38results about How to "Eliminate redundancy" patented technology

The invention relates to the technical field of data processing, and discloses an industrial one-stop multi-source data acquisition and monitoring system and equipment, and the system comprises a data acquisition module which is used for obtaining multi-source industrial monitoring data and generating a data stream input node; the data preprocessing module is used for processing the input data in the data flow model to generate a standardized data flow; the data conversion module is used for dynamically converting data nodes in the standardized data stream according to different data protocols to generate a compatible data stream; the task scheduling module is used for optimizing the data flow and distributing the data flow to a target node according to the compatible data flow in combination with a node priority and a dynamic scheduling strategy in the data flow model; and the exception detection module is used for carrying out exception state analysis on the optimized data stream and judging whether an exception exists or not according to a preset rule. According to the technical scheme, the compatibility and reliability of the monitoring system can be improved.

The invention discloses an unmanned aerial vehicle path planning method and system based on digital twinning. The method comprises the following steps: constructing a three-dimensional environment model based on first target environment data in first target multi-source environment perception data; constructing a digital twin based on the geometric model, the physical model, the control model, the target physical data hybrid dynamic model and the three-dimensional environment model; inputting the second target multi-source environment sensing data into the digital twin, and dynamically simulating the flight behavior of the physical unmanned aerial vehicle based on the second target multi-source environment sensing data to obtain simulated flight environment state data; and based on the simulated flight environment state data, performing path planning on the physical unmanned aerial vehicle in combination with a prediction control algorithm and a path planning algorithm to obtain a target flight path, transmitting the target flight path to the physical unmanned aerial vehicle, and controlling the flight of the physical unmanned aerial vehicle. The method can dynamically simulate the real-time change of the physical unmanned aerial vehicle, plan the optimal flight path and effectively avoid obstacles, and can be widely applied to the technical field of unmanned aerial vehicle path control.

The invention provides a multi-path embedded video coding method and system based on generative artificial intelligence. The multi-path embedded video coding method comprises the following steps: firstly, coding input cue words and demand parameters, and outputting cue word vectors and demand parameter code streams; then decoding the cue word vector and the demand parameter code stream, and outputting a feature vector and a feature map; extracting the feature vector, the feature map and the representation guide vector to obtain multi-embedded feature information; reconstructing a picture by using the obtained multi-embedded feature information through a generative AI and obtaining a video to be iterated; performing video quality evaluation on the video to be iterated; and generating a binary code stream for transmission through the current cue word vector, the demand parameter code stream and the multi-embedded feature information input. According to the method, important information characteristics are transmitted in a highly abstracted manner, and a high-definition video is reconstructed at a receiving end, so that the storage space required by video coding and the size of a transmission information stream are greatly reduced, and higher-definition transmission under a narrower bandwidth is realized or video coding under an extremely low bandwidth becomes possible.

The invention discloses a multi-object and multi-model deployment and control early warning system. The multi-object multi-model deployment and control early warning system comprises a multi-source data fusion module, an efficient algorithm and model optimization module, a real-time data processing and analysis module, a multi-model cooperation and decision module, an adaptive learning and evolution module and a security and privacy protection module. And a more comprehensive and deep judgment result is provided, so that the police officers can make more accurate decisions in the operation process. The multi-model structure of the system realizes cooperative work through the support of each module, and finally achieves the goals of accurate data, timely analysis and reliable early warning. And due to the close relationship between each module and multiple models, the system has higher decision-making ability and adaptability when facing a complex dynamic environment.

The invention provides a cuff-free continuous blood pressure signal reconstruction method and system, and relates to the technical field of blood pressure monitoring, and the method comprises the steps: obtaining a PPG signal and an ECG signal which are synchronized; respectively extracting a PPG feature vector and an ECG feature vector from the PPG signal and the ECG signal by using the trained attention sequential network, carrying out multi-modal fusion on the PPG feature vector and the ECG feature vector, and reconstructing a cuff-free continuous blood pressure signal based on the fused feature vectors; according to feature vector extraction, an information screening machine and a time sequence learning machine are additionally arranged between an encoder and a decoder and are respectively used for learning a channel probability distribution vector and a time sequence weight vector, guiding the decoder to decode high-level semantic features and decoding the sequential relation between PPG and ECG; the attention sequential network is provided on the basis of multi-modal fusion, and continuous blood pressure signal reconstruction is achieved by analyzing photoelectric volume pulse wave signals and electrocardiosignals.

The invention discloses a financial sharing evaluation system based on big data, relates to the technical field of enterprise management, effectively solves the integration problem of heterogeneous financial data through a data integration module, and ensures the consistency and accuracy of the data. The personalized index module allows user-defined financial evaluation indexes and templates, so that the personalized service capability of the system is improved, and the requirements of specific departments are better met; the intelligent analysis module uses a big data technology and algorithm optimization to collect and analyze data in real time and construct a financial prediction model, so that the intelligent level is improved, and the system can effectively predict and manage complex financial risks; on the whole, the system improves the accuracy and adaptability of financial assessment through high customization and intelligent analysis, and optimizes the risk management and decision-making efficiency of enterprises.

The invention provides a target identification method and system based on multi-source information fusion, and relates to the technical field of low-altitude target detection. The method comprises the following steps: acquiring target radar track data, interception equipment track data and position area information data; based on target radar track data, extracting a first feature in an RCS form dimension, and extracting a second feature in a motion dimension; based on the target radar track data and the track data of the monitoring equipment, determining a frequency spectrum monitoring correlation factor and regional position information features; performing feature fusion on the first feature, the second feature, the spectrum interception correlation factor and the regional position information feature to obtain a target feature, and identifying a target type; and identifying a target threat level based on the target type and the target radar track data. The method and the device are used in a target identification process based on multi-source information fusion, and the technical problem that the target threat degree cannot be accurately identified in a complex environment in the prior art is solved.

The invention relates to a machine learning-based power grid operation data cleaning method and system. The method comprises the following steps of obtaining real-time operation data of a power grid structure; sequentially performing low-dimensional representation and data reconstruction on the real-time operation data by adopting a pre-trained auto-encoder model, and performing screening based on the reconstructed data to obtain abnormal data and normal data; based on the abnormal data, adopting an Apriori algorithm to classify the abnormal data into abnormal state data and dirty data, and removing the dirty data; and adopting a pre-trained GRU recurrent neural network model to complement the abnormal data from which the dirty data is removed to obtain the complemented abnormal data, and completing the data cleaning process. Compared with the prior art, the method has the advantages of realizing efficient and low-threshold power grid operation data cleaning and the like.

The application discloses a tunnel boring machine space dynamic compensation seismic wave advanced detection method and system, comprising: receiving the reflection echo data of the current space position; compensating the reflection echo data according to the attenuation compensation factor of the current space position, wherein the attenuation compensation factor of the current space position is obtained by a trained neural network, and the prediction process comprises: after curve fitting of the attenuation compensation factor of the reflection echo data of the detected space position, the attenuation compensation factor of the next space position is predicted by using the trained neural network; the compensated reflection echo data is subjected to geological body imaging by using the equal travel time imaging method, and the geological body detection result is obtained according to the amplitude difference on the imaging profile. The space dynamic compensation algorithm of the mechanical wave attenuation is used to adaptively correct the attenuation compensation factor at the next tunneling surface, so as to compensate the received reflection echo data and enhance the identification ability of the reflection echo.

The invention belongs to the field of high-speed train fault diagnosis, and discloses a four-quadrant rectifier fault diagnosis method and device based on energy entropy optimization multi-feature fusion, a medium and equipment, and the method comprises the steps: firstly, carrying out the wavelet packet decomposition of a fault signal through employing different wavelet functions, and selecting an optimal frequency band according to the energy-information entropy ratio of a frequency band coefficient; secondly, calculating the energy entropy of each coefficient in the optimal frequency band to obtain a plurality of energy entropy features; and finally, carrying out local tangent space arrangement on the plurality of fault features in a spatial position, optimizing data distribution of the plurality of features, reducing abnormal values, and eliminating redundancy among the plurality of features. The result shows that the average diagnosis accuracy rate of the four-quadrant rectifier is 99.0625% through multi-feature optimization fusion. Compared with other methods, the fault diagnosis method provided by the invention has higher diagnosis rate and stronger robustness for output voltage, noise, training and testing.

The present invention discloses an image encoding and decoding method and device for eliminating spatial redundancy. An adjacent block attention model is set in the original encoding and decoding transformation network. When processing image features, information of adjacent (up, down, left, and right) positions can be introduced, thereby expanding the receptive field, better capturing and eliminating the redundancy of the image in space, and improving the image compression effect.

The invention discloses a fluid pulsation accelerator architecture for reinforcement learning of a large language model, and belongs to the technical field of artificial intelligence hardware acceleration. The system aims at solving the problems that when a current universal processor executes an RLHF working load, an instruction driving normal form is not matched with data flow calculation, fixed parallel granularity cannot adapt to a dynamic load, and the efficiency is low due to the fact that a framework does not sense data statistical characteristics. The core of the system is a liquid systolic array calculation fabric capable of being dynamically reconstructed, execution is triggered through data flow, and kernel boundaries and instruction overhead are eliminated. According to the system, a global lookup table subsystem is fused, and nonlinear calculation is optimized through merging query and parallel lookup by using data distribution prior; and a dynamic scheduling unit is configured, and elastic parallelism is realized by adopting a resource allocation algorithm supporting work stealing. According to the method, the throughput and the energy efficiency of RLHF training and a large language model reasoning stage can be remarkably improved, and a new design normal form is provided for a next-generation AI special computing architecture.

The invention discloses a quantitative grading method for carbon emission in a production stage of a power transformer. The method comprises the following specific steps: after data are collected, firstly calculating a production link carbon footprint and a product use link carbon footprint of a to-be-evaluated transformer product, and calculating a production carbon efficiency ratio of the to-be-evaluated transformer product on the basis of the production link carbon footprint and the product use link carbon footprint; secondly, through data standardization integration processing, calculating the information entropy and the information utility value of each evaluation index, thereby determining the index weight; and finally, calculating the square value of the weighted mahalanobis distance of the to-be-evaluated product, and substituting the square value into a centesimal system grading algorithm to obtain a product production carbon footprint evaluation score and a corresponding grade. According to the method, the transformer quality parameters, the production stage carbon footprint and the use stage carbon footprint are integrated and unified, a traditional carbon efficiency ratio concept is optimized, multi-dimensional and refined grading of the transformer production stage carbon footprint is achieved, and the defects of an existing method in the aspects of comprehensiveness and accuracy are overcome.

This invention discloses a method and system for injecting knowledge into large-scale models for vertical industries, belonging to the field of large-scale model knowledge injection technology. The method involves: collecting professional knowledge data from vertical industries, performing denoising, structured transformation, and terminology standardization; dividing the data into knowledge modules and associating them with vectorized indexes and metadata tags to form a vertical industry knowledge base; parsing the semantic feature vector of the input question using a dynamic routing classifier and routing it to the target knowledge module in the vertical industry knowledge base; employing a hybrid retrieval strategy to recall relevant document sets from the target knowledge module and deduplicate them to generate enhanced retrieval results; based on this, concatenating the enhanced retrieval results with the input question using industry-specific prompt templates; and inputting the LoRA-tuned large-scale model to generate accurate answers conforming to industry standards. This invention solves the problem of insufficient knowledge adaptability of large-scale models in vertical industries, improving the accuracy and applicability of large-scale models in vertical industries.

The invention belongs to the technical field of artificial intelligence, discloses an expert knowledge and data fused intelligent aid decision-making model training method, and aims to solve the problem that expert knowledge and data are difficult to synergistically utilize by an intelligent model, the expert knowledge is converted into a feature and strategy bound decision-making sequence, so that the expert knowledge becomes a data sample; a way is provided for utilizing expert knowledge; a business target state is adopted to evaluate the strategy effect, the expert sample weight and the parameter learning rate are dynamically adjusted according to the model training process, the expert knowledge utilization effect is improved, and meanwhile, new expert samples are dynamically injected in a model evaluation mode; a sample fusion matching mode is adopted to eliminate redundancy of expert samples and data samples, the model training process is accelerated, a sample cluster with a small number of samples is identified through a density clustering mode, the number of similar samples is increased by taking the sample cluster as a starting point, and the robustness of an auxiliary decision-making model is enhanced through continuous improvement of sample richness.

The utility model relates to a power assembly suspension and a vehicle with the same, the power assembly suspension comprises a power assembly suspension body of a claw-shaped structure, the front part of the power assembly suspension body is provided with a plurality of first end face bodies, and one or more first through hole groups are uniformly formed in the first end face bodies; second end face bodies are arranged on the side walls of the rear portion of the power assembly suspension body respectively, and one or more second through hole sets are formed in the second end face bodies. A third end face body is horizontally arranged on one side of the second end face body, and one or more third through hole sets are formed in the third end face body. The first end face body is connected with the second end face body and the third end face body through special-shaped variable cross-section smooth bodies. According to the power assembly suspension and the vehicle with the power assembly suspension, the arrangement structure of parts is compact, the structure is simplified, redundancy is removed, light weight is achieved, machining faces are reduced, and cost is low.

The present invention relates to the field of data analysis technology, and invents a sales big data analysis method based on cross-border e-commerce ERP, including: extracting enterprise resource data from an enterprise management platform, and extracting an inventory resource feature set from the enterprise resource data; performing clustering dimensionality reduction and correlation dimensionality reduction on the inventory resource feature set to obtain a product resource feature set; performing autocorrelation analysis and time series feature extraction on the product resource features to obtain product time series features; forward propagating the product time series features to obtain analysis product data, and extracting unsalable product inventory from the analysis product data; extracting product sales features corresponding to the unsalable product inventory from the enterprise resource data, and performing sales strategy analysis on the unsalable product inventory according to the product sales features to obtain product sales strategies. The present invention also proposes a sales big data analysis system based on cross-border e-commerce ERP. The present invention can improve the accuracy of sales analysis.

The present application relates to a kind of multi-test bench railway vehicle test data integrated management system, belong to railway vehicle test data management technical field.The data acquisition module of the system is based on adaptive matching mechanism obtains multi-test bench data and completes local pre-processing;Data processing module constructs data correlation rule, generates dynamic cross-test bench data correlation atlas, realizes multi-source data fusion and value-driven storage scheduling by hierarchical fusion strategy;Core management module generates test task scheduling scheme and completes priority adjustment, constructs digital twin mapping and generates whole-process traceability link;Interactive collaboration module generates adaptive data display scheme, pushes abnormal early warning and completes field operation interaction.The present application solves the problems of existing multi-test bench data integration, scattered management and poor traceability, realizes the standardization and intelligent control of whole-process railway vehicle test, and improves the test efficiency and data management reliability.

This invention discloses a topology optimization design method for hot-formed steel automotive main frame welded tubes, comprising: establishing an initial model of the automotive main frame welded tube, the initial model including an installation area, a central load-bearing area, a collision energy absorption area, a weld candidate area, an optimizable area, and a non-design area; performing topology optimization on the optimizable area to obtain material density distribution results; dividing the area into high-density, medium-density, and low-density areas based on the material density distribution results, and determining the collision energy absorption area based on the collision deformation trend; converting different areas into stiffeners, locally reinforced structures, variable cross-section transition structures, cross-section reduction areas, and induced crushing structures, respectively; determining the longitudinal weld position based on stress distribution and collision deformation trend; compensating and correcting the structural parameters according to the hot forming process, and outputting a design scheme for the hot-formed steel automotive main frame welded tube. This invention can improve material utilization, weld reliability, and hot-formability manufacturability.

The application discloses a kind of urban resilience spatial combination mode identification method based on space big data mining, belong to wisdom city and urban space planning technical field.This method obtains the spatial element vector data of research area and carries out grid division, extract the spatial element and its characteristics in each grid unit, quantization topological, direction, distance and so on configuration relationship;Using "element-feature-relation" triple to carry out standardization coding, construct spatial transaction dataset;Based on the improved FP-Growth algorithm, potential characteristic factors are mined, and the output typical spatial combination mode characteristic factor set is output by hierarchical merging, spatial continuity correction and spatial instance verification.The system includes data preprocessing, relationship quantization, coding, pattern mining, post-processing and output module.The application can improve the objectivity, reproducibility and interpretability of identification, and provide quantitative support for typhoon disaster prevention, resilience assessment and spatial planning decision.

The application discloses a kind of synergic network perception method and device, it is related to data communication technical field, method includes: each function branch of network perception collects raw data, respectively pre-processes and generates intermediate data, after data conversion and security processing, it is shared to other function branch;Each function branch generates perception result according to receiving and own intermediate data, and the perception result that can be utilized by other function branch is shared;Meanwhile, operation suggestion is generated according to perception result and is executed, and operation execution result is fed back to itself, and is shared to other function branch that produces influence.The application breaks the barrier of each function branch, so that each function branch maintains relatively independent while being able to cooperate with each other.The technical problem that the current dependence single network function branch realizes leads to not comprehensive consideration factor, the operation after perception is not accurate enough.

The invention discloses an electric energy quality data compression method and system based on dimension transformation, and the method comprises the steps: collecting electric energy quality data, calculating an electric energy period based on the time sequence autocorrelation of the electric energy quality data, carrying out the truncation and reorganization processing of the electric energy quality data based on the electric energy period, and obtaining a three-dimensional electric energy quality image. Performing color space conversion on the three-dimensional electric energy quality image to obtain a space conversion image, performing dimension conversion on the space conversion image by adopting a discrete cosine transform algorithm to obtain three-dimensional electric energy image energy distribution data, performing high-frequency energy removal processing on the three-dimensional electric energy image energy distribution data to obtain low-frequency energy distribution data, and performing high-frequency energy removal processing on the low-frequency energy distribution data to obtain low-frequency energy distribution data; and realizing primary compression, and performing secondary compression on the low-frequency energy distribution data by adopting a lossless compression algorithm to obtain power quality compression data. According to the electric energy quality data compression method and system based on dimension transformation provided by the embodiment of the invention, the compression effect on the electric energy quality data is improved.

This application provides an image classification method, training method, apparatus, and device based on a classification model, comprising: acquiring an image to be processed; inputting the image to be processed into a preset classification model; and acquiring a convolutional input vector from the input layer of the classification model; compressing and extracting features from the convolutional input vector based on a downsampling module in the classification model to obtain a downsampling output vector; extracting features from the downsampling output vector based on a normal module in the classification model to obtain an image feature vector; and classifying the image feature vector based on a fully connected layer in the classification model to obtain the image category of the image to be processed. This application reduces the amount of computational data required for feature extraction from the image to be processed, improves the computational speed of feature extraction, and ensures the comprehensiveness and accuracy of feature extraction.

The invention relates to the technical field of database architecture and management, and provides a database tool plug-in routing management method, which comprises the following steps of: dynamically distributing a globally unique namespace identifier for each functional plug-in; adding a namespace prefix for a front-end routing path of each plug-in according to the namespace identifier to realize centralized registration and life cycle management of routing; all application program interface requests are intercepted in a front-end communication link and a rear-end communication link, namespaces of corresponding plug-ins are automatically associated based on request paths, and standardized request and response processing flows including identity authentication and error processing are implemented; based on an event-driven architecture, an event bus supporting plug-in identifier tracking is established, and an event publishing and subscribing mechanism with context information is provided between plug-ins. According to the invention, routing conflicts can be eliminated, stable operation of the system is guaranteed, standardization of a communication process and centralization of management and control are realized, communication efficiency and consistency are improved, and plug-in development and system maintenance costs are greatly reduced.

The application discloses a glacier change prediction method and system, and belongs to the technical field of glacier monitoring; the application is based on a data-driven glacier mass balance model, considers the feedback mechanism of glacier morphological evolution on glacier mass balance and dynamic process, reduces the uncertainty of glacier change prediction results through iterative processing, and further improves the accuracy of future glacier change prediction. On this basis, the glacier mass balance model used by the application comprises a dimension reduction module, an attention mechanism module and a mapping module; the nonlinear mapping relationship between the glacier meteorological parameter sequence, the glacier morphological parameter and the glacier mass balance data is learned through the joint dimension reduction module and the attention mechanism module, the multiscale characteristics of the time series data can be fully mined, and the error of the glacier mass balance estimation is further reduced.

The invention relates to a method for constructing a knowledge graph in a sensing-communication and calculation-control integrated field. The method comprises the following steps: firstly, obtaining PDF and Word data sources from paper, standard and industry libraries, and analyzing and cleaning to generate a preprocessed data set; constructing a domain ontology layer taking sensing and calculation control integration as a root from top to bottom, and synchronously constructing a terminology library and a synonym dictionary; based on an ontology and term system, high-confidence entities are screened through entity boundary probabilities and term weights, and high-reliability entity pairs are obtained through position / type filtering and relation strength quantification; inputting the data and the preprocessed data into a large model, and extracting an initial triple by means of a cue word template; fusing the multi-source entities through the comprehensive similarity to obtain a standardized triple; and finally, importing a graph database, mapping into nodes / edges, and visually displaying, so that the problems of semantic unification and accurate association of cross-domain knowledge are solved.

This invention relates to the field of robot control technology, specifically a multimodal inspection robot swarm control method. The method includes: collecting heterogeneous environmental data such as laser point cloud sequences and visible light video streams through multiple sensing terminals; performing spatiotemporal alignment and feature-level fusion on these data to construct a unified spatiotemporal perception map containing information such as a semantically defined 3D environmental model; based on this map, using an improved distributed model predictive control algorithm with rolling optimization based on an online learning interactive cost function, generating and executing local control command sequences for each robot; and receiving incremental perception data returned by the robots and asynchronously updating the unified spatiotemporal perception map and the interactive cost function. This method improves the comprehensiveness of swarm environmental perception and control accuracy, and enhances the coordination and dynamic adaptability of robot swarm collaborative operations.

This application discloses a method and system for tracking cigarette product production information, relating to the field of tobacco technology. The method implements unique identification for each level of cigarette product through predefined coding rules; industrial IoT devices continuously collect the timestamps, status identifiers, and location coordinates corresponding to the codes to form a raw data stream; a stream processing engine based on Apache Flink performs real-time grouping, time-series alignment, and state merging on the raw data stream, outputting a structured event sequence; a graph database transforms the event sequence into a multi-dimensional relationship graph of coded nodes and process nodes, constructing a semantically complete full-process topology structure; a dynamic programming algorithm traverses the graph, extracting the complete flow path of any code according to the time sequence to generate a production trajectory chain; finally, a force-directed graph layout algorithm spatially rearranges the nodes and edges in the trajectory chain, achieving a natural balance in node distribution through physical simulation, and outputting an interactive visual graphic.

This invention provides a method and system for predicting multimodal drug interaction events based on compact representation learning. The method includes: acquiring multimodal features of each drug in a drug pair to be predicted, wherein the multimodal features include at least two of biological features, molecular structure features, and knowledge graph features; performing intramodal compact representation learning on each modal feature of each drug to eliminate redundant information within a single modality and generate compact sub-representations for each modality; using the compact sub-representations corresponding to the molecular structure features as anchors, performing cross-modal alignment of the compact sub-representations corresponding to the biological features and knowledge graph features through mutual information minimization constraints to eliminate intermodal redundant information; constructing a multimodal fusion representation for each drug based on the compact sub-representations of each modality; concatenating the multimodal fusion representations of the two drugs in the drug pair and inputting them into a classifier to output the prediction result of the interaction event type of the drug pair.