A multi-parameter regulation method and system for a vacuum insulation board insulation layer spraying robot

Abstract

The application provides a kind of vacuum insulation board insulating layer spraying robot multi-parameter regulation method and system, related to the technical field of multi-parameter regulation, which dynamically senses and regulates the interference relationship and trajectory deviation influence between multiple spray heads during spraying, effectively solving the problems of uneven spraying thickness and poor spatial consistency. The application can achieve fine adjustment of spray head parameters, allowing each spray head to maintain an ideal spraying state under different working conditions, thereby improving overall spraying precision and stability. By introducing a multi-parameter joint regulation mechanism, the application can significantly reduce spraying defects caused by factors such as attitude deviation and coupling interference, reducing rework rate and material waste.

Description

Technical Field

[0001] This invention relates to the field of multi-parameter control technology, and more specifically, to a multi-parameter control method and system for a vacuum insulation panel insulation layer spraying robot. Background Technology

[0002] With the continuous improvement of building energy efficiency standards and the widespread application of new insulation materials, vacuum insulation panels have become one of the core materials for high-performance insulation systems due to their excellent thermal resistance and good structural compatibility. In actual manufacturing, to ensure the sealing, durability, and thermal stability of the VIP edges and surface during subsequent use, a uniform insulation layer with a certain thickness control precision is often required to be sprayed onto its surface or edge areas. Currently, industrial spraying robots are widely used in this type of precision spraying operation, offering advantages such as high automation and good repeatability. However, due to the complex dimensions and large curvature variations in the corner areas of vacuum insulation panels, the nozzle array experiences coupling disturbance effects during multi-angle, multi-path operations, leading to localized thinning or over-thickness of the coating thickness, thus affecting the overall performance of the insulation panel. Meanwhile, existing spraying control methods are mostly based on single-parameter adjustment models, making it difficult to cope with the multi-dimensional precision feedback requirements under dynamic operating conditions and to achieve closed-loop adjustment of spraying quality.

[0003] In existing technologies, although some literature has proposed methods for correcting spraying paths based on robot pose feedback or using neural networks to predict and correct spraying thickness, there are still key technical deficiencies: First, the actual coupling relationship between nozzles is often not modeled by the system, making it difficult to effectively quantify the interference effect of adjacent nozzles, thus forming non-uniform coating accumulation or gaps in complex surface areas; Second, during dynamic spraying, the nozzle trajectory is affected by factors such as the lag in the dynamic response of the robotic arm and material adhesion rebound, and the actual spraying trajectory often deviates from the theoretical path, but most current control models only perform static compensation and lack continuous modeling and control of dynamic disturbance factors; Third, most existing spraying control systems are based on fixed parameter tables and lack adaptive update mechanisms, making it impossible to combine coating detection results for feedback correction, resulting in insufficient system control accuracy and difficulty in achieving synchronous optimization of spraying thickness and uniformity.

[0004] Therefore, there is an urgent need for an intelligent spraying control method that can integrate nozzle array relationship modeling, trajectory disturbance modeling, and multi-parameter feedback control. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention is proposed. This invention provides a multi-parameter control method and system for a robot spraying insulation layer onto a vacuum insulation panel.

[0006] According to one aspect of the present invention, a multi-parameter control method for a vacuum insulation panel insulation layer spraying robot is provided, comprising: modeling the spatial layout of the robot nozzle array and generating a nozzle array correlation matrix to represent the coupling effect between adjacent nozzles based on historical spraying data; acquiring nozzle trajectories in real time during the spraying process, calculating offset vectors based on pose errors, and constructing a trajectory disturbance influence factor matrix; constructing a three-parameter joint fine-tuning matrix based on the correlation matrix and the disturbance matrix; dynamically adjusting the current nozzle according to the nozzle number based on the parameter correction amount output by the joint fine-tuning matrix to achieve synchronous control of spraying thickness and spatial uniformity; updating the nozzle disturbance influence matrix based on the acquired actual coating thickness difference and uniformity deviation, and transmitting it back to the joint fine-tuning matrix to complete the next round of control.

[0007] Furthermore, the modeling of the spatial layout of the robot nozzle array includes: forming a set of spatial arrangement parameters by calibrating the relative position coordinates, pitch angle and array spacing of the robot nozzle array under a standard spraying task.

[0008] Furthermore, the construction of the nozzle array correlation matrix includes: performing statistical analysis on the overlapping segments of the trajectories of adjacent nozzles in the historical dataset, combining the coating thickness deviation value to determine the interference weight of each nozzle on its neighboring nozzles, and extracting it as a coupling weight vector; taking the spatial arrangement parameter group and the coupling weight vector as input, normalizing and fusing the spatial distance of the pairwise combination of nozzle numbers with the historical interference intensity to generate a nozzle array correlation matrix for representing the coupling relationship.

[0009] Furthermore, the calculation of the offset vector for the pose error includes: based on the current pose parameter set Q of the nozzle recorded by the spray controller, combined with the target pose parameter set Q0 set by the task, calculating the pose error term ΔQ of the nozzle in the six-degree-of-freedom space; by projecting the linear offset part of the pose error term ΔQ onto the spray axis direction, extracting the actual spray axis offset vector V of each nozzle as a physical quantity characterizing the actual deviation trend of the spray beam.

[0010] Furthermore, the three parameters include paint mixing parameters, spraying speed, and nozzle movement speed.

[0011] Furthermore, the construction of the joint fine-tuning matrix includes: using the nozzle number as the axis, extracting the nozzle coupling degree and offset trend from the correlation matrix and the perturbation influence factor matrix to construct a perturbation state tensor T; each dimension of the perturbation state tensor T corresponds to the interference direction, spraying displacement trend and physical position index between nozzles; setting a basic spraying parameter set P0 = {α0, β0, γ0}, which corresponds to the default paint mixing parameters, spraying speed and nozzle moving speed, respectively; based on the interference direction and perturbation offset direction of each nozzle in the perturbation state tensor T, constructing a nozzle parameter adjustment offset set ΔP = Δα, Δβ, Δγ}; using the nozzle number as the index, combining the basic parameter set P0 and the offset set ΔP, generating a three-parameter joint fine-tuning matrix M, where each row is the three-parameter adjustment result corresponding to the nozzle.

[0012] Furthermore, the dynamic adjustment of the current nozzle includes: extracting the parameter correction group corresponding to each nozzle based on the nozzle number index and the fine-tuning matrix M, which respectively correspond to the paint mixing parameters, spraying speed and nozzle moving speed settings of the current nozzle; and adjusting the paint mixing parameters, spraying speed and nozzle moving speed according to the parameter group instructions.

[0013] Furthermore, the actual coating thickness difference and uniformity deviation include: collecting thickness measurement data in the completed spraying area, and extracting the thickness difference and local uniformity deviation of the corresponding area of ​​each nozzle by comparing it with the preset target thickness template; and combining the thickness difference and local uniformity deviation of each nozzle to construct a deviation vector.

[0014] Furthermore, the update of the nozzle disturbance influence matrix includes: inputting the deviation vector into the trajectory disturbance influence factor matrix, and performing an exponential correction on the nozzle corresponding item in the disturbance matrix.

[0015] According to another aspect of the present invention, a multi-parameter control system for a vacuum insulation panel insulation layer spraying robot is provided, comprising: an array modeling module for modeling the spatial layout of the robot nozzle array and generating a nozzle array correlation matrix representing the coupling effect between adjacent nozzles based on historical spraying data;

[0016] The trajectory acquisition module is used to acquire the nozzle trajectory in real time during the spraying process, calculate the offset vector based on the pose error, and construct the trajectory disturbance influence factor matrix.

[0017] The fine-tuning matrix module is used to construct a three-parameter joint fine-tuning matrix based on the correlation matrix and the perturbation matrix;

[0018] The synchronous control module is used to dynamically adjust the current nozzle according to the nozzle number based on the parameter correction amount output by the joint fine-tuning matrix, so as to achieve synchronous control of the coating thickness and spatial uniformity.

[0019] The error feedback module is used to update the nozzle disturbance influence matrix based on the actual coating thickness difference and uniformity deviation, and then feed it back to the joint fine-tuning matrix to complete the next round of adjustment.

[0020] Compared with existing technologies, the multi-parameter control method and system for vacuum insulation panel insulation layer spraying robot provided by this invention effectively solves the problems of uneven spraying thickness and poor spatial consistency by dynamically sensing and controlling the interference relationship and trajectory deviation effects among multiple nozzles during the spraying process. It enables fine-tuning of nozzle parameters, allowing each nozzle to maintain an ideal spraying state under different working conditions, thereby improving overall spraying accuracy and stability. By introducing a multi-parameter joint control mechanism, this invention can significantly reduce spraying defects caused by factors such as posture deviation and coupling interference, reducing rework rate and material waste. Compared with traditional methods that struggle to cope with real-time disturbances and local error accumulation during the spraying process, this invention possesses stronger adaptive capabilities and control robustness, making it particularly suitable for high-requirement applications such as vacuum insulation panel insulation layers. It achieves simultaneous optimization of spraying thickness and spatial uniformity, effectively improving the forming quality and production yield of vacuum insulation panel insulation layers. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0022] Figure 1 This is a system block diagram of a multi-parameter control method for a vacuum insulation panel insulation layer spraying robot according to an embodiment of the present invention.

[0023] Figure 2 This is a flowchart illustrating the calculation of the offset vector based on pose error in the multi-parameter control method for a vacuum insulation panel insulation layer spraying robot according to an embodiment of the present invention.

[0024] Figure 3 This is a structural diagram of a multi-parameter control system for a vacuum insulation panel insulation layer spraying robot according to an embodiment of the present invention. Detailed Implementation

[0025] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0026] As mentioned in the background section, although some literature has proposed methods for correcting spraying paths based on robot pose feedback or using neural networks to predict and correct spraying thickness, key technical shortcomings remain: First, the actual coupling relationship between nozzles is often not modeled by the system, making it difficult to effectively quantify the interference effect of adjacent nozzles, resulting in non-uniform coating accumulation or gaps in complex surface areas; Second, during dynamic spraying, the nozzle trajectory is affected by factors such as the lag in the dynamic response of the robotic arm and material adhesion rebound, often causing errors between the actual spraying trajectory and the theoretical path. However, most current control models only perform static compensation, lacking continuous modeling and control of dynamic disturbance factors; Third, most existing spraying control systems are based on fixed parameter tables and lack adaptive update mechanisms, failing to combine coating detection results for feedback correction, resulting in insufficient system control accuracy and difficulty in achieving simultaneous optimization of spraying thickness and uniformity. Therefore, there is an urgent need for an intelligent spraying control method that integrates nozzle array relationship modeling, trajectory disturbance modeling, and multi-parameter feedback control.

[0027] Figure 1 This is a flowchart illustrating a multi-parameter control method for a robot spraying insulation layer on a vacuum insulation panel according to an embodiment of the present invention. Figure 1 As shown, the multi-parameter control method for vacuum insulation panel insulation layer spraying robot includes: S1: Modeling the spatial layout of the robot nozzle array and generating a nozzle array correlation matrix to represent the coupling effect between adjacent nozzles based on historical spraying data; S2: Real-time acquisition of nozzle trajectory during spraying, calculation of offset vector based on pose error, and construction of trajectory disturbance influence factor matrix; S3: Construction of three-parameter joint fine-tuning matrix based on correlation matrix and disturbance matrix; S4: Dynamically adjusting the current nozzle according to nozzle number based on parameter correction amount output by joint fine-tuning matrix to achieve synchronous control of spraying thickness and spatial uniformity; S5: Updating nozzle disturbance influence matrix based on actual coating thickness difference and uniformity deviation, and sending it back to joint fine-tuning matrix to complete the next round of control.

[0028] In this embodiment of the invention, step S1 specifically includes modeling the spatial layout of the robot nozzle array and generating a nozzle array correlation matrix based on historical spraying data to represent the coupling effect between adjacent nozzles. The correlation matrix is ​​used to quantify the degree of interference of the nozzles on the adjacent spraying trajectories under relative motion.

[0029] Preferably, the modeling of the spatial layout of the robot nozzle array includes: calibrating the relative position coordinates, pitch angle and array spacing of the robot nozzle array under a standard spraying task to form a spatial arrangement parameter set, which provides a spatial distribution basis for the nozzle array association modeling.

[0030] First, before modeling the nozzle array, it is necessary to clarify the structural form of the painting robot used in this invention and the basic principles of nozzle arrangement. In conventional painting operations, nozzles are generally arranged in a linear array or rectangular grid, suitable for the need for uniform coverage of large surface areas. However, these arrangements are often limited to a two-dimensional planar coordinate system in spatial modeling, lacking consideration of the "spatial offset" caused by spatial pitch angle, nozzle angle, and assembly errors in actual operations. Therefore, to more comprehensively depict the actual arrangement of nozzles in space, this invention uses a spatial configuration parameter set composed of "three-dimensional spatial coordinates + pitch attitude angle + array spacing" as the modeling basis. The acquisition of this parameter set relies on the combination of the robot's own motion control system and laser calibration device. Specifically, the motion control unit first moves the nozzles sequentially to multiple positions on a standard reference trajectory, and combines external positioning equipment to obtain the three-dimensional coordinates and attitude angle information of each nozzle in this process, thereby forming a multi-dimensional parameter record set including nozzle number, X / Y / Z coordinates, pitch angle, yaw angle, and distance to adjacent nozzles. The advantage of this approach is that, during modeling, not only is the location of the nozzle considered, but also the angle at which it faces the target area is clearly defined.

[0031] Next, based on the established nozzle spatial distribution model, a data model reflecting the mutual influence between nozzles needs to be constructed. Typically, in multi-nozzle collaborative operations, interactions such as overlapping coating areas, airflow disturbances, and edge tailing are inevitable between nozzles. These interactions often manifest as microscopic defects such as inconsistent coating thickness and rough edges within a short period. To accurately capture the true manifestations of these interactions, this invention designs a nozzle correlation matrix generation method based on historical spraying task records. Specifically, historical spraying trajectory data and corresponding thickness distribution records are extracted from archived spraying tasks to form a dataset. This dataset includes data samples of nozzle numbers, motion trajectories, control parameters, and finished product thickness distribution, used to deduce the degree of mutual interference between nozzles and the thickness offset response.

[0032] Ideally, statistical analysis is performed on the overlapping sections of the trajectories of adjacent nozzles in the historical dataset. Combined with the coating thickness deviation value, the interference weight of each nozzle on its neighboring nozzles is determined and extracted as a coupling weight vector.

[0033] It should be noted that, to further refine this interference trend, this invention proposes a coupled weight vector construction mechanism to quantify the influence of each nozzle on the coating quality of its neighboring nozzles. Specifically, all nozzle pairs with overlapping or intersecting trajectories (such as nozzle i and nozzle j) are first selected. Then, the mean and variance of the thickness deviation values ​​(i.e., the difference between theoretical and measured thickness) in their corresponding regions are statistically analyzed to summarize the interference intensity. For example, if nozzle j operates after nozzle i, and its coating thickness in the overlapping region is generally higher than the standard value, it indicates that the exhaust gas or solvent evaporation of nozzle i affects subsequent nozzles. This type of interference effect is characterized by a vector form, defined as W. ij represents the degree of influence of nozzle i on nozzle j. The interference weights between all nozzle pairs will be integrated into a two-dimensional matrix W∈R. n×n Where n is the total number of nozzles. It is worth noting that at this stage, the invention uses an asymmetric matrix representation, i.e., W. ij ≠W ji This is to accurately reflect the directional impact of the spraying order on the disturbance.

[0034] Finally, using the constructed spatial configuration parameter set and coupling weight vector as input, the spatial distance between each pair of nozzle numbers and the historical interference intensity are normalized and fused to generate a nozzle array correlation matrix to represent the coupling relationship.

[0035] As can be seen, this invention systematically establishes a high-dimensional, adjustable-precision nozzle array coupling model from four levels: nozzle space modeling, historical data extraction, interference weight derivation, and multi-factor fusion modeling. It not only reflects the potential impact of nozzle physical arrangement on coating consistency but also introduces historical coating performance as a weight reference, laying a solid data foundation for subsequent dynamic parameter control.

[0036] In this embodiment of the invention, S2 specifically includes real-time acquisition of the nozzle trajectory during the spraying process, calculation of the offset vector based on the pose error, and construction of a trajectory disturbance influence factor matrix to describe the spraying offset trend caused by disturbances in each nozzle. Its main purpose is to dynamically sense the degree of deviation between the actual execution trajectory of the nozzle and the task-set trajectory during the real-time execution of the spraying task.

[0037] Firstly, during the operation of the painting robot, each nozzle is typically mounted at the end of a multi-degree-of-freedom robotic arm. The nozzle moves along complex paths within the workspace, and its trajectory is affected by various factors such as drive errors, mechanical backlash, and inertial delays. Therefore, it is necessary to record the spatial position and attitude information of the nozzle in real time using a high-frequency sensing acquisition device.

[0038] Preferably, the step of calculating the offset vector based on the pose error includes: using the current pose parameter set Q of the nozzle recorded by the spray controller as a basis, and combining it with the target pose parameter set Q0 set by the task, to calculate the pose error term ΔQ of the nozzle in the six-degree-of-freedom space; and by projecting the linear offset part of the pose error term ΔQ onto the spray axis direction, extracting the actual spray axis offset vector V of each nozzle as a physical quantity characterizing the actual deviation trend of the spray bundle.

[0039] Specifically, the current pose parameter set Q of each nozzle contains six dimensions: three-dimensional linear position coordinates (x, y, z) and three-dimensional attitude angles (pitch, yaw, roll). These parameters can be obtained through an optical tracking system, encoder, or IMU (Inertial Measurement Unit). Compared to conventional processes that only pre-plan and offline record trajectory path points, this invention uses a real-time continuous acquisition strategy to dynamically acquire actual trajectory points with millisecond-level time resolution, ensuring data density and timeliness. Furthermore, to calculate the deviation between the actual motion state of the nozzle and the set trajectory, the target pose parameter set, i.e., the ideal nozzle trajectory predefined in the task, needs to be provided simultaneously. By comparing the current pose Q with the target pose point by point, the error vector ΔQ in the six-degree-of-freedom space can be solved.

[0040] As is known, the spraying behavior performed by the nozzle is essentially a directional distribution process of a material jet (such as a paint atomization flow), and the spraying effect is closely related to the axial consistency of the nozzle. Even a small spatial displacement, if the directionality changes, may lead to offset of the sprayed area, change of spraying angle, or uneven thickness. Therefore, this invention extracts only the linear displacement term (Δx, Δy, Δz) from the six-dimensional error ΔQ and projects it onto the current spraying direction axis of the nozzle (usually defined as the nozzle center axis, i.e., the Z direction). This projection operation is completed by vector dot product: if the unit direction of the spraying axis is m, then the offset vector is V = (Δx, Δy, Δz) × m × m, that is, the projection of the linear offset in the spraying direction multiplied by the unit vector of the direction. Through this operation, the three-dimensional spatial offset can be converted into an effective deviation of the spraying axis. This processing method can more realistically reflect the impact of nozzle deviation on the actual spraying thickness and trajectory path.

[0041] A better approach is to weight V based on the coupling weights of adjacent nozzles in the nozzle array correlation matrix to generate an interference adjustment vector group U, which is used to characterize the changing trend of single nozzle disturbance propagation to its neighboring area.

[0042] This physical vector V not only provides the directionality of the deviation trend (whether it's inward or outward deviation), but also reflects the magnitude of the deviation. In practical applications, when the V value is significantly positive or negative, it indicates that the nozzle currently has a deviation trend outside or inside the target spraying path, causing accumulation or loss of thickness. Therefore, the construction of the offset vector V is a core quantity constituting the physical representation in perturbation modeling.

[0043] Since painting robots typically employ a multi-nozzle array structure, the spraying areas of each nozzle overlap and influence each other. Even if the pose of one nozzle is disturbed, it will affect the thickness or uniformity of the coverage area of ​​adjacent nozzles. Therefore, the V vector cannot be viewed in isolation; its disturbance effect must be projected onto the neighboring area. To this end, this invention utilizes the nozzle array correlation matrix generated in the first step, where each element represents the coupling influence weight of nozzle i on nozzle j. By weighted diffusion of the V value vector of each nozzle, a disturbance adjustment vector group U can be formed. This process simulates the disturbance propagation path, i.e., the degree of influence of the actual offset of a nozzle on its neighboring spraying area. Compared to the traditional direct correction of the V vector, this scheme can more accurately capture the disturbance resonance and local amplification phenomenon under array coupling conditions. The establishment of the U vector group not only preserves the physical dimensions of V but also reasonably weights the error propagation direction, improving the accuracy of disturbance trend assessment, especially in multi-nozzle high-speed collaborative operation scenarios.

[0044] Finally, using the nozzle number as an index, a corresponding disturbance influence factor matrix D is constructed, with each row corresponding to the disturbance weight distribution of a nozzle.

[0045] S3: Based on the correlation matrix and the disturbance influence factor matrix, a three-parameter joint fine-tuning matrix is ​​constructed by utilizing the disturbance mapping relationship between the interference trend and the offset trend between nozzles. The joint fine-tuning matrix is ​​used to dynamically couple and adjust the paint mixing parameters, spraying speed and nozzle moving speed corresponding to each nozzle.

[0046] Preferably, using the nozzle number as the axis, the coupling degree and offset trend of the nozzles are extracted from the correlation matrix and the disturbance influence factor matrix respectively, and a disturbance state tensor T is constructed. Each dimension of the disturbance state tensor T corresponds to the interference direction between nozzles, the spraying displacement trend, and the physical position index. It is known that the nozzle number essentially serves as a unique identifier for the nozzle and can be mapped one-to-one with its physical position. For nozzle number i, its row in the correlation matrix represents the intensity of interference from all other nozzles; the higher the value, the greater the influence between nozzles. Simultaneously, the corresponding row in the disturbance influence factor matrix D represents the spraying offset trend of the nozzle in each direction. These two data sources jointly characterize the disturbance characteristics and spatial variation trend of nozzle i. To fuse data on these three-dimensional relationships, this invention constructs a perturbation state tensor T. This tensor can be understood as a three-dimensional matrix structure across multiple nozzles, with its dimensions corresponding to: the disturbance direction (e.g., the primary / secondary direction relationship between nozzles), the spraying offset trend direction (i.e., the physical perturbation vector, such as the X / Y / Z axes), and the nozzle physical number (used to locate nozzles in the physical array). By synthesizing the disturbance vector and offset vector of each nozzle i using tensors, the i-th layer of tensor T can be formed. This construction method not only preserves the spatial topology between nozzles but also effectively fuses their dynamic operating status. Compared to traditional two-dimensional matrix processing, this method utilizes tensor data structures to enhance feature representation capabilities, facilitating efficient indexing and retrieval in subsequent rule lookup and fine-tuning control.

[0047] As can be seen, this invention combines the two matrices mentioned above to extract the disturbance mapping relationship between the interference trend between nozzles and the spraying offset trend, thereby constructing a novel joint fine-tuning matrix. This matrix, on a nozzle-by-nozzle basis, links and adjusts three core control parameters for each nozzle: paint mixing parameters (e.g., the ratio of main and auxiliary paints, viscosity adjustment value), spraying speed (i.e., the speed at which the paint leaves the nozzle, affecting coating thickness), and nozzle movement speed (i.e., the linear velocity of the nozzle relative to the workpiece surface). This achieves dynamic adaptive adjustment for the coupling state of the nozzles. Each row of this matrix corresponds to one nozzle, and each column represents an adjustable parameter, ultimately forming a dynamic, time-varying linked control input. This ensures that the system maintains spraying uniformity and system stability even in multi-nozzle operating modes, solving the coupling imbalance problem present in traditional independent parameter control.

[0048] Preferably, a basic spraying parameter group P0 = {α0, β0, γ0} is set, which corresponds to the default paint mixing parameters, spraying speed, and nozzle movement speed, respectively. By bundling the three parameters into a basic parameter group, the coupling imbalance problem caused by independent adjustment of individual parameters can be avoided.

[0049] It should be noted that, to ensure a uniform and stable initial state of the control system during the startup of the spraying system or the initialization of a certain round of operations, this invention introduces a basic spraying parameter set setting strategy. This parameter set is typically set by process engineers based on the spraying type, paint properties, and target surface, providing a basic reference frame for the nozzle control module. The parameter set consists of three key indicators: paint mixing parameters, representing the mixing ratio or concentration of different paint components (such as the base agent and hardener); spray velocity, i.e., the linear velocity of the sprayed material ejected from the nozzle, directly affecting the coverage thickness and adhesion; and nozzle movement speed, i.e., the speed at which the nozzle moves along its trajectory, significantly impacting film uniformity and surface quality. In conventional systems, engineers may independently adjust a single parameter, such as adjusting the spray velocity to cope with airflow disturbances or increasing the nozzle movement speed alone to improve cycle time efficiency. However, in a spraying environment with dynamic coupling of multiple parameters, such single-parameter adjustment can disrupt the original control balance. For example, if the nozzle speed is increased but the spray velocity is not adjusted synchronously, it may lead to spray layer breakage or uneven thickness. Therefore, this invention proposes to bundle the three parameters into a whole for adjustment, forming a basic parameter set as a reference point for subsequent fine-tuning of the offset. During subsequent offset correction, the adjustment is no longer targeted at a single parameter, but rather the entire set of parameters is adjusted in a coordinated manner, ensuring overall system coupling balance and avoiding the risk of uncontrollable "one-size-fits-all" consequences. This design not only enhances the stability of the control system but also provides a unified reference for tensor mapping relationships, making parameter mapping more logically sound and feasible.

[0050] Furthermore, based on the interference direction and disturbance offset direction of each nozzle in the disturbance state tensor T, a nozzle parameter adjustment offset set ΔP = {Δα, Δβ, Δγ} is constructed. That is, by using a lookup table rule, the disturbance trend and interference trend exhibited by the nozzle in T are matched to a parameter offset mapping table, and ΔP is assigned to each nozzle as its dynamic adjustment amount based on P0. It should be noted that this invention determines the three-parameter adjustment amount through a pre-defined rule lookup table method, replacing the traditional control weight method, simplifying the logic and facilitating embedded implementation.

[0051] Traditional control systems typically use a "weighted coefficient multiplied by error" approach for fine-tuning, such as projecting the disturbance vector as an error term and then using a proportional factor for linear control. However, this method is often highly dependent on the proportional factor, making parameter tuning difficult, and it operates inefficiently in embedded controllers. This invention avoids the complexity of floating-point operations and weighted coefficient adjustments by using a lookup table. It directly and quickly obtains the adjustment values ​​of the three parameters through a discrete state-offset lookup table, greatly improving computational efficiency and system response speed, making it particularly suitable for resource-constrained industrial embedded control systems. The lookup table method also ensures that the adjustment range is limited to a safe operating range, avoiding over-adjustment or critical oscillations, and improving system robustness.

[0052] A better approach is to use the nozzle number as an index, combined with the basic parameter set P0 and the offset set ΔP, to generate a three-parameter joint fine-tuning matrix M. Each row represents the three-parameter adjustment result for the corresponding nozzle, that is, to represent the ternary parameter set of each nozzle as M. i =P 0i +ΔP i The final three-parameter joint fine-tuning matrix M is obtained and serves as the input to the linkage control module. Furthermore, this matrix can not only be refreshed periodically (e.g., before each task cycle) but also reconstructed in real time based on a disturbance trigger mechanism. That is, when the system detects a drastic change in the disturbance tensor T, it automatically invokes a lookup table mechanism to reset ΔP and reconstruct the joint fine-tuning matrix M. This structure enables the control system to possess high responsiveness, strong dynamics, and coupling stability, effectively supporting the coordinated operation of large-scale multi-nozzle systems.

[0053] S4: Based on the parameter correction amount output by the joint fine-tuning matrix, dynamically adjust the paint mixing parameters, spraying speed and moving speed corresponding to the current nozzle according to the nozzle number, so as to realize the synchronous control of the spraying thickness and its spatial uniformity.

[0054] Ideally, based on the nozzle number index and the joint fine-tuning matrix M, the parameter correction group corresponding to each nozzle is extracted, corresponding to the paint mixing parameters, spray speed, and nozzle movement speed setpoints for the current nozzle. It should be noted that using the joint fine-tuning matrix M directly as the scheduling reference source, binding the number index to the control parameters, avoids the traditional nozzle-parameter mapping table lookup process, improving control efficiency. Furthermore, during this operation, the nozzle number is used as a direct index pointer to quickly access the corresponding row in the joint fine-tuning matrix M, thereby obtaining the three correction values ​​for the current nozzle. This direct number retrieval method skips the traditional intermediate table lookup process, eliminating the need to first search for the nozzle ID and then map the control strategy, significantly reducing the response latency of the control system and facilitating parallel execution of control commands. In addition, this mechanism has good scalability; even if the number of nozzles is increased later, only the number of rows in the joint fine-tuning matrix M needs to be expanded, while the core control logic remains unchanged, avoiding the problem of redesigning the control path due to changes in system scale. The number index mechanism can also be implemented in the embedded controller as an array pointer or index address, making the system highly portable and real-time.

[0055] Preferably, based on the mixing ratio command of the paint mixing parameters in the parameter group, the opening combination of multiple component valves in the feeding unit is controlled to adjust the component mixing ratio of the paint in the current nozzle; based on the spray speed command of the spray speed in the parameter group, the internal pressure conveying system or pulse drive frequency of the nozzle is adjusted to output the material flow rate matching the mixing ratio; based on the movement speed command of the nozzle movement speed set value in the parameter group, the trajectory speed of the nozzle actuator is controlled, and the spray thickness and coverage uniformity are controlled according to the ratio relationship between the trajectory speed and the spray speed.

[0056] For example, the correction value for the spray velocity is provided by the joint fine-tuning matrix M, which is directly used as the input parameter for the pressure delivery mechanism or pulse driver within the nozzle. The required spray velocity varies significantly depending on the type of paint and the mixing ratio. If the spray velocity does not change synchronously after adjusting the mixing ratio, it may cause paint thinning, splashing, or sagging, affecting the coating quality. Spray velocity control can be achieved in two ways: if a pressure delivery system (such as a servo motor-driven pump) is used to deliver the paint, the pump speed can be adjusted according to instructions to achieve flow regulation; if a pulse-type nozzle (such as a piezoelectric nozzle) is used, the amount of paint pulsedly ejected can be adjusted by adjusting the drive frequency or pulse width. The system performs feedback adjustment and combines real-time sensor feedback (such as a flow meter or pressure gauge) with closed-loop correction to ensure that the output material flow rate of each nozzle matches its mixing ratio, achieving the dual purpose of material output consistency and coating thickness control.

[0057] Furthermore, a crucial speed ratio matching relationship exists between the nozzle's moving speed and its spraying speed. If the nozzle moves too slowly while the spraying speed is high, localized material accumulation will occur; conversely, moving too fast will lead to uneven coating and exposed substrate. This invention uses a three-parameter joint fine-tuning mechanism to simultaneously fine-tune the moving speed and spraying speed, ensuring that their ratio remains within the optimal range. The adjustment of the nozzle's moving speed is achieved by controlling an XYZ three-axis motion platform or a multi-joint robotic arm execution system. The speed adjustment command is given by the joint fine-tuning matrix M. After receiving this command, the control system uses it as the target speed adjustment factor for the trajectory planner, recalibrating the speed during trajectory calculation to synchronously adjust the speed along the predetermined path while maintaining trajectory stability. It is particularly noteworthy that traditional systems often decouple path planning from spraying speed adjustment, while this invention binds them together in the joint fine-tuning matrix M for joint adjustment, ensuring synchronous coupling of edge and motion on all nozzles, preventing issues such as ghosting, offset, and uneven thickness between sprayed patterns.

[0058] In summary, the synergistic application of the joint fine-tuning matrix M and the numbered indexing mechanism enables dynamic, three-parameter linkage control of multiple nozzles in the spraying system. This not only improves control accuracy and response speed but also significantly simplifies the control logic and enhances the system's robustness to external disturbances. By replacing the traditional mapping lookup method with a direct indexing mechanism, the controller can update nozzle parameters at a high frequency, adapting to rapidly changing interference environments. Furthermore, the integrated adjustment of the three parameters avoids coupling imbalances caused by single-parameter control, thus achieving significant results in actual coating thickness control and maintaining spatial uniformity.

[0059] S5: Based on the actual coating thickness difference and uniformity deviation obtained, update the nozzle disturbance influence matrix and send it back to the joint fine-tuning matrix to complete the next round of adjustment.

[0060] After completing one round of spraying, automatic detection will be activated to accurately collect data on coating thickness and uniformity of all sprayed areas.

[0061] Preferably, thickness measurement data is collected within the already sprayed area, and the thickness difference ΔH between the corresponding areas of each nozzle is extracted by comparing it with a preset target thickness template. i and local uniformity deviation ΔU i The thickness difference ΔH between each nozzle i With local uniformity deviation ΔU i Combined to construct the deviation vector ε i =ΔH i ,ΔU i The vector is input into the trajectory disturbance influence factor matrix to perform exponential correction on the nozzle-related terms in the disturbance matrix. In this embodiment, the exponential correction can be performed using an exponential decay coefficient, but this is not a unique limitation in this embodiment. The construction of this vector not only reflects the macroscopic thickness error of the nozzle coating result but also considers the local stability of the coating, constituting a multi-dimensional feedback input. Traditional spraying systems only use ΔH. i As a feedback reference value, the complex spatial characteristics of spray uniformity are ignored. This invention, however, incorporates ΔU... i These features enable the system to have greater spatial adaptability.

[0062] The updated disturbance influence matrix is ​​recombined with the original nozzle array correlation matrix for parameter tuning, dynamically generating a new round of three-parameter joint fine-tuning matrix. This three-parameter fine-tuning matrix is ​​then fed back to the nozzles for control and used as input for the next round of spraying, achieving adaptive iteration of the multi-parameter control loop. Unlike conventional fixed-value spraying control, this invention's system automatically generates a parameter matrix for each round of spraying based on feedback from the previous round, forming a closed-loop adaptive control system based on local spraying state—disturbance modeling—parameter correction—re-spraying feedback. This mechanism significantly reduces the system's dependence on human intervention and possesses dynamic adaptability to different materials, environments, and target template thicknesses.

[0063] In summary, the multi-parameter control method for vacuum insulation panel insulation layer spraying robot based on the embodiments of the present invention has been clarified. By dynamically sensing and controlling the interference relationship and trajectory deviation effects among multiple nozzles during the spraying process, it effectively solves the problems of uneven spraying thickness and poor spatial consistency. It enables fine-tuning of nozzle parameters, allowing each nozzle to maintain an ideal spraying state under different working conditions, thereby improving overall spraying accuracy and stability. By introducing a multi-parameter joint control mechanism, the present invention can significantly reduce spraying defects caused by factors such as posture deviation and coupling interference, reducing rework rate and material waste. Compared to traditional methods that struggle to cope with real-time disturbances and local error accumulation during the spraying process, the present invention possesses stronger adaptive capabilities and control robustness, making it particularly suitable for high-requirement applications such as vacuum insulation panel insulation layers. It achieves simultaneous optimization of spraying thickness and spatial uniformity, effectively improving the forming quality and production yield of vacuum insulation panel insulation layers.

[0064] Figure 3 This is a structural diagram of a multi-parameter control system for a vacuum insulation panel insulation layer spraying robot according to an embodiment of the present invention. Figure 3 As shown, the multi-parameter control system for vacuum insulation panel insulation layer spraying robot includes: an array modeling module, which is used to model the spatial layout of the robot nozzle array and generate a nozzle array correlation matrix to represent the coupling effect between adjacent nozzles based on historical spraying data.

[0065] The trajectory acquisition module is used to acquire the nozzle trajectory in real time during the spraying process, calculate the offset vector based on the pose error, and construct the trajectory disturbance influence factor matrix.

[0066] The fine-tuning matrix module is used to construct a three-parameter joint fine-tuning matrix based on the correlation matrix and the perturbation matrix;

[0067] The synchronous control module is used to dynamically adjust the current nozzle according to the nozzle number based on the parameter correction amount output by the joint fine-tuning matrix, so as to achieve synchronous control of the coating thickness and spatial uniformity.

[0068] The error feedback module is used to update the nozzle disturbance influence matrix based on the actual coating thickness difference and uniformity deviation, and then feed it back to the joint fine-tuning matrix to complete the next round of adjustment.

[0069] Here, those skilled in the art will understand that the specific operations of each step in the above-described method for multi-parameter control of vacuum insulation panel insulation layer spraying robot have been described in detail in the above description of the method for multi-parameter control of vacuum insulation panel insulation layer spraying robot with reference to Figures 1-2, and therefore, the repeated description will be omitted.

[0070] In summary, the multi-parameter control method for vacuum insulation panel insulation layer spraying robot based on the embodiments of the present invention has been clarified. It determines whether there are signs of infection at the patient's puncture site by performing time-series analysis on high-definition images of the puncture site at different time points within the target monitoring period. This improves the accuracy and timeliness of infection monitoring, helps to detect signs of infection early, and thus allows for timely measures to reduce the incidence and severity of infection.

Claims

1. A multi-parameter control method for a robot spraying insulation layer on a vacuum insulation panel, characterized in that, include: The spatial layout of the robot nozzle array is modeled, and a nozzle array correlation matrix is ​​generated based on historical spraying data to represent the coupling effect between adjacent nozzles. The nozzle trajectory is collected in real time during the spraying process, the offset vector is calculated based on the pose error, and the trajectory disturbance influence factor matrix is ​​constructed. Based on the correlation matrix and the perturbation matrix, a three-parameter joint fine-tuning matrix is ​​constructed. Based on the parameter correction amount output by the joint fine-tuning matrix, the current nozzle is dynamically adjusted according to the nozzle number to achieve synchronous control of coating thickness and spatial uniformity. Based on the actual coating thickness difference and uniformity deviation obtained, the nozzle disturbance influence matrix is ​​updated and fed back to the joint fine-tuning matrix to complete the next round of adjustment; The spatial layout modeling of the robot nozzle array includes: forming a set of spatial arrangement parameters by calibrating the relative position coordinates, pitch angle and array spacing of the robot nozzle array under a standard spraying task; The construction of the nozzle array correlation matrix includes: statistical analysis of the overlapping segments of the trajectories of adjacent nozzles in the historical dataset, combined with the coating thickness deviation value, determining the interference weight of each nozzle on its neighboring nozzles, and extracting it as a coupling weight vector; taking the spatial arrangement parameter group and the coupling weight vector as input, normalizing and fusing the spatial distance of the pairwise combination of nozzle numbers with the historical interference intensity, and generating a nozzle array correlation matrix to represent the coupling relationship. The three parameters include paint mixing parameters, spraying speed, and nozzle movement speed.

2. The multi-parameter control method for vacuum insulation panel insulation layer spraying robot according to claim 1, characterized in that, The calculation of the offset vector based on pose error includes: Based on the nozzle current pose parameter set Q recorded by the spray controller, combined with the target pose parameter set set by the task. Calculate the pose error term ΔQ of the nozzle in six-degree-of-freedom space; By projecting the linear offset portion of the pose error term ΔQ onto the spraying axis direction, the actual spraying axis offset vector V of each nozzle is extracted as a physical quantity characterizing the actual deviation trend of the spray beam.

3. The multi-parameter control method for vacuum insulation panel insulation layer spraying robot according to claim 2, characterized in that, The construction of the joint fine-tuning matrix includes: extracting the nozzle coupling degree and offset trend from the correlation matrix and the perturbation influence factor matrix, with the nozzle number as the axis, and constructing a perturbation state tensor T; each dimension of the perturbation state tensor T corresponds to the interference direction between nozzles, the spraying displacement trend, and the physical position index; and setting the basic spraying parameter set. These correspond to the default paint mixing parameters, spray velocity, and nozzle movement speed, respectively. Based on the interference direction and disturbance offset direction of each nozzle in the disturbance state tensor T, a nozzle parameter adjustment offset group is constructed. Using the nozzle number as an index, combined with the basic parameter group With offset group Generate a three-parameter joint fine-tuning matrix M, where each row represents the adjustment result of the three parameters corresponding to the nozzle.

4. The multi-parameter control method for vacuum insulation panel insulation layer spraying robot according to claim 1, characterized in that, The dynamic adjustment of the current nozzle includes: extracting the parameter correction group corresponding to each nozzle based on the nozzle number index and the fine-tuning matrix M, which respectively correspond to the paint mixing parameters, spraying speed and nozzle moving speed settings of the current nozzle; and adjusting the paint mixing parameters, spraying speed and nozzle moving speed according to the parameter group instructions.

5. The multi-parameter control method for vacuum insulation panel insulation layer spraying robot according to claim 1, characterized in that, The actual coating thickness difference and uniformity deviation include: collecting thickness measurement data in the completed spraying area, and extracting the thickness difference and local uniformity deviation of the corresponding area of ​​each nozzle by comparing it with the preset target thickness template; and combining the thickness difference and local uniformity deviation of each nozzle to construct a deviation vector.

6. The multi-parameter control method for vacuum insulation panel insulation layer spraying robot according to claim 5, characterized in that, The updated nozzle disturbance influence matrix includes: inputting the deviation vector into the trajectory disturbance influence factor matrix, and exponentially correcting the nozzle corresponding item in the disturbance matrix.

7. A multi-parameter control system for a vacuum insulation panel insulation layer spraying robot, comprising using the method described in any one of claims 1-6 to perform multi-parameter control of the vacuum insulation panel insulation layer spraying robot, characterized in that, include: The array modeling module is used to model the spatial layout of the robot nozzle array and generate a nozzle array correlation matrix to represent the coupling effect between adjacent nozzles based on historical spraying data. The trajectory acquisition module is used to acquire the nozzle trajectory in real time during the spraying process, calculate the offset vector based on the pose error, and construct the trajectory disturbance influence factor matrix. The fine-tuning matrix module is used to construct a three-parameter joint fine-tuning matrix based on the correlation matrix and the perturbation matrix; The synchronous control module is used to dynamically adjust the current nozzle according to the nozzle number based on the parameter correction amount output by the joint fine-tuning matrix, so as to achieve synchronous control of the coating thickness and spatial uniformity. The error feedback module is used to update the nozzle disturbance influence matrix based on the actual coating thickness difference and uniformity deviation, and then feed it back to the joint fine-tuning matrix to complete the next round of adjustment.

Images