Robot and method for determining activation mode thereof, and data processing device

Abstract

The present application discloses a robot, its activation mode determination method, and a data processing device, in which the robot selects left activation mode or right activation mode according to a judgment model using a camera, and when placed on a solar panel, the robot automatically selects the activation mode to clean the panel array.

Description

[Technical field]

[0001] This application claims priority to a Chinese patent application filed with the China Patent Office on April 30, 2022, bearing application number 202210469186.X and entitled "Robot and its startup mode determination method, data processing device," the entire contents of which are incorporated herein by reference.

[0002] The present application relates to a robot, an activation mode determination method thereof, and a data processing device. [Background technology]

[0003] After the cleaning robot is placed on the solar panel array, the method of controlling its cleaning process can be roughly divided into two types: manual control method and automatic control method. The manual control method requires one operator to be assigned to each robot, which is expensive and impractical. The automatic control method requires some fixed commands to be set for the robot to proceed along the planned route.

[0004] Since all solar panels are installed at an angle, the initial position of the robot on the solar panel is generally the lowest point of the solar panel array, and is preferably placed at the lower left or right corner of the array. However, the path of the robot on the panel array is related to its initial position on the panel array, and different initial positions of the robot will have different start-up modes and operation modes, and even different action command sets. In the prior art, the worker needs to manually set the start-up mode according to the setting position of the robot, which has the disadvantages that the operation of the worker is relatively complicated and requires the worker to learn the operation method, while once the worker makes a mistake in selecting the mode, the robot will make an error during its movement, deviating from the preset path, or the robot will select a wrong direction and fall off the solar panel.

[0005] Therefore, after the robot is placed on the solar panel, an operation mode determination method is required that obtains the robot's position on the solar panel array, determines its operation mode, and selects an appropriate command set. Summary of the Invention [Problem to be solved by the invention]

[0006] An object of the present invention is to provide a robot and a method for determining an activation mode thereof, which solves the problem that complex operations are required to select an activation mode of a robot. [Means for solving the problem]

[0007] In order to achieve the above object, the present invention provides a startup mode judgment method, which includes the steps of: constructing a startup mode judgment model using a deep learning algorithm; when a robot is placed on a solar panel array, collecting at least one real-time picture using cameras on both the left and right sides of the robot, and registering the real-time picture in the startup mode judgment model; and determining a startup mode in the initial state of the robot, which includes a left startup mode and a right startup mode.

[0008] The present application further provides a data processing device including a memory for storing executable program code, and a processor connected to the memory for reading the executable program code and operating a computer program corresponding to the executable program code, thereby executing the robot activation mode determination method described above.

[0009] The present application further provides a robot including the above data processing device.

[0010] Furthermore, the robot includes a vehicle body capable of moving over the solar panel array, and cameras installed on both the left and right sides of the vehicle body for collecting real-time images of the solar panel array, and the data processing device is installed inside the vehicle body and connected to the cameras. Effect of the Invention

[0011] The technical effect of the present invention is to use a camera to collect real-time images of both sides of the vehicle body in real time, register the real-time images in a mode judgment model, and further select a left startup mode or a right startup mode. When the robot is placed on the solar panel, it automatically selects the startup mode and executes the corresponding control command to clean. This eliminates the need for manual operation and the need to install a startup mode button, making the user's operation easier and effectively improving work efficiency and operation safety. [Brief description of the drawings]

[0012] Hereinafter, specific embodiments of the present application will be described in detail with reference to the drawings, so that the technical aspects and other beneficial effects of the present application will become apparent. [Figure 1] 1 is a route map for the robot described in Example 1 to travel on a solar panel array in left activation mode. [Diagram 2] 1 is a route map for the robot described in Example 1 to travel on a solar panel array in right-start mode. [Diagram 3] FIG. 2 is a side view of the robot described in the first embodiment. [Figure 4] 4 is an overall flowchart of a method for determining an activation mode of a robot in the first embodiment. [Diagram 5] 1 is a flowchart of steps for constructing an activation mode determination model using a deep learning algorithm in Example 1. [Figure 6] 1 is a flowchart of the steps of collecting two or more training samples in Example 1. [Figure 7]1 is a flowchart of the robot control steps described in the first embodiment. [Figure 8] FIG. 11 is a side view of the robot described in the second embodiment. [Figure 9] FIG. 11 is a plan view of the robot according to the second embodiment. [Figure 10] 11 is an overall flowchart of a method for determining an activation mode of a robot in the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, the technical aspects of the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. Based on the embodiments of the present application, other embodiments obtained by those skilled in the art without paying creative labor shall all be within the scope of protection of the present application.

[0014] In the following disclosure, many different embodiments or examples are provided to realize different structures of the present application. In order to simplify the disclosure of the present application, the following describes specific example components and installations. Of course, they are merely exemplary and are not intended to limit the present application. In addition, in different examples of the present application, reference numerals and / or reference alphabets may be repeated. Such repetition is for the purpose of brevity and clarity, and does not in itself dictate a relationship between the various embodiments and / or installations discussed. In addition, although the present application provides examples of various specific processes and materials, those skilled in the art may envision the application of other processes and / or the use of other materials.

[0015] Example 1 Example 1 of the present application discloses a robot for cleaning a solar panel array 1. Since all solar panels are installed at an angle, the initial position of the robot placed on the solar panel is generally the lowest point of the solar panel array, and it is preferable to place the robot in the lower left or right corner of the array with reference to Figures 1 and 2. The robot 2 in this example can automatically select a left-side activation mode or a right-side activation mode, eliminating the need for an operator to operate or set the mode, improving work efficiency.

[0016] As shown in Fig. 3, the robot 2 includes a vehicle body 21, a camera 22, and a data processing device (not shown), the vehicle body 21 is movable over the solar panel array 1, a cleaning unit 23, preferably a roll brush, is installed at the front or rear end of the vehicle body 21, and the data processing device is installed inside the vehicle body 21. Two cameras 22 are installed, one on each of the left and right sides of the vehicle body 21, and the two cameras 22 are for acquiring images of the solar panel array 1 on each of the left and right sides of the vehicle body 21 in real time and transmitting the image information to the data processing device. The information processing device is installed inside the vehicle body 21 and connected to the camera 22, and is capable of receiving and processing the image information sent from the camera 22 and selecting a left activation mode or a right activation mode according to the image information.

[0017] As shown in FIG. 1 and FIG. 2, the solar panel array 1 is an inclined array-type planar structure consisting of two or more solar panels, the solar panel array 1 has several rows and several columns, and the overall shape is generally rectangular or square. Therefore, the solar panel array 1 has four edge lines, including an upper edge line 11 and a lower edge line 12 parallel to each other, and a left edge line 13 and a right edge line 14 parallel to each other. When the robot 2 is placed on the solar panel array 1, the worker places the head of the body 21 facing the upper edge line 11, so that the cameras 22 on the left and right sides of the body 21 face the left edge line 13 and the right edge line 14 respectively to obtain images with the left edge line 13 and the right edge line 14, and the data processing device processes the images to determine whether the body 21 is at the lower left corner or the lower right corner of the solar panel array 1, and further determines whether to execute the left start mode or the right start mode. Each solar panel has four edge lines with metal bezels to protect the panel and to make the panel easier to identify. Depending on the judgment result, the left start mode or the right start mode is selected and the corresponding cleaning plan is executed. For convenience of description, the extension direction of the left edge line 13 is defined as a first direction X, and the extension direction of the top edge line 11 is defined as a second direction Y.

[0018] The solar panel array 1 is an inclined plane, and a suction device is usually installed on the bottom surface of the robot 2 to improve the grip of the robot 2 and prevent the robot 2 from slipping off the panel. However, there is usually a gap between two adjacent rows of panels, and if the gap is always below the robot 2 while the robot 2 is moving, the suction device will not work. Therefore, to prevent the robot 2 from slipping off, the robot 2 must not move above the gap between the two rows of solar panels 10 when cleaning the solar panels 10 of each row.

[0019] During cleaning, dust, dirt and dirty water will slide down along the surface of the panel, and in order to ensure the cleaning effect, the robot needs to start from the lower left or right corner of the solar panel array 1, run in a straight line along the left or right edge line of the solar panel array 1 from the bottom end of the array to its top end, then proceed along the array in a first direction, turn left or right when it reaches the edge of the array, and then run in the reverse direction along the first direction on the array. The height to which the robot's body has traveled in the first direction will be lower than the height to which the body previously traveled in the first direction.

[0020] Before the robot 2 is placed on the solar panel array 1, the dimensions of the solar panel array 1 are unknown parameters to the robot 2. Because the dimensions of the solar panel array 1 are different, the optimal travel route of the robot 2 on the panel array is different, the travel distance in the lateral direction each time is different, the number of turns on the panels in the same row is different, and the working mode is different. During its travel, the robot 2 needs to automatically calculate the dimensions of each solar panel 10 in the array of solar panels 10, particularly the length of the solar panels 10 in each row in the tilt direction, and calculate the number of times and distance the robot needs to travel back and forth on the solar panels 10 in each row according to the dimensions, and calculate the number of turns and position of the robot 2 to ensure that every corner of the solar panels 10 in each row can be cleaned.

[0021] The panel array area passed by the vehicle body each time it moves in the first direction may be defined as one cleaning passage, the width of each cleaning passage being less than or equal to the width of the cleaning portion 23, and when the vehicle body moves within a cleaning passage, the cleaning portion 23 may extend outside the cleaning passage.

[0022] At least one solar panel 10 in the same row may be divided into N cleaning paths extending in a first direction, extending from the left side of the array of solar panels 10 to the right side thereof, where N is the integer part of the quotient of the length of the solar panels 10 in the row on the inclined surface divided by the width of the cleaning unit 23 plus 1. For example, if the length of the panels in a row is 2 meters and the width of the roll brush is 0.7 meters, N is 3. Then, the solar panels 10 in the row need to be divided into three lateral cleaning paths, and the cleaning robot needs to make a total of three forward and backward runs until it can clean all the panels in the row without leaving any cleaning blind spots. The three cleaning passages are, from top to bottom, a first passage, a second passage and a third passage, respectively, and the widths of the three cleaning passages may be set to 0.7 meters, 0.6 meters and 0.7 meters, respectively, and when the vehicle body travels through the middle second passage, the edge portions at both ends of the roll brush will extend into the first passage and the third passage.

[0023] In this embodiment, when the vehicle body cleans the solar panel array 1, if the robot's initial position on the solar panel array (lower left corner or lower right corner) is different, the control instruction set executed thereafter will also be different, so it is a key issue for the robot to determine its initial position on the solar panel array and select the startup mode, selected left startup mode or right startup mode, accordingly.

[0024] This embodiment further provides an activation mode determination method for the above robot, and as shown in FIG. 4, the activation mode determination method specifically includes the following steps S100 to S400.

[0025] Step S100 uses a deep learning algorithm to build a start mode judgment model. As shown in Fig. 5, step S100 includes the following steps: Step S110 collects two or more training samples that include pictures with labels; Step S120 performs a grouping process on the training samples, and divides pictures with the same labels into the same group; Step S130 registers the grouped training samples into a convolutional neural network model, and trains the convolutional neural network algorithm to obtain a classifier, i.e., a start mode judgment model; in this embodiment, it is preferable to train the model using the caffe deep learning algorithm.

[0026] 6, step S110 specifically includes step S111 of setting the robot 2 at the lower left corner of the solar panel array 1 multiple times, step S112 of collecting at least one first picture using the camera 22 of the robot 2 after each setting, step S113 of setting a first label corresponding to the left start mode on each first picture, step S114 of setting the robot 2 at the lower right corner of the solar panel array 1 multiple times, step S115 of collecting at least one second picture using the camera 22 of the robot 2 after each setting, and step S116 of setting a second label corresponding to the right start mode on each second picture. In steps S112 and S115, a large number of pictures are collected in real time by both the two cameras on the left and right sides of the robot, generally more than 5000 pictures, and each picture is set with a label corresponding to the left start mode or the right start mode. In another embodiment, the steps S114 to S116 may be executed first, and then the steps S111 to S113 may be selected.

[0027] In step S200, when the robot 2 is placed on the solar panel array 1, at least one real-time picture is collected using the cameras 22 on both the left and right sides of the robot 2, and two cameras 22 are installed, one on each of the left and right sides of the vehicle body 21, for acquiring real-time images of the solar panel array 1 on both the left and right sides of the vehicle body 21 in real time. When the robot is placed in the lower left corner of the panel array and the head of the vehicle faces the upper edge line of the panel array, the left camera cannot capture an image of the panel array, but the right camera can capture an image of the panel array on the right side of the vehicle body. Similarly, when the robot is placed in the lower right corner of the panel array and the head of the vehicle faces the upper edge line of the panel array, the left camera can capture an image of the panel array on the left side of the vehicle body.

[0028] In step S300, the camera 22 transmits image information to a data processing device, and the real-time picture is registered in the startup mode judgment model. When a new picture is registered in the startup mode judgment model, the result output from the model is the type of label, and the model can determine by itself whether the label corresponding to the new picture is the first label or the second label.

[0029] Step S400 is to determine the starting mode of the robot 2 in the initial state, which includes a left starting mode and a right starting mode. Since the first label corresponds to the left starting mode and the second label corresponds to the right starting mode, the computer can determine the starting mode of the robot 2 in the initial state according to the type of the label in the previous step. After the starting mode is determined, the robot can execute the control command corresponding to the starting mode by itself, proceed on the solar panel array according to a pre-planned preferred path, and clean synchronously while proceeding, and the robot needs to clean every corner of the panel array while minimizing route overlap.

[0030] As shown in FIG. 7, the method for determining an activation mode of a robot according to the present embodiment may further include, after steps S100 to S400, a robot control step including the following steps S910 to S950.

[0031] In step S910, when the robot 2 proceeds to the upper edge line of the solar panel array 1, the robot 2 is controlled to rotate at a right angle on the spot, and if the startup mode is the left startup mode, the robot 2 is controlled to rotate at a right angle to the right, and if the startup mode is the right startup mode, the robot 2 is controlled to rotate at a right angle to the left.

[0032] Step S920 is a straight-line control step for controlling the vehicle body 21 to move linearly along the extension direction of the upper edge line of the solar panel array.

[0033] Step S930 is a turning control step in which, when the front end of the robot 2 advances to the left edge line 13 or the right edge line 14 of the solar panel array 1, it is determined whether the cleaning task is completed, and if completed, step S940 is executed, and if not, step S950 is executed.

[0034] Step S940 controls the robot 2 to stop moving forward.

[0035] In step S950, the vehicle body 21 is controlled to turn left or right in a U-shape, and the process returns to the straight ahead control step S920.

[0036] The purpose of steps S910-S950 is to achieve reciprocating cleaning over the solar panel array 1 by the vehicle body 21, where after the vehicle body 21 turns, the new cleaning trajectory is adjacent to or overlaps the previous linear cleaning trajectory. The robot travels over the solar panel array according to a pre-planned preferred path, cleaning synchronously as it travels, and the robot needs to clean every nook and cranny of the panel array without leaving any blind spots.

[0037] This embodiment further includes a data processing device, which includes a memory for storing executable program code, and a processor connected to the memory for reading the executable program code and running a computer program corresponding to the executable program code to realize at least one step of the robot activation mode determination method. The memory may also be used to store control instruction sets corresponding to the left activation mode and the right activation mode, and all of the operations realized by controlling the robot in steps S910 to S950 are realized by these control instruction sets.

[0038] The technical effect of this embodiment is that a camera is used to collect real-time images of both sides of the vehicle body in real time, the real-time images are registered in a mode judgment model, and the left startup mode or right startup mode is selected. When the robot is placed on the solar panel, it can automatically select the startup mode, execute the corresponding control instruction set, and automatically clean. This eliminates the need for manual operation and the need to install a startup mode button, making the user's operation easier and effectively improving work efficiency and operation safety.

[0039] Example 2 Example 2 includes all the technical aspects of Example 1, but one difference is that, as shown in Figures 8 to 9, the robot of Example 2 includes, in addition to the vehicle body 21, camera 22, cleaning unit 23 and data processing device of Example 1, a metal sensor 24 that is installed on both the left and right sides of the bottom of the vehicle body or is fixed to a side wall of the vehicle body and extends to the bottom surface of the vehicle body, and when the distance between the metal sensor and the metal bezel becomes equal to or less than a preset threshold value, a signal is generated from the metal sensor and transmitted to the data processing device.

[0040] For example, when the robot is placed at the lower left corner of the panel array, the metal sensor on the left side of the body can sense the left bezel and generate an electrical signal, while the metal sensor on the left side of the body is located on a solar panel and cannot generate an electrical signal. Therefore, when an electrical signal is generated from the metal sensor on the left side of the body and there is no signal from the metal sensor on the right side, the computer can determine that the initial position of the robot is at the lower left corner of the panel array and that the activation mode is the left activation mode. Similarly, when an electrical signal is generated from the metal sensor on the right side of the body and there is no signal from the metal sensor on the left side, the computer can determine that the initial position of the robot is at the lower right corner of the panel array and that the activation mode is the right activation mode.

[0041] The robot can determine the starting mode of the vehicle body by the camera 22 or the metal sensor 24, but since there is a certain probability of an error occurring during actual work, if the robot makes a decision by only one of them, the robot may fall off the panel due to an incorrect decision, which may cause a safety accident. In this embodiment, the two judgment results are calculated with a certain weight ratio to obtain a corrected judgment result, thereby making the judgment result of the starting mode more accurate.

[0042] Example 2 includes all the technical aspects of Example 1, but another difference is that, as shown in FIG. 10, Example 2 provides a method for determining an activation mode of a robot, which further includes the following steps S500 to S800 in addition to steps S100 to S400 of Example 1.

[0043] In step S500, when determining the startup mode in the initial state of the robot 2 in step S400, the data processing device may output the judgment result of the startup mode judgment model as a first group-specific parameter S1, and if the startup mode is the left startup mode, the first group-specific parameter is defined to be S1=0, and if the startup mode is the right startup mode, the first group-specific parameter is defined to be S1=1.

[0044] S600 re-determines the activation mode in the initial state of the robot 2 using the metal sensor 24, and outputs the second group-specific parameters. As shown in FIG. 8, re-determining the activation mode using the metal sensor 24 in step S600 includes the following steps S610 to S630. Step S610 installs one metal sensor 24 on each of the left and right sides of the bottom of the robot 2, and defines them as a left sensor and a right sensor. Step S620 synchronously collects electrical signals generated from the left sensor and the right sensor when the robot 2 is placed on the solar panel array 1. Step S630 determines that the activation mode in the initial state of the robot 2 is the left activation mode if there is a signal in the left sensor and no signal in the right sensor, and determines that the activation mode in the initial state of the robot 2 is the right activation mode if there is no signal in the left sensor and there is a signal in the right sensor.

[0045] When the start mode is the left start mode, the second group parameter is defined as S2=0, and when the start mode is the right start mode, the second group parameter is defined as S2=1. When the vehicle body 21 is close to the left edge line 13 or the right edge line 14, the height of the metal bezel at the edge position is low, the metal bezel is in the blind spot of the image sensor, the metal bezel does not exist in the image captured by the image sensor, and the start mode cannot be identified by the judgment model. An electrical signal is generated from the metal sensor 24 and transmitted to the output processing device, and the metal sensor 24 has the characteristics of a small detection range and high sensitivity, and the detection range is generally several tens of millimeters. The electrical signal indicates that the vehicle body 21 has approached the edge with the metal bezel of the solar panel array 1, and since the bezel position collectors are installed on both sides of the vehicle body 21, the electrical signal sent from the metal sensor 24 received by the data processing device means that the vehicle body 21 has already approached the left edge line 13 or the right edge line 14 of the solar panel array 1.

[0046] S700 calculates a correction parameter S according to the first group-specific parameters and the second group-specific parameters, S=K1*S1+K2*S2, where K1 and K2 are preset weight coefficients, and K1+K2=1. If the correction parameter S is approaching 0, it is determined that the activation mode is the left activation mode, and if the correction parameter S is approaching 1, it is determined that the activation mode is the right activation mode. By subtracting 0 from the correction parameter S, a first difference S-0 is obtained, and if the difference is less than a preset threshold value such as 0.1 or 0.2, it may be determined that S is approaching 0; similarly, by subtracting S from the correction parameter 1, a second difference 1-S is obtained, and if the difference is less than a preset threshold value such as 0.1 or 0.2, it may be determined that S is approaching 0. This makes it possible to determine whether S is approaching 0 or 1. By setting the correction parameter, it is possible to avoid erroneous determination caused by errors due to a single determination method, and the result is more accurate.

[0047] In step S800, according to the result of the correction parameter, the start-up mode in the initial state of the robot 2 is re-determined. The result of the determination in step S800 is more accurate than the result of the determination in step S400, and the probability of accidents occurring is effectively reduced.

[0048] The method for determining an activation mode of a robot according to the second embodiment further includes steps S910 to S950 described in the first embodiment after steps S100 to S800, but will not be described again here.

[0049] In the second embodiment, both the camera 22 and the metal sensor 24 are installed inside the robot 2, and the starting mode of the vehicle body can be determined by selecting only one of them. However, since there is a certain probability of errors occurring in the robot during actual operation, if the determination is made based on only one of them, the robot may fall off the panel due to an incorrect judgment, which may cause a safety accident.

[0050] In this embodiment, the robot 2 outputs one mode judgment result by the camera 22 and the mode judgment model, and outputs another mode judgment result by the metal sensor 24, and then calculates the two judgment results with a certain weight ratio to obtain a correction parameter, and automatically selects the startup mode according to the correction parameter, thereby making the judgment result of the startup mode more accurate. When the robot is placed on a solar panel, it can automatically select the startup mode and perform cleaning, eliminating the need for manual operation and installing a startup mode button, and improving work efficiency.

[0051] In the above embodiments, the description of each embodiment is given with a unique emphasis, and for the parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0052] Although the robot and its activation mode determination method, as well as the data processing device according to the embodiments of the present application, have been introduced in detail above, the present specification describes the principles and embodiments of the present application by giving specific examples, and the explanation of the above embodiments is merely to aid in understanding the technical aspects and core ideas of the present application, and those skilled in the art should understand that the technical aspects described in the above embodiments may be modified or some of the technical features therein may be replaced with equivalent ones. Such modifications or replacements do not deviate from the essence of the relevant technical aspects from the scope of the technical aspects in the embodiments of the present application.

Claims

1. A method for determining an activation mode of a robot, comprising: Positioning the robot at a lower left or right corner of a solar panel array; Building a startup mode judgment model using a deep learning algorithm; When the robot is placed on the solar panel array, collecting at least one real-time picture using cameras on both sides of the robot, and registering the real-time picture into the startup mode judgment model; and determining an activation mode in an initial state of the robot, the activation mode including a left activation mode and a right activation mode.

2. The step of constructing a startup mode judgment model using a deep learning algorithm includes: collecting two or more training samples that include pictures with labels; performing a grouping process on the training samples to classify the pictures with the same label into the same group; The method for determining an activation mode of a robot according to claim 1, further comprising: enrolling and training the grouped training samples in a convolutional neural network model to obtain an activation mode judgment model.

3. The step of collecting two or more training samples includes: placing the robot at a lower left corner of the solar panel array multiple times; After each setting, collecting at least one first picture using a camera of the robot; and placing a first label corresponding to the left activation mode on each of the first pictures.

4. The step of collecting two or more training samples includes: placing the robot at a lower right corner of the solar panel array multiple times; collecting at least one second picture using a camera of the robot after each setting; and placing a second label corresponding to the right activation mode on each of the second pictures.

5. outputting a first group-specific parameter when determining an activation mode in an initial state of the robot; determining an activation mode of the robot in an initial state using a metal sensor and outputting a second group-specific parameter; Calculating a correction parameter S according to the first group-specific parameters and the second group-specific parameters, where S=K1*S1+K2*S2, where K1 and K2 are preset weighting coefficients, and K1+K2=1; 2. The method of claim 1, further comprising the step of: re-determining an activation mode in an initial state of the robot according to a result of the correction parameter.

6. Re-determining the startup mode of the robot in the initial state using a metal sensor Installing a metal sensor on each of the left and right sides of the bottom of the robot, and defining them as a left sensor and a right sensor; 6. The method for determining an activation mode of a robot as described in claim 5, comprising the steps of: when the robot is placed on a solar panel array, synchronously collecting electrical signals generated from the left sensor and the right sensor; if there is a signal in the left sensor and there is no signal in the right sensor, determining that the activation mode in an initial state of the robot is the left activation mode; and if there is no signal in the left sensor and there is a signal in the right sensor, determining that the activation mode in an initial state of the robot is the right activation mode.

7. The robot includes a vehicle body capable of running on a solar panel array, After using the vehicle body to move to the upper edge of the solar panel array, If the activation mode is a left activation mode, controlling the vehicle body to turn right at a right angle, and if the activation mode is a right activation mode, controlling the vehicle body to turn left at a right angle; a straight-line control step of controlling the vehicle body so as to move linearly along an extension direction of an upper edge of the solar panel array; 2. The method for determining the startup mode of a robot as described in claim 1, further comprising: a turning control step of determining whether the cleaning task is completed when the front end of the robot advances to the left edge line or the right edge line of the solar panel array, and if completed, controlling the robot to stop progressing, and if not, controlling the body to turn in a U-shape to the left or right and returning to the straight-line control step.

8. 1. A data processing device, comprising: a memory for storing executable program code; and a processor connected to the memory, the processor reading the executable program code and running a computer program corresponding to the executable program code, thereby executing the robot activation mode determination method described in any one of claims 1 to 7.

9. A robot comprising the data processing device according to claim 8.

10. A vehicle body capable of moving on a solar panel array; and a camera installed on each of the left and right sides of the vehicle body for collecting real-time images of the solar panel array. The robot according to claim 9 , wherein the data processing device is installed inside the vehicle body and connected to the camera.

11. The solar panel array is an array-type planar structure consisting of two or more solar panels, The robot according to claim 10 , wherein a metal bezel is provided on an edge portion of each of the solar panels as a frame member for protecting the panel.

12. Further comprising a metal sensor installed on both the left and right sides of the bottom of the vehicle body or fixed to a side wall of the vehicle body and extending to the bottom of the vehicle body; 11. The robot according to claim 10, wherein when a distance between a metal sensor and a metal bezel falls below a preset threshold, a signal is generated from the metal sensor and transmitted to the data processing device.

Images