87 results about "Moving-average model" patented technology

The invention discloses a sudden snow disaster risk degree probability prediction method oriented to insufficient historical time sequences. The method comprises the following steps: step 1), constructing a multi-index risk index based on related data such as historical snow disasters; 2) performing time sequence modeling on the violent snow disaster risk index by adopting an autoregressive integral moving average model, and capturing trend and periodic characteristics of the violent snow disaster risk index; 3) discretizing a continuous risk index of modeling production into a plurality of risk levels, and constructing a state transition probability matrix; step 4), introducing space neighborhood influence and carrier exposure constraint, and constructing a Markov space weighted state transition model; 5) coupling CMIP6 climate scene data, and dynamically adjusting future state transition probability, and 6) fusing ARIMA trend prediction and Markov state transition probability, and outputting future multi-period snowstorm disaster risk level probability distribution.According to the method, the accuracy of snowstorm disaster risk prediction and the modeling ability and prediction adaptability of future climate change scenes are improved.

The invention provides a photovoltaic typical scene generation method and device based on source-load time sequence decomposition, and the method comprises the steps: collecting historical photovoltaic output time sequence data of a region, and carrying out the decomposition based on an empirical mode; the method comprises the following steps: acquiring load time sequence data of a region, and decomposing the load time sequence data based on a simplified autoregressive moving average model; performing feature extraction on the photovoltaic output periodic component, the photovoltaic output random component and the photovoltaic output trend component to obtain photovoltaic periodic features, photovoltaic random features and photovoltaic trend features; performing feature extraction on the load periodic component, the load random component and the load trend component to obtain a load periodic feature, a load random feature and a load basic feature; carrying out clustering analysis according to the photovoltaic periodic characteristics, the photovoltaic random characteristics, the photovoltaic trend characteristics, the load basic characteristics, the load periodic characteristics and the load random characteristics to obtain a photovoltaic typical scene; according to the method, the description of the rural source load characteristics is more accurate.

The invention relates to the technical field of lubricating oil quality monitoring, in particular to a lubricating oil quality on-line monitoring system and method, and the system comprises a data collection module which collects lubricating oil parameters in real time through a sensor array; the communication transmission module is used for constructing a data flow network and supporting multi-level interaction; the data analysis and diagnosis module is used for realizing state evaluation and fault early warning; the user interaction and visualization module provides a human-computer interaction interface and decision support; and the intelligent decision module realizes monitoring-maintenance closed-loop control, gives an alarm in time and realizes full-link intelligent control triggered by a maintenance action. According to the method, the prediction precision and the early warning capability of the degradation trend of the oil system are effectively improved through the improved autoregression moving average model, the real-time state of the oil is accurately reflected by dynamically adjusting the alarm threshold value, and the fault positioning precision is greatly improved through the space matching algorithm.

The invention relates to a multi-source meteorological data fusion technology, and discloses an air temperature reconstruction method fusing data of a micro-meteorological device, which improves the temporal-spatial resolution and precision of ground temperature under a complex terrain. The method comprises the following steps: densely deploying micrometeorological devices in a complex terrain area to obtain high-frequency observation data, carrying out quality control, and carrying out hierarchical processing in two dimensions of space and time by taking pattern forecast grid point data as an initial background field: in the spatial dimension, dynamically updating a fusion weight for grid points with observation stations by using geographical weighted regression, and carrying out data fusion; a residual machine learning model is combined with multi-topographic feature correction for grid points without observation stations, and a high-precision space fusion background field is generated; in the time dimension, errors after space fusion are decoupled into a trend term and a periodic term, an autoregressive integral moving average model is used for predicting a trend, a Fourier algorithm is used for correcting a periodic phase, and then time dimension machine learning correction is carried out on a grid point of an observation-free station. And finally, complete-process automatic, high-temporal-spatial-resolution and low-error complex terrain area air temperature reconstruction is realized.

The invention relates to the technical field of drilling, and provides a roofbolter feedforward-feedback model-free adaptive tracking control method and system based on rotation force. According to the method, a valve-controlled hydraulic propulsion control model for driving a hydraulic cylinder and a corresponding reversing valve in the working process of the hydraulic jumbolter is established, and a nonlinear autoregressive moving average model between the propulsion force in the working process of the hydraulic jumbolter and the output driving current of the reversing valve is obtained through the valve-controlled hydraulic propulsion control model; through a nonlinear autoregression moving average model of the hydraulic jumbolter, an output identification observer deployed with a data-driven adaptive control algorithm of the hydraulic jumbolter is established, and the output identification observer performs adaptive tracking control on the working state of the hydraulic jumbolter through the deployed data-driven adaptive control algorithm. The tracking control problem of the valve control electro-hydraulic servo unit under the complex working condition is effectively solved.

The invention discloses a digital asset management method and system based on multi-objective optimization, and the method comprises the steps: firstly constructing an enhanced fragmentation management network architecture, introducing an interaction frequency-based histocompatibility analysis method and a four-dimensional resource dynamic monitoring mechanism, and achieving the load prediction in combination with an autoregressive differential moving average model. Secondly, establishing a multi-objective optimization framework based on organizational affinity, load balancing and privacy protection, and adopting a differential privacy protection mechanism and dynamic privacy budget management to realize a Pareto optimal solution through a gradient descent method; and finally, designing an intelligent fragmentation decision algorithm and an enhanced load balancing strategy, constructing a layered traceability architecture based on a Merkle tree, and realizing full-life-cycle credible traceability in combination with zero-knowledge proof and a commitment mechanism. Therefore, inter-organization relation privacy and system load information security are protected, system resource configuration and cross-fragmentation cooperation efficiency are optimized, and the overall performance and availability of digital asset sharing service are maximized.

The invention relates to the technical field of artificial intelligence and machine learning, in particular to a deep neural network continuous learning method and system based on task-level attention, and the method comprises the steps of global experience pool construction, hierarchical feature enhancement network construction, index moving average model construction, task flow receiving, task sequential training, dynamic reasoning and the like. Core mechanisms such as a task attention module group and experience playback are fused in the sub-task sequential training and dynamic reasoning step, and efficient screening, storage and utilization of different task features are achieved. And by combining the optimized hierarchical feature enhancement network and the dynamic feature fusion design, the continuous learning comprehensive performance of the model is remarkably improved. According to the method, the efficient learning ability of the new task and the memory retention effect of the old task are both considered, and a brand new solution is provided for application of the deep neural network in a multi-task continuous learning environment.

The invention relates to the technical field of traffic services, in particular to a traffic service transport capacity demand prediction method and system, and the method comprises the steps: collecting historical operation data, track data, passenger flow counting data and external impact factor data for prediction, carrying out the preprocessing of the data, and dividing the data into a plurality of space attributions; constructing a time sequence feature, a periodic feature, a neighbor space feature and an instant available transport capacity feature set external factor feature based on the preprocessed data; generating short-term baseline prediction based on a seasonal autoregressive moving average model, constructing and training a space-time hybrid neural network model to extract spatial correlation and time dynamics of space attribution, outputting a space-time depth prediction result, and integrating the space-time depth prediction result with the short-term baseline prediction to obtain a demand prediction value of a future multi-step time period; and establishing an integer programming model comprising vehicle availability, shift constraints and transfer time constraints and solving the integer programming model to generate vehicle allocation or shift suggestions by taking the transport capacity level as the target and minimizing the scheduling cost.

The invention relates to the technical field of e-commerce, in particular to an e-commerce intelligent prediction management method and system based on big data, and the method comprises the following steps: obtaining a real-time message in a buffer area, generating a throughput rate sequence based on a sliding window, carrying out the difference and mapping of throughput rates of adjacent periods, and constructing a gradient slope sequence; and extracting a transient abrupt change ratio by means of a moving average model to generate a pre-trigger level identifier, extracting a set boundary extreme value, carrying out quantitative evaluation to generate discrete parameters, correcting a transfer matrix, and executing steady-state iterative solution to obtain prediction confidence to construct an elastic control instruction. According to the method, the sudden change trend of the business time domain dimension is accurately captured and static reference deviation interference is eliminated, so that deep analysis and real-time correction of the dynamic transfer characteristics of the core transaction link are realized, and the dilemma of research and judgment distortion and scheduling lag existing in a traditional rule when facing burst traffic is thoroughly avoided; and the full-link prediction precision and the elastic resource management and control efficiency in a complex transaction scene are remarkably improved.

The invention relates to the field of petroleum and natural gas exploitation engineering, and discloses a multi-layer gas injection working condition intelligent early warning method and system based on real-time data, a medium and equipment. The method comprises the steps that a corresponding dynamic attenuation or dynamic increasing dynamic reference model is built for each gas injection layer based on historical real-time gas injection amount data, obtaining the predicted gas injection amount of the layer so as to determine the real-time flow interval of each gas injection layer; a moving average model of wellhead gas injection pressure is established for each gas injection position, a pressure moving average value changing along with time is obtained so as to predict a wellhead gas injection pressure value at the current moment, and a real-time pressure interval of each gas injection position is determined; according to the real-time flow interval and the real-time pressure interval, whether the real-time flow and the real-time pressure collected at the current moment exceed the corresponding interval ranges or not is determined, if the real-time flow and the real-time pressure exceed the ranges and set time, early warning is carried out, the early warning process of each layer is achieved through one thread, and multi-layer synchronous early warning is achieved through multiple threads.

The invention relates to a battery replacement station battery mutual aid allocation method based on a battery demand prediction model and a DQN. The method comprises the following steps: acquiring a multi-dimensional battery data set of each battery replacement station; a deep neural network model and an autoregressive integral moving average model are adopted to predict the battery swap demand of each battery swap station; constructing a reinforcement learning environment which comprises a state space, an action space and a constraint condition; a deep Q network is adopted to construct a reinforcement learning model, a reward function is designed based on the battery demand satisfaction degree, the deployment cost and the battery health state, and an optimal deployment strategy is obtained through an intelligent agent and environment interaction training model; and formulating a battery allocation scheme according to an optimal allocation strategy output by the trained reinforcement learning model, and transmitting an allocation instruction to each battery swap station through a communication network to realize mutual aid allocation of batteries of the battery swap stations. The multi-objective optimization of mutual aid allocation of the batteries of the battery swap station is realized, the battery swap demand satisfaction rate of the battery swap station is improved, and the battery allocation cost is reduced.

The invention relates to the technical field of industrial control and equipment fault prediction, in particular to a tunnel furnace combustion array fault positioning method based on time sequence temperature data, which comprises the following steps: acquiring temperature data of each temperature zone, calculating a control loop deviation, determining neighborhood thermal compensation potential energy based on the control loop deviation and a preset asymmetric coupling coefficient, and determining the fault of a combustion array of a tunnel furnace according to the neighborhood thermal compensation potential energy. The convection covering strength is determined by combining the temperature change rate of the temperature zone and the noise reference, stagnation characteristics covered by heat flow are identified, and the performance degradation index of the combustion head corresponding to the temperature zone is determined based on the change characteristics of the convection covering strength along with time; and carrying out dynamic correction on the prediction residual error of the autoregression integral moving average model by using the performance degradation index, and carrying out fault positioning according to the autoregression integral moving average model after dynamic correction. According to the method, thermal compensation of the adjacent temperature zones can be stripped from the strong coupling thermal field, and accurate positioning of the early-stage micro-blockage fault of the combustion head covered by the heat flow is achieved.

The application provides a subjective question intelligent marking and learning situation analysis system based on writing characteristics, relates to the technical field of education, and the system architecture comprises a writing characteristic acquisition and analysis subsystem, an intelligent marking subsystem and a learning situation analysis and prediction subsystem. With the aid of quantum technology, all-round innovation and improvement are realized. The model constructed by quantum sensors and calculation makes micro dynamic data accurate and observable, time sequence characteristics deep and analyzable, and the quantum algorithm is remarkably effective in constructing the font aesthetics and multi-factor correlation model on a macro level. Quantum natural language processing and neural network cooperation, semantic style fusion innovation, knowledge graph reasoning with the aid of quantum graph algorithm, etc. can accurately judge and scientifically guide various types of questions. Quantum hashing and federal learning help to integrate multi-source data, quantum random walk and autoregressive moving average model realize psychological correlation prediction, accurately understand learning situation, effectively intervene to help students' physical and mental and academic progress, and greatly enhance the efficiency and value of the system in the field of education.

The invention provides a cutterhead load multi-gait prediction method and system for assisting TBM intelligent tunneling, and the method comprises the steps: carrying out the processing of real-time monitoring data through grading preprocessing, and obtaining standardized time sequence tunneling data; tunneling control parameters strongly related to cutterhead loads are screened based on grey relational analysis to serve as key tunneling features, time synchronization and vector splicing are conducted on the key tunneling features and geological indexes, and multi-source information state vectors are constructed; establishing a multivariable fractional order autoregression fractal integral moving average model; the method comprises the following steps: constructing a BiGRU-Seqseq-TAM network; inputting the multi-source information state vector into a BiGRU-Seqseq-TAM network, extracting time sequence features, and mapping and outputting mechanism parameters to be identified; and substituting into the model to reckon the predicted value of the cutterhead load in a plurality of time steps in the future. The invention also discloses a corresponding prediction system. According to the method, the capturing capability of the model on the dynamic evolution rule of the multi-gait load of the cutterhead is enhanced, and the transparency and interpretability of the model are remarkably improved.

The invention belongs to the technical field of lithium ion battery parameter identification, and particularly relates to an embedded optical fiber sensing lithium battery multi-physics field coupling parameter identification method and system. Comprising the following steps: acquiring temperature and strain signals in the lithium ion battery; performing periodic correlation analysis on the strain signal in the lithium ion battery, decomposing the strain signal into a periodic strain subitem and an irreversible aging strain subitem, and establishing a thermal-mechanical-aging multi-physics field coupling model of the lithium ion battery in combination with the temperature signal in the battery and the periodic strain subitem; and iteratively solving parameters of the thermal-mechanical-aging multi-physics field coupling model of the lithium ion battery based on an improved cuckoo search algorithm dynamically adjusted by a neural network. According to the lithium dendrite growth-related irreversible aging strain subitem extraction method based on the seasonal autoregressive moving average model and the improved cuckoo search algorithm based on dynamic adjustment of the neural network, effective identification of parameters is realized.

This invention belongs to the field of data caching and batch synchronization technology, and discloses a method and system for data caching and batch synchronization of a management system based on multi-level caching. The method includes: caching pre-acquired data using a caching strategy based on a pre-built caching architecture; extracting operation data from the cached data; and performing local operation processing and operation type judgment by combining a transaction engine and an operation classifier; acquiring the storage status data of the caching architecture in real time; optimizing the synchronization strategy using a decision tree algorithm; and performing batch synchronization processing by combining a chained version number mechanism; acquiring the running status data of the caching architecture; constructing a caching performance index system using the running status data; and performing data storage alarm processing by combining an autoregressive integral moving average model and association rules. This invention significantly optimizes the response speed of high-frequency operations and effectively reduces the load on backend servers, achieving efficient utilization of hardware resources.

This invention belongs to the field of data analysis technology for structural health monitoring in civil engineering, and particularly relates to a method for identifying time-varying cable forces in bridges based on nonlinear frequency-modal distribution. The method includes: firstly, extracting the free decay vibration response of bridge cables; iteratively decomposing the signal using the Hilbert vibration decomposition method to obtain the time-varying frequency; modeling the time-varying frequency based on an autoregressive moving average model, eliminating the endpoint effect of the time-varying frequency through predictive extension; using this as the initial input, further accurately extracting the instantaneous frequency, instantaneous amplitude, and time-varying damping ratio of each component using an adaptive frequency-modal decomposition algorithm; and outputting a stable time-varying cable force curve after calculating the foundation cable force. This invention has the advantages of good modal aliasing suppression and thorough elimination of endpoint effects, and is suitable for time-varying cable force monitoring scenarios for various cable-stayed bridges, such as cable-stayed bridges and suspension bridges.

The invention relates to the technical field of data prediction, and provides an ecological environment change prediction method based on human activity influence, and the method comprises the steps: obtaining environment monitoring data of a forest protection region, obtaining a poaching monitoring sequence of each monitoring point in the forest protection region according to the environment monitoring data, obtaining an ecological monitoring overall planning coefficient according to the poaching monitoring sequence, and predicting the ecological environment change. Obtaining an environment dynamic monitoring sample set according to the ecological monitoring overall planning coefficient and the wild animal quantity monitoring result data, and constructing an environment transient influence coefficient sequence according to the environment dynamic monitoring sample set; based on the environment transient influence coefficient sequence, utilizing an autoregression moving average model to obtain a prediction result of the environment transient influence coefficient of the forest protection area in a future period of time, and based on the prediction result of the environment transient influence coefficient, performing prediction analysis on ecological environment change. According to the method, the change of the ecological environment is predicted by constructing the environmental transient influence coefficient, so that the accuracy of an ecological environment change prediction result is improved.

The invention discloses a non-cellular intelligent beam management-oriented generative learning-assisted twinning enhancement pre-training method, which comprises the following steps of: firstly, constructing a long short-term memory network fusion diffusion model generative model as a channel prediction digital twinborn body, and predicting a future channel state in a probability mode; secondly, constructing an autoregressive comprehensive moving average model and a variational auto-encoder mixed model to carry out generative flow prediction, and accurately capturing the periodicity and the burstiness of the flow; and finally, jointly applying channel and flow prediction results to a monotonic Q value function decomposition beam selection algorithm after state enhancement. According to the invention, through prior knowledge provided by generative learning, the problems of difficult data sample acquisition, high training cost, slow convergence and poor dynamic traffic adaptability faced by training a beam selection algorithm in a real environment are solved. Compared with a traditional method, the training efficiency, the convergence speed and the final performance in a complex dynamic scene of a beam selection algorithm can be remarkably improved.

The invention discloses a shield cutter friction coefficient real-time monitoring and wear trend prediction method and device, and the method comprises the steps: transmitting a high-frequency pulse wave through an ultrasonic sensor embedded in a cutter working surface, collecting a reflection signal of a cutter-rock soil interface, extracting the multi-dimensional acoustic characteristics of the reflection signal, and synchronously collecting the temperature of a friction region; compensating the multi-dimensional acoustic features through a nonlinear regression model, inputting the compensated features into a support vector regression machine, outputting a real-time friction coefficient, and predicting a wear trend through an autoregressive integral moving average model based on a historical sequence of mu; the device comprises a sound-temperature integrated probe, an embedded processing unit and an anti-vibration sealing structure, and provides core data support for shield cutter service life management, maintenance strategy optimization and digital construction.

The invention relates to an offshore wind speed synthesis method and a wind power prediction method. The method is suitable for the technical field of offshore wind power. The technical problem to be solved by the invention is to provide an offshore wind speed synthesis method and a wind power prediction method. According to the technical scheme, the offshore wind speed synthesis method comprises the steps of generating wind speed data under a second time scale on the basis of an average wind speed under a first time scale in combination with a Weibull distribution model and wind speed time characteristics; and generating wind speed data under a third time scale based on the wind speed data under the second time scale in combination with a Vor-karman power spectral density function and an autoregressive moving average model.

The invention relates to the technical field of electricity market transaction, in particular to an electricity spot market price prediction and transaction optimization method. According to the technical scheme, the electric power spot market price prediction and transaction optimization method comprises a work flow of electric power spot market price prediction and transaction optimization; according to the method, the robustness and the prediction precision of a price prediction model in the face of market complexity and sudden events are remarkably improved, and a long-term and short-term memory network component can effectively capture and memorize a nonlinear dependency relationship from a historical price sequence and related multi-dimensional features by virtue of a gating mechanism; the model can understand a more abstract market dynamic mode, the autoregressive integral moving average model component is used for capturing inherent linear trends and short-term laws in a time sequence, and the model is allowed to dynamically allocate different weights for input information of past different time steps by setting an attention mechanism when prediction is performed each time.

The invention discloses an energy storage system self-healing and early warning method and device and a medium, and relates to the technical field. The method comprises the steps of collecting multi-dimensional operation data through a sensor network deployed in a layered manner; carrying out normalization and adaptive discretization on the multi-dimensional operation data by adopting a Bayesian network, and carrying out feature extraction by utilizing Kalman filtering to obtain enhanced feature data adaptive to a time sequence; the enhanced feature data is calculated through a collaborative model of a convolutional neural network and a Transform, and a fault is confirmed through a deep learning model; for the fault, constructing a state space, an action space and a reward function, and selecting an optimal strategy according to a reward value; and after the optimal strategy is executed, predicting a predicted value of a future preset time step through an autoregressive integral moving average model, and calculating a threshold value by combining a historical data mean value and a standard deviation in a time window. According to the method, real-time monitoring, fault prediction and performance optimization of the energy storage system are realized.

The application discloses a kind of pure angle target trajectory estimation methods, this method will typical target motion model be converted into the autoregressive moving average model of position component to realize state decoupling, by adjusting the relationship between pseudo-linear least square cost function and process noise energy to obtain regularized pseudo-linear least square model.To reduce the estimation deviation generated by the correlation between pseudo-linear measurement matrix and noise, the regularized instrumental variable least square position estimation is obtained by using instrumental variable estimation method.The method of the application obtains the analytical expression of batch estimation of azimuth position component in the framework of regularized least square estimation.The method has linear time calculation complexity, and its estimation performance is irrelevant to initial state.

The invention discloses a hydro-generator stator winding insulation peak-valley adaptive filtering prediction method and system, and belongs to the technical field of power equipment optimization control. The objective of the invention is to solve the problem of accurate prediction of dynamic characteristics of stator winding insulation of a hydro-generator under peak-valley load conditions. A data acquisition module acquires a temperature signal, a vibration signal and insulation resistance of a stator winding in real time; the self-adaptive filtering processing module receives the signal, fuses the temperature signal and the vibration signal to generate a composite reference signal, and performs self-adaptive filtering processing based on the obtained composite reference signal to obtain an output signal after self-adaptive filtering processing; the insulation state prediction module performs feature extraction, calculates an insulation state index based on the extracted features, and performs trend prediction on insulation parameters by adopting an autoregression integral moving average model to obtain an insulation resistance prediction curve; and an optimization control module triggers a load adjustment scheme to generate an output control instruction based on the insulation resistance prediction curve obtained in the step S3.

This invention relates to a method for predicting the abundance of squid resources in the Northwest Pacific based on a time series model, belonging to the field of marine fisheries analysis and processing technology. The method includes the following steps: dividing historical catch data of Northwest Pacific squid into training and test sets; using catch per unit catch effort to characterize resource abundance; calculating the average catch per unit catch effort for each fishing location and the monthly average catch per unit catch effort; performing a unit root test on the training set data to meet stationarity requirements; establishing a seasonal single autoregressive moving average model using the training set data; establishing a seasonal multivariate autoregressive moving average model by combining data from squid spawning grounds and feeding grounds; using the test set to perform white noise testing on the residuals, goodness of fit, and prediction accuracy evaluation to obtain the optimal model; and using the optimal model to predict the abundance of Northwest Pacific squid resources. This invention establishes a more accurate resource abundance prediction model, enabling more precise prediction of Northwest Pacific squid resource abundance.

The application discloses a kind of based on motion compensation time domain filtering method for image, comprising: using autoregressive moving average model to the obtained gyroscope error is modeled, the change rule of data is used to estimate model parameter, and gyroscope error model is established;Using the gyroscope data of removing error obtained by gyroscope error model, the rotation matrix of camera is calculated, and the rotation image is compensated in reverse using rotation matrix, and the de-rotation image is obtained.Through kalman filtering to image motion vector sequence, image random jitter is obtained, and the de-rotation image is compensated according to random jitter, and the stable image and the subjective motion vector of image are obtained;Using the motion target detection algorithm in difference map based on large map coordinate system to detect target in stable image.The application has high detection sensitivity, and can detect target and its trajectory quickly through a small amount of frame track correlation.

This application discloses a cross-regional metabolic prediction method for industrial solid waste based on a combined model, applicable to the field of industrial solid waste treatment. The method includes: converting monetary flows into physical flows to calculate the total flow of industrial solid waste; determining the contribution rate of each sector by calculating the total flow matrix through structural path analysis; using network control analysis to calculate the difference between driving force weights and pulling force weights to identify key sectors and their interactions; constructing a grey prediction model and an autoregressive differential moving average model, and determining weighting based on the residual variance of the grey prediction model and the autoregressive differential moving average model to establish a combined prediction model; inputting key control sectors and their interactions as constraints into the combined prediction model to obtain the optimized path for collaborative governance. This application reflects the inherent connection between industrial production sectors and industrial solid waste generation, obtaining the source path of solid waste generation.

The embodiment of the invention provides a method, device and equipment for resisting industrial network attacks and a computer storage medium. The method for defending the industrial network attack comprises the following steps: acquiring communication data packet information, network state information and side channel signals of an industrial network; according to the communication data packet information, a malicious data packet and signal-to-noise ratio information are obtained; inputting the side channel signal into an autoregression integral moving average model, and determining the attack protection capability of the side channel; determining a side channel attack defense strategy according to the malicious data packet, the signal-to-noise ratio information and the side channel attack defense capability; constructing a network state model based on the network state information, and inputting the network state model into a service attack denial model to obtain a service attack denial strategy; and deploying the side channel attack defense strategy and the service attack denial strategy to the industrial firewall so as to defend the industrial network attack. According to the invention, the comprehensiveness of industrial network attack resistance of the industrial network can be improved, and the application range is widened.

The invention discloses a physiological parameter monitoring task processing method and device, equipment and a storage medium, and relates to the technical field of physiological monitoring, and the method comprises the steps: determining each data set, composed of a video sequence containing a face region, of each local node, and based on a federated learning architecture, obtaining a data set of each local node; training by utilizing each data set to obtain each initial local model and an initial global model; respectively predicting the data set by using the initial local model and an index moving average model corresponding to the initial global model, and determining the difference between the corresponding first prediction result and the second prediction result as a model divergence value; and updating the initial global model based on each initial local model and the model divergence value, and processing a video sequence corresponding to the physiological parameter monitoring task through an updated target global model. Therefore, the privacy of the sensitive biological information can be guaranteed, the multi-source data is fully utilized, the model is optimized in combination with the model divergence value, and the generalization robustness of the multi-source data can be improved.