51 results about "Bayesian estimator" patented technology

The invention discloses a degradation track-based cylinder residual life dynamic prediction method and system. The method comprises the following steps of: preprocessing multi-source degradation original data to obtain a standardized training sample set; quantizing a cylinder performance degradation rate trend by adopting a drift coefficient, quantizing cylinder performance degradation random fluctuation intensity by adopting a diffusion coefficient, and establishing a Wiener process degradation model through the drift coefficient and the diffusion coefficient; substituting the standardized training sample set into the Wiener process degradation model, carrying out parameter self-learning by fusing Bayesian estimation and an EM algorithm, and outputting a final Wiener degradation model; inputting the real-time recorded data into a preset Wiener degradation model to obtain a prediction result of the residual life of the cylinder; dynamically updating the Wiener degradation model according to the prediction result of the residual life of the air cylinder and the real-time recorded data of the air cylinder at the next time step; the online dynamic accurate prediction of the residual life of the cylinder is realized, and the prediction precision and the model environment adaptability are effectively improved.

The invention relates to the technical field of coal mine safety monitoring, and discloses a coal mine tunnel laser radar three-dimensional dynamic monitoring method. Aiming at the problems that in the prior art, data of a laser radar is easy to interrupt or unreliable under a severe working condition, and monitoring continuity is poor due to insufficient multi-source data fusion, the method is characterized in that laser radar surface data and optical strain gauge point data are synchronously acquired; constructing a roadway digital twinborn model; carrying out fusion and assimilation on the multi-source data by utilizing a Transform mechanism to obtain initial roadway deformation data; judging the credibility of the laser radar data based on Bayesian estimation; when the data are missing or uncredible, deducing full-field deformation by utilizing a gating mechanism in combination with optical strain gauge point data and a digital twin model; and finally, generating and outputting a three-dimensional dynamic deformation field. The system is mainly used for real-time, continuous and high-reliability deformation monitoring and safety early warning of the coal mine tunnel.

The invention provides a wide-field magnetic field imaging method based on sparse sampling and Bayesian estimation, and belongs to the field of magnetic field imaging. The method comprises the following steps: constructing an NV color center system, measuring and recording information of a magnetic field to be measured; selecting a small number of discrete sampling points to measure magnetic field intensity data in an imaging field of view of a measured area; the method comprises the following steps of: constructing an error correlation function among pixel points in a measured area, establishing a Bayesian estimation model, and predicting the magnetic field intensity of unmeasured pixel points in a full field of view by taking magnetic field intensity data of a small number of discrete sampling points as observed quantity to obtain a preliminarily reconstructed magnetic field distribution image; selecting a plurality of reference points in the measured area, acquiring nominal magnetic field values and actually measured magnetic field values of the reference points, and calibrating the preliminarily reconstructed magnetic field distribution image by adopting a mean value proportion correction mode; on the premise that the number of sampling points is greatly reduced, magnetic field distribution reconstruction with high spatial resolution and high structural similarity is achieved, and therefore imaging efficiency and reliability are improved.

The invention discloses a noise enhancement nonlinear system joint detection and estimation method under a Bayesian framework, and belongs to the field of signal processing. Firstly, independent additive noise is added to a nonlinear system input signal, and noise-corrected nonlinear system output is obtained after the signal passes through a nonlinear system. And secondly, under the Bayesian criterion, judging which hypothesis in the binary hypotheses is established by utilizing the output of a nonlinear system of noise correction, and estimating unknown parameters in the signal of which the judgment result is H1. On the premise that the detection performance is not reduced, a noise enhancement nonlinear system joint detection and estimation model which minimizes the estimation risk is constructed. The additive noise is the optimal solution of the model and is random distribution formed by not more than two constant vectors. According to the method, noise enhancement and nonlinear system joint detection and estimation under the Bayesian framework are combined, and the Bayesian estimation risk is further reduced under the condition that the Bayesian detection cost is not increased.

The invention discloses an auditable asset allocation system and method based on time-varying risk preference estimation. The system comprises a data acquisition and preprocessing module; the time-varying risk preference estimation module is used for carrying out online Bayesian estimation on the risk preference time-varying hidden state by adopting a sequential Monte Carlo method based on the state space model; the dynamic asset allocation optimization module is used for taking the estimated value as an endogenous variable to be embedded into an optimization model to solve the optimal weight; and the compliance and auditing module is used for recording the full-link log in a tamper-proof manner and generating an interpretable report. According to the method, closed-loop optimization from dynamic risk preference estimation to asset allocation decision is realized, and auditing performance of the whole process is ensured.

The invention relates to a crane intelligent examination method and system based on thermal imaging and a medium in the technical field of crane training. The crane intelligent examination method comprises the steps that infrared heat source data and coordinates of crane structure nodes and skeleton node coordinates of an operator are obtained and processed to obtain a high-dimensional feature vector Z, and a preliminary score is obtained after linear mapping; and processing the X through a fuzzy logic rule and a Bayesian estimation method, and inputting the X and the X into the evaluation neural network to obtain a secondary score. And fusing the preliminary score and the secondary score to obtain a final score, and if the final score is greater than a threshold value, determining that the examination passes. According to the invention, by combining the infrared thermal imaging technology and the skeleton identification technology, dominant behaviors of examinees in operation and hidden influences of the dominant behaviors on equipment can be captured at the same time, and the potential bad operation habits are reflected in final scores, so that the real-time state and the actual capability of an operator can be comprehensively reflected.

The invention provides a high-speed rail 5G signal distribution circuit diagram generation method based on Bayesian estimation positioning, and the method comprises the steps: obtaining target network data, BDS positioning information and LBS ranging information, generating the positioning information of a current circuit library, and obtaining the positioning information of a historical circuit library; predicting first predicted position information at the next moment according to the current line library positioning information, and predicting second predicted position information according to the historical line library positioning information; target prediction position information at the next moment is determined through Bayesian estimation, and the actual position information is corrected into target position information; and generating a high-speed rail 5G signal distribution circuit diagram based on the target network data and the corrected target position information. According to the technical scheme provided by the embodiment of the invention, the historical experience position, the BDS positioning information and the LBS ranging information can be utilized, the next position can be predicted according to the Bayesian estimation to carry out position correction, and the high-speed rail 5G signal distribution circuit diagram can be accurately generated under the condition of reducing the hardware cost.

A signal processing apparatus includes: a data input unit to which image data is input; an output unit configured to output an output value based on the data input to the data input unit; an expectation feedback calculator configured to calculate a difference between an expectation based on the input data and the output value; and a Bayesian estimator to which information on the difference and information based on the image data are input and which is configured to perform machine learning in order to approximate the output value to the expectation based on the input information and to search for an optimum configuration.

Techniques for aerial vehicle tracking using dynamic aleatoric uncertainty covariance estimation are presented. The techniques include: obtaining an image depicting at least one aerial vehicle of interest; passing the image to a first machine learning subsystem, which provides at least one feature vector; inputting the at least one feature vector to a second machine learning subsystem, where the second machine learning subsystem is trained to provide detected aerial vehicle identification data sets (including respective aerial vehicle coordinates, respective aerial vehicle bounding box dimensions, and respective dynamic aleatoric uncertainty covariance values) corresponding to input feature vectors; providing at least one detected aerial vehicle identification data set to a recursive Bayesian estimator subsystem, from which at least one filtered set of aerial vehicle coordinates, representing a real-time location of a respective aerial vehicle of interest, is obtained; and outputting the at least one filtered set of aerial vehicle coordinates.

The present application relates to a kind of wind speed prediction method based on coal mine underground gas data fluctuation similarity, belong to coal mine underground safety monitoring technical field, the method includes: at least 3 adjacent methane sensors of deployment in target roadway, collect gas concentration time series data and carry out pre-processing;Identify gas abnormal fluctuation event, extract the local and time sequence characteristics of fluctuation waveform by CNN-LSTM network, output high-dimensional feature vector;Fluctuation signal of the same source is matched using improved Siamese neural network, and the position time difference of characteristic waveform is calculated by DTW algorithm;Calculate single group wind speed in combination with sensor spacing, and obtain preliminary wind speed by weighted average;Through bayesian estimation, the optimization result of the gas concentration of multiple devices is fused, and then error feedback is realized to realize online optimization of model.The present application does not need to rely on special wind speed equipment, and has high prediction accuracy and strong robustness, can provide reliable data support for underground gas management and ventilation control.

The invention relates to a multi-sensor intelligent refrigerator detection terminal and a data fusion method, and belongs to the technical field of intelligent manufacturing. The multi-sensor intelligent refrigerator detection terminal comprises a camera, a temperature and humidity sensor, a hydrogen sulfide gas sensor, a photosensitive sensor and a core control chip; the core control chip collects image data, temperature and humidity data and hydrogen sulfide gas concentration data measured by the camera, the temperature and humidity sensor and the hydrogen sulfide gas sensor; data output by the hydrogen sulfide gas sensor and the image sensor are collected, abnormal values in the data are removed according to a data abnormal value removing method of Euclidean distance, the processed data are fused by adopting a fusion algorithm, and whether the food freshness is met or not is judged according to the fused data. Abnormal data are removed through a data abnormal value removal method of Euclidean distance to obtain effective data, and then the freshness of food is judged through a multi-sensor data fusion method based on Bayesian estimation, so that the accuracy of food freshness collection by sensors is improved.

The invention discloses aerial vehicle tracking using dynamic accidental uncertainty. Techniques are provided for aerial vehicle tracking using dynamic accidental uncertainty covariance estimation. The techniques include obtaining an image depicting at least one aerial vehicle of interest; passing the image to a first machine learning subsystem, which provides at least one feature vector; inputting the at least one feature vector into a second machine learning subsystem, wherein the second machine learning subsystem is trained to provide a detected aerial vehicle identification data set (including respective aerial vehicle coordinates, respective aerial vehicle bounding box sizes, and respective dynamic accidental uncertainty covariance values) corresponding to the input feature vectors; providing the at least one detected set of aerial vehicle identification data to a recursive Bayesian estimator subsystem from which at least one set of filtered aerial vehicle coordinates representing a real-time location of a corresponding aerial vehicle of interest is obtained; and outputting at least one set of filtered aerial vehicle coordinates.

The application discloses a kind of automobile wheel positioning parameter bayesian estimation system and method, to solve automobile wheel positioning parameter bayesian estimation and need pre-calibration before detection problem.Automobile wheel positioning parameter bayesian estimation system mainly by camera (1), wheel target (2), target connection clamp (3), clamp (4) and tripod (5).Tribrach (5) is placed on horizontal ground, target connection clamp (3) is the section U-shaped cuboid of steel plate processing and two solid and 120-140 degree obtuse angle cylinder welding parts, automobile wheel positioning parameter bayesian estimation method is by automobile wheel positioning parameter bayesian estimation data acquisition, automobile wheel positioning parameter bayesian estimation training and test, automobile wheel positioning parameter bayesian estimation weight vector optimization etc.composition of step, provide a kind of structure simple, detection precision is high, need not pre-calibrated automobile wheel positioning parameter bayesian estimation system and method.

The application discloses a hydrogen energy storage system electrolytic cell selection method for complex multiple working conditions, comprising the following steps: according to existing large number platform data statistics results, determining different renewable energy hydrogen production scene probability distributions of various regions, multiple type electrolytic cell use probabilities and multiple type electrolytic cell use probabilities under different scenes, taking the data as input (prior probability matrix) of an electrolytic cell selection decision system, and then calculating electrolytic cell use probability matrix (posterior probability matrix) under different application service working conditions through a Bayesian estimation method according to the known prior probability. According to the dynamic correction of the Bayesian formula, the use probability of different types of electrolytic cells under different hydrogen production scenes is corrected, the multi-subject matching characteristics of the renewable energy hydrogen production and energy storage system are effectively ensured, the distribution problem of maximum resource utilization and maximum economic benefit in the renewable energy hydrogen production system is solved, and the economic construction of the hydrogen energy storage system applied to multiple types of scenes in the future can be effectively guaranteed.

Systems, methods, and computer program products described herein for building a prediction model with varying availability. Embodiments described herein build a prediction model that is initially trained with a selected subset of data characteristics (e.g., data characteristics that are commonly available in training data), and then augmented via Bayesian estimation with additional data characteristics when such data characteristics are available.

The application discloses a motor insulation residual life prediction method and system, the method comprises the following steps: taking the residual breakdown voltage as the degradation quantity, establishing a motor insulation degradation model based on a Wiener process and an Arrhenius formula, and constructing a motor insulation residual life prediction model based on the concept of first arrival time; replacing the unknown temperature change with an equivalent temperature, setting the model parameters as normal distribution random variables which are independent of each other; obtaining the joint prior distribution of the model parameters by using the maximum likelihood estimation method, and obtaining the posterior distribution approximation solution of the model parameters by using the Bayesian estimation and the Markov chain Monte Carlo method; obtaining the final motor insulation residual life prediction model by using the total probability formula, inputting the motor insulation degradation data of the motor winding under the time-varying temperature stress into the model, and obtaining the predicted motor insulation residual life. The application improves the applicability and flexibility, and makes the life prediction result more personalized and accurate.

The invention discloses a data-driven load model key parameter analysis method supporting trusted simulation analysis. On the basis of extracting multi-type load time-frequency domain features, a mechanism type dynamic feature knowledge base is constructed, a data driving model is converted into an interpretable equivalent load model through approximate Bayesian estimation and an alternating direction multiplier method, and the precision loss of dynamic feature response curves of the two models is compared in the same simulation scene, so that the accuracy of the equivalent load model is improved. Therefore, the simulation credibility of the data-driven load model is improved. According to the invention, the mechanism reconstruction of the data-driven load model is realized. According to the equivalent load model, a load physical mechanism structure is explicitly introduced while the data fitting precision is kept, so that a model result has clear physical meaning and engineering interpretability, and the model is more reasonable in load modeling expression.

The invention discloses a bus type pipe cutting system real-time control method based on digital twinning. The bus type pipe cutting system real-time control method comprises the following steps of 1, collecting and preprocessing multi-dimensional sensor data; 2, performing high-order differential processing based on the standardized sensor data set; 3, performing autoregression modeling according to the differential time sequence data to obtain a global time sequence prediction result; 4, collecting pipe cutting image data, and inputting the pipe cutting image data into the ResNeXt network for packet convolution and residual connection; 5, fusing and inputting the global time sequence prediction result and the cut tube image feature vector into an improved DeepAR model for variational reasoning and Bayesian estimation; 6, adjusting the control parameters according to the posterior distribution of the prediction result, and generating a corresponding control instruction; and 7, monitoring the difference value between the actual operation data and the prediction result, and correcting the control parameters. According to the invention, the ResNeXt network and the improved DeepAR model are fused, and the real-time accurate control of the cutting and management system is realized.

The application discloses a positioning and prediction method of radio signal equal-intensity boundary value based on a mobile vector set, comprising the following steps: constructing an offline map of an indoor scene in a server; dividing the scene into grids, collecting signal intensity values of Bluetooth beacons at the center of each grid, dividing the signal intensity values into intensity value regions, and obtaining intensity boundary lines after numerical fitting; a user carries a wireless device to move in the scene, cuts the intensity boundary lines, and obtains a dynamically updated mobile vector set; at each time, a position estimation region is constructed according to position estimation results of multiple mobile vectors; and multiple position estimation results are fused through Bayesian estimation, that is, the position at the current time is predicted by combining historical data before the current time. The application adopts a strategy of mapping first and then positioning, the user downloads a scene map through the server, and realizes information fusion calculation through a positioning algorithm, so that accurate positioning and prediction of the user are realized.

The invention belongs to the technical field of fault analysis, and particularly relates to a crane fault analysis method based on a knowledge graph, which comprises the following steps of: acquiring sensor data, vehicle conditions, working environments and geographic positions of a crane, inputting the sensor data, the vehicle conditions, the working environments and the geographic positions into a fault diagnosis network model, and obtaining diagnosis and prediction information of crane faults. The fault diagnosis network model construction method comprises the following steps: constructing a fault tree which comprises a top event, a middle event and a bottom event; converting the fault tree into a knowledge graph, mapping a top event into a failure mode entity, and mapping a bottom event into a fault root cause entity; integrating fault case data, and adding a fault case entity, variable entities such as a vehicle condition, an operation environment and a geographic position, and a sensor characterization entity; a naive Bayesian network is constructed after a knowledge graph relationship type is defined, nodes represent various entities and sensor parameters, and edges represent condition dependency relationships; on the basis of historical data discretization variables, Bayesian estimation is adopted to generate a conditional probability table;

The invention discloses a missile life prediction method under the condition of zero failure data, relates to the technical field of missile life prediction, and aims to solve the problem that life distribution fitting cannot be performed by using a traditional distribution method under the condition of zero failure in test data. The method comprises the following steps: collecting missile field detection data, arranging the missile field detection data into regular truncation data, and calculating Bayesian estimation of fault probability of each point in zero-fault truncation data by using a multilayer Bayesian method; selecting a Weibull distribution function as an empirical life distribution function of the missile, and calculating the Bayesian estimation of the fault probability of each point by using a weighted least square method to obtain shape parameter estimation and scale parameter estimation of the Weibull distribution function; and obtaining reliability estimation of the missile according to the shape parameter estimation and the scale parameter estimation of the Weibull distribution function. According to the method, the problem of zero failure existing in regular detection can be effectively solved, and product life distribution is obtained.

This invention provides a method for predicting the spatial distribution of demand for low-altitude take-off and landing facilities in a target area, relating to the field of infrastructure planning technology. The method includes: S1, analyzing and predicting historical demand data for four scenario types based on historical demand data of the target area; S2, using Monte Carlo simulation for parameter calibration and improving the prediction accuracy of the modified exponential growth model and logistic growth model through Bayesian estimation; S3, determining the demand for low-altitude passenger take-off and landing facilities and the demand for low-altitude cargo take-off and landing facilities, and optimizing the total demand for low-altitude take-off and landing stations in the target area; S4, planning the low-altitude infrastructure in the target area to achieve spatial matching between low-altitude take-off and landing facility nodes and demand hotspots. This invention integrates multiple dynamic technological progress models and multi-scenario analysis mechanisms to improve the accuracy and scientific nature of data prediction; it introduces a station reuse coefficient to achieve efficient sharing of passenger and freight facilities and optimize the planning and layout of low-altitude take-off and landing facilities.

Techniques for localizing a vehicle in real time using dynamic uncertainty estimates are presented. The techniques include obtaining a terrain image captured by the vehicle; passing the terrain image to a trained evidential deep learning neural network subsystem, from which a dynamic uncertainty value and a first feature vector are obtained in real time; for each of a plurality of candidate terrain locations, comparing the first feature vector to a respective second feature vector representative of a candidate terrain location, from which a respective similarity score is obtained; for at least one of the plurality of candidate terrain locations, updating in real time, by a recursive Bayesian estimator, a respective location weight based on the dynamic uncertainty value and the respective similarity score; estimating, in real time, a location of the vehicle based on the plurality of location weights; and providing the location of the vehicle.

A kind of ship model parameter online identification method and system based on bayesian estimation belong to ship model parameter identification neighborhood.Proposed to improve the precision of ship model parameter online identification.The technical points: the present application applies bayesian estimation theory to ship model parameter online identification, the posterior probability is obtained by likelihood function and parameter prior, the bayesian estimation result is obtained by maximum principle of posterior probability, then the data is weighted according to condition number and correlation coefficient, the precision of parameter identification is improved, the invention will be applied to real ship, and provide parameter basis for ship motion controller design and optimization.According to known online measurement data section, likelihood function is obtained;(according to bayesian maximum posterior theorem, the posterior probability is obtained by using prior and likelihood function, then the maximum value of posterior probability is obtained, the corresponding bayesian estimation result is obtained;The data is weighted according to condition number and correlation coefficient, the weighted posterior probability function is obtained, and the automatic weighting bayesian estimation method is obtained according to the above steps.

The application discloses a feedback information auxiliary Bayesian DOA estimation method and device based on a weighted atomic norm, and the method comprises the following steps: acquiring output information of a target tracker of a fusion center, and calculating a predicted direction of arrival (DOA) of each target according to the output information; constructing a prior interval of a DOA estimation value of each target according to the predicted DOA; and calculating the DOA estimation value of each target by using the prior interval and an atomic norm algorithm. The application converts the prior knowledge, i.e. the prior interval of the DOA estimation value of each target, into a semi-positive constraint, and then uses a meshless sparse method of atomic norm minimization to perform DOA estimation. The application combines array observation data and prior information, can obtain a high-resolution DOA estimation value, and reduces the calculation cost.

The application relates to a Bayesian DOA estimation method of an array distortion passive synthetic aperture sonar and belongs to the technical field of signal processing. The method comprises the following steps: performing layered probability modeling on array observation data, the model is used for applying a binary prior distribution to a signal vector to contain sparse induction characteristics; a variational Bayesian method is used to iteratively maximize the lower bound of the marginal likelihood function of array distortion parameters and hidden variables; unknown parameters are iteratively updated according to the estimated posterior distribution; and a spatial spectrum diagram is drawn according to the optimal estimation result, and each DOA is determined according to a peak value. The method solves the problems that a traditional passive synthetic aperture direction finding algorithm cannot estimate the direction of a random source and is sensitive to array distortion.

Techniques for aerial vehicle tracking using dynamic aleatoric uncertainty covariance estimation are presented. The techniques include: obtaining an image depicting at least one aerial vehicle of interest; passing the image to a first machine learning subsystem, which provides at least one feature vector; inputting the at least one feature vector to a second machine learning subsystem, where the second machine learning subsystem is trained to provide detected aerial vehicle identification data sets (including respective aerial vehicle coordinates, respective aerial vehicle bounding box dimensions, and respective dynamic aleatoric uncertainty covariance values) corresponding to input feature vectors; providing at least one detected aerial vehicle identification data set to a recursive Bayesian estimator subsystem, from which at least one filtered set of aerial vehicle coordinates, representing a real-time location of a respective aerial vehicle of interest, is obtained; and outputting the at least one filtered set of aerial vehicle coordinates.

This invention discloses an external force estimation and compensation method based on power flow discrimination and Bayesian estimation, relating to the field of precision servo control technology. The main steps include: calculating the motor power flow in real time and intelligently discriminating driving, pullback, and other operating modes based on its sign; employing a dual-model Bayesian estimation architecture to couple the estimation of fast dynamic states and slowly changing system parameters, achieving effective separation of internal and external disturbances; based on the high-confidence estimation results, comprehensively calculating friction, efficiency asymmetry, and inertia compensation terms to form a total feedforward compensation torque; this torque, after undergoing multiple safety constraints such as amplitude, rate of change, and energy integral, is superimposed on the torque command of the main controller in an open-loop feedforward manner, ultimately driving the motor. This invention achieves high-precision compensation for internal friction losses in the motor system and pure estimation of external forces through intelligent mode discrimination based on power flow and dual-model Bayesian collaborative estimation.

The invention provides a photovoltaic power station string fault intelligent diagnosis and arc detection method and system, and relates to the technical field of photovoltaic power station maintenance, and the method comprises the steps: collecting string electrical feature and environment feature data, and carrying out the fault judgment through a deep learning model combining adaptive multi-granularity feature extraction and graph neural network topology fusion; when it is determined that the arc fault exists, a cooperative detection range is determined based on fault feature strength, credibility evaluation is performed by adopting a hierarchical voting mechanism based on a consensus algorithm, and dynamic weighted fusion is performed by utilizing a recursive Bayesian estimation method, so that accurate positioning of the arc fault is realized, and fault detection accuracy and response speed are improved.