Photovoltaic panel unmanned cleaning method and system based on cooperation of mother vehicle and child vehicle

Abstract

The invention relates to a photovoltaic panel unmanned cleaning method and system based on cooperation of a mother vehicle and a child vehicle, belongs to the technical field of cleaning equipment, and solves the problems of high cleaning cost, low efficiency and poor cleaning effect of photovoltaic panels in the prior art. Comprising a mother vehicle, a lifting platform, a calculation unit, inertial navigation and accessories thereof, a storage battery, a camera, a laser radar, a son vehicle and a cleaning device, wherein the mother vehicle is provided with a bus system and carries a lifting platform, a calculation unit, an inertial navigator, accessories and a storage battery; the lifting platform performs controllable height change and carries the sub-vehicle, the camera and the laser radar; the calculation unit receives the information transmitted by each component, carries out calculation processing and sends out a control instruction; the inertial navigation obtains driving information of a mother vehicle; the sub-vehicle is provided with a control unit for controlling the sub-vehicle; the camera is arranged on the side edge of the lifting platform and used for roadside sensing. The laser radar obtains environment point cloud data; the cleaning device is arranged at the head of the sub-vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of cleaning equipment, and in particular to an unmanned cleaning method and system for photovoltaic panels based on the cooperation of a mother-child vehicle. Background Art

[0002] With the advantages of renewable resources, carbon neutrality, carbon compliance, green environmental protection, safety and reliability, the development of the photovoltaic industry has received increasing attention. In the production chain of the photovoltaic industry, the cleaning of photovoltaic panels occupies a pivotal position. Dirt attached to the surface of photovoltaic panels will cause the light transmittance of photovoltaic panels to decrease, reducing the power generation efficiency and service life of photovoltaic panel components.

[0003] At present, there are two main ways to clean photovoltaic panels: manual cleaning and mechanized cleaning vehicles. However, each of them has some significant shortcomings. Manual cleaning is costly, inefficient, and has poor cleaning effects. Traditional cleaning vehicles are large and heavy, require manual assistance, are difficult to operate, and are prone to damage to photovoltaic panels. They are also not suitable for photovoltaic production bases with dense photovoltaic panels or high altitudes.

[0004] Therefore, based on the above-mentioned pain points of the photovoltaic panel cleaning industry, this technical field urgently needs a photovoltaic panel cleaning method that can integrate the characteristics of unmanned, standardized, low-cost and flexible application. Summary of the invention

[0005] In order to solve the above-mentioned problems of the existing photovoltaic panel cleaning methods, an embodiment of the present invention provides an unmanned photovoltaic panel cleaning method and system based on the cooperation of a mother-child vehicle.

[0006] According to one embodiment of the present invention, a photovoltaic panel unmanned cleaning system based on a mother-child vehicle collaboration is provided, comprising: a mother vehicle, a lifting platform, a computing unit, an inertial navigation and its accessories, a battery, a camera, a laser radar, a child vehicle and a cleaning device, wherein

[0007] The mother vehicle is provided with a bus system for sending control information and data, and is equipped with a lifting platform, a computing unit, an inertial navigation system and its accessories and a battery;

[0008] The lifting platform can change its height in a controllable manner and is equipped with a sub-car, a camera and a laser radar. A rotating plate is arranged on one side, and the rotating plate is connected to the central part of the lifting platform through an axis. A servo motor is arranged on the axis to control the rotation angle and speed of the rotating plate.

[0009] The computing unit receives information transmitted by each component, performs computing and processing, and issues control instructions;

[0010] The inertial navigation obtains the driving information of the mother vehicle;

[0011] The sub-vehicle is provided with a control unit to control the sub-vehicle;

[0012] The camera is a planar camera, which is arranged on the side of the lifting platform and is used for roadside perception;

[0013] The laser radar acquires environmental point cloud data and transmits it to the computing unit;

[0014] The cleaning device is arranged at the head of the sub-trolley and is controlled by the control unit of the sub-trolley.

[0015] According to another embodiment of the present invention, there is provided a method for unmanned cleaning of photovoltaic panels based on mother-child cooperation, comprising the following steps:

[0016] Step S1, placing the lifting platform on the mother vehicle and installing each component in place;

[0017] Step S2, performing motion control debugging on the mother vehicle and the child vehicle;

[0018] Step S3, environmental perception, obtaining environmental point cloud data through the laser radar and transmitting it to the calculation unit, which calculates the position and center point coordinates of the photovoltaic panel;

[0019] Step S4, the calculation unit plans a feasible path as the driving trajectory of the mother vehicle;

[0020] Step S5, the mother vehicle drives to the bottom of the photovoltaic panel according to the mother vehicle driving trajectory;

[0021] Step S6, the lifting platform carried by the mother vehicle delivers the child vehicle on the lifting platform to the required height onto the photovoltaic panel;

[0022] Step S7, obtaining photovoltaic panel cleaning image information through a planar camera, and extracting the position information of the sub-vehicle on the photovoltaic panel and the pollutant information on the photovoltaic panel using an image detection algorithm based on the photovoltaic panel cleaning image information;

[0023] Step S8, the control unit of the sub-vehicle obtains a reference trajectory of the sub-vehicle according to the trajectory planning algorithm and the photovoltaic panel size planning, and the sub-vehicle performs full coverage cleaning work based on the reference trajectory;

[0024] Step S9, the control unit of the sub-vehicle plans a fixed-point cleaning path for the sub-vehicle based on the extracted pollutant information on the photovoltaic panel, and the sub-vehicle performs the fixed-point cleaning work based on the fixed-point cleaning path;

[0025] Step S10, determine whether there are any uncleaned photovoltaic panels through the information obtained by the plane camera. If so, return to step S2 to re-obtain the position information of the uncleaned photovoltaic panels, otherwise end the task and instruct the mother vehicle to return to the starting point.

[0026] Optionally, the step S3 specifically includes the following steps:

[0027] Step S3.1, scanning the surrounding working environment by laser radar, collecting point cloud data and transmitting it to the computing unit;

[0028] In step S3.2, the computing unit receives the point cloud data collected by the lidar, performs format conversion, reflection intensity screening, pass-through filtering, DBSCAN clustering, cluster screening, minimum bounding sphere calculation and coordinate conversion processing on the point cloud data to obtain the final coordinate result of the photovoltaic panel, including the position of the photovoltaic panel, the coordinates of the center point and the equation of the low side straight line of the photovoltaic panel.

[0029] Optionally, the step S3.2 specifically includes the following steps:

[0030] Step S3.2.1, the calculation unit uses Boolean indexing to obtain reflection intensity data in the point cloud data, and selects and retains the point cloud data whose reflection intensity data conforms to the photovoltaic panel data interval;

[0031] Step S3.2.2, filtering the screened point cloud data, using a statistical filtering method, by setting two parameters, the neighboring point and the standard deviation multiplier, to determine the noise points and filter them out, to obtain the filtered point cloud data;

[0032] Step S3.2.3, the calculation unit clusters the filtered point cloud data by setting the minimum spacing and the minimum number of clustering points to obtain clustered point cloud data;

[0033] Step S3.2.4, judging the category with the largest number of point clouds in the clustered point cloud data as the point cloud data of the photovoltaic panel category, and retaining it;

[0034] Step S3.2.5, the computing unit performs a minimum bounding sphere algorithm on the retained point cloud data of the photovoltaic panel category, finds the smallest sphere that can contain all the point cloud data, and finds the coordinates of the center of the sphere. The coordinates of the center of the sphere are retained and output as the center position coordinates of the photovoltaic panel, and the point cloud data of the lower side of the photovoltaic panel is fitted to generate the equation of the lower side straight line;

[0035] Step S3.2.6, after converting the center position coordinates of the photovoltaic panel through the coordinate points of the world coordinate system where the laser radar is located, the final coordinate results of the photovoltaic panel are obtained, including the position of the photovoltaic panel, the center point coordinates and the low side straight line equation of the photovoltaic panel.

[0036] Optionally, the step S5 specifically includes the following steps:

[0037] Step S5.1, the calculation unit uses a tracking algorithm to obtain the preview point of the mother vehicle at the next moment according to the calculated driving trajectory of the mother vehicle, and calculates the linear velocity and angular velocity of the mother vehicle, and sends them to the CAN bus of the mother vehicle;

[0038] Step S5.2, the mother vehicle drives to the position under the photovoltaic panel according to the received information, wherein during the driving process of the mother vehicle, the computing unit performs trajectory tracking control on the mother vehicle, calculates the tracking model error by receiving the deviation between the current mother vehicle position information and the target path from the inertial navigation at each control moment, and successively changes the driving trajectory of the mother vehicle to accurately navigate the mother vehicle. At the same time, the mother vehicle receives the fitting low-sideline data transmitted by the computing unit, and dynamically adjusts the direction angle of the mother vehicle so that its driving direction is always parallel to the low-sideline. The mother vehicle obtains the direction angle data of the laser radar in real time during driving to continuously correct the deviation.

[0039] Optionally, the step S8 specifically includes the following steps:

[0040] Step S8.1, the sub-vehicle control unit receives the four corner points of the photovoltaic panel transmitted by the calculation unit, calculates the size of the photovoltaic panel, and obtains the reference trajectory of the sub-vehicle according to the trajectory planning algorithm and the photovoltaic panel size planning;

[0041] Step S8.2, set a distance parameter ld as the preview distance, receive the speed, acceleration, position and posture of the sub-car from the odometer carried by the sub-car, draw a circle with ld as the radius based on the center position of the current sub-car's rear wheels, and the intersection of the circle and the reference trajectory is the sub-car preview point;

[0042] Step S8.3, obtain the angular velocity of the sub-car:

[0043] δ=(2L sinα) / ld

[0044] Among them, δ represents the angular velocity of the sub-car, L is the longitudinal distance between the sub-car and the preview point, and α is the longitudinal deviation angle between the sub-car position and the preview point;

[0045] Step S8.4, obtain the wheel angle that enables the sub-vehicle to accurately reach the pre-aiming point, transmit the wheel angle and speed information to the control unit that controls the sub-vehicle through the serial port, and after information processing by the callback function, generate a control signal to control the sub-vehicle to move to the pre-aiming point in sequence according to the reference trajectory. During the movement, the photovoltaic panel is cleaned by the cleaning device to perform full coverage cleaning.

[0046] Optionally, the step S9 specifically includes the following steps:

[0047] Step S9.1, extracting the pollutant information on the photovoltaic panel from the photovoltaic panel cleaning image information obtained by the planar camera, processing the information, and obtaining the coordinates of the stains, and transmitting the coordinates to the sub-vehicle as a new cleaning target point;

[0048] Step S9.2, after the control unit of the sub-vehicle obtains the new preview point, it calculates and obtains the shortest path that conforms to the vehicle dynamics as the fixed-point cleaning path. The sub-vehicle moves along the fixed-point cleaning path to the new cleaning target point. During the movement, the stains are cleaned by the cleaning device to perform the fixed-point cleaning work.

[0049] Compared with the prior art, an unmanned photovoltaic panel cleaning method and system based on mother-child vehicle collaboration provided in accordance with an embodiment of the present invention has at least the following beneficial effects: low cost, high efficiency, flexible application, ability to achieve standardization, etc.; saving human resources; and providing a new model for the photovoltaic panel cleaning industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. The features and advantages of the present invention can be more clearly understood by referring to the drawings. The drawings are schematic and should not be understood as limiting the present invention in any way. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 A schematic diagram of a communication system for an unmanned photovoltaic panel cleaning method based on mother-child vehicle collaboration according to an embodiment of the present invention.

[0052] Figure 2 A mother vehicle workflow diagram of an unmanned photovoltaic panel cleaning method based on mother-child vehicle collaboration provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0054] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0055] The following describes in detail with reference to the accompanying drawings a method and system for unmanned cleaning of photovoltaic panels based on mother-child vehicle collaboration according to an embodiment of the present invention.

[0056] like Figure 1 As shown, according to an embodiment of the present invention, a photovoltaic panel unmanned cleaning system based on the cooperation of a mother-child vehicle is provided, including: a mother vehicle, a lifting platform, a computing unit, an inertial navigation and its accessories, a battery, a camera, a laser radar, a child vehicle, a cleaning device and other components.

[0057] Among them, the mother vehicle is provided with a bus system for sending control information and data, and is equipped with a lifting platform, a computing unit, an inertial navigation and its accessories, a battery, etc.

[0058] The lifting platform can be controlled to change height and carry a sub-car, camera and laser radar. The lifting platform can be provided with a baffle that can be placed downward on one side so that the sub-car can drive through to the target photovoltaic panel. The baffle can be a rotating plate with circular motion function, which is connected to the central part of the lifting platform through an axis, and a servo motor is provided on the axis to control the rotation angle and speed of the rotating plate.

[0059] The computing unit can receive information transmitted by each component, perform computing and processing, and issue control instructions. The computing unit can be powered by a battery carried by the mother vehicle.

[0060] The inertial navigation obtains the driving information of the mother vehicle. The inertial navigation may include two antennas, which are respectively placed on the front and rear sides of the mother vehicle, and need to be debugged to reduce errors and select a suitable data transmission protocol to obtain the driving information of the mother vehicle.

[0061] The camera may be a planar camera and is placed on the side of the lifting platform for roadside perception.

[0062] The laser radar is set on the lifting platform to obtain environmental point cloud data and transmit it to the computing unit. The laser radar can be placed in an open area of ​​the lifting platform to ensure normal operation, and it is necessary to ensure that the laser radar's own accuracy can meet the work requirements.

[0063] The sub-vehicle is provided with a control unit to control the sub-vehicle.

[0064] The cleaning device is arranged on the head of the sub-trolley and is controlled by the control unit of the sub-trolley.

[0065] Among them, the communication method in the unmanned photovoltaic panel cleaning system based on the mother-child cooperation is as follows: Figure 1 As shown, UDP (User Datagram Protocol) is used to communicate between various components to realize component communication in the computing unit system.

[0066] like Figure 2As shown, a method for unmanned cleaning of photovoltaic panels based on mother-child vehicle collaboration according to an embodiment of the present invention includes the following steps.

[0067] Step S1, placing the lifting platform on the mother vehicle and installing each component in place.

[0068] Step S2, performing motion control debugging on the mother vehicle and the child vehicle, including enabling the mother vehicle to read inertial navigation and its own motion data information and perform motion control accordingly, and the child vehicle to automatically plan a trajectory and drive according to the received information.

[0069] Step S3, environmental perception, obtains environmental point cloud data through laser radar and transmits it to the calculation unit, which calculates the location of the photovoltaic panel, the center point and the lower side straight line equation, the location information of surrounding obstacles and other data.

[0070] The step S3 specifically includes the following steps.

[0071] Step S3.1, a laser radar is fixedly installed on the lifting platform. During the tracking process, the laser radar is used to scan the surrounding working environment and collect point cloud data for subsequent processing.

[0072] Step S3.2, the computing unit receives the point cloud data collected by the laser radar, and performs format conversion, reflection intensity screening, straight-through filtering, DBSCAN clustering, cluster screening, minimum bounding sphere calculation, coordinate conversion and other processing on the point cloud data to obtain the final coordinate result of the photovoltaic panel position. This step S3.2 specifically includes the following steps.

[0073] Step S3.2.1, first, by manually selecting the original point cloud data in the database, it is found that the reflection intensity values ​​of the photovoltaic panel surface are all between 0-20. Therefore, for the collected original point cloud, the calculation unit uses Boolean indexing to obtain the reflection intensity data in the point cloud data, and filters out the point cloud data whose reflection intensity data conforms to the photovoltaic panel data interval and retains it, and filters out the point cloud data with a reflection intensity between 20-255.

[0074] Step S3.2.2, secondly, filter the screened point cloud data, use statistical filtering to filter, define two parameters, neighborhood points and standard deviation multiplier, to determine noise points and filter them out, and obtain filtered point cloud data. Including, for each point in the point cloud data, by calculating the average value and standard deviation of each point in a certain area as statistical information, to determine whether it is a valid point or an outlier noise point, and filter out the outlier noise points. The neighborhood point is a parameter that limits the size of the neighborhood and the number of points, and the standard deviation multiplier is a specified multiplier of the standard deviation value. If the difference between the statistical numerical information of a point and the average value of the numerical information of each point in the adjacent area exceeds the product of the standard deviation and the standard deviation multiplier, the point will be determined as an interfering noise point and filtered out, and will not participate in the subsequent processing process.

[0075] Step S3.2.3, the computing unit then clusters the filtered point cloud data to obtain clustered point cloud data, including setting two parameters: minimum spacing and minimum number of cluster points. The minimum spacing specifies the minimum distance between adjacent points that are clustered into the same category to ensure that multiple types of objects are not clustered into one category. The minimum number of cluster points specifies the minimum number of points required to cluster into one category to prevent a small amount of noise from interfering with the clustering results.

[0076] Step S3.2.4, after filtering and clustering, according to actual verification, most of the interference points in the scene are successfully filtered out, and the number of point clouds other than photovoltaic panels in the obtained point cloud data is small and basically cannot be clustered. Even if it is clustered into one category as interference items, the point clouds included in the cluster are relatively limited. Therefore, the category with the largest number of point clouds in the clustered point cloud data can be determined as the point cloud data of the photovoltaic panel category and retained.

[0077] In step S3.2.5, the computing unit performs the minimum enclosing sphere algorithm processing on the retained point cloud data of the photovoltaic panel category, that is, finds the smallest sphere that can contain all the point cloud data, and finds the center coordinates of the sphere. The center coordinates are retained and output as the center position coordinates of the photovoltaic panel; the computing unit uses the point cloud data of the lower side of the photovoltaic panel to fit the straight line equation of the side.

[0078] Step S3.2.6, after converting the center position coordinates of the photovoltaic panel through the coordinate points of the world coordinate system where the laser radar is located, the final coordinate results of the photovoltaic panel are obtained, including the position of the photovoltaic panel, the coordinates of the center point and the lower edge coordinates of the photovoltaic panel, for use in planning the driving trajectory of the mother vehicle.

[0079] Step S4, after the calculation unit obtains the coordinates of the positions of the photovoltaic panels and surrounding obstacles, it plans a feasible path as the driving trajectory of the mother vehicle. Considering the obstacle avoidance requirement, the A* algorithm can be used as the planning algorithm for the driving trajectory of the mother vehicle.

[0080] In step S5, the mother vehicle receives the mother vehicle driving track from the calculation unit, and drives to the bottom of the photovoltaic panel according to the mother vehicle driving track. Step S5 specifically includes the following steps.

[0081] Step S5.1, the calculation unit uses a pure tracking algorithm to find the preview point of the mother vehicle at the next moment based on the calculated driving trajectory of the mother vehicle, thereby calculating the linear velocity and angular velocity of the mother vehicle and sending them to the CAN bus of the mother vehicle.

[0082] Step S5.2, the mother vehicle drives to the position below the photovoltaic panel according to the received mother vehicle driving trajectory. The driving and movement of the mother vehicle are controlled by the computing unit. In the process of the mother vehicle driving to the photovoltaic panel position according to the trajectory plan, the mother vehicle can be positioned in real time. The mother vehicle is equipped with high-precision inertial navigation, which can meet the high-precision positioning requirements. Centimeter-based positioning can be achieved through RTK differential positioning and real-time message analysis. The computing unit performs trajectory tracking control on the mother vehicle, calculates the tracking model error by receiving the deviation between the current mother vehicle position information and the target path from the inertial navigation at each control moment, and changes the trajectory plan one by one, so as to accurately navigate the mother vehicle. At the same time, the mother vehicle receives the fitting low-sideline data transmitted by the computing unit, dynamically adjusts the direction angle of the mother vehicle, so that its driving direction is always parallel to the low sideline, and the mother vehicle obtains the direction angle data of the laser radar in real time during driving to continuously correct the deviation.

[0083] Step S6, the lifting platform transports the sub-trolley. The sub-trolley is delivered to the required height on the photovoltaic panel by the lifting platform carried by the mother vehicle. Step S6 specifically includes that the laser ranging sensor arranged on the lifting platform reads the vertical distance between the lifting platform and the photovoltaic panel, and transmits the vertical distance back to the control unit deployed on the lifting platform. The control unit controls the forward and backward movement of the push rod motor of the lifting platform according to a simple following algorithm, and finally stops at the specified height. The computing unit then sends a signal to the rotating plate on the lifting platform to control the servo motor drive, so that the rotating plate overlaps the lifting platform with the photovoltaic panel, and ensures that the sub-trolley set on the lifting platform can smoothly travel to the photovoltaic panel.

[0084] Step S7, obtaining photovoltaic panel cleaning image information through a planar camera arranged on the side of the lifting platform, and extracting the position information of the sub-vehicle on the photovoltaic panel and the pollutant information on the photovoltaic panel by using an image detection algorithm based on the photovoltaic panel cleaning image information.

[0085] In this embodiment, a model constructed by an image detection algorithm can be used to extract image information. The model is constructed, including using a plane camera to collect a large number of overall images of photovoltaic panels when the sub-cars are running at different positions and angles above the photovoltaic panels, as well as common types of dirt on some photovoltaic panels such as tree branches and leaves, and is divided into two categories: a data set and a test set. The data set images are adjusted to a uniform resolution, the image truth information is obtained by manual annotation, and the data samples are increased by random cropping, scaling, and rotation processing and saved in a folder. The model is trained using the data samples obtained from the data set, and the data output of the training results is compared with the true value information. The training results are evaluated and continuously optimized by data such as intersection and union ratio. After the test set images can be successfully recognized, the code is rewritten for real vehicle deployment to obtain the final model that meets the actual application requirements of the recognition process. The computing unit extracts the position information of the sub-car on the photovoltaic panel and the pollutant information on the photovoltaic panel from the photovoltaic panel cleaning image information through the final model.

[0086] Step S8, the control unit of the sub-vehicle obtains the full coverage cleaning path planning of the sub-vehicle based on the pre-aiming point design, and the sub-vehicle performs the full coverage cleaning work based on the full coverage cleaning path planning. Specifically including the following steps. In the first full coverage cleaning process, ensure that the sub-vehicle can cover the photovoltaic panel to the maximum extent during the entire tracking process, combine the plane camera to perform real-time positioning of the sub-vehicle, and use image processing technology to realize instant detection of whether the sub-vehicle will exceed the photovoltaic panel.

[0087] Step S8.1, the sub-vehicle control unit receives the four corner points of the photovoltaic panel obtained by the computing unit through the image detection algorithm, calculates the size of the photovoltaic panel, and plans a reference trajectory based on the trajectory planning algorithm and the size of the photovoltaic panel.

[0088] Step S8.2, to ensure that the sub-car has good robustness, the sub-car control unit uses a pure tracking algorithm as the sub-car motion control algorithm. First, a distance parameter ld is set as the preview distance. After the motion control algorithm receives the sub-car's speed, acceleration, position, posture, and reference trajectory data from the sub-car's own odometer, it draws a circle with ld as the radius based on the center position of the current sub-car's rear wheel, and the intersection of the circle and the reference trajectory is used as the sub-car preview point.

[0089] Step S8.3, obtain the angular velocity of the sub-car through the angular velocity equation:

[0090] δ=(2L sinα) / ld

[0091] Among them, δ is the angular velocity of the sub-vehicle, L is the longitudinal distance between the sub-vehicle and the preview point, and α is the longitudinal deviation angle between the sub-vehicle position and the preview point. The wheel angle that enables the sub-vehicle to accurately reach the preview point can be obtained. The wheel angle and speed information are transmitted to the control unit that controls the sub-vehicle through the serial port. After the information is processed by the callback function, a control signal is generated to control the sub-vehicle to move to the preview point in sequence according to the reference trajectory. During the movement, the photovoltaic panel is cleaned by the cleaning device to perform full coverage cleaning.

[0092] Step S9, the control unit of the sub-vehicle plans a fixed-point cleaning path for the sub-vehicle based on the pollutant information extracted from the photovoltaic panels, and the sub-vehicle performs fixed-point cleaning based on the fixed-point cleaning path. After the sub-vehicle completes full-coverage tracking cleaning, the sub-vehicle fixed-point cleaning path is planned and fixed-point cleaning is performed to clean stains such as bird excrement and scale. Specifically, the following steps are included.

[0093] In step S9.1, the pollutant information on the photovoltaic panel extracted from the photovoltaic panel cleaning image information obtained from the planar camera is processed to obtain the coordinates of the stains, and the coordinates are transmitted to the sub-car as a new preview point.

[0094] Step S9.2, after the control unit of the sub-vehicle obtains the new preview point, it calculates and obtains the shortest path that conforms to the vehicle dynamics as the fixed-point cleaning path. The sub-vehicle moves along the fixed-point cleaning path to the new preview point in sequence. During the movement, the stains are cleaned by the cleaning device to perform the fixed-point cleaning work.

[0095] Step S10, determine whether there are any uncleaned photovoltaic panels through the information obtained by the plane camera. If so, return to step S2 to re-obtain the position information of the uncleaned photovoltaic panels, otherwise end the task and instruct the mother vehicle to return to the starting point.

[0096] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0097] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0098] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. An unmanned photovoltaic panel cleaning system based on mother-child cooperation, characterized in that: include: The mother vehicle, lifting platform, computing unit, inertial navigation and its accessories, battery, camera, laser radar, sub-vehicle and cleaning device include: The mother vehicle is provided with a bus system for sending control information and data, and is equipped with a lifting platform, a computing unit, an inertial navigation system and its accessories and a battery; The lifting platform can change its height in a controllable manner and is equipped with a sub-car, a camera and a laser radar. A rotating plate is arranged on one side, and the rotating plate is connected to the central part of the lifting platform through an axis. A servo motor is arranged on the axis to control the rotation angle and speed of the rotating plate. The computing unit receives information transmitted by each component, performs computing and processing, and issues control instructions; The inertial navigation obtains the driving information of the mother vehicle; The sub-vehicle is provided with a control unit to control the sub-vehicle; The camera is a planar camera, which is arranged on the side of the lifting platform and is used for roadside perception; The laser radar acquires environmental point cloud data and transmits it to the computing unit; The cleaning device is arranged at the head of the sub-trolley and is controlled by the control unit of the sub-trolley.

2. A cleaning method for the unmanned photovoltaic panel cleaning system based on the mother-child vehicle collaboration according to claim 1, characterized in that: The following steps are involved: Step S1, placing the lifting platform on the mother vehicle and installing each component in place; Step S2, performing motion control debugging on the mother vehicle and the child vehicle; Step S3, environmental perception, obtaining environmental point cloud data through the laser radar and transmitting it to the calculation unit, which calculates the position and center point coordinates of the photovoltaic panel; Step S4, the calculation unit plans a feasible path as the driving trajectory of the mother vehicle; Step S5, the mother vehicle drives to the bottom of the photovoltaic panel according to the mother vehicle driving trajectory; Step S6, the lifting platform carried by the mother vehicle delivers the child vehicle on the lifting platform to the required height onto the photovoltaic panel; Step S7, obtaining photovoltaic panel cleaning image information through a planar camera, and extracting the position information of the sub-vehicle on the photovoltaic panel and the pollutant information on the photovoltaic panel using an image detection algorithm based on the photovoltaic panel cleaning image information; Step S8, the control unit of the sub-vehicle obtains a reference trajectory of the sub-vehicle according to the trajectory planning algorithm and the photovoltaic panel size planning, and the sub-vehicle performs full coverage cleaning work based on the reference trajectory; Step S9, the control unit of the sub-vehicle plans a fixed-point cleaning path for the sub-vehicle based on the extracted pollutant information on the photovoltaic panel, and the sub-vehicle performs the fixed-point cleaning work based on the fixed-point cleaning path; Step S10, determine whether there are any uncleaned photovoltaic panels through the information obtained by the plane camera. If so, return to step S2 to re-obtain the position information of the uncleaned photovoltaic panels, otherwise end the task and instruct the mother vehicle to return to the starting point.

3. The cleaning method according to claim 2, characterized in that: The step S3 specifically comprises the following steps: Step S3.1, scanning the surrounding working environment by laser radar, collecting point cloud data and transmitting it to the computing unit; In step S3.2, the computing unit receives the point cloud data collected by the lidar, performs format conversion, reflection intensity screening, pass-through filtering, DBSCAN clustering, cluster screening, minimum bounding sphere calculation and coordinate conversion processing on the point cloud data to obtain the final coordinate result of the photovoltaic panel, including the position of the photovoltaic panel, the coordinates of the center point and the equation of the low side straight line of the photovoltaic panel.

4. The cleaning method according to claim 3, characterized in that: The step S3.2 specifically includes the following steps: Step S3.2.1, the calculation unit uses Boolean indexing to obtain reflection intensity data in the point cloud data, and selects and retains the point cloud data whose reflection intensity data conforms to the photovoltaic panel data interval; Step S3.2.2, filtering the screened point cloud data, using a statistical filtering method, by setting two parameters, the neighboring point and the standard deviation multiplier, to determine the noise points and filter them out, to obtain the filtered point cloud data; Step S3.2.3, the calculation unit clusters the filtered point cloud data by setting the minimum spacing and the minimum number of clustering points to obtain clustered point cloud data; Step S3.2.4, judging the category with the largest number of point clouds in the clustered point cloud data as the point cloud data of the photovoltaic panel category, and retaining it; Step S3.2.5, the computing unit performs a minimum bounding sphere algorithm on the retained point cloud data of the photovoltaic panel category, finds the smallest sphere that can contain all the point cloud data, and finds the coordinates of the center of the sphere. The coordinates of the center of the sphere are retained and output as the center position coordinates of the photovoltaic panel, and the point cloud data of the lower side of the photovoltaic panel is fitted to generate the equation of the lower side straight line; Step S3.2.6, after converting the center position coordinates of the photovoltaic panel through the coordinate points of the world coordinate system where the laser radar is located, the final coordinate results of the photovoltaic panel are obtained, including the position of the photovoltaic panel, the center point coordinates and the low side straight line equation of the photovoltaic panel.

5. The cleaning method according to claim 4, characterized in that: The step S5 specifically comprises the following steps: Step S5.1, the calculation unit uses a tracking algorithm to obtain the preview point of the mother vehicle at the next moment according to the calculated driving trajectory of the mother vehicle, and calculates the linear velocity and angular velocity of the mother vehicle, and sends them to the CAN bus of the mother vehicle; Step S5.2, the mother vehicle drives to the position under the photovoltaic panel according to the received information, wherein during the driving process of the mother vehicle, the computing unit performs trajectory tracking control on the mother vehicle, calculates the tracking model error by receiving the deviation between the current mother vehicle position information and the target path from the inertial navigation at each control moment, and successively changes the driving trajectory of the mother vehicle to accurately navigate the mother vehicle. At the same time, the mother vehicle receives the fitting low-sideline data transmitted by the computing unit, and dynamically adjusts the direction angle of the mother vehicle so that its driving direction is always parallel to the low-sideline. The mother vehicle obtains the direction angle data of the laser radar in real time during driving to continuously correct the deviation.

6. The cleaning method according to claim 5, characterized in that: The step S8 specifically includes the following steps: Step S8.1, the sub-vehicle control unit receives the four corner points of the photovoltaic panel transmitted by the calculation unit, calculates the size of the photovoltaic panel, and obtains the reference trajectory of the sub-vehicle according to the trajectory planning algorithm and the photovoltaic panel size planning; Step S8.2, set a distance parameter ld as the preview distance, receive the speed, acceleration, position and posture of the sub-car from the odometer carried by the sub-car, draw a circle with ld as the radius based on the center position of the current sub-car's rear wheels, and the intersection of the circle and the reference trajectory is the sub-car preview point; Step S8.3, obtain the angular velocity of the sub-car: δ=(2L sinα) / ld Among them, δ represents the angular velocity of the sub-car, L is the longitudinal distance between the sub-car and the preview point, and α is the longitudinal deviation angle between the sub-car position and the preview point; Step S8.4, obtain the wheel angle that enables the sub-vehicle to accurately reach the pre-aiming point, transmit the wheel angle and speed information to the control unit that controls the sub-vehicle through the serial port, and after information processing by the callback function, generate a control signal to control the sub-vehicle to move to the pre-aiming point in sequence according to the reference trajectory. During the movement, the photovoltaic panel is cleaned by the cleaning device to perform full coverage cleaning.

7. The cleaning method according to claim 6, characterized in that: The step S9 specifically includes the following steps: Step S9.1, extracting the pollutant information on the photovoltaic panel from the photovoltaic panel cleaning image information obtained by the planar camera, processing the information, and obtaining the coordinates of the stains, and transmitting the coordinates to the sub-vehicle as a new cleaning target point; Step S9.2, after the control unit of the sub-vehicle obtains the new preview point, it calculates and obtains the shortest path that conforms to the vehicle dynamics as the fixed-point cleaning path. The sub-vehicle moves along the fixed-point cleaning path to the new cleaning target point. During the movement, the stains are cleaned by the cleaning device to perform the fixed-point cleaning work.

Images