53 results about "Non linear prediction" patented technology

An efficient coding scheme with higher audio bandwidth and / or better audio quality at lower bitrates, wherein the scheme eliminates long-term and short-term frequency domain correlation in a signal via frequency domain predictors. The coding scheme compresses information consisting of coded low frequency components as well as a parametric representation for the high frequency components based on a non-linear model. Additionally, by working on the frequency domain representations of the signal (such as the MDCT representation which is naturally available to a PAC encoder and decoder), low pass and high pass signal components are easily obtained by windowing the appropriate ranges of frequencies in the signal. Furthermore, the power functions of the signal are replaced by corresponding convolution functions of the same order.

The invention discloses a robust dicyclic photovoltaic grid-connected control method based on power feedforward, which mainly comprises three parts: proportion integration (PI) control of exocyclic voltage, dead-beat control of endocyclic robust predictive current and power feedforward control, wherein the exocyclic PI control of the voltage is used for stabilizing the capacitive voltage of a direct-current side; the dead-beat control of the endocyclic robust predictive current is used for carrying out linear prediction on the voltage of a power grid at a next control period and carrying out non-linear prediction on grid-connected current by advanced control so as to obtain the command value of the grid-connected current at the next period; and then, PWM (pulse-width modulation) and grid-connected control are realized by the dead-beat control. According to the robust dicyclic photovoltaic grid-connected control method based on the power feedforward, a photovoltaic grid-connected inverter has higher robustness, wider stability margin and quicker dynamic response, and the requirement of grid-connected running of an inverter is better met.

A method and apparatus for non-linear prediction filtering are disclosed. In one embodiment, the method comprises performing motion compensation to generate a motion compensated prediction using a block from a previously coded frame, performing non-linear filtering on the motion compensated prediction in the transform domain with a non-linear filter as part of a fractional interpolation process to generate a motion compensated non-linear prediction, subtracting the motion compensated non-linear prediction from a block in a current frame to produce a residual frame, and coding the residual frame.

Disclosed is a general system and method, for prediction of binding of peptide-like ligands (peptides) to peptide-like receptors (receptors). Specifically this invention uses non-linear prediction models (including, but not limited to, artificial neural networks), sequence data form ligands and their respective receptors, and known ligand-receptor binding affinities. The representation of ligand-receptor interaction used along with the binding affinity of said interaction is used to train a determining means in a form of a predictive model. Prediction of binding affinity of a novel (not used for training of a predictive model) ligand-receptor interaction, involving a peptide and a particular receptor, involves the combining of representations of both peptide and receptor and presenting that representation to a previously trained predictive model. The system and method can be used as a single predictive model for determination of ligand binding to an individual receptor, or to a group of related receptors. This system and method was validated using data on peptide binding to major histocompatibility complex molecules (MHC) and artificial neural networks (ANN).

The invention provides a method of using genetic algorithm to optimize a BP neural network system. The genetic algorithm is used to optimize the weight and the threshold of the neural network, and optimized searching space is located in solution space. The searching space is used as the initial weight and the initial threshold of the searching of the neural network, and the local searching capability of the neural network is used to search an optimal solution in the searching space. The method provided by the invention is used to combine the advantages and the disadvantages of the BP neural network and the genetic algorithm, and then the convergence speed of the network is accelerated, and therefore precision of a model is effectively improved, and the optimized combination of the two research directions, namely the BP neural network and the genetic algorithm, is realized. A system is widely used for an intelligent detection field, a non-linear prediction field, a pattern recognition field, a robot control field, and other fields, and has good practicability.

The invention provides a lithium ion battery remaining life direct prediction method based on probability integration, relates to the technical field of lithium ion battery remaining life prediction and aims at solving the problem that a traditional MONESN method is unstable and lack of remaining life uncertainty expression. The method comprises the steps of firstly measuring the maximum capacity of a lithium ion battery in easy circulating period; adopting N MONESN models to predict the lithium ion battery remaining life and obtain N prediction results; performing uncertainty estimation and integration on the prediction results so as to obtain a lithium ion battery remaining life prediction result based on probability integration. The lithium ion battery remaining life direct prediction method fully plays the strong non-linear prediction capacity of the MONESN models and effectively solves the problem of instability of a traditional MONESN algorithm. In addition, uncertainty expression and management are achieved. The lithium ion battery remaining life direct prediction method is suitable for lithium ion battery remaining life prediction under the condition that the capacity can be directly measured and obtained.

The invention discloses an X-ray pulsar navigation positioning method and system based on nonlinear prediction strong tracking traceless Kalman filtering. The navigation positioning method comprises the following steps of taking a spacecraft position vector and a velocity vector as a navigation state variable, establishing a navigation system state model and obtaining a spacecraft state predictionvalue; determining a pulsar signal observation value and establishing a navigation system observation model; using a non-linear prediction strong tracking traceless Kalman filtering method to processa pulsar signal observation value and a spacecraft state prediction value, in the spacecraft state prediction stage, acquiring a minimum navigation system state model error according to a constraintfunction, and correcting a navigation system state model error in a quasi real-time mode; and in a spacecraft state updating stage, introducing a fading factor to suppress a noise interference, predicting and updating the state of a spacecraft. A spacecraft state model error is estimated and corrected, and simultaneously, the problems of filter divergence and the low precision of X-ray pulsar navigation caused by the noise interference are solved.

Inter-color image prediction is based on color grading modeling. Prediction is applied to the efficient coding of images and video signals of high dynamic range. Prediction models may include a color transformation matrix that models hue and saturation color changes and a non-linear function modeling color correction changes. Under the assumption that the color grading process uses a slope, offset, and power (SOP) operations, an example non linear prediction model is presented.

The invention relates to an airport freight traffic prediction analysis method based on an autoregressive integrating moving average (SARIMA) and RBF neural network integration combination model. According to the method, an airport freight traffic linear part is predicted by using seasonal SARIMA; a non-linear airport freight traffic part is predicted by using an RBF neural network; and then the non-linear prediction result is used as compensation of the linear prediction result, thereby obtaining a final prediction result.

The invention discloses a method for controlling non-linear prediction of a helicopter for spinning recovery. The method includes the following steps: after entering spinning, using a pre-trained helicopter requirement torque model to conduct real-time online prediction on current helicopter requirement torque; after entering the stage of spinning recovery, using a pre-trained engine dynamic parameter model to conduct real-time online prediction on current engine dynamic parameters, at the same time using online prediction results of the helicopter requirement torque model and the engine dynamic parameter model to resolve so as to reduce a difference between a helicopter requirement torque upon the connection of a clutch and a torque support provided by an engine, taking into consideration of rolling optimization of operation conditions for stability and safety of the engine, taking a first item of controlled variable sequence that is solved as the helicopter control variable that is currently input. The invention also discloses an apparatus for controlling non-linear prediction of the helicopter. According to the invention, the method and the apparatus can effectively shorten time delay at the stage of spinning recovery and reduce rotating speed transient downslide of helicopter rotors.

An observed signal that is corrupted with impulse noise is recorded in a signal processing system of an image processing system or a digital subscriber line (xDSL). The observed signal that is recorded by the signal processing system includes a noise component and data component. The signal processing system estimates the parameters of an alpha-stable distribution using a modified iteratively reweighted least squares (IRLS) technique. The estimated parameters define a probability density function that is used to model the noise component of the observed signal. Once the parameters of the alpha-stable distribution are estimated, the signal processing system uses them to estimate model coefficients of a non-linear prediction filter such as a Volterra filter. Using the model coefficients, the non-linear prediction filter estimates the data component of the observed signal.

Inter-color image prediction is based on color grading modeling. Prediction is applied to the efficient coding of images and video signals of high dynamic range. Prediction models may include a color transformation matrix that models hue and saturation color changes and a non-linear function modeling color correction changes. Under the assumption that the color grading process uses a slope, offset, and power (SOP) operations, an example non linear prediction model is presented.

Inter-color image prediction is based on color grading modeling. Prediction is applied to the efficient coding of images and video signals of high dynamic range. Prediction models may include a color transformation matrix that models hue and saturation color changes and a non-linear function modeling color correction changes. Under the assumption that the color grading process uses a slope, offset, and power (SOP) operations, an example non linear prediction model is presented.

A servo turnable controller based on a non-linear switching system is characterized in that a multi-model switching system is led into precise modeling and controlling of a servo turnable, global on-line rolling optimization calculation comprising a linear sub system and a non-linear sub system is adopted, a prediction function controller is applied to any sub controller, local prediction function controllers u2(k) and u4(k) are directly applied to linear models 2 and 4 shown as the formula, and non-linear prediction function controllers u1(k) and u3(k) are applied to non-linear models 1 and 3. As for specific turntables of other models, according to specific situations of non-linear factors, the modality of the switching system and the controllers based on sub system models can be increased, and operation methods are the same as the presented portion.

The invention provides a multi-response parameter optimization method based on a radial basis function neural network prediction model and improved WPCA (weighted principal component analysis). According to the method, a non-linear prediction model of a production process is built by adopting a radial basis function neural network, capacity prediction indexes of the neural network model are introduced, a WPCA algorithm is adjusted, response with high prediction capacity receives priority in improvement in multi-response parameter design, and the optimization effect of technological parameters is improved. The WPCA generally adopts linear regression to establish a relation model between a response variable and a controllable factor variable in the multi-response parameter optimization design, however, the fitting degree of a linear regression model is not high for a complicated non-linear production process, and modeling requirements for parameter design cannot be met. The method is applied to the multi-response parameter optimization design of a thermal polymerization process of an aluminum-metallized polypropylene film capacitor, so that a satisfying comprehensive optimization effect of two responses of capacitance and loss tangent value of the capacitor is realized.

The invention discloses a least square method support vector machine-based generalized prediction method in a lysozyme fermentation process. The prediction method comprises the following steps of establishing a non-linear prediction model, and training a least square method support vector machine by using production data with higher yield screened from tank fermentation; performing real-time linearization on the input and output non-linear prediction model, setting a reference trajectory, rolling-optimizing controller design, and intelligently embedding an LS-SVM (least square-support vector machine)-based generalized prediction control algorithm in the lysozyme fermentation process into an upper computer. According to the method, the least square method support vector machine and the generalized prediction control are combined, so the QP problem of time consumption of solving in the solving process with the model is avoided, the operation is simple, the convergence speed is speed, and the precision is high. A genetic algorithm and the rolling optimizing in the generalized prediction control are combined, so the robustness of a system is enhanced, and the lag and disturbance of the system are effectively overcome.

An observed signal that is corrupted with impulse noise is recorded in a signal processing system of an image processing system or a digital subscriber line (xDSL). The observed signal that is recorded by the signal processing system includes a noise component and data component. The signal processing system estimates the parameters of an alpha-stable distribution using a modified iteratively reweighted least squares (IRLS) technique. The estimated parameters define a probability density function that is used to model the noise component of the observed signal. Once the parameters of the alpha-stable distribution are estimated, the signal processing system uses them to estimate model coefficients of a non-linear prediction filter such as a Volterra filter. Using the model coefficients, the non-linear prediction filter estimates the data component of the observed signal.

The invention discloses a method for realizing the five-freedom-degree alternating current active magnetic bearing displacement self detection by utilizing a mixed kernel function support vector machine displacement predicating model. According to the method, magnetic bearing control current is used as an input sample, radial and axial displacement is used as an output sample, the sample data is collected, a mixed kernel function is selected, the performance parameters of the support vector machine are optimized through a particle swarm algorithm, the training sample and the performance parameters are utilized for training the least square support vector machine, and a non-linear predicting model is built. The predicting model is connected with a linear closed loop controller before being connected to a five-freedom-degree alternating current active magnetic bearing in series, the magnetic bearing displacement closed loop control is formed with a first and second expansion current hysteresis three-phase power inverter and a switch power amplifier, the self detection of a five-freedom-degree alternating current active magnetic bearing displacement-free sensor is realized, the cost of a magnetic bearing system is reduced, and the dynamic property of the system is improved.

The invention relates to a non-linear prediction control system in an internal thermal coupling distillation process, which comprises a field intelligent instrument and a DCS system which are directly connected with an internal thermal coupling distillation tower; the field intelligent instrument is connected with a storage device, a control station and an upper computer; the upper computer comprises a non-linear prediction controller which is used to roll, optimize and solve a control law and output a control variable; the non-linear prediction controller comprises a component deduction module, a model parameter self-adaptive correction fitting module, and a control law rolling, optimizing and solving module; the component deduction module is used to obtain temperature and pressure data from the intelligent instrument, and calculate the component concentration of all tower plates of the high-efficiency energy-saving distillation tower, the model parameter self-adaptive correction fitting module is used to adopt the component concentration data calculated by the component deduction module in a historical database and fits the model parameters on line; and the control law rolling, optimizing and solving module is used to optimize and solve the ideal value of the current control variable according to the current component concentration data, model functions and the current time operation variable. The invention also provides a non-linear prediction control method. The invention has good control effect and ideal control quality.

The invention provides a prediction method of the long-term aging performance degradation of high-Nb type GH4169 alloy. The method takes the long-term aging post-tissue of the alloy and the performance data of the long-term aging post-tissue of the alloy as a basis, and parameters, aging temperature T, aging time t and a hardness value of the alloy in a long-term aging process of the alloy are represented as a predictable function relationship. The predication method is suitable for hardness predication in the long-term aging process of the alloy and can provide reference data for designing and using the alloy. The prediction method is realized in the following steps: 1) carrying out heat treatment on the alloy; 2) implementing long-term aging experiments of different temperatures and time; and 3) constructing and verifying a long-term aging performance prediction model. When the method disclosed by the invention is adopted, the long-term tissue and performance of the alloy can be more accurately predicted, and a life prediction basis is provided for the use of the type of alloy. The prediction method can be applied to performance prediction in the long-term aging process of the hot-end component of the high-Nb type GH4169 alloy of the aerospace field.

The invention discloses a method for realizing displacement self detection of a three-DOF (Degree of Freedom) AC / DC (Alternating Current / Direct Current) hybrid magnetic bearing by utilizing a displacement prediction model of a hybrid kernel function SVM. The method is characterized in that magnetic bearing control current is used as an input sample, radial and axial displacements are used as output samples, sample data are collected, a hybrid kernel function is selected, performance parameters of an SVM are optimized through a PSO (Particle Swarm Optimization), an LS (Least Squares) SVM is trained by utilizing a training sample and the performance parameters, a non-linear prediction model is established, the prediction model is connected with a linear closed-loop controller before being connected to the three-DOF AC / DC hybrid magnetic bearing in series, magnetic bearing displacement closed-loop control is formed with an extended current hysteresis loop three-phase power inverter and a switch power amplifier, and self detection of a three-DOF AC / DC hybrid magnetic bearing displacement-free sensor is realized.

The invention discloses a vehicle stability index estimation method based on depth learning, which comprises a high-precision GPS system, a gyroscope and a front wheel angle sensor, a preliminary model of a long and short term memory (LSTM) neural network is established based on a software platform, the real vehicle experimental sample data is used to train the LSTM neural network to generate a time-delay non-linear prediction model, and after meeting the accuracy of the vehicle gauge level, a complete estimation module is packaged, and the centroid side deflection angle and the yaw rate valueof the vehicle are automatically output according to the real time input of the sensor information so as to realize the vehicle state estimation. According to the invention, the model has the capabilities of on-line learning and dynamic updating while estimating the vehicle state, the estimation precision is continuously improved through self-learning, and the development of the active safety control of the automobiles is promoted.

The invention relates to a prediction control algorithm based on a motion linear model and an area performance index. The prediction control algorithm comprises 1) linearization hypothesis; 2) a ship motion equation in an increment form; 3) the area performance index; 4) a prediction equation; 5) a constraint condition: 4) equation solution. The prediction control algorithm is advantageous in that the algorithm provided by the invention is a linear method, and the solution is simpler by comparing with a non-linear prediction control algorithm, and a result is reliable; the solution is simple, and calculation time is short, and therefore the prediction control algorithm is more suitable for a field having a high requirement on real-time performance like ship control; based on the area performance index, the prediction control algorithm meets a current green control concept, and therefore wear of a propeller is effectively prevented, and energy consumption is reduced.

Techniques for modeling a wellbore fluid that includes a base fluid and one or more fluid additives includes identifying a target viscosity profile of the wellbore fluid; determining an initial set of values of the fluid additives that are based at least in part on the target viscosity profile; determining, with one or more non-linear predictive models, a computed viscosity profile of the wellbore fluid and a computed set of values of the fluid additives based, at least in part, on the initial set of values of the fluid additives; comparing the computed viscosity profile and at least one of the computed set of values with a specified criteria of the wellbore fluid; and preparing, based on the comparison, an output including the computed viscosity profile and at least one of the computed set of values of a resultant wellbore fluid.

The invention discloses a chaotic characteristic analysis and non-linear prediction method of run-off. The method includes: collecting more than thirty nine run-off data; subjecting all the run-off data to wavelet transformation to generate a plurality of subsequences; respectively calculating a maximum lyapunov index for each subsequence with a small-data algorithm or other methods; judging the subsequences with the maximum lyapunov indexes being positive numbers as chaos time sequences, and respectively calculating delay time and embedded dimensions of the chaos time sequences; subjecting the corresponding subsequences with chaotic characteristic to phase-space reconstruction according to calculated delay time and the embedded dimensions, and taking reconstruction results as predicated data to be inputted to a neural network model to finish final prediction. The method has the advantage that when the run-off recorded data of rivers and seas are more than thirty nine, chaotic characteristic analysis and non-linear prediction can be performed.

The invention discloses an intelligent coagulation dosing method and device for a water purification plant, belongs to the technical field of coagulation in water purification production, and aims to solve the problems of poor anti-interference capability, high requirements on water quality and process conditions, troublesome adjustment and difficult maintenance for various water quality parameters. The method is based on collection of water plant operation big data. Non-linear modeling and non-linear prediction are carried out on the coagulation dosing process, the influence of system time lag is reduced through prediction control, the influence of raw water quality disturbance and model mismatch on a control system is eliminated, the dosing control precision and the robustness of the system are improved, and the purpose of intelligent dosing is achieved. The control model can overcome the main problem that an existing automatic dosing system is inaccurate in prediction, has high self-adaptive capacity, explores a new way for water plant coagulation dosing control, provides reliable basis for implementation of next coagulation dosing control, meanwhile, the dosing optimization method of the scheme is applied, the water quality assurance rate is increased, the labor is saved, and the coagulant cost is reduced.

The invention discloses a combined prediction method for short-time travel requirements of on online hailed car. The method comprises the following specific steps: acquiring historical travel demand data; based on the acquired historical travel demand data, establishing an ARIMA model and a BP neural network model, and performing online car-hailing short-term travel requirement prediction; performing weighted combination on the ARIMA model and the BP neural network model, and calculating a weight value of the weighted combination by utilizing a principle of minimum error in an approximate historical time period to obtain a final combined prediction model; and carrying out online car-hailing travel short-time travel requirement prediction according to the constructed combined prediction model. According to the method, the advantages of two linear and nonlinear prediction models are integrated; optimal estimation can be obtained through linear iteration based on historical data in the same time period, dynamic characteristics of online car hailing requirements can be reflected through the strong nonlinear mapping capacity of the BP neural network, overlarge errors of a single prediction model can be effectively reduced, and therefore the precision of online car-hailing short-time travel requirement prediction is improved.

The invention provides a method for detecting the length of a calcium carbide furnace electrode. The method comprises the steps: obtaining a non-linear prediction model of electrode consumption according to a model training sample set, wherein electrode temperature T, electrode power P, a furnace charge ratio R and calcium carbide output G are sample input variables, and the electrode consumption is a sample output variable; determining electrode consumption in a current working condition according to the non-linear prediction model of electrode consumption; and determining the length of an electrode according to the electrode consumption in the current working condition. The method can monitor the position changes of an electrode in real time and then guide pressing and releasing for electrodes, so the pressing and releasing amount for electrodes can be accurately controlled, and electrodes are made to work in a balance manner. The method guarantees stable production and energy-saving operation of a calcium carbide furnace.

The invention provides a non-linear prediction pitch control method based on wind speed in-advance measuring. The method comprises the steps of firstly, introducing a dynamic area to restrain the preset allowable range of the given pitch angle change rate of a wind generating set; secondly, within the preset allowable range, determining the limited control set of the given pitch angle change rateof a candidate wind generating set; thirdly, through the limited control set, searching the given pitch angle change rate sequence of an optimal wind generating set; and fourthly, taking the first element of the given pitch angle change rate sequence of the optimal wind generating set as the controller output. The non-linear prediction pitch control method based on wind speed in-advance measuringhas the important effect on unit safety, stability and efficient running.