23 results about "Deep belief network" patented technology

This invention discloses a rapid near-infrared analysis method for coal quality based on constrained learning and Adaptive Deep Belief Network (ADBN). First, a near-infrared spectrometer is used to acquire diffuse reflectance spectra of coal samples. Each sample is acquired 3-5 times and averaged to reduce random errors, and the true values ​​of corresponding coal quality parameters are determined. Next, the spectral data is preprocessed to eliminate noise, baseline drift, and scattering interference. A constrained learning algorithm is used to select feature bands strongly correlated with coal composition, constructing constrained spectral feature vectors. An Adaptive Deep Belief Network (ADBN) is then constructed. The spectral feature vectors are input into the model, incorporating physical constraints on coal quality parameters, spectral feature correlation constraints, and regularization constraints into the loss function. Finally, the trained model is used to predict the spectral composition of unknown coal samples, obtaining the coal composition content. This method effectively removes redundant bands and noise, solving the redundancy problem of traditional feature selection. ADBN can dynamically adjust the network structure to adapt to the inherent laws of coal quality, improving prediction accuracy and reliability.

The application discloses a low-altitude airspace multi-source data fusion congestion prediction and dynamic scheduling system, which solves the problems of incomplete data fusion, low congestion prediction accuracy, lack of coordination and real-time in existing low-altitude airspace traffic management. The system includes a perception layer, a data fusion layer, a prediction layer, a decision-making and scheduling layer, a control execution layer, and a blockchain storage layer. The perception layer collects multi-source data, the data fusion layer realizes data fusion through an improved deep belief network, the prediction layer adopts a mixed model of LSTM and graph neural network to output congestion prediction results, the decision-making and scheduling layer generates a dynamic scheduling scheme based on a multi-objective optimization algorithm, the control execution layer executes the scheme and feeds back data, and the blockchain storage layer guarantees data security and privacy. The application improves the accuracy of congestion prediction and the adaptability of the scheduling scheme, and realizes efficient, safe and coordinated operation of the low-altitude airspace.

Disclosed in the present invention are a method and system for predictive cruise control of a commercial vehicle. The method comprises: acquiring as input parameters road gradient information provided by a high-precision map and a target cruise vehicle speed corrected by a deep belief network-based driving style model; on the basis of the speed and torque of an engine, using a weighted moving average algorithm to perform averaging processing on power values within a short period of time, so as to generate a current engine power; on the basis of an automobile power balance equation, predicting a driving power demand for a next slope section on the basis of gradient values of a current position and a road ahead; and in view of a high-efficiency output power range of the engine itself, and on the basis of different power demands, generating a recommended gear and an optimal throttle opening for the next slope section. Therefore, in the method for predictive cruise control of a commercial vehicle of the present invention, a gear and a throttle can be adaptively adjusted on the basis of road gradients and driving habits, thereby improving the driving efficiency and fuel economy, and enhancing the personalization of driving experience.

This invention discloses an intelligent fall monitoring device, belonging to the field of fall monitoring technology, aiming to solve the problem of accuracy in fall monitoring and prediction for the elderly. It employs a collaborative artificial intelligence model architecture, integrating two different-dimensional artificial intelligence models for risk assessment. Model AI1, based on fuzzy logic, classifies fall risk levels into five levels: normal, low, medium, high, and emergency by real-time monitoring of vital signs such as heart rate, blood pressure, and blood oxygen saturation. Model AI2, based on deep belief networks, predicts risk by analyzing daily living activity patterns such as sitting, standing, walking, running, and jumping. The prediction results of the two models are input into a meta-model for comprehensive decision-making, generating the final low, medium, and high risk level prediction. For real-time fall monitoring, this invention proposes a hybrid monitoring algorithm, combining rule-based triggering based on acceleration and posture angle thresholds with AI fine-grained classification to achieve low latency and high accuracy.

This invention belongs to the field of intelligent energy supply technology for rail transit, specifically involving an energy allocation method and system for photovoltaic and energy storage systems in rail transit stations. It aims to solve the problems of existing technologies, such as single feature learning dimensions, poor balance in multi-objective optimization, and lagging dynamic feedback response. The invention includes: constructing a multi-agent spatiotemporal correlation matrix; employing a deep belief network with spatiotemporal dual-scale attention fusion for feature learning; establishing a three-dimensional energy flow coupled tensor to mine energy supply and demand relationships; constructing a multi-objective optimization model based on the weighted ideal point method to generate a preliminary strategy; and using adaptive gradient feedback to correct the model and dynamically adjust parameters. Through multiple rounds of iterative optimization, the final energy allocation strategy is obtained. This invention solves the problems of single feature learning dimensions, poor balance in multi-objective optimization, and lagging dynamic feedback response in existing technologies, improving the accuracy of energy supply and demand prediction, and achieving synergistic optimization of energy economy, environmental protection, and equipment reliability.

This invention discloses a method for preparing conductive copper paste with adaptive resistivity control, relating to the field of electronic conductive materials technology. The invention involves in-situ polydopamine coating of copper powder in a multi-level pH gradient buffer system to prepare an antioxidant copper powder premix; simultaneously, it synthesizes thermally responsive conductive microspheres with a core-shell structure; subsequently, the two are compounded and dispersed with an organic carrier to obtain an uncured slurry; microscopic characteristic data of the slurry are collected, and the initial curing process is intelligently predicted using a deep belief network with an attention mechanism and an improved sparrow search algorithm; closed-loop correction is performed through micro-trial and error and an improved particle swarm optimization algorithm to finally determine the optimal curing parameters for curing the main slurry. This invention not only solves the problem of easy oxidation of copper paste but also achieves accurate and adaptive control of the resistivity of a single-formulation slurry, improving batch deviation control rate and product consistency and reliability.

The application relates to a powder fuel mass flow rate prediction model based on physical information and data driving, and relates to the technical field of laser spectrum application. A laser system generates a laser beam, an optical modulation and shaping system modulates the laser beam, a signal detection system suppresses background light noise, a camera system collects a scattering image, an image preprocessing module performs clipping preprocessing, a feature extraction module learns image features through a deep belief network, an XGBoost regression prediction module considers physical constraint parameters, and the learned image features and real-time mass flow rate data are input into an XGBoost model for training calculation to construct a nonlinear mass flow rate prediction model. The scattering image and real-time mass flow rate are used as original data, physical constraint parameters are considered, image features are extracted through a deep belief network, and an XGBoost model is used for training to construct a nonlinear mass flow rate prediction model based on physical information and data driving, so that the accuracy of mass flow rate prediction is improved.

This invention discloses a visual inspection method and system for PTC starter production, relating to the field of automated visual inspection. It involves acquiring original images of the PTC starter to be inspected, enhancing them to generate standardized inspection images, extracting geometric structure and surface texture features based on a multi-scale decomposition algorithm to construct multi-dimensional feature vectors, inputting these vectors into a deep belief network model, and outputting defect classification results and location information. Based on the defect classification results, it generates control commands for qualified release, unqualified rejection, or adaptive adjustment of process parameters. This invention improves the accuracy and intelligence of PTC starter defect detection, standardizes defect judgment, promotes the linkage between inspection and production processes, and adapts to the online quality control needs of large-scale automated PTC starter production.

The application discloses a kind of based on multi-source data fusion and deep optimization network's fault diagnosis method, belong to wind turbine bearing fault diagnosis technical field, including: obtaining the multi-source sensor data of wind turbine under different working conditions, and the multi-source sensor data collected is preprocessed;Design multi-source data feature fusion algorithm based on correlation variance contribution, the multi-source sensor data after pre-processing is fused, and the fuzzy entropy value of the multi-source sensor data after fusion is extracted as the feature vector of input intelligent fault diagnosis model;Intelligent fault diagnosis model DBE based on optimized deep belief network is constructed, and the method and hippocampus optimization algorithm of greedy learning are used to train DBE;Based on the trained DBE, fault diagnosis is carried out.The method solves the problems of signal abnormal value and data missing under the influence of wind turbine variable working condition and external noise interference and fault diagnosis reliability, improves the accuracy of wind turbine fault diagnosis.

This invention relates to the field of distribution network fault detection technology and discloses a method for fault line selection in distribution networks based on SPWVD time-frequency analysis and deep belief networks. The method includes the following steps: collecting fault transient current signals of each line in a new distribution network; preprocessing the fault transient current signals; extracting time-frequency features using the optimal SPWVD time-frequency analysis method and generating a time-frequency spectrum matrix; based on the time-frequency spectrum matrix, using a waveform similarity algorithm to measure the similarity of each candidate line in the distribution network, completing the preliminary identification of the faulty line, and obtaining the preliminary identification result; performing normalization preprocessing on the time-frequency spectrum matrix and inputting it into a trained deep belief network; mining the deep nonlinear fault features of the time-frequency spectrum matrix through the deep belief network and outputting the fault probability distribution of each line to obtain a refined identification result; fusing the preliminary identification result and the refined identification result to determine the faulty line in the new distribution network; this scheme achieves accurate and rapid identification of faulty lines in new distribution networks.

The application discloses a warehouse scheduling decision method and device, a terminal equipment and a storage medium. The warehouse scheduling decision method comprises the following steps: when a cargo delivery request is detected, attribute feature data of a current intensive warehouse system is acquired; the attribute feature data is input into a pre-trained deep belief network model for scheduling decision, so as to generate a scheduling decision scheme, wherein the deep belief network model is composed of one or more of a hoist selection learning model, a shuttle selection learning model and a storage location priority learning model. The application solves the problem that the efficiency of warehouse scheduling operation is low, the utilization rate of equipment is low, and the total operation time of the system is long, and achieves the purpose of timely and efficient scheduling decision of the warehouse system.

The application provides a fault diagnosis method and device, electronic equipment, a storage medium and a program product, and relates to the technical field of power grids. The fault diagnosis method generates evidence weighted fusion data corresponding to the initial multi-source time series data set by performing weighted evidence fusion processing on the initial multi-source time series data set, simplifies the multi-source data fusion process without complex additional signal preprocessing, improves the fusion efficiency of multi-source data, solves the problem of poor accuracy of fixed evidence weight or subjective experience-based allocation in traditional methods, makes the data with high volatility and high correlation have higher priority after weighted evidence fusion, further inputs the evidence weighted fusion data into a deep belief network model for fault diagnosis, improves the accuracy of fault feature extraction, realizes high-precision fault classification, and improves the accuracy of fault diagnosis classification results.

This application provides a dynamic modeling and simulation method and system for the energy efficiency ratio (EER) of a solar-powered seawater desalination system, belonging to the technical field of seawater desalination and system modeling. First, this application acquires data on membrane surface resistance, selective permeability, and DC bus voltage ripple during the photovoltaic electrodialysis process. Second, the ripple data is Fourier transformed and concatenated with membrane parameters to generate an input matrix. This matrix is ​​then imported into a deep belief network, and a restricted Boltzmann machine is used to extract the unsteady-state ion impedance vector reflecting the influence of voltage fluctuations. Subsequently, a nonlinear regression model of this vector and unit water production energy consumption is established using a least-squares support vector machine. Finally, the water production rate and EER are calculated based on the predicted energy consumption and photovoltaic power, generating a dynamic simulation curve. This application can quantify the nonlinear influence of photovoltaic voltage ripple on membrane impedance through deep learning, significantly improving the accuracy of EER prediction for seawater desalination systems under fluctuating power supply conditions.

This invention provides an infrared brain-computer interface method with multimodal signals, belonging to the field of brain-computer interface technology, including the following steps: S1: Data acquisition, using a multimodal sensing system to simultaneously acquire the user's infrared spectral signals, electroencephalogram (EEG) signals, and eye movement signals; S2: Preprocessing the acquired multimodal signals; S3: Constructing a multimodal feature fusion network based on an attention mechanism; S4: Inputting the joint feature vector into a deep belief network for brain intention decoding and outputting corresponding control commands; S5: Based on the deviation between the decoding result and the actual control effect, dynamically adjusting the feature fusion weights and network parameters through a reinforcement learning algorithm to achieve closed-loop adaptive optimization; This invention inputs the joint feature vector into a deep belief network for brain intention decoding, enabling the deep belief network to better learn the mapping relationship between the joint feature vector and brain intention, thus improving the practicality and reliability of the system.

This invention discloses a multi-faceted and refined method for predicting electric vehicle charging load, relating to the field of load prediction technology. The method includes the following steps: Step S1: Constructing a system of influencing factors for electric vehicle charging load and quantifying the time series of each influencing factor; Step S2: Based on the system of influencing factors for electric vehicle charging load, using the information gain method to calculate the correlation between different influencing factors and electric vehicle charging load, and filtering them according to a descending order principle to determine the input factor type; Step S3: Selecting similar days for the electric vehicle charging load of the predicted day using grey relational analysis and DTW distance analysis to determine the input order of the input factors; Step S4: Building a prediction model using a deep belief network, and using the determined input factor type and the input order of the input factors as the input to the prediction model, outputting an effective prediction of the electric vehicle charging load.

The application discloses a signal processing method and system of an S-band narrowband digital processor of a phased array, and the method comprises the following steps: receiving and pre-processing service signals from a comprehensive processor, and converting the service signals into zero intermediate frequency signals; channelizing and decomposing the signals into a plurality of sub-channels with different center frequencies and bandwidths through a multi-phase filter group; performing deep-level feature extraction on the signals of the sub-channels by using a deep belief network which is stacked by a plurality of restricted Boltzmann machines; searching for optimal beam forming weights in a feature space by using a particle swarm optimization algorithm; finally, performing beam synthesis by weighting and summing the signals of the sub-channels according to the obtained optimal weights, and converting the result back into an analog intermediate frequency signal for subsequent transmission processing. The application can improve the processing performance and the accuracy of beam forming of a phased array system in S-band narrowband digital processing.

The application provides a CAN bus intrusion detection method and system based on a deep belief network, and belongs to the technical field of intrusion detection, and comprises the following steps: determining a controller area network (CAN) message rule filter and an initial detection model based on a deep belief network; using different preset training data sets and preset test data sets to train the CAN message rule filter and the initial detection model to obtain an intrusion combination detection model; preprocessing a CAN message to be detected to obtain a pretreated message; inputting the pretreated message into the intrusion combination detection model to output a CAN bus intrusion detection result. In a vehicle-mounted environment, the filter is constructed by using CAN messages, the detection precision is improved, the time cost in the detection process is reduced, the problem that an attack data set is difficult to obtain is overcome, all normal messages without attacks are used for model training, the detection accuracy for mixed attacks is improved, and better false positive rates and precision can be achieved.

The application provides a new energy bus driving range estimation method for working conditions and drivers, which extracts, performs principal component analysis and clustering on operation characteristic parameters of different personnel driving vehicles in sequence to obtain different working condition types which are related to objective factors such as road conditions and environment and subjective factors such as specific driver behaviors and habits, avoids the disadvantages in artificial explanation of principal components, and realizes as comprehensive coverage as possible of real complex working conditions by using machine learning. Through training and online application of a unit distance average energy consumption prediction model based on a deep belief network, the new energy bus vehicle can rapidly identify real-time working conditions according to actual driver and vehicle operation parameters, more accurately estimate the remaining driving range of the vehicle under the current working condition, and update the driving range estimation result as the vehicle position and working condition change.

This invention discloses a data fusion-based intelligent lupus monitoring system and its application in a smartwatch, covering the field of lupus monitoring. The system includes modules for data acquisition, preprocessing, fusion, disease assessment, early warning, storage and management, user interaction, model updating, personalized analysis, and telemedicine. Data is first collected from a smartwatch and medical devices, then denoised, normalized, and features extracted. A deep belief network is used to fuse the data to derive comprehensive health indicators. An LSTM and attention mechanism model is used to assess the disease condition. Threshold-based early warnings are set based on the assessment results and trend predictions. Distributed hash tables and blockchain are used to store and manage the data. A multi-terminal interface is developed, and the model is updated using federated learning and transfer learning. Lupus disease is assessed based on multiple datasets, and 5G communication technology enables remote doctor-patient interaction. This invention monitors the disease condition in real time, provides timely warnings of abnormalities, offers accurate disease assessment, optimizes the allocation of medical resources, promotes the development of telemedicine, and contributes to lupus research.

This invention discloses a nonlinear ultrasonic multi-harmonic multi-feature decoupling identification method and system for microcrack clusters, belonging to the field of ultrasonic nondestructive testing technology. The method includes: S1, acquiring the time-domain nonlinear ultrasonic response signal of the defect caused by multi-parameter feature coupling of the microcrack cluster; S2, data preprocessing, dividing the preprocessed data into training and testing sets; S3, constructing a deep belief network; and S4, importing the acquired microcrack cluster response signal into the deep belief network to establish a multi-harmonic nonlinear ultrasonic decoupling identification model. This invention leverages the powerful feature extraction and signal processing advantages of deep learning, utilizing four nonlinear ultrasonic effects generated by multi-parameter coupling of microcrack clusters to construct a deep belief network, classifying and identifying different distribution center deviations and different average discrete sizes of the microcrack clusters. This addresses the problems of traditional crack identification, which mainly relies on manual labor, resulting in high costs, long processing times, and low reliability.