CN111998887A - Detection device for parameter measurement - Google Patents

Detection device for parameter measurement Download PDF

Info

Publication number
CN111998887A
CN111998887A CN202010863414.2A CN202010863414A CN111998887A CN 111998887 A CN111998887 A CN 111998887A CN 202010863414 A CN202010863414 A CN 202010863414A CN 111998887 A CN111998887 A CN 111998887A
Authority
CN
China
Prior art keywords
neural network
longitudinal sliding
platform
sensor
prediction model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010863414.2A
Other languages
Chinese (zh)
Inventor
王建国
翁润庭
崔家兴
杨中员
刘伟
周恒瑞
丁晓红
王苏琪
张海江
陈亚娟
马从国
柏小颖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huaiyin Institute of Technology
Original Assignee
Huaiyin Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huaiyin Institute of Technology filed Critical Huaiyin Institute of Technology
Priority to CN202010863414.2A priority Critical patent/CN111998887A/en
Publication of CN111998887A publication Critical patent/CN111998887A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
    • G01D18/00—Testing or calibrating apparatus or arrangements provided for in groups G01D1/00 - G01D15/00
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24133—Distances to prototypes
    • G06F18/24137—Distances to cluster centroïds
    • G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"

Abstract

The invention relates to the technical field of measurement, and discloses a detection device for parameter measurement, which comprises a reference shelf table, a sensor rack plate and a base layer platform, wherein the reference shelf table and the sensor rack plate are arranged on the base layer platform, and the base layer platform is also provided with a first longitudinal sliding assembly and a transverse sliding assembly; the sensor rack plate is arranged on the transverse sliding component, and the sensor rack plate is provided with a measuring sensor which slides transversely along with the transverse sliding component and slides longitudinally along with the first longitudinal sliding component. The sensor rack board is also provided with an MSP430 singlechip detection unit, and a detection algorithm is arranged on the MSP430 singlechip detection unit to detect the measurement value of the measurement sensor. Compared with the prior art, the transverse movement and the longitudinal movement of the measuring sensor are realized through the first longitudinal sliding component and the transverse sliding component, and the problem of inaccurate jitter measurement in the measuring process of the sensor is solved; the measurement value of the measurement sensor can be accurately detected through the detection algorithm.

Description

Detection device for parameter measurement
Technical Field
The invention relates to the technical field of measurement, in particular to a detection device for parameter measurement.
Background
The sensor is an indispensable important element in the fields of industrial production, scientific research and the like as a source for acquiring natural information. In the prior art, there are various measuring sensors, such as temperature and humidity sensors, infrared sensors, photoelectric sensors, etc.
In the measurement process of the existing various measurement sensors, environmental factors have great influence on measurement, such as environmental pressure, environmental temperature and the like, and have certain influence on the measurement value of the measurement sensor, and how to accurately measure the actual parameter value of a reference object.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems in the prior art, the invention provides a detection device for parameter measurement, which is characterized in that a longitudinal sliding assembly and a transverse sliding assembly are matched to work, the longitudinal distance and the transverse distance between a measurement sensor and a reference object are rapidly adjusted, and meanwhile, an MSP430 single-chip microcomputer monitoring unit is used for accurately detecting the parameter value of the reference object measured by the measurement sensor.
The technical scheme is as follows: the invention provides a detection device for parameter measurement, which comprises a reference shelf table, a sensor rack plate and a base layer platform, wherein the reference shelf table and the sensor rack plate are arranged on the base layer platform; the sensor rack plate is arranged on the transverse sliding component, the sensor rack plate slides transversely along with the transverse sliding component, and the transverse sliding component slides longitudinally along with the first longitudinal sliding component; the sensor rack board is also provided with a measuring sensor and an MSP430 single-chip microcomputer monitoring unit, wherein the MSP430 single-chip microcomputer monitoring unit comprises a time sequence DRNN neural network prediction model, an ARIMA prediction model, a time sequence RBF neural network prediction model, a SOM neural network classifier, a plurality of ANFIS neural network models, a plurality of NARX neural network models and a metabolism GM grey predictor; the output values of the measuring sensors are respectively used as the input of a time series DRNN neural network prediction model, an ARIMA prediction model and a time series RBF neural network prediction model, the output of the time series DRNN neural network prediction model, the output of the ARIMA prediction model and the output of the time series RBF neural network prediction model are used as the input of a SOM neural network classifier, the SOM neural network classifier outputs a plurality of types of time series DRNN neural network prediction model, the output of the ARIMA prediction model and the output of the time series RBF neural network prediction model are respectively used as the input of a plurality of corresponding ANFIS neural network models, the outputs of the plurality of ANFIS neural network models are respectively used as the input of a plurality of corresponding NARX neural network models, the outputs of the plurality of NARX neural network models are used as the input of a metabolism GM grey predictor, and the output of the metabolism GM grey predictor is used as the measurement prediction value of a measurement sensor.
Further, first longitudinal sliding assembly includes first longitudinal sliding guide rail and first longitudinal sliding platform, first longitudinal sliding guide rail is fixed in through the support on the platform of basic unit, its lower surface is connected with first lead screw along its length direction rotation, threaded connection has first slider on the first lead screw, first longitudinal sliding platform cover is located on the first longitudinal sliding guide rail and its with first lead screw correspond the position with first slider fixed connection.
Further, the transverse sliding assembly comprises a first longitudinal sliding guide rail perpendicular to the first longitudinal sliding guide rail and a second lead screw rotatably connected to the upper surface of the first longitudinal sliding platform, a second sliding block is connected to the second lead screw in a threaded mode, a transverse sliding platform is fixed on the second sliding block, and the sensor rack plate is fixed on the transverse sliding platform.
Furthermore, two side edges of the first longitudinal sliding guide rail are provided with arc guide rails which are sunken inwards, and the first longitudinal sliding platform is matched with the corresponding positions of the arc guide rails.
Furthermore, a plurality of guide rail balls are arranged on the first longitudinal sliding platform in a rolling manner at the position matched with the arc guide rail.
Furthermore, a pair of longitudinal linear guide rails is further arranged on the upper surface of the first longitudinal sliding guide rail along the sliding direction of the first longitudinal sliding platform, a pair of strip-shaped protrusions is arranged at the positions, corresponding to the longitudinal linear guide rails, of the first longitudinal sliding platform, and each strip-shaped protrusion is matched with each longitudinal linear guide rail.
Furthermore, a second longitudinal sliding assembly is further arranged on the transverse sliding assembly, the second longitudinal sliding assembly is rotatably connected to the transverse sliding platform through a third lead screw perpendicular to the second lead screw, a third sliding block is in threaded connection with the third lead screw, the thread pitch of the third lead screw is smaller than that of the first lead screw, and the sensor rack plate is fixed on the transverse sliding platform.
Furthermore, driving mechanisms with the same structure are arranged on the first longitudinal sliding assembly, the transverse sliding assembly and the second longitudinal sliding assembly, the driving mechanisms are respectively a first stepping motor, a second stepping motor and a third stepping motor, one end of each of the first screw rod, the second screw rod and the third screw rod is provided with a main bevel gear and a slave bevel gear which are the same in structure, the output shafts of the 3 stepping motors are fixedly connected with the centers of the corresponding main bevel gears, the main bevel gears are meshed with the corresponding slave bevel gears, and the slave bevel gears are sleeved and fixed on the corresponding screw rods.
Furthermore, one end of the first screw rod, one end of the second screw rod and one end of the third screw rod are respectively fixed with a rotating handle.
Furthermore, the reference frame table further comprises a heightening fixed table fixed on the base platform, and a plurality of vertically arranged electric push rods fixed above the heightening fixed table, wherein the reference frame table is arranged at the top end of the electric push rods.
Has the advantages that:
the time series DRNN neural network prediction model is a dynamic regression neural network with feedback and the ability of adapting to time-varying characteristics, the network can directly and vividly reflect the dynamic change performance of the measured value of the measuring sensor and can more accurately predict the actual value of the measured value of the measuring sensor, the time series DRNN neural network model is a 3-layer network structure of 10-21-1, a hidden layer of the time series DRNN neural network model is a regression layer, and an output layer of the time series DRNN neural network prediction model is an output predicted value of the measuring sensor.
The method adopts an ARIMA prediction model to predict the measurement value of the measurement sensor, integrates the original time sequence variables of the factors such as the trend factor, the period factor, the random error and the like of the change of the measurement value of the measurement sensor, converts the non-stationary sequence into the stationary random sequence with zero mean value by the methods such as differential data conversion and the like, and performs data fitting and prediction on the measurement value of the measurement sensor by repeatedly identifying, diagnosing and comparing the model and selecting an ideal model. The method combines the advantages of autoregressive and moving average methods, has the characteristics of no data type constraint and strong applicability, and is a model with good short-term prediction effect on the measured value of the measuring sensor.
And thirdly, the SOM neural network classifier adopted by the invention is a data classification method. The method aims to divide a group of data sets in data spaces such as time sequence DRNN neural network prediction model output, ARIMA prediction model output and time sequence RBF neural network prediction model output into a plurality of subsets according to a similarity criterion, so that each subset of output characteristic normalization parameters of the data sets represents a certain characteristic of the whole data sample set, a SOM neural network classifier is established to classify the time sequence DRNN neural network prediction model output, the ARIMA prediction model output and the time sequence RBF neural network prediction model output characteristic normalization parameters to find reasonable sample subset division, the characteristics of different subsets of root normalization parameters are input into corresponding ANFIS neural network models to predict parameter values of a measurement sensor, and the accuracy of the measured value of the measurement sensor is improved.
The characteristic that a dynamic time sequence DRNN neural network prediction model, a trend prediction ARIMA prediction model and a static network time sequence RBF neural network prediction model are respectively adopted to form a complementary relation to realize simultaneous prediction of the measurement value of the measurement sensor is realized, a SOM neural network classifier is utilized to divide the sample subsets of the normalization parameters of the prediction value of the measurement sensor before an ANFIS neural network model, each subset adopts a corresponding ANFIS neural network model, the method can adopt the corresponding ANFIS neural network models according to the characteristics of each sub-normalization parameter to improve the prediction precision and the operation speed of the ANFIS neural network model, and the prediction method has better fitting precision and generalization capability.
And fifthly, the NARX neural network model adopted by the invention is a dynamic neural network model capable of effectively predicting the nonlinear and non-stationary time sequence of the measured value of the measuring sensor, and the prediction precision of the time sequence of the measured value of the measuring sensor can be improved under the condition that the non-stationary time sequence is reduced. Compared with the traditional prediction model method, the method has the advantages of good effect of processing the non-stationary time sequence, high calculation speed and high accuracy. Through the actual comparison of the experimental data of the measured values of the non-stationary measuring sensors, the method verifies the feasibility of the NARX neural network model for predicting the measured output time sequence of the measuring sensors. Meanwhile, the experimental result also proves that the NARX neural network model is more excellent in non-stationary time series prediction compared with the traditional prediction model.
The invention utilizes the NARX neural network model to establish the dynamic recursive network of the model by introducing the delay module and the output feedback, the output of the ANFIS neural network model is used as the input and the output vector delay feedback of the NARX neural network model is introduced into the network training to form a new input vector, and the NARX neural network model has good nonlinear mapping capability, the input of the NARX neural network model not only comprises the output data of the original ANFIS neural network model, but also comprises the output data of the NARX neural network model after training, the generalization capability of the network is improved, and the NARX neural network model has better prediction precision and self-adapting capability in the time sequence prediction of the measured value of the nonlinear measurement sensor compared with the traditional static neural network.
The time span of the measurement value of the measurement sensor is predicted by adopting a metabolism GM (1,1) gray predictor, the measurement value of a reference object measured by the measurement sensor at the future time can be predicted by using the metabolism GM (1,1) gray predictor according to the historical parameter value of the NARX neural network model, after the measurement value of the measurement sensor predicted by the method is added into an original number sequence output by the NARX neural network model, a data model at the beginning of the number sequence is correspondingly removed for modeling, and then the parameter is predicted. And the like, predicting the measurement value of the measurement sensor. The method is called a metabolism GM (1,1) grey predictor, can realize long-time prediction, and a user can more accurately grasp the change trend of the measuring value of the measuring sensor.
The invention is provided with the longitudinal sliding assembly and the transverse sliding assembly which are matched for use, so that the longitudinal and transverse distances between the measuring sensor and a reference object can be conveniently adjusted, the adjustment can be realized by a stepping motor or a hand-operated rotating handle, and the adjustment can be used according to specific conditions.
In the invention, two side edges of the first longitudinal sliding guide rail provided with the first longitudinal sliding assembly are provided with the inwards concave arc guide rails, and the first longitudinal sliding platform and the arc guide rails are arranged in a matching way at corresponding positions, so that the first longitudinal sliding platform cannot shake or incline when sliding on the first longitudinal sliding guide rail, and the first longitudinal sliding platform is relatively fixed to slide on the first longitudinal sliding guide rail, so that the first longitudinal sliding platform is more stable. The friction force is reduced through the guide rail ball, and the resistance in the sliding process is reduced. The upper surface of the first longitudinal sliding guide rail is also provided with a longitudinal linear guide rail, and the first longitudinal sliding platform is further relatively limited on the first longitudinal sliding guide rail.
The transverse sliding assembly is provided with the second longitudinal sliding assembly, the screw pitch of the screw rod on the second longitudinal sliding assembly is smaller than that of the screw rod on the first longitudinal sliding assembly, the fine adjustment effect is achieved, the adjustment is more accurate, and the distance between the front part and the rear part of the reference object can be adjusted in a small range.
Drawings
FIG. 1 is a block diagram of an MSP430 single-chip microcomputer monitoring unit system according to the present invention;
FIG. 2 is a schematic view of the overall structure of the present invention;
FIG. 3 is a schematic diagram of the overall structure of the sensor fine tuning positioning of the present invention;
FIG. 4 is a schematic view of a connection structure of the first longitudinal sliding platform and the first screw rod according to the present invention;
FIG. 5 is a schematic structural view of a first longitudinal sliding platform according to the present invention.
The system comprises a base platform 1, a sensor rack plate 2, a reference rack platform 3, a heightening fixed platform 101, an electric push rod 102, a first longitudinal sliding guide rail 201, a first longitudinal sliding platform 202, a support 203, a first screw rod 204, a first sliding block 205, a second screw rod 206, a second sliding block 207, a transverse sliding platform 208, an arc guide rail 209, a guide rail ball 210, a longitudinal linear guide rail 211, a strip-shaped bulge 212, a third screw rod 213, a third sliding block 214, a first stepping motor 215, a second stepping motor 217, a third stepping motor 218, a main bevel gear 218, a slave bevel gear 219, a rotary handle 220 and a foot placing groove 221.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The invention discloses a detection device for parameter measurement, which mainly comprises a reference shelf platform 3, a sensor rack plate 2 and a base layer platform 1, wherein the reference shelf platform 3 and the sensor rack plate 2 are arranged on the base layer platform 1, the base layer platform 1 is also provided with a first longitudinal sliding assembly and a transverse sliding assembly, the sensor rack plate 2 is arranged on the transverse sliding assembly, the sensor rack plate 2 slides transversely along with the transverse sliding assembly, and the transverse sliding assembly slides longitudinally along with the first longitudinal sliding assembly. The sensor mounting plate 2 is provided with a foot mounting groove 221 for facilitating fixing of the measuring sensor by a tripod or the like.
The rack table 3 further comprises a fixed base plate 101 fixed on the base platform 1, and a plurality of vertically arranged electric push rods 102 fixed above the fixed base plate 101, wherein the rack table 3 is arranged at the top end of the electric push rods 102.
The detection algorithm part of the detection device mainly comprises an MSP430 single-chip microcomputer monitoring unit, the MSP430 single-chip microcomputer monitoring unit is arranged on the sensor rack plate 2, and the MSP430 single-chip microcomputer monitoring unit is arranged on the sensor rack plate 2 to carry out longitudinal and transverse distance fine adjustment along with the first longitudinal sliding assembly, the transverse sliding assembly and the second longitudinal sliding assembly.
The MSP430 single-chip microcomputer monitoring unit comprises a time sequence DRNN neural network prediction model, an ARIMA prediction model, a time sequence RBF neural network prediction model, a SOM neural network classifier, a plurality of ANFIS neural network models, a plurality of NARX neural network models and a metabolism GM (1,1) gray predictor, and a detection algorithm is designed in the MSP430 single-chip microcomputer monitoring unit, and is shown in fig. 2. The measuring sensor detection algorithm design process comprises the following steps:
1. time series DRNN neural network prediction model design
The time series DRNN neural network prediction model is a dynamic regression neural network with feedback and the capability of adapting to time-varying characteristics, can directly and vividly reflect the dynamic change performance of a measured value of a measuring sensor and can more accurately predict the size of the measured value of the measuring sensor, and is a 3-layer network structure of 10-21-1, and a hidden layer of the time series DRNN neural network prediction model is a regression layer. In the time series DRNN neural network prediction model, the time series measurement value output by the measurement sensor is input into the time series DRNN neural network prediction model
Figure BDA0002648941250000061
Predicting model input vectors for a time series DRNN neural network, wherein Ii(t) is the input of the ith neuron of the time series DRNN neural network prediction model input layer at the t moment, and the output of the jth neuron of the regression layer is Xj(t)，Sj(t) is the sum of the j-th regression neuron inputs, and f (-) is a function of S, then O (t) is the output of the time series DRNN neural network prediction model. The output of the time series DRNN neural network prediction model is:
Figure BDA0002648941250000062
2. ARIMA predictive model design
The ARIMA prediction model is an autoregressive integrated moving average prediction model, which is a modeling method for predicting future measurement values of a measurement sensor based on historical data of the measurement values of the predicted measurement sensor, and analyzes a time series of the measurement values of the predicted measurement sensor. The method adopts the historical parameters of the measured values of the measuring sensor to analyze the time series of the measured values of the measuring sensor, and researches the autoregressive order (p), the difference times (d) and the moving average order (q) of the time series characteristics of the ARIMA prediction model. The ARIMA prediction model is written as: ARIMA (p, d, q). The equations for ARIMA dynamic predictive measurement sensor measurements with p, d, and q as parameters can be expressed as follows:
Figure BDA0002648941250000063
wherein, DeltadytDenotes ytThe sequence after d differential conversions,tis a random error with a variance of a constant σ2Normal distribution of phii(i ═ 1,2, …, p) and θj(j ═ 1,2, …, q) are parameters to be estimated for the ARIMA prediction model, and p and q are orders of the ARIMA dynamic prediction measurement sensor measurement model. The ARIMA prediction model predicts that the measurement value of the measurement sensor belongs to a linear model in nature, and the modeling and prediction comprise 4 steps of (1) sequence stabilization processing. If the measured value historical data sequence of the measuring sensor is not stable, if a certain increasing or decreasing trend exists, the measured value historical data of the measuring sensor needs to be subjected to differential processing. (2) And (5) identifying the model. The orders p, d and q of the ARIMA prediction model prediction measurement sensor measurement value model are determined through the autocorrelation coefficients and the partial autocorrelation coefficients. (3) Estimating parameters of the model and diagnosing the model. Obtaining estimated values of all parameters in an ARIMA prediction model prediction measurement sensor measurement value model by using maximum likelihood estimation, checking the estimated values including parameter significance check and residual error randomness check, judging whether the established measurement sensor measurement value model is available, and predicting the measurement sensor measurement value by using the ARIMA prediction model prediction measurement sensor measurement value model with selected proper parameters; and checks are made in the model to determine if the model is adequate and if not, the parameters are re-estimated. (4) Prediction of the measurement sensor measurements is made using a model with appropriate parameters. The ARIMA module of the time sequence analysis function in the SPSS statistical analysis software package is called by software to realize the whole modeling process of measuring sensor measurement value prediction.
3. Time series RBF neural network prediction model design
Measuring time series of values of sensor outputThe output of the time series RBF neural network prediction model is used as a primary prediction value of a measurement value of the measurement sensor for a period of time. The radial basis vector of the RBF neural network of the time series RBF neural network prediction model is H ═ H1,h2,…,hp]T，hpIs a basis function. The radial basis function commonly used in time series RBF neural networks is a gaussian function, and its expression is:
Figure BDA0002648941250000071
wherein, X is the time sequence value output by the measuring sensor, C is the coordinate vector of the central point of the Gaussian basis function of the hidden layer neuron,jthe width of the Gaussian base function of the jth neuron of the hidden layer; the output connection weight vector of the time series RBF neural network prediction model is wijThe output expression of the time series RBF neural network prediction model is as follows:
Figure BDA0002648941250000072
4. SOM neural network classifier design
The SOM neural network classifier is called a self-organizing feature mapping network, the network is a teacher-free self-organizing and self-learning network consisting of fully-connected neuron arrays, when a neural network receives an external input mode, the neural network is divided into different reaction areas, and each area has different response characteristics to the input mode. The method utilizes a SOM neural network classifier to classify output value samples of a time series DRNN neural network prediction model, an ARIMA prediction model and a time series RBF neural network prediction model, and various sample parameters are input into the corresponding ANFIS neural network models to predict the measurement value of the measurement sensor. The SOM neural network learning algorithm is as follows:
(1) and initializing the connection weight value. And (3) giving smaller weight to the connection weight from the N input neurons to the output neurons, wherein N of the network is 3, and the N is the output value of the time series DRNN neural network prediction model, the ARIMA prediction model and the time series RBF neural network prediction model.
(2) Calculating the Euclidean distance djI.e. the distance between the input sample X and each output neuron j:
Figure BDA0002648941250000081
and calculate a neuron j with the minimum distance*I.e. determining a certain unit k such that for any j there is
Figure BDA0002648941250000083
(3) Modifying output neuron j according to equation (2)*And the weight of its "neighbor neuron":
wij(t+1)＝wij(t)+η[xi(t)-wij(t)] (6)
(4) and calculating and outputting output value samples of the time series DRNN neural network prediction model, the ARIMA prediction model and the time series RBF neural network prediction model according to the following formula.
Figure BDA0002648941250000082
(5) The learning process is repeated by providing new learning samples.
5. Multiple ANFIS neural network model design
The ANFIS neural network model is an Adaptive Fuzzy Inference System ANFIS based on a neural network, also called an Adaptive neural-Fuzzy Inference System (Adaptive neural-Fuzzy Inference System), and organically combines the neural network and the Adaptive Fuzzy Inference System, thereby not only playing the advantages of the neural network and the Adaptive Fuzzy Inference System, but also making up the respective defects. The fuzzy membership function and the fuzzy rule in the ANFIS neural network model are obtained by learning historical data of known predicted values of measured values of a large number of measuring sensors, the input of the ANFIS neural network model is the output values of a time sequence DRNN neural network prediction model, an ARIMA prediction model and a time sequence RBF neural network prediction model, the output of the ANFIS neural network model is the predicted value of the measured values of the measuring sensors, and the ANFIS neural network model mainly comprises the following operation steps:
layer 1: and fuzzifying output values of the input time series DRNN neural network prediction model, the ARIMA prediction model and the time series RBF neural network prediction model.
Layer 2: and realizing rule operation, outputting the applicability of the rule, and multiplying the rule operation of the ANFIS neural network model by adopting multiplication.
Layer 3: the fitness of each rule is normalized.
Layer 4: the transfer function of each node is a linear function and represents a local linear model
Layer 5: the single node of the layer is a fixed node, and the output of the ANFIS neural network model is calculated.
The condition parameters determining the shape of the membership function and the conclusion parameters of the inference rule in the ANFIS neural network model can be trained through a learning process. The parameters are adjusted by an algorithm combining a linear least square estimation algorithm and gradient descent. In each iteration of the ANFIS neural network model, firstly, output values of a time sequence DRNN neural network prediction model, an ARIMA prediction model and a time sequence RBF neural network prediction model are transmitted to a layer 4 along the forward direction of the network, and a least square estimation algorithm is adopted to adjust conclusion parameters; the signal continues to propagate forward along the network to the output layer (i.e., layer 5). The ANFIS neural network model propagates the obtained error signal of the measured value of the measuring sensor along the network in the reverse direction, and updates the condition parameters by a gradient method. By adjusting the given condition parameters in the ANFIS neural network model in this way, the global optimum point of the conclusion parameters can be obtained, so that the dimension of the search space in the gradient method can be reduced, and the convergence rate of the ANFIS neural network model parameters can be increased. The output of the ANFIS neural network model is used as a predictor of measurement sensor measurements.
6. Multiple NARX neural network model design
The inputs of the NARX neural network models are the outputs of the corresponding ANFIS neural network models, the NARX neural network models realize the output prediction of the ANFIS neural network models, and the accuracy of the measurement value prediction of the measurement sensor is further improved. The NARX neural network model (nonlinear-Regression with External input neural network) is a dynamic feedforward neural network, the NARX neural network model is a nonlinear autoregressive network with ANFIS neural network model output as input, the NARX neural network model has a dynamic characteristic of multi-step time delay, and is connected with a plurality of layers of closed networks through NARX neural network model feedback, the NARX neural network model is a dynamic neural network which is most widely applied in a nonlinear dynamic system, and the performance of the NARX neural network model is generally superior to that of a full-Regression neural network. Before application, the delay order and the number of hidden layer neurons of the input and the output are generally determined in advance, and the current output of the NARX neural network model not only depends on the output y (t-n) of the past NARX neural network model, but also depends on the delay order of the current output of the input vector ANFIS neural network model. The NARX neural network model includes an input layer, an output layer, a hidden layer, and a time-cast layer. The ANFIS neural network model output is transmitted to the hidden layer through the time delay layer, the hidden layer processes signals output by the ANFIS neural network model and transmits the processed signals to the output layer, the output layer linearly weights the output signals of the hidden layer to obtain final output signals of the NARX neural network model, and the time delay layer delays signals fed back by the network and signals output by the input layer and transmits the delayed signals to the hidden layer. The NARX neural network model has the characteristics of non-linear mapping capability, good robustness, adaptability and the like. x (t) represents the external input to the NARX neural network, i.e., the output value of the ANFIS neural network model; m represents the delay order of the external input; y (t) is the output of the NARX neural network model, i.e. the output control quantity of the NARX neural network model for the next time period; n is the output delay order; s is the number of hidden layer neurons; the output of the jth implicit element can thus be found as:
Figure BDA0002648941250000091
in the above formula, wjiAs a connection weight between the ith input and the jth implicit neuron, bjIs the bias value of the jth implicit neuron, the value of the output y (t +1) of the NARX neural network model is:
y(t+1)＝f[y(t),y(t-1),…,y(t-n),x(t),x(t-1),…,x(t-m+1)；W] (14)
the input data of the NARX neural network model is the output of the ANFIS neural network model, the output of the NARX neural network model is the measured value of the measuring sensor, the number of the input layer, the number of the output layer and the number of the hidden layers of the NARX neural network model are respectively 1,1 and 10, the NARX neural network model realizes the secondary prediction of the output value of the ANFIS neural network model, and the dynamic performance, the rapidity, the accuracy and the reliability of the measured value of the measuring sensor are improved.
7. Metabolic GM (1,1) Gray predictor design
The metabolism GM (1,1) gray predictor has the advantages of less modeling information, convenient operation and higher modeling precision, thereby having wide application in various prediction fields. The metabolism GM (1,1) grey predictor takes the historical data output by the NARX neural network model as input, and the output is the predicted value measured by the next-stage measuring sensor. The metabolism GM (1,1) gray predictor is a differential equation established after historical data output by the NARX neural network model is generated, and the historical data output by the irregular NARX neural network model is changed into a more regular generation sequence for modeling, so the metabolism GM (1,1) gray predictor is actually a generation sequence model and is generally described by the differential equation. Since the solution of the metabolic GM (1,1) gray predictor is an exponential curve of the solution of the differential equation, it is required that the number series generated is incremental and close to an exponential curve. The measured values of the measuring sensors are positive values, the measured values become an increasing number series after being generated by once accumulation, and the historical data output by the NARX neural network model is set as follows:
x(0)＝(x(0)(1),x(0)(2)…x(0)(n)) (15)
the first generation is as follows:
x(1)＝(x(1)(1),x(1)(2)…x(1)(n)) (16)
for x(1)For a linear differential equation that can establish a variable to the first order as follows:
Figure BDA0002648941250000101
solving the differential equation, and obtaining a predicted value measured by the measuring sensor:
x(0)(k+1)＝x(1)(k+1)-x(1)(k) (18)
the metabolic GM (1,1) gray predictor must be equidistant, adjacent and not have jumps, and the latest NARX neural network model output data is used as a reference point to remove the oldest data prediction value and measure the sensor measurement value at the next stage. The latest NARX neural network model output value can be used for modeling in the measurement sensor measurement value prediction, so that the measurement sensor measurement value of the next phase can be predicted. After the measured value of the measuring sensor at one stage is predicted by the method, the measured value of the measuring sensor is added into the original sequence, a data model at the beginning of the sequence is correspondingly removed, and the prediction of the measured value of the measuring sensor at the next stage in the future is predicted. And so on, predicting the future measurement value of the measurement sensor.
In this embodiment, the first longitudinal sliding assembly includes a first longitudinal sliding guide rail 201 and a first longitudinal sliding platform 202, the first longitudinal sliding guide rail 201 is fixed on the base platform 1 through a bracket 203, a first lead screw 204 is rotatably connected to a lower surface of the first longitudinal sliding guide rail along a length direction of the first longitudinal sliding guide rail, a first slider 205 is connected to the first lead screw 204 in a threaded manner, the first longitudinal sliding platform 202 is sleeved on the first longitudinal sliding guide rail 201 and is fixedly connected to the first slider 205 at a position corresponding to the first lead screw 204, and a structure of the first longitudinal sliding platform 202 is shown in fig. 3 and fig. 4.
The transverse sliding assembly comprises a second lead screw 206 which is perpendicular to the first longitudinal sliding guide rail 201 and is rotatably connected to the upper surface of the first longitudinal sliding platform 202, a second sliding block 207 is connected to the second lead screw 206 in a threaded manner, a transverse sliding platform 208 is fixed on the second sliding block 207, and the sensor rack plate 2 is fixed on the transverse sliding platform 208.
Further, in order to solve the problem that the first longitudinal sliding platform 202 slides on the first longitudinal sliding rail 201 to shake, two side edges of the first longitudinal sliding rail 201 are provided with inwardly recessed arc rails 209, and the first longitudinal sliding platform 202 is arranged in a position corresponding to the arc rails 209 in a matching manner, so that the first longitudinal sliding platform 202 is provided with arc-shaped protrusions protruding toward the arc rails 209 in a position corresponding to the arc rails 209, and thus, the arc rails 209 are arranged in a matching manner with the arc-shaped protrusions, when the first longitudinal sliding platform 202 slides on the first longitudinal sliding rail 201, the upper and lower parts of the first longitudinal sliding platform 202 are limited, the first longitudinal sliding platform 202 cannot slide obliquely, and the first longitudinal sliding platform 202 is difficult to shake.
Further, in order to reduce the friction force between the first longitudinal sliding platform 202 and the first longitudinal sliding guide rail 201, a plurality of guide rail balls 210 are further arranged on the first longitudinal sliding platform 202 in a rolling manner at positions matched with the circular arc guide rail 209. Thus, when the first longitudinal sliding platform 202 slides on the first longitudinal sliding rail 201, the rail ball 210 can reduce the friction between the two.
Further, in order to stabilize the first longitudinal sliding platform 202 more, a pair of longitudinal linear guide rails 211 is further arranged on the upper surface of the first longitudinal sliding guide rail 201 along the sliding direction of the first longitudinal sliding platform 202, a pair of strip-shaped protrusions 212 is arranged at positions of the first longitudinal sliding platform 202 corresponding to the longitudinal linear guide rails 211, each strip-shaped protrusion 212 is arranged in a matching manner with each longitudinal linear guide rail 211, when the first longitudinal sliding platform 202 slides on the first longitudinal sliding guide rail 201, the strip-shaped protrusions 212 on the lower surface of the first longitudinal sliding platform 202 are limited in sliding in the longitudinal linear guide rails 211, the first longitudinal sliding platform 202 is limited in the left-right direction, and no inclination occurs.
Further, when the distance between the measuring sensor and the reference object is adjusted, fine adjustment is needed to be performed on the front-back distance adjustment sometimes, the adjustment pitch of the first longitudinal sliding assembly is large, and it is difficult to achieve the required adjustment accuracy sometimes, so a second longitudinal sliding assembly is further arranged on the transverse sliding assembly, the second longitudinal sliding assembly comprises a third screw rod 213 which is rotatably connected to the transverse sliding platform 208 and perpendicular to the second screw rod 206, a third slide block 214 is connected to the third screw rod 213 in a threaded manner, and the pitch of the third screw rod 213 is smaller than that of the first screw rod 204. The sensor mounting plate 2 is fixed to the lateral sliding platform 208.
Further, in order to realize that the first longitudinal sliding platform 202, the transverse sliding platform 208 and the sensor rack plate 2 slide on the corresponding slide rails to adjust the longitudinal and transverse distances of the measuring sensor, the first longitudinal sliding assembly, the transverse sliding assembly and the second longitudinal sliding assembly are all provided with driving mechanisms with the same structure, the driving mechanisms are respectively a first stepping motor 215, a second stepping motor 216 and a third stepping motor 217, one end of each of the first screw rod 204, the second screw rod 206 and the third screw rod 213 is provided with a main bevel gear 218 and a slave bevel gear 219 with the same structure, 3 output shafts of the stepping motors are all fixedly connected with the center of the corresponding main bevel gear 218, the main bevel gear 218 is meshed with the corresponding slave bevel gear 219, and the slave bevel gear 219 is sleeved and fixed on the corresponding screw rod.
The stepping motor drives the main bevel gear 218 to rotate, the main bevel gear 218 drives the auxiliary bevel gear 219 to rotate so as to realize rotation of the screw rod, the screw rod is in threaded connection with the sliding block, relative position conversion of the sliding block on the screw rod is realized by rotation of the screw rod, and then the first longitudinal sliding platform 202, the transverse sliding platform 208 and the sensor rack plate 2 slide on the corresponding first screw rod 204, the second screw rod 206 and the third screw rod 213.
Further, if the first longitudinal sliding platform 202, the transverse sliding platform 208 and the sensor mounting plate 2 do not slide on the corresponding sliding rails electrically to adjust the longitudinal and transverse distances of the measuring sensor, a rotating handle 220 may be respectively fixed at one end of the first lead screw 204, the second lead screw 206 and the third lead screw 213, and the rotating handle 220 is used to rotate the corresponding lead screw, so as to slide the first longitudinal sliding platform 202, the transverse sliding platform 208 and the sensor mounting plate 2 on the corresponding lead screw.
The working principle is as follows:
firstly, a measuring sensor is fixed on a sensor rack plate 2 through a tripod, a reference object is arranged on a reference object rack table 3, the distance between the measuring sensor and the reference object is adjusted through a first longitudinal sliding assembly, a transverse sliding assembly and a second longitudinal sliding assembly, the measuring sensor is moved to a set position according to a preset requirement, and a first stepping motor 215 is started to drive a first screw rod 204 to rotate so as to realize preliminary adjustment of the longitudinal distance between the reference object and the measuring sensor; then, the second stepping motor 216 is started to drive the second screw rod 204 to rotate, so that the transverse distance between the measuring sensor and the reference object is adjusted; and finally, the third stepping motor 217 is started again to drive the third screw rod 213 to rotate so as to finely adjust the longitudinal distance between the reference object and the measuring sensor, so that the position of the measuring sensor is accurately positioned.
When the measurement sensor measures the parameter value of a certain point of the reference object, the measurement value of the measurement sensor is detected through a detection algorithm in the MSP430 single-chip microcomputer monitoring unit, so that the parameter value measured by the measurement sensor is more approximate to the accurate parameter value of the reference object.
The above embodiments are merely illustrative of the technical concepts and features of the present invention, and the purpose of the embodiments is to enable those skilled in the art to understand the contents of the present invention and implement the present invention, and not to limit the protection scope of the present invention. All equivalent changes and modifications made according to the spirit of the present invention should be covered within the protection scope of the present invention.

Claims (10)

1. A detection device for parameter measurement comprises a reference shelf platform (3), a sensor mounting plate (2) and a base layer platform (1), wherein the reference shelf platform (3) and the sensor mounting plate (2) are arranged on the base layer platform (1), and the detection device is characterized in that a first longitudinal sliding assembly and a transverse sliding assembly are further arranged on the base layer platform (1); the sensor rack plate (2) is arranged on the transverse sliding component, the sensor rack plate (2) slides transversely along with the transverse sliding component, and the transverse sliding component slides longitudinally along with the first longitudinal sliding component; the sensor rack board (2) is also provided with a measuring sensor and an MSP430 single-chip microcomputer monitoring unit, wherein the MSP430 single-chip microcomputer monitoring unit comprises a time sequence DRNN neural network prediction model, an ARIMA prediction model, a time sequence RBF neural network prediction model, a SOM neural network classifier, a plurality of ANFIS neural network models, a plurality of NARX neural network models and a metabolism GM (1,1) gray predictor; the output values of the measuring sensors are respectively used as the input of a time series DRNN neural network prediction model, an ARIMA prediction model and a time series RBF neural network prediction model, the output of the time series DRNN neural network prediction model, the output of the ARIMA prediction model and the output of the time series RBF neural network prediction model are used as the input of the SOM neural network classifier, the SOM neural network classifier outputs a plurality of types of time series DRNN neural network prediction model, the output of the ARIMA prediction model and the output of the time series RBF neural network prediction model are respectively used as the input of a plurality of corresponding ANFIS neural network models, the outputs of the plurality of ANFIS neural network models are respectively used as the input of a plurality of corresponding NARX neural network models, the outputs of the plurality of NARX neural network models are used as the input of a metabolism GM (1,1) gray predictor, and the output of the metabolism GM (1,1) gray predictor is used as the measurement prediction value of the measurement sensor.
2. The detection device for parameter measurement according to claim 1, wherein the first longitudinal sliding assembly includes a first longitudinal sliding guide rail (201) and a first longitudinal sliding platform (202), the first longitudinal sliding guide rail (201) is fixed on the substrate platform (1) through a support (203), a first lead screw (204) is rotatably connected to a lower surface of the first longitudinal sliding guide rail along a length direction of the first longitudinal sliding guide rail, a first slider (205) is connected to the first lead screw (204) in a threaded manner, and the first longitudinal sliding platform (202) is sleeved on the first longitudinal sliding guide rail (201) and is fixedly connected to the first slider (205) at a position corresponding to the first lead screw (204).
3. The detection device for parameter measurement according to claim 2, wherein the lateral sliding assembly comprises a second lead screw (206) perpendicular to the first longitudinal sliding guide rail (201) and rotatably connected to the upper surface of the first longitudinal sliding platform (202), a second sliding block (207) is connected to the second lead screw (206) in a threaded manner, a lateral sliding platform (208) is fixed on the second sliding block (207), and the sensor mounting plate (2) is fixed on the lateral sliding platform (208).
4. The detecting device for parameter measurement according to claim 2, characterized in that two side edges of the first longitudinal sliding guide rail (201) are provided with inwardly recessed arc guide rails (209), and the first longitudinal sliding platform (202) is provided in a matching manner with the corresponding positions of the arc guide rails (209).
5. A detection device for parameter measurement according to claim 4, characterized in that a plurality of guide balls (210) are further arranged on the first longitudinal sliding platform (202) in a rolling manner at the position matched with the circular arc guide rail (209).
6. The detection device for parameter measurement according to claim 2, wherein a pair of longitudinal linear guide rails (211) is further disposed on the upper surface of the first longitudinal sliding guide rail (201) along the sliding direction of the first longitudinal sliding platform (202), a pair of strip-shaped protrusions (212) is disposed on the first longitudinal sliding platform (202) corresponding to the longitudinal linear guide rails (211), and each strip-shaped protrusion (212) is disposed in a matching manner with each longitudinal linear guide rail (211).
7. The detection device for parameter measurement according to claim 3, wherein a second longitudinal sliding assembly is further disposed on the transverse sliding assembly, the second longitudinal sliding assembly includes a third lead screw (213) perpendicular to the second lead screw (206) and rotatably connected to the transverse sliding platform (208), a third sliding block (214) is connected to the third lead screw (213) in a threaded manner, a thread pitch of the third lead screw (213) is smaller than a thread pitch of the first lead screw (204), and the sensor mounting plate (2) is fixed to the transverse sliding platform (208).
8. The detection device for parameter measurement according to claim 7, wherein the first longitudinal sliding assembly, the transverse sliding assembly and the second longitudinal sliding assembly are all provided with driving mechanisms with the same structure, the driving mechanisms are respectively a first stepping motor (215), a second stepping motor (216) and a third stepping motor (217), one ends of the first screw rod (204), the second screw rod (206) and the third screw rod (213) are all provided with a main bevel gear (218) and a secondary bevel gear (219) with the same structure, 3 output shafts of the stepping motors are all fixedly connected with centers of the corresponding main bevel gears (218), the main bevel gears (218) are engaged with the corresponding secondary bevel gears (219), and the secondary bevel gears (219) are sleeved and fixed on the corresponding screw rods.
9. The detecting device for parameter measurement according to claim 7, wherein one end of the first lead screw (204), one end of the second lead screw (206) and one end of the third lead screw (213) are respectively fixed with a rotating handle (220).
10. A sensing apparatus for parameter measurement according to any one of claims 1 to 9, wherein the rack table (3) further comprises a fixed base (101) fixed on the base platform (1), and a plurality of vertically arranged electric push rods (102) fixed above the fixed base (101), wherein the rack table (3) is arranged on top of the electric push rods (102).
CN202010863414.2A 2020-08-25 2020-08-25 Detection device for parameter measurement Pending CN111998887A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010863414.2A CN111998887A (en) 2020-08-25 2020-08-25 Detection device for parameter measurement

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010863414.2A CN111998887A (en) 2020-08-25 2020-08-25 Detection device for parameter measurement

Publications (1)

Publication Number Publication Date
CN111998887A true CN111998887A (en) 2020-11-27

Family

ID=73470816

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010863414.2A Pending CN111998887A (en) 2020-08-25 2020-08-25 Detection device for parameter measurement

Country Status (1)

Country Link
CN (1) CN111998887A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114184839A (en) * 2021-12-10 2022-03-15 江苏集萃智能集成电路设计技术研究所有限公司 ADC frequency spectrum testing method for inhibiting signal source jitter

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103344186A (en) * 2013-07-17 2013-10-09 广西柳工机械股份有限公司 Loading machine unloading parameter testing device and method
CN203337107U (en) * 2013-07-17 2013-12-11 广西柳工机械股份有限公司 Loader unloading parameter testing device
CN108091111A (en) * 2017-11-20 2018-05-29 淮阴工学院 A kind of oil truck oil and gas leakage intelligent early-warning system
CN110119086A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of tomato greenhouse environmental parameter intelligent monitoring device based on ANFIS neural network
CN110119169A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of tomato greenhouse temperature intelligent early warning system based on minimum vector machine
CN110119767A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of cucumber green house temperature intelligent detection device based on LVQ neural network
CN110119766A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of multiple groups close the green pepper greenhouse temperature intelligence prior-warning device of intelligent model
CN111444947A (en) * 2020-03-20 2020-07-24 淮阴工学院 Bridge settlement intelligent detection system

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103344186A (en) * 2013-07-17 2013-10-09 广西柳工机械股份有限公司 Loading machine unloading parameter testing device and method
CN203337107U (en) * 2013-07-17 2013-12-11 广西柳工机械股份有限公司 Loader unloading parameter testing device
CN108091111A (en) * 2017-11-20 2018-05-29 淮阴工学院 A kind of oil truck oil and gas leakage intelligent early-warning system
CN110119086A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of tomato greenhouse environmental parameter intelligent monitoring device based on ANFIS neural network
CN110119169A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of tomato greenhouse temperature intelligent early warning system based on minimum vector machine
CN110119767A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of cucumber green house temperature intelligent detection device based on LVQ neural network
CN110119766A (en) * 2019-04-19 2019-08-13 淮阴工学院 A kind of multiple groups close the green pepper greenhouse temperature intelligence prior-warning device of intelligent model
CN111444947A (en) * 2020-03-20 2020-07-24 淮阴工学院 Bridge settlement intelligent detection system

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114184839A (en) * 2021-12-10 2022-03-15 江苏集萃智能集成电路设计技术研究所有限公司 ADC frequency spectrum testing method for inhibiting signal source jitter
CN114184839B (en) * 2021-12-10 2024-03-01 江苏集萃智能集成电路设计技术研究所有限公司 ADC frequency spectrum testing method for inhibiting signal source jitter

Similar Documents

Publication Publication Date Title
Binu et al. Rider-deep-LSTM network for hybrid distance score-based fault prediction in analog circuits
CN110263479B (en) Atmospheric pollution factor concentration space-time distribution prediction method and system
CN109060001B (en) Multi-working-condition process soft measurement modeling method based on feature transfer learning
CN107451102B (en) Method for predicting concentration of butane at bottom of debutanizer tower based on improved self-training algorithm semi-supervised Gaussian process regression soft measurement modeling
Derks et al. Robustness analysis of radial base function and multi-layered feed-forward neural network models
Casdagli et al. State space reconstruction in the presence of noise
CN107015165B (en) Lithium battery residual life prediction method based on sparse coefficient multi-core correlation vector machine
Dudul Prediction of a Lorenz chaotic attractor using two-layer perceptron neural network
CN112000008B (en) Pipeline monitoring system based on neural network
CN112185104B (en) Traffic big data restoration method based on countermeasure autoencoder
Cheng et al. Estimating the shift size in the process mean with support vector regression and neural networks
CN111998887A (en) Detection device for parameter measurement
CN111079856B (en) Multi-period intermittent process soft measurement modeling method based on CSJITL-RVM
CN110045606A (en) A kind of increment space-time learning method for distributed parameter system line modeling
Smith et al. Online simultaneous semi-parametric dynamics model learning
CN111982041B (en) Displacement detection system for measurement
CN111982042A (en) Displacement detection device for parameter measurement
CN111998885A (en) Parameter calibration system for measurement
CN115062764B (en) Intelligent illuminance adjustment and environmental parameter Internet of things big data system
CN116662925A (en) Industrial process soft measurement method based on weighted sparse neural network
CN114970745B (en) Intelligent security and environment big data system of Internet of things
Zhang et al. An ELM based online soft sensing approach for alumina concentration detection
Vagale et al. Time series forecasting of mobile robot motion sensors using LSTM networks.
Haoguang et al. Study on soft sensing technology of penicillin fermentation based on PLS and SVR
Peng et al. The robust and efficient adaptive normal direction support vector regression

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination