Spraying method and spraying device for environment-friendly waterproof paint

Abstract

The invention relates to the technical field of material spraying, in particular to an environment-friendly waterproof paint spraying method and a spraying device thereof.The method comprises the steps that in the second process, spraying pressure, spraying flow and a spraying area are obtained in real time when spraying operation is executed, in the first process, the spraying pressure, the spraying flow and the spraying area are obtained again through multiple iterations according to the spraying unevenness of the spraying area, so that the spraying unevenness is smaller than or equal to a first preset threshold value. Reducing the number of iterations of the spraying control model in the first process according to the movement track of the nozzle in the second process and the track prediction error of the prediction track in the first process; and the spraying pressure and the spraying flow which are obtained again after the first process is finished serve as the spraying pressure and the spraying flow at the time point T when the spraying operation is executed in the second process. By means of the waterproof paint spraying device, spraying personnel can conduct uniform spraying operation in a narrow area which is difficult to enter, and the one-time construction qualification rate of waterproof paint is increased.

Description

Technical Field

[0001] The invention relates to the technical field of material spraying, and in particular to an environmentally friendly waterproof coating spraying method and a spraying device thereof. Background Art

[0002] When spraying waterproof coatings on houses in a community, sprayers are generally required to manually control the spraying device to perform the spraying operation. Especially for some narrow areas that are inaccessible to personnel, such as corners or around pipes, it is often difficult to complete the spraying task relatively evenly in one go. It is usually necessary to spray repeatedly on some areas where the spraying is uneven or the spraying thickness is insufficient, as well as narrow and difficult-to-access areas. On the one hand, there is over-spraying, and on the other hand, there is prone to obvious uneven spraying, resulting in low manual spraying efficiency and unsatisfactory spraying results. Summary of the invention

[0003] In order to solve the above problems, the present invention provides an environmentally friendly waterproof coating spraying method and a spraying device thereof.

[0004] The environmentally friendly waterproof coating spraying method and the spraying device of the present invention adopt the following technical scheme: An embodiment of the present invention provides an environmentally friendly waterproof coating spraying method, the method comprising the following steps: Train a spray control model that inputs spray height, spray pressure, and paint flow rate, and outputs spray thickness and spray area; In the second process, a spraying operation is performed, and the spraying pressure, spraying flow rate and spraying area at each moment are obtained according to the spraying height H and the spraying thickness L at each moment, including: inputting the variable parameters consisting of the spraying pressure and the spraying flow rate and the spraying height H into the spraying control model, and recording the spraying thickness output by the spraying control model as L1, updating the variable parameters through multiple iterations of the spraying control model until the loss function F is minimized, and using the updated variable parameters and the spraying area output by the spraying control model as the spraying pressure, spraying flow rate and spraying area at each moment, wherein F is the difference between L1 and L; The first process includes: D1: Get the predicted trajectory of the nozzle between the current moment and time point T, where time point T is a preset time point after the current moment, and the predicted trajectory includes the spraying height at time point T; D2: According to the spraying height H and spraying thickness L at time point T, the spraying pressure, spraying flow rate and spraying area at time point T are obtained; according to the spraying area at time point T, the spraying unevenness is obtained; D3: According to the spraying unevenness, the spraying pressure, spraying flow rate, spraying area, and spraying unevenness at the time point T are obtained by repeatedly executing D2, until the first process is terminated when the spraying unevenness is less than or equal to the first preset threshold value; In the process of repeatedly executing D2, the variable parameter of F each time reaching the minimum value is recorded as the control parameter; according to the motion trajectory of the nozzle in the second process and the trajectory prediction error of the predicted trajectory in the first process and the control parameter when reaching the minimum value all times, the number of iterations of the spray control model in the first process is reduced; The spraying pressure and spraying flow rate regained after the first process is completed are used as the spraying pressure and spraying flow rate at time point T when the spraying operation is performed in the second process.

[0005] Preferably, the spraying pressure, spraying flow rate, spraying area, and spraying unevenness at time point T are obtained by repeatedly executing D2 according to the spraying unevenness, and the first process is terminated when the spraying unevenness is less than or equal to a first preset threshold value, and the specific steps include the following: After the first execution of D2, if the spraying unevenness is less than or equal to the first preset threshold, the first process is directly terminated. If the spraying unevenness is greater than the first preset threshold, D2 is executed again according to the spraying unevenness and the spraying pressure, spraying flow rate and spraying area at the time point T are re-obtained. The spraying unevenness is obtained again according to the re-obtained spraying area, and so on, until the first process is terminated when the spraying unevenness is less than or equal to the first preset threshold; According to the spraying unevenness, D2 is executed again to re-obtain the spraying pressure, spraying flow rate and spraying area at the time point T, wherein the specific process also includes: iterating the spraying control model multiple times to update the variable parameters until the loss function F is minimized; wherein the method for obtaining the loss function F specifically includes: the spraying thickness at the time point T is also represented as L, ; Wherein w represents the spraying suppression coefficient, and the spraying suppression coefficient is negatively correlated with the spraying unevenness.

[0006] Preferably, the specific steps of obtaining the spraying suppression coefficient include: The variable parameters input by the spray control model are obtained and recorded as the first parameter feature point. The spray pressure and spray flow rate obtained in D2 are recorded as the second parameter feature point. The Euclidean distance between the first parameter feature point and the second parameter feature point is recorded as the first feature of the first parameter feature point. The spray suppression coefficient is negatively correlated with the spray unevenness and positively correlated with the first feature.

[0007] Preferably, the specific steps for obtaining the spraying unevenness are as follows: For the spraying areas obtained at all times before time point T, the sum of the spraying thickness of each position in all spraying areas is recorded as the sprayed thickness of each position; The spraying area obtained at time point T is recorded as the current spraying area. The spraying thickness of each position in the current spraying area plus the sprayed thickness of each position is taken as the predicted spraying thickness of each position. The spraying unevenness is positively correlated with the variance of the predicted spraying thickness of all positions in the current spraying area.

[0008] Preferably, the specific steps of obtaining the trajectory prediction error between the motion trajectory of the nozzle in the second process and the predicted trajectory in the first process are as follows: When the spraying task is executed in the second process, the spraying pressure, spraying flow rate and spraying area obtained at each moment are recorded as actual parameter data points; the spraying pressure, spraying flow rate and spraying area obtained at each moment on the predicted trajectory in the first process are recorded as predicted parameter data points; The Euclidean distance between the nozzle position at each moment on the predicted trajectory in the first process and the nozzle position at each moment in the second process is recorded as the first difference; the Euclidean distance between the predicted parameter data point at each moment in the first process and the actual parameter data point at each moment in the second process is recorded as the second difference; the spraying error of the predicted trajectory is positively correlated with the first difference and the second difference, respectively.

[0009] Preferably, the method of reducing the number of iterations of the spray control model in the first process according to the motion trajectory of the nozzle in the second process and the trajectory prediction error of the predicted trajectory in the first process and the control parameters when all times reach the minimum value includes the following specific steps: When the trajectory prediction error is greater than the second preset threshold, the spraying height and spraying thickness obtained in the second process are recorded as H2 and L2 respectively; H2 and L2 are used as target feature points; At the same time, for the spraying height H and the spraying thickness L at the time point T during the repeated execution of D2, the values ​​of the spraying height H and the spraying thickness L at the time point T are reset to H2 and L2 respectively; For each control parameter when F reaches a minimum value, the spraying height and spraying thickness corresponding to the control parameter are recorded as parameter characteristic points of the control parameter; In the process of repeatedly executing D2, for the variable parameters input in each iteration of the spray control model, the iteration suppression priority of the variable parameters is obtained according to the distance between the variable parameters and all control parameters, the distance between the parameter feature point of the control parameter when F reaches the minimum value all times and the target feature point, and the trajectory prediction error; The number of iterations of the spray control model in the first process is reduced according to the iteration suppression priority of the variable parameter.

[0010] Preferably, the iterative suppression priority of the variable parameter is obtained according to the distance between the variable parameter and all control parameters, the distance between the parameter feature point of the control parameter when F reaches the minimum value for all times and the target feature point, and the trajectory prediction error, and the specific steps include the following: The iterative suppression coefficient of each control parameter is obtained by using the distance between the parameter feature point of the control parameter when F reaches the minimum value for all times and the target feature point; The Euclidean distance between the variable parameter and each control parameter is obtained, the iterative suppression coefficient of each control parameter is normalized, the normalized iterative suppression coefficient of each control parameter is used as the weight, the Euclidean distance between the variable parameter and all control parameters is weighted summed, and the product of the weighted summation result and the trajectory prediction error is used as the iterative suppression priority of the variable parameter.

[0011] Preferably, the iterative suppression coefficient of each control parameter is obtained by using the distance between the parameter feature point of the control parameter when F reaches the minimum value all times and the target feature point, and the specific steps include the following: For the parameter feature points of the control parameters when F reaches the minimum value all the times; perform mean shift clustering on all the parameter feature points to obtain all categories, and all the control parameters corresponding to all the parameter feature points in the same category are taken as a control parameter category, and the mean of all the parameter feature points in the same control parameter category is recorded as the average feature point; the Euclidean distance between the average feature point and the target feature point is taken as the iterative suppression coefficient of each control parameter in each control parameter category.

[0012] Preferably, the method of reducing the number of iterations of the spray control model in the first process according to the iteration suppression priority of the variable parameter includes the following specific steps: The mean of the iteration suppression priorities of the variable parameters in the most recent iterations is obtained, and the variable parameters input into the spray control model in the current iteration are recorded as the current variable parameters; if the iteration suppression priority of the current variable parameter is greater than the mean, several other variable parameters closest to the current variable parameter are randomly selected, and among the iteration suppression priorities of the other variable parameters, the variable parameter with the smallest iteration suppression priority is used as the input of the spray control model.

[0013] An embodiment of the present invention also provides an environmentally friendly waterproof coating spraying device, which includes a nozzle, a memory, a processor, and a computer program stored in the memory and executable on the processor. An ultrasonic ranging sensor and an acceleration sensor are installed on the nozzle, which are used to obtain the spraying height and the movement trajectory of the nozzle, respectively. When the processor executes the computer program, the steps of the above-mentioned environmentally friendly waterproof coating spraying method are implemented.

[0014] The beneficial effects of the technical solution of the present invention are: The present invention obtains the spraying pressure, spraying flow rate and spraying area at each moment according to the spraying height H and spraying thickness L at each moment, so that the spraying pressure and spraying flow rate of the spraying device can be automatically controlled during the spraying process to ensure a suitable spraying thickness.

[0015] Furthermore, according to the spraying unevenness, the spraying pressure, spraying flow rate, spraying area, and spraying unevenness at the time point T are regained by repeatedly executing D2 until the spraying unevenness is less than or equal to the first preset threshold value; this process recalculates the spraying pressure and spraying flow rate multiple times so that the finally obtained spraying pressure and spraying flow rate can ensure that the waterproof coating sprayed on the wall or the ground has as uniform a thickness as possible.

[0016] Furthermore, according to the trajectory prediction error of the predicted trajectory of the nozzle in the second process and the control parameters when all times reach the minimum value, the number of iterations of the spray control model in the first process is reduced; the spray pressure and spray flow rate obtained after the first process is completed are used as the spray pressure and spray flow rate at time point T when the spraying operation is performed in the second process. This process makes the iterative process in the first process adapt to the spraying process in the second process, avoiding the situation that the spray pressure and spray flow rate obtained at the end of the first process cannot be used for the spraying process in the second process due to the large trajectory prediction error in the first process. On the other hand, the spray prediction process of the first process is intervened and adjusted through the spraying process of the second process, so that the spraying personnel can spray in a small and difficult-to-access space. Even if the same position is sprayed repeatedly, the uniformity of the spraying can be guaranteed and the waterproof coating has the required thickness. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] 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 or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flow chart of the steps of an environmentally friendly waterproof coating spraying method provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of an environmentally friendly waterproof coating spraying method and a spraying device thereof proposed according to the present invention, its specific implementation method, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0020] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0021] The specific scheme of an environmentally friendly waterproof coating spraying method and a spraying device thereof provided by the present invention is described in detail below in conjunction with the accompanying drawings.

[0022] Embodiment 1: The present embodiment provides an environmentally friendly waterproof coating spraying device, which comprises a nozzle, a feed pipe and a waterproof coating storage tank. One end of the feed pipe is connected to the waterproof coating storage tank, and the other end is connected to the nozzle. A flow control valve is installed at the connection between the feed pipe and the waterproof coating storage tank to control the flow of the coating. The waterproof coating storage tank pressurizes the coating therein and delivers it to the feed pipe, and the coating in the feed pipe is sprayed out from the nozzle. The waterproof coating anti-splash spraying device and control method thereof provided by CN116603653A discloses the nozzle structure, which will not be described in detail in this embodiment.

[0023] Furthermore, in this embodiment, an ultrasonic ranging sensor and an acceleration sensor are installed on the nozzle, and the spraying device also includes a computer, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it reads the data output by the ultrasonic ranging sensor and the acceleration sensor, and implements an environmentally friendly waterproof coating spraying method.

[0024] The spraying device is already available, and the use and installation methods of the ultrasonic distance measuring sensor and the acceleration sensor are also well-known and conventional, and will not be described in detail in this embodiment.

[0025] It should be noted that in this embodiment, the feed pipe is divided into two parts, one part connected to the nozzle is a hard pipe (such as a metal pipe), and the other part connected to the waterproof coating storage tank is a hose, and the hose and the hard pipe are connected together using a metal clamp. The purpose of using a hard pipe is to facilitate the sprayer to extend the nozzle into a narrow and difficult-to-access area (such as the corners of old communities, around pipelines, etc.).

[0026] Embodiment 2: See also Figure 1 , which shows a flow chart of the steps of an environmentally friendly waterproof coating spraying method provided in this embodiment, the method comprising the following steps: Step S001, training input spraying height, spraying pressure, paint flow rate, output spraying thickness and spraying area of ​​the spraying control model.

[0027] When spraying waterproof paint on the walls or floors of buildings, the sprayer holds a spray device (specifically, the hard tube part of the handheld spray device), and then turns on the spray device. The waterproof paint sprays out from the nozzle of the spray device. The sprayed waterproof paint forms a waterproof coating of a certain thickness and a certain spraying area on the wall or floor, making the wall or floor waterproof.

[0028] When spraying at any point on the wall or the ground, the thickness of the waterproof coating formed and the spraying area are related to the paint flow, paint pressure and spraying height of the spraying device. For example, the greater the paint flow, the greater the spraying thickness; the greater the spraying height, the larger the spraying area, the more obvious the atomization, and the smaller the spraying thickness; appropriate paint pressure can reduce the atomization effect and make the spraying thickness higher.

[0029] During spraying, in order to ensure the uniformity of spraying and improve the pass rate of waterproof coating construction in one time, it is necessary to control the spraying pressure and nozzle flow.

[0030] This embodiment controls the spraying process by training a spraying control model.

[0031] As an example, a method for training a spray control model includes: In a laboratory environment, the same type of spray device is used to conduct a spray test on the surface of an object: the nozzle of the spray device is placed at different heights (the height refers to the distance from the nozzle to the surface of the object, also recorded as the spray height). At each height, different spray pressures and paint flows are used for spraying. Any point on the surface of the object is selected as a test point. After one second of spraying at the test point, the spray thickness and spray area of ​​the test point are obtained. Each set of spray height, spray pressure, and paint flow is recorded as a sample, and the corresponding spray thickness and spray area are used as sample labels.

[0032] All samples and labels obtained in the laboratory environment constitute the dataset.

[0033] As an example, a method for obtaining the spraying height is as follows: an ultrasonic distance measuring sensor on the nozzle is used to measure the distance from the nozzle to the surface of the object, and the distance is used as the spraying height.

[0034] As an example, the method for obtaining the spraying thickness is: In a laboratory environment, a laser distance sensor is installed above the test point. The vertical distance from the laser distance sensor to the test point is obtained before spraying, and the vertical distance from the laser distance sensor to the test point is obtained again after spraying. The absolute value of the difference between the two distances is used as the spraying thickness.

[0035] The spraying thickness in this embodiment refers to the thickness of the spraying per second when the spraying is carried out continuously at a fixed point at the test point.

[0036] As an optional example, the method for obtaining the spraying area includes: The spraying area in this embodiment is circular, and the radius of the circular area is measured; the radius of the circular area is used as the value of the spraying area.

[0037] In another preferred example, the method for obtaining the spraying area includes: Considering that the thickness of the waterproof material coating at different positions in the spraying area may be different; obtain the coating thickness at each position in the spraying area, and fit the coating thickness at all positions in the spraying area into a Gaussian model; the mean, variance and radius of the Gaussian model form a three-dimensional vector and use it as the value of the spraying area. At this time, the coating thickness at each position in the spraying area can be obtained according to the value of the spraying area.

[0038] In other examples, the Gaussian model can be replaced by a Gaussian mixture model so that the spraying area can more accurately describe the coating thickness (or spraying thickness) at different locations in the spraying area.

[0039] The spray control model in this embodiment is a fully connected neural network model. In this embodiment, the fully connected neural network model has four intermediate layers, each with three neurons. The input of the fully connected neural network model is spray height, spray pressure, and paint flow, and the output is spray thickness and spray area. Other neural network models can be used in other embodiments, or the fully connected neural network model can be set to other network structures, which is not specifically limited in this embodiment.

[0040] The spray control model is trained using the above-obtained data set, and the trained spray control model is obtained; the loss function used in the training is the mean square error loss function, and the stochastic gradient descent algorithm is used during the training process to update the neuron parameters in the spray control model. The specific training process is a well-known technology and will not be described in detail in this embodiment.

[0041] Step S002: using the spraying control model to obtain the spraying pressure, spraying flow rate and spraying area at each moment according to the spraying height and spraying thickness at each moment.

[0042] In addition to the ultrasonic distance measuring sensor for obtaining the spraying height, the nozzle in this embodiment is also equipped with an acceleration sensor for the acceleration of the nozzle. In this embodiment, only two-dimensional acceleration data in a plane parallel to the wall or the ground is obtained, and the motion trajectory of the nozzle is obtained according to the acceleration and spraying height of the nozzle using an inertial navigation algorithm.

[0043] When the spraying personnel are spraying, the goal of the spraying operation is to spray a layer of waterproof coating with a thickness of L0 on the wall or the ground. It should also be noted that this embodiment is for spraying on the wall or the ground of a residential area, and does not strictly require that the thickness of the waterproof coating at each position on the wall or the ground is exactly the same. A certain error in the thickness of the waterproof coating is allowed. The purpose of this embodiment is to ensure the uniformity of the spraying as much as possible.

[0044] When the sprayer starts spraying at the current moment (in this embodiment, 0.5 seconds is a moment), the ultrasonic ranging sensor is used to obtain the spraying height H of the nozzle at the current position at the current moment. The spraying pressure, spraying flow rate and spraying area at the current moment are obtained according to the spraying height H and the spraying thickness L at the current position.

[0045] As an example, the spraying pressure, spraying flow rate and spraying area at the current moment are obtained according to the spraying height H and the spraying thickness L at the current position, including the following method: The spray thickness L=L0 at the current position of the nozzle.

[0046] The spraying pressure and spraying flow rate are regarded as a set of variable parameters (or as a set of neuron parameters), and the variable parameters are randomly initialized. The variable parameters and the spraying height H are input into the spraying control model, and the spraying control model outputs the spraying thickness and the spraying area, which are recorded as L1 and S1 respectively. The neuron parameters in the spraying control model are kept unchanged (that is, the neuron parameters are frozen so that they cannot be updated and changed). The variable parameters are regarded as neurons that can be updated, and the variable parameters are updated using the following loss function. The loss function F is: The variable parameters in the spray control model are updated multiple times by using the stochastic gradient descent algorithm. When F is the smallest, the variable parameters and the spray area S1 are obtained. The variable parameters at this time are used as the spray pressure and spray flow rate of the spray device at the current moment, and the current position is sprayed with the spray pressure and spray flow rate. The spray area formed on the wall or the ground at the current position is S1.

[0047] As an example, the method for determining whether F is minimum is: obtain the values ​​of F after the most recent several (for example, 20) iterative updates, obtain the maximum value max and the minimum value min among these values, and when (max-min) / max is less than the preset discrimination threshold, it is determined that F has reached the minimum, otherwise it has not reached the minimum. In particular, if the most recent iterative update number does not meet 20 times, it is also determined that it has not reached the minimum. The preset discrimination threshold in this embodiment is 20%, and the discrimination threshold can be set to other values ​​in other embodiments, which is not limited in this embodiment.

[0048] At the next moment, the nozzle moves to the next position, and the spraying height H is obtained again at the next moment. The above method is used to obtain the spraying pressure and spraying flow rate during spraying at the next moment according to the spraying height H, as well as the spraying area formed on the wall or the ground at the next position.

[0049] In some optional embodiments, by analogy, the spraying is continuously performed at subsequent moments until the spraying is completed. The process automatically controls the spraying pressure and the spraying flow rate so that the waterproof coating reaches the expected thickness.

[0050] However, in this embodiment, considering that the movement of the nozzle during the spraying process is manually controlled, especially when spraying in some narrow spaces that the sprayers cannot enter or are difficult to access, such as corners, around pipes and other places that are difficult to access manually, the sprayers are often required to extend the nozzle into the narrow space and then manually control the nozzle to spray. Therefore, the following problems may exist: the spraying area formed during the next spraying may overlap with the spraying area formed during the previous spraying, resulting in the same position being sprayed multiple times, making the thickness of the waterproof coating at some positions thicker or uneven, especially in a narrow space, because the sprayers cannot observe the spraying conditions at each position, it is easy to have uneven thickness.

[0051] In this embodiment, during the spraying process at the subsequent moments after the first moment, for the current moment in the spraying process, the ultrasonic distance measuring sensor is used to obtain the spraying height of the nozzle at the current position, which is still expressed as H. The spraying thickness L at the current moment where the nozzle is located (that is, the current position) is obtained. At this time, L is not necessarily equal to L0 because the current position may have been sprayed when spraying was performed before the current moment (that is, the current position is in the spraying area during the previous spraying).

[0052] As an example, the method for obtaining the spraying thickness L at the current position includes: Get the spraying area of ​​each position at all previous moments. Each spraying area contains the spraying thickness of each position in the spraying area. Get the spraying thickness of each spraying area at the current position. The sum of the spraying thickness of all spraying areas at the current position is recorded as the sprayed thickness L1. Then the spraying thickness L at the current position = L0-L1, which means that the previous position needs to be sprayed with a waterproof layer with a thickness of L.

[0053] It should be noted that the spraying thickness in this embodiment is the thickness at the designated point after one second of spraying, but in the actual spraying process, the same position will not stay for one second. At this time, the spraying thickness is the stay time at the same position multiplied by the thickness after one second of spraying.

[0054] Note that if the current position is not included in any spraying area, the spraying thickness of the spraying area at the current position is equal to 0. When the spraying thickness L is less than 0, L=0.

[0055] Furthermore, using the above method, the spraying pressure and spraying flow rate at the current moment (or current position) and the spraying area formed on the wall or the ground are obtained according to the spraying height H and the spraying thickness L at the current position.

[0056] Step S003, iterating the spray control model multiple times according to the unevenness of the spray area, and re-obtaining the spray pressure, spray flow rate and spray area.

[0057] So far, the above provides a method for obtaining the spraying pressure and the spraying flow rate at each moment in the spraying process at subsequent moments after the first moment. In some feasible embodiments, the spraying process can be completed according to this method.

[0058] However, in the present embodiment, considering that the spraying process, especially the spraying process in a narrow space, is controlled by the spraying personnel, there may be repeated spraying in certain positions or certain local areas. Therefore, for the spraying area at each moment in the spraying process in the subsequent moments, these spraying areas overlap and overlap, which leads to excessive spraying thickness at certain positions, or uneven spraying thickness. The reason is that the above method can only ensure to a certain extent that the coating thickness of the waterproof material at the position where the nozzle is located is equal to or close to L0, but does not ensure that the coating thickness of the waterproof material at all positions in the spraying area formed during spraying is equal to or close to L0.

[0059] (I) To address this problem, an optional embodiment provided by the present invention includes: After obtaining the spraying pressure and spraying flow rate at the current moment and the spraying area formed on the wall or the ground according to step S002, spraying is not performed according to the spraying pressure and spraying flow rate, but the following processing is performed: (1) Obtain the spraying unevenness based on the spraying area.

[0060] As an example, the spraying unevenness is obtained according to the spraying area, including the following methods: For the spraying areas obtained at all times before the current moment (excluding the current moment), the acquisition method is the same as step S002; the sum of the spraying thickness of each position in all spraying areas is recorded as the sprayed thickness of each position; For each position in the spraying area at the current moment (recorded as the current spraying area), the spraying thickness of each position in the current spraying area plus the sprayed thickness of each position is taken as the predicted spraying thickness of each position; the ratio of the variance of the predicted spraying thickness of all positions in the current spraying area to L0 is taken as the spraying unevenness.

[0061] In other examples, the difference between the average of the predicted spraying thicknesses at all positions in the current spraying area and L0 is obtained, and the ratio of the absolute value of the difference to L0 is used as the spraying unevenness.

[0062] In some other embodiments, the product of the variance of the predicted spraying thickness and the absolute value of the difference may be used as the spraying non-uniformity.

[0063] When the spraying unevenness is less than or equal to the first preset threshold, the spraying pressure and spraying flow rate obtained at the current moment are used as the spraying pressure and spraying flow rate at the current moment. In this embodiment, the first preset threshold is set to 0.3, and in other embodiments, the first preset threshold can be set to other values, which are not specifically limited in this embodiment.

[0064] When the spraying unevenness is greater than the first preset threshold, the following processing is performed: (2) Re-obtain the spraying pressure, spraying flow rate, and spraying area based on the spraying unevenness.

[0065] As an example, the method of re-obtaining the spraying pressure and spraying flow rate, as well as the spraying area according to the spraying unevenness includes: It should be noted that, at the same spraying height, different spraying pressures and spraying flow rates may correspond to the same or similar spraying thicknesses, but may correspond to different spraying areas. For example, a smaller spraying pressure and a larger spraying flow rate may correspond to the same or similar spraying thicknesses as a larger spraying pressure and a smaller spraying flow rate, but may correspond to different spraying areas.

[0066] In the above step S002, the spraying area obtained according to the spraying height H and the spraying thickness L may be only one of the local optimal spraying areas among all the spraying areas. Although the local optimal spraying area can ensure that the thickness of the waterproof coating at the current moment (or the current position) is equal to L0, the thickness of the waterproof coating at all positions in the local optimal spraying area may not be equal to L0.

[0067] Therefore, the process of re-obtaining the spraying area in this embodiment is equivalent to re-obtaining other local optimal spraying areas based on the spraying height H and the spraying thickness L. The specific method includes: The spraying pressure and spraying flow rate at the current moment obtained in step S002 are recorded as reference spraying pressure and reference spraying flow rate respectively.

[0068] The same as the method in step S002: The spraying pressure and spraying flow rate are regarded as a set of variable parameters (or as a set of neuron parameters), and the variable parameters are randomly initialized. The variable parameters and the spraying height H are input into the spraying control model, and the spraying control model outputs the spraying thickness and the spraying area, which are recorded as L1 and S1 respectively. The neuron parameters in the spraying control model are kept unchanged (that is, the neuron parameters are frozen so that they cannot be updated and changed). The variable parameters are regarded as neurons that can be updated, and the variable parameters are updated using the following loss function. The loss function F is: The stochastic gradient descent algorithm is used to iteratively update the variable parameters until F is minimized.

[0069] Wherein w represents the spray suppression coefficient. As an example, the current method for obtaining the spray suppression coefficient w includes: In the process of multiple iterative updates of variable parameters, the spraying pressure and spraying flow rate (that is, the variable parameters before each update) input by the spraying control model are obtained and recorded as the first parameter characteristic point, the reference spraying pressure and the reference spraying flow rate are recorded as the second parameter characteristic point, and the Euclidean distance between the first parameter characteristic point and the second parameter characteristic point is recorded as the first characteristic x of the first parameter characteristic point. Let w=exp(-M / (x+1)), where M represents the spraying unevenness, and exp() represents an exponential function with a natural constant as the base.

[0070] The spray suppression coefficient w obtained above is positively correlated with the first feature. When the first feature x is smaller (that is, the variable parameters before updating are close to the reference spray pressure and the reference spray flow), w is smaller, and F is larger at this time, skipping the local optimal situation represented by the reference spray pressure and the reference spray flow, or in other words, it is more necessary to suppress the interference of the spray pressure output by the spray control model when the above-mentioned reference spray pressure and the reference spray flow are used as the input of the spray control model in multiple iterations; when the first feature x is larger (that is, the variable parameters before updating are far away from the reference spray pressure and the reference spray flow), w is larger, and F is smaller at this time, and it can reach the minimum at a position far away from the reference spray pressure and the reference spray flow, or in other words, it is less necessary to suppress the interference of the spray pressure output when the reference spray pressure and the reference spray flow are input into the spray control model in multiple iterations.

[0071] In addition, when the spraying unevenness M is larger, w is smaller, and it is more necessary to make w bring obvious inhibitory effect. When the spraying unevenness is smaller, w is larger, and it is more necessary to make w bring insignificant inhibitory effect.

[0072] When F is minimum, the variable parameters and spraying area S1 at this time are obtained, and the spraying area S1 at this time is used as the re-acquired spraying area at the current moment, and the variable parameters at this time are recorded as the re-acquired spraying pressure and spraying flow at the current moment.

[0073] (3) Repeat the above (1) and (2) processes, that is: obtain the spraying unevenness again according to the re-acquired spraying area, obtain the spraying pressure and spraying flow rate again according to the spraying unevenness, and the spraying area, and so on, until the spraying unevenness is less than or equal to the first preset threshold value, then stop the repeated operation, and use the spraying pressure and spraying flow rate finally obtained as the spraying pressure and spraying flow rate at the current moment of spraying.

[0074] In the process (3), whenever F is the smallest, the spraying pressure and spraying flow rate included in the variable parameters are recorded as a set of control parameters, and the spraying height H input and the output spraying thickness L1 of the spraying control model are recorded as parameter characteristic points of the control parameters.

[0075] (II) The preferred embodiments of the present invention take into account: In the optional embodiment of (I) above, when the spraying pressure, spraying flow rate and spraying area obtained at the current moment have a large spraying unevenness, it is necessary to repeat the processes (1) and (2) multiple times, and each time it is repeated, it is necessary to update the variable parameters through multiple iterations of the spraying control model, and finally regain the spraying pressure and spraying flow rate at the current moment. Therefore, in step S003, when regaining the spraying pressure and spraying flow rate, a certain amount of calculation time is required, during which the spraying personnel may change the position of the nozzle by manipulating the spraying device.

[0076] For the optional embodiment in (I), when executing step S003, if the spraying speed of the sprayer is slow and the position change of the nozzle within the calculated time can be ignored, then the spraying pressure and spraying flow rate re-obtained by the optional embodiment in (I) can still be used as the spraying pressure and spraying flow rate at the current moment (the current moment here refers to the current moment after a certain calculated time). If the spraying speed of the sprayer is fast and the position change of the nozzle within the calculated time cannot be ignored, then the spraying pressure and spraying flow rate re-obtained by the optional embodiment in (I) have a delay, that is, the spraying pressure and spraying flow rate obtained by the optional embodiment in (I) cannot be used as the spraying pressure and spraying flow rate at the current moment (the current moment here refers to the current moment after a certain calculated time).

[0077] The preferred embodiment of the present invention is as follows: For the current moment in the spraying operation, the motion trajectory of the nozzle at the current moment is obtained. The motion trajectory is based on the acceleration sensor and the spraying height and is obtained through the inertial navigation method. The motion trajectory described in the preferred embodiment represents the position and spraying height of the nozzle at each moment in the recent period of time (for example, within 15 seconds).

[0078] The trajectory is predicted based on the motion trajectory to obtain a predicted trajectory between the current time point and the time point T, where the time point T is a preset time point after the current moment. In this embodiment, the time point T is a time point 3 seconds after the current moment.

[0079] The prediction algorithm used in trajectory prediction in this embodiment is the extended Kalman filter (EKF) algorithm. In other embodiments, LSTM neural network can also be used for prediction.

[0080] The predicted trajectory includes the predicted spraying height at each moment between the current time point and time point T (including time point T). The predicted spraying height corresponding to time point T is still recorded as H, and the spraying thickness at the location of the nozzle at time point T is obtained, which is still recorded as L (the acquisition method is the same as step S002).

[0081] It should be noted that the spraying pressure, spraying flow rate and spraying area at each moment are obtained according to the spraying height and spraying thickness at each moment in the above step S002. In the optional embodiment of step S003 (i), the spraying pressure, spraying flow rate and spraying area are re-obtained according to the spraying unevenness in the spraying area.

[0082] The same as steps (1), (2) and (3) in the optional embodiment of (I). This preferred embodiment obtains the spraying pressure, spraying flow rate and spraying area at time point T according to the spraying height and spraying thickness L at time point T, and re-obtains the spraying pressure, spraying flow rate and spraying area according to the spraying unevenness in the spraying area.

[0083] The spraying pressure and spraying flow rate obtained at this time are the spraying pressure and spraying flow rate during spraying at time point T. In some embodiments, when the spraying process reaches time point T, the spraying pressure and spraying flow rate during spraying obtained at time point T are used as the spraying pressure and spraying flow rate during spraying.

[0084] Step S004, the process of multiple iterations of the spray control model runs in the first process, and the spraying task runs in the second process. The second process is used to intervene and adjust the first process. The spraying pressure and spraying flow rate are regained after the first process ends, and are used as the spraying pressure and spraying flow rate when performing the spraying operation in the second process.

[0085] It should be noted that the preferred embodiment of the above step S003 is executed while, that is, the preferred embodiment of step S003 continuously iterates the spray control model using (1), (2) and (3), and also continues to execute the spray task. The spray control model is continuously iterated using (1), (2) and (3) to update the variable parameters, and this process is executed in the first process of the computer; the spray task is executed in the second process of the computer, and both the first process and the second process start to execute at the same time at the current moment. When the first process is completed, the spray pressure and spray flow rate at the time point T can be obtained.

[0086] Specifically: when executing the spraying task in the second process, the spraying height and spraying thickness at each moment are obtained, and the spraying pressure, spraying flow rate and spraying area at each moment are obtained according to the spraying height and spraying thickness according to step S002, and these three are recorded as the actual parameter data points at each moment in the second process.

[0087] The first process includes the predicted trajectory from the current time to time point T. According to the spraying height and spraying thickness at each moment on the predicted trajectory, the spraying pressure, spraying flow rate and spraying area at each moment (that is, at each moment between the current time and time point T) are obtained, and these three are recorded as the predicted parameter data points at each moment in the first process.

[0088] The Euclidean distance between the nozzle position at each moment on the predicted trajectory in the first process and the nozzle position at each moment in the second process is recorded as the first difference. The Euclidean distance between the predicted parameter data point at each moment in the first process and the actual parameter data point at each moment in the second process is recorded as the second difference. The mean of the movement distance of the nozzle at two adjacent moments is obtained, and the ratio of the first difference to the mean is recorded as the first coefficient. Any set of spraying pressure, spraying flow rate and spraying area in the data set is recorded as the sample parameter data point, the maximum Euclidean distance between the sample parameter data points is obtained, and the ratio of the second difference to the maximum Euclidean distance is recorded as the second coefficient. The product of the first coefficient and the second coefficient is recorded as the spraying error of the predicted trajectory. The purpose of obtaining the first coefficient and the second coefficient in this process is to remove the dimension of the first difference and the second difference. In other embodiments, other methods can also be used to remove the dimension of the first difference and the second difference, which is not limited in this embodiment.

[0089] The spraying error of the predicted trajectory is used to adjust the iterative process in the first process. It includes: When the spraying error of the predicted trajectory is less than or equal to the second preset threshold value, it indicates that there is a small difference between the trajectory prediction in the first process and the movement trajectory of the nozzle when the spraying task is executed in the second process, including the difference in position information on the trajectory and the difference in the spraying results. At this time, the spraying pressure and spraying flow rate at the time point T obtained at the end of the first process can be applied to the spraying task in the second process, that is, when the spraying task is executed to the time point T, the spraying pressure and spraying flow rate of the spraying device are set to the spraying pressure and spraying flow rate obtained at the end of the first process.

[0090] In this embodiment, the second preset threshold value is 0.1 as an example for description. In other embodiments, the second preset threshold value may be set to other values, which is not specifically limited in this embodiment.

[0091] When the spraying error of the predicted trajectory is greater than the second preset threshold, it means that the trajectory prediction error in the first process is large, which will cause the spraying pressure and spraying flow rate obtained at the end of the first process to be unsuitable for the spraying device of the nozzle in the second process at time point T. At this time, the present embodiment performs the following processing: The spraying height and spraying thickness obtained at this moment (i.e., the moment when the trajectory prediction error is greater than the second preset threshold) are obtained and recorded as H2 and L2 respectively. At this moment, H and L in step S003 (2) are reset to H2 and L2 respectively.

[0092] It should be noted that when H and L in (2) of step S003 are reset, the purpose is to make the iterative process in the first process adapt to the spraying process of the second process when predicting the spraying error of the trajectory; the disadvantage is that once H and L in (2) of step S003 are reset, the iterative process in the first process needs to pay more iterations, resulting in the inability to end the first process before time point T (that is, the spraying pressure and spraying flow required for spraying at time point T cannot be obtained before time point T).

[0093] This embodiment provides a method for reducing the number of iterations, including: According to the description in the above step S003, when the first process is executed, multiple groups of control parameters are obtained, and each group of control parameters corresponds to a parameter feature point. All parameter feature points are clustered by mean shift to obtain all categories. Each analogy contains at least one parameter feature point. All control parameters corresponding to all parameter feature points in the same category are regarded as a control parameter category, and the mean of all parameter feature points in the same control parameter category is recorded as the average feature point. Each control parameter category corresponds to an average feature point. The control parameter category is some local optimal spraying pressures and spraying flows on the spraying control model. These local optimal spraying pressures and spraying flows are iteratively obtained based on the spraying height and spraying thickness included in the average feature point.

[0094] H2 and L2 are regarded as a target feature point, and the Euclidean distance between the average feature point and the target feature point is used as the iteration suppression coefficient of each control parameter in each control parameter category. Considering that the relationship between the spraying pressure, spraying flow, spraying height and spraying thickness reflected by the spraying control model will not change suddenly and has local similarity (that is, under the same or similar spraying height and spraying thickness, the corresponding spraying pressure and spraying flow will not change too obviously), the larger the iteration suppression coefficient is, the less the control parameters in each control parameter category need to participate in the iteration of the spraying control model and the update of the variable parameters; the smaller the iteration suppression coefficient is, the more the control parameters in each control parameter category need to participate in the iteration of the spraying control model and the update of the variable parameters.

[0095] Then, when the first process iterates the spray control model, it obtains the input variable parameters, the spray pressure and spray flow included in the variable parameters, obtains the Euclidean distance between the variable parameters and each control parameter, and normalizes the iteration suppression coefficient of each control parameter. This embodiment uses the softmax algorithm for normalization, takes the normalized iteration suppression coefficient of each control parameter as the weight, and performs weighted summation on the Euclidean distances between the variable parameters and all control parameters. The product of the weighted summation result and the trajectory prediction error is used as the iteration suppression priority of the variable parameter.

[0096] In the above-mentioned iterative suppression priority process, variable parameters that are far away from the control parameters may need more iterations when iterating. The more difficult it is to use the already obtained control parameters to reduce the number of iterations, the more likely it is that the first process will end after multiple iterations. The greater the iterative suppression priority at this time, the more it is necessary to avoid it from participating in the iteration of the spray control model and the update of the variable parameters; the greater the trajectory prediction error, the greater the difference between the actual spraying process represented by the second process and the predicted spraying process contained in the first process. At this time, after setting H and L in (2) of step S003 to H2 and L2 respectively, the iterative process of the first process will be reset to a large extent, thereby requiring the first process to spend more iteration time. Therefore, the fewer the number of iterations required, the greater the iterative suppression priority.

[0097] Reduce the number of iterations according to the iteration suppression priority of the variable parameter.

[0098] As an example, reducing the number of iterations according to the iteration suppression priority of the variable parameter includes: Obtain the iteration suppression priority of the variable parameters in the most recent several iterations (for example, 30 times), obtain the mean of these iteration suppression priorities, and if the iteration suppression priority of the variable parameters of the current input spray control model is greater than the mean, then re-initialize the variable parameters randomly and perform subsequent iterations.

[0099] As another example, reducing the number of iterations according to the iteration suppression priority of the input parameter includes: A number of (for example, 100) other variable parameters that are closest to the current variable parameter are randomly selected, and among the iteration suppression priorities of these variable parameters, the variable parameter with the smallest iteration suppression priority is used as the input of the spraying control model.

[0100] As a preferred example, reducing the number of iterations according to the iteration suppression priority of the variable parameter includes: Obtain the iteration suppression priority of the variable parameters in the most recent several iterations (for example, 30 times), obtain the mean of these iteration suppression priorities, if the iteration suppression priority of the variable parameters currently input into the spray control model is greater than the mean, randomly select several other variable parameters that are closest to the current variable parameters, and among the iteration suppression priorities of these variable parameters, the variable parameter with the smallest iteration suppression priority is used as the input of the spray control model.

[0101] At this point, the iterative process in the first process is intervened and adjusted through the iterative suppression priority of the variable parameters to reduce the number of iterations. On the one hand, the iterative process in the first process is adapted to the spraying process in the second process, avoiding the situation where the spraying pressure and spraying flow obtained at the end of the first process cannot be used for the spraying process in the second process due to the large trajectory prediction error in the first process. On the other hand, by recalculating the spraying pressure and spraying flow in the first process many times, the spraying pressure and spraying flow finally obtained can ensure that the waterproof coating sprayed on the wall or the ground has as uniform a thickness as possible. On the other hand, by combining the spraying process of the second process with the spraying prediction process of the first process, the spraying personnel can spray in a small and difficult-to-access space while ensuring the uniformity of the spraying.

[0102] Furthermore, when the first process ends, that is, when (3) in the preferred embodiment of step S003 ends, on the one hand, the obtained spraying pressure and spraying flow rate are recorded, and when the spraying task in the second process reaches time point T, the spraying pressure and spraying flow rate recorded at the end of the first process are used as the spraying pressure and spraying flow rate of the spraying device. On the other hand, when the first process ends, this moment is used as the current moment, and the method in step S003 (ii) is similar to that in step S003, the first process is started again, including: The predicted trajectory between the current time point and time point T is obtained again, and at this time point T is still a preset time point after the current moment. The predicted spraying height corresponding to time point T is still recorded as H, and the spraying thickness at the location of the nozzle at time point T is obtained, which is still recorded as L (the acquisition method is the same as step S002). The spraying pressure, spraying flow rate and spraying area at time point T are obtained according to the spraying height and spraying thickness L at time point T, and the spraying pressure, spraying flow rate and spraying area are re-obtained according to the spraying unevenness under the spraying area.

[0103] At the same time, the second process continues the spraying task, specifically including: obtaining the spraying height and spraying thickness at each moment, obtaining the spraying pressure, spraying flow and spraying area at each moment according to the spraying height and spraying thickness, and spraying according to the spraying pressure and spraying flow at each moment.

[0104] Finally, according to the method in step S004, the second process is intervened and adjusted in the first process to reduce the number of iterations of the first process. When the second process ends again, the second process is started again in the same way, and so on, until the spraying process ends.

[0105] Embodiment three: In the above embodiment, the spraying pressure and the spraying flow rate can be re-obtained through the unevenness every time the first process ends (equivalent to every certain time period).

[0106] As a feasible example, methods of reducing the length of the time period (or further terminating the first process as soon as possible) include: When spraying, the sprayer should try to avoid unpredictable spraying trajectories, such as swinging or moving the nozzle regularly, and spraying at a slower speed, such as moving the nozzle 10 cm per second, to avoid large spraying errors in the predicted trajectory in step S004. Alternatively, use a smart car or drone to carry a spraying device, and then spray according to the set motion trajectory; it should be noted that this method may not be applicable to small areas in old communities.

[0107] As a preferred example, the method of reducing the length of the time period (or further terminating the first process as soon as possible) includes: The time difference between the start and end of the first process is recorded as T1. Then, when the first process is started again next time, the corresponding time point T is set to the time point corresponding to the time period T1 after the current time.

[0108] In other optional examples, methods for reducing the length of the time period also include: increasing the computing power of the computer, or reducing the number of parameters of the spray control model, or pruning the spray control model. Even in some old communities or small areas that are difficult for people to enter, when the requirements for the accuracy of spraying uniformity or spraying thickness are not high, the spray control model can be set to a polynomial model.

[0109] It should also be noted that if the first process has not ended when it reaches time point T, the variable parameters iteratively updated by the spray control model at time point T are directly used as the spray pressure and spray flow required for spraying at time point T, and the first process is ended. In other embodiments, the spray pressure and spray flow obtained at time point T in the second process can also be used as the spray pressure and spray flow required for spraying at time point T, and the first process is ended.

[0110] Embodiment 4: Step S004 includes the following process: when executing the spraying task in the second process, the spraying height and spraying thickness at each moment are obtained, and the spraying pressure, spraying flow rate and spraying area at each moment are obtained according to the spraying height and spraying thickness.

[0111] In this embodiment, the degree of uneven spraying is obtained according to the spraying area. When the degree of uneven spraying is greater than the first preset threshold value, it indicates that there may be uneven spraying when spraying according to the spraying pressure and spraying flow rate. Although this uneven situation can be solved by re-obtaining the spraying pressure and spraying flow rate in the first process, the first process at this time may not be finished yet. In this embodiment, in order to avoid this uneven situation, the spraying thickness is reduced by 10%, and then the spraying pressure and spraying flow rate are re-obtained. Although this method reduces the spraying thickness, the spraying pressure, spraying flow rate and spraying area calculated at the next moment of spraying can correct the thickness reduced at the previous moment. The reason is that: as can be seen from step S002, the calculation method of the spraying thickness is based on the previously sprayed thickness (i.e., the formula L=L0-L1 in step S002), so when the sprayed thickness is reduced, the spraying thickness at the next moment will increase, thereby correcting the thickness reduced at the previous moment.

[0112] Embodiment five: In the process (3) of step S003, whenever F is minimum, the spraying pressure and spraying flow rate included in the variable parameters are recorded as a set of control parameters, and the spraying height H and the output spraying thickness L1 of the spraying control model as input are recorded as parameter feature points of the control parameters.

[0113] It should be noted that when processes (1) and (2) are repeated, spraying areas with large spraying unevenness may be obtained multiple times, which makes it impossible to quickly obtain the spraying pressure and spraying flow rate at the current spraying moment.

[0114] In this embodiment, whenever F reaches the minimum, the spray unevenness corresponding to the spray area output by the spray control model is recorded, and the spray suppression coefficient w calculated before F reaches the minimum each time is recorded. All recorded spray unevennesses are normalized using the softmax method, and the normalized spray unevenness is used as the weight to perform weighted summation on all recorded spray suppression coefficients w, which is recorded as the historical average suppression coefficient w0. When F reaches the minimum again, the spray unevenness is sprayed again according to the spray area output by the spray control model. If the spray unevenness is still greater than the first preset threshold, the spray suppression coefficient w is obtained again using the method in (2), and the minimum value between w0 and w is used as the spray suppression coefficient for subsequent iterations. This process obtains the spray suppression coefficient for the next iteration by combining the distribution of the spray suppression coefficient in the historical iteration process, thereby avoiding the situation where spray areas with large spray unevenness are obtained multiple times.

[0115] Embodiment six: In this embodiment, the following method is used to reduce the atomization of the waterproof coating during the spraying process as much as possible: In step S003 (2), after the spray suppression coefficient w is calculated, the spray height and spray pressure obtained at this time are recorded as h0 and p0 respectively, and the dimensions of h0 and p0 are removed to obtain h1 and p1. w×exp(-h0 / p0) is used as the updated spray suppression coefficient. The updated spray suppression coefficient is used to update the variable parameters in subsequent iterations.

[0116] The larger h0 is and the smaller p0 is, the more likely it is that atomization will occur. At this time, the smaller the updated spray suppression coefficient is, the more likely it is to avoid spraying with a smaller spray pressure during the spraying process. This embodiment avoids atomization as much as possible while avoiding uneven spraying, reduces the emission of volatile organic compounds (VOCs), and ensures efficient application of the coating.

[0117] It should also be noted that all embodiments of the present invention are implemented in windless weather. When there is wind in the spraying scene, the atomization of the waterproof coating will become serious. In some embodiments, the spraying operation area can be enclosed by a windshield before the spraying operation is carried out.

[0118] As an example, the method of removing the dimension of h0 and p0 is: obtain the maximum value of all spraying heights in the data set, and the ratio of h0 to the maximum value is taken as h1; obtain the maximum value of all spraying pressures in the data set, and the ratio of p0 to the maximum value is taken as p1.

[0119] In other embodiments, this method can be used to remove the influence of dimension when two different physical quantities are used for calculation. This embodiment will not be described in detail one by one.

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for spraying an environmentally friendly waterproof coating, characterized in that: The method comprises the following steps: Train a spray control model that inputs spray height, spray pressure, and paint flow rate, and outputs spray thickness and spray area; In the second process, a spraying operation is performed, and the spraying pressure, spraying flow rate and spraying area at each moment are obtained according to the spraying height H and the spraying thickness L at each moment, including: inputting the variable parameters consisting of the spraying pressure and the spraying flow rate and the spraying height H into the spraying control model, and recording the spraying thickness output by the spraying control model as L1, updating the variable parameters through multiple iterations of the spraying control model until the loss function F is minimized, and using the updated variable parameters and the spraying area output by the spraying control model as the spraying pressure, spraying flow rate and spraying area at each moment, wherein F is the difference between L1 and L; The first process includes: D1: Get the predicted trajectory of the nozzle between the current moment and time point T, where time point T is a preset time point after the current moment, and the predicted trajectory includes the spraying height at time point T; D2: According to the spraying height H and spraying thickness L at time point T, the spraying pressure, spraying flow rate and spraying area at time point T are obtained; according to the spraying area at time point T, the spraying unevenness is obtained; D3: According to the spraying unevenness, the spraying pressure, spraying flow rate, spraying area, and spraying unevenness at the time point T are obtained by repeatedly executing D2, until the first process is terminated when the spraying unevenness is less than or equal to the first preset threshold value; In the process of repeatedly executing D2, the variable parameter of F each time reaching the minimum value is recorded as the control parameter; according to the motion trajectory of the nozzle in the second process and the trajectory prediction error of the predicted trajectory in the first process and the control parameter when reaching the minimum value all times, the number of iterations of the spray control model in the first process is reduced; The spraying pressure and spraying flow rate regained after the first process is completed are used as the spraying pressure and spraying flow rate at time point T when the spraying operation is performed in the second process.

2. The method for spraying an environmentally friendly waterproof coating according to claim 1, characterized in that: The method of re-obtaining the spraying pressure, spraying flow rate, spraying area, and spraying unevenness at the time point T by repeatedly executing D2 according to the spraying unevenness until the spraying unevenness is less than or equal to the first preset threshold value ends the first process, including the following specific steps: After the first execution of D2, if the spraying unevenness is less than or equal to the first preset threshold, the first process is directly terminated. If the spraying unevenness is greater than the first preset threshold, D2 is executed again according to the spraying unevenness and the spraying pressure, spraying flow rate and spraying area at the time point T are re-obtained. The spraying unevenness is obtained again according to the re-obtained spraying area, and so on, until the first process is terminated when the spraying unevenness is less than or equal to the first preset threshold; According to the spraying unevenness, D2 is executed again to re-obtain the spraying pressure, spraying flow rate and spraying area at the time point T, wherein the specific process also includes: iterating the spraying control model multiple times to update the variable parameters until the loss function F is minimized; wherein the method for obtaining the loss function F specifically includes: the spraying thickness at the time point T is also represented as L, ; Wherein w represents the spraying suppression coefficient, and the spraying suppression coefficient is negatively correlated with the spraying unevenness.

3. The method for spraying an environmentally friendly waterproof coating according to claim 2, characterized in that: The specific steps of obtaining the spraying suppression coefficient include: The variable parameters input by the spray control model are obtained and recorded as the first parameter feature point. The spray pressure and spray flow rate obtained in D2 are recorded as the second parameter feature point. The Euclidean distance between the first parameter feature point and the second parameter feature point is recorded as the first feature of the first parameter feature point. The spray suppression coefficient is negatively correlated with the spray unevenness and positively correlated with the first feature.

4. An environmentally friendly waterproof coating spraying method according to any one of claims 1, 2 or 3, characterized in that: The specific steps for obtaining the spraying unevenness are as follows: For the spraying areas obtained at all times before time point T, the sum of the spraying thickness of each position in all spraying areas is recorded as the sprayed thickness of each position; The spraying area obtained at time point T is recorded as the current spraying area. The spraying thickness of each position in the current spraying area plus the sprayed thickness of each position is taken as the predicted spraying thickness of each position. The spraying unevenness is positively correlated with the variance of the predicted spraying thickness of all positions in the current spraying area.

5. The method for spraying an environmentally friendly waterproof coating according to claim 1, characterized in that: The specific steps of obtaining the trajectory prediction error between the motion trajectory of the nozzle in the second process and the predicted trajectory in the first process are as follows: When the spraying task is executed in the second process, the spraying pressure, spraying flow rate and spraying area obtained at each moment are recorded as actual parameter data points; the spraying pressure, spraying flow rate and spraying area obtained at each moment on the predicted trajectory in the first process are recorded as predicted parameter data points; The Euclidean distance between the nozzle position at each moment on the predicted trajectory in the first process and the nozzle position at each moment in the second process is recorded as the first difference; the Euclidean distance between the predicted parameter data point at each moment in the first process and the actual parameter data point at each moment in the second process is recorded as the second difference; the spraying error of the predicted trajectory is positively correlated with the first difference and the second difference, respectively.

6. The method for spraying an environmentally friendly waterproof coating according to claim 1, characterized in that: The method of reducing the number of iterations of the spray control model in the first process according to the motion trajectory of the nozzle in the second process and the trajectory prediction error of the predicted trajectory in the first process and the control parameters when all times reach the minimum value includes the following specific steps: When the trajectory prediction error is greater than the second preset threshold, the spraying height and spraying thickness obtained in the second process are recorded as H2 and L2 respectively; H2 and L2 are used as target feature points; At the same time, for the spraying height H and the spraying thickness L at the time point T during the repeated execution of D2, the values ​​of the spraying height H and the spraying thickness L at the time point T are reset to H2 and L2 respectively; For each control parameter when F reaches a minimum value, the spraying height and spraying thickness corresponding to the control parameter are recorded as parameter characteristic points of the control parameter; In the process of repeatedly executing D2, for the variable parameters input in each iteration of the spray control model, the iteration suppression priority of the variable parameters is obtained according to the distance between the variable parameters and all control parameters, the distance between the parameter feature point of the control parameter when F reaches the minimum value all times and the target feature point, and the trajectory prediction error; The number of iterations of the spray control model in the first process is reduced according to the iteration suppression priority of the variable parameter.

7. The method for spraying an environmentally friendly waterproof coating according to claim 6, characterized in that: The specific steps of obtaining the iterative suppression priority of the variable parameter according to the distance between the variable parameter and all the control parameters, the distance between the parameter feature point of the control parameter when F reaches the minimum value for all times and the target feature point, and the trajectory prediction error are as follows: The iterative suppression coefficient of each control parameter is obtained by using the distance between the parameter feature point of the control parameter when F reaches the minimum value for all times and the target feature point; The Euclidean distance between the variable parameter and each control parameter is obtained, the iterative suppression coefficient of each control parameter is normalized, the normalized iterative suppression coefficient of each control parameter is used as the weight, the Euclidean distance between the variable parameter and all control parameters is weighted summed, and the product of the weighted summation result and the trajectory prediction error is used as the iterative suppression priority of the variable parameter.

8. The method for spraying an environmentally friendly waterproof coating according to claim 7, characterized in that: The iterative suppression coefficient of each control parameter is obtained by using the distance between the parameter feature point and the target feature point when the control parameter reaches the minimum value all the times, and the specific steps include the following: For the parameter feature points of the control parameters when F reaches the minimum value all the times; perform mean shift clustering on all the parameter feature points to obtain all categories, and all the control parameters corresponding to all the parameter feature points in the same category are taken as a control parameter category, and the mean of all the parameter feature points in the same control parameter category is recorded as the average feature point; the Euclidean distance between the average feature point and the target feature point is taken as the iterative suppression coefficient of each control parameter in each control parameter category.

9. The method for spraying an environmentally friendly waterproof coating according to claim 6, characterized in that: The method of reducing the number of iterations of the spraying control model in the first process according to the iteration suppression priority of the variable parameter includes the following specific steps: The mean of the iteration suppression priorities of the variable parameters in the most recent iterations is obtained, and the variable parameters input into the spray control model in the current iteration are recorded as the current variable parameters; if the iteration suppression priority of the current variable parameter is greater than the mean, several other variable parameters closest to the current variable parameter are randomly selected, and among the iteration suppression priorities of the other variable parameters, the variable parameter with the smallest iteration suppression priority is used as the input of the spray control model.

10. An environmentally friendly waterproof coating spraying device, comprising a spray head, a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: An ultrasonic distance measuring sensor and an acceleration sensor are installed on the nozzle, which are used to obtain the spraying height and the movement trajectory of the nozzle respectively. When the processor executes the computer program, the steps of an environmentally friendly waterproof coating spraying method as described in any one of claims 1 to 9 are implemented.

Images