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

The invention provides a mobile robot formation control method based on leader-follow. The method is formed by a global positioning system, a wireless communication system, an algorithm processing system, and a speed control system. The global positioning system obtains the pose information of each robot and sends the pose information to an arithmetic processing system through a wireless communication system, and the formation motion control is finally realized through the information interaction with the speed control system. In a control algorithm, firstly a leader-follow formation motion model is established, a follow robot motion control rate is given, then a follow robot trajectory prediction model is established, a nonlinear least squares method prediction model is employed, a prediction model parameter is optimized by using an improved particle swarm algorithm, a communication data abnormal range is defined, and a prediction point is started to substitute an abnormal point so as to ensure formation motion. According to the method, the prediction model is introduced, the formation order deviation phenomenon caused by temporary communication abnormality, the reliability of follow robot motion is ensured, and the stability of the formation is greatly improved.

The invention relates to a photovoltaic array fault diagnosis and early warning method comprising the following steps: combining an Elman nerve network optimized by a non-linear least square method and a decision tree with experience knowledge so as to form a fault diagnosis model; collecting present photovoltaic array operation data and meteorology data, and computing errors when compared with historical normal state data; using the fault diagnosis model to obtain the corresponding fault type and credibility when the error is bigger than a threshold; finally integrally evaluating so as to obtain the final fault type credibility, and selectively carrying out fault early warning according to the credibility values; updating a fault knowledge base according to the field actual measurement conditions. The method combines the LM-Elman nerve network and the decision tree with experience knowledge so as to built the fault diagnosis model, thus improving the history data sensitivity, providing better prediction effect when compared with a BP network, and improving the network convergence speed and training precision; the experience knowledge is supplemented, thus providing stronger robustness; the method can timely detect and diagnose, thus reducing fault incidence rate, and ensuring the photovoltaic power station to stably work.

A method is disclosed for estimating clutch engagement characteristics of a friction clutch system in a vehicle powertrain. A dynamic model of the system is used under conditions that cause clutch slipping. Algebraic equations defining a functional relationship between clutch torque and an engagement angle have characteristic parameters that are estimated using a non-linear least squares technique. A non-linear least squares technique iteratively minimizes the difference between a measured output clutch disk speed and an output clutch disk speed from the system dynamic model for the same inputs until a small insignificant error is reached. Parameter estimates are used to compile an estimated clutch engagement characteristic.

A method for determining chromatograph peak form parameter and overlapped peak area includes setting time value and chromatographic signal value at point for chromatographic elution curve and peak number, using n as total number for point, using Gauss model as objective function to fit said elution curve to be determined, utilizing nonlinear least square method to solve out optimum parameter and using the corresponding relation of fitting elution curve to elution curve to be determined to work out peak form parameter and overlapped peak area of chromatograph .

The present invention provides a chemometric equation to estimate fluid density, viscosity, dielectric constant and resistivity for a formation fluid sample downhole. The chemometric estimates can be used directly as estimated values for fluid density, viscosity, dielectric constant and resistivity for a formation fluid sample downhole. The chemometric estimates can also be plugged into a Levenberg-Marquardt (LM) non-linear least squares fit, as an initial estimate of the parameter to be estimated by the LM fit. If the initial parameter estimate is too far from the actual parameter values, the LM algorithm may take a long time to converge or even fail to converge entirely. The present invention estimates an initial value of a parameter that provides a high probability that the LM algorithm will converge to a global minimum.

The invention provides a method for assessing a technology for a large numerical control machine, comprising the following steps: obtaining frequency response function curve of the large numerical control machine through structure composition dynamic stiffness experiments; employing nonlinear least square method to identify dynamical parameter through curve least square fitting; establishing a non-linear dynamical model of machining process of the numerical control machine by identifying machining technical parameters of cutting force dynamic model in a cutting force experiment; then employing the method of random sampling to simulate tooling motion trail and calculating the number of failpoints to work out probability of technological reliability.With the method of the invention, technological reliability of the numerical control machine can be accurately assessed according to the existing structural characteristics, processing technic characteristics and processing conditions of the large numerical control machine, so that accuracy and high efficiency of technological reliability assessment can be improved and a new method is provided for on-line reliability assessment of the large numerical control machine.

The invention provides a system calibration method for a measuring head of a coordinate measuring machine based on a master ball. The system calibration method comprises the steps of a) putting the standard ball with known radius R on a working platform of the coordinate measuring machine; b) installing the measuring head at a movable end of the coordinate measuring machine, controlling the coordinate measuring machine to move the measuring head according to a proper movement trajectory and completing the point scanning of the master ball on the working platform within a measurement range, wherein the point scanning approximately and uniformly covers the surface of the entire master ball to obtain N groups of data, N is larger than or equal to 200 and each group of data comprises the measuring point numerical value of the measuring head, the three-coordinate position reading of the coordinate measuring machine and the rotating angle of a gyrator; c) on the basis of the operation in the previous step, selecting partial data of the N groups of data, performing coordinate conversion, bringing the data into a spherical constraint equation and solving an over-determined nonlinear equation set by using an improved nonlinear least square method to obtain parameters to be calibrated and complete the system calibration.

The invention discloses a line shape control method for short line method segment prefabrication construction and relates to the field of control of bridge construction. The method comprises the following steps of: calculating a theoretical prefabrication line shape according to the line shape of bridge design and the construction sequence by a tangent displacement method; building the global coordinate system of the prefabrication line shape and the local coordinate systems of all prefabrication segments; converting segment control points between the local coordinate systems and the global coordinate system by a matrix calculation method; processing errors by a non-linear least square method, and correcting coordinates of segment nodes in the global coordinate system; and adjusting the line shapes of subsequent segments by a direct adjustment method or a segmentation adjustment method according to the errors. According to the method, the errors are processed by the non-linear least square method, so the method is high in calculation speed; the influence of position errors of matched segments, measurement errors of a control point coordinate, and template errors on the prefabrication line shape is comprehensively considered, the control accuracy of the line shape is improve obviously, and the popularization of a short line method segment construction method in China is accelerated.

The invention provides a measuring method for the relative pose of a non-cooperative spatial circular object. The method comprises the following steps: providing left and right cameras to shoot left and right images formed by the spatial circular object, wherein images of the spatial circular object comprise a left ellipse and a right ellipse; carrying out non-linear least square fitting to detect ellipses or approximate ellipses and using the detected ellipses or approximate ellipses as candidate ellipses; recognizing the left ellipse and the right ellipse from the candidate ellipses according to epipolar constraint relationship between the left ellipse and the right ellipse; respectively determining a left projection equation of a spatial circle and a right projection equation of the spatial circle according to an equation coefficient of the left ellipse and the equation coefficient of the right ellipse generated by fitting the left ellipse and the right ellipse so as to obtain two sets of pose solutions of the spatial circle; and converting one of the two sets of pose solutions of the spatial circle to the spatial coordinate system of the other set according to the positional and attitude relationship between the left and right cameras and determining the optimal pose solutions of the spatial circular object according to the principle that an included angle is the smallest. The measuring method provided by the invention can overcome the problem of ambiguity in the process of pose resolution.

The invention discloses a positioning method based on a non-linear fitting micro-seismic source or acoustic emission source. The method is characterized in that: a plurality of acoustic emission sensors are installed near an object to be measured, according to a coordinate figure of a known sensor position and P, S wave arriving time, and according to a distance operation formula and P, S wave speed and time relations, a position coordinate of a micro-seismic source or acoustic emission source is obtained through a non-linear least square regression method like an acceleration simplex method or a Marquardt method. According to the positioning method based on a non-linear fitting micro-seismic source or acoustic emission source, measuring a speed or setting a given speed in advance is not needed, influence to positioning caused by acoustic velocity measurement can be avoided, and positioning precision is raised. The method is convenient and practical in practical engineering application compared with a traditional method.

A temperature monitoring technique for collecting radiation intensity (blackbody emission) across a broad wavelength range. A solid state spectrometer (26) acquires spectra from a sample (10) in real time and resolves the spectra to a radiation intensity (I) versus wavelength (λ) curve. This curve is fitted to Planck's equation using a non-linear least squares fitting analysis. The system can be configured to self-calibrate and then lock in an amplitude value (A) which is used in subsequent temperature measurements by fitting to the blackbody emission curve. Preferably, the spectrometer (26) is flat field corrected (36) in an initial step to counteract variations in the spectrometer response with wavelength.

The invention discloses a phase noise measuring method based on a phase noise mathematical model. The phase noise measuring method comprises the following steps: a power spectrum model of an oscillator signal is used for measuring data through a signal power spectrum, and a nonlinear least square method is adopted for curve fitting, so as to obtain a parameter in the power spectrum model; then, according to a relation between the signal power spectrum and a phase noise power law model, the obtained parameter is substituted into the phase noise power law model, so as to obtain a phase noise measuring result of the tested signal. Compared with the existing phase noise measuring method, the phase noise measuring method based on the phase noise mathematical model has the main advantage of avoiding the influence of a phase noise extraction circuit of hardware on the measuring performance.

The invention relates to the field of optical detection and discloses a measuring and calculating method for geometrical parameters of a large-caliber off-axis aspheric optical element. The problem that the geometrical parameters of an existing large-caliber off-axis aspheric optical element are difficult to accurately measure is solved. An adjustment compensator, an interferometer and the aspheric element to be tested are coaxial. A laser tracker is used for respectively measuring coordinates of the outer surface of the compensator and coordinates of the aspheric mirror to be detected. Coordinate data measured on the cylindrical surface of the compensator are used for fitting the optical axis direction of the aspheric element. According to the fitted optical axis direction, coordinate measuring points on the aspheric element to be tested under a measuring coordinate system are rotated so that an optical axis corresponding to the measuring points after rotating can be in parallel with a Z axis of the measuring coordinate system. Non-linear least square fitting is performed on measuring point coordinates of pyramid reflection mirror sphere centers at the contact positions of the aspheric mirror after rotating according to an equation for a surface, so that the geometrical parameters of the aspheric optical element to be tested are obtained. The method is simple in measuring process and suitable for measuring and calculating the geometrical parameters of the large-caliber off-axis aspheric element.

The invention provides a phase model based ship line spectrum noise source position identification method. On an existing single-hydrophone ship-radiated noise measuring platform, instantaneous phase and time-varying amplitude of a Doppler signal are obtained by utilizing a higher-order ambiguous function method according to Doppler effect generated by relative movement between a ship and a measuring hydrophone during measurement, then parameters of the Doppler signal are estimated by using a weighted nonlinear least square method, and a reference sound source is combined, so that positioning of a ship low-frequency line spectrum noise source is achieved. The method is simple to implement, and measuring accuracy of the method is higher than that by a characteristic method.

The invention discloses a monocular camera and inertial measurement unit relative mounting angle calibrating method and belongs to the field of sensor calibration. The method comprises the following steps of, firstly, rotating a measurement module to collect calibrating data, wherein the measurement module is composed of a monocular camera and an inertial measurement unit; secondly, extracting relative attitude information from images collected by the monocular camera; thirdly, solving a mounting matrix through a nonlinear least square method. With dependence on absolute reference information,the monocular camera and inertial measurement unit relative mounting angle calibrating method can avoid calibrating errors resulting from absolute reference information measurement errors; besides, during the calibrating process, the data collecting period is short and the calibrating precision is high, so that the monocular camera and inertial measurement unit relative mounting angle calibratingmethod can meet the requirements of combined configuration of the monocular camera and the inertia measurement unit on spatial alignment during practical application.

The invention provides a two-dimensional wind-induced vibration coefficient calculation method for an ultra-standard oversized cooling tower. The method comprises the steps of a, building a geometric model, wherein the full-size geometric model of the cooling tower and a computational domain of a flow field where the cooling tower is located are built based on the actual size of the cooling tower; b, mesh generation, wherein the computational domain is divided into grids by means of the local cipher technique; c, numerical simulation computation, wherein a turbulence model is selected, boundary conditions are set, and numerical wind tunnel simulation analysis is conducted to obtain tower tube outer surface wind pressure time interval; d, solving wind-induced response, wherein a finite element model of the cooling tower is built, and wind-induced dynamic response of the cooling tower is analyzed; e, computing a wind-induced vibration coefficient, wherein the tower tube two-dimensional wind-induced vibration coefficient is computed based on structural response time interval; f, analysis fitting, wherein fitting of the tower tube wind-induced vibration coefficient is conducted based on a nonlinear least square principle, and finally a two-dimensional wind-induced vibration coefficient fitting formula is provided.

A method of and a system for stabilizing High Voltage Power Supply (HVPS) DC and AC voltages in multi-energy X-ray computed tomography scanners are provided. The method comprises generating filter ratios, computing DC and AC voltages, and feeding back the computed DC and AC voltages to the commanded voltages. The filtered ratios including an air ratio and a copper ratio are modeled as nonlinear functions of the DC and AC voltages. Computing DC and AC voltages include computing an m-ratio and an n-ratio. The parameters of the nonlinear model comprise an exponent parameter and a set of polynomial coefficients. The parameters are determined by a calibration procedure, which performs scanning at different combination of DC and AC voltages. The optimal parameters are obtained through a nonlinear least square minimization, which is solved through a brute force search over the exponent parameter and a closed form solution of the polynomial coefficients. Feeding back the computed DC and AC voltages include comparing the computed voltages with commanded voltages, integrating the difference between the computed voltages and commanded voltages, and adding the integrated voltage differences to the commanded voltages.

The invention provides a pose calibration method of a motion mechanism for a track capturing wind tunnel test. According to the method, the six-rod parallel connected motion mechanism is used in the track capturing wind tunnel test, relations between a model mass-center coordinate as well as a model attitude angle and a motor code disk value are established, and thus, unknown parameters in the six-rod parallel connected motion mechanism in the track capturing wind tunnel test are obtained. A mathematical model structure of the six-rod parallel connected motion mechanism of the track capturing wind tunnel test is determined by theory derivation, the center coordinate of a hook joint of a static platform along the whole guide rail is measured and calculated, the relation between the center coordinate of the hook joint of the static platform and the motor code disk value is fit, the pose of a calibration point is measured and a nonlinear least square method is utilized to obtain the unknown parameters in the parameterized model, and pose calibration of the motion mechanism in the wind tunnel test is realized. Thus, the method can be used for pose calibration of the six-rod parallel connected motion mechanism of the track capturing wind tunnel test, and the pose positioning accuracy is high.

The invention discloses a numerical control rotary table geometric error measurement method based on a four-base-station laser tracking system, and the method comprises the steps: building a self-calibration coordinate system, and calibrating the position of a tracking interferometer; respectively placing the target mirrors at three non-coplanar points on a rotary table, keeping a certain distancefrom the rotary table, controlling the rotary table to rotate according to a certain angle interval, and solving by using a nonlinear least square method to obtain coordinates of a measurement pointunder the self-calibration coordinate system; after the measurement is finished, fitting to obtain a circle center of a space circle formed by the three points and a radius fitting function; establishing a coordinate system of the rotary table; subtracting the coordinates of the theoretical point from the coordinates of the actual measurement point under the coordinate system of the rotary table to obtain the spatial position errors of the measurement point, and obtaining the six geometric errors of the rotary table by utilizing the spatial position errors of the three points at the same position through establishing a six-position-related geometric error model of the rotary table. The method has the advantages of short detection time, high detection precision and low detection uncertainty, and can be used for rotary table error detection of a precision numerical control machine tool.

The invention relates to a method for calculating electric field probe rotation offset through a circular polarization antenna axial ratio directional diagram. According to the method, the nonlinearity least square method is used for guiding out a mathematic model of the relation between the offset after electric field probe rotation and the circular polarization antenna axial ratio directional diagram, the Gauss-Newton iterative method is used for carrying out numerical solution on the offset after electric field probe rotation to obtain electric field probe rotation offset delta x and delta y for compensation for near-far field conversion, and therefore an accurate circular polarization antenna far field directional diagram and the accurate axial ratio directional diagram are obtained. According to the method, the accurate value of the delta x and the delta y of electric field probe rotation offset can be obtained without arranging heavy optical measurement equipment in an erecting mode, only two times of plane near field measurement is needed, and the workload is reduced. An electric field probe used for measurement just needs to be rotated by 90 degrees instead of being rotated by 360 degrees, and electric field probe rotation offset can be automatically compensated.

The invention relates to a method for arranging displacement measuring points, comprising the following steps: establishing a finite element model and calculating theoretical values of the positions of displacement measuring points presenting a bridge state; fitting the theoretical values by adopting the polynomial of nonlinear least squares to obtain a regression equation; carrying out F detection and t detection on the regression equation, and selecting a regression curve with the best fitting as a theoretical value curve; abstracting a plurality of points from measured values of the displacement measuring points to form different groups, and then also fitting the points of each group to obtain a plurality of measured value curves; defining the closeness degree between the theoretical value curves and the measured value curves as Euclidean distance and Chebishov distance; calculating the Euclidean distance and the Chebishov distance of each measured value curve and theoretical valuecurve, and setting the maximum Chebishov distance; and under the condition of according with the deviation, analyzing and comparing the Euclidean distance values of each measured value curve and theoretical value curve so as to determine the best displacement measuring point arranging position.

A method of locating and characterizing defects on semiconductor using a scanner device and a high-magnification imaging device comprises the steps of scanning (A) a test wafer a plurality of times with the scanner device, recording the scanner device coordinates of defects and the markers in the standard patterns, analyzing the coordinates to identify the standard patterns and; loading and aligning (B) the test wafer in both the average predicted coordinates and the actual coordinates for each of the located patterns, and then averaging over the multiple sets of actual coordinates; then using a non-linear least-squares program to calculate a set of alignment transformation parameters that converts the average predicted coordinates as nearly as possible to the actual coordinates.

The invention discloses a brillouin spectrum denoising method based on wavelet transform. The brillouin spectrum denoising method comprises the following steps of: S1, selecting a wavelet function with orthogonality, symmetry, compact support, urgent attenuation and smoothness; carrying out N-layer wavelet decomposition on a brillouin spectrum; S2, selecting a threshold value for a high-frequency coefficient at each layer; S3, carrying out threshold value quantization on the high-frequency coefficient at each layer; S4, carrying out lorentz curve fitting on a low-frequency coefficient at the N-th layer by adopting a nonlinear least square method; and S5, carrying out wavelet inverse transformation on the high-frequency coefficient at each layer, which is subjected to threshold value quantization and the low-frequency coefficient at the N-th layer, which is fitted. When the signal to noise ratio of the brillouin spectrum is small, the brillouin spectrum denoising method provided by the invention can obtain high-precision brillouin spectrum through less wavelet decomposition layers, and can effectively reduce the calculation quantity while effectively removing the noise.

The invention discloses a method for calibrating an accelerometer under an unsupported state, which is applied to an inertial navigation system with the capacity of initially aligning with a static base attitude. The method comprises the following steps of: firstly, simplifying an error model, repeatedly changing the carrier position, and coarsely and finely aligning with each position by using the static base to acquire the attitude information; and calculating the three-axis component of an acceleration of gravity g projected under a carrier coordinate system according to the attitude information, and recording the measured value of the accelerometer at each position. The invention provides a 10-position calibrating scheme, which comprises the following steps of: acquiring zero offset to degenerate the error model into a linear model by identifying the zero offset through a nonlinear least square method according to the calculated value and measured value of the projected component; and mounting an error and scale factor coupled matrix by identifying through a linear least square method. Therefore, the key parameters of the accelerometer are calibrated. The method has the advantages of capacities of realizing unsupported calibration of the accelerometer and effectively estimating the key parameters of the accelerometer with time-varying performance, along with simple process and high speed.

The invention belongs to the technical field of vehicle parameter measurement, and particularly relates to a recognition method for rotational inertia around z axis and tire cornering stiffness of an automobile. The method comprises the following steps: (1) operating steering wheel angle impulse test according to an automobile operation stability testing method to acquire data, wherein the data comprises a steering wheel angle, yaw velocity, lateral acceleration and a vehicle speed signal; (2) recognizing a transfer function of a vehicle system equivalent two-degree-of-freedom vehicle model through a least square method by using the data acquired by the step (1); (3) recognizing the yaw rotational inertia around the z axis and the tire cornering stiffness of the automobile included in the transfer function obtained by the step (2) through a nonlinear least square method. According to the method, the use of other large-sized experimental instruments is avoided, and the research cost is effectively reduced; the data used by the method is convenient to measure so as to finally facilitate modeling; the speed is high, quick development of research work is facilitated, and the method has higher practicability.

A method is disclosed for computing clutch engagement characteristics of a friction clutch system in a vehicle powertrain. A dynamic model of the system is used under conditions that cause clutch slipping. Algebraic equations defining a functional relationship between clutch torque and an engagement angle have characteristic parameters that are estimated using a non-linear least squares technique. The non-linear least squares technique repetitively minimizes the difference between a measured clutch output speed and a clutch output speed from the system dynamic model for the same inputs until a small insignificant error is reached. Parameter estimates are used to update each engagement characteristic.

The invention discloses a comprehensive subway energy consumption forecasting method based on a time sequence. The comprehensive subway energy consumption forecasting method based on the time sequence directly uses sample data of a subway energy management system to build a total subway energy consumption forecasting model based on the time sequence. The comprehensive subway energy consumption forecasting method based on the time sequence gives sufficient consideration to the subway energy consumption periodicity, builds the forecasting model according to the energy consumption forecasting time, and combines a long autoregression model method with a nonlinear least square method to estimate parameters, wherein the long autoregression model method is used for the primary parameter estimation, and the nonlinear least square method is used for the precise parameter estimation. The comprehensive subway energy consumption forecasting method based on the time sequence uses the data recorded in the subway energy management system to directly build the subway energy consumption forecasting model based on the time sequence, does not need to spend energy to research the energy consumption structure of the subway system and avoids the influence to the subway energy consumption forecasting precision due to incomplete consideration to energy consumption influence factors.