9 results about "Non-linear least squares" patented technology

This invention discloses a method for determining the structural error during kinematic calibration simulation of a parallel mechanism, applicable to the calibration of industrial parallel robots. The method involves establishing structural constraint equations for the parallel mechanism to determine independent variables in the structural parameters; combining the structural constraint equations and independent variable values ​​to calculate the values ​​of non-independent variables in the structural parameters; calculating the motion of each prismatic joint under ideal structural parameter conditions with a given target pose, establishing the correlation between pose parameters and motion; given an initial structural error value, calculating the structural error value conforming to geometric constraints using constraint equations and nonlinear least squares method, and calculating the norm of the residuals. If the residuals meet the requirements, the structural error for kinematic calibration simulation of the parallel mechanism can be generated; if the residual value does not meet the requirements, the generated structural error is used as input, and a new structural error value is generated again using constraint equations and nonlinear least squares method, and the residual calculation is repeated until the residual value meets the requirements, thus generating a structural error that meets the requirements and accurately verifying the structural error identification accuracy of the error model.

This invention relates to the field of highway maintenance technology, specifically to a highway asphalt pavement performance prediction and maintenance decision-making system and method. It includes the following steps: S1, acquiring pavement images and identifying disease types using a deep learning segmentation network; S2, describing the time-varying evolution of each disease index using a Lotka-Volterra type differential equation system; S3, estimating the parameter set using historical detection data and maintenance records for corresponding time periods through nonlinear least squares inference; S4, given the current disease quantification vector, numerically integrating the differential equation system to predict the change trajectory of each disease index in future time periods; S5, identifying promoting diseases based on the influence coefficient matrix, simulating the dynamic impact of different maintenance measures on the entire disease using the differential equation system, and selecting the optimal maintenance timing and measure combination based on maximizing the comprehensive maintenance benefit function.

This invention specifically relates to a single-station passive localization method based on an improved genetic-particle swarm optimization algorithm, comprising: measuring the phase difference using an interferometer; obtaining an estimate of the phase difference change rate using a Kalman filter algorithm; further obtaining a coarse estimate of the target radiation source position using the improved genetic-particle swarm optimization algorithm based on the estimated phase difference change rate; and using the coarse estimate of the target radiation source position as prior information, performing iterative calculations using a nonlinear least squares method to obtain a fine estimate of the target radiation source position. This method does not require phase difference de-ambiguity resolution or azimuth angle prior information, and can be implemented using only a single-baseline interferometer formed by two array elements. It has advantages such as no need to measure angles, no need for phase difference de-ambiguity resolution, fast convergence speed, and high localization performance.

This invention relates to the field of UAV flight control and power transmission line operation technology, and discloses a UAV hoisting control method for high-voltage power lines. This method utilizes a full-tensor magnetic field gradient measurement array, an airborne inertial measurement unit (IMU), and a suspension-point IMU to calculate the UAV's position and yaw state relative to the high-voltage conductor using an analytical model of the high-voltage line magnetic field and nonlinear least squares method. Furthermore, it combines error state extended Kalman filtering, dual IMU yaw angle calculation, a disturbance observer, and dynamic electromagnetic anomaly fuse protection to achieve anti-sway control during hoisting. This invention is beneficial for improving the reliability of state perception, control stability, and safety of UAV hoisting operations in environments with strong near-field electromagnetic interference from high-voltage power lines.

This invention discloses a TDOA positioning method, server, and system based on clock and frequency offset compensation. The method includes: acquiring local TOEs recorded when a base station periodically transmits a broadcast signal, and local TOAs recorded by other base stations when they receive corresponding broadcast signals; dynamically and continuously estimating the frequency and clock offsets between base stations based on the acquired TOEs and TOAs; acquiring the local TOAs recorded by each base station when receiving signals transmitted by the tag to be located; combining the TOAs of each base station to form an initial TDOA; compensating the initial TDOA based on the estimated frequency and clock offsets between base stations to obtain a correct TDOA; and using a nonlinear least squares method to achieve TDOA positioning based on the correct TDOA. This invention can improve the positioning accuracy of TDOA and has broad application value in the field of passive target positioning.

A non-equilibrium mass transfer model 79 This invention relates to a method for predicting the migration of selenium (Se) in granite fractures and its application, falling within the field of radionuclide migration prediction technology. By coupling the convection-dispersion equation with a two-point kinetic adsorption model and introducing key parameters such as the retardation factor, dispersion coefficient, and kinetic rate constant, this invention achieves the prediction of selenium migration in granite fractures. 79 A precise description of the nonequilibrium migration process of Se under complex hydrochemical conditions is presented. The model incorporates nonlinear least squares method for parameter inversion, significantly improving the fitting accuracy and mechanism analysis capability for the "breakthrough-plateau-tail" characteristics in the breakthrough curve. Validation demonstrates high stability and reliability under various hydrochemical conditions and effective extrapolation to 10-10. 3 -10 5 Predicting long-term migration behavior on an annual timescale provides insights for deep geological disposal environments of high-level radioactive waste. 79 Se's long-term migration risk assessment provides an efficient and interpretable analytical tool.

The application discloses a high-resistance fault detection method, device and equipment based on arc current change trend, wherein the detection method first collects a feeder zero sequence current signal by using a signal detection device; secondly, the zero sequence current signal is processed by using a variational mode decomposition to obtain a component after filtering high-frequency modes; thirdly, the current instantaneous amplitude of the filtered component is extracted by using a windowed fast Fourier transform; then, the change trend of the arc current is quantitatively estimated by using a nonlinear least square method; finally, the high-resistance fault is detected by using the sign of the coefficient in the nonlinear function. The application uses the nonlinear least square method to quantitatively estimate the overall change trend of the arc current. Meanwhile, the irregularity and randomness of the arc current can be weakened. This stage can improve the robustness of the proposed HIF detection method.

The application discloses a kind of based on thermal inertia model ultra-stable crystal oscillator dynamic temperature compensation method, comprising: through second pulse signal trigger sampling logic, real-time collection ultra-stable crystal oscillator's ambient temperature and frequency deviation;Third-order static polynomial model including multiple static temperature coefficients to be estimated is constructed;The output of third-order static polynomial model is used as excitation, and first-order thermal inertia model including thermal time constant to be estimated is constructed;With the error minimization between measured historical frequency deviation and predicted frequency deviation as target, third-order static polynomial model and first-order thermal inertia model are combined, and optimal parameter vector is estimated by nonlinear least square method, and the optimal parameter vector estimated and real-time ambient temperature are used to calculate real-time dynamic predicted frequency deviation, and the predicted frequency deviation is subtracted from real-time frequency deviation, and the frequency signal after compensation is output.The application can accurately compensate the frequency drift of crystal oscillator under dynamic temperature change, and improve frequency stability.

A parameter identification method for fractional-order time-delay systems based on the Gegenbauer wavelet operation matrix is ​​presented, addressing the problem that existing parameter identification methods cannot identify fractional-order systems with time delays. This method constructs a Gegenbauer wavelet operation matrix for both fractional-order integral operators and time-delay operators. Then, the original fractional-order time-delay differential equation is transformed into a system of algebraic equations, thus avoiding direct numerical approximation of fractional-order differential terms. Furthermore, a nonlinear least squares method is used to jointly estimate the fractional-order order, time-delay parameters, and system parameters. Finally, the optimal fractional-order order and model parameters are determined based on minimizing the error between the model output and the actual output. This method effectively reduces solution complexity and improves the numerical stability and accuracy of parameter identification. It provides an effective tool for high-precision modeling and analysis of fractional-order time-delay systems.