141 results about "Nonlinear correlation" patented technology

A system for detecting subtle differences in a signal in a set of linearly and / or non-linearly related signals that characterize a sensor-instrumented machine, process or living system. The system employs an improved similarity operator for signal differentiation. Signals or data representative of several linearly and / or non-linearly related parameters that describe a machine, process or living system are input to the inventive system, which compares the input to acceptable modeled states. If one or more of the input signals or data are different than expected, given the relationships between the parameters, the inventive system will indicate that difference. The system can provide expected parameter values, as well as the differences between expected and input signals; or the system can provide raw measures of similarity between the collection of input signals and the collection of acceptable modeled states. The system can be embodied in software or in a micro-controller.

A method of identifying and localizing objects belonging to one of three or more classes, includes deriving vectors, each being mapped to one of the objects, where each of the vectors is an element of an N-dimensional space. The method includes training an ensemble of binary classifiers with a CISS technique, using an ECOC technique. For each object corresponding to a class, the method includes calculating a probability that the associated vector belongs to a particular class, using an ECOC probability estimation technique. In another embodiment, increased detection accuracy is achieved by using images obtained with different contrast methods. A nonlinear dimensional reduction technique, Kernel PCA, was employed to extract features from the multi-contrast composite image. The Kernel PCA preprocessing shows improvements over traditional linear PCA preprocessing possibly due to its ability to capture high-order, nonlinear correlations in the high dimensional image space.

A method and system for combining a feedback control and a feedforward control in a linear MPC to minimize effect of model uncertainty. An externally computed feedforward signal, which is more accurate and reliable, can be utilized in association with the MPC, A steady state relation between system parameters can be determined in order to compute the feedforward signal for a set of actuators associated with a non-linear system. A feedback MPC controller can then be designed. A state observer can be configured as an unknown input observer to estimate the effect of the feedforward signal. A strategy for manipulating the constraints of the MPC feedback signal can be implemented. A resulting control action for the actuators can be provided as a sum of corresponding feedback and feedforward signal while ensuring the constraints satisfaction.

The invention discloses a vibration and audio signal-based high-speed train track damage detecting method, belongs to the field of signal detection and processing as well as safety monitoring, and solves the problems of low detection speed and single detection method in the conventional train track damage detection. The method comprises the following steps: 1, acquiring vibration signals and audio signals of a train track through sensors arranged at train track detection points; 2, respectively extracting information characteristics included in the vibration signals and the audio signals; 3, respectively obtaining a nonlinear correlation curve of the vibration signals and a nonlinear correlation curve of the audio signals by using a nonlinear correlation analysis method; 4, respectively analyzing the information of the two nonlinear correlation curves obtained in the step 3 so as to respectively obtain minimum values of the two nonlinear correlation curves; and 5, carrying out data fusion on the two minimum values and corresponding information thereof so as to obtain a damage coefficient, and looking up a table to obtain the damage degree according to the coefficient. The method is suitable for detecting the damage on railway train tracks and monitoring the safety operation of trains.

Methods and apparatus for high-deposition sputtering are described. A sputtering source includes an anode and a cathode assembly that is positioned adjacent to the anode. The cathode assembly includes a sputtering target. An ionization source generates a weakly-ionized plasma proximate to the anode and the cathode assembly. A power supply produces an electric field between the anode and the cathode assembly that creates a strongly-ionized plasma from the weakly-ionized plasma. The strongly-ionized plasma includes a first plurality of ions that impact the sputtering target to generate sufficient thermal energy in the sputtering target to cause a sputtering yield of the sputtering target to be non-linearly related to a temperature of the sputtering target.

A method of identifying and localizing objects belonging to one of three or more classes, includes deriving vectors, each being mapped to one of the objects, where each of the vectors is an element of an N-dimensional space. The method includes training an ensemble of binary classifiers with a CISS technique, using an ECOC technique. For each object corresponding to a class, the method includes calculating a probability that the associated vector belongs to a particular class, using an ECOC probability estimation technique. In another embodiment, increased detection accuracy is achieved by using images obtained with different contrast methods. A nonlinear dimensional reduction technique, Kernel PCA, was employed to extract features from the multi-contrast composite image. The Kernel PCA preprocessing shows improvements over traditional linear PCA preprocessing possibly due to its ability to capture high-order, nonlinear correlations in the high dimensional image space.

An LDPC unit decoder included in a signal decoding device is provided with a parity checking unit that is a multiplier for performing multiplication of a check matrix (parity check matrix) and temporary estimated values of an encoded signal, computed by a temporary estimated value computation unit. A check matrix (parity check matrix) holding unit holds the check matrix (parity check matrix). If s and t are natural numbers, s≧t≧2, among s columns extracted from this check matrix (parity check matrix), t columns or less have a linearly independent relationship. These s columns are multiplied at locations where error occurrence frequency is relatively high, with regard to the temporary estimated values. According to this mode, the matrix is composed to include t columns that are linearly independent, that is, t columns that are not linearly dependent. In this way, in a decoded signal series, when errors occur at t locations, since it is possible to avoid multiplying all the linear dependent columns at locations where errors occur, errors with a relatively high frequency of occurrence can be preferably detected.

A plurality of acoustic source and / or detector elements are employed in a scanning mode to acoustically illuminate and acquire acoustic data from numerous sites within a larger target area. Based on the acoustic data collected in the scanning mode, a localised site within the target area is selected as the target site for focused acoustic illumination and / or probing. Elements of the acoustic source / detector array are focused on the selected target site in an automated fashion. Thereafter, the target area is periodically scanned and the acoustic focus repositioned, as required, to maintain the focus of the acoustic source at the desired target site.

A method for automatically characterizing non-linearities of a perpendicular read-back signal of a recording system is disclosed. The method includes using dibit extraction to obtain a read-back signal having a main pulse and a plurality of echoes where the read-back signal exhibits a baseline shift. An area under the first echo is integrated to obtain a first area where the integrating subtracts any baseline shift area within the first echo and where the first echo is associated with a first non-linearity. The method integrates an area of the read-back signal under the main pulse to obtain a second area where the integrating subtracts any baseline shift area within the main pulse. A first parameter is computed that characterizes the first non-linearity based on the first area and the second area. The method may be applied to characterize several non-linearities.

The embodiment of the present invention provides a method and device of constructing a building energy consumption predication model. The method comprises the steps of obtaining an energy consumptioninfluence factor set; dividing the energy consumption influence factor set into a linear correlation influence factor set and a nonlinear correlation influence factor set; separately constructing thecorresponding Bayes network models; based on the corresponding Bayes network models, grouping into the first primary influence factors, the first secondary influence factors, the second primary influence factors and the second secondary influence factors; constructing the BP neural network training models; based on the training sample data, training the BP neural network training models separately; based on the preset test sample data, separately predicting and verifying the trained BP neural network training models, and outputting the prediction result values; if an error of the prediction result values is within a preset error range, outputting an energy consumption predication model of the linear correlation influence factors and an energy consumption predication model of the nonlinearcorrelation influence factors.

The invention provides a color image authentication method and system based on hypercomplex number encrypted domain sparse representation. The color image authentication method comprises the steps of generating a phase mask according to a Logistic chaotic sequence, carrying out double random phase encryption on a color image by adopting quaternion Fourier transform so as to acquire ciphertext data, carrying out sparse representation on a component of the ciphertext data by using a sparse matrix so as to acquire a sparse ciphertext, wherein the component comprises a real component and an imaginary component, decrypting the sparse ciphertext, calculating a nonlinear correlation coefficient of a decrypted image, and carrying out authentication on the color image according to a centralized peak value. The color image authentication method based on the hypercomplex number encrypted domain sparse representation in the embodiment of the invention has high security, and a certain degree of noises can be resisted, thereby being suitable for being applied to the fields of secret communication and authentication of images.

The invention discloses a target identification device and method based on ghost imaging calculation. The device is formed by a mode-locked laser, a laser beam expander, a diaphragm, a high-speed spatial light modulator, an imaging lens, a high-speed photodiode, a digital acquisition module and the like. The method comprises the steps of utilizing a random speckle generated by the high-speed spatial light modulator for sampling a target to be identified, adopting the ghost imaging calculation for rebuilding the target, and realizing target identification under extra-low sampling ratio through nonlinear correlation. According to the target identification device and the method based on the ghost imaging calculation provided by the invention, an information acquisition method of the ghost imaging calculation is adopted, and a nonlinear correlation test is combined, so that not only is target identification with single pixel and low sampling ratio realized and is the spectral region of target identification expanded, but also the size of a detector is reduced, and the cost of an imaging system is reduced.

Systems and methods are provided for predicting and optimizing the probability of an outcome event. In a specific embodiment, the disclosure is directed to a multi-phase communication system configured to perform predictive analyses during stages based on input received from a user. In a particular implementation, there may be a first communication phase configured to accept limited input from a user to establish linear dependency between input and an outcome event for the purpose of an agent assignment, followed by a second communication phase to provide sequential predictive analyses based on natural conversation data between a user and agent. In a specific embodiment, the second communication phase may implement a second predictive model trained to identify non-linear dependencies between communication data and an outcome event. Herein is also described a graphical user interface for representing scores corresponding to the probability of outcome events, among other features.

The invention discloses a reservoir physical property parameter prediction method combined with deep learning, and the method comprises the steps: introducing the nonlinear correlation between an MICquantitative measurement physical property parameter and a logging curve, and selecting the logging curve which is obvious in response to the physical property parameter; introducing CEEMDAN to decompose the physical property parameter data sequence to obtain an IMF component and a residual RES component of an intrinsic mode function, and subjecting the physical property parameter data sequence tostationary processing; introducing SE to evaluate the complexity of each IMF component and RES margin, and recombining component sequences with similar entropy values to obtain a new intrinsic mode component; carrying out normalization processing on the new intrinsic mode component data and then dividing the new intrinsic mode component data into a training set and a test set; introducing an LSTMrecurrent neural network to establish a prediction model for the reconstructed new component, and obtaining a prediction value of each new intrinsic mode component; and carrying out inverse normalization on the prediction value of each new intrinsic mode component, and carrying out superposition reconstruction to obtain a physical parameter prediction result. According to the method, the modelingnumber of redundant information and prediction components is reduced, and the prediction precision and the prediction speed are improved.

The invention discloses an electric power system static security assessment method considering wind speed correlation. According to wind speed historical data, cut-off pair copula is used to realize wind speed modeling with non-linear correlation; then according to fan power characteristics, electric power system quick static security assessment is realized. The electric power system static security assessment method can be used for processing wind speeds with any probability characteristics in an electric power system and is suitable for quickly assessing the influence of connected wind power plants on the static security of the electric power system in medium-term and long-term plans.

The invention discloses a comprehensive monitoring method for a large coal-fired unit combustion system based on dynamic characteristic and static characteristic synergistic analysis of. According toan idea of system distribution, a combustion system is used as an upper integrated system, and is physically divided into various lower-level sub-devices according to system structural characteristics. For the lower-level sub-devices, dynamic information and static information of each sub-device are extracted by using slow feature analysis, and independent monitoring of each sub-device is realized. In the upper layer, process characteristics extracted from all lower-layer sub-devices are synthesized, nonlinear relations between different device variable groups are taken into account, and the dynamic and static information of the whole combustion system and the nonlinear correlations between sub-device variable groups are extracted by using kernel slow feature analysis, so that the monitoring of the overall process state of the entire combustion system is realized. Through the comprehensive monitoring method for the large coal-fired unit combustion system based on the dynamic characteristic and static characteristic synergistic analysis, the synergistic comprehensive detection of multiple devices is realized, The normal working condition switching and process faults of a system andthe sub-devices can be effectively distinguished through the extracted dynamic information and the static information, and the accuracy of the industrial large-scale system state monitoring is improved.

The invention belongs to the technical field of wind speed prediction, and discloses a wind speed prediction method and system based on a neighborhood gate long-short-term memory network. The Pearsoncorrelation coefficient and the maximum information coefficient are respectively adopted to explore the linear and nonlinear correlation among variables to screen the wind speed correlation factors. On the basis of correlation analysis, the Granger causality test is used to explore the statistical causality of wind speed and wind speed factors. The structure of causality is divided into five types, and all types of causality are unified into an equivalent tree causality structure by the method of 'decomposition-dummy variable-pruning'. Aiming at the causality structure of equivalent tree, a long-term and short-term memory network model based on neighborhood gates is proposed to predict wind speed. The prediction method (NLSTM) of the invention accurately considers the causality between thewind speed and the wind speed factors, effectively improves the prediction accuracy of the wind speed, and plays a vital role in the application of the wind power and the dispatching of the power grid.

Estimation and imaging of linear and nonlinear propagation and scattering parameters in a material object where the material parameters for wave propagation and scattering has a nonlinear dependence on the wave field amplitude. The methods comprise transmitting at least two pulse complexes composed of co-propagating high frequency (HF) and low frequency (LF) pulses along at least one LF and HF transmit beam axis, where said HF pulse propagates close to the crest or trough of the LF pulse along at least one HF transmit beam, and where one of the amplitude and polarity of the LF pulse varies between at least two transmitted pulse complexes. At least one HF receive beam crosses the HF transmit beam at an angle >20 deg to provide at least two HF cross-beam receive signals from at least two transmitted pulse complexes with different LF pulses. The HF cross-beam receive signals are processed to estimate one or both of i) a nonlinear propagation delay (NPD), and ii) a nonlinear pulse form distortion (PFD) of the transmitted HF pulse for said cross-beam observation cell.

A near field communications (NFC) device is disclosed that detects a presence of another NFC capable device within its magnetic field. The NFC device observes signal metrics of an observed detection signal at various intervals. The NFC device determines a statistical relationship based upon at least two first signal metrics from among the signal metrics to determine an estimate of at least one second signal metric from among the signal metrics. The NFC device compares a difference between the estimate of the signal metric of the at least one second signal metric and the at least one second signal metric. The NFC capable device makes a first determination that another NFC device is present within its magnetic field when the difference indicates that the at least one second signal metric is non-linearly related to the at least two first signal metrics.

A method, and program for implementing such method, for use in estimating a conditional probability distribution of a current signal state and / or a future signal state for a non-linear random dynamic signal process includes providing sensor measurement data associated with the non-linear random dynamic signal process. A filter operating on the sensor measurement data by directly discretizing both amplitude and signal state domain for an unnormalized or normalized conditional distribution evolution equation is defined. The discretization of the signal state domain results in creation of a grid comprising a plurality of cells and the discretization in amplitude results in a distribution of particles among the cells via a particle count for each cell.

The invention belongs to the technical field of runoff prediction, and discloses a runoff probability prediction method and system based on deep learning, and the method comprises the steps: employinga maximum information coefficient to analyze the linear and nonlinear correlation between variables, so as to screen a runoff correlation factor; building an extreme gradient boosting tree model on the basis of correlation analysis, and inputting runoff correlation factors into a trained XGB model to complete runoff point prediction; inputting a point prediction result obtained by the XGB model into a GPR model, and performing secondary prediction to obtain a runoff probability prediction result; selecting confidence and acquiring a runoff interval prediction result under the corresponding confidence through Gaussian distribution; and optimizing hyper-parameters in the XGB model and the GPR model by adopting a Bayesian optimization algorithm. A high-precision runoff point prediction result, an appropriate runoff prediction interval and reliable runoff probability prediction distribution can be obtained, and the prediction method plays a crucial role in utilization of water resourcesand reservoir scheduling.

An electronic circuit for adjusting a property of an input signal of a non-linear element is disclosed. A detector detects a strength of the input signal. An analog controller forms a control signal as a piecewise approximation of a non-linear dependency between the property and the strength of the input signal detected by the detector. An adjusting circuit adjusts the property of the input signal according to the control signal from a controller.

The invention discloses a line transformation relation abnormity judgment method based on correlation between electric quantity and line loss. The line transformation relation abnormity judgment method includes a data acquisition module which is used for acquiring electric quantity information, a data processing module which is used for analyzing correlation between electric quantity and line loss, and a data display module which is used for displaying abnormal line transformation relations, wherein the data acquisition module, the data processing module and the data display module are sequentially connected. The line transformation relation abnormity judgment method utilizes power consumption, voltage, current and other data of a line and a transformer provided by a power consumption information acquisition system, and utilizes a big data analysis technology and a transformer area line loss non-linear correlation analysis method to construct the line transformation relation error correction model independent of the local communication relation and find out the line transformation relation with statistical errors, so as to improve the accuracy of topological information of the power transmission line and provide technical guidance for conducting line transformation relation on-site troubleshooting, so that the calculation precision of line loss is improved, and technical support is provided for lean marketing management.

The invention discloses a behavior recognition method of nuclear covariance descriptors based on dense tracks. The objective of the invention is to solve a problem of low behavior recognition accuracy caused by nonlinear correlation between different characteristics which are not considered in the prior art. The method comprises steps of 1), extracting dense tracks, extracting characteristics of each pixel point in a track cuboid and acquiring a base layer characteristic matrix; 2), calculating a nuclear covariance matrix of the base layer characteristic matrix, mapping the nuclear covariance matrix to the euclidean space and acquiring vectorization characteristic representation; 3), by use of all characteristic representation in the track cuboid and constructing nuclear covariance descriptors based on dense tracks; and 4), using a BOW model to code the nuclear covariance matrix descriptors, acquiring a codon histogram, training a SVM by use of the codon histogram of a training set, testing the codon histogram of the training set in the trained SVM and acquiring a behavior recognition result. According to the invention, description ability of behaviors is further improved and the method can be used in complex environment like video monitoring.

The invention belongs to the field of ultrasonic imaging, and discloses a synthetic aperture ultrasonic imaging system for calculating delay time based on non-linear correlation. The synthetic aperture ultrasonic imaging system for calculating the delay time based on the non-linear correlation comprises a channel receiving and amplifying module, a channel amplifying and driving module, an analog / digital (A / D) conversion module, a digital / analog (D / A) conversion module, a sampling control module, a wave beam synthesizing module, a wave beam forming module, an emission control module, a universal serial bus (USB) subsidiary communication module and a USB communication module. The wave beam synthesizing module calculates the delay time of reconstruction of a synthetic aperture based on the non-linear correlation, and can solve the problem that large deviation exists between a synthetic aperture focusing algorithm based on the delay superposition and reality when the delay time of a transducer and theoretical calculation time differ so much.

The invention provides a dimension reduction model training method and device and electronic equipment, and belongs to the technical field of artificial intelligence. The embodiment of the invention provides a dimension reduction model training method, the device and electronic equipment. First, feature data is acquired, the feature data is input into a dimension reduction algorithm; a preset hierarchical number of feature results is output, the feature data are input into a to-be-trained dimension reduction model; training data of which the number is the same as that of the feature results ata preset level is obtained; a loss function value is determined according to the training data and the feature result; training data, the feature result and the feature data are used for training a to-be-trained dimension reduction model; ending training until the loss function value converges, obtaining a trained dimension reduction model. The training model can carry out dimension reduction ondata of the non-linear correlation feature, and carry out dimension reduction by the trained dimension reduction model when new feature data is obtained, so as to predict the dimension of the new feature data.

The invention discloses a method for recognizing a nonlinear correlation between equipment failures and electric quantity information. The method comprises the steps as follows: selecting equipment operation related information of multiple groups of transformers with same models in different time periods as samples; establishing an information base containing equipment electric information amounts of each group of samples and current equipment states corresponding to the equipment electric information amounts; calculating a correlation coefficient between the electric information amounts of each group of samples and the current equipment states corresponding to the electric information amounts through a nonlinear correlation recognition algorithm; and analyzing the difference between different samples through matlab software according to the correlation coefficients obtained through calculation. The method has the benefits as follows: the correlation between the equipment failures and the operating state quantities of the transformers can be analyzed rapidly and effectively, the accurate extraction of failure information is facilitated, the failure diagnostic accuracy is improved, meanwhile, the validity and unbiasedness of the nonlinear correlation recognition algorithm are verified, and the method is high in practicability.

The invention discloses a fitting method for calculating the main coefficients of the boundary conditions of a groove wall based on experimental data, aiming at solving the present situation of lacking accurate calculation of the transonic wind tunnel slot wall boundary condition coefficients. In this method, the least square solution of the overdetermined equations is obtained by fitting the experimental data of the pressure coefficient near the wall and the deflection angle along the flow direction, which exclude the bad points, and the number of the main systems in the boundary condition ofthe wall is obtained by the least square method. The invention needs to eliminate the bad points of the original data, and then subtracts the reference data of the empty wind tunnel from the pressurecoefficient to eliminate the local fluctuation caused by slotting; based on the pressure coefficient-deflection angle nonlinear correlation,an overdetermined equation is established, the least squaresolution of the overdetermined equations is obtained by using the pseudo-inverse of the coefficient matrix, and the number of coefficients of the gradient term, the first order term and the second order term of the deflection angle in the boundary condition of the channel wall is obtained. The invention can improve the accuracy of the boundary conditions of the groove wall and guide the aerodynamic design and correction of the groove wall.