned aerial vehicle and the tool box unmanned aerial vehicle continuously transmit docking progress data (such as relative distance, attitude deviation, and whether contact, etc.) to the ground control station. The ground control station displays these data in real time through a visual interface. If there are problems such as path deviation and size mismatch, it can also generate emergency correction instructions and issue them to the corresponding unmanned aerial vehicle for adjustment.

[0041] Further, the implementation steps of the flexible docking of the tool box unmanned aerial vehicle body and the work unmanned aerial vehicle body include:

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

[0043] Specifically, when the end tool is not in contact with the target replacement tool, the flight speed of the work drone is obtained in real time. It can be understood that the flight speed of the work drone during docking is set in advance according to the task requirement. When the real-time flight speed reaches the pre-set speed, it indicates that the distance between the end tool and the target replacement tool is close. At this time, the compliant docking mode is started.

[0044] (2) the toolbox drone determines the compliant control parameters according to the size information of the target replacement tool, and the compliant control parameters include the elastic coefficient and the damping coefficient of the target replacement tool;

[0045] Specifically, the mapping table of the size information and the compliant control parameters of each replacement tool is pre-stored in the toolbox drone. The elastic coefficient is positively correlated with the diameter of the target replacement tool guide positioning column, and the damping coefficient is positively correlated with the tool base area. Based on the diameter and the base area in the size information, the corresponding elastic coefficient and damping coefficient are directly found from the mapping table.

[0046] (3) the toolbox drone obtains the relative position deviation and the attitude tilt angle of the work drone, and calculates the compensation force based on the elastic coefficient and the damping coefficient;

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

[0048] Further, because the distance between the docking points of the two machines is fixed, the attitude tilt angle can be converted into vertical position deviation compensation by trigonometric function, and the total vertical deviation is obtained. The vertical direction compensation force is calculated based on the elastic coefficient, the damping coefficient and the total vertical deviation, as the compensation force.

[0049] Further, the vertical direction deviation compensation can be calculated according to the following formula:

[0050] ;

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

[0052] is the roll direction deviation;

[0053] pitch direction deviation.

[0054] Further, the compensation force can be calculated according to the following formula:

[0055]

[0056] wherein, is an elastic coefficient;

[0057] is a damping coefficient;

[0058] is a relative vertical deviation;

[0059] is a vertical direction deviation compensation;

[0060] is a relative vertical speed between the two machines.

[0061] (4) Adjusting the hovering attitude of the toolbox UAV according to the compensation force, the direction of the compensation force being opposite to the flight direction of the working UAV.

[0062] Specifically, the compensation force is converted into a height adjustment amount through a force-height conversion formula, and the height adjustment amount is superimposed on the current height instruction by the flight control system of the toolbox UAV, so as to realize height adjustment through adjusting the total throttle output.

[0063] By monitoring the flight speed of the working UAV in real time to start the compliant docking mode, it is ensured that the buffer control is activated in advance in the safe stage when the end tool approaches but does not contact the target replacement tool, so as to avoid equipment collision damage caused by high-speed impact; the design that the vertical compensation force is opposite to the flight direction of the working UAV can offset the approaching impact force through the reverse force. Based on the size of the target replacement tool, the adaptive elastic coefficient and damping coefficient are retrieved from the pre-stored mapping table, so as to solve the problem of insufficient adaptability of a single parameter. At the same time, the relative position deviation and attitude inclination angle are collected through UWB positioning and binocular vision, and the attitude deviation is converted into a vertical compensation amount through a trigonometric function, so as to ensure the coaxial accuracy of docking. The toolbox UAV actively converts the compensation force into a throttle adjustment instruction through force-height conversion to adjust the hovering attitude, instead of relying on unilateral adjustment of the working UAV, which not only reduces the control load of the working UAV, but also realizes the cooperation of the two machines, finally significantly improves the success rate of aerial tool switching in high-altitude environment, reduces equipment loss, and provides reliable support for UAV cooperative operation.

[0064] Further, the system further comprises:

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

[0066] ​Specifically, the world coordinate system adopts a right-hand Cartesian coordinate system, taking the starting point of the task as the origin, the X-axis pointing to the horizontal direction of the task area, the Y-axis being perpendicular to the X-axis and parallel to the ground, and the Z-axis being perpendicular to the ground upward, to construct the world coordinate system. The work unmanned aerial vehicle is equipped with an RTK (Real-Time Kinematic) module and an IMU (Inertial Measurement Unit). The RTK module obtains the coordinates of the work unmanned aerial vehicle in the world coordinate system in real time through satellite signals, and the IMU collects acceleration and angular velocity data of the unmanned aerial vehicle; the system fuses the two types of data through Kalman filtering algorithm, and outputs the first position vector of the work unmanned aerial vehicle.

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

[0068] (2) calculating a relative position deviation vector according to the first position vector and the second position vector;

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

[0070] (3) real-time acquiring a change rate of the relative position deviation vector, and calculating a compliant enhancement control law of the toolbox unmanned aerial vehicle based on the elastic coefficient, the damping coefficient and the change rate;

[0071] Specifically, the change rate of the relative position deviation vector is the derivative of the deviation vector with respect to time, reflecting the dynamic change trend of the deviation, and the compliant enhancement control law of the lower unmanned aerial vehicle can be calculated through the following formula:

[0072] ;

[0073] wherein, is the expected position of the toolbox unmanned aerial vehicle,

[0074] is the relative position deviation vector;

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

[0076] , , are the mass, damping and elastic coefficient of the ideal mass-damping-spring system respectively

[0077] The acceleration vector of the end of the manipulator arm of the work unmanned aerial vehicle.

[0078] The expected speed of the lower unmanned aerial vehicle is obtained by integration , and the expected position and the expected position .

[0079] Further, after obtaining the compliant augmentation control law of the tool box unmanned aerial vehicle, the expected position of the tool box unmanned aerial vehicle output by the compliant augmentation control law is taken as a target position, and the tool box unmanned aerial vehicle is controlled to fly to the target position to realize flexible docking.

[0080] Further, the positioning information of the lower tool box unmanned aerial vehicle is obtained, including:

[0081] (1) obtaining real-time images of the tool box unmanned aerial vehicle;

[0082] (2) identifying the position features in the real-time images;

[0083] Specifically, a camera is mounted on the work unmanned aerial vehicle, and the real-time images of the tool box unmanned aerial vehicle are captured in real time by the camera. A circular visual marker is arranged at the end of the manipulator arm of the work unmanned aerial vehicle, and a square positioning frame is arranged at the top of the replacement tool. The digital templates of the two types of features are pre-stored in the recognition algorithm library.

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

[0085] (3) calculating the relative position and the relative attitude angle between the work unmanned aerial vehicle and the replacement tool according to the position features;

[0086] Specifically, based on the triangulation principle of binocular vision, the feature point parallax collected by the left and right cameras is used to calculate the three-dimensional relative position in combination with the camera baseline distance and the focal length. The rotation state of the square positioning frame in the three-dimensional space is fitted through the pixel coordinates of the four vertices of the square positioning frame in the image, and the roll angle, pitch angle and yaw angle of the work unmanned aerial vehicle relative to the replacement tool are calculated by using the PnP algorithm as the relative attitude angle.

[0087] (4) generating a motion control instruction of the work unmanned aerial vehicle according to the relative position and generating an adjustment instruction of the replacement tool according to the relative attitude angle.

[0088] Specifically, the position of the replacement tool in the body coordinate system is represented as The position of the work drone in the inertial coordinate system is represented as (in the room measured by vicon, and measured by RTK outdoors), is the rotation matrix of the body coordinate system of the work drone to the world coordinate system. For the position of the work drone end and the replacement tool in the world coordinate system, there are the following expressions:

[0089] ;

[0090] ;

[0091] From the above expressions, we can get:

[0092] ;

[0093] ;

[0094] wherein, is the position of the replacement tool in the body coordinate system;

[0095] is the position of the work drone end in the world coordinate system;

[0096] is the position of the replacement tool in the world coordinate system;

[0097] is the rotation matrix;

[0098] is the position of the work drone in the inertial coordinate system;

[0099] is the position of the work drone arm end in the body coordinate system.

[0100] Let . Let . Therefore, the automatic docking control law is set to , Under the above control law, the system can achieve stability and the final error converges. Wherein, is the desired position of the work drone, is the position of the work drone arm end in the body coordinate system, is the target position in the body coordinate system, is a constant greater than 0, used to control the gain, is the position of the work drone in the inertial coordinate system, is the desired position of the work drone, is a constant greater than 0, used to control the gain.

[0101] Further, the system provided by the embodiment further comprises:

[0102] (1) acquiring a real-time image of the tool box UAV;

[0103] Specifically, in combination with the above description, details are not repeated here.

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

[0105] Specifically, the real-time image is subjected to grayscale processing (to reduce the amount of calculation), Gaussian filtering to remove noise, and then histogram equalization to enhance the edge contrast. The Canny operator is used to extract the image edge. In combination with the prior shape features of the replacement tool (such as a square base and a cylindrical column), the Hough transform is used to screen out straight lines / curve segments that conform to the tool contour. The screened edge contour is subjected to minimum circumscribed rectangle fitting 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. In this way, background interference (such as the sky, clouds, and ground scenery) can be excluded, and only the effective contour of the replacement tool is retained to provide a clear tracking target for subsequent displacement calculation.

[0106] (3) determining the displacement change of the dynamic tracking box in the continuous frames of real-time images, and calculating the relative linear velocity between the work UAV and the replacement tool based on the displacement change;

[0107] Specifically, for the dynamic tracking box in the continuous two frames of images, the pixel displacement of the center coordinates is calculated. In combination with the camera internal parameter and the distance from the replacement tool to the camera, the pixel displacement is converted into actual physical displacement, and the relative linear velocity is calculated according to the ratio of the actual physical displacement to time.

[0108] Further, if the displacement change of the tracking real-time image is tracked, when the original tracking box ID is not detected for N consecutive frames, or the matching degree of the tracking box and the prior shape of the replacement tool is less than a threshold value, it is judged that the dynamic tracking box is lost. A motion trajectory is fitted based on the displacement changes of the previous M frames (a quadratic polynomial is used for prediction), a predicted tracking box position is generated, a template matching algorithm is started, and an area matching the edge template (prestored high-definition contour) of the replacement tool is searched in the predicted area. If the matching degree is greater than a threshold value, the tracking box is regenerated and the calculation is continued. In addition, if the dynamic tracking box is lost for too long, the wide-angle camera is triggered to assist in shooting at this time, the search range is expanded, the relative position relationship between the GPS position of the work UAV and the tool box UAV is combined, the search area is narrowed, and if the tracking is still not restored, the docking is stopped.

[0109] (4) generating a speed closed-loop control instruction according to the relative linear velocity, and controlling the work UAV based on the speed closed-loop control instruction.

[0110] Specifically, the deviation of the target relative linear velocity and the actual relative linear velocity is taken as an input, a PID controller is used to calculate the speed adjustment amount of the work unmanned aerial vehicle as a speed closed-loop control instruction, so as to ensure that the work unmanned aerial vehicle approaches the replacement tool at a stable speed and avoid collision or docking deviation caused by excessive relative speed.

[0111] Further, the implementation step of determining the target replacement tool comprises:

[0112] (1) Collecting a real-time image of the tool box unmanned aerial vehicle, and identifying tool regions of each replacement tool in the real-time image based on a target detection algorithm;

[0113] Specifically, the collected real-time image is preprocessed first, the real-time image is input into the target detection algorithm, and all detected tool regions are output, and low-confidence targets are filtered through a confidence threshold to retain valid tool regions. In this way, the occlusion problem caused by dense arrangement of tools can be solved, and all replacement tools can be identified completely to provide accurate region positioning for subsequent contour extraction and size calculation.

[0114] (2) Performing edge detection on each tool region to extract contour features of each tool region;

[0115] Specifically, the determined image of each tool region is cropped to obtain an image containing only a single tool region; the cropped tool region image is subjected to grayscale processing to convert the color image into a grayscale image; a suitable edge detection algorithm is selected to perform edge detection on the grayscale tool region image to obtain edge information of the tool; the detected edge information is subjected to denoising and connection processing to eliminate edge breakage and false edges; the tool contour features formed by the processed edge information are extracted, including parameters such as the contour length, area, number of inflection points, and curvature change; the extracted contour features are subjected to standardization processing to eliminate the influence of image size, scaling ratio, and other factors on the contour feature parameters.

[0116] Further, the original real-time image is cropped to generate an independent tool region image for each tool region. During the cropping process, it is ensured that the tool region is completely contained in the cropped image, and the distance between the image edge and the tool edge is not less than a preset number of pixels, so as to avoid cropping off the key edge information of the tool. The image of a single tool region is subjected to image preprocessing, and the preprocessed image is further subjected to contour feature extraction. The steps related to contour feature extraction are described in the related art, which will not be described here.

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

[0118] Specifically, a mapping relationship between pixel size and actual size is established by using camera calibration parameters, pixel size of each contour feature is calculated based on contour features combined with pixel size, and actual size of the contour feature is further obtained according to the mapping relationship. By comparing the parameter information in the replacement tool in the task, the replacement tool with the same actual size and parameter information is determined as the target replacement tool.

[0119] In addition, if there are multiple contour fusion cases, in the fusion area, the gradient direction of each edge pixel is calculated, the gradient direction is K-means clustered, the pixels are divided into two groups of candidate edges, the gradient direction difference and the Euclidean distance of adjacent pixels are calculated for each group of candidate edges, the edge segments with a continuity score greater than a score threshold are retained, and the 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 to cause edge fusion, the tangent directions of the two arcs (differing by 180°) can be distinguished by gradient direction clustering, and the fusion point can be accurately segmented by combining inflection point detection, so that the sub-contours after segmentation are consistent with the edges of a single tool with a degree of coincidence ≥ 90%. In this way, the gradient features of adjacent pixels are used to realize local separation of the fused contour, solve the problem of edge sticking caused by physical contact or imaging interference, and provide preliminary segmentation results for subsequent trajectory prediction.

[0121] Further, for the tool region in the continuous M 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, the matching degree of the center coordinates of the sub-contour obtained by the current frame segmentation and the tool trajectories in the historical database is calculated, if the matching error of a sub-contour and a historical tool trajectory is less than an error threshold, the sub-contour is attributed to the tool, and the shape similarity (≥ 80% is considered consistent) of the sub-contour and the historical shape features (such as width-height ratio, convex hull defects) of the matching tool is calculated, if they are inconsistent, re-segmentation is triggered.

[0122] Further, weights are set for pixel feature analysis and contour trajectory analysis respectively, the final confidence is calculated based on the weighted sum of the two, the finally segmented sub-contour is output, and the segmentation error (compared with the contour of the subsequent non-fusion frame) is fed back to the pixel feature analysis module to dynamically optimize the clustering threshold and the inflection point detection parameters. Combining the pixel features in the spatial domain and the trajectory rules in the time domain, a complementary verification decision mechanism is formed, which effectively solves the contour fusion problem in complex scenes and ensures the reliability of tool recognition and size calculation.

[0123] Further, the coordinate positions of the work unmanned aerial vehicle and the tool box unmanned aerial vehicle are respectively determined, the installation position parameters of the target tool in the tool box unmanned aerial vehicle body coordinate system and the attitude deflection angle of the tool installation are determined by the tool box unmanned aerial vehicle, the coordinates of the docking point in the local coordinate system of the replacement tool are obtained, the coordinates of the docking point in the tool box unmanned aerial vehicle body coordinate system are calculated through the coordinate transformation formula, the coordinates of the docking point in the world coordinate system are calculated by substituting the coordinates of the docking point in the tool box unmanned aerial vehicle body coordinate system into the world coordinate conversion formula, and the coordinates of the docking point in the world coordinate system are taken as the end point of the path planning.

[0124] In addition, the system provided by the embodiment further comprises:

[0125] When the work unmanned aerial vehicle is docked with the target replacement tool, the target replacement tool in the tool box unmanned aerial vehicle is powered on;

[0126] After the work unmanned aerial vehicle completes the last work, before the target replacement tool for the next work is docked, a power-off instruction is sent to the original end tool to disconnect the original end tool from the work unmanned aerial vehicle.

[0127] Specifically, when the work unmanned aerial vehicle locks the target replacement tool through visual positioning (such as recognizing the tool exclusive ID code) and enters the docking preparation stage, the ground control station or the tool box unmanned aerial vehicle flight control system sends a power-on instruction to the independent power supply module of the target replacement tool, and only activates the electromagnet power supply of the tool. The modular power supply unit of the tool box unmanned aerial vehicle adopts a one-tool-one-circuit design, and each replacement tool corresponds to an independent relay switch. At the same time of powering on the target tool, the system sends a power-off instruction to the relays of all non-target tools to cut off the power supply of the electromagnet and the controller, ensuring that the non-target tools have no magnetic attraction capability. The current value of each tool power supply circuit can also be collected in real time through a current sensor. If the non-target tool circuit current > 0, an alarm is triggered immediately and the power supply module is restarted to forcibly cut off the abnormal circuit to prevent mis-supply caused by relay failure. In this way, the magnetic attraction capability of the non-target tool electromagnet can be completely eliminated, the additional attraction force caused by other tools being powered on during docking of the work unmanned aerial vehicle can be avoided, and problems such as mispositioning of the tool, collision damage, etc. can be prevented.

[0128] Optionally, when the relative perpendicular distance between the docking port and the target replacement tool guide positioning column is less than a threshold value and there is an axis deviation, the work unmanned aerial vehicle generates and sends a docking signal to the tool box unmanned aerial vehicle and the ground control station, the ground control station sends a target tool identity verification instruction to the tool box unmanned aerial vehicle, the tool box unmanned aerial vehicle verifies the target tool by comparing the tool hardware ID, feeds back an identity verification pass signal to the ground control station after verification, and the ground control station generates and issues a target tool power-on instruction; the tool box unmanned aerial vehicle receives the target tool power-on instruction, controls the target tool relay to close and power on, and sends a disconnection instruction to the non-target tool relay at the same time.

[0129] Further, when the job scene requires replacement tools to be on standby, and all replacement tools need to be powered, a magnetic shielding cover is added outside the electromagnet of each replacement tool to avoid magnetic field interference between adjacent tools; an isolation transformer and common mode inductor are used to achieve electrical isolation of the power supply loop, and a double-layer shielding net is used for the communication cable to reduce electromagnetic interference; on the software side, an identity verification instruction is used to ensure that only the target tool responds to the docking, the electromagnet current of the target tool is dynamically adjusted according to the relative distance, and the magnetic field is monitored in real time with the help of a Hall sensor. When an abnormality occurs, the current of the non-target tool is reduced, and the current of the target tool is compensated. This path reduces the false adsorption rate and shortens the tool switching time while ensuring the continuous power supply of all tools, meeting the demand for rapid response.

[0130] Further, the system automatically or manually selects a power supply path according to the type of the job task. The default path for regular tasks is path one, which aims to be low in power consumption and high in safety. For urgent tasks, the system switches to path two to achieve rapid response of the tools. When switching paths, the system first cuts off the power supply of all tools within 100 ms, and then reconfigures the power supply state according to the target path to avoid power supply conflicts or interference during the switching process, and to ensure the adaptability and stability of the tool power supply control in different scenarios.

[0131] Optionally, after the work unmanned aerial vehicle detects that the work is completed and hovers directly above the original tool placement hole, it generates and sends an original tool recovery ready signal to the ground control station; the ground control station generates an original tool power-off instruction and issues it to the work unmanned aerial vehicle; the work unmanned aerial vehicle receives the original tool power-off instruction to control the mechanical arm to cut off the power supply of the original tool, and at the same time detects the load force of the mechanical arm to determine whether the separation is complete according to the load force of the mechanical arm.

[0132] Further, before the work unmanned aerial vehicle completes the last work and docks the next target tool, it flies to the original tool recovery area, hovers, and sends an original tool recovery request. After the system confirms the position, it sends a power-off instruction to the original tool power supply interface, cuts off the power supply of the magnetic attraction or mechanical lock, and the original tool is recovered under the action of gravity or mechanical unlocking. The load force is detected by the end force sensor of the mechanical arm, and the connection is confirmed to be disconnected when the load threshold is reached. If it is stuck, the power is cut off again and the auxiliary separation is shaken to ensure that the original tool is completely separated and the interface is cleared for the next docking.

[0133] The system for switching the end tool in the air provided by the embodiment can directly obtain the positioning information of the tool box UAV and the size information of the replacement tool in the air through the work UAV, and can determine and switch the target replacement tool without returning to the ground. Compared with the ground fixed tool box scheme, the tool replacement time is greatly shortened, the efficiency loss and safety risk of long-distance round trip are avoided, the high-altitude and long-distance operation scene is adapted, the multi-tool cooperation demand in the complex tasks such as wind turbine maintenance and power inspection is met, and the cost and scheduling risk of multi-device cooperative operation are reduced. On the other hand, the conical boss guide positioning column of the replacement tool is in shape cooperation with the conical groove of the work UAV, combined with the flexible docking control of the tool box UAV, according to the speed start mode of the work UAV, the elastic and damping coefficients are matched according to the tool size, the deviation is obtained and the reverse compensation force is calculated to adjust the posture, the docking deviation is effectively compensated, the hard contact loss is reduced, and the power supply scheme of the target tool directional power supply combined with the non-target tool power-off or the whole tool power supply combined with the magnetic shielding and dynamic current regulation is matched to avoid misabsorption and ensure reliable docking. The success rate and efficiency of the tool switching in the air are improved, and support is provided for continuous and efficient operation in high-altitude and high-risk fields.

[0134] Corresponding to the foregoing embodiment of the system for switching the end tool in the air, the application also provides an embodiment of a method for switching the end tool in the air.

[0135] Figure 2 The flowchart of the second embodiment of the method for switching the end tool in the air provided by the application is shown in Figure 2 The method provided by the embodiment comprises:

[0136] S201, obtaining the positioning information of the tool box UAV and the size information of a plurality of replacement tools;

[0137] S202, determining a target replacement tool according to the size information;

[0138] S203, determining the relative position between the work UAV and the tool box UAV according to the positioning information;

[0139] S204, controlling the work UAV and the tool box UAV to flexibly dock the target replacement tool based on the relative position.

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

[0141] The functions and effects of each step in the above method are implemented in detail in the implementation process of the corresponding steps in the above system, which will not be described here.

[0142] The above only is the preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present 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, contact the target replacement tool based on the positioning information, and realize end tool switching of the work unmanned aerial vehicle; A tool box unmanned aerial vehicle, which is provided with a plurality of replacement tools on a top plane, and the top of each replacement tool is provided with a guide positioning column, and the bottom of the work unmanned aerial vehicle is provided with a docking interface, which is connected with the guide positioning column in a shape fitting manner, and the tool box unmanned aerial vehicle and the work unmanned aerial vehicle are flexibly docked when they contact to replace the end tool; wherein the flexible docking of the tool box unmanned aerial vehicle and the work unmanned aerial vehicle comprises: starting a compliant docking mode according to the real-time flight speed of the work unmanned aerial vehicle when the end tool does not contact the target replacement tool; the tool box unmanned aerial vehicle determines compliant control parameters according to the size information of the target replacement tool, the compliant control parameters comprising 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; and adjusts the hovering attitude of the tool box unmanned aerial vehicle according to the compensation force, the direction of the compensation force being opposite to the flight direction of the work unmanned aerial vehicle.

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 system further comprises: Determine 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; Calculate a relative position deviation vector according to the first position vector and the second position vector; Real-time acquire a change rate of the relative position deviation vector, and calculate a compliant enhanced control law of the tool box unmanned aerial vehicle based on the elastic coefficient, the damping coefficient and the change rate.

4. The system of claim 1, wherein, The acquisition of the positioning information of the underlying tool box unmanned aerial vehicle comprises: Acquire real-time images of the tool box unmanned aerial vehicle; Identify position features in the real-time images; Calculate a relative position and a relative attitude angle between the work unmanned aerial vehicle and the replacement tool according to the position features; Generate a motion control instruction of the work unmanned aerial vehicle according to the relative position, and generate an adjustment instruction of the replacement tool according to the relative attitude angle.

5. The system of claim 1, wherein, The system further comprises: Acquire real-time images of the tool box unmanned aerial vehicle; Extract an edge contour of the replacement tool in the real-time images based on an edge detection algorithm, and generate a dynamic tracking frame; 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.

6. 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.

7. 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.

8. 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.

9. A method of enabling over-the-air end tool switching, the method being implemented based on any of the systems of claims 1-8, characterized by, 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