An intelligent door and window adaptive installation method with installation error self-compensation function

Abstract

The application discloses a kind of intelligent door and window adaptive installation methods with installation error self-compensation function, belong to intelligent construction and automation installation technical field;The method comprises: starting multi-dimensional sensing array to collect the space parameter between door and window frame and wall hole;Reconstruct the actual topography of hole and dynamically generate optimal installation target model, obtain actual pose matrix by coordinate transformation;Actual pose is compared with target pose, generate six-degree-of-freedom compound error and decouple into translation component and rotation component;Establish deformation prediction model, combined with error component and elastic pre-compensation amount to execute hierarchical compensation decision, and obtain compensation instruction containing pose adjustment component and internal force decoupling component by calculation;Drive four-corner movable adjusting mechanism to move synchronously, through real-time retest and cyclic comparison until parameter reaches the standard;The application realizes self-sensing, self-decoupling and self-compensation of installation error, improves installation precision and efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction and automated installation technology, specifically to an intelligent door and window self-adaptive installation method with self-compensation function for installation errors. Background Technology

[0002] Door and window installation is a key process in construction engineering, and its installation accuracy directly affects the sealing performance, smoothness of opening and closing, and service life of doors and windows.

[0003] Currently, the industry generally adopts manual measurement and adjustment methods. Workers use tools such as spirit levels and tape measures to measure gaps, levelness, and diagonal deviations, judge the compensation amount based on experience, and manually adjust hinges or connectors. However, conventional processes still have obvious limitations. The detection and compensation methods are relatively crude and cannot obtain the global spatial error distribution. Different operators can have judgment deviations of several millimeters. After adjustment, the compensation effect cannot be automatically verified, and repeated trial and error are required, resulting in inconsistent installation quality and low efficiency.

[0004] Meanwhile, existing error compensation methods treat door and window frames as ideal rigid bodies, failing to consider the elastic deformation coupling effect of large-sized frames in multi-point adjustment. The calculated compensation commands have unpredictable residual errors compared to the actual pose changes. Furthermore, the preset ideal rectangular installation target does not match the actual irregular shape of the wall opening. Forcing a close approach may result in local over-tightness or over-looseness. Moreover, there is an internal force coupling problem when the four corner adjustment mechanisms move synchronously. Existing decoupling control strategies are difficult to achieve true six-degree-of-freedom coordinated adjustment.

[0005] Therefore, there is an urgent need in this field for an intelligent door and window adaptive installation method with self-compensation function for installation errors, so as to realize elastic deformation compensation, opening shape adaptation and internal force decoupling control, and break through the existing technical bottlenecks. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent door and window adaptive installation method with self-compensation function for installation errors, so as to solve the technical problems in the existing methods, such as the difficulty in obtaining global errors by manual measurement, ignoring the elastic deformation coupling effect of the frame, the mismatch between the preset ideal rectangular target and the actual shape of the opening, and the difficulty in coordinated adjustment due to the internal force coupling of the four corner adjustment mechanism.

[0007] The objective of this invention can be achieved through the following technical solutions: An adaptive installation method for smart doors and windows with self-compensation function for installation errors includes the following steps: The controller activates the multi-dimensional sensor array to collect the original spatial parameter set between the door / window frame and the wall opening. The original spatial parameter set includes the gap value, horizontal tilt angle value, vertical tilt angle value, and diagonal length value. The controller receives the original set of spatial parameters, reconstructs the actual shape of the opening based on the collected gap values, and generates the optimal installation target model; at the same time, it calculates the current actual pose matrix of the door and window frame; and establishes an error model containing translation and rotation components based on the deviation between the actual pose matrix and the optimal installation target model. The controller compares the actual pose matrix with the theoretical pose matrix of the optimal installation target model to generate a six-degree-of-freedom composite installation error vector. Based on the orthogonal decoupling characteristics of the error model, the six-degree-of-freedom composite installation error vector is decoupled into translation error components and rotation error components. The controller calculates the elastic additional deformation as a pre-compensation amount based on the translation error component and rotation error component, combined with the stiffness characteristics and deformation prediction model of the door and window frame, and solves the compensation command including the pose adjustment component and the internal force decoupling component. The controller drives the movable adjustment mechanisms at the four corners of the door and window frame to move synchronously according to the compensation command. At the same time, it uses a multi-dimensional sensor array to remeasure the current spatial parameters in real time, and compares the remeasured parameters with the various indicators of the optimal installation target model in a loop until all parameters meet the standards, thus completing the adaptive installation.

[0008] Preferably, the process of generating the optimal installation target model specifically includes: Based on the collected gap values ​​from multiple discrete points, a discrete gap distribution map of the hole edge is constructed. Interpolation fitting is performed on the discrete gap distribution map to reconstruct the continuous actual shape contour of the wall opening; feature parameters are extracted from the continuous actual shape contour, including the local concavity and convexity of each side, the overall tilt angle, and the corner offset. Calculate the standard deviation of the gap and the gap deviation at each corner point based on the multiple discrete point gap values; With the goal of minimizing the mean square error of the gap and the corner gap deviation, and constrained by the horizontal tolerance, vertical tolerance and diagonal length difference tolerance in the preset installation quality specifications, the optimal installation target model is dynamically generated based on the aforementioned characteristic parameters. The theoretical pose matrix corresponding to the optimal installation target model and the error model use the same coordinate system reference.

[0009] Preferably, the process of calculating the current actual pose matrix of the door and window frame specifically includes: Based on the collected gap values, the rectangular outline is fitted using the least squares method to determine the coordinates of the four corner points of the fitted rectangle of the wall opening, and then the actual geometric center of the fitted rectangle is calculated. A three-dimensional spatial coordinate system is established with the actual geometric center as the origin, which serves as the reference coordinate system for error analysis; The collected horizontal tilt angle, vertical tilt angle, and diagonal length values ​​are transformed into the coordinate system to obtain the six degrees of freedom parameters of the door and window frame in the coordinate system. The six degrees of freedom parameters are used to construct the actual pose matrix of the door and window frame.

[0010] Furthermore, the process of transforming to the coordinate system via coordinate transformation specifically includes: Using the geometric center of the fitted rectangle as a reference point, calculate the rotation matrix and translation vector corresponding to the deviation between the current actual pose and the theoretical pose of the door and window frame; The rotation matrix and translation vector are solved by singular value decomposition to obtain the optimal rigid body transformation parameters of the door and window frame from the current pose to the theoretical pose. Based on the optimal rigid body transformation parameters, the six degrees of freedom parameters of the door and window frame are transformed into the coordinate system.

[0011] Preferably, the process of obtaining the elastic additional deformation as the pre-compensation amount specifically includes: Based on the translational and rotational error components, the required basic adjustment displacement for each adjustment point is estimated using an inverse kinematics model. Based on the stiffness matrix parameters of the door and window frames, the basic adjustment displacement of each adjustment point is converted into the corresponding estimated adjustment force. Substitute the estimated adjustment force into the deformation prediction model to output the additional elastic deformation of the frame generated at each adjustment point under the action of the estimated adjustment force. The negative value of the elastic additional deformation is used as the pre-compensation amount.

[0012] Preferably, the process of calculating the compensation command, which includes pose adjustment components and internal force decoupling components, specifically includes: Calculate the magnitude ratio of the translation error component to the rotation error component, and based on the comparison result of the ratio with a preset threshold range, determine whether the current error type is translation-dominant, rotation-dominant, or a hybrid type. The compensation order for each error component is determined according to a preset priority rule; For single error types dominated by translation or rotation, a direct reverse compensation strategy is adopted; for mixed error types, a collaborative compensation strategy is adopted. The translation error component, rotation error component, and pre-compensation amount are weighted and superimposed to obtain the comprehensive compensation amount, and the weight coefficients are dynamically adjusted according to the error type. The compensation commands for each movable adjustment mechanism are obtained by solving the inverse kinematics model.

[0013] Furthermore, the preset priority rule is as follows: First priority: Diagonal length deviation caused by rotational error component. When the diagonal length deviation exceeds ±2mm, the rotational error component should be compensated first. Second priority: horizontal and vertical deviations caused by translation error components. When the horizontal or vertical deviation exceeds ±1mm / m, the translation error components should be compensated first. Third priority: gap uniformity deviation. When the gap uniformity deviation exceeds ±1mm, compensate for the remaining translation error component and rotation error component.

[0014] Preferably, the process of cyclically comparing the retest parameters with the various indicators of the optimal installation target model specifically includes: The current spatial parameters are reacquired using a multi-dimensional sensor array; The gap values, horizontal tilt angle values, vertical tilt angle values, and diagonal length values ​​obtained from the remeasurement are compared with the target gap distribution, target levelness, target verticality, and target diagonal length difference in the optimal installation target model, and the deviations are calculated. If all deviations fall within the preset error threshold range, the installation is considered complete; otherwise, return to perform error decoupling and subsequent steps, and proceed to the next iteration.

[0015] The beneficial effects of this invention are as follows: This invention collects global spatial parameters between door and window frames and wall openings through a multi-dimensional sensor array, and calculates the actual pose matrix of six degrees of freedom by combining coordinate transformation. It breaks through the limitation of traditional manual measurement that cannot obtain the global error distribution, realizes the digital and holographic perception of installation errors, and provides a data foundation for subsequent accurate compensation. The actual shape of the opening is reconstructed based on the collected gap values, and the optimal installation target model is dynamically generated in conjunction with the preset installation quality specifications. This replaces the traditional fixed ideal rectangular target, solves the problem of local over-tightness or over-looseness caused by the mismatch between the preset target and the irregular shape of the opening, and realizes the adaptive adjustment of the installation target to the on-site working conditions. By obtaining the stiffness matrix of the door and window frame and establishing a deformation prediction model, the elastic additional deformation is calculated and its negative value is used as the pre-compensation amount. This overcomes the shortcomings of the traditional method that treats the door and window frame as an ideal rigid body and ignores the elastic deformation coupling effect, eliminates the unpredictable residual error caused by the deformation of the frame, and improves the accuracy and reliability of the pose adjustment. The compensation command, which includes pose adjustment component and internal force decoupling component, is obtained by hierarchical compensation decision algorithm. The internal force decoupling component is obtained by inverse calculation of deformation prediction model. This solves the problem that internal force coupling makes it difficult to achieve six degrees of freedom coordinated adjustment when the four corner adjustment mechanism moves synchronously, and realizes the decoupling and coordinated control of adjustment force. By retesting spatial parameters in real time and comparing them cyclically with the optimal installation target model until all parameters meet the standards, the drawbacks of traditional methods, such as the inability to automatically verify the compensation effect after adjustment and the need for repeated trial and error, are overcome, thereby improving installation efficiency and quality consistency. Attached Figure Description

[0016] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.

[0017] Figure 1 This is a schematic diagram of the overall process of the intelligent door and window adaptive installation method with self-compensation function for installation errors as described in this invention; Figure 2 This is a schematic diagram of the orthogonal decoupling of the error model described in this invention; Figure 3 This is a schematic diagram illustrating the execution logic of the hierarchical compensation decision algorithm described in this invention; Figure 4 This is a schematic diagram of the forward and inverse solutions of the deformation prediction model described in this invention; Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0019] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0020] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.

[0021] Example 1 Please see Figures 1-4 This embodiment provides an adaptive installation method for smart doors and windows with self-compensation function for installation errors, including the following steps: S1. The controller starts the multi-dimensional sensor array to collect the original spatial parameter set between the door and window frame and the wall opening. The original spatial parameter set includes the gap value, horizontal tilt angle value, vertical tilt angle value, and diagonal length value. First, a multi-dimensional sensor array is pre-deployed at the four corners and the midpoints of each side of the door and window frame; the multi-dimensional sensor array includes a laser displacement sensor, an electronic tilt sensor and a laser rangefinder, which are used to collect the gap value between the door and window frame and the wall opening, the horizontal tilt value, the vertical tilt value and the diagonal length value, respectively. After the controller is powered on, it sends a start command to the multi-dimensional sensor array, triggering all sensors to synchronously acquire data. Specifically: along the four edges of the door and window frame, at multiple preset discrete measurement points, the gap between the outer edge of the door and window frame and the inner wall of the wall opening is measured by a laser displacement sensor, and the gap data and corresponding spatial coordinates of each measurement point are recorded; the horizontal tilt angle of the door and window frame relative to the horizontal reference plane and the vertical tilt angle relative to the vertical reference plane are acquired by an electronic tilt sensor; the actual length of the two diagonals of the door and window frame is measured by a laser rangefinder; the controller receives all the above acquired data to form an original spatial parameter set, which includes: the gap value, horizontal tilt angle value, vertical tilt angle value, length of the two diagonals, and spatial coordinate information corresponding to each measurement point; the original spatial parameter set serves as the basic input data for subsequent pose calculation, shape reconstruction, and error analysis, and is transmitted to the controller's built-in data processing unit for caching and preprocessing; This step involves using a multi-dimensional sensor array to collect the gap, horizontal tilt angle, vertical tilt angle, and diagonal length between the door / window frame and the wall opening. The gap is measured at multiple points around the door / window frame, and the spatial coordinate information of each measurement point is recorded. This completes the spatial parameter perception and digital acquisition at the door / window installation site, laying the data foundation for subsequent actual pose matrix calculation, actual opening shape reconstruction, and error model establishment.

[0022] S2. The controller receives the original set of spatial parameters, reconstructs the actual shape of the opening based on the collected gap values, and generates the optimal installation target model; at the same time, it calculates the current actual pose matrix of the door and window frame; based on the deviation between the actual pose matrix and the optimal installation target model, it establishes an error model containing translation and rotation components. The process of reconstructing the actual shape of the opening based on the collected gap values ​​and generating the optimal installation target model is as follows: First, the controller constructs a discrete gap distribution map of the opening edge based on the collected discrete gap values ​​from multiple discrete points. Specifically, N measurement points are evenly spaced along the upper edge of the door and window frame from left to right, and the horizontal coordinates of each measurement point are recorded. and their corresponding gap values ,by x-axis Using the ordinate as the vertical axis, plot the discrete point sequence of the upper edge in a Cartesian coordinate system. Similarly, N measurement points are placed at equal intervals from left to right along the lower edge, and a discrete point sequence is plotted with the horizontal position as the x-coordinate and the gap value as the y-coordinate. M measurement points are placed at equal intervals from top to bottom along the left edge, and a discrete point sequence is plotted with the vertical position as the y-coordinate and the gap value as the x-coordinate. M measurement points are placed at equal intervals from top to bottom along the right edge, and a discrete point sequence is plotted with the vertical position as the y-coordinate and the gap value as the x-coordinate. Secondly, the controller performs interpolation fitting on the discrete gap distribution map to reconstruct the continuous actual shape contour of the wall opening; for the discrete point sequence at the upper edge, cubic spline interpolation is used in each sub-interval. Construct a cubic polynomial above: ; satisfy And the first derivative and second derivative The curves are continuous at internal nodes, and the coefficients are obtained by solving a tridiagonal system of equations. Similarly, cubic spline interpolation is performed on the lower, left, and right edges to obtain four continuous contour curves. Based on the continuous contours, the feature parameters of each side are extracted: the local concavity / convexity is the peak-valley difference of the contour curve, i.e. The overall tilt angle is the angle corresponding to the slope of the line connecting the two endpoints of the contour curve, i.e. The corner offset is the difference between the measured gap values ​​at the four corner points and the corresponding corner gap values ​​of the fitted rectangle. Next, the controller calculates the standard deviation of the gap and the gap deviation at each corner point based on the gap values ​​at multiple discrete points; the formula for calculating the standard deviation of the gap is: ; in This represents the average gap value. It reflects the overall uniformity of the gap distribution; the gap deviation at each corner point is the difference between the measured gap value and the target gap value at the four corner points: These correspond to the four corner points: top left, top right, bottom left, and bottom right, respectively. Finally, the controller aims to minimize the standard deviation of the clearance and the corner clearance deviation, while being constrained by the horizontal tolerance, vertical tolerance, and diagonal length difference tolerance in the preset installation quality specifications. Based on the extracted feature parameters, it dynamically generates the optimal installation target model. The first step is to define the decision variables; let the overall translation adjustment of the upper edge of the door and window frame be... The overall translation adjustment amount of the lower edge is The overall translation adjustment amount of the left edge is The overall translation adjustment amount of the right edge is The unit is millimeters; the overall translation adjustment amount represents the distance that the entire door and window frame is translated towards the wall opening. A positive value indicates a reduction in the gap, and a negative value indicates an increase in the gap. The second step is to establish the target gap distribution function; based on the continuous actual shape contour of the opening obtained from step S2, the target gap distribution function of the upper edge is defined as follows: ,in The upper edge of the opening is in a horizontal position. The measured gap value at the location; similarly, the lower edge left edge right edge Therefore, the target gap distribution consists of four decision variables. , , , Uniquely certain; The third step is to construct the objective function. The objective function consists of two weighted terms: the first term is the root mean square error of the gap distribution, reflecting the uniformity of the gap distribution; the second term is the sum of the target gap deviations at the four corner points, reflecting the corner alignment accuracy. The expression for the objective function is: ;in, For the first The standard deviation of the target gap values ​​at each measurement point on the edge of the strip. For the first The deviation between the target clearance value and the ideal clearance value at each corner point and Let be the weighting coefficient, satisfying , usually take , ; The fourth step is to set constraints; convert the tolerance requirements in the preset installation quality specifications into constraints on the target gap distribution: Horizontalness constraint: The difference between the maximum and minimum values ​​of the upper edge target gap shall not exceed [a certain value]. ,in The width of the door and window frame, i.e. The same applies to the bottom edge; Verticality constraint: The difference between the maximum and minimum values ​​of the target gap on the left edge shall not exceed [a certain value]. ,in The height of the door and window frames, i.e. The same applies to the right edge; Diagonal length difference constraint: The theoretical lengths of the two diagonals of the door and window frame are calculated from the target gap values ​​at the four corner points, and the absolute value of the difference between the two lengths shall not exceed 2mm. The fifth step is to solve the optimization problem; the optimization problem described above has four decision variables. The objective function is continuously differentiable, and the constraints are linear inequalities. A sequential quadratic programming method is used to solve the problem, yielding the optimal decision variable values ​​that minimize the objective function, and thus the optimal objective gap distribution. , , , Together with the corresponding target horizontality, target verticality, and target diagonal length difference, they constitute the optimal installation target model; The theoretical pose matrix corresponding to the optimal installation target model is expressed with the geometric center of the fitted rectangle as the reference, and uses the same coordinate system reference as the error model. The specific process for calculating the current actual pose matrix of the door and window frames is as follows: The controller first measures the gap at multiple points around the door and window frame based on the collected gap values. It then fits a rectangular outline using the least squares method to determine the coordinates of the four corner points of the fitted rectangle for the wall opening, and finally calculates the actual geometric center of the fitted rectangle. Specifically, let the set of measurement points on the upper edge be... ,in The horizontal position coordinates are For the corresponding gap values; the least squares method is used to fit the straight line. The slope is solved by minimizing the sum of squared residuals. and intercept That is, minimizing the objective function: ;right and Taking the partial derivatives of each equation and setting them to zero, we obtain the normal system of equations: Solving for the parameters of the fitted line at the upper edge yields the results. , Similarly, the fitted line at the lower edge can be obtained. Left edge fitted line Right edge fitted line Intersect the two horizontal fitted lines (top and bottom) with the two vertical fitted lines (left and right) to obtain the coordinates of the four corner points; taking the top left corner point as an example, its coordinates... satisfy: The solution is: Similarly, calculate the top right corner point. , lower left corner point bottom right corner The actual geometric center of the fitted rectangle Calculate the arithmetic mean of the coordinates of the four corner points: ; Then, the controller establishes a three-dimensional spatial coordinate system with the actual geometric center of the fitted rectangle as the origin, serving as the reference coordinate system for error analysis; specifically, with... The origin is the coordinate system, and the positive X-axis is defined as the horizontal direction to the right (unit direction vector). The positive Y-axis direction is the vertically upward direction (unit direction vector). The positive Z-axis direction is defined by the right-hand rule, with the direction perpendicular to the plane of the door and window frame pointing outwards. Establish a right-handed rectangular coordinate system The coordinates of any point P in space in this coordinate system Satisfy the transformation relationship: Where R is the initial rotation matrix; this coordinate system serves as the unified benchmark for all subsequent pose calculations, error analyses, and compensation decisions. Next, the collected horizontal tilt angle, vertical tilt angle, and diagonal length values ​​are transformed into the coordinate system to obtain the six degrees of freedom parameters of the door and window frame in the coordinate system. First, taking the geometric center of the fitted rectangle as the reference point, calculate the rotation matrix and translation vector corresponding to the deviation between the current actual pose and the theoretical pose of the door and window frame. Specifically, let the current actual pose matrix of the door and window frame be... The preset theoretical reference pose matrix is Both are 4×4 homogeneous transformation matrices: in and It is a 3×3 rotation matrix. and It is a 3×1 translation vector; The controller first uses the four corner points of the door and window frames as feature points, and then uses the actual pose matrix... Extract the coordinates of the four corner points in the actual pose. From the theoretical pose matrix Extract the coordinates of the four corner points in the theoretical pose. The initial point set consists of four sets of feature point pairs. Then, the controller calculates the deviation transformation matrix from the theoretical pose to the current actual pose. : ;Will Expand ;therefore: ' Wherein, the deviation rotation matrix Deviation translation vector The deviation transformation matrix This describes the deviation relationship from the theoretical pose to the current actual pose; in order to obtain the transformation parameters from the current actual pose to the theoretical pose, the controller needs to solve the rotation matrix. Translation vector , so that: The controller uses the above deviation relationship as an initial estimate and substitutes it into the next step to solve for the optimal solution; The singular value decomposition method is used to solve for the rotation matrix and translation vector, thereby obtaining the optimal rigid body transformation parameters for the door and window frame to transform from the current pose to the theoretical pose. Specifically, the controller uses singular value decomposition to solve for the optimal rotation matrix. Translation vector This minimizes the following objective function: The specific calculation process is as follows: The first step is to calculate the centroids of the two point sets: ; The second step is to decentralize the point set: The third step is to calculate the covariance matrix. : ; Fourth step, for Perform singular value decomposition: ; in and It is a unitary matrix. It is a diagonal matrix; Step 5: Calculate the rotation matrix : ;like Then it is corrected to: ; Step 6: Calculate the translation vector : ; Thus, the optimal rotation matrix is ​​obtained. Translation vector , which are the optimal rigid body transformation parameters for the door and window frame to transform from the current actual pose to the theoretical pose; Based on the optimal rigid body transformation parameters, the six degrees of freedom parameters of the door and window frame are transformed into the coordinate system. Specifically, the controller transforms the six degrees of freedom parameters of the door and window frame into the coordinate system based on the aforementioned optimal rigid body transformation parameters. The six degrees of freedom parameters include: translation along the X-axis. Translation along the Y-axis Translation along the Z-axis Rotation angle about the X-axis (Pitch angle), rotation angle about the Y-axis (Roll angle), rotation angle about the Z-axis (Yaw angle); where the translation amount is directly derived from the translation vector. Given: The rotation angle is determined by the rotation matrix. The inverse solution yields the result, given the rotation matrix. The elements are The three rotation angles are calculated using the following formula: , , ,in To consider the arctangent function in the quadrant, the six degrees of freedom parameters of the door and window frame in this coordinate system are obtained. ; Finally, the controller constructs the actual pose matrix of the door and window frame from the six degrees of freedom parameters. Specifically, the controller first calculates the rotation matrix based on the three rotation angles. The elements of the rotation matrix are composed of the sine and cosine function values ​​of the rotation angles around the X-axis, Y-axis, and Z-axis. Then, the translation vector is combined with the rotation matrix to form a homogeneous transformation matrix of four rows and four columns. The rotation matrix is ​​placed in the top left three rows and three columns, the translation vector is placed in the top right three rows and one column, and the fourth row is fixed as zero, zero, zero, one. This four-row and four-column matrix is ​​the actual pose matrix of the door and window frame in the current coordinate system. Based on the deviation between the actual pose matrix and the optimal installation target model, an error model including translation and rotation components is established. The specific process is as follows: The controller calculates the deviation transformation matrix between the actual pose matrix and the theoretical pose matrix corresponding to the optimal installation target model. This deviation transformation matrix describes the deviation relationship between the door and window frame from the theoretical pose to the current actual pose. Then, the translation and rotation components are separated from this deviation transformation matrix to obtain three translation error components and three rotation error components, which together constitute a six-degree-of-freedom error vector. This error model has orthogonal decoupling characteristics, and the error components of the six degrees of freedom can be expressed and processed independently, providing a theoretical basis for subsequent error decoupling. This step reconstructs the continuous shape of the opening based on discrete gap values, aiming to minimize the gap mean square error and corner deviation. Constrained by installation quality specifications, it dynamically generates the optimal installation target model. Then, it obtains the actual geometric center of the opening by fitting the rectangular contour using the least squares method and establishes a three-dimensional coordinate system. It uses the singular value decomposition method to solve for the optimal rigid body transformation parameters and converts the six degrees of freedom parameters of the door and window frame into the actual pose matrix.

[0023] S3. The controller compares the actual pose matrix with the theoretical pose matrix of the optimal installation target model, generates a six-degree-of-freedom composite installation error vector, and decouples the six-degree-of-freedom composite installation error vector into translation error components and rotation error components based on the orthogonal decoupling characteristics of the error model. First, the actual pose matrix is ​​compared with the theoretical pose matrix of the optimal installation target model to generate a six-degree-of-freedom composite installation error vector, specifically including: The controller obtains the current actual pose matrix of the door and window frame calculated in step S2, and the theoretical pose matrix corresponding to the optimal installation target model dynamically generated in step S2. Both are homogeneous transformation matrices with four rows and four columns. Each matrix consists of a three-row, three-column rotation matrix in the upper left corner and a three-row, one-column translation vector in the upper right corner. The controller multiplies the actual pose matrix by the inverse of the theoretical pose matrix to calculate the error transformation matrix from the theoretical pose to the current actual pose. The error transformation matrix is ​​also composed of two parts: the error rotation matrix and the error translation vector, which describes the deviation relationship between the theoretical pose and the actual pose. The controller extracts a six-degree-of-freedom composite installation error vector from the error transformation matrix, specifically including: directly extracting three translation error components from the error translation vector, corresponding to the position deviations along the X-axis, Y-axis, and Z-axis, respectively; and extracting three rotation error components from the error rotation matrix through inverse solution of the arctangent function, corresponding to the angular deviations around the X-axis, Y-axis, and Z-axis, respectively; the three translation error components and the three rotation error components together constitute the six-degree-of-freedom composite installation error vector. Then, relying on the orthogonal decoupling characteristics of the error model, the six-degree-of-freedom composite installation error vector is decoupled into translational error components and rotational error components, specifically including: Based on the orthogonal decoupling characteristics of the error model established in step S2, the controller decouples the six-degree-of-freedom composite installation error vector into independent translational and rotational error components. The translational error component includes three positional deviations along the X, Y, and Z axes, representing the positional deviations of the door / window frame in three-dimensional space. The rotational error component includes three angular deviations around the X, Y, and Z axes, representing the attitude deviations of the door / window frame in three-dimensional space. The specific decoupling process is as follows: The translational error component is extracted from the error transformation matrix. From the error rotation matrix The rotation error components are solved by inversely using the arctangent function: , , This yields the complete expression for the six-degree-of-freedom composite installation error vector. ; Because the error model has orthogonal decoupling characteristics, the error components in the six degrees of freedom can be expressed and processed independently. The translation error component and the rotation error component are not coupled to each other, so they can be used as independent inputs for subsequent compensation decisions. This step generates a six-degree-of-freedom composite installation error vector by comparing the actual pose matrix with the theoretical pose matrix of the optimal installation target model. Then, it decouples the error vector into translational and rotational error components using the orthogonal decoupling characteristics of the error model, providing clear error input for subsequent deformation prediction and hierarchical compensation decision-making.

[0024] S4. The controller calculates the elastic additional deformation as the pre-compensation amount based on the translation error component and rotation error component, combined with the stiffness characteristics and deformation prediction model of the door and window frame, and solves the compensation command including the pose adjustment component and the internal force decoupling component. The calculated additional elastic deformation, used as a pre-compensation amount, specifically includes: First, based on the translational and rotational error components, the required basic adjustment displacement for each adjustment point is estimated using an inverse kinematics model: Suppose that the four adjustment points are located near the four corner points of the door and window frame, and their spatial coordinates are respectively... When door and window frames need to generate translational error components and rotational error components At that time, the first The basic adjustment displacement required for each adjustment point Calculate using the following formula: ;in This represents the vector cross product operation; from this, the basic adjustment displacement vectors of the four adjustment points are obtained. ; The controller converts the basic adjustment displacement of each adjustment point into the corresponding estimated adjustment force based on the stiffness matrix parameters of the door and window frame. The specific method for obtaining the stiffness matrix parameters is as follows: A finite element model is established based on the material and geometric dimensions of the specific door and window frame; a unit force is applied at each of the four corner points of the door and window frame, perpendicular to the plane of the door and window frame; the displacement response of each corner point under this force is calculated through finite element static analysis; the applied unit force is then... At the adjustment point, the first The displacement response generated by each adjustment point is denoted as . Then the stiffness matrix It is a 4×4 matrix, and its elements It describes the linear relationship between adjustment force and displacement; for standard door and window frames, the pre-calculated stiffness matrix parameters can be stored in the controller and directly called without the need for online real-time calculation; Specifically, the estimated adjustment force vector is obtained by left-multiplying the stiffness matrix by the foundation adjustment displacement vector: ;in , For the first The controller will apply the estimated adjustment force vector to each adjustment point; Substituting into the deformation prediction model, the output is the additional elastic deformation of the frame generated at each adjustment point under the predicted adjustment force: ;in , For the first The additional elastic deformation generated at each adjustment point under the estimated adjustment force; the controller uses the negative value of the additional elastic deformation as the pre-compensation amount for each adjustment point: ;in The pre-compensation amount is used to be superimposed in reverse in subsequent instructions to counteract the elastic deformation of the frame caused by the adjustment force. The specific construction process of the deformation prediction model is as follows: Finite element model construction method: For door and window frames in specific application scenarios, a finite element model is established according to the following method; in terms of material properties, the typical material for door and window frames is thermally broken aluminum alloy, and its elastic modulus is... Poisson's ratio Or it may be plastic steel, whose elastic modulus is... Poisson's ratio Users can input the corresponding material parameters according to the actual door and window frame material, or obtain them through actual measurement using a material testing machine; regarding geometric dimensions, they should be based on the actual width of the door and window frame to be installed. ,high A three-dimensional geometric model is established based on the profile cross-sectional parameters. For standard-sized door and window frames, parametric models with different size ranges can be pre-built, while for non-standard specifications, models are built according to the actual dimensions. Regarding element types, for profiles with uniform wall thickness, it is recommended to use four-node shell elements for mesh generation, with a mesh size controlled between 10-20 mm; for profiles with significant wall thickness variations, it is recommended to use hexahedral solid elements, with a mesh size controlled between 5-10 mm. For boundary conditions, fixed constraints are applied to the contact surface between the door / window frame and the wall opening, constraining all six degrees of freedom. If elastic gaskets or expanding foam are used to fill the gap between the door / window frame and the wall, elastic support boundary conditions can be applied to the contact surface, with the spring stiffness taken as... It can be adjusted according to the actual filling material; Compliance matrix calculation: Apply a unit adjustment force at each of the four adjustment points. The direction of the force is perpendicular to the plane of the door and window frame; through finite element static analysis, the displacement response along the Z-axis of each adjustment point under the action of a unit force is calculated. ,in Indicates the first When a unit force is applied to the adjustment point, the first... The displacement generated at each adjustment point; the displacement response coefficients at each adjustment point are combined into a 4×4 compliance matrix. Its elements ; Alternative solution: Experimental calibration method; if the user does not have the conditions for finite element analysis, the experimental calibration method can be used as an alternative; the specific operation is as follows: place the door and window frame on a rigid platform, and apply known forces at the four adjustment points respectively. Displacement sensors were used to measure the displacement response of each adjustment point under the corresponding force, and the compliance matrix was obtained by linear regression fitting. The scope of application of this method is completely equivalent to that of the finite element method, and there is no need to establish a finite element model, which lowers the implementation threshold. Deformation prediction: When the adjustment force vector is estimated Given the amount of elastic additional deformation generated at each adjustment point. Calculated by left-multiplying the adjustment force vector by the compliance matrix: ;in , For the first The additional elastic deformation along the Z-axis generated by each adjustment point under the estimated adjustment force; for door and window frames of different specifications, the corresponding compliance matrix can be established and stored in the controller according to the above method, and the corresponding compliance matrix can be called according to the actual door and window frame specifications during installation. The calculation yields a compensation command that includes pose adjustment components and internal force decoupling components, specifically including: First, the magnitude ratio of the translation error component to the rotation error component is calculated. Based on the comparison result of the ratio with a preset threshold range, it is determined whether the current error type is translation-dominant, rotation-dominant, or a hybrid type. Calculate the order of magnitude ratio of the translation error component to the rotation error component; the magnitude of the translation error component is denoted as... The unit is millimeters; the modulus of the rotational error component is denoted as... The unit is milliradians; calculate the ratio. The controller has preset threshold ranges stored internally. ,in Based on the comparison result between the ratio and the threshold interval, determine the current error type: If it is determined to be translation-dominant, it means that translation error is dominant and rotation error is negligible; if If it is determined to be rotation-dominated, it indicates that rotation error is dominant and translation error is negligible; if The result is classified as a mixed type, indicating that both translation and rotation errors are not negligible and require coordinated processing. The compensation order for each error component is determined according to a preset priority rule. This preset priority rule establishes a mapping relationship between installation quality indicators and error components, as detailed below: First priority: Diagonal length deviation caused by rotational error component; Diagonal length deviation is mainly affected by the rotational attitude of the door and window frame, especially the rotational error around the Z-axis, which will directly lead to the inconsistency of the lengths of the two diagonals; The controller detects the diagonal length deviation. If it exceeds ±2mm, it is determined that the rotational error component needs to be compensated first, and the compensation of the translational error component is temporarily suspended until the diagonal length deviation meets the standard. Second priority: horizontal and vertical deviations caused by translation error components; horizontal and vertical deviations are mainly affected by the translation error of the door and window frame in the Y-axis direction and the rotation error around the X and Y axes, with translation error being the dominant factor; after the diagonal length deviation meets the standard, the controller detects the horizontal and vertical deviations. If either exceeds ±1mm / m, it is determined that the translation error component needs to be compensated first, and the fine adjustment of the rotation error component is temporarily suspended until the horizontal and vertical deviations meet the standard. Third priority: gap uniformity deviation; gap uniformity deviation is affected by the combined effects of translation error component and rotation error component; after the deviations of the first two priorities meet the standards, the controller detects the gap uniformity deviation. If it exceeds ±1mm, the remaining translation error component and rotation error component are compensated in a coordinated manner to make the gaps around the door and window frames evenly distributed. Through the above mapping rules, the controller maps the diagonal length deviation to the rotation error component, and the horizontal and vertical deviations to the translation error component, thereby clarifying the compensation order of each error component: first compensate the rotation error component to ensure the diagonal, then compensate the translation error component to ensure the horizontality and verticality, and finally compensate the remaining translation and rotation errors in a coordinated manner to ensure the uniformity of the gap. For single error types dominated by translation or rotation, a direct reverse compensation strategy is adopted, that is, the negative value of the error component is directly used as the compensation amount, and the influence of non-dominant error components is ignored to simplify the calculation; for mixed error types, a collaborative compensation strategy is adopted, that is, translation and rotation error components are considered at the same time, and the comprehensive compensation amount is calculated by weighting to ensure that the errors of the six degrees of freedom are collaboratively corrected. The collaborative compensation strategy specifically includes the following steps: Step 1, constructing the collaborative compensation objective function: ,in The translation error component vector, The rotation error component vector, , These are the Jacobian matrices for the translation and rotation directions, respectively. The variable is the cooperative displacement increment of the regulating mechanism to be solved. The second step is to set collaborative constraints: Furthermore, the displacement difference between each adjustment mechanism does not exceed the preset cooperative tolerance threshold. To avoid generating additional internal stress; The third step involves solving the objective function using the weighted least squares method or a quadratic programming algorithm to obtain the coordinated displacement increment that simultaneously compensates for translation and rotation errors. ; The fourth step is to verify the consistency between the obtained cooperative displacement increment and the independent compensation quantities corresponding to the translation error component and rotation error component. If the deviation between the cooperative compensation quantity and the independent compensation quantity exceeds the set threshold, the weight coefficient is adjusted and the solution is recalculated until the convergence condition is met. Through the above steps, synchronous and coordinated compensation of translation and rotation errors under mixed error types can be achieved; The translation error component, rotation error component, and pre-compensation amount are weighted and superimposed to obtain the comprehensive compensation amount, and the weighting coefficients are dynamically adjusted according to the error type. Translation error components Rotational error components and pre-compensation amount Perform weighted superposition; let the weighting coefficients of the translation error components be... The weighting coefficients for the rotation error components are: The weighting coefficient of the pre-compensation amount is ,satisfy The values ​​of each weighting coefficient are dynamically adjusted according to the current error type: when translation is dominant. , and Correspondingly reduced; rotation-dominant type , and The corresponding decrease; in the case of mixed types The weighted and aggregated total compensation is: ;in The adjustment point displacement command is obtained by solving the translation error components through the inverse kinematics model. The adjustment point displacement command is obtained by solving the rotational error component through the inverse kinematics model; The compensation commands for each movable adjustment mechanism are obtained based on the inverse kinematics model: The instruction set consists of two parts: the first part is the pose adjustment component, which is used to adjust the overall pose of the door and window frame to make it approach the theoretical pose corresponding to the optimal installation target model; the second part is the internal force decoupling component, which is used to counteract the internal force coupling effect caused by the elastic deformation of the frame when the four corner adjustment mechanisms move synchronously. The physical mechanism of internal force coupling is as follows: When the four adjustment mechanisms operate independently, the force applied to each adjustment point will not only cause displacement at that point, but will also be transmitted through the elasticity of the frame to other adjustment points, causing additional displacement. For example, when the upper left corner is pushed alone, since the frame is not an ideal rigid body, the upper right corner will also produce a small displacement. This phenomenon of adjustment at one point and response everywhere is called internal force coupling. If this coupling effect is ignored, there will be an unpredictable deviation between the actual displacement and the commanded displacement of each adjustment point, resulting in a decrease in installation accuracy. The objective of solving the internal force decoupling components is to make the regulating forces among the four regulating points independent of each other, that is, the additional displacement produced by the force at any regulating point at the other regulating points is zero; mathematically, this is expressed as: let the compliance matrix be... The target internal force decoupling displacement vector is The desired regulating force vector is The decoupling objective requires In A diagonally dominated matrix, meaning that the off-diagonal elements are as small as possible; The specific solution process is as follows: The first step is to set the decoupling target; to decouple the target's internal forces into displacements. It equals the negative value of the pre-compensation amount at each adjustment point, that is ,in This refers to the additional elastic deformation caused by the estimated adjustment force. The physical meaning of this setting is: by actively applying a reverse deformation, the additional distortion caused by the coupling of the adjustment force is offset, so that the frame can be restored to a state without additional internal stress. The second step is to construct the optimization objective function; with the goal of minimizing the coupling between the regulating forces, the following objective function is established: The first term is the displacement tracking error, and the second term is the cross-coupling penalty term between the adjustment forces. This is the weighting coefficient, typically ranging from 0.1 to 0.5; The third step is to set constraints; and to adjust the force vector for internal force decoupling. The following constraints should be met: Amplitude constraint ,in The maximum output force of the adjustment mechanism is determined by hardware parameters, typically ranging from 500 to 2000 N; directional constraints. , The direction is perpendicular to the plane of the door and window frame, that is, along the Z-axis; balance constraint ,in The weight of the door and window frames is expressed in units of... That is, the resultant force of the four adjusting forces is balanced by the weight of the door and window frame; The fourth step is to solve for the optimal internal force decoupling adjustment force. A quadratic programming algorithm is used to solve the above optimization problem to obtain the optimal internal force decoupling adjustment force vector. ; The fifth step is to convert the internal force decoupling displacement components; through the inverse operation of the compliance matrix, the optimal internal force decoupling adjustment force is converted into internal force decoupling displacement components: ,in , For the first The internal force decoupling displacement component of each adjustment point, that is, the amount of displacement that needs to be added on top of the basic adjustment displacement; The internal force decoupling displacement components obtained through the above steps can decouple the adjustment forces between the four adjustment points, avoiding the generation of additional bending moments or torsional deformations. The criteria for judging whether decoupling is complete are: the displacement cross-influence coefficient between each adjustment point is less than a preset threshold, which is usually taken as 0.05, i.e., less than 5%, or the maximum equivalent stress inside the frame is less than 30% of the material yield strength. Through the above steps, synchronous and coordinated compensation of translation and rotation errors under mixed error types can be achieved; This step calculates the pre-compensation amount by obtaining the stiffness matrix and establishing a deformation prediction model. Through a hierarchical compensation decision algorithm, the compensation strategy is dynamically selected according to the error type and priority rules, and the compensation command containing the pose adjustment component and the internal force decoupling component is obtained, providing accurate execution instructions for subsequent mechanism drive and closed-loop verification.

[0025] S5. The controller drives the movable adjustment mechanisms at the four corners of the door and window frame to move synchronously according to the compensation command. At the same time, it uses a multi-dimensional sensor array to remeasure the current spatial parameters in real time, and compares the remeasured parameters with the various indicators of the optimal installation target model in a loop until all parameters meet the standards, thus completing the adaptive installation. First, the controller sends the compensation command obtained from step S4 to the movable adjustment mechanism at the four corners of the door and window frame; the movable adjustment mechanism is a precision adjustment device driven by an electric push rod or a servo motor, which is installed near the four corners of the door and window frame respectively, and each adjustment mechanism can independently perform displacement adjustment in three directions: X-axis, Y-axis and Z-axis. The movable adjustment mechanism is installed as follows: four movable adjustment mechanisms are fixedly installed near the four corners of the wall opening, and the ends of their push rods are movably connected to the four corners of the door and window frame; each adjustment mechanism can independently perform displacement adjustment along the X-axis, Y-axis, and Z-axis, and the corners of the door and window frame are displaced in the corresponding directions by the extension and retraction of the push rod; the compensation command sent by the controller is the target displacement value of each adjustment mechanism in the three directions, and the driver of the adjustment mechanism drives the push rod to move to the target position according to the command; The controller synchronously drives the four movable adjustment mechanisms according to the compensation command. Specifically, the controller superimposes the pose adjustment component and internal force decoupling component of each adjustment mechanism to generate the final execution displacement command of the adjustment mechanism, and sends them to the corresponding adjustment mechanism driver. All adjustment mechanisms execute the displacement command synchronously, while adjusting the pose of the door and window frame and canceling the internal force coupling distortion. During the adjustment mechanism's operation, the controller re-measures the current spatial parameters in real time through a multi-dimensional sensor array. Specifically, the controller reactivates the laser displacement sensor, electronic tilt sensor, and laser rangefinder to re-collect gap values ​​at multiple discrete measurement points along the perimeter of the door and window frames. Simultaneously, it collects the current horizontal tilt angle, vertical tilt angle, and the lengths of the two diagonals to form a set of re-measured spatial parameters.

[0026] The controller iteratively compares the retested parameters with various indicators of the optimal installation target model; the specific process is as follows: First, the current spatial parameters are reacquired through a multi-dimensional sensor array, including the gap values, horizontal tilt angle values, vertical tilt angle values, length values ​​of the two diagonals, and spatial coordinate information of each measurement point at multiple discrete points. Then, the controller compares the parameters obtained from the remeasurement with the corresponding indicators in the optimal installation target model and calculates the deviations: it compares the gap values ​​at each remeasured point with the target gap distribution in the optimal installation target model and calculates the gap deviation value at each point; it compares the remeasured horizontal tilt angle value with the target levelness in the optimal installation target model and calculates the levelness deviation; it compares the remeasured vertical tilt angle value with the target verticality in the optimal installation target model and calculates the verticality deviation; and it compares the remeasured length values ​​of the two diagonals with the difference in the target diagonal length in the optimal installation target model and calculates the diagonal length deviation. The controller determines whether all deviations fall within the preset error threshold range; the preset error threshold range is consistent with the installation quality specifications on which the optimal installation target model is generated, specifically: the allowable range for diagonal length deviation is ±2 mm, the allowable range for horizontal deviation is ±1 mm per meter, the allowable range for vertical deviation is ±1 mm per meter, and the allowable range for gap uniformity deviation is ±1 mm. If all deviations fall within the preset error threshold range, the installation is considered complete, the controller outputs an installation success signal, and the adaptive installation process ends. If any deviation exceeds the preset error threshold, the compensation is deemed unsatisfactory, and the controller automatically returns to perform error decoupling and subsequent steps to enter the next iteration. Specifically, the controller jumps to step S3, compares the actual pose matrix of the current retest with the theoretical pose matrix of the optimal installation target model, generates a new six-degree-of-freedom composite installation error vector, decouples it into translation error components and rotation error components, recalculates the pre-compensation amount and compensation command, and drives the adjustment mechanism to move again until all parameters meet the standard. This step achieves precise adjustment of the door and window frame position and real-time cancellation of internal force coupling by driving the four corner movable adjustment mechanisms to move synchronously according to the compensation command. By re-measuring spatial parameters in real time and comparing them cyclically with the optimal installation target model, a closed-loop adaptive installation mechanism of perception, decision-making, execution and verification is formed to ensure that the installation quality meets the specifications and overcomes the drawbacks of traditional methods that cannot automatically verify the compensation effect after adjustment and require repeated trial and error.

[0027] Example 2 describes an adaptive installation method for smart doors and windows with self-compensation for installation errors. The difference from Example 1 is that: a lightweight sensor configuration is used in the perception stage to simplify the spatial parameter acquisition process; a simplified algorithm is used in the pose calculation and error modeling stage to reduce computational complexity; and a lightweight compensation strategy is used in the elastic compensation and hierarchical decision-making stage to adapt to resource-constrained control terminals. Specifically, it includes the following steps: S1. Collect the set of lightweight space parameters between door and window frames and wall openings; Raw spatial parameters are collected using a simplified multidimensional sensor array. Only the gap values ​​at the four corners of the door and window frames, the overall horizontal tilt angle, the overall vertical tilt angle, and the lengths of the two diagonals are collected, eliminating redundant measurement points at the midpoints of each side. The sensor array is equipped with only four laser displacement sensors and two electronic tilt sensors, reducing sensor hardware costs and data acquisition volume. The collected lightweight spatial parameter set serves as the basic input data for subsequent calculations. S2. Simplify coordinate transformation calculation of the actual pose matrix of door and window frames and quickly generate the optimal installation target model; First, based on the gap values ​​at the four corner points, the geometric center of the fitted rectangle for the wall opening is directly calculated. Specifically, it is obtained by taking the arithmetic mean of the coordinates of the four corner points, omitting the complex calculation process of fitting multiple points using the least squares method. A three-dimensional spatial coordinate system is established with the geometric center as the origin. A simplified coordinate transformation method is used to calculate the six degrees of freedom parameters of the door and window frame. The translation is directly taken as the average of the gap values ​​at each corner point, and the rotation angle is estimated only through the horizontal and vertical tilt values, omitting the iterative calculation process of singular value decomposition. Based on the gap values ​​at the four corner points, the gap distribution of each edge is estimated using a linear interpolation method, replacing the cubic spline interpolation fitting in Example 1; the overall tilt angle and corner offset of each edge are extracted as feature parameters, omitting the fine extraction of local concavity and convexity; with the minimization of corner gap deviation as the optimization objective, and constrained by the diagonal length difference tolerance in the installation quality specification, the optimal installation target model is quickly generated, and the gap uniformity is gradually corrected in subsequent iterations as a secondary objective; S3. Generate a composite error vector and decouple it into translation and rotation components; The actual pose matrix is ​​compared with the theoretical pose matrix of the optimal installation target model to generate a six-degree-of-freedom composite installation error vector. Relying on the orthogonal decoupling characteristics of the error model, it is quickly decoupled into translation error components and rotation error components, simplifying the calculation process of the error inverse solution. S4. Load the preset stiffness compensation parameters and quickly generate compensation instructions using a table lookup method; A pre-set stiffness matrix parameter library for typical door and window frame specifications is provided. The stiffness matrix can be directly obtained from the parameter library based on the material and size specifications of the current door and window frame, replacing the real-time finite element analysis or online calibration in Example 1. The pre-compensation amount is quickly obtained by looking up the pre-set compensation coefficient table, and the calculation formula is simplified to a linear proportional relationship. The hierarchical compensation decision algorithm is simplified to rule-based judgment: the ratio of translation error component to rotation error component is preset to a fixed threshold to directly determine the error type; the compensation order is fixed as follows: first compensate for diagonal length deviation, then compensate for horizontal and vertical deviation, and finally compensate for gap uniformity deviation; the compensation strategy uniformly adopts a direct reverse compensation strategy, omitting the weighted superposition calculation of collaborative compensation for mixed errors; The compensation command is simplified to include only the pose adjustment component. The internal force decoupling component is quickly superimposed by looking up a preset coupling compensation coefficient table, without the need for real-time reverse calculation of the deformation prediction model. S5. The drive adjustment mechanism operates synchronously, and installation is completed through a simple cyclic comparison. The movable adjustment mechanisms at the four corners of the door and window frame are driven to move synchronously according to the compensation command. The spatial parameters are remeasured in real time through a simplified multi-dimensional sensor array. The remeasured parameters are compared with the core indicators of the optimal installation target model. If the core indicators meet the standards, the installation is considered complete. If they do not meet the standards, the process returns to the simple error decoupling and subsequent steps, and enters the next iteration. This embodiment significantly reduces the demand for sensor hardware, computing resources, and storage resources while ensuring core installation accuracy by adopting lightweight sensor configuration, simplified algorithms, and a pre-set parameter library. This enables the adaptive installation method to be adapted to cost-sensitive door and window installation scenarios and low-power embedded control terminals.

[0028] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A self-adaptive installation method for intelligent doors and windows with self-compensation function for installation errors, characterized in that, Includes the following steps: The controller activates the multi-dimensional sensor array to collect the original spatial parameter set between the door / window frame and the wall opening. The original spatial parameter set includes the gap value, horizontal tilt angle value, vertical tilt angle value, and diagonal length value. The controller receives the original set of spatial parameters, reconstructs the actual shape of the opening based on the collected gap values, and generates the optimal installation target model; at the same time, it calculates the current actual pose matrix of the door and window frames. Based on the deviation between the actual pose matrix and the optimal installation target model, an error model containing translation and rotation components is established. The controller compares the actual pose matrix with the theoretical pose matrix of the optimal installation target model to generate a six-degree-of-freedom composite installation error vector. Based on the orthogonal decoupling characteristics of the error model, the six-degree-of-freedom composite installation error vector is decoupled into translation error components and rotation error components. The controller calculates the elastic additional deformation as a pre-compensation amount based on the translation error component and rotation error component, combined with the stiffness characteristics and deformation prediction model of the door and window frame, and solves the compensation command including the pose adjustment component and the internal force decoupling component. The controller drives the movable adjustment mechanisms at the four corners of the door and window frame to move synchronously according to the compensation command. At the same time, it uses a multi-dimensional sensor array to remeasure the current spatial parameters in real time, and compares the remeasured parameters with the various indicators of the optimal installation target model in a loop until all parameters meet the standards, thus completing the adaptive installation.

2. The method according to claim 1, characterized in that, The process of generating the optimal installation target model specifically includes: Based on the collected gap values ​​from multiple discrete points, a discrete gap distribution map of the hole edge is constructed. Interpolation fitting is performed on the discrete gap distribution map to reconstruct the continuous actual shape contour of the wall opening; feature parameters are extracted from the continuous actual shape contour, including the local concavity and convexity of each side, the overall tilt angle, and the corner offset. Calculate the standard deviation of the gap and the gap deviation at each corner point based on the multiple discrete point gap values; With the goal of minimizing the mean square error of the gap and the corner gap deviation, and constrained by the horizontal tolerance, vertical tolerance and diagonal length difference tolerance in the preset installation quality specifications, the optimal installation target model is dynamically generated based on the aforementioned characteristic parameters. The theoretical pose matrix corresponding to the optimal installation target model and the error model use the same coordinate system reference.

3. The method according to claim 1, characterized in that, The process of calculating the current actual pose matrix of the door and window frames specifically includes: Based on the collected gap values, the rectangular outline is fitted using the least squares method to determine the coordinates of the four corner points of the fitted rectangle of the wall opening, and then the actual geometric center of the fitted rectangle is calculated. A three-dimensional spatial coordinate system is established with the actual geometric center as the origin, which serves as the reference coordinate system for error analysis; The collected horizontal tilt angle, vertical tilt angle, and diagonal length values ​​are transformed into the coordinate system to obtain the six degrees of freedom parameters of the door and window frame in the coordinate system. The six degrees of freedom parameters are used to construct the actual pose matrix of the door and window frame.

4. The method according to claim 3, characterized in that, The process of transforming to the coordinate system via coordinate transformation specifically includes: Using the geometric center of the fitted rectangle as a reference point, calculate the rotation matrix and translation vector corresponding to the deviation between the current actual pose and the theoretical pose of the door and window frame; The rotation matrix and translation vector are solved by singular value decomposition to obtain the optimal rigid body transformation parameters of the door and window frame from the current pose to the theoretical pose. Based on the optimal rigid body transformation parameters, the six degrees of freedom parameters of the door and window frame are transformed into the coordinate system.

5. The method according to claim 1, characterized in that, The process of obtaining the elastic additional deformation as the pre-compensation amount specifically includes: Based on the translational and rotational error components, the required basic adjustment displacement for each adjustment point is estimated using an inverse kinematics model. Based on the stiffness matrix parameters of the door and window frames, the basic adjustment displacement of each adjustment point is converted into the corresponding estimated adjustment force. Substitute the estimated adjustment force into the deformation prediction model and output the additional elastic deformation of the frame generated at each adjustment point under the action of the estimated adjustment force. The negative value of the elastic additional deformation is used as the pre-compensation amount.

6. The method according to claim 1, characterized in that, The process of calculating and obtaining a compensation command that includes pose adjustment components and internal force decoupling components specifically includes: Calculate the magnitude ratio of the translation error component to the rotation error component, and based on the comparison result of the ratio with a preset threshold range, determine whether the current error type is translation-dominant, rotation-dominant, or a hybrid type. The compensation order for each error component is determined according to a preset priority rule; For single error types dominated by translation or rotation, a direct reverse compensation strategy is adopted; for mixed error types, a collaborative compensation strategy is adopted. The translation error component, rotation error component, and pre-compensation amount are weighted and superimposed to obtain the comprehensive compensation amount, and the weight coefficients are dynamically adjusted according to the error type. The compensation commands for each movable adjustment mechanism are obtained by solving the inverse kinematics model.

7. The method according to claim 6, characterized in that, The preset priority rule is as follows: First priority: Diagonal length deviation caused by rotational error component. When the diagonal length deviation exceeds ±2mm, the rotational error component should be compensated first. Second priority: horizontal and vertical deviations caused by translation error components. When the horizontal or vertical deviation exceeds ±1mm / m, the translation error components should be compensated first. Third priority: gap uniformity deviation. When the gap uniformity deviation exceeds ±1mm, compensate for the remaining translation error component and rotation error component.

8. The method according to claim 1, characterized in that, The process of cyclically comparing the retest parameters with the various indicators of the optimal installation target model specifically includes: The current spatial parameters are reacquired using a multi-dimensional sensor array; The gap values, horizontal tilt angle values, vertical tilt angle values, and diagonal length values ​​obtained from the remeasurement are compared with the target gap distribution, target levelness, target verticality, and target diagonal length difference in the optimal installation target model, and the deviations are calculated. If all deviations fall within the preset error threshold range, the installation is considered complete; otherwise, return to perform error decoupling and subsequent steps, and proceed to the next iteration.

Images