117 results about "Model order" patented technology

A model creation tool is provided that automatically evaluates potential econometric models and, given a set of historical data on which the model is intended to operate and other constraints, automatically identifies an optimum model. The optimum model may be an autoregressive model. The optimum model is defined by a set of optimum model parameters and optimum model coefficients. The optimum model parameters may include a model order such as an autoregressive model order, a seasonality function parameter, a trend function parameter, and a parameter that sets a tolerance that is used when performing principal components dimension reduction operations on the historical data.

A communications system includes a transmitter that transmits a modulated signal having encoded communications data over a communications channel. The transmitter adjusts one of at least modulation and coding at the transmitter based on the received channel state information of the transmitted signal. A receiver determines received signal metrics from the modulated signal. A noise power estimator estimates the noise power of the received communications signal by collecting N data samples from the communications signals, forming a covariance matrix of the N data samples based on a model order estimate, computing the eigenvalue decomposition of the covariance matrix and ranking resultant eigenvalues from the minimum to the maximum for determining the noise power of the received signal. At least one of modulation and coding are adjusted based on the monitored link quality of the communications channel to enhance use of the available channel capacity.

SAR images are improved by a method for acquiring a synthetic aperture image from a sequence of periodic pulse returns where the sequence of periodic pulse returns is interspersed with interrupts, i.e. missing pulses. The interrupts mark the start and end of one or more segments, where the segments contain the periodic pulse returns form the SAR image. The method comprises the steps of:converting said pulse returns into a digital stream;performing an azimuth deskew on said digital stream to obtain a deskewed digital stream;forming a forward-backward data matrix from the deskewed digital stream for one or more segments;forming an average segment covariance from the forward-backward data matrix;computing a model order for the average segment covariance;computing one or more linear prediction coefficients using data contained in the forward backward data matrix, and model order;using the linear prediction coefficients to compute missing pulse returns belonging within the interrupts.The computation for extrapolating the missing pulse returns is introduced after the Stolt interpolator in RMA processing. In computing the model order, eigenvalues are found and compared to a threshold. Roots of a linear prediction polynomial are computed, then stabilized to obtain stabilized roots. Linear prediction coefficients are reconstituted using the stabilized roots. Sub-bands are used to decrease computing time for the missing pulse returns.

Allocating limited manufacturing resources to achieve a feasible production plan that is consistent with customer demand is a difficult and common problem faced in many manufacturing industries. For large-scale multi-stage manufacturing systems, existing methods are typically based either on allocating limited resources sequentially, according to a priority ranked list of production starts, or linear programming based models. The output of such planning models is a production plan which specifies the quantity of each part to produce at each plant, using resources available to the enterprise. Typically there are lot-sizing rule defining the permissible production start quantities. The invention disclosed herein is a method for applying these rules to the lot-sizing of production starts within a linear program. It employs advanced heuristics that consider both established operational objectives (e.g. customer service, short lead times, low inventory, and prioritized allocation of supply and capacity) and lot-size rules to efficiently compute a feasible production plan for the division.

The invention discloses a method for estimating city expressway traffic states based on mobile detection of smartphones, wherein a city expressway cell transmission model is built firstly, a observing network is built by adopting a smartphone to rapidly detect parameters, then a state space model based on a lighthill-whitham-richards (LWR) traffic flow model is designed, a traffic state and a boundary flux are synchronously estimated by utilizing three-step type recursive filters algorithm, then are coalesced with upstream and downstream subsection boundary flux by adopting a weighted average algorithm, and traffic parameter estimation is upgraded, thereby achieving real-time distributed estimation of city expressway network traffic state. The method for estimating the city expressway traffic states based on the mobile detection of the smartphones can collect average speed information of vehicles on any time and space positions of a loop, enables traffic estimation not to be restrained by the position of a detector, can achieve synchronization estimation of traffic density and boundary flux by designing a state-space model and a three-step recursive filter, and achieves the problem of large range expressway network traffic estimation by being coalesced with subsection boundary flux, reduces model order, and improves efficiency of algorithm.

The invention relates to a modal parameter identification method based on response signal time-frequency joint distribution characteristics. According to the modal parameter identification method based on the response signal time-frequency joint distribution characteristics, signal analysis and structural modal parameter identification are carried out directly through a structural vibration response. The modal parameter identification method based on the response signal time-frequency joint distribution characteristics comprises the steps of firstly carrying out complex wavelet continuous transformation on a structural response signal, obtaining energy distribution characteristics of various wavelet transformation domains (a real domain, a virtual domain, a modal domain and a phase domain), obtaining a time average wavelet energy spectrum through a wavelet transformation coefficient, therefore carrying out quantification on selection of model orders and the scale corresponding to each order modality, on the basis, obtaining the optimum scale required by parameter identification, achieving pre-identification of modal frequency through the corresponding relation of the scale and the frequency, finally extracting a wavelet transformation coefficient slice at the specific scale, carrying out linear fitting through an amplitude value and a phase component, and achieving structural identification of inherent frequency and a damping ratio. As simulation and experiment results show, even if an external incentive function is not included, accurate identification of structural modal parameters can be achieved through the modal parameter identification method based on the response signal time-frequency joint distribution characteristics.

An apparatus for filtering malicious multimedia data using sequential processing and a method thereof are provided. The apparatus includes: a maliciousness classification model training unit extracting a predetermined feature from at least one or more types of moving pictures and then, through machine training, generating a maliciousness determination model for each of at least one or more classes; a malicious data classification unit sequentially inputting input moving pictures for which maliciousness is required to be determined, to the maliciousness determination model, and determining the maliciousness class of the input moving pictures, based on a probability that data at a determination time of the input moving pictures belongs to a predetermined maliciousness class, and an accumulated maliciousness probability to a current time; and a malicious information filtering unit cutting off service if the maliciousness class belongs to a predetermined reference maliciousness class.

An array has Nel array elements for receiving a plurality of communication signals from different sources. A noise power estimator at each array element receives a signal and estimates the noise power by collecting Ns data samples, forming a covariance matrix of the Ns data samples based on a model order estimate, computing the eigenvalue decomposition of the covariance matrix and ranking eigenvalues from the minimum to the maximum for the determining the noise power of the received signal at the respective array element. A pencil-based blind source separation (BSS) processor receives the signals and noise estimates from each array element and noise power estimator and removes the noise.

The invention relates to a three-dimensional mechanical in-process model sequential modeling method based on removal feature recognition, which mainly aims to solve the technical problems of repetitive work and big application limitation in the prior art. In the method, first of all, the removal features are classified into boundary feature, inset feature, communication feature, thread feature and loop feature. On the basis of the classification, the method includes the steps of: 1) producing a workblank; 2) judging whether a next process is necessary, and if yes, entering the third step; 3) recognizing the removal features in the process, which includes selecting a feature type and recognizing the feature; and 4) modeling the in-process models. All the in-process models are assembled to form a whole process model and are organized by a standard structure of a process all-information model. The method makes full use of the geometrical characteristics of the original components, solves the problem of high quality requirement to the original component model design in three dimensional process design, speedily creates the in-process models according to the practical mechanical processing sequence and improves the overall efficiency of the process design system.

The invention discloses an output-only linear time-varying structure modal parameter identification method and belongs to the technical field of structural dynamics. Firstly, a cost function of a least squares support vector machine vector time-varying autoregressive model is deduced; secondly, a function space is built by means of a Wendland compactly supported radial basis function; a regular factor is determined through the non-parameter method based on Gamma testing, and a basis function width reduction coefficient is given on the basis of actual experiences; a time-varying autoregressive model order is determined according to the Bayesian information criterion and the Akaike information criterion; a function space order is determined according to the ratio of residual sum of squares to sequence sum of squares; finally, the matrix expression of the least squares support vector machine vector time-varying autoregressive model is solved according to the cost function, modal frequency of a system is solved according to a time freezing method, and linear time-varying structure modal parameter identification is finished. The method can improve calculation efficiency, improves system robustness, and is widely used in linear time-varying structure modal identification in structural dynamic engineering application.

The invention discloses a signal noise reducing method for modal parameter identification. The signal noise reducing method for modal parameter identification comprises the steps that 1, a Hankel matrix is built through a pulse response signal of a noise-containing structure; 2, a rank of the Hankel matrix is resolved, an order determination index is resolved according to the rank of the Hankel matrix, and a model order is determined through the order determination index; 3, the Hankel matrix is processed through the order determination index and structure low rank approximation to obtain a rebuilt matrix processed through low rank approximation; 4, the step 2 and the step 3 are repeatedly performed until the convergent standard is met, and therefore a noise reducing signal is obtained; 5, modal parameter identification is performed through the noise reducing signal. The signal noise reducing method for modal parameter identification has the advantages that the fact that a Frobenius norm of a difference between the Hankel matrix before being processed through noise reducing and the Hankel matrix after being processed through noise reducing approaches to be the minimum can be achieved by setting the mode of the convergent standard and structure low rank approximation, that is, the improvement of the precision of the noise reducing signal can be achieved.

The invention discloses an iteration-adaptive-algorithm-based method for detecting a coherent MIMO radar target. First, a relation model among a first echo signal, a transmitting signal reflection coefficients and noise interference of the target is confirmed, wherein the relation model is observed on the n observation point, secondly, the relation model is linearized, further the reflection coefficients of the target are initialized through a delayed superposition DAS algorithm, and finally, a bayesian model order selecting tool is used for optimizing the reflection coefficients of the target. An iteration adaptive algorithm can effectively solve existing encountered problems of moving target parameter detection and greatly improves accuracy of detecting of the coherent MIMO radar moving target.

The invention relates to the field of identification of electric power system low-frequency oscillation modes, and in particular relates to an electric power system low-frequency oscillation mode identification method based on Matrix Pencil. Aiming at solving the problem that measured data obtained by ac actual system is usually influenced by site environment and other factors and thereby is signal data with a certain signal to noise ratio, the method proposes to utilize ESPRIT to improve the Matrix Pencil algorithm, decompose a signal space into a signal sub-space and a noise sub-space by directly adopting a data matrix composed of measured data as a basis, accurately estimate a model order, and detect oscillation frequencies, decay factors, oscillation amplitudes, phases and other information of the low-frequency oscillation signals of an electric power system in different oscillation modes, and thereby computational efficiency and low-frequency oscillation identification capability can be effectively improved. The method is applicable to the electrical power systems and other relevant departments for identifying low-frequency oscillation modes of the electrical power systems.

A method of equalizing a signal received over transmission channel defined by BEM coefficients of a basis expansion model of its channel taps, comprising the step of approximately solving the relation (I) for x[n] by an iterative method, n being the index of time, y[n] being the received signal, x[n] being the equalized signal, Bm[n] being the mth basis function of the basis expansion model, M being the model order of the basis expansion model, and blm, being the BEM coefficient of the mth of the basis function of the / th channel tap, and w[n] being optional noise.y[n]=∑m=0MBm[n]·(∑l=0L-1blmx[n-l])+w[n](I)

The invention discloses a resting state complex fMRI data ICA-CNN classification framework of patients and healthy people, and belongs to the field of biomedical signal processing. According to the method, an interested functional network separated from resting state complex fMRI data by ICA is taken as a research object, and classification of patients and healthy people is realized by utilizing 2D CNN learning characteristics with few parameters; data augmentation is carried out by utilizing ICA results obtained under multiple groups of model orders, and the problem of fMRI data shortage is solved. Compared with an existing 3D CNN network, the training amount is reduced, and the accuracy is improved. For example, aiming at a plurality of fMRI data collected in an 82 tested resting state,by applying the DMN component extracted by ICA, the slice identification accuracy is higher than that of 3DCNN by (0.728 vs 0.701), and the tested identification accuracy obtained by a tested decisionis further improved by (0.914 vs 0.701).

The invention relates to a synchronous phasor self-adaptation calculation method based on verification. The synchronous phasor self-adaptation calculation method based on verification comprises the following steps that firstly, initialization is carried out, wherein the wave sampling number of each cycle, a signal model order and a time interval of two data windows in a steady state algorithm and a dynamic state algorithm are determined; secondly, a model for electric power signals in the steady state algorithm is established; thirdly, a synchronous phasor of the electric power signals is solved through a DFT result of the two data windows; fourthly, whether a calculation result is correct or not is verified by comparing a back-stepping value and a practically-measured value, the process is ended if verification succeeds, and the fifth step is carried out if the process fails; fifthly, the synchronous phasor is calculated with a dynamic algorithm; sixthly, verification is carried out again, a result of the dynamic algorithm serves as a finial result if verification succeeds, and the result of the steady algorithm serves as a final result if verification fails. Thus, the calculation modes are switched in a self-adaptation mode; according to the synchronous phasor self-adaptation calculation method based on verification, the calculation precision of a steady situation and the dynamic performance of a transient state situation are considered, and the requirements for accuracy and rapidity are considered. The synchronous phasor self-adaptation calculation method based on verification can be widely applied to synchronous phasor calculation of an electric system.

The invention relates to a high resolution seismic wavelet extracting method based on high-order statistics and an ARMA (autoregressive moving average) model, which belongs to the field of seismic signal processing. The high resolution seismic wavelet extracting method provided by the invention is characterized in that: under the precondition of performing ARMA parameter simplified modeling on seismic wavelets, SVD (singular value decomposition) based on autocorrelation function is adopted to determine an order of AR part, an MA order determining method is provided for integrating information content criterion function in a high-order cumulant MA order determining method, and the MA order determining accuracy rate in a seismic wavelet ARMA model is improved; an SV-TLS (singular value decomposition - total least squares estimation) and a cumulant method are respectively adopted to estimate wavelet parameters; and under the precondition of ensuring wavelet precision, the model order is decreased as far as possible to improve the operation efficiency and to finally realize seismic wavelet extraction in high efficiency and high precision. Through the data simulation verification and the practical seismic data processing demonstration, the method provided by the invention is proved to effectively improve the estimated precision and extracting efficiency for the seismic wavelets and to have obvious effect even under short-time seismic data and strong noise pollution.

The invention discloses a dynamic brain functional magnetic resonance imaging method and system. The method includes the steps that high-time low-space-resolution navigation data S<NAV> and high-space low-time-resolution dynamic image data S are collected in (k,t) space; the navigation data S<NAV> are decomposed through singular values to obtain a model order L and a time primary function phi[l](t), and Tikhonov regularization constraint is conducted by combining with the image data S to solve a space primary function cl(k); interpolation recovery with the high-time high-space coverage density is conducted on the (k,t) space through the model order L, the time primary function phi[l](t) and the space primary function cl(k), Fourier inversion is then conducted, and therefore a high-time-resolution and high-space-resolution dynamic brain function magnetic resonance image is obtained. By means of the method and the system, quality of the reconstructed image is improved, and meanwhile the high-time-resolution and high-space-resolution dynamic brain function magnetic resonance image is obtained.

The invention discloses a speaker verification system creation method based on a model sequence adaptive technique. The method comprises the steps that voices of a first speaker to an nth speaker are entered sequentially, wherein j takes from 1 to n, and the following steps are executed: when the voice of the jth speaker is entered, a universal background model and an attenuation speaker model of the jth speaker are created, and the attenuation speaker models of all the created speakers are updated; after the voice of the nth speaker is entered, the universal background model of the nth speaker and the attenuation speaker models of the first speaker to the nth speaker are finally updated; and a speaker model is obtained to form a speaker verification system. The method is based on the initial universal background model, utilizes channel information and speaker sufficiency statistical information which are included in the sequentially entered voices of the n speakers, and sequentially adapts to the universal background model and the speaker model of the corresponding speaker, so that the speaker verification system used for speaker identity verification is created.

The present invention belongs to the field of microelectromechanical MEMS and electronic technology, and is especially one kind of order reducing method for system model with parameters. The model order reducing technology can raise simulating and verifying speed of system model effectively and improve the design scheme of circuit and device timely. The present invention establishes model order reducing method to parameter system. Transfer function of parameter system is multiple series developed to constitute son projection matrix, and the son projection matrix is orthogonalized to constitute projection matrix. The projection matrix is used in reducing order of the original parameter system and the obtained reduced-order system is independent on some special values of the system parameters and can maintain order reducing precision on different values. The present invention has greatly raised order reducing precision and efficiency.

The invention discloses a multivariable time-delay system identification method based on a step test, and provides the identification method in a frequency domain mainly for a transfer function model having a pure lag step. The method is suitable for most chemical processes ( such as a pressure reduction heating furnace). The multivariable time-delay system identification method based on the step test comprises a system test unit used for stimulating an identified object to obtain input / output data for identifying; a data processing unit for carrying out noise reduction processing and the like on the data obtained through system test to enable the obtained data to be less influenced by noise, and calculating frequency domain response of the system according to the data obtained through processing; and a parameter estimation unit for carrying out linear processing on the pure lag step, then, estimating parameters through a least square method, and finally, optimizing each model parameter through iteration. The advantages of the method are that the method not only can be applied to an ordinary open-loop system, but also can be directly applied to closed-loop identification without increasing model order; identification precision is high; and the method can provide a high-precision model for a controller design engineer to design a controller.

The invention relates to the field of time sequence analysis, and discloses an air conditioner load prediction method and device, and the method comprises the steps: collecting a plurality of samplesto form a training set, and carrying out the processing of the samples in the training set, and obtaining a data sequence; Training an ARMA (p, q) model by adopting the data sequence; According to theARMA model order p, load values in the data sequence are sequentially used as ARMA model input, and an estimated value of the load value at the next moment corresponding to the input load value is output; Training the SVR model according to the load value in the data sequence and the difference value between the load value and the corresponding estimated value; According to the ARMA model order p, taking the load value in the current time period as the input of the ARMA model, and outputting an initial predicted value of a to-be-predicted moment adjacent to the current time period; And takingthe load value in the current time period as the input of the SVR model, outputting an initial predicted value correction value, and taking the difference value between the initial predicted value and the correction value as the actual predicted value of the to-be-predicted moment.

An abstraction model supporting multiple hierarchical levels is inputted into a generalized static timing analysis of a hierarchical IC chip design to analyze and optimize the design of circuits integral to the chip containing a plurality of macro abstracts. An electrical network, synthesized for an internal abstract interconnect segment, is performed only once per macro and is applied to multiple instances of the macro abstract model in the IC chip design. The synthesized electrical network is a resistive capacitive or a resistive inductive capacitive network or a combination thereof. The synthesized electrical network is then used to match impulse response transfer functions of the network and the abstract interconnect segment's timing model. This network is stitched with the electrical parasitics of external interconnect segments connected to macro primary outputs. Various model order reductions are then performed on the electrical parasitics of external interconnects prior to network stitching. A static timing analysis is performed on the final network.

The invention discloses a method for determining the order of an unknown model based on the traversing and identification of a genetic algorithm, and belongs to the technical field of undetermined system identification modeling. The method is characterized by comprising three stages of initial setting, increasing identification and order determination, wherein the initial setting stage is used for setting the search range of the unknown model order and comprises three steps of acquiring frequency domain response data, setting an upper limit and a lower limit for the model order and initializing the order; the increasing identification stage is used for searching for and identifying all model structures in the set stage and comprises three steps of identifying by the genetic algorithm, recording the identification results and increasing the order; and the order determination stage is used for finding the optimal model structure by a cost function and comprises two steps of optimizing by the cost function and determining order combinations, wherein through the three stages, all possible order combinations of the unknown model can be traversed and identified, each order combination is approached to the greatest degree by using the genetic algorithm, and then the order combination corresponding to the minimum cost function is found from the identification results so that the order of the unknown model can be determined. In the method, the genetic algorithm is used to perform order traversing system identification on the frequency domain response data of the unknown model and find the model structure approaching the experimental data most, and thus, the order of the unknown model is determined by using experimental data and an optimizing means.

The invention discloses a control design method of a hypersonic flight vehicle. The method comprises the following steps that a longitudinal model of the hypersonic flight vehicle is given; finite time height control design to the hypersonic flight vehicle is achieved according to an engine dynamic equation described by a second-order system model; according to the sufficient condition of controller gain selecting, a flight speed and a flight path angle are converged to a given value within the finite time, and therefore the finite time speed control design to the hypersonic flight vehicle is achieved. According to the motion control design method of the hypersonic flight vehicle, based on the control method of a finite time control technology, the controller design method of dynamic inverse and the finite time control technology is combined, height control and speed control are seen as two sub systems for designing controllers respectively, and the difficult problems that the model order of the hypersonic flight vehicle is high and finite time control can not be achieved easily are solved.

The embodiment of the invention provides a bridge monitoring data modal identification method and device, and the method comprises the steps: simplifying an original Hankel matrix, and obtaining a simplified Hankel matrix; creating an autocorrelation matrix according to the simplified Hankel matrix; determining a modal order according to the autocorrelation matrix; calculating a projection matrixaccording to the simplified Hankel matrix, and weighting the projection matrix to obtain a weighted projection matrix; singular value decomposition and the modal order are carried out on the weightedprojection matrix to obtain an observable matrix; and identifying modal parameters according to the observable matrix. An autocorrelation matrix is created through the simplified Hankel matrix, and the model order is accurately selected by utilizing the relative difference between eigenvalues of the autocorrelation matrix, so that an unimportant mode in an estimation result is eliminated, and pseudo-mode interference caused by noise and calculation errors is avoided.

The model order reducing technology is one very useful technology for raising circuit simulating and verifying speed in improving circuit design scheme. The present invention establishes one kind of directly projecting method of reducing order of non-linear circuit model, and features the combination of direct projection and variation analysis in control theory. The sub-system obtained through variation analysis is used in constituting son projection matrix V1, V2, ..., Vn; and the relation between the original system and the son system is then used in constituting united projection matrix V. The projection matrix V is finally used in realizing direct projection of one approximate non-linear system on the original non-linear system. Compared with the order reducing the sub-system method, the present invention has greatly raised order reducing precision and efficiency.

The present invention relates to a full-system model establishing method when a multi-port direct current system an alternating current system are interconnected, especially to an estimation method of the full-system electromechanical oscillation mode after the multi-port direct current system is accessed to the alternating current system. The method comprises the following steps: S1: establishing the transfer function model of the multi-port direct current system and the transfer function model of an alternating current system, and determining the coupling relation between the transfer function model of the multi-port direct current system and the transfer function model of an alternating current system; and S2: assessing the influence of the dynamic interaction effect of the multi-port direct current system and the alternating current system on the system electromechanical oscillation mode according to the coupling relation. The present invention provides a derivation method capable of predicting the electrical power system electromechanical oscillation mode after accessing the multi-port direct current system and a rapid estimation method capable of predicting the electrical power system electromechanical oscillation mode after accessing the multi-port direct current system, so that the model order and the computation complexity in the feature value analysis are reduced, and a simple and effective practical method without losing accuracy is provided for the engineering practice.

A structure local defect recognition method based on a vector autoregression model comprises the steps: 1, collecting acceleration responses of a defect structure and a defect-free structure, and decomposing the collected acceleration responses into a plurality of samples with the same data length; 2, determining the VAR model order of each sample obtained in the step 1 by using an AIC criterion;3, according to the VAR model order determined in the step 2, establishing each sample VAR model; 4, extracting diagonal elements of the coefficient matrix in each VAR model in the step 3, and defining a brand-new feature vector f; 5, respectively calculating Mahalanobis distances of a defect state and a defect-free state according to the vector f in the step 4; and 6, performing subsequent structural defect identification according to the Mahalanobis distances of the defect state and the defect-free state calculated in the step 5. According to the method, the vibration response of the structure can be effectively utilized, the vibration response of the structure is analyzed by means of the vector autoregression model, and the local defect identification of the structure is quickly and effectively completed.

The invention provides a vector fitting and balanced truncation method based electromagnetic compatible macro model modeling method. The vector fitting and balanced truncation method based electromagnetic compatible macro model modeling method comprises measuring port scattering parameters to obtain scattering parameters under different frequency points and obtaining a state space model due to conversion through a vector fitting method; constructing a passivity judgment matrix according to a state equation coefficient matrix and performing model passivity judgment and enhancement; calculating controllability and observability Grammian matrixes according to the state equation coefficient matrix to achieve system balance; calculating an HSV (Hankel Singular Value) and a corresponding curvature spectrum of every frequency point and confirming orders of a reduced-order model; confirming a reduced order macro module; obtaining a general module. According to the vector fitting and the balanced truncation method based electromagnetic compatible macro model modeling method, a wideband system model of an original circuit can be fit only according to testing data within a small frequency band, the accuracy is high, the fitting time is short, the model orders can be confirmed rapidly after order reduction, the electromagnetic compatible macro model modeling time is greatly reduced, and a converted SPICE (Simulation Program with Integrated Circuit Emphasis) net list has universality.