31 results about "Error function" patented technology

This invention provides a method for calculating the minimum delay time between two seismic traces: Calculating a first error value between the two seismic traces under at least three first delay times (either preset or obtained after the last iteration update); fitting an error function between the delay time and the error value; calculating a second delay time and a second error value at the extreme point of the error function; if the first difference between the second delay time and a threshold delay time of any of the first delay times is greater than or equal to a first threshold, or the second difference between the second error value and the first error value is greater than or equal to a second threshold, then the second delay time is used to replace one of the edge delay times in the current first delay time. All the above steps are repeated until the first difference is less than the first threshold and the second difference is less than the second threshold, and this second delay time is determined as the minimum delay time between the two seismic traces. This invention improves the calculation speed for calculating the minimum delay time between two seismic traces.

The application belongs to the technical field of load identification of structural health monitoring, and provides a concentrated load positioning method based on principal stress constraint and strain gradient trajectory identification, which comprises the following steps: arranging strain sensors in a monitoring area of a planar structure plate and collecting strain data of measuring points, constructing a strain field inversion function based on a radial basis function and establishing an error optimization function containing a fitting error term and a smooth constraint, introducing a principal stress direction constraint and a gradient projection smooth constraint to form a comprehensive error function and solving to obtain an optimal strain field distribution, then calculating strain gradient vectors of each grid node and tracking a gradient trajectory through a gradient descent method, and finally performing an iterative clustering analysis based on distance statistics on a trajectory end point to output a cluster center as a concentrated load application position identification result. The application can improve the concentrated load positioning accuracy, enhance the anti-noise capability, and realize fast, stable and large-sample-free load position identification.

The application discloses a spacecraft magnetic moment attitude control method based on a non-inverse iteration algorithm, and the method establishes a linear periodic model for a spacecraft attitude control system which only relies on a magnetic moment; on the basis, according to a linear quadratic optimal control theory, an optimal controller design problem is converted into a solution problem of a discrete periodic Riccati equation associated with the optimal controller design problem; on the basis of a gradient descent principle, an error function is constructed, and a non-inverse iteration algorithm based on the gradient descent principle is proposed to solve the discrete periodic Riccati equation, so that a matrix inversion operation in a traditional solution method is avoided, and the calculation efficiency and the numerical stability are effectively improved; the obtained Riccati equation solution is used to obtain an optimal state feedback control law, so that the spacecraft attitude optimal control only using the magnetic moment is realized, and a new solution scheme is provided for an attitude control problem of a limited actuator in a space task.

The application provides a large model quantization method based on double-scale scaling and a text information processing method, comprising: splitting a weight matrix to be quantized into two low-bit matrices, introducing column scaling vectors for the two low-bit matrices respectively, and establishing a quantization parameter system containing row scaling factors, column scaling factors and biases; taking an error function as an optimization objective, jointly optimizing the quantization values of the low-bit matrices, the column scaling factors, the row scaling factors and the biases through an iterative algorithm until the error converges to a minimum; and absorbing and fusing the optimized column scaling factors into associated upper projection matrix weights to generate a deployable quantization model. The application avoids additional calculation and storage overhead, thereby improving the weight quantization precision without gradient training.

The present application relates to the technical field of mechanical arm control, and particularly relates to a mechanical arm trajectory drawing method and system based on an AZTND model; the method comprises the following steps: acquiring mechanical arm curve drawing task parameters and an AZTND solver; establishing a mechanical arm time-varying kinematics equation, introducing joint angular velocity constraints and acceleration constraints, and converting the curve drawing problem into a time-varying nonlinear equation; defining an error function as the residual of the time-varying nonlinear equation, constructing an AZTND dynamics equation, and determining the parameters of a segmented adaptive coefficient function according to the curve drawing stage; using an ODE numerical solution method to solve the AZTND dynamics equation in real time, obtaining a mechanical arm joint angular velocity control instruction, and driving a mechanical arm end effector to track an expected curve trajectory. The present application solves the problems of slow convergence speed, weak anti-interference ability and low geometric precision in traditional control in time-varying trajectory tracking, and realizes high-precision real-time drawing of complex artistic curves.

This invention provides a control method and system for a five-degree-of-freedom marine crane in a non-inertial frame, relating to the field of crane automatic control technology. The method includes: acquiring the real-time state variables of the crane in a non-inertial frame; inputting the state variables into an adaptive radial basis function neural network to obtain a lumped disturbance estimate, the neural network being designed based on a five-degree-of-freedom dynamic model established in the ship's coordinate system using the Lagrange equations and the principle of virtual work, considering the ship's six-degree-of-freedom motion and installation position offset; calculating a tracking error function based on the state variables and the desired trajectory; inputting the disturbance estimate and the error function into a feedback controller to calculate and generate a control torque, which is then output to the actuator to drive the crane to track the desired trajectory and suppress load sway. This invention achieves direct modeling and control in a non-inertial frame, solving the coordinate transformation disturbance and singularity problems caused by inertial frame modeling, and improving trajectory tracking accuracy and disturbance rejection robustness.

The application discloses a kind of wide dynamic range signal measurement error calibration method and system, the present application includes on harmonic signal measurement experimental platform for specified harmonic different amplitude, different temperature is tested multiple times and the amplitude data of harmonic signal measurement device output is collected;Nonlinear parametrization function about amplitude correction parameter vector L is constructed and fitting error function;Combining the amplitude data collected and nonlinear parametrization function and fitting error function, amplitude correction parameter vector L is iteratively solved based on gradient descent method and substituted into nonlinear parametrization function to obtain temperature correction function, whether continue to iteratively solve amplitude correction parameter vector L according to the correction result of temperature correction function is judged.The present application aims to solve the error problem influenced by temperature change in power system harmonic signal measurement, and improve the accuracy and reliability of harmonic signal measurement in power system by effective error correction.

The application discloses a kind of super long spliced guide rail straightness prediction methods considering multi-source error, belong to precision machinery manufacturing and measurement technical field. Including: the mathematical model of guide rail surface initial manufacturing error is established;Subsection independent integration method is applied, and the deflection deformation analytical model of guide rail caused by gravity load is established;By parameterization modeling and finite element simulation are combined, the parameterization simulation model of splicing surface geometric error and bolt pretightening force coupling effect is established, and splicing error function is extracted;Based on linear superposition principle, the above three independent error models are algebraically superimposed, and the guide rail comprehensive straightness prediction model is constructed.The strong nonlinear coupling effect of splicing interface is accurately quantified by parameterization means, forming a set of high-precision forward-looking prediction method that can be used in design or installation stage, which provides an effective theoretical tool for the precision design and process control of super long spliced guide rail.

This application provides a method, apparatus, and electronic device for processing feature data of a neural network model. The method includes: acquiring feature data; performing fixed-point processing on the feature data to obtain first fixed-point data; scaling the first fixed-point data to obtain second fixed-point data; and performing approximate calculation on the second fixed-point data using a preset error function to obtain approximate result data; and integrating the approximate result data and the first fixed-point data to obtain nonlinear result data. This solution reduces hardware implementation complexity while improving the computational efficiency of model feature data, making it suitable for lightweight deployment in real-time scenarios.

This invention relates to a predetermined-time fuzzy backstepping control method for a fractional-order decoupled dual-mass MEMS gyroscope with delay constraints and actuator faults, belonging to the field of MEMS gyroscope control technology. The method includes: S2: establishing a fractional-order motion equation for the decoupled dual-mass MEMS gyroscope with actuator faults based on the Lagrange equations to eliminate the influence of linear acceleration and vibration at the same frequency; S3: constructing a β-cut 2 type fuzzy logic system to address the uncertainties and nonlinear terms of the system model; superimposing a delay error function on a specified time function to ensure that the error does not violate the constraints after a finite time; using a fractional-order hyperbolic tangent tracking differentiator in the controller design to handle direct fractional derivatives and repeated derivatives; and finally, combining the β-cut 2 type fuzzy logic system to establish a control law with an adaptive law for fault parameters.

The application relates to the technical field of angle measurement, and discloses a double-encoder mutual calibration method based on a shearing principle, two encoders are coaxially and rigidly connected, one reference reading head and two movable reading heads are installed on the encoder 1, one reference reading head is installed on the encoder 2, two sets of shearing difference functions of the positioning error of one of the encoders are constructed by reading the double-encoder measurement values and performing difference operation, the error function is obtained by shearing reconstruction and weighted fusion on the function; the error function of the other encoder is obtained by substituting the error function into the double-encoder difference value, and the mutual calibration between the double encoders is realized; the application solves the limitations of the traditional encoder calibration steps being complicated and the experimental requirements being high.

This invention discloses a data analysis-based method for predicting soil fluoride content, belonging to the field of data prediction. The method includes: S1, acquiring environmental parameters of soil samples and recording the soil fluoride content under the current environmental parameters; S2, randomly shuffling the environmental parameter data, preprocessing the shuffled data, and dividing it into training and test sets; S3, constructing a first soil fluoride content prediction model using a backpropagation (BP) neural network; S4, optimizing the number of hidden layers and learning rate factor of the first prediction model using an improved grey goose optimization algorithm; S5, reconstructing the prediction model using the optimal number of hidden layers and learning rate factor, and finally training it on the entire training set to obtain a second prediction model; S6, inputting the test set into the second prediction model, iterating the training process, calculating the prediction effect using the mean squared error function, until the target prediction error for soil fluoride content is reached, and outputting the soil fluoride content prediction result.

The application provides a deep learning model segmentation quantization method and system for multimodal feature distribution, the method comprising: collecting a feature tensor to be quantized in a quantization parameter, and determining a feature quantization range; constructing a quantization error function based on the feature quantization range, minimizing the quantization error function based on a golden section search algorithm, and determining an optimal breakpoint; dividing the feature region according to the optimal breakpoint and setting a quantization step, and constructing a segmented quantization operator; implementing segmented symmetric linear quantization according to the feature quantization range, determining the region to which the floating point value belongs through the segmented quantization operator, and performing differential mapping based on the region; and performing multiplication and addition operation on the quantized integer and the quantized weight to obtain an integer result and output. The application proposes a segmented quantization strategy based on region division, aiming to realize accurate quantization of multimodal feature distribution through an adaptive quantization step allocation mechanism.

The application provides an online self-adaptive calibration method and system for low power factor error of an electric energy meter, comprising the following steps: measuring initial phase error at multiple preset low power factor calibration points to establish a data set of power factor and error; constructing an initial error function based on the data set; calculating a second derivative sign change point of the function as a boundary point to divide multiple calibration subintervals; adopting a weighted sum model of Legendre polynomial and hyperbolic tangent function to fit a phase error compensation formula in each subinterval; calculating a compensation value according to a real-time power factor position and a corresponding subinterval compensation formula during actual operation of the electric energy meter; and if the power factor is near the boundary point, then smoothing and fusing adjacent interval results through a Gaussian weighting function to realize online correction of the phase error.

The application provides a data processing method and device based on contour constraint, a medium and computer equipment. The data processing method based on contour constraint comprises the following steps: acquiring texture image data of an object with a mark point patch pasted thereon, wherein the texture image data comprises a plurality of continuous image frames, and each image frame comprises a mark point pattern of the mark point patch; extracting contour three-dimensional information of the mark point pattern; constructing a contour error function model of the mark point pattern, and optimizing the contour three-dimensional information and pose information through the contour error function model; and performing three-dimensional point cloud splicing through the optimized pose information. The existing point cloud data can be automatically corrected by defining the error function model, the optimization precision of the three-dimensional point cloud splicing is improved, and the robustness of the three-dimensional point cloud splicing processing is more stable and robust.

The application discloses an A0 mode arrival time difference deep learning-based damage acoustic emission positioning method, which comprises the following steps: 1. collecting acoustic emission signals of a metal plate in a working state through acoustic emission sensors on four vertices of a rectangular region to be measured; 2. performing time-frequency transformation on the collected signals by using continuous wavelet transformation to extract wavelet coefficients at an A0 mode characteristic frequency; 3. calculating the maximum distance of wave propagation according to the geometric structure of the region to be measured, using the maximum distance to inversely calculate the maximum arrival time difference between a signal initial arrival mode and the A0 mode, and then using the maximum time difference to select a wavelet coefficient characteristic section; 4. inputting wavelet coefficients of sensor groups into a deep learning model after training to obtain the time difference of acoustic emission signals arriving at the sensor groups; and 5. substituting the A0 arrival time difference output by the deep learning model in step 4 into an error function, using a genetic algorithm to optimize the function, and evaluating the position of the damage. The application can accurately identify the position of the damage of the metal plate.

This invention relates to the field of image recognition technology, and more particularly to a method and system for identifying scalp anomalies under near-field illumination. The method first utilizes the Navier-Stokes equations and dispersion-free physical constraints to repair the texture diffusion of specular highlights generated by near-field illumination, restoring details of hair follicle openings. Then, based on camera intrinsic parameters, images from various viewpoints are projected onto a virtual skull spherical model. Thin-plate spline interpolation is used to compensate for skin elastic deformation, and an adaptive weight matrix based on an error function is introduced to perform multi-band fusion, constructing a high-precision global panoramic image. On this basis, a semantic mask generated by a convolutional neural network, combined with a distance transformation algorithm and pre-calibrated pixel conversion factors, is used to extract quantitative indicators such as hair follicle density and hair shaft physical diameter. Finally, feature point registration is used to achieve spatial alignment of temporal data, and a correlation mapping between the anomaly evolution vector and the hair follicle physiological cycle is established.

The present application relates to the technical field of computer-aided processing, in particular to a kind of probability characterization method of ceramic matrix composite component mechanical property parameter, comprising: establishing target finite element model;Data set is constructed;Neural network model is constructed, and training is carried out;Error function is constructed;Likelihood function is constructed, and probability distribution model is obtained.The present application can provide accurate parameter basis for the uncertainty quantification, performance dispersion analysis and performance prediction of ceramic matrix composite through the analysis and probability modeling of posterior distribution.

A device for federated machine learning, comprising: a communication system; one or more memories configured to store: values of model parameters of a machine learning (ML) model, and phase values; and one or more processors are configured to: apply the ML model, using the values of the model parameters, to model input data to determine model output data; determine an error function based on the model output data and expected output data; calculate a gradient of the error function; generate a plurality of frequency-domain values based on the gradient; generate modified frequency-domain values based on the phase values and the frequency-domain values; generate a time-domain digital signal based on the modified frequency-domain values; wherein the communication system is configured to transmit an analog radio frequency (RF) signal based on the time-domain digital signal.

This invention relates to the field of underwater robotic arm control technology, specifically to a method and system for trajectory planning and control of an underwater robotic arm based on the PNVRGZND model. The method includes: obtaining the joint angle vectors, desired velocity vectors, and Jacobian matrix of a six-DOF cascaded underwater robotic arm; transforming the kinematic equations into time-varying linear equations and defining an error function; constructing a PNVRGZND model containing a novel double-integral structure, novel time-varying parameters that are monotonically increasing and bounded, and a novel nonlinear activation function containing exponential, linear, and power terms; substituting the derivative of the error function into the PNVRGZND model to obtain the dynamic evolution equation of the joint angular velocity vector; determining known quantities based on the real-time state of the robotic arm and the desired trajectory; numerically solving the dynamic evolution equation to obtain the joint angular velocity vector and outputting it to the joint actuator. This invention significantly improves the trajectory tracking accuracy and robustness of underwater robotic arms in environments with strong interference.

This application belongs to the field of grating measurement and provides an error compensation method for grating interferometer measurement. The method includes: constructing an angle error function based on the angle error measured by the grating interferometer; constructing a relative displacement error function based on the relative displacement error measured by the grating interferometer; constructing an explicit compensation model based on the angle error function and the relative displacement error function; constructing an implicit compensation model based on the dataset formed by the angle error and the relative displacement error; obtaining the explicit accuracy mean based on the explicit compensation model and the implicit accuracy mean based on the implicit compensation model; when the implicit accuracy mean is not less than the explicit accuracy mean, the explicit compensation model is used to compensate for the error in the grating interferometer measurement. This method ensures the reliability of the selected model by analyzing the consistency between the internal mapping of the model and physical laws, thereby achieving high-precision, high-robustness, and highly interpretable displacement measurement compensation.

A kind of iron concentration prediction method based on water supply pipeline wall layered structure and automatic calibration, comprising: integrated pipeline hydraulic model, iron concentration release model and microscopic electron microscope layered observation result, construct pipe network iron concentration layered multi-species reaction model, simulate the migration and transformation of iron ions between corrosion base layer, loose porous layer, hard shell layer and surface layer;Multi-objective comprehensive error function is constructed, and the difference between simulation result and observation data is evaluated;Three-stage differential evolution calibration is executed: the model is used as water quality simulation engine, the comprehensive error function is used as optimization target, and the model parameters are optimized by iteration method;First stage global search locates advantage area, second stage local optimization, third stage fine tuning, and after each stage ends, the subsequent parameter search boundary is dynamically contracted based on optimal solution and boundary tightening coefficient;Iron concentration spatiotemporal distribution prediction is output.The method of the application realizes high-precision prediction of complex pipe network iron concentration and efficient automatic calibration of model parameters.

The invention relates to a method for determining obstacles in an image (B) of the environment (U) of an ego-vehicle (2) taken by a camera, the procedure is carried out by a neural network and includes the following measures: a) Dividing the image into different image areas (B1, B2, B3), where each image area is assigned to one of the three categories “passable”, “impassable” or “unknown”, b) Predicting a lower boundary of an impassable obstacle located lowest in the image by minimizing an error function (L1), where: - if the predicted lower limit (YPRB) is located above an actual lower limit (YTRB) of the lowest image area (B2) of the category "impractible" present in the image (B), this is taken into account in the error function (L1) according to measure c1), - if the predicted lower limit (YPRB) is located below an actual upper limit (YFT) of the lowest image area (B1) of category (B1) ‘passable’ in the image (B), this is taken into account in the error function (L1) according to measure c2), - if the predicted lower limit (YPRB) is located in an image area (B3) of the category “unknown”, this is not taken into account in the error function (L1) according to measure c3).

This invention provides a data processing method and system based on the ratio difference of a current transformer, belonging to the field of current transformer technology. The method includes: synchronously acquiring the primary and secondary current signals of the transformer, obtaining a current waveform dataset after aperiodic component subtraction and wavelet denoising; calibrating the dual electromagnetic circuits of the magnetic valve-type transformer separately, completing the error data through bicubic interpolation and fitting an error function model; extracting the amplitude and phase angles of the fundamental and harmonic waves through ChirpZ transform, and obtaining the frequency domain error by combining FFT transform with the error model; calculating the initial ratio difference and compensating for harmonic flux linkage, inversely transforming the frequency domain error to the time domain, and weighted synthesizing to obtain the final ratio difference dataset, thereby improving the ratio difference measurement accuracy and adapting to the structural characteristics of the dual electromagnetic circuits.

A kind of self-adapting feature matching heterogeneous face point cloud registration method, first, the source face point cloud and target face point cloud are preprocessed, the initial pose deviation of both is quickly reduced through rough registration;Then the ISS key points of the point cloud after rough registration are extracted, the normal vector and curvature value of each key point are calculated, and the registration primitive covering multidimensional geometric features is constructed;Then, combined with the inherent geometric law of face, an adaptive weight distribution mechanism is designed, by real-time analysis of the curvature difference of key points, normal angle and point cloud data characteristics, the weight proportion of Euclidean distance, normal consistency and curvature similarity is dynamically adjusted;Then, based on the adaptive weight, a weighted error function is constructed, and the optimization process of weighted point-to-plane ICP is iteratively executed;When the iteration termination condition is reached, the final transformation matrix is output, and the high-precision fine registration of heterogeneous face point cloud is completed.The present application realizes accurate and robust alignment of heterogeneous face point cloud.

The application provides a temperature prediction method, device, medium and product for quickly measuring body temperature, wherein the method comprises the following steps: acquiring temperature collection data, the temperature collection data comprising collection temperatures corresponding to different sampling time points; performing fitting operation on a temperature rise curve model according to a set fitting method, the temperature rise curve model being a double exponential model; the fitting operation comprising: constructing an error function according to the difference between the collection temperature corresponding to the sampling time point and a predicted temperature obtained based on the temperature rise curve model, and determining an optimal model parameter combination when the error function takes the minimum value; and substituting the optimal model parameter combination and a target time point into the temperature rise curve model to obtain a target temperature. The application overcomes the limitation of existing human body temperature measurement schemes that rely on fixed correction factors, and improves the accuracy of temperature prediction and the anti-noise interference capability.

The application discloses a high-dimensional spherical harmonic calibration method and system for improving robot absolute positioning accuracy, and belongs to the technical field of robots. The method is suitable for calibration of a robot of a rotary joint type or a robot with periodic errors, can greatly reduce periodic errors of a mechanical arm caused by deformation, axial center deviation and the like, and can complete calibration through simple linear equation solving, and is high in calculation efficiency. Specific steps include the following steps that S1, end pose residual errors and joint angles of the mechanical arm are acquired through calibration; S2, a residual error function is constructed; S3, the residual error function is fitted according to data quantity and actual requirements; and S4, error compensation is carried out based on a fitting result of the residual error function. The method calibrates the robot of the rotary joint type or the robot with the periodic errors through the high-dimensional spherical harmonic model, converts the periodic errors into simple linear equation solving, is easier to calculate, and thus greatly improves calibration accuracy and calibration efficiency.

In various example embodiments, model calibration techniques are used to identify one or more external structural loads on a structure. Initially, material or geometry attributes of elements of a model are adjusted to minimize discrepancy between measured responses and modeled responses. Thereafter, the model is used to search for external structural loads that minimize discrepancy between the measured responses and the modeled responses. Discrepancy may be measured by an error function that looks to dynamic responses and / or static responses. A solution that minimizes discrepancy may be determined using a genetic algorithm that generates a set of proposed external structural loads, applies the set to the model to produce the modeled responses, computes an error function that measures the difference between the measured responses and the modeled responses, and evolves the solutions to minimize an error produced by the error function, with such operation proceeding until a stopping criteria is reached.

The invention provides an early-arrival wave equation travel time inversion method for optimizing an error function. The early-arrival wave equation travel time inversion method for optimizing the error function comprises the following steps: step 1, inputting field observation early-arrival wave seismic data, seismic wavelets and an initial speed model; step 2, carrying out numerical solution on the wave equation to obtain a seismic source wave field and simulated seismic data; 3, calculating the travel time difference between the observed seismic data and the simulated seismic data, and constructing an accompanying seismic source; 4, the wave equation is subjected to numerical solution, and an accompanying wave field is obtained; step 5, calculating a gradient and obtaining a speed update quantity; step 6, constructing an optimized error function according to the travel time difference and the seismic data; and step 7, when the inversion result converges, outputting the inversion result. According to the early arrival wave equation travel time inversion method for optimizing the error function, the defect that wave equation inversion is prone to falling into a local minimum value is overcome, and inversion convergence and speed inversion precision are improved.