A building block module switching method and system for flexible equipment

Abstract

The present invention discloses a building block module switching method and system for flexible equipment, the method comprising: obtaining equipment parameters of a to-be-cut flexible equipment, and performing feature analysis on the equipment parameters to obtain equipment feature values; matching a trajectory optimization model corresponding to the equipment feature values ​​to call a switching trajectory from a preset trajectory library through the trajectory optimization model; simulating a switching process corresponding to the to-be-cut flexible equipment according to the switching trajectory; adjusting the switching trajectory according to the switching process, and switching the to-be-cut flexible equipment according to the adjusted switching trajectory to improve the switching effect and efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of flexible switching control, and in particular to a building block module switching method and system for flexible equipment. Background Art

[0002] In order to realize the fast and smooth switching of the building block modules of the flexible equipment, it is necessary to design the switching trajectory corresponding to the building block modules. In the prior art, the motion characteristics of the modules and the deformation characteristics of the flexible equipment composed of the modules are often not taken into account during the optimization process of the switching trajectory, which results in that the switching trajectory finally generated cannot realize the fast and smooth switching of the building block modules.

[0003] Furthermore, in the process of optimizing the switching trajectory, the prior art does not take into account that due to the different structural dimensions and material properties of the modules, the module deformation and stress distribution under different switching trajectories should also be different, resulting in the selection of the same switching trajectory to switch different flexible devices, which usually reduces the switching effect. In addition, the prior art often uses the same pre-set switching trajectory to switch the modules of the flexible device, without adaptively optimizing the switching trajectory according to the characteristics of the flexible device, resulting in the switching effect not fitting the characteristics of the flexible device itself, reducing the efficiency and effect of the switching. Summary of the invention

[0004] In order to solve the above technical problems, the present invention discloses a building block module switching method and system for flexible equipment, which are used to improve the reliability of switching.

[0005] In order to achieve the above object, the present invention discloses a building block module switching method for flexible equipment, comprising:

[0006] Acquire equipment parameters of the flexible equipment to be switched, and perform characteristic analysis on the equipment parameters to obtain equipment characteristic values;

[0007] Matching the trajectory optimization model corresponding to the device characteristic value, so as to call the switching trajectory from a preset trajectory library through the trajectory optimization model;

[0008] Simulating the switching process corresponding to the flexible device to be switched according to the switching trajectory;

[0009] The switching trajectory is adjusted according to the switching process, and the flexible device to be switched is switched according to the adjusted switching trajectory.

[0010] The present invention discloses a building block module switching method for flexible equipment, which obtains the equipment parameters of the flexible equipment for feature analysis to match the corresponding trajectory optimization model to the flexible equipment, and then calls the switching trajectory corresponding to the flexible equipment through the trajectory optimization model to match different switching trajectories for flexible equipment with different characteristics, thereby improving the adaptability of the switching, and thereby improving the reliability of the switching. Furthermore, after calling the switching trajectory, in order to further improve the fit between the switching trajectory and the flexible equipment, the process of the switching of the flexible equipment to be switched according to the switching trajectory is simulated, so as to provide a reference for the formulation of the actual switching plan according to the switching process, and then the switching trajectory is adjusted to further improve the effect of the flexible equipment when it is switched according to the optimized switching trajectory.

[0011] As a preferred example, the step of obtaining device parameters of the flexible device to be switched and performing feature analysis on the device parameters to obtain device feature values ​​includes:

[0012] Obtaining the vibration mode parameters of the flexible device to be switched; wherein the vibration mode parameters include vibration mode parameters and deformation recovery parameters;

[0013] Extracting modal parameters from the vibration shape parameters using singular value decomposition, and constructing a stiffness matrix and a mass matrix according to the modal parameters to obtain eigenvalue parameters;

[0014] A characteristic vector is obtained according to the characteristic value parameter, and the characteristic vector is normalized to obtain the device characteristic value; wherein the characteristic vector includes curvature, length and rotation angle.

[0015] In order to improve the efficiency of switching, that is, to match the flexible device to the corresponding switching trajectory as quickly as possible, the present invention performs feature analysis on the device parameters, extracts corresponding feature values, reduces the amount of data to be processed in the subsequent matching process, and thus improves the efficiency of trajectory matching.

[0016] As a preferred example, calling the cutting trajectory from a preset trajectory library through the trajectory optimization model includes:

[0017] Acquire the trajectory parameter features corresponding to each trajectory parameter combination in the trajectory library, and calculate the similarity between each trajectory parameter feature and the device feature value through a mapping relationship preset in the trajectory optimization model;

[0018] When the similarity exceeds a preset threshold, a preset trajectory optimization rule is called to optimize the current trajectory parameter combination, and a switching trajectory corresponding to the optimized trajectory parameter combination is generated.

[0019] Based on the fact that different flexible devices have different device characteristics and different mapping relationships between different devices and different trajectories, the present invention pre-trains a trajectory model corresponding to the flexible device type to save the mapping relationship between the current type of device and the corresponding switching trajectory, and then obtains the corresponding switching trajectory from the saved multiple preset switching trajectories according to the mapping relationship, thereby ensuring the switching efficiency and improving the switching reliability.

[0020] As a preferred example, simulating the switching process corresponding to the flexible device to be switched according to the switching trajectory includes:

[0021] Constructing a virtual prototype geometric model according to the discrete points of the cutting trajectory, and meshing the virtual prototype geometric model through twenty-node hexahedral units to obtain a meshed model;

[0022] A strain energy density function is used to establish a hyperelastic constitutive equation, and deformation field data and stress field data of the meshing model are calculated by using the hyperelastic constitutive equation and an implicit dynamics solver;

[0023] Performing feature extraction on the deformation field data and the stress field data through a five-layer deep autoencoder to obtain a feature mapping relationship between the deformation field data and the stress field data;

[0024] Processing the data collected by the sensor according to the characteristic mapping relationship to obtain the displacement data and strain data of the flexible device to be switched during the switching process;

[0025] The displacement data and the strain data are fused by a Kalman filter to obtain an actual deformation curve and an actual stress distribution curve;

[0026] The reliability of the cutting trajectory is verified according to the actual deformation curve, the actual stress distribution curve, the deformation field data and the stress field data.

[0027] The present invention simulates the cutting process by constructing the virtual prototype geometric model, ensures the reliability of the trajectory optimization result through virtual prototype verification, and further optimizes the trajectory by utilizing the simulation to achieve the unity of virtual and actual, thereby improving the reliability of the optimized trajectory.

[0028] As a preferred example, the step of adjusting the switching trajectory according to the switching process, and switching the flexible device to be switched according to the adjusted switching trajectory includes:

[0029] When the reliability of the cutting trajectory is verified, a fuzzy evaluator based on a confidence interval is used to calculate the deformation evaluation parameter and the stress value evaluation parameter corresponding to the cutting trajectory, and a parameter sequence to be adjusted is generated according to the comparison results of the deformation evaluation parameter and the stress value evaluation parameter with the preset thresholds respectively;

[0030] Training a random forest regressor according to the parameter sequence to be adjusted, predicting a parameter adjustment amount through the random forest regressor, and adjusting the trajectory parameters of the cutting trajectory according to the parameter adjustment amount to obtain the updated discrete point coordinates of the cutting trajectory;

[0031] The updated discrete point coordinates of the switching trajectory are fitted using a Bezier curve, and trapezoidal speed planning is performed under the maximum speed and acceleration constraints according to the Bezier curve to obtain a speed planning curve;

[0032] The curvature value at each discrete point of the speed planning curve is calculated, and a transition point is inserted in the curvature change rate exceeding limit region by Hermite interpolation to obtain an adjusted switching trajectory.

[0033] After verifying that the cutting trajectory is reliable, the present invention further adjusts the cutting trajectory using a random forest regressor and a fuzzy evaluator, so that the adjusted cutting trajectory is more consistent with the actual motion trajectory, thereby improving the effect of cutting. Furthermore, in the process of adjusting the cutting trajectory, a transition point is inserted into the curve according to the curvature value of each discrete point in the cutting curve corresponding to the adjusted cutting trajectory, thereby ensuring the smoothness of the curve and improving the smoothness of the cutting.

[0034] As a preferred example, before obtaining the equipment parameters of the flexible equipment to be switched, the method includes:

[0035] Obtaining material parameters of the building block module to construct a deformation statics model corresponding to the flexible device composed of the building block module and performing a shearing analysis on the deformation statics model to obtain a module deformation cloud map and a module stress cloud map corresponding to each preset initial shearing trajectory of the flexible device;

[0036] Respectively extracting the discrete points of each of the initial cutting trajectories in the module deformation cloud map and the rotation angles in the module stress cloud map, and calculating the cutting speed and acceleration corresponding to each of the initial cutting trajectories at each of the discrete points, so as to obtain the module deformation change curve and the module stress change curve in the cutting process corresponding to the flexible device under different initial cutting trajectories;

[0037] Selecting a plurality of first switching trajectories from a plurality of the initial switching trajectories according to the module deformation variation curve and the module stress variation curve;

[0038] The first trajectory parameters in each of the first switching trajectories are optimized according to a preset trajectory optimization algorithm, and a plurality of optimized trajectory parameter combinations are obtained and saved in a preset trajectory library.

[0039] The present invention aims at the problem that the building block modules of flexible equipment are prone to excessive deformation and excessive stress during the rapid switching process. The present invention obtains the module structural parameters and material properties, establishes a deformation statics model that takes into account the module connection stiffness and damping, and uses finite element analysis to obtain the module deformation and stress distribution under different switching trajectories, and calculates the speed and acceleration distribution during the switching process. If the module deformation exceeds the limit or the stress exceeds the limit, the switching trajectory is judged to be infeasible and the trajectory is optimized. With the switching time as the target and the deformation and stress as the constraints, the trajectory parameters are coded and optimized to obtain a smooth switching trajectory that meets the requirements, and then the switching effect is guaranteed according to the smooth switching trajectory.

[0040] As a preferred example, the construction of a deformation statics model corresponding to the flexible device composed of the building block modules and the switching analysis of the deformation statics model to obtain the module deformation cloud map and module stress cloud map corresponding to each initial switching trajectory of the flexible device include:

[0041] Obtaining the structural dimension parameters and material elastic modulus of the building block module to calculate the stiffness matrix corresponding to the building block module; wherein the structural dimension parameters include the module contact area and the number of bolts; the stiffness matrix is ​​determined by the material elastic modulus, the module contact area and the number of bolts;

[0042] Obtaining the unit node stiffness value through the stiffness matrix and mapping it to the mesh node to establish a cubic spline curve interpolation equation; wherein the mesh node is divided using hexahedral units and the mesh size in the contact area is reduced to half of the basic mesh size;

[0043] Calculating the curvature value according to the cubic spline curve interpolation equation to obtain the strain field calculation equation; wherein the curvature value is determined by the curve tangent vector and the normal vector;

[0044] A nonlinear stress-strain equation group is established according to the strain field calculation equation to obtain the deformation field displacement distribution function and the stress field distribution function corresponding to the building block module; wherein the equation group is solved using the Newton iteration method until the residual is less than a preset value.

[0045] The present invention can reflect the deformation and stress of each part of the module based on the deformation field displacement distribution and stress field distribution, which has important guiding significance for the module structure design. Then, by adjusting the module cross-sectional size and material parameters to reduce stress concentration, a flexible equipment deformation statics model considering the building block module connection stiffness and damping is established. Then, by performing shearing analysis on the deformation statics model, the module deformation cloud map and module stress cloud map are constructed to ensure the accuracy of the subsequently established shearing trajectory.

[0046] As a preferred example, the calculation of the switching speed and acceleration corresponding to each of the initial switching trajectories at each of the discrete points to obtain the module deformation change curve and the module stress change curve during the switching process of the flexible device under different initial switching trajectories includes:

[0047] Obtaining a trajectory curve corresponding to each of the initial cutting trajectories, and obtaining a coordinate data set of discrete points on the trajectory curve; wherein different initial cutting trajectories correspond to different trajectory shapes, trajectory lengths, and trajectory curvatures;

[0048] The speed distribution value is obtained by fitting the discrete points with a Bezier curve; wherein the speed distribution value is obtained by calculating the mapping relationship between the curve parameters and the switching speed;

[0049] A kinematic equation is established according to the velocity distribution value, and the acceleration value of the discrete point is solved by using the Runge-Kutta integral formula; wherein the acceleration value is obtained by integrating the velocity with respect to time;

[0050] A state equation is constructed according to the acceleration value of the discrete point and the deformation on the module deformation cloud map, and a deformation time function is obtained by solving the deformation time function through the variational principle; wherein the deformation time function includes an acceleration coupling term;

[0051] A stress state equation is constructed for the deformation time function and the stress value in the module stress cloud map, and the stress time function is obtained by solving it through a fixed-step integrator; wherein the stress time function imposes fixed constraints and load constraints on the deformation boundary.

[0052] According to the obtained deformation cloud maps and stress cloud maps of the module under different cutting trajectory shapes, lengths and curvatures, the present invention calculates the cutting velocity distribution and acceleration distribution at the discrete points of each trajectory according to the discrete points and angles in the deformation cloud maps and stress cloud maps of different cutting trajectories, and performs numerical integration in combination with the static analysis results of the module deformation to obtain the curves of the module deformation and stress variation with time during the cutting process, so as to analyze the influence of the entire cutting trajectory on the module and improve the accuracy of the cutting trajectory analysis.

[0053] As a preferred example, the method of selecting a plurality of first switching trajectories from a plurality of the initial switching trajectories according to the module deformation change curve and the module stress change curve includes:

[0054] Constructing a constraint equation according to the preset maximum deformation of the flexible device and the current deformation of the module corresponding to each of the initial switching trajectories, so as to obtain deformation constraint optimization parameters through gradient descent iterative calculation;

[0055] A multi-objective optimization function of a deformation objective function and a stress objective function is established according to the deformation constraint optimization parameters, and a mapping relationship between a trajectory parameter vector and a constraint response vector corresponding to each of the initial cutting trajectories is obtained through a support vector regressor of a Gaussian kernel function;

[0056] According to the mapping relationship, a constrained optimization equation group is constructed using Lagrange multipliers, and the optimal trajectory parameters and Lagrange multipliers corresponding to each of the initial switching trajectories are obtained by solving them through an alternating iteration method;

[0057] If the norm difference between the optimal trajectory parameter and the initial trajectory parameter corresponding to the initial cutting trajectory is less than a preset convergence threshold, storing the optimal trajectory parameter as the first trajectory parameter of the first cutting trajectory;

[0058] If the norm difference is greater than a preset threshold, the number of iterations is increased based on the original number of iterations to continue optimizing until convergence.

[0059] The present invention optimizes the initially set trajectory by using the module deformation and module stress, thereby improving the smoothness and reliability of the trajectory.

[0060] As a preferred example, optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm includes:

[0061] Constructing a chromosome coding sequence corresponding to each of the first trajectory parameters; wherein the chromosome coding sequence adopts a binary coding method to obtain an initial trajectory parameter population; the first trajectory parameters include a trajectory shape parameter, a trajectory length parameter, a trajectory curvature parameter and a trajectory turning angle parameter;

[0062] A gradient boosting tree regressor is established for the initial trajectory parameter population, so as to obtain a mapping relationship between the motion time and the first trajectory parameter through the gradient boosting tree regressor; wherein the mapping relationship is used to calculate the fitness value corresponding to the parameter value after chromosome decoding;

[0063] A roulette wheel selector is used to select high-quality chromosomes within a preset interval, to perform a single-point crossover operation on the high-quality chromosomes to obtain offspring chromosomes, and to form a new trajectory parameter population according to the Gaussian distribution mutation of the offspring chromosomes;

[0064] Calculating the fitness value of the new trajectory parameter population, and obtaining the optimal chromosome decoding parameter when the continuous change of the fitness value is less than a preset threshold;

[0065] Each of the first trajectory parameters is optimized according to the optimal chromosome decoding parameters and the chromosome coding sequence.

[0066] The present invention encodes the shape, length, curvature and angle parameters of the cutting trajectory, obtains an optimized cutting trajectory parameter combination through selection, crossover and mutation operations, and improves the reliability of the trajectory.

[0067] As a preferred example, the step of optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm further includes:

[0068] The interpolation operation is performed using the cubic spline basis function according to the trajectory control points to obtain a piecewise cubic polynomial function expressing the trajectory curve;

[0069] Performing Fourier series expansion on the piecewise cubic polynomial function, smoothing the velocity component through a Butterworth low-pass filter and a Wiener filter, and obtaining the acceleration component from the derivative of the velocity component;

[0070] Calculate the curvature and stress value at each point according to the trajectory curve, and if the stress value is greater than the allowable stress value, add a trajectory control point at the point; wherein the spacing of the added trajectory control points is half of the original spacing;

[0071] The continuity of the trajectory curve is verified. If the derivative continuity index exceeds a preset threshold, the spline function order is increased to a preset number, and the updated trajectory curve is obtained by solving the spline interpolation coefficient matrix.

[0072] The present invention generates a smooth and continuous cutting trajectory curve by combining spline interpolation with optimized cutting trajectory parameters, and performs speed and acceleration smoothing on the trajectory curve to obtain a smooth cutting trajectory that meets the preset module deformation and stress requirements, thereby improving the reliability of the trajectory.

[0073] On the other hand, the present invention also discloses a building block module switching system for flexible equipment, including a feature analysis module, a trajectory matching module, a switching simulation module and a switching adjustment module;

[0074] The characteristic analysis module is used to obtain the equipment parameters of the flexible equipment to be switched, and perform characteristic analysis on the equipment parameters to obtain equipment characteristic values;

[0075] The trajectory matching module is used to match the trajectory optimization model corresponding to the device characteristic value, so as to call the switching trajectory from the preset trajectory library through the trajectory optimization model;

[0076] The switching simulation module is used to simulate the switching process corresponding to the flexible device to be switched according to the switching trajectory;

[0077] The switching adjustment module is used to adjust the switching trajectory according to the switching process, and switch the flexible device to be switched according to the adjusted switching trajectory.

[0078] The present invention discloses a building block module switching system for flexible equipment, which obtains the equipment parameters of the flexible equipment for feature analysis to match the corresponding trajectory optimization model to the flexible equipment, and then calls the switching trajectory corresponding to the flexible equipment through the trajectory optimization model to match different switching trajectories for flexible equipment with different characteristics, thereby improving the adaptability of the switching, and thereby improving the reliability of the switching. Furthermore, after calling the switching trajectory, in order to further improve the fit between the switching trajectory and the flexible equipment, the process of the switching of the flexible equipment to be switched according to the switching trajectory is simulated, so as to provide a reference for the formulation of the actual switching plan according to the switching process, and then the switching trajectory is adjusted to further improve the effect of the flexible equipment when it is switched according to the optimized switching trajectory. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 : A schematic diagram of a flow chart of a building block type module switching method for a flexible device disclosed in an embodiment of the present invention;

[0080] Figure 2 : is another flow chart of a building block type module switching method for a flexible device disclosed in an embodiment of the present invention;

[0081] Figure 3 : A structural schematic diagram of a building block type module switching system for flexible equipment disclosed in an embodiment of the present invention;

[0082] Figure 4 : is another structural schematic diagram of a building block type module switching system for flexible equipment disclosed in an embodiment of the present invention;

[0083] Figure 5 : A flow chart of a building block module switching method for flexible equipment disclosed in another embodiment of the present invention. DETAILED DESCRIPTION

[0084] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0085] Embodiment 1

[0086] This embodiment discloses a modular switching method for a flexible device. Specifically, the implementation process of the switching method can refer to Figure 1 , mainly including steps 101 to 104, wherein the steps are mainly:

[0087] Step 101: Obtain device parameters of the flexible device to be switched, and perform feature analysis on the device parameters to obtain device feature values.

[0088] In this embodiment, the step mainly includes: obtaining the vibration mode parameters of the flexible device to be cut; wherein the vibration mode parameters include vibration mode modal parameters and deformation recovery parameters; extracting modal parameters from the vibration mode parameters by singular value decomposition, and constructing a stiffness matrix and a mass matrix according to the modal parameters to obtain eigenvalue parameters; obtaining eigenvectors according to the eigenvalue parameters, and normalizing the eigenvectors to obtain the eigenvalues ​​of the device; wherein the eigenvectors include curvature, length and rotation angle.

[0089] In this embodiment, in order to improve the efficiency of switching, that is, to match the flexible device to the corresponding switching trajectory as quickly as possible, this step performs feature analysis on the device parameters, extracts the corresponding feature values, reduces the amount of data to be processed in the subsequent matching process, and thus improves the efficiency of trajectory matching.

[0090] Step 102: matching the trajectory optimization model corresponding to the equipment characteristic value, so as to call the switching trajectory from a preset trajectory library through the trajectory optimization model.

[0091] In this embodiment, this step mainly includes: obtaining the trajectory parameter features corresponding to each trajectory parameter combination in the trajectory library, and calculating the similarity between each trajectory parameter feature and the device characteristic value through the mapping relationship preset in the trajectory optimization model; when the similarity exceeds a preset threshold, calling the preset trajectory optimization rule to optimize the current trajectory parameter combination, and generating the switching trajectory corresponding to the optimized trajectory parameter combination.

[0092] In this embodiment, this step is based on the fact that different flexible devices have different device characteristics and different mapping relationships between different devices and different trajectories. Therefore, a trajectory model corresponding to the flexible device type is pre-trained to save the mapping relationship between the current type of device and the corresponding switching trajectory, and then the corresponding switching trajectory is obtained from the saved multiple preset switching trajectories according to the mapping relationship, which not only ensures the switching efficiency but also improves the switching reliability.

[0093] Step 103: simulating the switching process corresponding to the flexible device to be switched according to the switching trajectory.

[0094] In this embodiment, the step mainly includes: constructing a virtual prototype geometric model according to the discrete points of the cutting trajectory, and meshing the virtual prototype geometric model through twenty-node hexahedral units to obtain a meshed model; establishing a hyperelastic constitutive equation using a strain energy density function, and calculating the deformation field data and stress field data of the meshed model through the hyperelastic constitutive equation and the implicit dynamics solver; extracting features of the deformation field data and the stress field data through a five-layer deep autoencoder to obtain a feature mapping relationship between the deformation field data and the stress field data; processing the data collected by the sensor according to the feature mapping relationship to obtain the displacement data and strain data of the flexible device to be cut during the cutting process; fusing the displacement data and the strain data through a Kalman filter to obtain an actual deformation curve and an actual stress distribution curve; and verifying the reliability of the cutting trajectory according to the actual deformation curve, the actual stress distribution curve, the deformation field data and the stress field data.

[0095] In this embodiment, this step simulates the cutting process by constructing the virtual prototype geometric model, and ensures the reliability of the trajectory optimization result through virtual prototype verification. Further, the simulation is used to optimize the trajectory to achieve the unity of virtual and actual, thereby improving the reliability of the optimized trajectory.

[0096] Step 104: adjusting the switching trajectory according to the switching process, and switching the flexible device to be switched according to the adjusted switching trajectory.

[0097] In this embodiment, this step mainly includes: when verifying that the cutting trajectory is reliable, using a fuzzy evaluator based on a confidence interval to calculate the deformation evaluation parameter and the stress value evaluation parameter corresponding to the cutting trajectory, and generating a parameter sequence to be adjusted according to the comparison results of the deformation evaluation parameter and the stress value evaluation parameter with preset thresholds respectively; training a random forest regressor according to the parameter sequence to be adjusted, predicting the parameter adjustment amount through the random forest regressor, and adjusting the trajectory parameters of the cutting trajectory according to the parameter adjustment amount to obtain the discrete point coordinates of the updated cutting trajectory; fitting the discrete point coordinates of the updated cutting trajectory with a Bezier curve, and performing trapezoidal speed planning under the maximum speed and acceleration constraints according to the Bezier curve to obtain a speed planning curve; calculating the curvature value at each discrete point of the speed planning curve, inserting transition points in the curvature change rate exceeding limit region through Hermite interpolation, and obtaining the adjusted cutting trajectory.

[0098] In this embodiment, after verifying that the cutting trajectory is reliable, this step further adjusts the cutting trajectory using a random forest regressor and a fuzzy evaluator, so that the adjusted cutting trajectory is more consistent with the actual motion trajectory, thereby improving the effect of cutting. Furthermore, in the process of adjusting the cutting trajectory, transition points are inserted into the curve according to the curvature values ​​of each discrete point in the cutting curve corresponding to the adjusted cutting trajectory, thereby ensuring the smoothness of the curve and improving the smoothness of the cutting.

[0099] like Figure 1 Before calling the switching trajectory from the trajectory library, this embodiment also provides a construction process of each switching trajectory in the trajectory library. Specifically, the construction process of the switching trajectory can be referred to Figure 2 , mainly including steps 201 to 204, wherein the steps are mainly:

[0100] Step 201: Obtain material parameters of the building block module to construct a deformation statics model corresponding to the flexible device composed of the building block module and perform shear analysis on the deformation statics model to obtain the module deformation cloud map and module stress cloud map corresponding to each preset initial shear trajectory of the flexible device.

[0101] In this embodiment, the step mainly includes: obtaining the structural dimension parameters and material elastic modulus of the building block module to calculate the stiffness matrix corresponding to the building block module; wherein the structural dimension parameters include the module contact area and the number of bolts; the stiffness matrix is ​​determined by the material elastic modulus, the module contact area and the number of bolts; the unit node stiffness value is obtained through the stiffness matrix and mapped to the grid node to establish a cubic spline curve interpolation equation; wherein the grid node is divided into hexahedral units and the grid size in the contact area is reduced to half of the basic grid size; the curvature value is calculated according to the cubic spline curve interpolation equation to obtain the strain field calculation equation; wherein the curvature value is determined by the curve tangent vector and the normal vector; a nonlinear stress-strain equation group is established according to the strain field calculation equation to obtain the deformation field displacement distribution function and the stress field distribution function corresponding to the building block module; wherein the equation group is solved using the Newton iteration method until the residual is less than a preset value.

[0102] In this embodiment, this step can reflect the deformation and stress amount of each part of the module based on the deformation field displacement distribution and stress field distribution, which has important guiding significance for the module structure design. Then, by adjusting the module cross-sectional size and material parameters to reduce stress concentration, a flexible equipment deformation statics model considering the building block module connection stiffness and damping is established, and then, by performing shear analysis on the deformation statics model, the module deformation cloud map and module stress cloud map are constructed to ensure the accuracy of the subsequently established shear trajectory.

[0103] Step 202: respectively extracting discrete points of each of the initial cutting trajectories in the module deformation cloud map and the rotation angles in the module stress cloud map, and calculating the cutting speed and acceleration corresponding to each of the initial cutting trajectories at each of the discrete points, so as to obtain the module deformation change curve and the module stress change curve in the cutting process corresponding to the flexible device under different initial cutting trajectories.

[0104] In this embodiment, this step mainly includes: obtaining a trajectory curve corresponding to each of the initial cutting trajectories, and obtaining a coordinate data set of discrete points on the trajectory curve; wherein different initial cutting trajectories correspond to different trajectory shapes, trajectory lengths and trajectory curvatures; fitting the discrete points with a Bezier curve to obtain a velocity distribution value; wherein the velocity distribution value is calculated by a mapping relationship between curve parameters and cutting velocities; establishing a kinematic equation according to the velocity distribution value, and using the Runge-Kutta integral formula to solve the acceleration value of the discrete point; wherein the acceleration value is calculated by integrating the velocity over time; constructing a state equation according to the acceleration value of the discrete point and the deformation on the module deformation cloud map, and solving the deformation time function by the variational principle; wherein the deformation time function includes an acceleration coupling term; constructing a stress state equation for the deformation time function and the stress value in the module stress cloud map, and solving the stress time function by a fixed step integrator; wherein the stress time function imposes fixed constraints and load constraints on the deformation boundary.

[0105] In this embodiment, this step calculates the cutting velocity distribution and acceleration distribution at each discrete point of the trajectory according to the obtained deformation cloud map and stress cloud map of the module under different cutting trajectory shapes, lengths and curvatures, and the discrete points and angles in the deformation cloud map and stress cloud map of different cutting trajectories. The numerical integration is combined with the static analysis results of the module deformation to obtain the curve of the module deformation and stress change over time during the cutting process, so as to analyze the impact of the entire cutting trajectory on the module and improve the accuracy of the cutting trajectory analysis.

[0106] Step 203: selecting a plurality of first switching trajectories from a plurality of the initial switching trajectories according to the module deformation variation curve and the module stress variation curve.

[0107] In this embodiment, this step mainly includes: constructing a constraint equation based on the preset maximum deformation of the flexible device and the current deformation of the module corresponding to each of the initial cutting trajectories, so as to obtain the deformation constraint optimization parameters through gradient descent iterative calculation; establishing a multi-objective optimization function of the deformation objective function and the stress objective function according to the deformation constraint optimization parameters, and obtaining the mapping relationship between the trajectory parameter vector and the constraint response vector corresponding to each of the initial cutting trajectories through the support vector regressor of the Gaussian kernel function; constructing a set of constraint optimization equations using Lagrange multipliers according to the mapping relationship, and solving the optimal trajectory parameters and Lagrange multipliers corresponding to each of the initial cutting trajectories through the alternating iteration method; if the norm difference between the optimal trajectory parameters and the initial trajectory parameters corresponding to the initial cutting trajectory is less than a preset convergence threshold, then storing the optimal trajectory parameters as the first trajectory parameters of the first cutting trajectory; if the norm difference is greater than the preset threshold, then increasing the number of iterations based on the original number of iterations to continue optimizing until convergence.

[0108] In this embodiment, this step optimizes the initially set trajectory based on the module deformation and module stress to improve the smoothness and reliability of the trajectory.

[0109] Step 204: optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm, obtaining a plurality of optimized trajectory parameter combinations and saving them in a preset trajectory library.

[0110] In this embodiment, this step mainly includes: constructing a chromosome coding sequence corresponding to each of the first trajectory parameters; wherein the chromosome coding sequence adopts a binary coding method to obtain an initial trajectory parameter population; the first trajectory parameters include trajectory shape parameters, trajectory length parameters, trajectory curvature parameters and trajectory turning angle parameters; establishing a gradient boosting tree regressor for the initial trajectory parameter population, so as to obtain a mapping relationship between movement time and the first trajectory parameter through the gradient boosting tree regressor; wherein the mapping relationship is used to calculate the fitness value corresponding to the parameter value after chromosome decoding; using a roulette selector to select high-quality chromosomes within a preset interval, so as to perform a single-point crossover operation through the high-quality chromosomes to obtain offspring chromosomes, and form a new trajectory parameter population according to the Gaussian distribution mutation of the offspring chromosomes; calculating the fitness value of the new trajectory parameter population, and obtaining the optimal chromosome decoding parameters when the continuous change of the fitness value is less than a preset threshold; optimizing each of the first trajectory parameters according to the optimal chromosome decoding parameters and the chromosome coding sequence.

[0111] Furthermore, a cubic spline basis function is used to perform interpolation operations according to the trajectory control points to obtain a piecewise cubic polynomial function that expresses the trajectory curve; a Fourier series expansion is performed on the piecewise cubic polynomial function, the velocity component is smoothed by a Butterworth low-pass filter and a Wiener filter, and the acceleration component is obtained from the derivative of the velocity component; the curvature and stress value at each point are calculated according to the trajectory curve, and if the stress value is greater than the allowable stress value, a trajectory control point is added at the point; wherein the spacing of the added trajectory control points is half of the original spacing; the continuity of the trajectory curve is verified, and if the derivative continuity index exceeds a preset threshold, the spline function order is increased to a preset number, and the updated trajectory curve is obtained by solving the spline interpolation coefficient matrix.

[0112] In this embodiment, this step encodes the shape, length, curvature and angle parameters of the cutting trajectory, and obtains an optimized cutting trajectory parameter combination through selection, crossover and mutation operations to improve the reliability of the trajectory. Furthermore, by performing spline interpolation on the optimized cutting trajectory parameter combination, a smooth and continuous cutting trajectory curve is generated, and the trajectory curve is smoothed in terms of velocity and acceleration to obtain a smooth cutting trajectory that meets the preset module deformation and stress requirements, thereby improving the reliability of the trajectory.

[0113] On the other hand, this embodiment also discloses a modular switching system for flexible devices. For the specific structural composition of the switching system, please refer to Figure 3 , including a feature analysis module 301 , a trajectory matching module 302 , a switching simulation module 303 and a switching adjustment module 304 .

[0114] The characteristic analysis module 301 is used to obtain equipment parameters of the flexible equipment to be switched, and perform characteristic analysis on the equipment parameters to obtain equipment characteristic values.

[0115] The trajectory matching module 302 is used to match the trajectory optimization model corresponding to the device characteristic value, so as to call the switching trajectory from a preset trajectory library through the trajectory optimization model.

[0116] The switching simulation module 303 is used to simulate the switching process corresponding to the flexible device to be switched according to the switching trajectory.

[0117] The switching adjustment module 304 is used to adjust the switching trajectory according to the switching process, and switch the flexible device to be switched according to the adjusted switching trajectory.

[0118] In this embodiment, the switching system further includes: Figure 4 The module switching analysis module 401 , the switching trajectory analysis module 402 , the switching trajectory screening module 403 and the trajectory parameter optimization module 404 are shown.

[0119] The module switching analysis module 401 is used to obtain material parameters of the building block module, so as to construct a deformation statics model corresponding to the flexible device composed of the building block module and perform switching analysis on the deformation statics model to obtain the module deformation cloud map and module stress cloud map corresponding to each preset initial switching trajectory of the flexible device.

[0120] The cutting trajectory analysis module 402 is used to respectively extract the discrete points of each of the initial cutting trajectories in the module deformation cloud map and the rotation angles in the module stress cloud map, and calculate the cutting speed and acceleration corresponding to each of the initial cutting trajectories at each of the discrete points, so as to obtain the module deformation change curve and module stress change curve of the flexible device in the cutting process corresponding to different initial cutting trajectories.

[0121] The switching trajectory screening module 403 is used to screen a plurality of first switching trajectories from a plurality of the initial switching trajectories according to the module deformation variation curve and the module stress variation curve.

[0122] The trajectory parameter optimization module 404 is used to optimize the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm, obtain a plurality of optimized trajectory parameter combinations and save them into a preset trajectory library.

[0123] The present embodiment discloses a building block module switching method and system for flexible equipment, which obtains the equipment parameters of the flexible equipment for feature analysis to match the corresponding trajectory optimization model to the flexible equipment, and then calls the switching trajectory corresponding to the flexible equipment through the trajectory optimization model to match different switching trajectories for flexible equipment with different characteristics, thereby improving the degree of adaptability of the switching, and thereby improving the reliability of the switching. Furthermore, after calling the switching trajectory, in order to further improve the fit between the switching trajectory and the flexible equipment, the process of the switching of the flexible equipment to be switched according to the switching trajectory is simulated, so as to provide a reference for the formulation of the actual switching plan according to the switching process, and then the switching trajectory is adjusted to further improve the effect of the flexible equipment when it is switched according to the optimized switching trajectory. Furthermore, in order to solve the problem that the building block modules of flexible equipment are prone to excessive deformation and excessive stress during rapid switching, the present invention obtains the module structural parameters and material properties, establishes a deformation statics model that takes into account the module connection stiffness and damping, and uses finite element analysis to obtain the module deformation and stress distribution under different switching trajectories, and calculates the speed and acceleration distribution during the switching process. If the module deformation exceeds the limit or the stress exceeds the limit, the switching trajectory is judged to be infeasible and the trajectory is optimized. With the switching time as the target and the deformation and stress as the constraints, the trajectory parameters are coded and optimized to obtain a smooth switching trajectory that meets the requirements, and then the switching effect is guaranteed according to the smooth switching trajectory.

[0124] Embodiment 2

[0125] In the fast and smooth cutting of building block modules for flexible equipment, the cutting trajectory optimization problem of building block modules needs to consider the motion characteristics of the module and the deformation characteristics of the flexible equipment. Among them, how to establish a deformation statics model of module connection stiffness and damping according to the structural dimensions and material properties of the module, so as to obtain the module deformation and stress distribution under different cutting trajectories, is a technical difficulty. And how to calculate the velocity and acceleration distribution at each discrete point of the trajectory, and combine the static analysis results to obtain the module deformation and stress change curve over time during the cutting process, is a technical problem to be solved. If the deformation exceeds the limit of the flexible equipment or the stress exceeds the allowable stress of the material, how to optimize the cutting trajectory is a key step. In addition, in the face of different flexible equipment, how to optimize and generate several trajectories is a necessary solution step. When there is a new cutting task, how to adaptively optimize the cutting trajectory according to the characteristics of the flexible equipment to improve the efficiency and adaptability of trajectory planning is a technical goal to be achieved in this embodiment.

[0126] In order to solve the above technical difficulties and technical problems and achieve the corresponding technical goals, specifically, in a certain implementation mode provided in this embodiment, a method for fast and smooth switching of building block modules for flexible devices is disclosed, wherein the implementation process of the switching method is referred to Figure 5 , mainly including step 501 to step 508, the steps are:

[0127] Step 501: Obtain module parameters of the building block module to establish a static deformation model of the flexible device taking into account the connection stiffness and damping of the building block module, and obtain the module deformation cloud map and stress cloud map of the building block module under different cutting trajectories by performing finite element analysis on the static deformation model of the flexible device.

[0128] In this embodiment, this step mainly includes: obtaining the material parameters of the building block module to construct a deformation statics model corresponding to the flexible device composed of the building block module and performing a shearing analysis on the deformation statics model to obtain the module deformation cloud map and module stress cloud map corresponding to each preset initial shearing trajectory of the flexible device.

[0129] Specifically, the stiffness matrix is ​​calculated according to the contact surface area and the number of bolts between the building block modules; wherein, the calculation expression of the stiffness matrix is: K=EA / L, wherein K is the stiffness matrix (which takes into account the stiffness and coupling effects in different directions), E is the elastic modulus of the module material (which is a physical quantity describing the stress-strain relationship of the module within the elastic deformation range), A is the contact area (the area between the two contact surfaces, which affects the magnitude of the stiffness), and L is the module length (the length of the module in the force direction, which affects the magnitude of the stiffness). Then, a preload F is applied to the contact surface and the damping coefficient is calculated: C=2mξωn, wherein C is the damping coefficient, which is a physical quantity used to describe the energy dissipation capacity of the system during vibration, m is the module mass, ξ is the damping ratio, which describes the proportional relationship between damping and critical damping, and ωn is the natural frequency, that is, the free vibration frequency when there is no external excitation. The building block module dynamics equation is obtained by combining the stiffness matrix with the damping matrix, and the building block module contact structure is meshed using hexahedral units. The minimum size of the hexahedral unit is one tenth of the minimum characteristic size of the module, and the mesh size in the contact area is reduced to one half of the basic mesh size.

[0130] The unit node stiffness value is obtained from the stiffness matrix and mapped to the grid node. The Coulomb contact calculation unit is established through the contact surface friction coefficient μ and the preload F. The cubic spline curve interpolation equation is constructed according to the coordinate data of the cutting trajectory, and the curvature is calculated using the variational principle. Next, the strain field: ε=κh / 2 calculation equation is established, h is the module section height, and displacement and load constraints are applied to the deformation boundary. A nonlinear stress-strain relationship is established based on the strain field, and the Newton iteration method is used to solve the nonlinear equation corresponding to the nonlinear stress-strain relationship until the residual is less than the set value, and then the deformation field displacement distribution function and the stress field distribution function are obtained, so as to calculate the displacement of each node and the Mises stress value through the deformation field displacement distribution function and the stress field distribution function.

[0131] In a specific application scenario, the calculation of the connection stiffness of the building block module takes into account the material properties and geometric parameters of the module. For example, when the elastic modulus of the module is 70GPa, the contact area is 50mm×50mm, and the module length is 100mm, the stiffness value of a single connection node is 17500N / mm, and the contact surface stiffness is significantly improved when the preload is 1000N. The damping characteristics of the building block module are derived from the internal friction and structural deformation of the material. For example, in the robot joint module, it is manifested as an under-damped characteristic with a natural frequency of 5Hz and a damping ratio of 0.05. The damping coefficient matrix reflects the energy dissipation during the movement of the module. Selecting the appropriate unit size during the meshing process is crucial to the calculation accuracy. For example, in the design of industrial robot arms, when the minimum characteristic size of the module is 10mm, the basic grid size is 1mm, and the contact area grid size is 0.5mm to ensure the accuracy of contact stress calculation. The node stiffness value is obtained from the stiffness matrix and assigned to the grid unit. The friction coefficient of the contact surface is 0.15, and the preload force is 1000N to establish the Coulomb contact constraint. The shape of the cutting trajectory curve determines the deformation degree of the module. In the trajectory planning of the flexible manipulator end of the industrial robot arm design, a smooth and continuous deformation curve is obtained by fitting the trajectory points with a cubic spline curve. For the arc trajectory with a radius of 100mm, the curvature value remains constant at 0.01 / mm. When the module section height is 20mm, the maximum strain of the outer surface is 0.0001. Displacement constraints and load constraints are applied to the deformation boundary to ensure the stability of the deformation process.

[0132] The nonlinear stress-strain relationship reflects the material properties during large deformation. When the strain is less than 0.001, it is approximately a linear relationship. When the strain exceeds 0.001, it shows obvious nonlinear characteristics. The converged solution is obtained by iterative calculation. The iteration is stopped when the residual is less than 0.0001. The deformation field displacement distribution reflects the deformation of each part of the module. The maximum displacement occurs at the free end. The Mises stress distribution reveals the stress concentration area, and the maximum stress occurs at the fixed end. The deformation field stress distribution has important guiding significance for the design of different module structures. For example, in the design of industrial robots, stress concentration is reduced by adjusting the module cross-sectional size and material parameters. In the design of medical robots, deformation accuracy and positioning accuracy are emphasized. In the field of aerospace, structural lightweight and reliability are emphasized. There are significant differences in the design focus and constraints in different application scenarios.

[0133] Step 502: According to the module deformation cloud map and stress cloud map, the number of discrete points and rotation angles in the deformation cloud map and stress cloud map of different modules are obtained, and the switching velocity distribution and acceleration distribution at each discrete point of the trajectory are calculated, so as to perform numerical integration in combination with the static analysis results of the module deformation, and obtain the module deformation and stress change curves over time during the switching process.

[0134] In this embodiment, this step mainly includes: extracting the discrete points of each of the initial cutting trajectories in the module deformation cloud map and the turning angles in the module stress cloud map respectively, and calculating the cutting speed and acceleration corresponding to each of the initial cutting trajectories at each of the discrete points, so as to obtain the module deformation change curve and module stress change curve of the flexible device in the cutting process corresponding to different initial cutting trajectories.

[0135] Specifically, a coordinate data set of discrete points on the trajectory curve is obtained according to the shape of the cutting trajectory, and the curvature value is calculated for the discrete points. The calculation expression of the curvature value is: k=|y''| / (1+y'^2)^(3 / 2), where y'' is the second-order derivative of the trajectory and y' is the first-order derivative of the trajectory. The Bezier curve is used for smooth fitting, and the key points of the trajectory are selected as control points to construct a cubic Bezier curve. The curve parameters and the cutting speed satisfy the mapping relationship: v=ds / dt, s is the arc length parameter of the curve, and t is the time parameter, so as to calculate the cutting speed distribution value at each discrete point. A kinematic equation is established for the speed value of the discrete points of the cutting trajectory. The expression of the kinematic equation is: dx / dt=v(t), x is the position vector, and v is the velocity vector. The fourth-order Runge-Kutta integral formula is used to solve the integral of the velocity over time. The calculation expression of the fourth-order Runge-Kutta integral formula is: y n+1 =y n+h / 6(k1+2k2+2k3+k4), h is the integration step, k1 to k4 are the integration coefficients.

[0136] Calculate the acceleration value of each discrete point: a=dv / dt, and obtain the deformation and Mises stress value at the corresponding point in the deformation cloud map and stress cloud map. Construct the state equation based on the deformation of the discrete points on the deformation cloud map and the static analysis results. The state equation is: du / dt=F(u,t), u is the deformation displacement vector, and F is the deformation field function.

[0137] The variational principle is used to establish the integral interval constraint, so as to solve the deformation time-varying function through the implicit integrator and obtain the deformation time curve including the acceleration coupling term. The state equation is constructed for the stress value of the discrete point in the stress cloud diagram and the static analysis results: dσ / dt=G(σ,u,t), σ is the stress tensor, G is the stress field function, and u is the deformation displacement field. Fixed constraints and load constraints are applied to the deformation boundary, and the fourth-order fixed step integrator is used to solve the stress time-varying function. The integration step is one-fourth of the minimum characteristic time, and the relative error is controlled within one thousandth. The shape of the cutting trajectory determines the motion characteristics of the entire flexible mechanism. In the design of the flexible arm of the industrial robot, the curvature value reflects the curvature of the trajectory. The curvature of the straight segment is zero, and the curvature of the arc segment is constant. When the curvature value exceeds the maximum curvature allowed by the module, it will cause excessive deformation. Taking the arc trajectory with a radius of 200 mm as an example, when the discrete point interval is 5 mm, the curvature value is 0.005 per mm, and the velocity distribution in the arc segment remains constant. The selection of integral step size during kinematic solution has a significant impact on the calculation accuracy. In the application of medical surgical robots, when the end effector of the robotic arm moves along the preset trajectory, the cutting speed varies in the range of 0.1-1 meters per second, and the integral step size is 0.001 seconds. The Runge-Kutta integral ensures the stability of numerical calculations. The acceleration value reflects the rate of change of the speed. The acceleration value is large at the corner of the trajectory, causing a large inertial load. The deformation field calculation involves complex mechanical processes. In the aerospace flexible mechanism, the deformation obtained by static analysis is used as the initial condition. The inertial force during the movement will cause additional deformation. For an aluminum alloy module with a length of 500 mm and a cross-section of 20×5 mm, the static maximum deformation is 2 mm. Due to the influence of inertial force in the dynamic process, the maximum deformation increases to 3 mm, and the deformation changes periodically over time. The stress field distribution reflects the internal stress state of the structure. In industrial robot applications, the module material is aluminum alloy with an elastic modulus of 70 GPa, a yield strength of 200 MPa, and a static maximum stress of 100 MPa, located near the fixed end. During the dynamic process, due to the influence of inertial loads, the maximum stress increases to 150 MPa, and the stress fluctuation amplitude is 50 MPa. The stress field boundary constraints include full constraints at the fixed end and load constraints at the free end. The coupling effect of deformation and stress plays an important role in the dynamic characteristics of flexible mechanisms. When the module reciprocates at a frequency of 5 Hz, both the deformation and stress values ​​show obvious dynamic response characteristics. The deformation phase lags behind the load phase, and stress fluctuations cause fatigue effects in the structure. In medical robot applications, the dynamic response amplitude is reduced by optimizing trajectory planning and velocity distribution. In the aerospace field, attention is paid to the lightweight and reliability of the structure.

[0138] Step 503: If the module deformation exceeds the preset maximum deformation of the flexible device, or the module stress exceeds the preset allowable material stress, it is determined that the cutting trajectory is not feasible, and the cutting trajectory optimization is performed; otherwise, the cutting trajectory is used as the optimized initial cutting trajectory, and the next step of optimizing the cutting trajectory parameters is entered.

[0139] In this embodiment, the step mainly includes: selecting a plurality of first switching trajectories from a plurality of the initial switching trajectories according to the module deformation change curve and the module stress change curve;

[0140] Specifically, a constraint equation is constructed according to the preset maximum deformation of the flexible device and the current deformation of the module: f(x)=||u(x)-umax||, where u(x) is the deformation vector, and umax is the preset maximum deformation vector. A gradient descent iterative calculation with a learning rate of 0.01 is used, and the Mises stress is compared with the allowable stress of the material, and then displacement constraints are imposed on the boundary of the deformation field and stress constraints are imposed on the boundary of the stress field. Furthermore, a multi-objective optimization function is established for the cutting trajectory parameters: F(x)=w1f1(x)+w2f2(x), where f1(x) is the deformation objective function, f2(x) is the stress objective function, w1 and w2 are weight coefficients, and the support vector regressor of the Gaussian kernel function is used to construct the mapping relationship between the trajectory parameter vector x and the constraint response vector y: y=Σ(αi-αi*)K(xi,x), where αi and αi* are Lagrange multipliers, and K(xi,x) is the Gaussian kernel function, so as to solve the optimal trajectory shape parameters under the preset deformation and allowable stress constraints. The displacement and stress values ​​in the deformation cloud map and stress cloud map are calculated according to the optimized trajectory parameters, and the Lagrange multiplier λ is used to construct the constrained optimization equation group. The optimal trajectory parameter x and the Lagrange multiplier λ are solved by the alternating iteration method, and the trajectory parameter combination that meets the constraint conditions is obtained under the dual constraints of deformation and stress. Among them, the convergence is judged based on the trajectory optimization result. If the norm difference between the trajectory parameter vector x0 before optimization and the trajectory parameter vector x after optimization ||x-x0|| is less than the preset convergence threshold ε=0.001, the set of trajectory parameters is stored as the initial cutting trajectory. If the norm difference is greater than the preset threshold, 50 times are added to the original number of iterations to continue the optimization calculation until the convergence condition is met or the maximum number of iterations is 500. In some implementations of this embodiment, the deformation constraint of flexible equipment and the material stress constraint play a key role in the manufacture of industrial robots. Taking the arm structure of an industrial robot as an example, the preset maximum deformation limit of the flexible module made of aluminum alloy is within 2 mm, and the allowable stress is limited to within 100 MPa. A smaller learning rate of 0.01 is used in the gradient descent optimization process to ensure convergence stability, and the displacement and stress boundary constraints are used to ensure that the optimization results meet the actual engineering needs. Trajectory parameter optimization involves multi-objective trade-offs. In the application of medical surgical robots, the weight of the deformation objective function is 0.6, and the weight of the stress objective function is 0.4, which reflects the higher requirements for positioning accuracy. The selection of Gaussian kernel function parameters of the support vector regressor has a significant impact on the mapping relationship. When the kernel width parameter is 0.1, it has good generalization performance. The trajectory shape parameters include geometric features such as curvature and length. The Lagrange multiplier method realizes constrained optimization solution in the design of aerospace flexible mechanisms. Taking the solar panel deployment mechanism as an example, the initial value of the Lagrange multiplier of the deformation constraint and the stress constraint is 1.0. In the alternating iteration process, the weight of the deformation constraint gradually increases with the number of iterations, and the weight of the stress constraint remains constant, ensuring that the final trajectory parameters meet the stiffness and strength requirements.The convergence judgment criterion reflects the accuracy requirements of optimization calculations in surgical robot applications. Convergence is determined when the difference in the norm of the trajectory parameter vector is less than 0.001. Taking the robotic arm deployment process as an example, the curvature value before optimization is 0.01 per millimeter and the length is 300 millimeters. After 200 iterations of optimization, the curvature value drops to 0.008 per millimeter and the length increases to 320 millimeters, meeting the convergence conditions and serving as the initial cutting trajectory. Multi-objective optimization has different focuses in different application scenarios. Industrial robots pay more attention to the balance between motion accuracy and efficiency, medical robots emphasize position accuracy and safety, and the aerospace field pursues the unity of lightweight and reliability. The selection of optimization parameters and the setting of constraint conditions need to match the specific application requirements. Under the guidance of deformation cloud maps and stress cloud maps, trajectory parameter combinations that meet multiple constraints are obtained through iterative optimization. The optimization process under complex constraint conditions reflects the trade-offs in engineering practice. Taking the robotic arm deployment mechanism as an example, the preset deformation constraint and the allowable stress constraint restrict each other. Increasing the structural stiffness will lead to an increase in mass, and reducing the stress level will limit the motion performance. Through multi-objective optimization and constraint solving, a balance point is found between conflicting indicators to achieve comprehensive optimization of the flexible mechanism in terms of stiffness, strength and motion performance.

[0141] Step 504: Taking the cutting motion time as the optimization target, the module deformation and stress constraint as the optimization conditions, the cutting trajectory shape, length, curvature and angle parameters are encoded, and the optimized cutting trajectory parameter combination is obtained through selection, crossover and mutation operations.

[0142] In this embodiment, this step mainly includes: optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm, obtaining a plurality of optimized trajectory parameter combinations and saving them into a preset trajectory library.

[0143] Specifically, the chromosome coding sequence is constructed according to the shape, length, curvature and turning parameters of the cutting trajectory, where the trajectory shape parameter is coded in 16-bit binary, the length parameter is coded in 12-bit binary, the curvature parameter is coded in 10-bit binary, and the turning parameter is coded in 8-bit binary. The actual parameter value is calculated by x=xmin+(xmax-xmin)×b / 2n, where xmin and xmax are parameter boundary values, b is a binary value, and n is the number of coding bits. The population size is set to 100, the mutation probability is 0.01, and the initial trajectory parameter population is generated under the constraints that the deformation is less than the preset value and the stress is less than the allowable value. The fitness evaluation is performed on the trajectory parameter population, and the mapping relationship f(x) between the motion time and the trajectory parameter is established through a gradient boosting tree regressor with a maximum depth of 6 and a number of trees of 100. The least squares loss function is used to train the regressor, and the motion time corresponding to the parameter value after chromosome decoding is calculated as the fitness value. The selection probability is calculated according to the fitness value, and a roulette selector is used to generate random numbers in the interval [0,1] to select high-quality chromosomes. According to the selection results, the crossover operation is performed. The single-point crossover operator is used to randomly select the crossover point position in the chromosome sequence. The gene sequence after the crossover point of the two parent chromosomes is exchanged to generate the daughter chromosome. The gene position of the daughter chromosome is mutated through a Gaussian distribution with a mean of 0 and a standard deviation of 0.1. The deformation and stress value under the decoding parameters of the daughter chromosome are calculated, and the chromosomes that exceed the constraint range are eliminated. The fitness value is calculated for the newly generated trajectory parameter population. The optimal fitness value is less than 0.001 for 5 consecutive generations as the convergence criterion. The optimal chromosome after convergence is decoded and restored to obtain the trajectory shape, length, curvature and angle parameter combination. If the convergence condition is not met, the roulette selection operation is returned to continue the evolutionary calculation until the maximum iteration generation is 500 generations. In certain implementations of this embodiment, the cutting trajectory parameter encoding reflects the trade-off between parameter accuracy and computational efficiency in industrial robot applications. The trajectory shape uses 16-bit encoding to provide higher shape accuracy. In the smooth transition from a straight line segment to an arc segment, the shape parameter changes from 0 to 1, and the accuracy reaches 0.00002. The 12-bit encoding of the length parameter covers the motion range of 300 to 500 mm. The 10-bit encoding of the curvature parameter corresponds to a curvature change of 0 to 0.02 per millimeter. The 8-bit encoding of the angle parameter meets the requirements of 0 to 90 degrees. Medical surgical robots have higher requirements for trajectory optimization. The depth selection of the gradient boosting tree reflects the complexity of the model. The decision tree with a depth of 6 performs segmented fitting on the nonlinear relationship between trajectory parameters and motion time. The integration of 100 trees improves the prediction stability. In the optimization of the motion trajectory of surgical instruments, the motion time is reduced from 2 seconds to 1.5 seconds, while keeping the end position accuracy within 0.1 mm.The selection operation of the genetic algorithm achieves the retention of high-quality solutions in aerospace flexible mechanisms. The population size of 100 ensures the breadth of the search space, and the mutation probability of 0.01 strikes a balance between maintaining population diversity and convergence. Taking the solar panel deployment mechanism as an example, the roulette selector selects the parent chromosome according to the probability interval of the normalized fitness value of 0.2 to 0.8, and high-quality individuals get more inheritance opportunities. Crossover and mutation operations generate new solution spaces in the optimization design of industrial robotic arms. Single-point crossover randomly selects crossover points in the 46-bit coding sequence, and parent chromosomes exchange gene fragments to produce offspring. Gaussian mutation introduces random perturbations while maintaining parameter continuity. The standard deviation of 0.1 controls the amplitude of variation. Chromosomes with deformation exceeding 2 mm or stress exceeding 100 MPa are eliminated to ensure that the offspring meet the constraints. The convergence judgment criteria are different in different application scenarios. Industrial robots focus on optimization efficiency. When the change of the optimal value for five consecutive generations is less than 0.001, it is considered convergence. Medical robots emphasize optimization accuracy. The convergence threshold is reduced to 0.0005. Spacecraft mechanisms require higher reliability. The iteration generation is increased to 1000 generations. In the optimization of the robot arm trajectory, the three indicators of motion time, deformation and stress are optimized at the same time. The chromosome encoding and decoding process reflects the accuracy of parameter mapping. In industrial robot applications, the 16-bit shape parameters are converted from binary to the real number domain of 0 to 1, with a resolution of 0.00002. In medical robots, the 12-bit length parameters are mapped to the range of 300 to 500 mm with an accuracy of 0.1 mm. In spacecraft, the 10-bit curvature parameters correspond to the range of 0 to 0.02 per mm, which accurately describes the structural deformation characteristics.

[0144] Step 505: Generate a smooth and continuous cutting trajectory curve by combining spline interpolation with the optimized cutting trajectory parameters, and perform velocity and acceleration smoothing on the trajectory curve to obtain a smooth cutting trajectory that meets the preset module deformation and stress requirements.

[0145] In this embodiment, this step mainly includes: optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm, obtaining a plurality of optimized trajectory parameter combinations and saving them into a preset trajectory library.

[0146] Specifically, a cubic spline interpolation function is constructed based on the optimized combination of cutting trajectory parameters. The node density ρ=ρ0×(1+10κ) is increased by the curvature threshold κ>0.01 / mm, ρ0 is the basic node density, and the cubic spline basis function is used to perform interpolation operations at the trajectory control points. A piecewise cubic polynomial is constructed: P(t)=ΣPiNi(t) to express the trajectory curve, where Pi is the coordinate of the control point and t is the parameter variable. The polynomial function is expanded into a Fourier series form. The frequency domain processing is performed on the velocity component of the Fourier series expansion. The high-frequency terms are truncated using a Butterworth low-pass filter with a cutoff frequency of one-tenth of the sampling frequency. The velocity curve is smoothed by a Wiener filter with a signal-to-noise ratio of 20. The acceleration component is obtained by taking the derivative of the smoothed velocity curve v(t): a(t)=dv / dt. The same filtering and smoothing operation is performed on the acceleration curve to achieve a continuous transition between velocity and acceleration. The curvature at each point is calculated based on the smoothed trajectory curve, and control points are added at the locations where the deformation is greater than the preset value or the stress is greater than the allowable value. The spacing between the control points is half of the original spacing, and the trajectory curve shape is updated by solving the new spline interpolation coefficient matrix. The continuity of the optimized trajectory is verified. The position continuity requires that the position error of adjacent segments is less than 0.1mm, the velocity continuity requires that the derivative error is less than 0.1m / s, and the acceleration continuity requires that the second-order derivative error is less than 0.1m / s². If the curve derivative continuity index exceeds the requirement, the spline function order is increased from cubic to quintic, and the spline interpolation coefficient is recalculated and the continuity index is verified. Spline interpolation reflects the relationship between node distribution and curve smoothness in industrial robot trajectory planning. The curvature threshold reflects the degree of curvature of the trajectory. When the curvature exceeds 0.01 per millimeter, the node density is increased. The basic node spacing of the arc segment with a radius of 100 mm is 5 mm. After considering the curvature factor, the node spacing is reduced to 2 mm. The cubic spline basis function ensures the continuity of the second-order derivative of the curve. Speed ​​smoothing is concerned with the stability of motion in medical surgical robots. Fourier series expansion decomposes the trajectory curve into different frequency components. When the sampling frequency is 1000 Hz, the cutoff frequency is 100 Hz to eliminate high-frequency oscillation components. The Wiener filter with a signal-to-noise ratio of 20 effectively suppresses noise while maintaining signal characteristics. The speed curve accelerates from 0 to 500 mm / s more smoothly. Curvature and stress calculations reflect the structural deformation characteristics in the design of aerospace flexible mechanisms. The elastic modulus of the aluminum alloy module is 70 GPa, the section height is 20 mm, and the curvature of 0.01 per mm corresponds to a bending stress of 7 MPa. The control point density is increased in dangerous areas where the stress exceeds 100 MPa. The original control point spacing is reduced from 10 mm to 5 mm. The update of the spline interpolation coefficient reduces local stress concentration.Continuity verification ensures the smoothness of the motion trajectory in industrial robot applications. The position continuity requirement of 0.1 mm error limit is suitable for precision assembly operations. The speed continuity limit of 0.1 m / s avoids sudden changes in speed. The acceleration continuity limit of 0.1 m / s2 reduces the impact load. The quintic spline function provides higher-order derivative continuity. Different application scenarios have different requirements for trajectory smoothness. Industrial robots focus on the balance between motion efficiency and accuracy, and use larger node spacing to improve calculation speed. Medical robots emphasize motion stability and perform more stringent speed smoothing. Spacecraft mechanisms focus on structural safety and reduce stress levels by refining control points. Continuity verification indicators are adjusted accordingly with application requirements. Multiple optimization objectives constrain each other. Increasing node density improves curve accuracy but reduces calculation efficiency. Speed ​​smoothing improves stability but prolongs motion time. Stress control enhances safety but limits motion flexibility. Continuity constraints improve smoothness but increase optimization difficulty. Balance is achieved between multiple objectives by adjusting various parameters.

[0147] Step 506: According to the deformation characteristics and vibration frequencies of different flexible devices and the optimized cutting trajectory parameter combinations, different trajectory optimization models are established respectively, and a trajectory parameter library and a trajectory generation rule library are formed. If there is a new cutting task, according to the vibration mode and deformation recovery ability of the flexible device, the corresponding optimization model and generation rule are retrieved and matched in the trajectory parameter library and the trajectory generation rule library to optimize the cutting trajectory.

[0148] In this embodiment, this step mainly includes: obtaining equipment parameters of the flexible equipment to be switched, and performing feature analysis on the equipment parameters, so as to call the switching trajectory from the trajectory library through the trajectory optimization model.

[0149] Specifically, the first ten modal parameters are extracted by singular value decomposition matrix decomposition according to the vibration characteristics of flexible equipment. The generalized eigenvalue problem Kφ=ω²Mφ is constructed through the stiffness matrix K and the mass matrix M, where φ is the vibration vector and ω is the natural frequency. The eigenvectors including curvature κ, length L, and angle θ are extracted from the optimized trajectory parameter combination, and the feature data are normalized and stored in the parameter library. A five-layer deep neural network is established for the trajectory parameter characteristics and vibration parameters, and three convolution layers are used to extract features. The convolution kernel sizes are 5×5, 3×3, and 3×3, respectively. Feature mapping is achieved through two fully connected layers, and the number of hidden layer nodes is 256 and 128, respectively. The feature similarity is calculated based on the Euclidean distance, the training batch is 64, and the learning rate is 0.001. The trained weights and the similarity threshold of 0.9 are recorded in the rule library. According to the calculation of the structural response characteristic vector of the flexible equipment of the new switching task, the cosine similarity is used to retrieve the trajectory parameter combination with the highest matching degree in the parameter library. If the similarity exceeds 0.9, the corresponding optimization rule in the rule library is directly applied to adjust the trajectory parameters. The optimization rule includes the parameter correction amount The mapping relationship f( ,b). Calculate the structural response based on the optimized trajectory parameters, including the maximum deformation umax, natural frequency ω, and damping ratio ξ. If the response characteristics meet the preset index requirements, the optimization parameters and rules are stored in the parameter library and rule library respectively. If they exceed the preset range, a genetic planner with a crossover probability of 0.8 and a mutation probability of 0.1 is used to generate new optimization rules. Among them, the fitness function is calculated using the weighted index deviation function.

[0150] The vibration characteristics of flexible equipment are reflected in industrial robots as multi-order modal coupling. The first ten modes extracted by singular value decomposition cover the main vibration characteristics. Taking the six-axis robot as an example, the first-order natural frequency of 5 Hz corresponds to the overall swing, the second-order 8 Hz corresponds to the arm bending, and the high-order modes reflect the local deformation characteristics. The trajectory feature vector contains key parameters such as the maximum curvature of 0.01 per millimeter, the trajectory length of 500 mm, and the maximum turning angle of 45 degrees. Deep neural networks realize parameter mapping and feature extraction in medical surgical robots. The three-layer convolution structure corresponds to different scale features. The 5×5 convolution kernel extracts the large-scale trajectory shape, and the 3×3 convolution kernel focuses on local detail features. The number of nodes in the fully connected layer is reduced from 256 to 128, reflecting the layer-by-layer compression of the feature dimension. The batch size of 64 during training balances the computational efficiency and optimization effect, and the learning rate of 0.001 ensures convergence stability. Similarity retrieval reflects knowledge reuse and optimization efficiency in aerospace flexible mechanisms. The cosine similarity threshold of 0.9 ensures matching accuracy. Taking the solar panel deployment mechanism as an example, the deformation threshold of 2 mm, the frequency range of 5-10 Hz, and the damping ratio of 0.05-0.1 constitute the response feature space. The optimization rule maps the parameter deviation to the constraint boundary, realizing rapid parameter adjustment. Genetic programming balances multiple optimization goals in industrial robot applications. The crossover probability of 0.8 ensures the inheritance of high-quality rules, and the mutation probability of 0.1 introduces new search directions. The deformation weight of 0.4, the frequency weight of 0.3, and the damping weight of 0.3 in the fitness function reflect the importance of different indicators. The dynamic update of the rule base improves the optimization adaptability. Different application scenarios have differentiated requirements for parameter optimization. Industrial robots emphasize motion efficiency, medical robots focus on positioning accuracy, and spacecraft mechanisms focus on reliability. The rule base records the mapping relationship between scene characteristics and optimization strategies, and establishes the association between the dynamic characteristics of flexible mechanisms and trajectory parameters through deep learning. The process of establishing the parameter library and the rule library reflects the accumulation of knowledge and the refinement of experience. Industrial robots have accumulated optimization parameters under different loads and speeds, medical robots have recorded the trajectory rules of various surgical actions, and spacecraft have stored the deformation characteristics in a weightless environment. Data from multiple fields complement each other to form a complete knowledge system.

[0151] Step 507: Develop a virtual prototype simulation system, import the optimized discrete points of the cutting trajectory into the virtual prototype, simulate the module movement, deformation and stress during the cutting process, and compare them with the actual deformation of the flexible device to verify the reliability of the cutting trajectory.

[0152] In this embodiment, this step mainly includes: simulating the switching process corresponding to the flexible device to be switched according to the switching trajectory.

[0153] Specifically, the geometric model of the virtual prototype is constructed according to the discrete points of the optimized trajectory, and the grid is divided by twenty-node hexahedral units. The unit size is one-tenth of the characteristic size, and the grid distortion rate is controlled above 0.8. Displacement boundary conditions are applied at discrete points, and the deformation field and stress field are calculated using the strain energy density function and the implicit dynamics solver with a time step of 0.001 seconds. A five-layer deep autoencoder is used to extract features based on the simulation results of the virtual prototype. The number of nodes in the input layer and the output layer is the same as the number of grid nodes, and the number of nodes in the three hidden layers is 1024, 256, and 1024 respectively. The mean square error loss function is used to train the network, and the learning rate is set to 0.001. The compressed feature vector is mapped to the measured data feature in the 256-dimensional feature space, and the verification is completed when the feature reconstruction error is less than 1%. According to the measured data collected by the displacement sensor and strain gauge of the flexible equipment, the sensors are arranged at the maximum deformation and stress concentration area. The db4 wavelet basis function is used to perform five-layer decomposition and noise reduction on the sensor data. The multi-sensor data is fused through the Kalman filter with a preset state equation, where the displacement data weight is 0.6 and the strain data weight is 0.4, and the actual deformation curve and stress distribution curve are calculated. The root mean square error distribution is calculated for the simulation data and the measured data, and the node displacement error index e1 and the unit stress error index e2 are used. If e1 exceeds 0.1 mm or e2 exceeds 1 MPa, the grid size of the error-exceeding area is reduced to half of the original size, and the virtual prototype simulation results are recalculated until the error indicators of all areas meet the requirements or the grid encryption times reach three times. Virtual prototype meshing reflects the balance between accuracy and efficiency in industrial robots. Hexahedral units have higher computational accuracy than tetrahedral units. Taking the robotic arm module as an example, the grid size corresponding to the characteristic size of 100 mm is 10 mm. The grid distortion rate of 0.8 ensures the quality of the unit shape. In the hyperelastic constitutive equation, C1 takes 2 MPa and C2 takes 0.5 MPa to describe the nonlinear characteristics of the material. The time step of 0.001 second ensures the stability of dynamic calculation. Deep autoencoders realize feature extraction and dimensionality compression in medical surgical robots. The number of input layer nodes corresponds to the number of grid nodes, reaching 10,000. The feature dimension is reduced to 256 through three hidden layers, and the compression ratio reaches 97.4%. The mean square error loss function is reduced to 0.0001, indicating good reconstruction accuracy. The feature mapping error is less than 1%, which verifies the effectiveness of data dimensionality reduction. The sensors are arranged in aerospace flexible structures to focus on the structural response characteristics. The displacement sensor is installed at the maximum deformation position to measure the millimeter-level displacement. The strain gauge is arranged in the stress concentration area to monitor the megapascal-level strain. The db4 wavelet five-layer decomposition realizes data noise reduction, and the signal-to-noise ratio is improved from 10 to 30. The displacement data weight of 0.6 reflects the importance of deformation characteristics, and the strain data weight of 0.4 reflects the necessity of stress monitoring.Error evaluation verifies simulation accuracy in industrial robot applications. The displacement error index of 0.1 mm meets the motion accuracy requirements. The stress error index of 1 MPa ensures the reliability of strength evaluation. The grid size in high error areas is halved to improve local calculation accuracy. The maximum three encryptions balance accuracy improvement and calculation efficiency. Different application scenarios have different requirements for simulation verification. Industrial robots focus on motion accuracy verification, medical robots emphasize deformation prediction accuracy, and spacecraft focus on structural reliability verification. Error control standards need to match specific applications. The reliability of trajectory optimization results is ensured through virtual prototype verification. The fusion of simulation data and measured data reflects the unity of virtual and actual. Industrial robot motion control performs trajectory planning based on simulation prediction. Medical robot surgical operations adjust motion parameters based on real-time feedback. Spacecraft in-orbit work combines ground test data to evaluate structural status. The comprehensive analysis of multi-source data provides a comprehensive performance evaluation basis and realizes reliability verification of optimized trajectories.

[0154] Step 508: If the verification result is higher than the preset reliability, continue to adjust the switching trajectory to obtain a smooth switching trajectory that meets the preset module deformation and stress requirements, which provides a reference for the formulation of an actual switching plan, so that the building block module for flexible equipment can be quickly and smoothly switched.

[0155] In this embodiment, this step mainly includes: adjusting the switching trajectory according to the switching process, and switching the flexible device to be switched according to the adjusted switching trajectory.

[0156] Specifically, the reliability index is calculated according to the verification results of the virtual prototype, and the fuzzy evaluator based on the 95% confidence interval is used to calculate the evaluation parameters through the triangle membership function: μ(x)=max(0,1-|x-x0| / d), x0 is the ideal value, d is the allowable deviation, the deformation evaluation weight is 0.6, and the stress value evaluation weight is 0.4. The evaluation value is compared with the preset threshold of 0.85, and the mapping matrix between the parameter evaluation value and the adjustment priority is established to generate the parameter sequence to be adjusted. For the trajectory parameters to be adjusted, a random forest regressor with a tree depth of 8 and a number of trees of 100 is used. The data that meets the constraints in the most recent 50 sets of historical optimization data are selected for training. The regression features include three dimensions: curvature, velocity, and acceleration. The learning rate is set to 0.01. The parameter adjustment amount is predicted under the deformation constraint and stress constraint. The sensitivity of parameter update is verified by forward difference, and the coordinates of discrete points of the updated trajectory are calculated. According to the discrete points of the updated trajectory, a cubic Bezier curve: P(t)=ΣPiBi(t) is used for smooth fitting. The spacing of control points in the area where the curvature is greater than 0.01 per millimeter is reduced to one-fourth of the reference spacing. Trapezoidal speed planning is performed based on the maximum speed and acceleration constraints. A quintic polynomial is used to achieve smooth speed transition in the acceleration and deceleration sections, and the continuity characteristics of the speed planning curve are calculated. The curvature value at each point of the smoothed trajectory is calculated: κ=(x'y''-y'x'') / ((x'²+y'²)^(3 / 2)). The curvature change rate limit of adjacent points is 0.001 per millimeter per meter. The cubic Hermite interpolation is used to insert transition points in the area where the change rate exceeds the limit. The interpolation function keeps the first-order derivative of the curvature continuous. It is verified that the curvature derivative error is less than 0.0001 per millimeter per meter to achieve the overall smoothness of the trajectory curve. Reliability evaluation in industrial robots reflects the complexity of comprehensive judgment of multiple indicators. The 95% confidence interval reflects the statistical credibility of the evaluation results. The triangular membership function is symmetrically distributed around the ideal value of 0.8. The allowable deviation of 0.2 defines the fuzzy interval range. The weight of 0.6 for deformation emphasizes the position accuracy. The weight of 0.4 for stress value reflects the structural safety. The evaluation threshold of 0.85 ensures the reliability of the optimization results. Random forest regression realizes intelligent decision-making for parameter optimization in medical surgical robots. The tree depth of 8 provides sufficient feature stratification capabilities. 100 decision trees ensure the stability of the prediction results. Effective samples are selected from the most recent 50 sets of historical data for training. The curvature feature reflects the trajectory shape, the speed feature reflects the motion characteristics, the acceleration feature represents the dynamic response, and the learning rate of 0.01 ensures the stability of parameter adjustment.Bezier curve fitting achieves smooth transition of trajectory in aerospace flexible mechanisms. The curvature threshold of 0.01 per millimeter identifies the area of ​​drastic change of trajectory. The control point spacing is reduced from 4 mm to 1 mm to improve the local fitting accuracy. The trapezoidal velocity planning limits the maximum speed to 500 mm per second and the maximum acceleration to 1000 mm per square second. The fifth-order polynomial ensures the continuity of the second-order derivative of the velocity curve. Curvature continuity verification ensures the smoothness of movement in industrial applications. The curvature change rate limit of 0.001 per millimeter per meter avoids sudden changes in trajectory. Cubic Hermite interpolation inserts 5 control points in the transition area to maintain the continuity of the curvature derivative. The derivative error limit of 0.0001 per millimeter per meter ensures the overall smoothness. Different application scenarios have different requirements for trajectory optimization. Industrial robots focus on the balance between motion efficiency and accuracy, medical robots emphasize trajectory smoothness and accurate positioning, and spacecraft focus on structural deformation and stress control. The optimization goals and constraints need to match the specific applications. The reliability of trajectory planning is ensured through multi-level optimization and verification. The parameter optimization process embodies the unity of theory and practice. Evaluation indicators in industrial robot applications guide the optimization direction, historical experience in medical robots guides parameter adjustment, and dynamic constraints in spacecraft regulate trajectory morphology. The organic combination of multiple technical links realizes the intelligent trajectory planning and meets the actual needs of different application scenarios.

[0157] The present embodiment discloses a method for fast and smooth switching of building block modules for flexible equipment. Aiming at the problem that the building block modules of flexible equipment are prone to excessive deformation and stress exceeding the limit during the fast switching process, the module structural parameters and material properties are obtained, and a deformation statics model considering the module connection stiffness and damping is established. The module deformation and stress distribution under different switching trajectories are obtained by finite element analysis, and the speed and acceleration distribution during the switching process are calculated. If the module deformation exceeds the limit or the stress exceeds the limit, the switching trajectory is judged to be infeasible, and the trajectory optimization is performed. With the switching time as the target and the deformation and stress as the constraints, the trajectory parameters are coded and optimized to obtain a smooth switching trajectory that meets the requirements. According to the characteristics of different flexible devices, a trajectory optimization model and a generation rule base are established to achieve fast optimization of the switching trajectory. The trajectory reliability is verified by virtual prototype simulation, providing a reference for the formulation of actual switching schemes. The present invention effectively solves the deformation and stress control problems in the fast and smooth switching process of the building block modules of flexible equipment, and improves the switching efficiency and reliability.

[0158] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A modular module switching method for flexible equipment, characterized in that: include: The device parameters of the flexible device to be switched are obtained, and the device parameters are analyzed to obtain the device characteristic values; wherein, before obtaining the device parameters of the flexible device to be switched, the method includes: obtaining the material parameters of the building block module to construct a deformation statics model corresponding to the flexible device composed of the building block module and performing switching analysis on the deformation statics model to obtain the module deformation cloud map and module stress cloud map corresponding to each preset initial switching trajectory of the flexible device; extracting the discrete points of each initial switching trajectory in the module deformation cloud map and the discrete points of each initial switching trajectory in the module stress cloud map; The turning angle in the force cloud diagram is calculated, and the cutting speed and acceleration corresponding to each of the initial cutting trajectories at each of the discrete points are calculated to obtain the module deformation change curve and the module stress change curve of the flexible device in the cutting process corresponding to different initial cutting trajectories; multiple first cutting trajectories are selected from the multiple initial cutting trajectories according to the module deformation change curve and the module stress change curve; the first trajectory parameters in each of the first cutting trajectories are optimized according to the preset trajectory optimization algorithm, and the optimized multiple trajectory parameter combinations are obtained and saved in the preset trajectory library; Matching the trajectory optimization model corresponding to the device characteristic value, so as to call the cutting trajectory from the preset trajectory library through the trajectory optimization model; wherein, the trajectory optimization model corresponding to the current device characteristic value is obtained from a plurality of pre-trained trajectory optimization models corresponding to the device characteristic values ​​of the flexible device; each of the trajectory optimization models saves the mapping relationship between the device characteristic value and the corresponding cutting trajectory; obtaining the trajectory parameter features corresponding to each trajectory parameter combination in the trajectory library, and calculating the similarity between each trajectory parameter feature and the device characteristic value through the mapping relationship preset in the trajectory optimization model; when the similarity exceeds the preset threshold, calling the preset trajectory optimization rule to optimize the current trajectory parameter combination, and generating the cutting trajectory corresponding to the optimized trajectory parameter combination; Simulating the switching process corresponding to the flexible device to be switched according to the switching trajectory; The switching trajectory is adjusted according to the switching process, and the flexible device to be switched is switched according to the adjusted switching trajectory.

2. A modular module switching method for flexible equipment according to claim 1, characterized in that: The step of obtaining device parameters of the flexible device to be switched and performing feature analysis on the device parameters to obtain device feature values ​​includes: Obtaining the vibration mode parameters of the flexible device to be switched; wherein the vibration mode parameters include vibration mode parameters and deformation recovery parameters; Extracting modal parameters from the vibration shape parameters using singular value decomposition, and constructing a stiffness matrix and a mass matrix according to the modal parameters to obtain eigenvalue parameters; A characteristic vector is obtained according to the characteristic value parameter, and the characteristic vector is normalized to obtain the device characteristic value; wherein the characteristic vector includes curvature, length and rotation angle.

3. A modular module switching method for flexible equipment according to claim 1, characterized in that: The simulating the switching process corresponding to the flexible device to be switched according to the switching trajectory includes: Constructing a virtual prototype geometric model according to the discrete points of the cutting trajectory, and meshing the virtual prototype geometric model through twenty-node hexahedral units to obtain a meshed model; A strain energy density function is used to establish a hyperelastic constitutive equation, and deformation field data and stress field data of the meshing model are calculated by using the hyperelastic constitutive equation and an implicit dynamics solver; Performing feature extraction on the deformation field data and the stress field data through a five-layer deep autoencoder to obtain a feature mapping relationship between the deformation field data and the stress field data; Processing the data collected by the sensor according to the characteristic mapping relationship to obtain the displacement data and strain data of the flexible device to be switched during the switching process; The displacement data and the strain data are fused by a Kalman filter to obtain an actual deformation curve and an actual stress distribution curve; The reliability of the cutting trajectory is verified according to the actual deformation curve, the actual stress distribution curve, the deformation field data and the stress field data.

4. A modular module switching method for flexible equipment according to claim 3, characterized in that: The step of adjusting the switching trajectory according to the switching process, and switching the flexible device to be switched according to the adjusted switching trajectory, includes: When the reliability of the cutting trajectory is verified, a fuzzy evaluator based on a confidence interval is used to calculate the deformation evaluation parameter and the stress value evaluation parameter corresponding to the cutting trajectory, and a parameter sequence to be adjusted is generated according to the comparison results of the deformation evaluation parameter and the stress value evaluation parameter with the preset thresholds respectively; Training a random forest regressor according to the parameter sequence to be adjusted, predicting a parameter adjustment amount through the random forest regressor, and adjusting the trajectory parameters of the cutting trajectory according to the parameter adjustment amount to obtain the updated discrete point coordinates of the cutting trajectory; The updated discrete point coordinates of the switching trajectory are fitted using a Bezier curve, and trapezoidal speed planning is performed under the maximum speed and acceleration constraints according to the Bezier curve to obtain a speed planning curve; The curvature value at each discrete point of the speed planning curve is calculated, and a transition point is inserted in the curvature change rate exceeding limit region by Hermite interpolation to obtain an adjusted switching trajectory.

5. The modular module switching method for flexible equipment according to claim 1, characterized in that: The step of constructing a deformation statics model corresponding to the flexible device composed of the building block modules and performing a shearing analysis on the deformation statics model to obtain a module deformation cloud map and a module stress cloud map corresponding to each initial shearing trajectory of the flexible device includes: Obtaining the structural dimension parameters and material elastic modulus of the building block module to calculate the stiffness matrix corresponding to the building block module; wherein the structural dimension parameters include the module contact area and the number of bolts; the stiffness matrix is ​​determined by the material elastic modulus, the module contact area and the number of bolts; Obtaining the unit node stiffness value through the stiffness matrix and mapping it to the mesh node to establish a cubic spline curve interpolation equation; wherein the mesh node is divided using hexahedral units and the mesh size in the contact area is reduced to half of the basic mesh size; Calculating the curvature value according to the cubic spline curve interpolation equation to obtain the strain field calculation equation; wherein the curvature value is determined by the curve tangent vector and the normal vector; A nonlinear stress-strain equation group is established according to the strain field calculation equation to obtain the deformation field displacement distribution function and the stress field distribution function corresponding to the building block module; wherein the equation group is solved using the Newton iteration method until the residual is less than a preset value.

6. A modular module switching method for flexible equipment according to claim 1, characterized in that: The calculating of the switching speed and acceleration corresponding to each of the initial switching trajectories at each of the discrete points to obtain the module deformation change curve and the module stress change curve during the switching process of the flexible device under different initial switching trajectories includes: Obtaining a trajectory curve corresponding to each of the initial cutting trajectories, and obtaining a coordinate data set of discrete points on the trajectory curve; wherein different initial cutting trajectories correspond to different trajectory shapes, trajectory lengths, and trajectory curvatures; The speed distribution value is obtained by fitting the discrete points with a Bezier curve; wherein the speed distribution value is obtained by calculating the mapping relationship between the curve parameters and the switching speed; A kinematic equation is established according to the velocity distribution value, and the acceleration value of the discrete point is solved by using the Runge-Kutta integral formula; wherein the acceleration value is obtained by integrating the velocity with respect to time; A state equation is constructed according to the acceleration value of the discrete point and the deformation on the module deformation cloud map, and a deformation time function is obtained by solving the deformation time function through the variational principle; wherein the deformation time function includes an acceleration coupling term; A stress state equation is constructed for the deformation time function and the stress value in the module stress cloud map, and the stress time function is obtained by solving it through a fixed-step integrator; wherein the stress time function imposes fixed constraints and load constraints on the deformation boundary.

7. A modular module switching method for flexible equipment according to claim 1, characterized in that: The selecting a plurality of first switching trajectories from a plurality of the initial switching trajectories according to the module deformation change curve and the module stress change curve comprises: Constructing a constraint equation according to the preset maximum deformation of the flexible device and the current deformation of the module corresponding to each of the initial switching trajectories, so as to obtain deformation constraint optimization parameters through gradient descent iterative calculation; A multi-objective optimization function of a deformation objective function and a stress objective function is established according to the deformation constraint optimization parameters, and a mapping relationship between a trajectory parameter vector and a constraint response vector corresponding to each of the initial cutting trajectories is obtained through a support vector regressor of a Gaussian kernel function; According to the mapping relationship, a constrained optimization equation group is constructed using Lagrange multipliers, and the optimal trajectory parameters and Lagrange multipliers corresponding to each of the initial switching trajectories are obtained by solving them through an alternating iteration method; If the norm difference between the optimal trajectory parameter and the initial trajectory parameter corresponding to the initial cutting trajectory is less than a preset convergence threshold, storing the optimal trajectory parameter as the first trajectory parameter of the first cutting trajectory; If the norm difference is greater than a preset threshold, the number of iterations is increased based on the original number of iterations to continue optimizing until convergence.

8. A modular module switching method for flexible equipment according to claim 7, characterized in that: The step of optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm includes: Constructing a chromosome coding sequence corresponding to each of the first trajectory parameters; wherein the chromosome coding sequence adopts a binary coding method to obtain an initial trajectory parameter population; the first trajectory parameters include a trajectory shape parameter, a trajectory length parameter, a trajectory curvature parameter and a trajectory turning angle parameter; A gradient boosting tree regressor is established for the initial trajectory parameter population, so as to obtain a mapping relationship between the motion time and the first trajectory parameter through the gradient boosting tree regressor; wherein the mapping relationship is used to calculate the fitness value corresponding to the parameter value after chromosome decoding; A roulette wheel selector is used to select high-quality chromosomes within a preset interval, to perform a single-point crossover operation on the high-quality chromosomes to obtain offspring chromosomes, and to form a new trajectory parameter population according to the Gaussian distribution mutation of the offspring chromosomes; Calculating the fitness value of the new trajectory parameter population, and obtaining the optimal chromosome decoding parameter when the continuous change of the fitness value is less than a preset threshold; Each of the first trajectory parameters is optimized according to the optimal chromosome decoding parameters and the chromosome coding sequence.

9. A modular module switching method for flexible equipment according to claim 8, characterized in that: The step of optimizing the first trajectory parameters in each of the first switching trajectories according to a preset trajectory optimization algorithm further includes: The interpolation operation is performed using the cubic spline basis function according to the trajectory control points to obtain a piecewise cubic polynomial function expressing the trajectory curve; Performing Fourier series expansion on the piecewise cubic polynomial function, smoothing the velocity component through a Butterworth low-pass filter and a Wiener filter, and obtaining the acceleration component from the derivative of the velocity component; Calculate the curvature and stress value at each point according to the trajectory curve, and if the stress value is greater than the allowable stress value, add a trajectory control point at the point; wherein the spacing of the added trajectory control points is half of the original spacing; The continuity of the trajectory curve is verified. If the derivative continuity index exceeds a preset threshold, the spline function order is increased to a preset number, and the updated trajectory curve is obtained by solving the spline interpolation coefficient matrix.

10. A modular switching system for flexible equipment, characterized in that: It includes feature analysis module, trajectory matching module, switching simulation module and switching adjustment module; The characteristic analysis module is used to obtain the equipment parameters of the flexible equipment to be cut, and perform characteristic analysis on the equipment parameters to obtain the equipment characteristic values; wherein, before obtaining the equipment parameters of the flexible equipment to be cut, it includes: obtaining the material parameters of the building block module to construct a deformation statics model corresponding to the flexible equipment composed of the building block module and performing cutting analysis on the deformation statics model to obtain the module deformation cloud map and module stress cloud map corresponding to each preset initial cutting trajectory of the flexible equipment; respectively extracting the discrete points of each initial cutting trajectory in the module deformation cloud map and the The turning angle in the module stress cloud map is calculated, and the cutting speed and acceleration corresponding to each of the initial cutting trajectories at each of the discrete points are calculated to obtain the module deformation change curve and the module stress change curve of the flexible device in the cutting process corresponding to different initial cutting trajectories; multiple first cutting trajectories are selected from multiple initial cutting trajectories according to the module deformation change curve and the module stress change curve; the first trajectory parameters in each of the first cutting trajectories are optimized according to a preset trajectory optimization algorithm, and multiple optimized trajectory parameter combinations are obtained and saved in a preset trajectory library; The trajectory matching module is used to match the trajectory optimization model corresponding to the device characteristic value, so as to call the cutting trajectory from the preset trajectory library through the trajectory optimization model; wherein, the trajectory optimization model corresponding to the current device characteristic value is obtained from a plurality of pre-trained trajectory optimization models corresponding to the device characteristic values ​​of the flexible device; each of the trajectory optimization models saves the mapping relationship between the device characteristic value and the corresponding cutting trajectory; the trajectory parameter feature corresponding to each trajectory parameter combination in the trajectory library is obtained, and the similarity between each trajectory parameter feature and the device characteristic value is calculated through the mapping relationship preset in the trajectory optimization model; when the similarity exceeds the preset threshold, the preset trajectory optimization rule is called to optimize the current trajectory parameter combination, and the cutting trajectory corresponding to the optimized trajectory parameter combination is generated; The switching simulation module is used to simulate the switching process corresponding to the flexible device to be switched according to the switching trajectory; The switching adjustment module is used to adjust the switching trajectory according to the switching process, and switch the flexible device to be switched according to the adjusted switching trajectory.

Images