Method and system for cross-domain scheduling of distributed photovoltaic cleaning trolley by rotor unmanned aerial vehicle

Abstract

The invention discloses a method and system for cross-domain scheduling of a distributed photovoltaic cleaning trolley through a rotor unmanned aerial vehicle, and belongs to the technical field of unmanned aerial vehicle application, and the system comprises a cleaning trolley, a rotor unmanned aerial vehicle, a magnetic induction positioning device, an ultrasonic sensor, a visual sensor and a GPS module. The method comprises the following steps: a cleaning trolley returns to a parking area under the navigation of a GPS module, and the relative position of the cleaning trolley and the center of the parking area is finely adjusted; the rotor unmanned aerial vehicle flies to the position above the parking area through positioning of a GPS module, the cleaning trolley is detected through YOLOV5m, the relative position of the rotor unmanned aerial vehicle and the center of the parking area is adjusted, and coarse adjustment including a hovering mode and second-stage coarse adjustment are carried out; the rotor unmanned aerial vehicle enters a fine adjustment mode, and when the ultrasonic ranging height is smaller than # imgabs0 #, a motor of the rotor unmanned aerial vehicle is closed for landing; according to the invention, accurate positioning and auxiliary landing of the unmanned aerial vehicle and the cleaning trolley can be realized, and accurate cross-domain scheduling of the cloth type photovoltaic cleaning trolley is realized.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle application technology, and in particular to a method and system for a rotor-wing unmanned aerial vehicle to dispatch distributed photovoltaic cleaning vehicles across domains. Background Art

[0002] In recent years, photovoltaic power generation has been increasingly used, and the cleaning of solar panels has attracted much attention. For the cleaning of distributed photovoltaic panels, the tracked photovoltaic cleaning vehicle has the best cleaning effect and is the most widely used. However, this cleaning solution cannot be autonomously scheduled across domains and is generally costly.

[0003] With the development of drone technology, drones have been widely used in transportation. The existing drone transportation of photovoltaic cleaning vehicles has at least the following problems: (1) Reliance on a single GPS positioning results in large positioning errors and low accuracy; (2) There is a lack of effective auxiliary means when the UAV is landing, which can easily lead to target loss and recognition failure due to changes in viewing angle; (3) Since photovoltaic panels are often tilted, the existing structure of the UAV and the cleaning vehicle is simple when docked and fixed, and it is impossible to effectively adjust the relative position and maintain a stable connection, which cannot meet the needs of precise positioning, stable landing, and reliable docking and transportation. Summary of the invention

[0004] One of the purposes of the present invention is to provide a system for cross-domain dispatching of distributed photovoltaic cleaning vehicles by rotorcraft UAVs to solve the above-mentioned problems.

[0005] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is as follows: a system for cross-domain dispatching of distributed photovoltaic cleaning vehicles by rotorcraft UAVs, comprising a cleaning vehicle, a rotorcraft UAV, a magnetic induction positioning device, an ultrasonic sensor, a visual sensor and a GPS module, wherein the magnetic induction positioning device is composed of a detection coil and a permanent magnet, the permanent magnet is located at the center of the parking area on the photovoltaic panel, and the detection coil is arranged at the bottom of the cleaning vehicle; the ultrasonic sensor and the visual sensor are both arranged at the bottom of the main engine of the rotorcraft UAV, the GPS module is installed inside the rotorcraft UAV, and a flight pad for landing the rotorcraft UAV is arranged on the top of the cleaning vehicle.

[0006] The magnetic induction positioning device composed of detection coils and permanent magnets is used for the precise parking of the cleaning trolley relative to the photovoltaic panel; while the self-centering clamping mechanism is used for the relative position adjustment and connection between the drone and the trolley.

[0007] As a preferred technical solution, a self-centering gripping mechanism is also provided on the landing pad, and the self-centering gripping mechanism includes a return spring provided with a guide shaft, a symmetrically centered crank slider, a V-block, a frame, a driving servo, a magnet and a magnetic sensor, wherein the magnet is placed on the positioning column; the magnetic sensor is located at the center of the frame, and the frame is provided with threaded holes for screw connection with the trolley; there are two V-blocks and they are symmetrically arranged, and the V-blocks are connected to the guide shaft, and the return spring is responsible for the return movement; the output shaft of the driving servo is connected to the symmetrically centered crank slider, and the V-block is driven to move by the linkage rod of the symmetrically centered crank slider.

[0008] After the UAV lands on the landing pad, the driving servo drives the symmetrical centrifugal crank slider, driving the V-shaped block to clamp the four positioning columns of the UAV along the guide shaft; the magnet is placed on the positioning column; the magnetic sensor is located at the center of the frame, and transmits a contact signal after contacting the magnet of the positioning column; the driving servo receives the contact signal and drives the symmetrical centrifugal crank slider.

[0009] The second object of the present invention is to provide a method for cross-domain dispatching of distributed photovoltaic cleaning vehicles using the above system, comprising the following steps: S100, the cleaning trolley returns to the parking area under the navigation of the GPS module, and further fine-adjusts the relative position of the cleaning trolley and the center of the parking area according to the magnetic induction positioning device; S200, the rotor drone flies above the parking area through the positioning of the GPS module, detects the cleaning car through YOLOV5m, and adjusts the relative position of the rotor drone and the center of the flying pad, entering the first stage of coarse adjustment including hovering mode; The S300 and the rotor UAV perform the second stage of rough adjustment based on the auxiliary vehicle logo and ultrasonic ranging results read by the detection; S400, rotor drone enters fine adjustment mode, when ultrasonic ranging height is less than When the rotor UAV motor is turned off for landing.

[0010] The ultrasonic ranging height refers to the relative distance between the center of the downward-looking camera of the rotorcraft UAV and the landing pad in the vertical direction, wherein: It means: whether to send a "re-land" command, ; It means: the altitude at which autonomous descent begins. .

[0011] As a preferred technical solution, in step S100, the specific method is: First, the parking area on the solar panel is square, and the center of the parking area corresponds to the GPS coordinates of the parking area. The cleaning car receives the GPS coordinates of the parking area and navigates to the parking area based on GPS positioning technology; Then, the coil serving as a detector is located at the bottom of the cleaning trolley, and a permanent magnet is selected as a magnetic field source with stable strength and installed at the center of the parking area of ​​the photovoltaic panel; when the cleaning trolley moves, the position of the cleaning trolley relative to the center of the parking area changes, generating a changing magnetic field, which will induce an electromotive force on the detection coil; by measuring the size and direction of the electromotive force, the relative position of the rotorcraft UAV and the center of the parking area is determined; after the adjustment is completed, the electromotive force is 0, and a parking completion signal is sent to the control center; after receiving the parking completion signal, the control center sends a start command to the rotorcraft UAV.

[0012] As a preferred technical solution, in step S200, the specific method is: The rotary-wing UAV uses GPS for preliminary navigation and positioning to hover above the center of the parking area; the downward-looking camera carried by the rotary-wing UAV is opened with OpenCV, and the cleaning vehicle is detected through the YOLOV5s model; at the same time, the ultrasonic sensor performs distance measurement to obtain the relative height between the center of the downward-looking camera of the rotary-wing UAV and the landing pad; The landing pad on the top of the cleaning trolley is square; an auxiliary vehicle tag is set at the midpoint of each of the four sides of the landing pad, and the auxiliary vehicle tag is AprilTag, which is different from the ID number of the guide landmark, and the ID numbers of the four auxiliary vehicle tags are also different; a center tag is set at the center of the landing pad, and the center tag is AprilTag; a nested center tag is also set in the middle of the center tag, and the nested center tag is AprilTag, and the ID number of the nested center tag has not been used before; a) The rotor UAV adopts adaptive control. At each hovering height, the rotor UAV sets The time threshold, If the rotor drone searches for any auxiliary vehicle logo 5-8 times within the detection time, it goes to step S300 (i.e. Whether it can be identified within this time interval, according to this result, enter S300); b) If only the cleaning cart is detected, then according to the coordinate information of the cleaning cart detected by YOLOV5m and the ranging result of the ultrasonic sensor, set The threshold is used to adjust the horizontal position. After completion, it descends at a speed of 20.0-20.5cm / s , enter the hover mode again; Repeat the above steps a) and b) until any auxiliary vehicle logo is recognized.

[0013] Considering that the altitude may be too high at the beginning of landing and the code cannot be recognized, we should land gradually, lower the altitude and repeat the steps of stop recognition-horizontal adjustment-descent. Figure 4 .

[0014] As a preferred technical solution, in step S300, the specific method is: In step S300, the specific method is: c) the rotor UAV adopts adaptive control, lands at a speed of 12.0-12.5cm / s in the height direction, and adjusts the horizontal plane according to the auxiliary vehicle logo read by OpenCV; If the center mark is read 5-8 times during the above step c), the process proceeds to step S400; otherwise, the motion state of the above step c) is maintained; There is a height threshold during the landing phase , It means: whether to send a "re-land" command, ; When the ultrasonic sensor distance measurement result is less than , but the center mark is still not detected, the rotor UAV sends a "re-landing" command to the control center, and returns to the height when step S300 is just entered, and re-performs landing detection and adjustment.

[0015] As a preferred technical solution, in step S400, the specific method is: the unmanned landing speed of the rotor is 8.0-8.3 cm / s; when the nested center mark is detected, the landing speed becomes 4.0-4.2 cm / s. At this stage, when the ultrasonic ranging height is less than , the rotors are unmanned and the motors are turned off, and the aircraft lands by its own gravity; After the rotary-wing UAV completes precise landing, the rotary-wing UAV adopts a self-centering gripping mechanism to grab the cleaning trolley.

[0016] First, the present invention controls the rotor UAV in stages and layers to accurately determine the landing position of the UAV; Then, after the rotor UAV lands, the relative position of the rotor UAV and the distributed photovoltaic cleaning vehicle is further adjusted through the self-centering clamping structure configured on the landing pad, and the four legs of the rotor UAV are clamped to fix it; The present invention achieves precise landing of the drone on top of the trolley by precise parking of the trolley and precise landing of the drone through visual recognition, thereby achieving a certain error range for the drone to land on top of the trolley. Within this error range, the drone can be adjusted by the self-centering gripper of the present invention. Since the photovoltaic panel is inclined, the landing platform of the drone is also inclined, which is prone to the risk of slipping. The present invention relies on the self-centering gripper to allow the relative position error between the drone and the trolley and protect the drone, thereby reducing or even preventing its risk of slipping and facilitating subsequent cross-domain scheduling.

[0017] The present invention uses a magnetic induction positioning device to further accurately adjust the position of the cleaning vehicle based on the initial GPS positioning, thereby solving the problem of large errors and low accuracy of GPS positioning alone. The present invention uses GPS positioning, machine vision to assist landing, and calls OpenCV to capture and identify the Apriltag label; The invention uses multiple Apriltag tags to cooperate with each other and combine with ultrasonic ranging to control the landing state of the rotary wing UAV in stages, solving the problem of target loss and recognition failure due to a sudden decrease in the viewing angle during the landing process of the UAV. The present invention uses a self-centering clamp structure to further adjust the relative position of the rotor UAV and the distributed photovoltaic cleaning vehicle, and clamps the four legs of the rotor UAV by the clamp for preliminary fixation, and then further fixes them by an electromagnetic induction device.

[0018] Compared with the prior art, the advantages of the present invention are: (1) Accurate positioning: The present invention adopts a multi-stage, multi-technique integrated positioning method: first, GPS navigation is used to make the distributed photovoltaic cleaning vehicle return to the parking area, and then the magnetic induction positioning device is used to fine-tune its relative position with the center C; during the landing process, the rotor UAV is first located above the parking area by GPS, and then OpenCV is used to parse the guide landmark AprilTag (including the guide landmarks located at the four corners of the parking area, the auxiliary vehicle landmarks at the midpoints of the four sides of the flying apron, the center landmark of the flying apron, and the nested center landmark nested in the center of the center landmark) for position adjustment, and ultrasonic sensor ranging assistance is combined at the same time; the problem that the existing technology relies only on a single GPS positioning, resulting in large positioning errors and low precision, is solved; (2) Assisted landing: The present invention sets detection and adjustment mechanisms at different stages during the landing process of the rotor UAV, utilizes a variety of landmark tags and adaptive control technologies, and combines ultrasonic ranging to control the landing state in stages, effectively solving the problems of target loss and recognition failure caused by changes in viewing angle, as well as the lack of effective auxiliary means during the landing of the UAV and the problems that are easily caused by changes in viewing angle. (3) Stable and reliable docking: After the rotor UAV lands, the present invention utilizes the self-centering clamping structure configured on the landing pad to further adjust the relative position with the distributed photovoltaic cleaning vehicle, and clamps the four positioning columns of the rotor UAV for fixation, which solves the problem of the prior art that the UAV and the cleaning vehicle have a simple structure and cannot effectively adjust the relative position and maintain a stable connection when docking and fixing. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of a parking area according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a flight pad according to an embodiment of the present invention; Figure 3is a flow chart of the present invention; Figure 4 is a specific flow chart of step S200; Figure 5 is a specific flow chart of step S300 and step S400; Figure 6 This is the structural diagram of the cleaning trolley; Figure 7 It is the structural diagram of the self-centering gripping mechanism; Figure 8 for Figure 7 AA section view; Fig. 9 for Figure 7 The structural diagram of the centrally symmetrical crank slider; Fig.10 It is an overall schematic diagram of the implementation of the embodiment of the present invention; Fig.11 Schematic diagram of the structure of a rotary-wing UAV Fig.12 It is a cross-sectional view of the positioning column; In the figure: 1. positioning column a; 2. return spring; 3. symmetrical centrifugal crank slider; 4. V-block; 5. frame; 6. driving servo; 7. magnet; 8. magnetic sensor; 9. positioning column b; 10. positioning column c; 11. positioning column d. DETAILED DESCRIPTION

[0020] The present invention will be further described below in conjunction with embodiments.

[0021] Embodiment 1: See also Figure 3 , a method for cross-domain dispatching of distributed photovoltaic cleaning vehicles by a rotary-wing UAV, comprising the following steps: S100, the distributed photovoltaic cleaning vehicle returns to the parking area under GPS navigation, and further fine-tunes the relative position of the distributed photovoltaic cleaning vehicle and the center C of the parking area according to the magnetic induction positioning device, as follows: In this step S100, Figure 1 ,The parking area on the solar panel is square, with a size of 80cm×80cm; the center C of the parking area corresponds to the GPS coordinates of the parking area; In this embodiment, after the distributed photovoltaic cleaning vehicle completes the cleaning task on the current solar panel, it sends a task completion signal to the control center; the control center traverses the array storing all the solar panel parking areas in the area according to the received signal, and feeds back the GPS coordinates of the parking area of ​​the currently numbered solar panel to the cleaning vehicle and the rotor drone in standby state; The distributed photovoltaic cleaning vehicle receives the GPS coordinates of the parking area and navigates to the parking area based on GPS positioning technology; The coil used as the detector is located at the bottom of the cleaning trolley. A permanent magnet is selected as a magnetic field source with stable strength and installed at the center of the parking area of ​​the photovoltaic panel. When the cleaning trolley moves, the position of the cleaning trolley relative to the center of the parking area changes, generating a changing magnetic field, which will induce an electromotive force on the detection coil. By measuring the magnitude and direction of the electromotive force, the relative position of the rotorcraft and the center of the parking area is determined. After the adjustment is completed, the electromotive force is 0, and a parking completion signal is sent to the control center. After receiving the parking completion signal, the control center sends a start command to the rotorcraft. S200, the rotor UAV flies above the parking area through GPS positioning, detects the distributed photovoltaic cleaning vehicle through YOLOV5m, and adjusts the relative position of the rotor UAV relative to the center C of the parking area, entering the first stage of coarse adjustment including hovering mode, as follows: The rotor UAV uses GPS for preliminary navigation and positioning to hover above the center C of the parking area; OpenCV is used to open the downward-looking camera carried by the rotor UAV, and the distributed photovoltaic cleaning vehicle is detected through the YOLOV5s model; at the same time, the ultrasonic sensor performs distance measurement to obtain the relative height between the center of the rotor UAV downward-looking camera and the landing pad; It should be noted that the experimental environment of this embodiment is located in Chengdu, Sichuan. According to relevant data, the optimal installation angle of distributed solar panels at this longitude and latitude is 16°. Fig.10 As shown in the figure, the lightweight model YOLOV5m used in the present invention can better balance the model size and detection speed under the experimental conditions. When training the YoloV5m model, only one class is marked. In order to improve the detection performance, images of distributed photovoltaic cleaning vehicles under different lighting conditions, different shooting angles and not shooting at the same height are collected to enrich the data set when the photovoltaic panel with an inclination angle of 16° is placed. Combination Figure 2 The landing pad on the top of the distributed photovoltaic cleaning vehicle is a 50cm×50cm square; the auxiliary vehicle tags are AprilTags of the TAG36H11 series with a size of 15cm×15cm, which are different from the ID numbers of the guide landmarks, and the ID numbers of the four auxiliary vehicle tags are also different, and are placed at the midpoints of the four sides of the landing pad; the center tag is an AprilTag of the TAG16H5 series with a size of 10cm×10cm, which is placed in the center of the landing pad; the nested center tag is a 5cm×5cm AprilTag of the TAG36H11 series, and the ID number of this tag has not been used before, and it is nested in the middle of the center tag; In step S200, the rotor UAV adopts adaptive control to suppress the uncertainty of the UAV dynamics and external interference caused by changing environmental factors; the rotor UAV sets If the rotor drone searches for any auxiliary vehicle logo 5-8 times during the detection time, it goes to step S300; if only the distributed photovoltaic cleaning vehicle is detected, the coordinate information of the distributed photovoltaic cleaning vehicle detected by YOLOV5m and the ultrasonic sensor ranging result are set. The threshold is used to adjust the horizontal position. After completion, it descends at a speed of 20cm / s. , enter the hover mode again and repeat the above steps until any auxiliary vehicle logo is recognized; It should be noted that the purpose of setting the hovering link is to more accurately determine the relative position of the current rotorcraft and the center C, and to better adjust the horizontal plane direction; the relative position of the landmark and the nested center landmark can be obtained by knowing the ID number of the guiding landmark; Combination Figure 4 In step S200, the rotor drone is landing at a speed of 20 cm / s. According to the icon detected by the downward camera and the adaptive control, the rotor drone still makes adjustments in the horizontal direction. If any auxiliary vehicle logo is searched for 5-8 times in this stage, step S300 is entered; S300, the rotor unmanned vehicle performs the second stage of rough adjustment based on the auxiliary vehicle logo and ultrasonic ranging results read by the detection; Combination Figure 2 The rotor drone adopts adaptive control, lands at a speed of 12 cm / s in the height direction, and adjusts the horizontal plane according to the auxiliary vehicle logo read by OpenCV; if the center logo is read 5-8 times during this process, it enters step S400; otherwise, it maintains the above motion state; Combination Figure 5 It should be noted that there is a height threshold during the landing phase. ; When the ultrasonic sensor distance measurement result is less than , but the center mark is still not detected, the rotorcraft sends a "re-landing" command to the control center, and returns to the altitude just entered into step S300, and re-performs landing detection and adjustment; S400, rotor unmanned enters fine adjustment mode, when ultrasonic ranging height is less than When the unmanned rotor motor is shut down, the aircraft will land. In step S400, Figure 5 The rotor unmanned landing speed is 8cm / s; when the nested center mark is detected, the landing speed becomes 4cm / s. At this stage, combined with Fig.10 , when the ultrasonic ranging height is less than , the rotors are unmanned and the motors are turned off, and the aircraft lands by its own gravity; It should be noted that Affected by the FOV of the downward-looking camera and the wind speed of the working environment, the selection should be based on the actual situation; After the rotor drone completes the precise landing, the rotor drone grabs the distributed photovoltaic cleaning vehicle. The specific operation process is as follows: See also Figure 6 The distributed photovoltaic cleaning vehicle includes a self-centering gripping mechanism disposed on the top and a roller brush disposed on the bottom, wherein the self-centering gripping mechanism is as follows: Figure 7 and Figure 8 As shown, it is composed of a return spring 2 with a guide shaft, a symmetrically aligned crank slider 3, a V-shaped block 4, a frame 5, a driving steering gear 6, a magnet 7 and a magnetic sensor 8, wherein the driving steering gear 6 drives the symmetrically aligned crank slider 3 to drive the V-shaped block 4 to clamp the positioning column a1 along the guide shaft of the return spring 2; the magnet 7 is placed on the positioning column a1; the magnetic sensor 8 is located at the center of the frame 5, and transmits a contact signal after contacting the magnet 7 of the positioning column a1; the frame 5 is provided with a threaded hole for screw connection with the distributed photovoltaic cleaning trolley; combined with Figure 6 The self-centering gripping mechanism is installed at A, B, C, and D of the landing pad respectively; combined with Fig.11 The rotor UAV includes a positioning column a1, a positioning column b9, a positioning column c10, and a positioning column d11; the positioning column b9 is connected to the self-centering clamping mechanism at B, the positioning column c10 is connected to the self-centering clamping mechanism at C, and the positioning column d11 is connected to the self-centering clamping mechanism at D. The above connection methods are the same as the connection working method of the positioning column a1 and the self-centering clamping mechanism at A described in detail above.

[0022] Among them, see Fig. 9 The symmetrical centrifugal crank slider 3 includes a driving crank 31, a connecting rod 32 and a linkage rod 33; when working, the steering gear drives the crank to rotate counterclockwise, so that the connecting rod drives the linkage rod to move toward the center along the frame limit groove; in this process, the linkage rod drives the V-shaped block 4 to move toward the center to clamp the positioning column a1; After the rotor UAV completes precise landing, the magnetic sensor is triggered, and the four driving servos 6 for self-centering gripping are started, driving the symmetrical centering crank slider 3 to rotate, so that the V-shaped block 4 moves along the guide axis to the middle, so that the four positioning columns of the rotor UAV are adjusted in the horizontal and longitudinal positions; after the four positioning columns have been completely clamped, the four driving servos 6 stop working; the motor of the rotor UAV is started, and the distributed photovoltaic cleaning vehicle is carried to the GPS coordinate point given by the control center; It should be noted that the installation angle of the distributed solar panels in this experimental environment is 16°. At this height, the motor is turned off and the drone is lowered to the solar panel with an inclination of 16°. The error between the center of the rotorcraft and the center of the landing pad can be adjusted by the self-centering clamp. The motor of the unmanned rotor starts, and the distributed photovoltaic cleaning vehicle is carried to the GPS coordinate point given by the control center; It should be noted that the GPS coordinates given by the control center are the coordinates of the parking area corresponding to the next solar panel that needs to be cleaned; for the cross-domain scheduling of distributed photovoltaic cleaning vehicles, please refer to Figure 3 ; The unmanned rotor motor is started, and according to the method for cross-domain dispatching of distributed photovoltaic cleaning vehicles by a rotor drone disclosed in the present invention, the system enters step S100.

[0023] See also Figure 3 After the system completes step S400, that is, the rotor drone motor is turned off, the rotor drone's cross-domain dispatching of the distributed photovoltaic cleaning car task has been completed, the drone falls freely, and the magnet installed on its feet triggers the magnetic sensor in the middle of the self-centering fixture of the cleaning car, driving the steering gear 6 to drive the symmetrical centering crank slider 3, so that the self-centering fixture clamps the positioning column, and driving the steering gear 6 to stop working. The rotor drone has completed the landing of the distributed photovoltaic cleaning car; the rotor drone motor is started, and under GPS navigation, the rotor drone flies back to the standby point.

[0024] according to Fig.11 The rotor drone comprises four identical positioning columns, namely positioning column a1, positioning column b9, positioning column c10, and positioning column d11. The structure of the positioning column is as follows: Fig.12 As shown; the positioning column is set at the position of the four feet of the rotor UAV, and the foot positioning column is installed with a magnet. According to the working principle of the magnetic sensor, when the magnetic field is enhanced to a certain extent, it outputs a high voltage (equivalent to the switch function). After the rotor UAV falls on the top of the cleaning trolley, due to the close distance and strong magnetic field, the magnetic sensor is driven to output a high voltage, start the driving servo 6, and drive the symmetrical centering crank slider to adjust the positioning column (that is, the position of the UAV relative to the trolley) through the fixture and clamp the positioning column; after the driving servo 6 works for 20s to 25s, it stops working. At this time, the relative position adjustment of the trolley with respect to the UAV has been completed, and the trolley and the UAV are connected, and the next location can be reached through GPS.

[0025] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A system for cross-domain dispatching of distributed photovoltaic cleaning vehicles by rotary-wing drones, characterized in that: It includes a cleaning trolley, a rotary-wing UAV, a magnetic induction positioning device, an ultrasonic sensor, a visual sensor and a GPS module, wherein the magnetic induction positioning device is composed of a detection coil and a permanent magnet, the permanent magnet is located at the center of the parking area on the photovoltaic panel, and the detection coil is arranged at the bottom of the cleaning trolley; the ultrasonic sensor and the visual sensor are both arranged at the bottom of the main body of the rotary-wing UAV, the GPS module is installed inside the rotary-wing UAV, and a landing pad for the rotary-wing UAV to land is arranged on the top of the cleaning trolley.

2. The system according to claim 1, characterized in that A self-centering clamping mechanism is also provided on the landing pad, and the self-centering clamping mechanism includes a return spring provided with a guide shaft, a symmetrically centered crank slider, two symmetrically arranged V-blocks, a frame, a driving servo, a magnet and a magnetic sensor. The rotor UAV is provided with a positioning column, wherein the magnet is placed on the positioning column; the magnetic sensor is located at the center of the frame, and the frame is provided with a threaded hole for screw connection with the trolley; there are two V-blocks and they are symmetrically arranged, and the V-blocks are connected to the guide shaft, and the return spring is responsible for the return movement; the output shaft of the driving servo is connected to the symmetrically centered crank slider, and the V-block is driven to move by the linkage rod of the symmetrically centered crank slider.

3. A method for cross-domain dispatching of distributed photovoltaic cleaning vehicles using the system of claim 1 or 2, characterized in that: The steps include: S100, the cleaning trolley returns to the parking area under the navigation of the GPS module, and further fine-adjusts the relative position of the cleaning trolley and the center of the parking area according to the magnetic induction positioning device; S200, the rotor drone flies above the parking area through the positioning of the GPS module, detects the cleaning car through YOLOV5m, and adjusts the relative position of the rotor drone and the center of the flying pad, entering the first stage of coarse adjustment including hovering mode; The S300 and the rotor UAV perform the second stage of rough adjustment based on the auxiliary vehicle logo and ultrasonic ranging results read by the detection; S400, rotor drone enters fine adjustment mode, when ultrasonic ranging height is less than When the rotor UAV motor is turned off for landing, the ultrasonic ranging height refers to the relative distance between the center of the rotor UAV's downward-looking camera and the landing pad in the vertical direction. It means: the altitude at which autonomous descent begins. .

4. The method according to claim 3, characterized in that In step S100, the specific method is: First, the parking area on the solar panel is square, and the center of the parking area corresponds to the GPS coordinates of the parking area. The cleaning car receives the GPS coordinates of the parking area and navigates to the parking area based on GPS positioning technology; Then, the coil serving as a detector is located at the bottom of the cleaning trolley, and a permanent magnet is selected as a magnetic field source with stable strength and installed at the center of the parking area of ​​the photovoltaic panel; when the cleaning trolley moves, the position of the cleaning trolley relative to the center of the parking area changes, generating a changing magnetic field, which will induce an electromotive force on the detection coil; by measuring the size and direction of the electromotive force, the relative position of the rotorcraft and the center of the parking area is determined; after the adjustment is completed, the electromotive force is 0, and a parking completion signal is sent to the control center; after receiving the parking completion signal, the control center sends a start command to the rotorcraft.

5. The method according to claim 3, characterized in that: In step S200, the specific method is: The rotary-wing UAV uses GPS for preliminary navigation and positioning to hover above the center of the parking area; the downward-looking camera carried by the rotary-wing UAV is opened with OpenCV, and the cleaning vehicle is detected through the YOLOV5s model; at the same time, the ultrasonic sensor performs distance measurement to obtain the relative height between the center of the downward-looking camera of the rotary-wing UAV and the landing pad; The landing pad on the top of the cleaning trolley is square; an auxiliary vehicle tag is set at the midpoint of each of the four sides of the landing pad, and the auxiliary vehicle tag is AprilTag, which is different from the ID number of the guide landmark, and the ID numbers of the four auxiliary vehicle tags are also different; a center tag is set at the center of the landing pad, and the center tag is AprilTag; a nested center tag is also set in the middle of the center tag, and the nested center tag is AprilTag, and the ID number of the nested center tag has not been used before; a) The rotor UAV adopts adaptive control. At each hovering height, the rotor UAV sets The time threshold, If the rotor drone searches for any auxiliary vehicle logo 5-8 times within the detection time, it goes to step S300; b) If only the cleaning vehicle is detected, set the coordinate information of the cleaning vehicle detected by YOLOV5m and the ranging result of the ultrasonic sensor The threshold is used to adjust the horizontal position. After completion, it descends at a speed of 20.0-20.5cm / s , enter the hover mode again; Repeat the above steps a) and b) until any auxiliary vehicle logo is recognized.

6. The method according to claim 5, characterized in that In step S300, the specific method is as follows: c) the rotor drone adopts adaptive control, lands at a speed of 12.0-12.5 cm / s in the height direction, and makes adjustments on the horizontal plane according to the auxiliary vehicle logo read by OpenCV; If the center mark is read 5-8 times during the above step c), the process proceeds to step S400; otherwise, the motion state of the above step c) is maintained; There is a height threshold during the landing phase , It means: whether to send the "re-landing" command, ; When the ultrasonic sensor distance measurement result is less than , but the center mark is still not detected, the rotor UAV sends a "re-landing" command to the control center, and returns to the altitude just entered into step S300, and re-performs landing detection and adjustment.

7. The method according to claim 5, characterized in that In step S400, the specific method is: the unmanned landing speed of the rotor is 8.0-8.3 cm / s; when the nested center mark is detected, the landing speed becomes 4.0-4.2 cm / s. At this stage, when the ultrasonic ranging height is less than , the rotors are unmanned and the motors are turned off, and the aircraft lands by its own gravity; After the rotary-wing UAV completes precise landing, the rotary-wing UAV adopts a self-centering gripping mechanism to grab the cleaning trolley.

Images