35results about How to "Reduce transfer volume" patented technology

The invention relates to the technical field of big data processing, and discloses a cloud computing and edge computing-based big data co-processing method and a cloud server, which are applied to a co-processing system consisting of an edge node cluster and the cloud server. And obtaining target data through a preprocessing strategy consisting of data filtering, format conversion, increment extraction and lightweight compression. According to the method, the compression ratio larger than or equal to 62% is achieved in combination with the improved LZ4 algorithm, only differential data with service value is transmitted, the invalid data transmission amount is greatly reduced, and the network bandwidth occupation pressure is remarkably reduced; and a dynamic task scheduling link, which is fused with a genetic-reinforcement learning hybrid algorithm of bidirectional LSTM load prediction, can accurately predict edge node computing power fluctuation, dynamically divide task types, match optimal processing nodes, control real-time task response delay within a reasonable range, and realize maximization of computing power resource utilization rate at the same time.

The invention discloses a cloud edge hierarchical collaborative information system integrated resource scheduling system. The system comprises an acquisition module. Data preliminary cleaning and feature extraction are locally executed through the edge end, only key feature data are uploaded, network transmission quantity is reduced, cloud computing pressure is relieved, real-time tasks can respond quickly, task timeout caused by transmission delay is avoided, feature vectors of the edge end and the cloud end are dynamically analyzed based on a scheduling decision model of reinforcement learning, and the real-time performance of the edge end is improved. Task requirements and resource states are intelligently matched, the scheduling accuracy is improved, model parameters are dynamically updated through task execution state data (such as duration and resource utilization rate deviation), the model is adaptively optimized along with environment changes (such as edge end hardware aging and cloud load fluctuation), the efficient scheduling capacity is kept for a long time, and the scheduling efficiency is improved. Layered processing and intelligent scheduling avoid edge end overload or cloud end resource idling, resource allocation of a CPU, a memory and the like is optimized, resource waste is reduced, and the system collaboration efficiency is integrally improved.

The utility model relates to remote monitoring control technical field discloses a kind of remote monitoring control device for EIGA equipment, including total data processing mechanism, the end of total data processing mechanism is provided with probe mechanism, the probe mechanism includes frame component, the surface of frame component is connected with the end of total data processing mechanism, the inside of frame component is provided with self-cleaning component, the self-cleaning component includes outer extension cylinder, the surface of outer extension cylinder is provided with air hole, the inner wall of air hole is fixedly connected with dust baffle, the outer surface of outer extension cylinder is fixedly connected with helical rib, the inner wall of outer extension cylinder is rotatably connected with baffle. In the utility model, self-cleaning design reduces the interference of dust to monitoring, without disassembly cleaning, the layout of multiple probe multiple type monitoring element realizes multi-region multidimensional synchronous monitoring, hierarchical data processing and transmission mode according to demand networking, reduce data transmission volume and total control operation pressure.

The invention relates to a DPU-based fog cloud collaborative Internet of Things intrusion detection system and method. The system comprises an edge node used for receiving target data sent by Internet of Things equipment, encoding the target data through a pre-trained stacked auto-encoder to obtain a plurality of encoding features, determining a prediction capability index of each encoding feature, and determining a target encoding feature according to each prediction capability index; the cloud server is used for receiving the target coding features sent by the edge node, inputting the target coding features into a pre-trained Internet of Things intrusion identification model, extracting target global features, target local features and target time features associated with the target coding features and Internet of Things intrusion through different branch networks in the Internet of Things intrusion identification model, and sending the extracted target global features, target local features and target time features to the edge node; and outputting an Internet of Things intrusion detection result according to the target global feature, the target local feature and the target time feature. By adopting the system, Internet of Things intrusion detection with low resource occupation, high precision and high efficiency can be realized.

The present application relates to a kind of pneumonia detection model training method, specifically to a kind of hierarchical topology pneumonia detection model based on collaborative learning system and method.It overcomes the difficulty of existing pneumonia detection model training system, low efficiency and the problem that local detection node cannot achieve optimal detection performance.First, each node uses local pneumonia data to pre-train respective pneumonia detection model locally, then, center node is based on the feedback data of local model to build hierarchical topology collaborative learning system;After that, each cluster node and child node uses local pneumonia data to train respective pneumonia detection model locally;Finally, collaborative training is carried out, including intra-cluster collaborative training and inter-cluster collaborative training;Intra-cluster collaborative training includes cluster node aggregation training and child node random interactive training.The present application has small data transmission, storage and calculation pressure, and is easy to implement;At the same time, different client is provided with different model parameters, so that each hospital detection node achieves local detection performance maximization.

The present application belongs to the technical field of geophysical exploration, and specifically discloses a noise real-time monitoring method and device for seismic exploration. The device comprises a terminal control module, a wireless transmission module and a noise monitoring module; the noise monitoring module is in communication connection with the terminal control module through the wireless transmission module; the terminal control module calculates the minimum time interval of the data packet sent by the noise monitoring module and configures information for the noise monitoring module; the noise monitoring module collects noise signals, performs root mean square value calculation according to the set window length, performs consistency analysis on the continuous noise signals and multiple signal models, obtains the source type of the noise, and transmits the analysis results of the multi-channel collected noise to the terminal control module; and the terminal control module displays the analysis results sent by the noise monitoring module in real time. The present application can realize real-time monitoring of noise in the whole region and automatic determination of the interference source type, and can perform excitation control determination or disaster warning.

This invention discloses a cross-language network message encapsulation method based on function signatures, belonging to the field of cross-language network message encapsulation technology, including the following steps: S1, function signature parsing; S2, parameter preparation; S3, intelligent serialization; S4, message header assembly; S5, checksum calculation; S6, message assembly and transmission; S7, cross-language parsing. This cross-language network message encapsulation method based on function signatures uses the high 3 bits of the first byte to identify the data type and the low 5 bits to store control information or decimal values, achieving optimal encoding for values ​​in different ranges. It also maintains an independent string cache table for each communication session, with strings as keys and unique integers as indexes. When transmitting a string for the first time, the index and string are transmitted simultaneously; when transmitting the same string subsequently, only the index is transmitted, allowing the receiving end to quickly look up and restore the string, significantly reducing the amount of duplicate strings transmitted.

The embodiment of the application provides a kind of market data processing method, system and computer readable medium, in which according to the characteristics of two different types of data in market data, different ways are used for targeted differential processing, for the transaction information therein, data processing equipment obtains transaction information from data source equipment full amount, writes into database after processing the transaction information to save, for portal equipment to obtain and push to user at any time, and for the order book information therein, it is not full amount acquisition, but the form of subscription is used to establish long connection by portal equipment and the data source equipment, when obtaining the market subscription request of user, it is obtained and pushed from data source equipment through long connection. Therefore, the processing amount and transmission amount of market data as a whole are reduced, the demand for computing resources and bandwidth resources can be reduced, and the cost is reduced.

The invention discloses a wireless federated learning energy consumption optimization method and a medium, and belongs to the technical field of wireless communication, and the method comprises the steps: building a bidirectional model compression frame based on a wireless federated learning system model, and building a joint optimization problem with the minimization of the total energy consumption of the system as a target and a model convergence condition as a constraint; solving the joint optimization problem by adopting an alternating iterative optimization algorithm, and solving a sub-problem of optimal bandwidth allocation and a sub-problem of optimal bidirectional sparseness parameters; and executing compressed global model issuing and local model uploading between the central server and the client according to the solved optimal bandwidth allocation and the optimal bidirectional sparseness parameter. According to the method, the minimization of the total energy consumption of the system is realized on the premise of ensuring the model precision by constructing the bidirectional model compression framework and the bandwidth joint optimization, the method is particularly suitable for large-scale deployment, the communication overhead can be reduced, the model performance of federal learning can be maintained, and the collaborative optimization of the energy efficiency and the precision can be realized.

This invention discloses a hybrid gated loop unit circuit based on memristors. The invention includes a memristor circuit, a weight mapping circuit, a hybrid gate circuit, a candidate activation state circuit, and an output circuit. The signal output terminal of the memristor circuit is connected to the signal input terminal of the weight mapping circuit. Relying on an in-memory computing architecture, it can directly complete the entire computation process of the hybrid gated loop unit and its variant models at the hardware level, eliminating the need for frequent data transfer and interaction. This fundamentally reduces the amount of data transmission, effectively breaking through the bottleneck limitations of data transmission in traditional architectures. It integrates the update gate and reset gate in traditional gated loop units into a single hybrid gate structure. Through the gate value output by the hybrid gate, the retention ratio of historical hidden information can be flexibly and dynamically adjusted, while precisely controlling the degree of fusion between new and historical information, making the information processing of the entire unit more efficient and dynamically adaptable.

This invention relates to a federated learning training system and method based on layered privacy protection. Distributed clients deployed locally by multiple participants perform model training based on local data, outputting encrypted model parameters. A federated coordination center receives and aggregates the encrypted parameters uploaded by the clients, generates and distributes a global model, and the application service layer provides training services to end users based on the aggregated global model. The privacy preprocessing module of the distributed clients implements differentiated privacy protection during the training process based on data sensitivity, and the adaptive privacy budget allocation module of the federated coordination center dynamically adjusts the privacy protection strength of each client. This invention significantly reduces the risk of data leakage, meets compliance requirements, significantly improves the accuracy of personalized recommendations and skills assessment, greatly enhances the coherence and operability of the entire training process, effectively reduces hardware investment costs, breaks down resource barriers between institutions, lays the foundation for long-term cross-institutional collaboration, and continuously iterates and optimizes system performance.

The invention provides a pumped storage power station 5G + BIM intelligent interaction management system, and the system comprises a sensing layer which is used for collecting earthwork vehicle transportation data, dam rolling data, underground cavern construction data and monitoring data, and achieves the dynamic data updating of a BIM model through a 5G camera, a sensor and construction data; the network layer transmits dynamic data of the BIM model through a 5G private network, and allocates communication resources according to service requirements by adopting a network slicing strategy; the data layer is used for fusing dynamic data of the BIM model, including BIM geometric data, sensor real-time parameters and construction data, and constructing a dynamic digital twinborn database; and the application layer performs construction state early warning and construction scheduling based on design parameters and sensor parameters and a BIM model, generates a decision instruction through a deep learning algorithm, and realizes data deep fusion, network resource optimization and emergency capability improvement by fusing 5G high-reliability communication, BIM dynamic modeling and AI intelligent decision.

The application provides a wafer map generation method and system, the method comprising: obtaining test data of a target wafer, the test data comprising coordinate information and test result values of chips on the target wafer; generating a wafer map according to the test data, dividing the wafer map into a plurality of subunits along a horizontal direction or a vertical direction; obtaining attribute data of each subunit, the attribute data comprising coordinate positions, chip numbers and test data of the chips of the subunit; and independently storing the attribute data of each subunit of the wafer map. According to the technical scheme of the application, when data in the wafer map is called subsequently, the data in a certain region (subunit) in the wafer map can be called pertinently, the amount of data transmission is reduced, and the efficiency of calling wafer map data is improved.

The embodiment of the invention discloses a voice data processing method and device. According to the method, voice activity detection is adopted at the equipment terminal to screen out the first audio data stream only containing effective voice segments, and the invalid audio transmission quantity is reduced. According to the method, the problem that the voice frames processed by an equipment terminal are possibly discontinuous due to limited computing power of the equipment terminal is solved, and in order to solve the problem, the method adopts a mode of calculating the mute duration at a server side to determine the truncation point between effective voice segments, so that the completeness of sentences in audio data is ensured, and the processing precision of subsequent services is improved. Through the method, the contradiction between the transmission bandwidth and the processing precision can be effectively balanced. The voice data processing device provided by the embodiment of the invention also has the above beneficial effects.

This invention belongs to the field of mining engineering and addresses the problems of large data transmission volumes, incomplete risk assessments, and delayed early warnings in traditional mine safety management systems. It proposes a new mine safety management system. The data acquisition module collects multi-dimensional raw data, including mine environmental parameters and geomechanical data, through sensors. The edge computing module deploys nodes in the acquisition area to locally clean, denoise, and extract features from the raw data, generating secondary data which is then transmitted via the information transmission module. The data processing module incorporates a risk emergence prediction component using multimodal fusion and dynamic Bayesian networks. It integrates multimodal features and dynamically simulates risk evolution to calculate risk levels. The data response module pushes early warning information to personnel at different levels through multiple channels based on the risk level. This system reduces data transmission volume, improves the comprehensiveness of risk assessment, and enables proactive risk prediction, meeting the needs of real-time mine safety monitoring and efficient management.

The application discloses a transformer state monitoring and data transmission method, system, device and medium based on cloud edge cooperation, belongs to the technical field of power equipment monitoring, and comprises the following steps: collecting transformer data through an edge layer; judging whether matched fault physical characteristics exist in a fault characteristic library; triggering a diagnostic rule engine to obtain a first diagnostic result when the matched fault physical characteristics exist; calling a lightweight diagnostic model to obtain a second diagnostic result; calculating the final probability of various faults according to the first diagnostic result and the second diagnostic result; determining a health state according to the final probability; determining a data transmission level according to the health state; determining data to be transmitted according to the data transmission level; predicting available bandwidth and determining a transmission mode; and transmitting the data to be transmitted from the edge layer to a cloud layer. Through the cloud edge cooperation architecture and the data hierarchical transmission mechanism, the application meets the real-time requirement of data transmission and maintains the fault diagnosis capability when communication is interrupted.

The invention relates to a crane health state real-time monitoring method and system, and the method comprises the steps: obtaining original operation data of a crane, and carrying out the calculation to obtain a time domain feature and a frequency domain feature; a structure including a physical topological graph is constructed, node weights are obtained through calculation according to the physical topological graph structure, feature fusion is carried out on the time domain features, the frequency domain features and the node weights, and a fault diagnosis result is generated in combination with the spectral analysis graph; a life prediction model is constructed, and pre-training and fine tuning are carried out; and inputting the real-time time domain feature and the frequency domain feature into the fine-tuned service life prediction model, and outputting the remaining service life. According to the method, the fault diagnosis precision is improved, the actual crane structure can be better adapted, and the generalization ability of the model is improved.

The application discloses a digital communication signal symbol timing deviation estimation method and device, belonging to the technical field of digital communication, comprising: generating periodic peak points without inter-symbol interference in time domain through frequency domain truncation and equal interval modulus summation mechanism; sinc interpolation up-sampling ensures distortionless reconstruction of band-limited signal and complete reservation of spectral information under low sampling rate. Through specific interval reservation and zero operation on the frequency domain sequence after discrete Fourier transform, periodic reference points without inter-symbol interference are generated in subsequent processing; these reference points only carry current symbol information, eliminating adjacent symbol interference. Through equal interval extraction of points in the discrete sequence after twice frequency domain transformation and summation of modulus values, when the first extraction point position is aligned with the real timing deviation, the sum value reaches the maximum; using the monotone corresponding relationship, high-precision timing deviation estimation can be obtained through simple search, realizing effective estimation of symbol timing deviation.

The application discloses an industrial Internet of Things edge device-oriented communication protocol optimization method and system, relates to the technical field of industrial Internet of Things communication, and aims to solve the technical problems that the existing industrial Internet of Things communication protocol faces network congestion, data transmission delay, state drift and difficulty in coping with burst traffic impact when processing massive data. The method comprises the following steps: determining the running situation in which the edge device is currently located; according to the running situation, individualizing setting a corresponding report triggering condition for at least one target parameter currently monitored by the edge device; based on the report triggering condition after the individualized setting, analyzing continuous sampling values of the at least one target parameter to identify whether there is a to-be-reported event meeting the report triggering condition; in response to identifying the to-be-reported event, generating a simplified report data packet; and in the communication sending queue of the edge device, preferentially scheduling the simplified report data packet for sending.

The invention discloses a system performance simulation verification method based on information data analysis, and relates to the technical field of system performance testing and simulation. The system performance simulation verification method based on information data analysis comprises the six steps of information data collection, preprocessing, simulation model construction, model calibration, performance simulation verification and result analysis, the real-time collection and historical data are combined, a high-precision simulation model is constructed, performance verification is carried out in various load scenes, and the system performance simulation verification efficiency is improved. And finally generating a verification report containing performance bottleneck diagnosis and optimization suggestions. According to the system performance simulation verification method based on information data analysis, the accuracy and comprehensiveness of system performance simulation can be effectively improved, the actual test cost is reduced, and a scientific basis is provided for system optimization.

An encoding device (1570) is provided with a circuit (1571) and a memory (1572) connected to the circuit (1571), and the circuit (1571), during operation, acquires a plurality of networks constituting a three-dimensional data generation model, encodes the plurality of networks, and outputs a bit stream including the plurality of encoded networks.

The utility model discloses a temperature diagnosis system for a cable head, which comprises a current sensor, a wireless temperature sensor, an environment temperature sensor and a signal processing unit, the current sensor is arranged on a protection layer of the cable head to collect cable load current data, and the wireless temperature sensor is arranged at the position of the cable head to measure the temperature of a cable metal layer. The wireless temperature sensor is in signal connection with the current sensor, the current sensor sends a signal to the wireless temperature sensor to wake up the wireless temperature sensor to perform temperature measurement when detecting current abnormity, and the wireless temperature sensor is in a sleep state after sending temperature information to the signal processing unit; the environment temperature sensor is arranged on the outer surface of the cable to measure the current environment temperature, and the signal processing unit is connected with the wireless temperature sensor and the environment temperature sensor to compare the temperature detected by the wireless temperature sensor with the temperature detected by the environment temperature sensor. And when the temperature difference between the two exceeds a temperature threshold, an alarm is triggered.

The invention belongs to the technical field of thermal power plant monitoring, and discloses a fan blade vibration real-time monitoring system and method based on edge computing, which are deployed in a fan site through edge computing nodes and integrate multiple modules to realize data localization real-time processing, greatly reduce data transmission delay, ensure that hidden dangers such as abnormal vibration can be diagnosed in time, and improve the monitoring efficiency. Abnormal vibration and even fracture accidents caused by fatigue load, corrosion and the like of the blade are avoided, and the timeliness and reliability of monitoring are improved. The multi-dimensional sensor network collects vibration signals and auxiliary monitoring signals, and cooperates with the data fusion unit to integrate multi-source data, so that the signal quality is effectively improved, and data support is provided for accurate abnormality diagnosis. The collaborative architecture of the edge computing node and the cloud processing platform enables the edge side to undertake preprocessing, feature extraction and abnormality diagnosis tasks with high real-time requirements, and the cloud end focuses on deep model training, historical data mining and long-term trend analysis, thereby reducing the remote data transmission amount, reducing the bandwidth occupation and transmission cost, and improving the data transmission efficiency. And real-time monitoring and long-term trend control are considered, and the operation performance of the whole monitoring system is obviously optimized.

The invention relates to the technical field of micro-grid scheduling, and discloses a micro-grid dynamic scheduling method and system based on edge calculation and electric fingerprint identification. The method comprises the following steps: collecting current and voltage waveforms of a user total inlet, capturing a transient waveform data segment, and identifying the type and real-time power of an electric appliance; electric appliance types and real-time power are received and aggregated, and a load detail table is constructed; a schedulable load resource pool is constructed based on the load detail list, and the micro-grid central controller calculates a load regulation and control instruction according to the schedulable load resource pool; and the edge computing node selects an electric appliance from the load detail list to execute control and feeds back the cut-down power to the micro-grid central controller. According to the method, transient waveforms are collected in a non-intrusive mode at a user total entrance through the electric fingerprint identification technology, V-I track features, harmonic components and starting transient features are extracted for template matching, accurate identification of each electric appliance type and real-time power is achieved, and the technical problem that a traditional dispatching system can only obtain the total power is solved.

The invention discloses a high slope surface deformation monitoring method based on edge AI and related equipment, and relates to the technical field of civil engineering safety monitoring and geological disaster early warning, and the method comprises the steps: collecting a specific target image in real time, processing the specific target image based on a preset lightweight AI model, obtaining displacement data corresponding to a specific target, and calculating the deformation of the specific target according to the displacement data. And sending the displacement data to a rear-end monitoring platform, so that the rear-end monitoring platform performs real-time monitoring on the slope surface deformation of the high slope based on the displacement data to obtain a monitoring result of the slope surface deformation of the high slope. High slope surface deformation monitoring is completed based on cooperation of the rear-end monitoring platform and the front-end acquisition unit, the front-end acquisition unit only needs to upload displacement data instead of image data, the data transmission amount is greatly reduced, meanwhile, the front-end acquisition unit processes the acquired data in real time through marginalization design, the data processing efficiency is improved, and then the data processing efficiency is improved. In a complex field environment, high-efficiency and high-precision deformation monitoring is carried out.

This application provides a method for live streaming of satellite cloud images, relating to the field of satellite monitoring technology. The method includes the following steps: acquiring multiple cloud images observed by satellite; pre-setting grids of different sizes for the cloud images and classifying the cloud images to obtain cloud images of different resolutions; cropping the cloud images of corresponding resolutions according to the pre-set grids to obtain corresponding slices; grouping each slice according to time and creating video slices; and loading relevant information from the video slices onto the user terminal to display satellite cloud image animations according to the window size and position. This method can deliver high spatiotemporal resolution meteorological satellite cloud images to the user terminal in a timely manner, providing users with low-latency and high-timeliness cloud image services, thereby achieving live streaming of satellite cloud image animations, meeting weather monitoring needs, and providing technical support for meteorological forecasting services.

The application discloses a photovoltaic area fire early warning method and system based on unmanned aerial vehicle multi-spectral imaging and deep learning. The method comprises the following steps: the unmanned aerial vehicle dynamically optimizes a route based on historical warning data to perform cruising collection, and generates a hyperspectral data cube; after being preprocessed, a light three-dimensional convolutional neural network is deployed on an edge to extract spectral spatial features and calculate a spectral reflectivity ratio change rate, and a first fire probability map is output; a suspected fire point is uploaded to the cloud, a second fire probability map is output through a multi-scale feature fusion network; a real fire point confirmation label is generated by filtering based on a screened feature band, combined with a reflection constraint and an activity track; time series spectral data of the real fire point is called, a sliding window anomaly detection is combined with a thermal runaway physical model to map and output a warning level; the cloud and the edge cooperate to generate and send warning information. Through the edge-cloud cooperative architecture and spectral change rate detection, the application realizes early and accurate warning of photovoltaic area fire.

The invention relates to the technical field of network optimization, and discloses an acceleration system and method based on game traffic feature recognition and intelligent path optimization, and the system comprises a traffic feature recognition engine which is deployed at a user client, the sub-flow classification module is used for identifying and classifying sub-flows with different priorities in a game data flow in real time through deep packet detection and a machine learning model; the real-time network probe module is composed of a plurality of acceleration nodes distributed globally, and is used for continuously detecting the network quality among the nodes and constructing a dynamic network quality topological graph; and the intelligent routing decision engine is used for receiving the sub-traffic priority information output by the traffic feature recognition engine and the network quality topological graph output by the real-time network probe module, and dynamically calculating an optimal transmission path according to the sub-traffic priority. According to the acceleration system and method based on game traffic feature recognition and intelligent path optimization, the network delay, jitter and packet loss rate of the game can be remarkably reduced, and the stability and smoothness of the game are improved.

The invention relates to a distribution network line on-line monitoring method and monitoring system based on a multi-hop self-organizing network, and the method comprises the steps: carrying out the pre-analysis of collected data at a node position, and obtaining a pre-analysis result; sending the pre-analysis result to the cloud for analysis; a data transmission link is selected according to the analysis result, the data transmission link comprises a self-organizing network line and a backbone network line, and when the collected data is pre-analyzed at the node position, the distillation network deployed at the node position is used for pre-analyzing the collected data; and when the analysis result is a basic fault, selecting a self-organizing network line and / or a backbone network line to transmit the analysis result, and when the analysis result is a range fault, selecting the backbone network line to transmit the analysis result. According to the distribution network line on-line monitoring method and monitoring system based on the multi-hop self-organizing network disclosed by the invention, the data delay can be reduced by combining a localized analysis mode with dynamic optimization of data transmission.

The application discloses a kind of based on digital pathology section local high-precision magnifying glass system, comprising: main view module, for loading and display digital pathology whole section image;Magnifying glass module, including nested sub-viewport, for specified local area in the current field of view of the main view module is carried out high-power display;Wherein, the magnifying glass module is shared with the main view module by cloning mechanism Image resources;Display area layer, which is overlaid on the display interface of the main view module, is used to visually mark the corresponding position of the local area in the main view module;Control handle layer, integrated in the display area layer, is used to receive user input to interactively adjust the local area, realize local magnification based on original high-resolution data;Non-repeated loading of image resources significantly reduces server and network pressure;Main view and local view are highly consistent, to avoid misdiagnosis.