Photovoltaic robot path planning method, photovoltaic robot, device and medium

Abstract

The present application relates to a kind of photovoltaic robot path planning method, photovoltaic robot, equipment and medium, according to machine cleaning performance index setting adjacent path spacing, then generate internal cleaning path based on photovoltaic panel map, adjacent path spacing, internal cleaning path is discretized into several path points with expected pose;Acceptance path point is set on each sub cleaning path of internal cleaning path, and the parameter of acceptance path point includes expected cleaning area;Internal cleaning path is executed, real-time monitoring pose error between actual pose and expected pose, determine cleaning performance error according to actual cleaning area and expected cleaning area;When pose error is greater than pose tolerance threshold or cleaning performance error is lower than clean tolerance threshold, generate and execute target internal cleaning path based on pose error and / or cleaning performance error. Improve the execution speed and cleaning efficiency of photovoltaic robot to cleaning path, while ensuring that cleaning quality reaches expected standard.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic panel cleaning, and in particular to a path planning method for a photovoltaic robot, a photovoltaic robot, equipment and a medium. Background Art

[0002] Solar photovoltaic panels, also known as solar panels, are devices that convert sunlight directly into electricity. To optimize their efficiency in receiving sunlight, they are typically installed at an angle on structures such as rooftops. However, long-term outdoor exposure can lead to the accumulation of dust and dirt, which significantly reduces the panels' efficiency in converting sunlight into electricity. Therefore, regular cleaning of the panels is crucial to maintaining their high performance. Photovoltaic robots can effectively improve the removal of dust, dirt, and snow from the panels' surfaces by automating cleaning operations, eliminating the high costs and risks of panel damage and overhead work associated with manual operation.

[0003] In the diverse environments of photovoltaic power plants, photovoltaic robots face numerous challenges, such as adapting to PV panels of varying sizes and cleaning tilted panels. To achieve precise path planning, the robots rely on advanced perception modules, which collect environmental data to ensure the reliability of their cleaning paths.

[0004] For example, the Chinese invention patent application publication number CN117032266A discloses a photovoltaic cleaning robot driven by a brushless motor and its path planning method. The robot rotates the photovoltaic panel map coordinate system constructed based on satellite positioning signals, converts the bottom edge of the photovoltaic panel in the photovoltaic panel map to the X-axis of the positioning coordinate system, establishes a mapping relationship between each coordinate position in the photovoltaic panel map and the positioning coordinate system, and thus constructs the traversal path of the photovoltaic cleaning robot in the positioning coordinate system according to the rotated photovoltaic panel map. Each coordinate point in the path thus obtained can correspond one-to-one with the physical coordinate position in the photovoltaic map, and the traversal path of the photovoltaic cleaning robot can be quickly and completely planned directly through a general traversal algorithm. (Safety indicator + posture adjustment). The Chinese invention patent application publication number CN11644809A discloses a photovoltaic robot positioning and navigation method and device based on SLAM technology. The method includes: obtaining an environmental point cloud map of the photovoltaic robot through a rotatable laser radar installed on the photovoltaic robot; obtaining a scene image of the photovoltaic robot through a visual camera installed on the photovoltaic robot; fusing the environmental point cloud map and the scene image at the same time, using SLAM mapping to obtain an omnidirectional three-dimensional photovoltaic site map; based on the omnidirectional three-dimensional photovoltaic site map, combined with the obstacle avoidance sensor installed on the photovoltaic robot, positioning, identifying and navigationally controlling the photovoltaic robot to control the photovoltaic robot to reach the destination. The Chinese invention patent application publication number CN114877892 A discloses a fusion positioning method for a photovoltaic robot, which belongs to the field of robot vision technology, including: obtaining relevant data sources during the movement of the photovoltaic robot through multiple data source acquisition devices installed on the photovoltaic robot; processing part of the data in the obtained relevant data sources to obtain new data; establishing an observation model based on an extended Kalman filter; inputting part of the relevant data sources during the movement of the photovoltaic robot and the new data obtained by processing part of the data in the obtained relevant data sources into the observation model to obtain multiple observation quantities; adjusting relevant parameters of the extended Kalman filter; obtaining the final fusion positioning data of the photovoltaic robot. The present invention inputs the processed GPS data and INS data into the observation model for fusion, which can avoid the saturation and short-term loss of INS data due to robot vibration and the failure of GPS positioning due to occlusion and other reasons.

[0005] However, in complex and diverse environments, photovoltaic robots are still not flexible enough when cleaning photovoltaic arrays, and are prone to problems such as incomplete coverage and excessive energy consumption. Summary of the Invention

[0006] The main purpose of the present invention is to provide a path planning method for a photovoltaic robot, a photovoltaic robot, a device, and a medium. In order to solve the above-mentioned technical problems, the present invention specifically adopts the following technical solutions:

[0007] A first aspect of the present invention is to provide a path planning method for a photovoltaic robot, wherein the photovoltaic robot includes at least two perception modules, wherein the perception modules include a visual perception module, and the method includes:

[0008] S201: obtaining a machine cleaning performance index, and setting an adjacent path spacing between adjacent sub-cleaning paths according to the machine cleaning performance index;

[0009] S202: Acquire a photovoltaic panel map, generate an internal cleaning path based on the photovoltaic panel map and the adjacent path spacing, and discretize the internal cleaning path into a plurality of path points, wherein parameters of the path points include expected postures;

[0010] S203: setting an acceptance path point on each sub-cleaning path of the internal cleaning path, wherein parameters of the acceptance path point include an expected cleaning area;

[0011] S204 controls the photovoltaic robot to execute the internal cleaning path; monitors in real time the posture error between the actual posture acquired by the perception module and the expected posture; obtains the actual cleaning area based on the visual perception module at each acceptance path point, and determines the cleaning performance error based on the actual cleaning area and the expected cleaning area;

[0012] S205: When the posture error is greater than a posture tolerance threshold or the cleaning performance error is greater than a cleaning tolerance threshold, a target internal cleaning path is generated based on the posture error and / or the cleaning performance error, and the photovoltaic robot is controlled to execute the target internal cleaning path.

[0013] In some embodiments, the method further includes: when the posture error is less than a posture tolerance threshold and the cleaning performance error is less than a cleanliness tolerance threshold, controlling the photovoltaic robot to continue executing the internal cleaning path.

[0014] In some embodiments, the target internal cleaning path includes a posture correction path and an optimized cleaning path, and S205 includes: generating a posture correction path of the photovoltaic robot based on the posture error; and / or updating a machine cleaning performance index based on the cleaning performance error; updating an adjacent path spacing between adjacent sub-cleaning paths according to the updated machine cleaning performance index; generating the optimized cleaning path based on the updated adjacent path spacing, and controlling the photovoltaic robot to execute the optimized cleaning path.

[0015] In some embodiments, the S205 also includes: controlling the photovoltaic robot to move according to the posture correction path; obtaining the corrected posture based on the perception module, if the posture error between the corrected posture and the expected posture is less than a preset correction error threshold, the posture correction is successful; controlling the photovoltaic robot to execute the optimized cleaning path.

[0016] In some embodiments, S204 also includes: when the photovoltaic robot reaches the acceptance path point, obtaining the image acceptance data collected by the visual perception module; based on a preset recognition model, identifying the final clean area image in the image acceptance data; and calculating the actual cleaning area corresponding to the final clean area image based on a preset mapping algorithm.

[0017] In some embodiments, the cleaning performance index includes a cleaning force index and a cleaning coverage index, and S201 also includes: determining the minimum number of repeated cleaning times based on the cleaning force index and the expected cleaning standard; and determining the adjacent path spacing between adjacent sub-cleaning paths based on the cleaning coverage index and the minimum number of repeated cleaning times.

[0018] In some embodiments, the method also includes: obtaining the fluctuation frequency of the actual posture; when the fluctuation frequency is not within a preset frequency range, identifying whether there is a high-risk area on the cleaning path based on the environmental data collected by the perception module, and if there is no high-risk area on the cleaning path, controlling the photovoltaic robot to continue executing the internal cleaning path; if there is a high-risk area on the cleaning path, controlling the photovoltaic robot to bypass the high-risk area and continue executing the internal cleaning path.

[0019] A second aspect of the present invention is to provide a photovoltaic robot, comprising:

[0020] A moving device for moving between photovoltaic panels;

[0021] A detection device, mounted on the mobile device, comprising at least two sensing modules, the sensing modules being used to collect motion data or environmental data of the photovoltaic robot;

[0022] a cleaning device, mounted on the mobile device, for cleaning the photovoltaic panel;

[0023] A control system is connected to the mobile device, the detection device and the cleaning device, and is used to implement the steps of the path planning method for the photovoltaic robot provided in any embodiment of the present invention.

[0024] A third aspect of the present invention is to provide a computer device, comprising a memory and a processor;

[0025] The memory is used to store computer programs;

[0026] The processor is used to execute the computer program and implement the steps of the photovoltaic robot path planning method provided in any embodiment of the present invention when executing the computer program.

[0027] The fourth aspect of the present invention is to provide a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor performs the steps of the path planning method for a photovoltaic robot as provided in any embodiment of the present invention.

[0028] Beneficial technical effects:

[0029] The present invention provides a highly adaptable path planning method for photovoltaic robots that improves cleaning efficiency while ensuring cleaning quality. Adapting to various photovoltaic robots with different cleaning ranges and capabilities, the method flexibly determines the spacing between adjacent paths, allowing for customized generation of internal cleaning paths that meet the desired cleaning effect. This ensures that cleaning path planning can be scalably and precisely adapted to the needs of different cleaning tasks. Furthermore, during the photovoltaic panel cleaning process, the execution of the cleaning task is monitored in real time from two dimensions: path deviation and panel cleaning effectiveness. Different processing modes are adopted for varying degrees of path execution error.

[0030] One is the slow response mode. When cleaning the internal area of ​​the photovoltaic panel array, the photovoltaic robot's movement in this area is relatively safe, and in order to ensure the cleaning quality, some areas will be cleaned repeatedly. As a result, the photovoltaic robot enjoys a certain degree of freedom in path execution and can tolerate path execution errors that do not affect the cleaning quality and movement safety, avoiding frequent adjustments to the cleaning path, reducing unnecessary repetitions of the cleaning path, improving the photovoltaic robot's execution speed and cleaning efficiency of the cleaning path, and saving energy and time.

[0031] The second is a rapid response mode. When cleaning quality and travel safety cannot be effectively guaranteed, the position posture and path spacing are adjusted promptly based on real-time data to improve the effectiveness and flexibility of the cleaning task. For example, the posture error generated by real-time monitoring generates a posture correction path to improve the reliability and safety of the operation; the cleaning results of fixed-point monitoring are used to adaptively adjust the spacing between adjacent paths to generate an optimized cleaning path, ensuring that the cleaning work meets the expected standards, improving overall cleaning quality, and thus enhancing the sustainability of the cleaning task.

[0032] Furthermore, the fluctuation frequency of the actual posture is evaluated to identify position mutations caused by high-risk areas (such as damaged areas and slipping areas), and the cleaning path is flexibly adjusted. Abnormal fluctuations in data within the tolerance range of path execution error are additionally monitored to adapt to different environmental conditions and photovoltaic panel conditions, thereby improving the adaptability, safety and reliability of the photovoltaic robot. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the embodiments or the description of the prior art. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the various elements or parts are not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without inventive work.

[0034] Figure 1 Schematic diagrams of two array arrangements of photovoltaic panels provided by embodiments of the present invention;

[0035] Figure 2 is a schematic diagram of a photovoltaic panel map provided by an embodiment of the present invention;

[0036] Figure 3 is a schematic flow chart of a photovoltaic array cleaning method provided in an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram of a cleaning path of a photovoltaic robot provided by an embodiment of the present invention;

[0038] Figure 5 This is a partial schematic diagram of a cleaning path of a photovoltaic robot provided by an embodiment of the present invention;

[0039] Figure 6 is a schematic flow chart of a path planning method for a photovoltaic robot provided in an embodiment of the present invention;

[0040] Figure 7 This is a schematic diagram of an acceptance path point provided by an embodiment of the present invention;

[0041] Figure 8 It is a schematic block diagram of the structure of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0043] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0044] Herein, suffixes such as "module," "component," or "unit" used to represent elements are only used to facilitate description of the present invention and have no specific meaning. Therefore, "module," "component," or "unit" may be used interchangeably.

[0045] As used herein, terms such as "upper," "lower," "inner," "outer," "front," "back," "one end," and "the other end" indicate positions or locations based on those shown in the accompanying drawings. These terms are intended solely to facilitate and simplify the description of the present invention and are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0046] As used herein, unless otherwise expressly specified or limited, the terms "installed," "provided with," and "connected" should be understood broadly. For example, "connected" may refer to a fixed connection, a detachable connection, or an integral connection; it may refer to a mechanical connection, a direct connection, an indirect connection via an intermediate medium, or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention on a case-by-case basis.

[0047] As used herein, "and / or" includes any and all combinations of one or more of the associated listed items.

[0048] Herein, "plurality" means two or more than two, ie, it includes two, three, four, five, etc.

[0049] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0050] In order to adapt to the complex environment of photovoltaic power stations, photovoltaic robots need to call a large number of perception modules during the cleaning process to collect and process a large amount of data in real time, and use these data to ensure that they are on the planned path to ensure that the photovoltaic robots clean the photovoltaic panels according to the planned path. This process will consume a lot of electricity.

[0051] Furthermore, when the PV robot deviates from its path, it must be immediately corrected to the planned path to ensure safety and cleaning coverage during high-altitude operations. Frequent posture and position corrections further exacerbate energy consumption and reduce the efficiency of path execution. Hereinafter, posture and position are collectively referred to as pose.

[0052] Therefore, while ensuring safety, this invention optimizes the photovoltaic array cleaning strategy and the photovoltaic robot's path planning algorithm. It also refines the photovoltaic robot's energy consumption management, improving its flexibility, adaptability, and energy efficiency, ensuring it can stably and efficiently complete cleaning tasks in a variety of complex environments. This will not only help improve the operation and maintenance of photovoltaic power plants, but will also promote the further development and application of photovoltaic technology.

[0053] A photovoltaic array is a DC power generation unit composed of multiple photovoltaic modules or panels mechanically and electrically assembled together with a fixed support structure. The photovoltaic panel map in the embodiments of the present invention is a photovoltaic panel layout diagram or photovoltaic panel array diagram, which is a map used to record the arrangement and layout of photovoltaic panels.

[0054] In some embodiments, photovoltaic panel layout information input by a user is obtained, and a photovoltaic panel map is generated based on the photovoltaic panel layout information. The photovoltaic panel layout information is a key parameter in photovoltaic panel array installation and may include photovoltaic panel size information, panel tilt information, inter-panel gap information, array arrangement information, and the like.

[0055] Specifically, PV panel size information refers to the physical dimensions of the panels, including length, width, and thickness. Panel tilt information refers to the angle between the panels and the horizontal plane, also known as the installation inclination. Inter-panel gap information refers to the size of the gap between panels. Array layout information refers to the arrangement of the panels within the array.

[0056] For example, the array arrangement information includes the distribution direction of the photovoltaic panels, see Figure 1 , Figure 1 Schematic diagram of two array arrangements of photovoltaic panels provided by the embodiment of the present invention, such as Figure 1 As shown, photovoltaic panels can be placed vertically or horizontally.

[0057] For example, if the user inputs the photovoltaic panel size information "the height of the photovoltaic panel is 1600mm, the width of the photovoltaic panel is 800mm", the inter-panel gap information "the gap between two photovoltaic panels is 40mm", and the array arrangement information "the vertical arrangement of 4 rows and 16 columns of photovoltaic panels", please refer to Figure 2 , Figure 2 This is a schematic diagram of a photovoltaic panel map provided by an embodiment of the present invention. According to the photovoltaic panel layout information in this example, the following can be mapped and generated: Figure 2 Photovoltaic panel map shown.

[0058] In some embodiments, the photovoltaic robot includes at least two sensing modules. The sensing modules are used to enable the photovoltaic robot to perceive its surrounding environment, collect motion data and environmental data, and thereby understand the actual position of the robot and the cleaning status of the photovoltaic panels. The sensing modules can be various types of visual sensing modules, speed sensing modules, displacement sensing modules, etc. Specifically, they can include wheel odometers, inertial measurement units, visual sensors, ultrasonic sensors, environmental sensors, satellite positioning systems, lidar, etc.

[0059] The present invention proposes a highly adaptable photovoltaic array cleaning method that takes into account both cleaning effect and endurance. Based on the algorithm-generated internal cleaning path, the boundary compensation cleaning path is extended to explore, and the cleaning strategy for the internal area and boundary area of ​​the photovoltaic panel array is adaptively customized. All perception modules are used for precise boundary exploration at the first boundary cleaning station in the trust interval area, and some perception modules are used for simple edge exploration at other boundary cleaning stations. By distinguishing between the main edge exploration cleaning station and the secondary edge exploration cleaning station, a differentiated perception module scheduling strategy is implemented. While ensuring high coverage and high cleanliness of the photovoltaic panel cleaning, the repetitive work in the edge exploration process is reduced, energy is saved, and the resource utilization of the photovoltaic robot is optimized, effectively improving the cleaning efficiency and sustainable operation capability of the photovoltaic robot, and ensuring the long-term stable operation and maximum energy output of the photovoltaic power station.

[0060] See also Figure 3 , Figure 3 is a schematic flow chart of a photovoltaic array cleaning method provided by an embodiment of the present invention, such as Figure 3 As shown, the method is applied to a photovoltaic robot and includes steps S101 to S104.

[0061] S101 obtains a photovoltaic panel map and generates an internal cleaning path based on the photovoltaic panel map. The internal cleaning path passes through a plurality of cleaning stations, including a boundary cleaning station with a preset distance value from any boundary to be cleaned in the photovoltaic panel map.

[0062] The internal cleaning path consists of multiple sub-cleaning paths, each covering an inner area of ​​the PV panel map that is relatively far from the boundary. To ensure that the internal cleaning path fully covers the inner area, it passes through several cleaning stations. Cleaning stations are the speed change or turning points of the PV robot along the internal cleaning path. Correspondingly, boundary cleaning stations are cleaning stations that change speed or turn near any boundary to be cleaned on the PV panel map.

[0063] Specifically, the photovoltaic panel map is divided into an internal area and a boundary area by a preset distance value, and then the cleaning strategy for the internal area and boundary area of ​​the photovoltaic panel array is adaptively customized. The specific value of the preset distance value can be flexibly determined according to actual needs. For example, the preset distance value can be 50 cm, and a virtual internal area boundary is formed at a distance of 50 cm from the boundary. The area within the virtual internal area boundary is the internal area, and the area outside the virtual internal area boundary is the boundary area. Then, based on the path planning algorithm, an internal cleaning path is generated for the internal area of ​​the photovoltaic panel map. Correspondingly, the cleaning station on the virtual internal area boundary is the boundary cleaning station. The path planning algorithm can refer to the relevant technology and is not limited here.

[0064] It should be understood that when executing an internal cleaning path, the photovoltaic robot can only clean the inner area. Within this area, the photovoltaic robot is away from the boundary, without the risk of falling from height, and travel is relatively safe. Furthermore, the photovoltaic panels in the inner area are installed more neatly and regularly than at the boundary, with fewer obstacles, and the corresponding map is more accurate. Therefore, the inner area eliminates most complex interference factors in the environment, making the path planning algorithm more reliable and practical in this area.

[0065] S102 controls the photovoltaic robot to execute the internal cleaning path, and when arriving at the boundary cleaning station, detects whether there is a main edge cleaning station for the same boundary to be cleaned within the trust distance area.

[0066] The photovoltaic robot follows the internal cleaning path to complete the cleaning of the internal area. When it reaches a boundary cleaning station, it cleans the boundary area. It first determines the boundary to be cleaned within a preset distance value of the boundary cleaning station. Then, it detects whether there is a primary edge cleaning station within the trust distance of the boundary cleaning station to determine the perception module scheduling strategy for the boundary cleaning station.

[0067] The trust distance area is defined as the area extending from the current boundary cleaning station along the same boundary into the cleaned area by a preset trust distance. This preset trust distance can be flexibly set based on actual conditions. For example, the trust distance area can be set to 1 meter, which is the area extending from the boundary cleaning station along the same boundary into the cleaned area by 1 meter. It should be noted that the trust distance area is the cleaned area. The main edge cleaning station is searched within the path previously executed by the photovoltaic robot, which improves the reliability of the reference cleaning path.

[0068] For example, the trust distance area is the length area or width area of ​​a single photovoltaic panel laid out on the photovoltaic panel map. Since the shape of the photovoltaic panel is a regular rectangle, according to the arrangement of the photovoltaic panel array, the long side or wide side of the rectangle is selected as the trust distance area, such as Figure 2 As shown in the figure, when the photovoltaic panels are placed vertically, the trust spacing area is the length area of ​​a single photovoltaic panel.

[0069] It should be understood that the trust interval region is defined based on the structural characteristics and layout patterns of the photovoltaic panels in the photovoltaic panel map. Multiple boundary cleaning stations within this interval region have a high degree of similarity with the boundary cleaning paths of the boundary to be cleaned. Therefore, multiple secondary edge cleaning stations near the same boundary within the trust interval region can trust the reference cleaning path of the primary edge cleaning station for simple edge cleaning. In other words, only one deep edge cleaning can be performed, and other boundary cleaning stations can refer to the deep edge cleaning data to quickly, accurately, and efficiently complete the cleaning of the boundary area, thus improving the safety of simple edge cleaning during boundary cleaning of the photovoltaic cleaning machine.

[0070] In step S103 , if there is no main edge detection cleaning station for the same boundary to be cleaned, the boundary cleaning station is set as the main edge detection cleaning station, all the perception modules are enabled to perform deep edge detection on the boundary to be cleaned, and a reference cleaning path for the main edge detection cleaning station is generated.

[0071] Specifically, if there is no main edge cleaning station within the trust distance area, the current boundary cleaning station is set as the main edge cleaning station, and all perception modules are enabled to perform deep edge detection on the boundary to be cleaned, achieving high-precision boundary recognition and posture speed control, ensuring full coverage and cleaning of the boundary during the edge detection process, while ensuring the safety of the photovoltaic robot, and generating a reference cleaning path based on the deep edge detection data. It should be understood that there are multiple reference cleaning paths for the boundary to be cleaned, and each reference cleaning path is only effective when the boundary to be cleaned is within a certain area, that is, when arriving at the boundary cleaning station, when the main edge cleaning station for the same boundary to be cleaned is detected within the trust distance area, the reference cleaning path of the main edge cleaning station is effective.

[0072] It should be noted that all the perception modules enabled in this embodiment are all the perception modules used for boundary recognition and posture and speed control.

[0073] In some embodiments, S103 includes: controlling the photovoltaic robot to move toward the boundary to be cleaned at a preset edge detection value based on a first speed, wherein the preset edge detection value is greater than the preset distance value; enabling all the perception modules to collect multiple environmental data in real time, performing boundary identification based on the multiple environmental data, and controlling the photovoltaic robot to stop moving when the boundary is identified; obtaining the moving path of the photovoltaic robot as a reference cleaning path for the main edge detection cleaning station.

[0074] The first speed is determined by the remaining distance between the PV robot and the boundary when it detects the boundary, and the buffer distance required for the PV robot to brake. When the PV robot is traveling at the first speed and detects the boundary, it can stop before reaching the edge safety distance. The specific speed value can be flexibly set according to actual needs and is not limited here. For example, if the remaining distance is 18cm, the buffer distance is 10cm, and the edge safety distance is 8cm, then the PV robot can stop 8cm from the edge of the boundary when traveling at the first speed, ensuring the safety of the PV robot.

[0075] The preset edge detection value is greater than the preset distance value. This allows the robot to identify map anomalies when the PV panel map does not match the actual PV panel array, allowing it to fully clean the PV panels while ensuring safety, thereby improving the coverage and cleanliness of the PV panel cleaning. For example, if the preset edge detection value is set to 100cm and the preset distance value is 50cm, all perception modules are enabled during deep edge detection, enabling accurate detection of boundaries and allowing the PV robot to freely explore within a limited range of 50cm.

[0076] In some embodiments, the boundary detection of the depth detection can be performed by the visual perception module through image recognition and the infrared sensing perception module through cliff detection.

[0077] In step S104, if there is a main edge detection cleaning station for the same boundary to be cleaned, the boundary cleaning station is set as a secondary edge detection cleaning station, the photovoltaic robot is controlled to obtain and execute the reference cleaning path of the main edge detection cleaning station, and part of the perception module is enabled to perform simple edge detection on the boundary to be cleaned.

[0078] Specifically, if there is a main edge cleaning station within the trust distance area, the current boundary cleaning station is set as a slave edge cleaning station, and the photovoltaic robot is controlled to execute the reference cleaning path of the main edge cleaning station to reduce repetitive work in the edge detection process, and some perception modules are enabled to perform simple edge detection on the boundary to be cleaned, avoiding safety risks caused by local damage to the photovoltaic panels, and achieving high-precision, low-risk, short-time, and low-power efficient boundary cleaning.

[0079] It should be understood that there can be at most one primary edge cleaning station within the trust distance area. When the trust distance area is 1m, if the first boundary cleaning station detects that there are no primary edge cleaning stations within 1m for the same boundary to be cleaned, the first boundary cleaning station will be set as the first primary edge cleaning station. Correspondingly, any second boundary cleaning stations within 1m of the same boundary will detect the first primary edge cleaning station and be set as secondary edge cleaning stations. Furthermore, if the photovoltaic robot reaches a third boundary cleaning station 1m away, the first primary edge cleaning station is not within the trust distance area of ​​the third boundary cleaning station, and therefore the third boundary cleaning station is set as the second primary edge cleaning station.

[0080] In some embodiments, after the simple edge detection is completed at the edge detection cleaning station, an actual cleaning path for the edge detection cleaning station is generated.

[0081] It should be noted that some of the perception modules enabled in this embodiment are those used for boundary recognition and posture speed control. One or more can be selected according to actual needs. The purpose is to use only specific perception modules for simple edge detection to achieve energy saving and consumption saving, optimize the resource utilization of the photovoltaic robot, and effectively improve the sustainable operation capability of the photovoltaic robot.

[0082] In some embodiments, S104 also includes: controlling the photovoltaic robot to execute the reference cleaning path based on the second speed, enabling any perception module to collect environmental data in real time, and performing boundary identification based on the environmental data; when the boundary is identified and / or the reference cleaning path is completed, controlling the photovoltaic robot to stop moving.

[0083] The second speed is determined by the remaining distance between the PV robot and the boundary when it detects it, and the buffer distance required for the PV robot to brake. The second speed can be greater than or equal to the first speed. When the PV robot is traveling at the second speed and detects the boundary, it can at least stop before reaching the edge. The specific speed value can be flexibly set according to actual needs and is not limited here.

[0084] For example, if the remaining distance is 18 cm and the buffer distance is 10 cm, the photovoltaic robot can stop at least when it is close to the edge of the boundary when traveling at the second speed, thereby ensuring the safety of the photovoltaic robot.

[0085] In some embodiments, the photovoltaic robot travels at a third speed when performing an internal cleaning path, with both the first and second speeds being less than the third speed. For example, the third speed is 0.4 m / s, the first speed is 0.1 m / s, and the second speed is 0.2 m / s. By adapting different speeds to different cleaning paths, the photovoltaic robot's cleaning efficiency is effectively improved while ensuring its safety.

[0086] In some embodiments, the boundary detection of the simple edge detection can be performed by the visual perception module through image recognition of the boundary and the infrared sensing perception module through cliff detection to recognize the boundary.

[0087] Therefore, by distinguishing between the main edge cleaning station and the slave edge cleaning station, a differentiated perception module scheduling strategy is implemented, the cleaning path is optimized, and repetitive work is reduced. The photovoltaic robot can dynamically adjust the cleaning path according to the actual environment and cleaning needs to ensure the efficiency and coverage of the cleaning work.

[0088] See also Figures 4 and 5 , Figure 4 This is a schematic diagram of a cleaning path of a photovoltaic robot provided by an embodiment of the present invention. Figure 5 This is a partial schematic diagram of a cleaning path of a photovoltaic robot provided by an embodiment of the present invention. Figure 4 、 Figure 5 As shown, the dotted line is the cleaning path, including the internal cleaning path of the internal area and the boundary cleaning path of the boundary area, wherein the internal cleaning path includes multiple sub-cleaning paths and multiple transition paths connecting the multiple sub-cleaning paths, and the boundary cleaning path includes the reference cleaning path of the main edge cleaning station.

[0089] In addition, the boundary cleaning path may further include an actual cleaning path from the edge detection cleaning station, a cleaning path of an independent edge detection station, and a verification cleaning path of a verification edge detection station, which are not shown in the figure.

[0090] An embodiment of the present invention also provides an intelligent switching mechanism for edge detection strategies to cope with diverse boundary scenarios caused by complex photovoltaic panel layouts and photovoltaic panel damage. While simplifying the use of perception modules, it avoids security issues caused by data uniformity or accuracy limitations, thereby improving the safety of simple edge detection in boundary cleaning of photovoltaic cleaning machines.

[0091] In some embodiments, a first-level pre-inspection is performed on the environmental data collected by the pre-inspection perception module, that is, the environmental matching degree between the main edge cleaning site and the slave edge cleaning site is judged based on the site pre-inspection data and the site actual data collected by the pre-inspection perception module, and whether to execute the reference cleaning path is determined based on the environmental matching degree.

[0092] The method further includes: when cleaning the main edge detection site, enabling the pre-inspection sensing module, obtaining the environmental data collected by the pre-inspection sensing module and storing it as site pre-inspection data; when cleaning the secondary edge detection site, enabling the pre-inspection sensing module, and using the environmental data collected by the pre-inspection sensing module as site actual data; comparing the data error between the site pre-inspection data and the site actual data, and if the data error is less than a preset environmental error threshold, controlling the photovoltaic robot to execute the reference cleaning path.

[0093] The pre-inspection perception module can be a visual perception module, a laser radar, an ultrasonic sensor, an infrared sensor, etc., and can be flexibly determined based on the perception module equipped by the photovoltaic robot. A perception module that is more easily able to determine the similarity of environmental data in adjacent locations can be selected. This embodiment uses the visual perception module as an example for description.

[0094] Specifically, at the main edge cleaning site, the pre-inspection perception module is enabled, which is responsible for collecting the site pre-inspection data of the boundary cleaning site, such as photovoltaic panel characteristics, surrounding environment characteristics, photovoltaic panel boundary characteristics, etc. The environmental data collected by the pre-inspection perception module is stored as site pre-inspection data for subsequent path planning and verification. When at the slave edge cleaning site, the pre-inspection perception module is enabled again to collect the actual site data of the current environment. The site pre-inspection data is compared with the actual site data, and the data error between the two is calculated. If the data error is less than the preset environmental error threshold, it means that the environments of the main edge cleaning site and the slave edge cleaning site match, the reference cleaning path can be trusted, and the photovoltaic robot is controlled to perform the cleaning task according to the reference cleaning path. Among them, the preset environmental error threshold can be flexibly set according to actual needs and is not limited here.

[0095] In some embodiments, when the pre-check fails, deep edge detection is performed at the cleaning station, and the reliability of the reference cleaning path is verified again based on the environmental data and the cleaning path during the deep edge detection process.

[0096] Specifically, after comparing the data error between the site pre-inspection data and the site actual data, it also includes: if the data error is greater than the preset environmental error threshold, updating the edge detection cleaning site to a verification edge detection site, replacing other perception modules as verification perception modules, obtaining the environmental data collected by the verification perception module and storing it as site verification data; performing deep edge detection on the boundary to be cleaned, generating a verification cleaning path for the verification edge detection site; comparing the verification cleaning path with the reference cleaning path to obtain a path error; if the path error is less than the preset path error threshold, retaining the reference cleaning path, updating the verification perception module to a pre-inspection perception module, and updating the site verification data to the site pre-inspection data. Among them, the preset path error threshold can be flexibly set according to actual needs and is not limited here.

[0097] It should be understood that replacing other perception modules as verification perception modules can avoid misjudgment due to pre-inspection perception module failure and eliminate data interference from the pre-inspection perception module. When the path error is confirmed to be less than the preset path error threshold, it is confirmed that the pre-inspection perception module has detected an error. Replacing the site pre-inspection data and pre-inspection perception module avoids security issues caused by data uniqueness or accuracy limitations and improves the accuracy of reliability testing.

[0098] In some embodiments, after the simple edge detection is completed, a secondary re-inspection is performed on the actual cleaning path from the edge detection cleaning station, that is, the similarity between the actual cleaning path and the reference cleaning path is compared, and whether the reference cleaning path is retained is determined based on the similarity.

[0099] The method includes: controlling the photovoltaic robot to stop moving, obtaining an actual cleaning path of the photovoltaic robot; comparing the actual cleaning path with a reference cleaning path to obtain a path error; and retaining the reference cleaning path if the path error is less than a preset path error threshold. The preset path error threshold can be flexibly set according to actual needs and is not limited here.

[0100] It should be understood that before the reference cleaning path is completed, the perception module recognizes the boundary in advance and controls the photovoltaic robot to stop moving, resulting in errors. There are two specific reasons. One is that the reference cleaning path is accurate, but due to different sensor recognition sensitivities or unstable data, the boundary is recognized in advance and the photovoltaic robot is controlled to stop moving; the other is that the photovoltaic panel is damaged, resulting in an inaccurate reference cleaning path.

[0101] Therefore, by comparing the actual cleaning path with the reference cleaning path, the path error is calculated, and the magnitude of the path error is used to distinguish the above two causes and confirm whether the reference cleaning path is trustworthy.

[0102] In some embodiments, the primary pre-check scheme and the secondary re-check scheme may be enabled simultaneously, or only one of the checks may be enabled to confirm whether the reference cleaning path is trustworthy.

[0103] In some embodiments, the method further includes: if the path error is greater than the preset path error threshold, deleting the reference cleaning path, and updating all boundary cleaning stations within the preset spacing area to independent edge detection stations; when reaching the independent edge detection station, enabling all the perception modules to perform deep edge detection on the boundary to be cleaned, and generating a cleaning path for each independent edge detection station.

[0104] The preset spacing area is the area extending from the current boundary cleaning station along the same boundary to the uncleaned area by a preset untrusted distance. The preset untrusted distance can be flexibly set based on actual conditions. For example, the preset spacing area can be set to 1 meter, that is, the area extending from the boundary cleaning station along the same boundary to the uncleaned area by 1 meter.

[0105] It should be understood that when the reliability of the reference cleaning path is found to be in doubt in the first-level pre-inspection plan and the second-level re-inspection plan, the differentiated perception module scheduling strategy will be suspended within the preset spacing area, and all perception modules will be used for precise boundary exploration, with safety as the priority, to ensure cleaning coverage and improve the reliability and practicality of the photovoltaic robot.

[0106] Exemplarily, the preset spacing area is the length or width of a single photovoltaic panel as laid out on the photovoltaic panel map. Since photovoltaic panels are rectangular in shape, the long side or wide side of the rectangle is selected as the preset spacing area based on the arrangement of the photovoltaic panels. If the reliability of the reference cleaning path is questionable, the differentiated perception module scheduling strategy is suspended within a photovoltaic panel area, and the size of the preset non-trust distance is reasonably limited while ensuring safety.

[0107] In some embodiments, the present invention provides a highly adaptable path planning method for photovoltaic robots that improves cleaning efficiency while ensuring cleaning quality. Adapting to various photovoltaic robots with different cleaning ranges and capabilities, the method flexibly determines the spacing between adjacent paths, allowing for customized generation of internal cleaning paths that meet the desired cleaning effect. This ensures that cleaning path planning can be scalably and precisely adapted to the needs of different cleaning tasks. Furthermore, during the photovoltaic panel cleaning process, the execution of the cleaning task is monitored in real time from two dimensions: path deviation and panel cleaning effectiveness. Different processing modes are adopted for different degrees of path execution error.

[0108] In some embodiments, the photovoltaic robot includes at least two perception modules, wherein the perception module includes a visual perception module. Figure 6 , Figure 6 is a schematic flow chart of a path planning method for a photovoltaic robot provided by an embodiment of the present invention, such as Figure 6 As shown, it includes steps S201 to S205.

[0109] S201 obtains a machine cleaning performance index, and sets an adjacent path spacing between adjacent sub-cleaning paths according to the machine cleaning performance index.

[0110] Cleaning performance indicators refer to one or more parameters used to evaluate the cleaning efficiency of a photovoltaic robot. These parameters are then used to comprehensively determine the spacing between adjacent sub-cleaning paths. The smaller the spacing, the more times the photovoltaic robot repeatedly cleans the same area, resulting in better cleaning results. Cleaning efficiency is reflected in aspects such as cleaning power and endurance.

[0111] In some embodiments, the cleaning performance indicator includes a cleaning force indicator and a cleaning coverage indicator. S201 further includes: determining a minimum number of repeated cleanings based on the cleaning force indicator and the expected cleaning standard; and determining the adjacent path spacing between adjacent sub-cleaning paths based on the cleaning coverage indicator and the minimum number of repeated cleanings.

[0112] Among them, the cleaning power index refers to the degree of dirt removal that the photovoltaic robot can achieve in a single cleaning. It is one of the parameters for evaluating the cleaning ability of the photovoltaic robot. The higher the cleaning power, the fewer cleaning times can be used to achieve the same cleaning effect.

[0113] The cleaning coverage rate refers to the area that the cleaning device of the photovoltaic robot can cover. A larger coverage area means that cleaning work on a large area can be completed faster.

[0114] Among them, the expected cleaning standards may include the degree of cleanliness, the area that can be cleaned on a single charge, and other requirements put forward by users from different angles for the entire process of photovoltaic panel cleaning.

[0115] For example, based on the cleaning power index, the minimum number of repeated cleanings required to achieve the expected cleaning standard for the same area can be calculated. Based on the cleaning coverage index and the minimum number of repeated cleanings, the distance between each sub-cleaning path, that is, the spacing between adjacent paths, can be calculated so that after the photovoltaic panel is cleaned the minimum number of repeated cleanings, the cleanliness of the photovoltaic panel surface can reach the expected cleaning standard, thereby improving the overall cleaning quality.

[0116] For example, if the cleaning force required for a complete and thorough cleaning is 100%, the cleaning force index of the photovoltaic robot is 40%, and the expected cleaning standard cleanliness is 80%, then the minimum number of repeated cleanings is 2. When the cleaning coverage index is 0.4m, the adjacent path spacing can be set to 0.2m so that each area to be cleaned is cleaned twice. It should be understood that the area where the cleaning coverage index exceeds the adjacent path spacing is the area of ​​the repeated cleaning area.

[0117] In some embodiments, the cleaning performance index includes a battery life index. Based on the battery life index and photovoltaic panel map of the photovoltaic robot, the power required for the photovoltaic robot to clean the current photovoltaic panel map can be calculated. According to the number of charging times corresponding to the expected cleaning standard, the minimum path spacing between each sub-cleaning path is calculated; when the minimum path spacing is greater than the cleaning coverage index, an insufficient cleaning signal is generated to remind the user.

[0118] Furthermore, when the minimum path spacing is less than the cleaning coverage index, the adjacent path spacing is determined based on the cleaning coverage index, the cleaning force index and the minimum path spacing, and the final adjacent path spacing is greater than the minimum path spacing to take into account both energy consumption and cleaning quality.

[0119] It should be understood that developing different cleaning paths based on cleaning performance indicators can effectively adapt to various types of photovoltaic robots with different cleaning ranges and cleaning capabilities. For example, as the number of uses increases, the photovoltaic robot may experience varying degrees of wear and tear. For example, if the cleaning ability decreases, repeated cleaning is needed to ensure cleaning results, and a smaller spacing between adjacent paths can be set accordingly. For another example, if the battery life decreases, repeated cleaning is needed to reduce unnecessary energy consumption, and a larger spacing between adjacent paths can be set accordingly. In this way, the cleaning path of the photovoltaic robot can be planned with high adaptability, while taking into account both cleaning quality and battery life.

[0120] S202 obtains a photovoltaic panel map, generates an internal cleaning path based on the photovoltaic panel map and the adjacent path spacing, and discretizes the internal cleaning path into a plurality of path points, wherein parameters of the path points include expected postures.

[0121] The internal cleaning path consists of multiple sub-paths, each covering an internal area of ​​the PV panel map that is relatively far from the panel boundary. To achieve precise position control, the internal cleaning path is discretized into several path points, for example, with a distance of approximately 5 mm between them. Each path point is assigned appropriate path point parameters such as position, posture, speed, and angular velocity based on the cleaning task. These parameters are used to determine in real time whether the PV robot has deviated from the path and the degree of deviation, ensuring smooth and efficient movement.

[0122] For example, based on the path planning algorithm and the adjacent path spacing, an internal cleaning path is generated for the internal area of ​​the photovoltaic panel map. The path spacing between each sub-cleaning path of the internal cleaning path is the adjacent path spacing. This allows for customized generation of an internal cleaning path that meets the desired cleaning effect, ensuring that cleaning path planning can be scalably and accurately adapted to different cleaning tasks and different photovoltaic robots. For details, please refer to the description of other embodiments of the present invention and will not be repeated here.

[0123] It should be understood that the path points and cleaning stations in the embodiments of the present invention are different. To fully clean the photovoltaic panel array, the internal cleaning path must fully cover the internal area. Therefore, multiple turning or speed change locations are present in the cleaning path. Cleaning stations are set up at these locations to control the timely speed or turning of the photovoltaic robot, thereby accurately executing the internal cleaning path. Furthermore, to ensure smooth movement of the photovoltaic robot, speed changes or turns are completed gradually during the movement. Correspondingly, the parameters of the path points around the cleaning stations are generated based on the speed or direction to be adjusted, ensuring that the photovoltaic robot can smoothly and accurately complete speed changes or turns.

[0124] For example, the photovoltaic robot changes speed at cleaning station A, accelerating from 0.1m / s to 0.4m / s. The corresponding cleaning path within the preset speed change distance before cleaning station A or the preset speed change distance after cleaning station A is the acceleration path. The acceleration path is discretized into path points q and p. The photovoltaic robot passes through path points q and p in sequence. Correspondingly, the speed parameter value of path point q is 0.2m / s, and the speed parameter value of path point p is 0.3m / s. Therefore, when the photovoltaic robot reaches cleaning station A after traveling the preset speed change distance, it accelerates to 0.4m / s; or after traveling the preset speed change distance from cleaning station A, the photovoltaic robot accelerates to 0.4m / s.

[0125] S203 sets an acceptance path point on each sub-cleaning path of the internal cleaning path, wherein parameters of the acceptance path point include an expected cleaning area.

[0126] The acceptance path point is the location where the cleaning quality is checked during the cleaning process. The expected cleaning area is the area expected to have been cleaned in the image captured at the current acceptance path point.

[0127] For example, to improve image availability, the relative angles between acceptance path points are smaller than a preset angle value, ensuring that the angle differences between acceptance path points are controllable, thereby ensuring the continuity and coverage of images between adjacent sub-cleaning paths, and ensuring that every part of the photovoltaic panel can be effectively cleaned and inspected, thereby ensuring that the cleaning quality of the photovoltaic panel meets the expected standards and improving the reliability and effectiveness of the entire cleaning process. Figure 7 , Figure 7 This is a schematic diagram of an acceptance path point provided by an embodiment of the present invention. Figure 7 As shown, the relative angles between the acceptance path points are close to zero.

[0128] S204 controls the photovoltaic robot to execute the internal cleaning path; monitors in real time the posture error between the actual posture acquired by the perception module and the expected posture; at each acceptance path point, obtains the actual cleaning area based on the visual perception module, and determines the cleaning performance error according to the actual cleaning area and the expected cleaning area.

[0129] Specifically, the photovoltaic robot is controlled to operate along a predetermined internal cleaning path while simultaneously monitoring its actual position, as collected by the perception module, in real time. By comparing this actual position data with the expected position data, the position error can be measured. Furthermore, the current cleaning performance of the photovoltaic robot is evaluated at acceptance points along the path using images to determine the actual cleaning area. Based on the actual and expected cleaning areas, any cleaning performance errors in the photovoltaic robot's cleaning performance indicators can be determined.

[0130] That is to say, during the process of the photovoltaic robot executing the internal cleaning path, through real-time analysis of posture errors and fixed-point monitoring of cleaning effectiveness, the execution of the cleaning task is monitored in real time from two dimensions: path deviation and photovoltaic panel cleaning effect. The cleaning strategy of the photovoltaic robot is adjusted appropriately in practice to ensure cleaning quality and travel safety, so as to ensure that the cleaning work of the photovoltaic panel achieves the expected results.

[0131] In some embodiments, when the photovoltaic robot reaches the acceptance path point, the image acceptance data collected by the visual perception module is obtained; based on a preset recognition model, the final clean area image in the image acceptance data is identified; and based on a preset mapping algorithm, the actual cleaning area corresponding to the final clean area image is calculated.

[0132] Specifically, acceptance points are set along each sub-cleaning path. When the photovoltaic robot reaches an acceptance point, it acquires and stores image acceptance data collected by the visual perception module for subsequent cleaning effect evaluation and quality control. A pre-trained recognition model is used to identify the final clean area image from the image acceptance data, showing the complete cleaning path. A pre-set mapping algorithm is then used to calculate the actual cleaned area corresponding to the final clean area image.

[0133] Among them, the preset mapping algorithm is generated based on the proportional relationship between the image area and the actual area.

[0134] Among them, the preset recognition model is a cleanliness recognition model obtained based on machine learning or deep learning technology, which can analyze and classify images. For example, the training process of the preset recognition model uses three different types of photovoltaic panel images: fully clean, partially clean, and unclean. Different cleaning powers are distinguished by parameters such as photovoltaic panel cleanliness, dust adhesion, and transmittance. The corresponding data such as brightness, texture features, and grayscale values ​​of different parameters in the image data are also different. These pictures are used as training data to help the model learn how to identify and distinguish photovoltaic panels with different cleaning powers. The cleanliness recognition model obtained by training can then receive image acceptance data as input and distinguish images of fully clean, partially clean, or uncleaned areas in the current image acceptance data.

[0135] For example, when dust adhesion is low, the transmittance increases, corresponding to a higher degree of cleanliness. Such photovoltaic panels can reflect more sunlight, making the corresponding areas in the image appear brighter. Conversely, when dust adhesion is high, the transmittance decreases, corresponding to a lower degree of cleanliness. Dirt or dust on the photovoltaic panel surface absorbs more light, causing the corresponding areas in the image to appear darker. The shadows and irregular shapes of dust may also cause dark spots or uneven textures in the image.

[0136] It should be understood that in order to adapt to various types of photovoltaic robots with different cleaning ranges and cleaning capabilities, the photovoltaic robot may need to repeatedly clean the same area. The final cleaning area is the area that has been repeatedly cleaned. For example, each area to be cleaned needs to be cleaned twice, then the final cleaning area is the area that has been cleaned twice.

[0137] In some embodiments, a photovoltaic robot path planning method provided by an embodiment of the present invention further includes a slow response step S206: when the posture error is less than the posture tolerance threshold and the cleaning performance error is less than the cleanliness tolerance threshold, controlling the photovoltaic robot to continue executing the internal cleaning path.

[0138] Among them, the posture tolerance threshold can be determined based on the preset distance value between the internal area and the boundary area. For example, when the preset distance value is 50 cm, the risk of falling of the photovoltaic robot increases once it deviates by 50 cm. Therefore, in order to ensure the safe movement of the photovoltaic robot, the posture tolerance threshold is set to 50 cm.

[0139] Furthermore, photovoltaic panels are composed of multiple electronic boards. The installation gaps between the electronic boards will form multiple lines. The visual perception module will collect a limited image and fit the lines in the image into multiple white lines. Furthermore, the limited image includes at least one baseline and two reference lines that are updated in real time. The changes in the photovoltaic robot's posture are determined based on the changes in these lines in the field of view. When the photovoltaic robot's position offset is too large, it is difficult to return to the expected position through the visual perception module. In this case, the posture tolerance threshold can also be set based on the posture calibration capability of the perception module.

[0140] Among them, the cleanliness tolerance threshold can be flexibly set according to actual needs.

[0141] It should be noted that when setting the adjacent path spacing between adjacent sub-cleaning paths based on the machine's cleaning performance indicators, it is often difficult to precisely match the cleaning power indicator with the expected cleaning standard through the number of repeated cleanings. For example, if the cleaning power indicator is 50% and the expected cleaning standard is 80% cleanliness, then in order to ensure cleaning quality, the minimum number of repeated cleanings is 2, and theoretically the actual cleaning level can reach 100%. Therefore, the cleanliness level of the fully cleaned area has exceeded the expected cleaning standard. From an overall perspective, even if there are small areas that are not fully cleaned, the power generation efficiency of the photovoltaic panel has reached a significantly improved level compared to the expected cleaning standard, and the minor cleaning omissions caused by this deviation can be tolerated.

[0142] Furthermore, the redundancy rate is determined based on the minimum number of repeated cleanings, the cleaning power index, and the expected cleaning standard. The cleanliness tolerance threshold can also be flexibly set based on the redundancy rate. For example, the cleanliness tolerance threshold can be set to 0.5 square meters.

[0143] It should be understood that when cleaning the inner area of ​​the photovoltaic panel array, the movement of the photovoltaic robot in this area is relatively safe, and some areas will be cleaned repeatedly to ensure the cleaning quality, so the photovoltaic robot is allowed to have a certain degree of freedom in the execution of the path.

[0144] That is to say, even if there is a path execution error within this degree of freedom, the repeated cleaning strategy can ensure that the cleaning quality is maintained stably, and the regular paving of the internal area and the low risk of falling can ensure the safe movement of the photovoltaic robot. The embodiment of the present invention uses the posture tolerance threshold and the cleanliness tolerance threshold to monitor in real time whether the current path execution error is within this degree of freedom from the two dimensions of path deviation and photovoltaic panel cleaning effect. The path execution error that can be tolerated corresponding to this degree of freedom is not adjusted, avoiding frequent adjustments to the cleaning path, reducing unnecessary repetitions of the cleaning path, improving the execution speed and cleaning efficiency of the photovoltaic robot on the cleaning path, saving energy and time, and at the same time still having reliable guarantees for cleaning quality and movement safety.

[0145] In some embodiments, a path planning method for a photovoltaic robot provided by an embodiment of the present invention includes a quick response step S205: when the posture error is greater than the posture tolerance threshold or the cleaning performance error is greater than the cleanliness tolerance threshold, a target internal cleaning path is generated based on the posture error and / or the cleaning performance error, and the photovoltaic robot is controlled to execute the target internal cleaning path.

[0146] Specifically, if the posture error exceeds a preset posture tolerance threshold, or if the cleaning performance error exceeds a preset cleanliness tolerance threshold, the robot's actual position is deemed to have deviated significantly from the intended path, or the cleaning quality is no longer up to standard. A new target internal cleaning path is generated based on both the posture error and the cleaning performance error to correct the path execution error. The PV robot then continues its cleaning task according to this newly generated target internal cleaning path.

[0147] It should be understood that when cleaning quality and travel safety are difficult to be effectively guaranteed, the path execution exceeds the allowed degrees of freedom, and the position posture and path spacing are adjusted in time according to real-time data to ensure cleaning quality and safety and improve the effectiveness and flexibility of cleaning tasks.

[0148] In some embodiments, the target internal cleaning path includes a posture correction path and an optimized cleaning path, and S205 includes: generating a posture correction path of the photovoltaic robot based on the posture error; and / or updating a machine cleaning performance index based on the cleaning performance error; updating an adjacent path spacing between adjacent sub-cleaning paths according to the updated machine cleaning performance index; generating the optimized cleaning path based on the updated adjacent path spacing, and controlling the photovoltaic robot to execute the optimized cleaning path.

[0149] The posture correction path is a path for correcting posture errors, and the optimized cleaning path is a path obtained by adjusting the distance between adjacent paths of the unexecuted internal cleaning path.

[0150] Specifically, when the posture error exceeds the posture tolerance threshold, a posture correction path is generated based on the posture error of the photovoltaic robot using the path planning algorithm to correct the photovoltaic robot to the expected posture and ensure its safety. When the cleaning performance error exceeds the cleanliness tolerance threshold, an adjacent path spacing that is more suitable for the current application scenario is determined. Based on the path planning algorithm and the updated adjacent path spacing, an optimized cleaning path is generated to ensure that the cleaning quality meets the expected standards.

[0151] In some embodiments, when the posture error is greater than the posture tolerance threshold and the cleaning performance error is less than the clean tolerance threshold, or when the cleaning performance error is greater than the clean tolerance threshold and the posture error is less than the posture tolerance threshold, a posture correction path and an optimized cleaning path can be generated respectively. The posture error and the cleaning performance error can be eliminated through a one-time path adjustment, thereby avoiding subsequent frequent adjustments to the cleaning path, reducing unnecessary repetitions of the cleaning path, and improving cleaning efficiency while ensuring cleaning quality.

[0152] It should be understood that the photovoltaic array cleaning method provided in embodiments of the present invention can dynamically adapt to changes in the cleaning performance of the photovoltaic robot. For example, as the photovoltaic robot completes its cleaning task, its cleaning performance index may gradually decrease, in which case the adjacent path spacing may need to be reduced. For another example, if the acquired machine cleaning performance index is inaccurate, the cleaning performance error exceeding the cleanliness tolerance threshold may indicate that the machine cleaning performance index is too high or too low. Accordingly, the adjacent path spacing may be increased or decreased to reduce repeated cleaning or missed cleaning.

[0153] In some embodiments, the S205 also includes: controlling the photovoltaic robot to move according to the posture correction path; obtaining the corrected posture based on the perception module, if the posture error between the corrected posture and the expected posture is less than a preset correction error threshold, the posture correction is successful; controlling the photovoltaic robot to execute the optimized cleaning path.

[0154] It should be understood that a posture correction path is generated and executed based on the posture error. The matching of the corrected posture with the expected posture is used to determine whether the photovoltaic robot has been corrected to the desired position within the photovoltaic array, thereby ensuring the accuracy of the posture correction. Furthermore, if the posture error between the corrected posture and the expected posture is greater than a preset correction error threshold, other perception modules are called upon to re-perform posture correction until the posture correction is successful.

[0155] In some embodiments, the target internal cleaning path includes a compensatory cleaning path, and S205 also includes: when the cleaning performance error is lower than the clean tolerance threshold, generating and executing a compensatory cleaning path for the area within the image range captured by the acceptance path point to perform compensatory cleaning on the insufficiently cleaned area, ensuring that the cleaning work meets the expected standards and improving the overall cleaning quality.

[0156] In some embodiments, a minimum spacing value of the adjacent path spacing between adjacent sub-cleaning paths is set. When the updated adjacent path spacing is less than the fault spacing value, the cleaning performance index of the photovoltaic robot is too low. Continuing cleaning may result in ineffective cleaning that wastes electricity. The internal cleaning path can be stopped and the user can be reminded.

[0157] In some embodiments, if the high-frequency monitoring shows that the posture error in the path is greater than the posture tolerance threshold or the cleaning performance error is greater than the cleanliness tolerance threshold, considering special circumstances such as map distortion and machine failure, the internal cleaning path is stopped and the user is reminded.

[0158] In some embodiments, the method also includes: obtaining the fluctuation frequency of the actual posture; when the fluctuation frequency is not within a preset frequency range, identifying whether there is a high-risk area on the cleaning path based on the environmental data collected by the perception module, and if there is no high-risk area on the cleaning path, controlling the photovoltaic robot to continue executing the internal cleaning path; if there is a high-risk area on the cleaning path, controlling the photovoltaic robot to bypass the high-risk area and continue executing the internal cleaning path.

[0159] When the photovoltaic robot is moving smoothly and normally, the posture changes should be within the preset frequency range. The fluctuation frequency of the actual posture should be evaluated, and the posture mutations caused by high-risk areas (such as damaged areas and slipping areas) should be identified in combination with the environmental data collected by the perception module. The cleaning path should be flexibly adjusted to bypass the high-risk areas, and additional monitoring should be performed on abnormal fluctuations in data within the tolerance range of path execution errors to adapt to different environmental conditions and photovoltaic panel conditions, thereby improving the adaptability, safety and reliability of the photovoltaic robot.

[0160] Through this sophisticated monitoring and analysis, manual intervention is reduced, operation and maintenance costs are lowered, energy waste can be minimized and the risk of incomplete cleaning can be avoided, thereby improving the economy and sustainability of the photovoltaic power generation system.

[0161] An embodiment of the present invention further provides a photovoltaic robot, comprising:

[0162] A moving device for moving between photovoltaic panels;

[0163] A detection device is installed on the mobile device, and the detection device includes at least two perception modules, and the perception modules are used to collect motion data or environmental data of the photovoltaic robot; wherein, the perception modules may include a visual perception module.

[0164] a cleaning device, mounted on the mobile device, for cleaning the photovoltaic panel;

[0165] A control system, connecting the mobile device, the detection device and the cleaning device, is used to implement the steps of the photovoltaic array cleaning method and / or the photovoltaic robot path planning method provided in any embodiment of the present invention. For details, please refer to the description of the photovoltaic array cleaning method and / or the photovoltaic robot path planning method in other embodiments of the present invention, which will not be repeated here.

[0166] See also Figure 8 , Figure 8 1 is a schematic block diagram of the structure of a computer device provided by an embodiment of the present invention. The computer device may be a server.

[0167] The computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0168] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, which, when executed, can cause the processor to execute any one of the photovoltaic array cleaning methods and / or photovoltaic robot path planning methods.

[0169] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0170] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any one of the photovoltaic array cleaning methods and / or photovoltaic robot path planning methods.

[0171] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 8 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0172] It should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0173] In one embodiment, the method is applied to a photovoltaic robot, wherein the photovoltaic robot includes at least two perception modules, and the processor is configured to run a computer program stored in a memory to implement the following steps:

[0174] S101 obtains a photovoltaic panel map and generates an internal cleaning path based on the photovoltaic panel map, wherein the internal cleaning path passes through a plurality of cleaning stations, and the cleaning stations include boundary cleaning stations with a preset distance value from any boundary to be cleaned on the photovoltaic panel map;

[0175] S102 controls the photovoltaic robot to execute the internal cleaning path, and when arriving at the boundary cleaning station, detects whether there is a main edge cleaning station on the same boundary to be cleaned within the trust distance area;

[0176] S103: if there is no main edge detection cleaning station for the same boundary to be cleaned, setting the boundary cleaning station as the main edge detection cleaning station, enabling all the perception modules to perform deep edge detection on the boundary to be cleaned, and generating a reference cleaning path for the main edge detection cleaning station;

[0177] In step S104, if there is a main edge detection cleaning station for the same boundary to be cleaned, the boundary cleaning station is set as a secondary edge detection cleaning station, the photovoltaic robot is controlled to obtain and execute the reference cleaning path of the main edge detection cleaning station, and part of the perception module is enabled to perform simple edge detection on the boundary to be cleaned.

[0178] In some embodiments, the trust distance area is a length area or a width area of ​​a single photovoltaic panel laid out on the photovoltaic panel map.

[0179] In some embodiments, when implementing the step S103, the processor is also used to implement: controlling the photovoltaic robot to move toward the boundary to be cleaned at a preset edge detection value based on a first speed, wherein the preset edge detection value is greater than the preset distance value; enabling all the perception modules to collect multiple environmental data in real time, performing boundary identification based on the multiple environmental data, and controlling the photovoltaic robot to stop moving when the boundary is identified; obtaining the moving path of the photovoltaic robot as a reference cleaning path for the main edge detection cleaning station.

[0180] In some embodiments, when implementing the step S104, the processor is also used to implement: controlling the photovoltaic robot to execute the reference cleaning path based on the second speed, enabling any perception module to collect environmental data in real time, and performing boundary identification based on the environmental data; when the boundary is identified and / or the reference cleaning path is completed, controlling the photovoltaic robot to stop moving.

[0181] In some embodiments, the method further includes: when at the main edge cleaning site, enabling the pre-inspection sensing module, obtaining the environmental data collected by the pre-inspection sensing module and storing it as site pre-inspection data; when at the slave edge cleaning site, enabling the pre-inspection sensing module, and using the environmental data collected based on the pre-inspection sensing module as site actual data; comparing the data error between the site pre-inspection data and the site actual data, if the data error is less than a preset environmental error threshold, controlling the photovoltaic robot to execute the reference cleaning path; and / or, after controlling the photovoltaic robot to stop moving, obtaining the actual cleaning path of the photovoltaic robot; comparing the actual cleaning path with the reference cleaning path to obtain a path error, if the path error is less than a preset path error threshold, retaining the reference cleaning path.

[0182] In some embodiments, after implementing the comparison of the data error between the site pre-inspection data and the site actual data, the processor is further used to implement: if the data error is greater than the preset environmental error threshold, the edge detection cleaning site is updated to the verification edge detection site, other perception modules are replaced as verification perception modules, and the environmental data collected by the verification perception module is obtained and stored as site verification data; deep edge detection is performed on the boundary to be cleaned to generate a verification cleaning path for the verification edge detection site; the verification cleaning path is compared with the reference cleaning path to obtain a path error; if the path error is less than the preset path error threshold, the reference cleaning path is retained, the verification perception module is updated to the pre-inspection perception module, and the site verification data is updated to the site pre-inspection data.

[0183] In some embodiments, the processor is also used to implement: if the path error is greater than the preset path error threshold, the reference cleaning path is deleted, and all boundary cleaning stations within the preset spacing area are updated to independent edge detection stations; when reaching the independent edge detection station, all the perception modules are enabled to perform deep edge detection on the boundary to be cleaned, and a cleaning path for each independent edge detection station is generated.

[0184] In one embodiment, the photovoltaic robot includes at least two perception modules, which may include a visual perception module. The processor is configured to run a computer program stored in a memory to implement the following steps:

[0185] S201: obtaining a machine cleaning performance index, and setting an adjacent path spacing between adjacent sub-cleaning paths according to the machine cleaning performance index;

[0186] S202: Acquire a photovoltaic panel map, generate an internal cleaning path based on the photovoltaic panel map and the adjacent path spacing, and discretize the internal cleaning path into a plurality of path points, wherein parameters of the path points include expected postures;

[0187] S203: setting an acceptance path point on each sub-cleaning path of the internal cleaning path, wherein parameters of the acceptance path point include an expected cleaning area;

[0188] S204 controls the photovoltaic robot to execute the internal cleaning path; monitors in real time the posture error between the actual posture acquired by the perception module and the expected posture; obtains the actual cleaning area based on the visual perception module at each acceptance path point, and determines the cleaning performance error based on the actual cleaning area and the expected cleaning area;

[0189] S205: When the posture error is greater than a posture tolerance threshold or the cleaning performance error is greater than a cleaning tolerance threshold, a target internal cleaning path is generated based on the posture error and / or the cleaning performance error, and the photovoltaic robot is controlled to execute the target internal cleaning path.

[0190] In some embodiments, the processor is further configured to control the photovoltaic robot to continue executing the internal cleaning path when the posture error is less than a posture tolerance threshold and the cleaning performance error is less than a cleanliness tolerance threshold.

[0191] In some embodiments, the target internal cleaning path includes a posture correction path and an optimized cleaning path. When implementing the step S205, the processor is also used to implement: generating a posture correction path of the photovoltaic robot based on the posture error; and / or updating the machine cleaning performance index based on the cleaning performance error; updating the adjacent path spacing between adjacent sub-cleaning paths according to the updated machine cleaning performance index; generating the optimized cleaning path based on the updated adjacent path spacing, and controlling the photovoltaic robot to execute the optimized cleaning path.

[0192] In some embodiments, when implementing the step S205, the processor is also used to implement: controlling the photovoltaic robot to move according to the posture correction path; obtaining the corrected posture based on the perception module, if the posture error between the corrected posture and the expected posture is less than a preset correction error threshold, the posture correction is successful; controlling the photovoltaic robot to execute the optimized cleaning path.

[0193] In some embodiments, when implementing the step S204, the processor is also used to implement: when the photovoltaic robot reaches the acceptance path point, obtaining the image acceptance data collected by the visual perception module; based on a preset recognition model, identifying the final clean area image in the image acceptance data; and calculating the actual cleaning area corresponding to the final clean area image based on a preset mapping algorithm.

[0194] In some embodiments, the cleaning performance index includes a cleaning force index and a cleaning coverage index. When implementing step S201, the processor is also used to implement: determining the minimum number of repeated cleanings based on the cleaning force index and the expected cleaning standard; determining the adjacent path spacing between adjacent sub-cleaning paths based on the cleaning coverage index and the minimum number of repeated cleanings.

[0195] In some embodiments, the processor is also used to implement: obtaining the fluctuation frequency of the actual posture; when the fluctuation frequency is not within a preset frequency range, identifying whether there is a high-risk area on the cleaning path based on the environmental data collected by the perception module; if there is no high-risk area on the cleaning path, controlling the photovoltaic robot to continue executing the internal cleaning path; if there is a high-risk area on the cleaning path, controlling the photovoltaic robot to bypass the high-risk area and continue executing the internal cleaning path.

[0196] An embodiment of the present invention also provides a computer-readable storage medium, which stores a computer program, and the computer program includes program instructions. The processor executes the program instructions to implement any one of the photovoltaic array cleaning methods and / or photovoltaic robot path planning methods provided in the embodiments of the present invention.

[0197] The computer-readable storage medium may be an internal storage unit of the computer device described in the aforementioned embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a flash memory card, etc., equipped on the computer device.

[0198] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A path planning method for a photovoltaic robot, characterized in that: The photovoltaic robot includes at least two perception modules, each of which includes a visual perception module. The method includes: S201 obtains a machine cleaning performance index and sets an adjacent path spacing between adjacent sub-cleaning paths based on the machine cleaning performance index; the cleaning performance index includes a cleaning force index and a cleaning coverage index; S201 further includes: determining a minimum number of repeated cleanings based on the cleaning force index and an expected cleaning standard; determining the adjacent path spacing between adjacent sub-cleaning paths based on the cleaning coverage index and the minimum number of repeated cleanings; S202: Acquire a photovoltaic panel map, generate an internal cleaning path based on the photovoltaic panel map and the adjacent path spacing, and discretize the internal cleaning path into a plurality of path points, wherein parameters of the path points include expected postures; S203: setting an acceptance path point on each sub-cleaning path of the internal cleaning path, wherein parameters of the acceptance path point include an expected cleaning area; S204 controls the photovoltaic robot to execute the internal cleaning path; monitors in real time the posture error between the actual posture acquired by the perception module and the expected posture; obtains the actual cleaning area based on the visual perception module at each acceptance path point, and determines the cleaning performance error based on the actual cleaning area and the expected cleaning area; S205: when the posture error is greater than a posture tolerance threshold or the cleaning performance error is greater than a cleaning tolerance threshold, generating a target internal cleaning path based on the posture error and / or the cleaning performance error, and controlling the photovoltaic robot to execute the target internal cleaning path; Among them, the posture tolerance threshold is determined based on a preset distance value between the internal area and the boundary area; the cleanliness tolerance threshold is set based on the redundancy rate determined by the cleaning performance index, and the redundancy rate is determined according to the minimum number of repeated cleanings, the cleaning power index, and the expected cleaning standard.

2. The method according to claim 1, characterized in that The method further comprises: When the posture error is smaller than a posture tolerance threshold and the cleaning performance error is smaller than a cleanliness tolerance threshold, the photovoltaic robot is controlled to continue executing the internal cleaning path.

3. The method according to claim 1, characterized in that The target internal cleaning path includes a posture correction path and an optimized cleaning path, and S205 includes: generating a posture correction path of the photovoltaic robot based on the posture error; and / or, updating a machine cleaning performance indicator based on the cleaning performance error; updating adjacent path spacings between adjacent sub-cleaning paths according to the updated machine cleaning performance index; The optimized cleaning path is generated based on the updated adjacent path spacing, and the photovoltaic robot is controlled to execute the optimized cleaning path.

4. The method according to claim 3, characterized in that The S205 further includes: Controlling the photovoltaic robot to move according to the posture correction path; Acquire a corrected posture based on the perception module, and if a posture error between the corrected posture and the expected posture is less than a preset correction error threshold, the posture correction is successful; The photovoltaic robot is controlled to execute the optimized cleaning path.

5. The method according to claim 1, wherein The S204 further includes: When the photovoltaic robot reaches the acceptance path point, acquiring image acceptance data collected by the visual perception module; Based on a preset recognition model, identifying the final clean area image in the image acceptance data; The actual cleaning area corresponding to the final cleaning region image is calculated based on a preset mapping algorithm.

6. The method according to claim 1, characterized in that The method further comprises: Obtaining a fluctuation frequency of the actual posture; When the fluctuation frequency is not within a preset frequency range, identifying whether there is a high-risk area on the cleaning path based on the environmental data collected by the perception module, and if there is no high-risk area on the cleaning path, controlling the photovoltaic robot to continue executing the internal cleaning path; If there is a high-risk area on the cleaning path, the photovoltaic robot is controlled to bypass the high-risk area and continue to execute the internal cleaning path.

7. A photovoltaic robot, characterized in that: The photovoltaic robot comprises: A moving device for moving between photovoltaic panels; A detection device, mounted on the mobile device, comprising at least two perception modules, the perception modules being used to collect motion data or environmental data of the photovoltaic robot, the perception modules comprising a visual perception module; a cleaning device, mounted on the mobile device, for cleaning the photovoltaic panel; A control system, connecting the mobile device, the detection device and the cleaning device, is used to implement the path planning method of the photovoltaic robot according to any one of claims 1 to 6.

8. A computer device, characterized in that: The computer device includes a memory and a processor; The memory is used to store computer programs; The processor is configured to execute the computer program and implement the path planning method for the photovoltaic robot according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor enables the processor to implement the path planning method for the photovoltaic robot according to any one of claims 1 to 6.

Images