52results about How to "Reduce time delay" patented technology

The invention relates to the technical field of geological disaster monitoring, in particular to a landslide geological disaster monitoring and early warning system based on multi-means fusion, which comprises a data acquisition module, a data fusion module, a landslide prediction module, a landslide disaster intelligent partition and response module and an early warning transmitting module, the data acquisition module acquires landslide related data in real time; the data fusion module is used for fusing landslide related data from different sources; the landslide prediction module is used for predicting the occurrence probability of the landslide; the landslide disaster intelligent partition and response module combines landslide related data and geologic structures monitored in real time, dynamically adjusts post-disaster response measures, and generates early warning information; and the early warning signal sending module automatically sends an early warning signal based on the generated early warning information. According to the invention, potential risks can be found in time in the early stage of landslide occurrence, and the accuracy and timeliness of disaster early warning are greatly improved.

The invention relates to the technical field of data processing, and discloses an intelligent bridge reinforcement decision generation method, device and equipment based on damage identification. The method comprises the following steps: carrying out bridge damage identification and evaluation on a bridge state data set to obtain a bridge damage characteristic spectrum; performing intelligent decision analysis on the bridge damage characteristic spectrum, generating a preliminary reinforcement scheme, and obtaining a candidate reinforcement strategy set; performing parametric modeling and multi-scale simulation on the candidate reinforcement strategy set, and performing coupling calculation evaluation on a reinforcement effect to obtain an optimized reinforcement scheme; virtual construction planning is conducted on the optimized reinforcement scheme, a construction process model is generated through an intelligent scheduling algorithm, and a target construction scheme is obtained; and monitoring and feeding back the target construction scheme in real time, and dynamically optimizing the construction process through an adaptive control algorithm to obtain a target intelligent reinforcement implementation scheme. The intelligent bridge reinforcement decision generation efficiency and accuracy based on damage identification are improved.

The invention discloses a method and equipment for improving data nonstop efficiency based on a large model, and a medium, belongs to the technical field of large models and data processing, and aims to solve the technical problem of how to reduce time delay of data transmission and processing in a large data processing process, ensure integrity and consistency of data, and improve data nonstop efficiency. According to the technical scheme, the method comprises the steps that multi-source data collection and processing are conducted, specifically, real-time data streams are integrated and synchronized in batches, data are efficiently collected from a database, an API, a file and Internet of Things equipment through a data source adapter, then preprocessing and optimization preprocessing are conducted on the data, and an original data set is obtained; feature engineering: performing feature extraction and feature conversion on the original data set to obtain a pre-training data set; training and optimizing a large model; model deployment and integration: the trained model is exported as a SavedModel format of deployable TensorFlow or a TorchScript format of PyTorch, and real-time data pushing is realized through an API (Application Program Interface) according to a data direct requirement; and a real-time analysis engine.

The invention relates to the technical field of network security analysis, in particular to a big data network signal security early warning system. The system comprises a network signal big data stream transmission module, a network signal ratio difference detection module, a network signal division result safety early warning module and a network signal abnormity safety response module. Data acquisition equipment can be deployed at an entrance and an exit of each network equipment, and network signal data real-time acquisition and big data stream type transmission conversion can be carried out, so that a network signal big data stream can be obtained; carrying out time sequence flow size and access frequency statistical analysis on the network signal big data flow, and carrying out network signal ratio difference detection division and abnormal safety early warning processing at the same time to generate a network signal abnormal safety early warning signal; and uploading the network signal abnormity safety early warning signal to a cloud platform to carry out abnormity safety response processing so as to generate a network signal abnormity safety response protection strategy. According to the invention, network signal abnormity can be efficiently monitored in real time and early warning can be provided.

The invention relates to the technical field of distribution management, in particular to an optimization method formed through secondary feedback of a rider, and the method comprises the steps: S1, building a networking platform, and collecting various information of the rider in real time; s2, a feedback mechanism is established, and a rider feeds back important information encountered in the distribution process in real time; s3, collecting the work data of the rider, and generating a personalized work capability evaluation model; s4, the networked map software collects road information, and establishes a route prediction mechanism in combination with feedback data and historical data; s5, monitoring distribution information of the rider in real time, and providing a processing scheme in case of an emergency; and S6, obtaining distribution information, and learning and optimizing an evaluation model and route planning of the rider. The method is beneficial for greatly improving the flexibility and response speed of the platform to cope with emergencies, ensuring dynamic optimization of the distribution plan of the rider, improving the consumer experience, enhancing the participation sense of the rider, and improving the working enthusiasm and initiative of the rider.

The invention provides a voice recognition method and device, a storage medium and electronic equipment, and the method comprises the steps: calling a target pre-training model, carrying out the feature extraction of target voice data, and obtaining the voice feature of each target voice frame in the target voice data; based on the voice feature of each target voice frame, performing clustering processing on the target voice data to obtain a plurality of clustered voice features so as to determine at least one text recognition result; when the target voice data meets the streaming audio data recognition condition, determining audio data representation features under the target voice data based on the plurality of clustering voice features and the at least one text recognition result; and calling a target attention model, and determining a decoding feature of each text recognition result in the at least one text recognition result based on the audio data representation features under the target voice data, thereby determining a target voice recognition result of the target voice data. According to the embodiment of the invention, voice recognition can be quickly carried out on the target voice data.

The invention discloses an information retrieval method and device, electronic equipment, a storage medium and a computer program product, and the method comprises the steps: receiving a first query statement input by a user; judging whether a first list contains the first query statement or not to obtain a judgment result; the first list comprises one or more query statements, and each query statement in the first list is associated with zero or one or more answers in a set database; when the judgment result represents that the first list contains the first query statement and the first query statement is associated with one or more answers in the set database, randomly selecting a first answer from the one or more answers, and calculating a first score based on user interaction information of the first answer; and when the first score is greater than a first score threshold value, outputting the first answer as a retrieval result of the first query statement.

The invention relates to the field of gas pipe network leakage early warning, in particular to a gas pipe network leakage early warning method based on edge calculation, which is applied to an equipment terminal, and comprises the following steps: responding to a leakage monitoring request, and receiving pipe network current data acquired by an acquisition end, the pipe network current data comprising pressure current data and leakage current data; calling a preset screening model to process the pipe network current data to obtain abnormal deviation values, and obtaining effective pipe network data based on the abnormal deviation values and the corresponding data fluctuation thresholds; receiving pipe network annotation data, obtaining hidden layer output values of the pipe network annotation data according to the initial evaluation model, and determining a state output weight based on each hidden layer output value to obtain a pipe network state evaluation model; and inputting the effective pipe network data into the pipe network state evaluation model for evaluation, generating a state evaluation result, processing an abnormal state evaluation result based on a preset early warning strategy, generating leakage evaluation data, and sending the leakage evaluation data to a data center.

The invention belongs to the technical field of textile traceability, and particularly relates to a textile traceability method and system based on a block chain, and the method comprises the steps: analyzing the textile data integrity based on a block chain network, judging whether the textile data is complete or not, analyzing the textile traceability access confidence, and judging whether the textile traceability access behavior is credible or not. And analyzing the textile traceability query result, comparing to obtain textile traceability data, analyzing the textile traceability data state, and comparing to obtain a textile traceability data risk point. According to the method, the problems of inaccurate and missing data possibly existing in traditional textile data traceability related information and incomplete judgment of the textile data traceability access verification state are solved, the whole production resume of a product can be restored more accurately, risk points in a supply chain can be found and positioned favorably, and the quality of the product is improved. And a basis is provided for improvement.

The invention relates to a website marketing method based on intelligent matching geographic data and newly-opened enterprise information, which comprises the following steps: acquiring detailed information of a business website, creating a basic snapshot information base, and acquiring and preprocessing the newly-opened enterprise information at regular time; converting the website position and the enterprise registration address into latitude and longitude coordinates, and based on a preset matching distance range, screening out a matching website set meeting conditions by using a Haversine formula; performing clustering on stock enterprises corresponding to the matching network point set through a K-Means clustering algorithm, and matching a clustering center in combination with the characteristics of a new enterprise to obtain an optimal network point list; presetting website scoring indexes, calculating website comprehensive scores according to the website scoring indexes, and selecting the website with the highest score as an optimal matching website; and pushing the information of the new enterprise to a customer manager corresponding to the optimal matching network to carry out marketing activities, and if the information is returned, carrying out matching again. According to the invention, the accuracy and efficiency of website marketing are improved, and the service experience of newly opened enterprises is optimized.

The invention provides a model reasoning method and device, a storage medium and computer equipment. The method comprises the following steps: acquiring a pre-stored allocation table and an address recovery table of a target model; wherein the allocation table is a target model used for recording an execution sequence corresponding to nodes in a computational graph captured by the target model in a preprocessing process, and the address recovery table is a mapping relation table of a kernel and parameters constructed according to the computational graph; according to the execution sequence in the allocation table, an address recovery table is inquired, address information of nodes in the computational graph is obtained, and the address information comprises a kernel address and a parameter address of each node; according to the address information of each node, recovering a computational graph of the target model in the reasoning process; and based on the calculation graph, executing reasoning calculation of the target model. Therefore, the model can be inferred according to the computational graph, the operator graph is prevented from being recaptured when the model is inferred, the time delay of cold start of the model is reduced, and the cold start efficiency of the model is improved.

The embodiment of the invention discloses a response method and device for an interrupt signal of a virtual machine, a terminal and a storage medium. The method comprises the following steps: in response to a target trigger operation of external equipment of a virtual machine, obtaining target interrupt information matched with the target trigger operation from a pre-configured interrupt information set, and determining a target virtual processor for processing the target trigger operation; sending a target interrupt signal to the target virtual processor according to the target interrupt information through a target message interrupt controller matched with the target virtual processor; the target interrupt signal has a corresponding first physical address. And determining a second physical address matched with the first physical address according to pre-configured address mapping, and executing the target interrupt signal based on the second physical address through the physical processor to complete interrupt response of the target trigger operation. Through the embodiment of the invention, frequent operation of the virtual machine monitor is avoided, the system overhead is reduced, and the operation efficiency of the virtual machine is improved.

The present invention relates to a method for rapid stroke identification combining TOPSIS and artificial intelligence. First, the existence of unilateral muscle strength decline is judged by the degree of horizontal raising of the user's arms, and the muscle strength difference is graded by the angle between the arms; then the degree of slurred speech of the user is analyzed by artificial intelligence speech recognition and combined with the error word rate calculation method; then the position coordinates of the facial feature points are detected by the artificial intelligence facial feature point recognition model, the degree of facial droop is calculated by the TOPSIS method, and finally the risk degree of stroke is rapidly identified by the TOPSIS method by comprehensively grading the muscle strength difference, the degree of slurred speech and the degree of facial droop. The present invention can enable patients and their families to effectively and quickly identify stroke symptoms at home, and automatically shunt stroke patients to corresponding stroke centers for treatment according to different severity of stroke, thereby improving the emergency rescue efficiency and quality of emergency medical rescue services.

The invention provides a wireless multi-source data fusion method for multi-component mixing spray gun control, which comprises the following steps of: constructing a time sequence vector V, carrying out dimension raising processing and preprocessing on the time sequence vector, sending a signal to a wireless channel, superposing modulation signals of different sensor nodes to obtain a superposed signal, and fusing the superposed signal into a multi-component mixing spray gun; the method comprises the following steps of: acquiring a superposed signal, constructing a mean square error according to the superposed signal and a preset spraying decision function, constructing and solving an optimization problem to obtain an optimal sensor emission signal, sending the optimal sensor emission signal to a decision controller, and directly performing a spraying control decision according to the received signal by the decision controller to control spraying parameters of a multi-component mixing spray gun; according to the spray gun control system, dimension raising processing and preprocessing are conducted on the signals collected by the sensors, the preprocessed signals are directly transmitted to the decision controller, the decision controller directly receives the decision result, the spray gun is controlled more rapidly, and the control precision and effect of the spray gun are improved.

The application relates to a vehicle-end data processing method, a vehicle-end device and a vehicle-end data processing system. The method comprises the following steps: obtaining a target communication message in a communication line during vehicle operation; in the case that a current computing task is received from a cloud server, converting the current computing task into a current processing instruction; processing the target communication message according to the current processing instruction to obtain a current computing result, and uploading the current computing result to the cloud server. The method can realize real-time processing of the target communication message, reduce the time delay of data transmission to the cloud for processing, timely monitor abnormal conditions of the vehicle state, avoid leakage of vehicle-end privacy information, and ensure data security.

The invention provides a ship berthing and unberthing method and device for a wharf and relates to the technical field of wharfs, and the ship berthing and unberthing method for the wharf comprises the steps that when a ship berths in the first direction, whether the first distance between the bow of the ship and a second trestle meets the preset safety distance or not is judged, and if yes, the first trestle is stopped; controlling the ship to enter a berth water area between the first trestle and the second trestle through a reverse cable for connecting the ship with the first abutment and a ship adjusting device; when the ship leaves the berth, the bow is connected with the second abutment through a reverse cable, and the ship is controlled to retreat; and when the ship is driven out of the berth water area by a preset second distance, the stern of the ship is connected with the first abutment through an inverted cable, and the ship is controlled to rotate out of the berth water area through a ship adjusting device. The ship berthing and unberthing efficiency of a wharf can be effectively improved, and the safety risk of berthing and unberthing of a large ship in a harbor basin with a narrow space is reduced.

This invention relates to the field of high-altitude work equipment technology and discloses a method for reconstructing the flow field in a supersonic isolation section by integrating CNN and FCNN. The method includes the following steps: acquiring wall pressure data and schlieren image sequences in a wind tunnel test system; using synchronous light signals for time alignment and preprocessing; constructing a training dataset containing normalized pressure feature vectors and a matrix of real flow field schlieren images; constructing a parallel dual-branch deep learning model; using a convolutional neural network branch to reshape the pressure data into a matrix to extract spatial structure features; and using a fully connected neural network branch to directly extract numerical correlation features. The outputs of both are merged in a feature fusion layer. After optimizing the model parameters using the training dataset, the real-time acquired wall pressure is input into the trained model, and the reconstructed flow field grayscale image is output. This invention solves the problem of difficulty in reconstructing complex flow fields from sparse pressure data and achieves high-precision real-time monitoring of the flow field in a supersonic isolation section.

The present invention discloses a lightweight remote sensing image small target detection method; through a lightweight convolution module LFC, the amount of parameters and the amount of calculation required for the convolution operation are reduced to complete feature extraction and downsampling under the premise of limited computing resources and storage space; a receptive field expansion module RFEB is adopted to expand the receptive field of the backbone network through multi-branch dilated convolution, and make full use of local and global context information to solve the problem of insufficient feature representation of small targets in remote sensing images; a hierarchical fusion attention module HFAB is used to achieve feature fusion from coarse-grained to fine-grained by progressively stacking spatial attention and self-attention, so that the feature representation of small targets is richer; the method provided by the present invention reduces the amount of model parameters and calculation overhead while maintaining the target detection accuracy, effectively reduces the calculation complexity, and improves the feasibility of the target detection algorithm in edge computing and real-time monitoring scenarios.

The invention provides a multi-source data fusion report generation method and system, and relates to the technical field of data processing, and the method comprises the steps: extracting a first filling lattice in a first report demand, the first filling lattice having a first filling logic; obtaining a first filling parameter of the first filling lattice after the expiration date, and introducing a multi-layer verification mechanism for verification in combination with the first filling logic to obtain a first verification result; when the first verification result reaches a first predetermined constraint, generating a first sequence based on a first filling parameter, the first sequence including a first transverse sequence and a first longitudinal sequence; a multi-parallel verification mechanism is introduced to verify the first transverse sequence or the first longitudinal sequence, and a second verification result is obtained; and when the second verification result reaches a second predetermined constraint, generating a target report. The technical problems that in the prior art, the data quality and accuracy of different data sources are difficult to unify, data verification is insufficient, and then the accuracy and effectiveness of generated reports are poor are solved.

The application relates to the technical field of financial informatization, and discloses a financial supervision rule dynamic distillation method, device, medium and equipment, which comprises the following steps: preprocessing unstructured rule text, and converting the unstructured rule text into structured training samples based on a logic tree annotation system; generating teacher output results based on the structured training samples, and performing distillation training on a student model by using the structured training samples and the teacher output results to obtain a distillation model; performing constraint decoding on the output process of the distillation model to generate candidate structured executable rules, and performing multidimensional effectiveness verification to determine target structured executable rules, which are deployed into a rule engine of a target business system. The above method forms a complete automatic closed loop from rule text acquisition, structured conversion, model distillation to rule generation and deployment, realizes instant analysis of unstructured rule text, structured extraction and automatic deployment, shortens a rule updating cycle, and reduces maintenance cost.

The invention discloses a parenteral nursing medicine formula optimization system based on data fusion, and relates to the technical field of parenteral nursing. Constructing an infusion amount change curve of different preparation components, performing time sequence alignment on the target physiological parameters and the infusion amount change curve, extracting a fusion feature vector, and inputting the fusion feature vector into a patient metabolism model to output a deviation index, extracting local fusion feature vectors of different preparation components, combining the local fusion feature vectors with the patient metabolism model to output a contribution index, generating an evaluation result according to the deviation index and the contribution index, updating model parameters in the patient metabolism model, and generating formula optimization suggestions; multi-source data can be integrated, the metabolic state of a patient can be monitored in real time, a medication formula is dynamically optimized, and the safety and effectiveness of parenteral nursing are improved.

Embodiments of the present application provide a vehicle-cloud collaborative perception defense method, electronic equipment, storage medium and program product. It relates to the technical field of vehicle-cloud collaboration. The method is applied to a vehicle client, a local model is deployed on the vehicle client, the local model is used for attack detection and active defense response, and the vehicle-cloud collaborative perception defense method comprises the following steps: feature extraction and reasoning are performed on multi-source sensor data by the local model to obtain a predicted logic value and a local prediction probability distribution corresponding to the predicted logic value; a variational free energy is calculated based on a global knowledge prior distribution received from a cloud server and the local prediction probability distribution; the variational free energy and the predicted logic value are sent to the cloud server, and a global logic value and a global prediction probability distribution returned by the cloud server are received; and parameter optimization is performed on the local model with the aim of minimizing a composite loss. The present application is used to improve the detection performance of the global model and enhance the effect of vehicle-cloud collaborative perception defense capability.

The invention relates to the technical field of chip optimization, in particular to a chip performance optimization method and device, electronic equipment and a storage medium, and the method comprises the steps: planning input data and weight data through employing a three-dimensional MAC array additionally arranged in a target chip, and obtaining a distribution scheme; according to the distribution scheme and the trapezoidal shift register array, carrying out time-sharing loading on the input data, and shifting the input data subjected to time-sharing loading to obtain an input data stream; and according to the input data stream and the weight data, performing convolution calculation on each multiplier-adder unit in the three-dimensional MAC array in sequence to obtain an optimized pipeline calculation result of the target chip. According to the invention, the computing architecture based on the three-dimensional MAC array is optimized, the proportion of read-write operation is reduced, and the overall computing efficiency of the chip is improved; by optimizing the data processing flow, it is ensured that the calculation process can be continuously carried out, the proportion of calculation duration is increased, and time delay caused by read-write operation is shortened.

The application discloses a multi-transmit CPU instruction splicing method, comprising: an early splicing step, used for splicing instructions according to PC information and Fetch Group information predicted by a branch prediction part and writing the Fetch Group information into a Fetch Group queue in a fetch stage; and a read Fetch Group queue and then optimize splicing timing through multiple continuous time periods; a speculation splicing step, used for performing instruction splicing processing on the Fetch Group information to obtain a branch prediction result in a branch prediction pipeline and a fetch pipeline parallel process; wherein branch prediction information is given in stages in the first N cycles of the branch prediction pipeline; and the fetch pipeline dynamically updates the Fetch Group information according to the branch prediction information. The application performs instruction splicing preparation according to branch prediction information in the fetch stage in advance, and significantly reduces time delay in a subsequent processing stage. According to the length and transmission bandwidth of the Fetch Group information, the application dynamically adjusts a data reading and splicing method in multiple continuous time periods, and maximizes utilization of resources.

This invention relates to the field of intelligent manufacturing technology and discloses an artificial intelligence-based intelligent manufacturing production line collaborative control system, including a multi-source data sensing module, a digital twin virtual-real mapping module, a multi-agent collaborative decision-making module, and a human-machine interaction module. The multi-source data sensing module collects and preprocesses multi-source heterogeneous data across the entire production line, constructing a real-time database. The digital twin virtual-real mapping module builds a 1:1 high-precision digital twin model of the production line based on real-time data, enabling bidirectional real-time data interaction between the virtual and real worlds. It also determines the disturbance level of the production line by calculating multi-dimensional disturbance coefficients and introducing a time decay function to accumulate the effective disturbance influence coefficient. The multi-agent collaborative decision-making module matches and executes hierarchical collaborative control strategies according to the disturbance level, generating a comprehensive collaborative control scheme. This achieves accurate quantification and hierarchical collaborative response to production disturbances, significantly improving the adaptive capability and operational stability of the intelligent manufacturing production line.

The present invention relates to a method, a device, a program product and a storage medium for ensuring data consistency. The method comprises: receiving data to be compared sent by a RocketMQ message queue with a delay, querying first summary information of the received data to be compared; obtaining second summary information from Redis; comparing the first summary information with the second summary information to see whether they are consistent, and if they are consistent, obtaining third summary information from an Elasticsearch server; comparing the first summary information with the third summary information to see whether they are consistent, and if they are consistent, deleting the second summary information in Redis, otherwise judging whether the comparison count is greater than a predetermined value; if it is greater than the predetermined value, the comparison fails, deleting the second summary information in Redis, and saving the business data that failed the comparison to a database, otherwise synchronizing the business data corresponding to the first summary information to the Elasticsearch server again, and sending the data to be compared to the message queue at the same time, and performing the comparison after a delay.

Embodiments of the present application provide a data access method and device, electronic equipment and storage medium. The data access method is applied to a computer device. An operating system of the computer device includes a user-mode file system, an interceptor, a sandbox and a first cache area. The interceptor, the sandbox and the first cache area are all set in a kernel mode of the operating system. The method comprises the following steps: receiving data access information sent by an application program through the interceptor, and sending the data access information to the sandbox; acquiring target data in the first cache area and sending the target data to the interceptor through the sandbox when it is detected according to the data access information that the target data exists in the first cache area; the user-mode file system comprises a second cache area set in a user mode of the operating system, and the first cache area shares data in the second cache area; and returning the target data to the application program through the interceptor. The data access path can be reduced, the access delay can be reduced, and the data access efficiency can be improved.

The invention discloses a frequency measurement method and device, a storage medium and electronic equipment. The method comprises the following steps: carrying out analog-to-digital sampling on a signal to be measured at a first sampling rate to obtain a first sampling signal, and carrying out first N-point fast Fourier transform on the first sampling signal to obtain a first frequency; carrying out sampling down-sampling on the first sampling signal based on a preset sampling rate to obtain a second sampling signal sampled under a second sampling rate corresponding to the preset sampling rate; performing second N-point fast Fourier transform on the second sampling signal to obtain a second frequency; under the condition that frequency aliasing does not occur, determining a frequency measurement value of the to-be-measured signal according to the second frequency; and under the condition that frequency aliasing occurs, performing de-aliasing according to the second sampling rate and the first frequency to obtain a frequency after de-aliasing, and determining a frequency measurement value of the to-be-measured signal according to the frequency after de-aliasing and the first frequency. According to the technical scheme provided by the invention, the interference signal radiation delay time in the electronic warfare system can be shortened, and hardware resources required by frequency measurement of the system are saved.

The invention discloses a municipal engineering traffic accident emergency processing method, which collects accident related information in real time through equipment such as a sensor, an automobile data recorder and a traffic camera, and transmits the accident related information to an emergency resource scheduling platform in real time through a wireless network or an Internet of Things platform to ensure rapid and accurate transmission of the information. Therefore, the time delay of manual information confirmation and transmission is greatly reduced; the AI algorithm and the real-time traffic data are utilized, the driving speed and the optimal driving time of the emergency resources are calculated, and the optimal distribution strategy of the emergency resources is further calculated, so that the high efficiency and the reasonability of resource distribution are ensured, and the rescue response time can be remarkably shortened; the emergency resource scheduling platform can perform information sharing and real-time linkage with a plurality of departments such as police, a medical emergency center, a traffic management department and a fire brigade, so that efficient coordination among the departments is ensured, and the problems of response delay and unreasonable resource allocation caused by information asymmetry in a traditional system are avoided.

The present invention relates to the field of edge computing technology, and specifically, to an intelligent edge computing method, system and terminal based on AI artificial intelligence. It includes the following steps: collecting operation data through edge devices; using long short-term memory network models based on operation data to predict equipment failures, and training the long short-term memory network models on the cloud; deploying the trained long short-term memory network models to edge devices; the long short-term memory network models in the edge devices continuously analyze the newly collected data in real time and respond quickly. The design of the present invention can realize real-time analysis and rapid response to newly collected data by deploying the trained long short-term memory network model to the edge device. This greatly reduces the time delay for data to be transmitted to the cloud and then returned to the processing results, which is particularly important for application scenarios that require real-time decision-making.