Elevator dangerous situation active identification and intelligent rescue system

Abstract

The invention discloses an active dangerous situation recognition and intelligent rescue system for an elevator, belongs to the technical field of elevator monitoring, and solves the problems that the data processing amount of a data processing module is increased and the adverse risk judgment accuracy is insufficient due to the fact that an existing system does not carry out pre-recognition processing on multi-source heterogeneous data. The system comprises a data acquisition end, a cloud processing platform, a response feedback module and a visual terminal, and the cloud processing platform identifies and analyzes a monitoring information set based on a dangerous situation identification model and calculates an elevator dangerous situation value based on a dangerous situation risk factor; according to the method, the data acquisition end pre-identifies the elevator monitoring data based on the federated learning algorithm in combination with the relation attention mechanism, so that data distributed pre-identification processing is achieved, the method is suitable for multi-source heterogeneous elevator monitoring scenes possibly involving different main body data, the efficiency and accuracy of data processing of the dangerous situation identification model are improved, and the safety of the dangerous situation identification model is improved. Therefore, the active recognition response speed and the comprehensive evaluation precision of the elevator dangerous situation are ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of elevator monitoring, and in particular relates to an elevator emergency active identification and intelligent rescue system. Background Art

[0002] With the acceleration of urbanization and the increasing number of high-rise buildings, elevators / lifts, as an indispensable vertical transportation tool in modern society, have received extensive attention for their safety and reliability. However, elevators still face many potential risks during operation, which may cause passengers to be trapped, injured or even endanger their lives.

[0003] When elevators / lifts are in operation, common risk types include mechanical failure risk, electrical failure risk, human factor risk, and environmental factor risk. At present, elevator risk identification mainly relies on regular manual inspection and monitoring equipment. Manual inspections are highly subjective, inefficient, and difficult to fully cover, and some potential risks may not be discovered in time.

[0004] The invention patent application with the publication number CN115818388A discloses a system for monitoring the adverse risk of elevator operation. The elevator control module collects risk assessment data in real time during the process of controlling the elevator operation; the data upload module uploads the risk assessment data to the database, and the data processing module uses the risk assessment data in the database to identify whether adverse risks have occurred and pushes adverse risk prompt information to the maintenance terminal when adverse risks exist. In addition, it also identifies whether the adverse risks have been eliminated based on the maintenance operation information fed back by the maintenance terminal and the risk assessment data after the maintenance is completed. However, the existing system only uploads the risk assessment data to the database, and does not perform pre-identification processing on multi-source heterogeneous data, which results in an increase in the data processing volume of the data processing module and insufficient accuracy in adverse risk assessment. In response to the above problems, an elevator emergency active identification and intelligent rescue system is proposed. Summary of the invention

[0005] The purpose of the present invention is to address the shortcomings of the prior art and provide an elevator crisis active identification and intelligent rescue system, which solves the problem that the existing system only uploads risk assessment data to the database without pre-identifying and processing multi-source heterogeneous data, resulting in an increase in the data processing volume of the data processing module and insufficient accuracy in adverse risk assessment.

[0006] The present invention is implemented as follows: an elevator emergency active identification and intelligent rescue system, the system comprising: The data acquisition end is used to obtain multi-source heterogeneous elevator monitoring data in real time, integrate and normalize the elevator monitoring data, pre-identify the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, and upload the monitoring information set after data pre-identification to the cloud processing platform; A cloud processing platform, the cloud processing platform is used to identify and store monitoring information sets, update an elevator operation database based on the monitoring information sets, build and train a danger recognition model based on the elevator operation database, the danger recognition model recognizes and analyzes the monitoring information sets, outputs a danger risk factor, and calculates an elevator danger value based on the danger risk factor; A response feedback module, wherein the response feedback module responds to the elevator danger value, indexes the danger risk factor in the monitoring information set based on the elevator danger value, triggers an alarm feedback signal based on the danger risk factor, and traverses the elevator operation database to generate an intelligent rescue plan corresponding to the alarm feedback signal; The visualization terminal is respectively connected to the data acquisition end, the cloud processing platform, and the response feedback module, and is used to visualize the pre-identification results of elevator monitoring data, crisis risk factors, and intelligent rescue plans.

[0007] Preferably, the data acquisition terminal includes: An elevator sensor group, wherein the elevator sensor group is distributedly deployed in the elevator operation environment, and the elevator sensor group is used to obtain elevator mechanical operation data, electrical operation data, and environmental monitoring data in real time; A video monitoring terminal, which is used to obtain real-time monitoring data inside and outside the elevator; A front-end controller is connected to the elevator sensor group and the video monitoring terminal for communication. The front-end controller is used to collect elevator mechanical operation data, electrical operation data, environmental monitoring data, and internal and external real-time monitoring data and process them into elevator monitoring data, and integrate and normalize the elevator monitoring data; The multi-layer perception module pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, and uploads the monitoring information set after pre-identification to the cloud processing platform.

[0008] Preferably, the method for pre-identifying elevator monitoring data based on a federated learning algorithm combined with a relational attention mechanism includes: Load and integrate the normalized elevator monitoring data, use the principal component analysis method to assign weights to the data types of the elevator monitoring data, and decompose the elevator monitoring data based on the sliding window method, where the sliding window width is 1.5 seconds and the sliding step is 0.5 seconds; Perform spectrum analysis on the elevator monitoring data in each sliding window to determine the position, quantity, amplitude and bandwidth of the noise in the signal, judge whether the weight value of the data type corresponding to the signal exceeds the preset filtering threshold, and determine the data filtering algorithm based on the weight value of the data type corresponding to the signal; If the weight value of the signal corresponding to the data type does not exceed the preset filtering threshold, the Daubechies filtering algorithm of DB6 order is used to perform discrete wavelet decomposition on the elevator monitoring data; If the weight value of the signal corresponding to the data type exceeds the preset filtering threshold, the elevator monitoring data is decomposed and reconstructed based on the biorthogonal wavelet basis algorithm; Integrate the elevator monitoring data after filtering, and perform feature type clustering on the elevator monitoring data based on the federated learning algorithm combined with fuzzy clustering iteration to obtain at least one set of sub-cluster centers; Among them, when the feature type clustering of elevator monitoring data is performed based on the federated learning algorithm combined with fuzzy clustering iteration, the optimization objective function is expressed as: (3); in, represents the optimization objective function of feature type clustering, Monitoring data for elevators With sub-cluster centers The membership matrix of The matrix of order, Sub-cluster center The number of Monitoring data for elevators With sub-cluster centers The Euclidean distance of is the fuzzy index; Membership Matrix The constraints are: (4); Computing elevator monitoring data based on relational attention mechanism With sub-cluster centers The closeness of the elevator monitoring data is judged based on the preset closeness threshold Whether it belongs to the sub-cluster center ; Among them, elevator monitoring data With sub-cluster centers The closeness is calculated by the following formula: (5); in, Indicates closeness, Indicates elevator monitoring data Pair cluster center The membership matrix of Indicates elevator monitoring data For data types The membership matrix of Calculate subcluster centers Internal elevator monitoring data The cluster entropy value of With sub-cluster centers The cluster difference judgment threshold is used to extract the sub-cluster center Pre-identify abnormal data and integrate sub-cluster centers Elevator monitoring data , pre-identify abnormal data and obtain the monitoring information set.

[0009] Preferably, when decomposing and reconstructing the elevator monitoring data based on the biorthogonal wavelet basis algorithm, the CDF5 / 3 wavelet basis and the number of decomposition layers are selected to decompose the elevator monitoring data, and the CDF5 / 3 wavelet basis is used to perform convolution operations on two groups of filters, and upsampling processing is performed, and the elevator monitoring data of each decomposition layer is sequentially decomposed into approximate coefficients and detail coefficients, at least one group of detail coefficients is integrated, and the shrinkage threshold of the detail coefficient is determined based on the cross-validation criterion combined with the random undersampling algorithm, and the detail coefficient whose coefficient amplitude is less than the shrinkage threshold is set to 0, and the reconstruction of the elevator monitoring data is completed; The contraction threshold is determined by the following formula: (1); (2); in, represents the shrinkage threshold, Indicates the corresponding data type The weight value of is the maximum value of detail coefficient, is the mean value of detail coefficient, is the standard deviation of detail coefficient, is the amount of noise in the signal, is the number of decomposition layers, is the initial threshold for contraction, Indicates the detail correction value, is the undersampling factor of the noise quantity.

[0010] Preferably, the cloud processing platform includes: A database management unit, used for identifying and storing monitoring information sets, and updating an elevator operation database based on the monitoring information sets; A model building unit, used to build and train a danger recognition model based on an elevator operation database; The danger value determination unit is used to load the danger recognition model, identify and analyze the monitoring information set based on the danger recognition model, output the danger risk factor, and calculate the elevator danger value based on the danger risk factor.

[0011] Preferably, when the model construction unit constructs and trains the crisis recognition model based on the elevator operation database, the extreme learning machine model is used as the initial model. The initial model includes an input layer, a hidden layer, and an output layer. The hidden layer is frozen, and the hidden layer is replaced by an improved Transformer architecture. The improved Transformer architecture includes an autoencoder, a feature extraction layer, a feature fusion layer, and a decoder. The autoencoder and the decoder are both composed of a three-layer self-attention mechanism, a feedforward neural network, a global pooling layer, and an average pooling layer. The entropy weight method is introduced in the autoencoder to obtain the objective weight of the feature, and the game theory weighting method is introduced in the feature fusion layer to calculate the comprehensive weight of each indicator. The multi-layer perception mechanism and the neighbor propagation clustering algorithm are introduced in the decoder to calculate the elevator crisis value. The feature extraction layer includes 1 BiLSTM layer, 3 attention layers and 1 CFCs layer, and the feature fusion layer is a federated learning architecture including a softmax layer.

[0012] Preferably, a method for constructing and training a danger identification model based on an elevator operation database includes: Load the pre-built crisis recognition model, preset the connection weights between the input layer, improved Transformer architecture, and output layer in the crisis recognition model, as well as the number of neurons, activation function, and loss function; Traverse the elevator operation database, import training samples from the elevator operation database, and divide the training samples into a training set and a test set; The training set is preprocessed using a feature optimization strategy, and the preprocessed training set is input into the Transformer architecture. The Transformer architecture is iteratively trained, and the risk factor recognition accuracy and the critical situation value accuracy of the Transformer architecture under different activation functions are calculated. The optimal risk factor recognition accuracy and the critical situation value accuracy are selected as the optimal activation function of the Transformer architecture. Based on the optimal activation function of the Transformer architecture as a precondition, set the initial learning rate and training rounds, select the adam optimizer to adaptively adjust the Transformer architecture hyperparameters, and output a converged crisis recognition model; Obtain a test set, use the test set to test the crisis recognition model, output the test results, and determine whether the test results meet the preset test threshold. If they meet the preset test threshold, output a converged crisis recognition model.

[0013] Preferably, the method for identifying and analyzing a monitoring information set using a crisis recognition model comprises: The monitoring information set is loaded, and the input layer extracts the pre-identified abnormal data in the monitoring information set. The autoencoder assigns objective weights to the feature factors of the pre-identified abnormal data based on the entropy weight method to obtain the pre-identified abnormal data after the abnormal feature factors are assigned weights. The pre-identified abnormal data weighted by the abnormal feature factor is propagated through the BiLSTM layer, the attention layer, and the CFCs layer in the feature extraction layer. The BiLSTM layer and the attention layer extract the feature vector of the pre-identified abnormal data, and use CFCs to reduce the feature extraction calculation amount and speed up the inference speed, and output the feature vector set; Load the feature vector set. The feature fusion layer calculates the comprehensive weight of each abnormal feature factor based on the game theory weighting method. The federated learning architecture dynamically aggregates the abnormal feature factors globally based on the comprehensive weight of the abnormal feature factors, extracts the critical risk factors through the softmax layer, and outputs the feature fusion set containing the critical risk factors. A feature fusion set is obtained and input into a decoder. The decoder calculates the elevator danger value based on a multi-layer perception mechanism and a neighbor propagation clustering algorithm.

[0014] Preferably, the elevator danger value is calculated by the following formula: (6); (7); in, is the elevator danger value, represents the activation function of the decoder's multi-layer perception mechanism, Indicates the current risk factor The input representation is, Indicates the current risk factor The comprehensive weight of Indicates the current risk factor The bias term, is the number of risk factors for critical situations, Represents the current crisis risk factor based on the neighbor propagation clustering algorithm The cumulative distribution function of is the input mean of the critical risk factor, represents the damping coefficient of the neighbor propagation clustering algorithm, Indicates the maximum number of iterations of the Affinity Propagation clustering algorithm.

[0015] Compared with the prior art, the embodiments of the present application have the following beneficial effects: In the present invention, the data acquisition end pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, thereby realizing distributed data pre-identification processing, which is suitable for elevator monitoring scenarios with multi-source heterogeneity and may involve data from different subjects, and improves the efficiency and accuracy of data processing of the danger recognition model. The cloud processing platform identifies and analyzes the monitoring information set based on the danger recognition model, thereby ensuring the active identification response speed and comprehensive evaluation accuracy of elevator dangers, and realizing distributed data pre-identification processing.

[0016] In the present invention, when pre-identifying elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, the filtering algorithm is dynamically selected according to the weight value of the data type, and the most suitable filtering method can be used for data of different importance. This not only improves the filtering effect, but also avoids the limitations of a single filtering algorithm. By flexibly adjusting the filtering strategy, it ensures more refined processing of key data, while efficiently filtering secondary data, thereby improving the efficiency of overall data processing.

[0017] In the present invention, a danger identification model and a training method are provided. The extreme learning machine model is used as the initial model, and its advantages of fast learning and high efficiency are utilized to replace the hidden layer with an improved Transformer architecture. The extreme learning machine has advantages in processing large-scale data and complex nonlinear problems, while the Transformer architecture performs well in processing sequence data and capturing long sequence dependencies. This combination method gives full play to the advantages of both, so that the model can not only quickly learn data features, but also effectively process the time series information in the elevator operation data, improve the accuracy of danger identification, introduce the entropy weight method in the autoencoder to obtain the objective weight of the feature, and the entropy weight method can determine the weight of the feature according to the information entropy of the data, avoiding the influence of subjective factors, and making the feature weight more objective and accurate. In the feature fusion layer, the game theory weighting method is introduced to calculate the comprehensive weight of each indicator. The game theory weighting method can take into account the mutual relationship and influence between different indicators, and determine the optimal weight of each indicator through the game method, so as to more comprehensively and accurately reflect the characteristics and risk factors of the elevator operation data.

[0018] In the present invention, intelligent rescue plans are automatically generated based on the real-time monitored danger values, avoiding the subjectivity and uncertainty of manual judgment, ensuring that the rescue measures match the actual danger of the elevator, improving the effectiveness and success rate of the rescue, and ensuring the safety of passengers and the normal operation of the elevator equipment to the greatest extent. The full process automation from data acquisition, pre-identification, danger analysis to rescue plan generation greatly shortens the response time, can respond to elevator dangers in a timely manner, reduce the time passengers are trapped due to elevator failures, reduce the losses and impacts caused by accidents, and improve user experience and social satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a structural schematic diagram of the elevator emergency active identification and intelligent rescue system provided by the present invention; Figure 2 The schematic diagram of the implementation process of the pre-identification method of elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism is shown; Figure 3 A schematic diagram of the implementation process of a method for building and training a danger identification model based on an elevator operation database is shown; Figure 4A schematic diagram of the implementation process of the method for identifying and analyzing monitoring information sets using a crisis identification model is shown.

[0020] In the figure: 100, data acquisition terminal; 110, elevator sensor group; 120, video monitoring terminal; 130, front-end controller; 140, multi-layer perception module; 200, cloud processing platform; 210, database management unit; 220, model building unit; 230, danger value determination unit; 300, response feedback module; 400, visualization terminal. DETAILED DESCRIPTION

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of this application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0022] The existing system only uploads the risk assessment data to the database, and does not perform pre-identification processing on the multi-source heterogeneous data, which leads to an increase in the data processing volume of the data processing module and insufficient accuracy in the adverse risk assessment. In view of the above problems, an elevator emergency active identification and intelligent rescue system is proposed. The system includes a data acquisition terminal 100, a cloud processing platform 200, a response feedback module 300, and a visualization terminal 400. When working, first, the data acquisition terminal 100 acquires multi-source heterogeneous elevator monitoring data in real time, integrates and normalizes the elevator monitoring data, and pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism. The cloud processing platform 200 builds and trains a emergency recognition model based on the elevator operation database. The emergency recognition model identifies and analyzes the monitoring information set, outputs the emergency risk factor, and calculates the elevator emergency value based on the emergency risk factor. Finally, the response feedback module 300 responds to the elevator emergency value, indexes the emergency risk factor in the monitoring information set based on the elevator emergency value, triggers an alarm feedback signal based on the emergency risk factor, and traverses the elevator operation database to generate an intelligent rescue plan corresponding to the alarm feedback signal.

[0023] In the embodiment of the present invention, the data acquisition terminal 100 pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, thereby realizing distributed data pre-identification processing, which is suitable for elevator monitoring scenarios with multi-source heterogeneity and may involve different subject data, and improves the efficiency and accuracy of the data processing of the danger identification model. The cloud processing platform 200 identifies and analyzes the monitoring information set based on the danger identification model, thereby ensuring the active identification response speed and comprehensive evaluation accuracy of the elevator danger, and realizing distributed data pre-identification processing. It overcomes the problem that the existing system only uploads the risk assessment data to the database, and does not pre-identify the multi-source heterogeneous data, resulting in an increase in the data processing volume of the data processing module and insufficient accuracy in the adverse risk assessment.

[0024] The embodiment of the present invention provides an elevator emergency active identification and intelligent rescue system. Figure 1 The schematic diagram of the implementation process of the elevator emergency active identification and intelligent rescue system is shown, and the elevator emergency active identification and intelligent rescue system includes: The data acquisition terminal 100 is used to acquire multi-source heterogeneous elevator monitoring data in real time, integrate and normalize the elevator monitoring data, pre-identify the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, and upload the monitoring information set after data pre-identification to the cloud processing platform 200; Wherein, the data acquisition terminal 100 includes: An elevator sensor group 110, wherein the elevator sensor group 110 is distributedly deployed in the elevator operation environment, and the elevator sensor group 110 is used to obtain elevator mechanical operation data, electrical operation data, and environmental monitoring data in real time; It should be noted that the elevator sensor group 110 includes but is not limited to acceleration sensors, gravity sensors, vibration sensors, temperature sensors, humidity sensors, wind speed sensors, etc. The elevator sensor group 110 integrates multiple monitoring functions. In addition to monitoring the basic operating parameters of the elevator (such as speed, position, door status, etc.), it can also diagnose and analyze elevator faults. It can determine whether the elevator is in normal operation according to preset algorithms and rules, and take timely measures, such as alarming, stopping operation, etc., so as to obtain multi-source heterogeneous elevator mechanical operation data, electrical operation data, and environmental monitoring data.

[0025] The video monitoring terminal 120 is used to obtain real-time monitoring data inside and outside the elevator. It should be noted that the video monitoring terminal 120 is installed in the elevator car and the elevator shaft, etc., to monitor the situation inside the elevator and the opening and closing status of the elevator door in real time. Through video image analysis technology, some abnormal behaviors can also be identified, such as passengers fighting, whether there are foreign objects in the car blocking the elevator door from closing, etc. The video monitoring terminal 120 can be a high-definition network camera; A front-end controller 130, wherein the front-end controller 130 is respectively connected to the elevator sensor group 110 and the video monitoring terminal 120 for communication. The front-end controller 130 is used to collect elevator mechanical operation data, electrical operation data, environmental monitoring data, and internal and external real-time monitoring data and process them into elevator monitoring data, and integrate and normalize the elevator monitoring data; The multi-layer perception module 140 is used to pre-identify the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, and upload the monitoring information set after the pre-identification of the data to the cloud processing platform 200.

[0026] A cloud processing platform 200, the cloud processing platform 200 is used to identify and store monitoring information sets, update an elevator operation database based on the monitoring information sets, build and train a danger recognition model based on the elevator operation database, the danger recognition model recognizes and analyzes the monitoring information sets, outputs a danger risk factor, and calculates an elevator danger value based on the danger risk factor; A response feedback module 300, wherein the response feedback module 300 responds to the elevator danger value, indexes the danger risk factor in the monitoring information set based on the elevator danger value, triggers an alarm feedback signal based on the danger risk factor, and traverses the elevator operation database to generate an intelligent rescue plan corresponding to the alarm feedback signal; It should be noted that the response feedback module 300 first receives the elevator emergency value calculated from the cloud processing platform 200. The elevator emergency value is a quantitative representation of the current operation risk of the elevator. The elevator emergency value is calculated based on the analysis and processing of the emergency identification monitoring information set in the cloud processing platform 200 and combined with the emergency risk factor. The response feedback module 300 will judge whether the elevator is in an emergency state and the severity of the emergency according to the elevator emergency value. The emergency risk factor is output by the emergency identification model of the cloud processing platform 200 after analyzing the monitoring information set. The emergency risk factor contains various key information and features related to the elevator emergency. By indexing the emergency risk factor, the response feedback module 300 can clearly identify the specific factors that lead to the elevator emergency, and provide a basis for the subsequent alarm and rescue plan generation. According to the indexed emergency risk factor and the corresponding elevator emergency value, the response feedback module 300 will trigger the alarm feedback signal. The triggering of the alarm feedback signal is based on the preset threshold and rules. While triggering the alarm feedback signal, the response feedback module 300 will traverse the elevator operation database. The elevator operation database stores a large amount of elevator operation data, historical fault records, maintenance information, and corresponding rescue measures, etc. The response feedback module 300 will search for matching historical data and rescue experience in the elevator operation database based on the indexed crisis risk factors, thereby generating an intelligent rescue plan corresponding to the current crisis situation.

[0027] In the embodiment of the present invention, an intelligent rescue plan is automatically generated based on the elevator danger value monitored in real time, avoiding the subjectivity and uncertainty of manual judgment, ensuring that the rescue measures match the actual elevator danger, improving the effectiveness and success rate of rescue, and ensuring the safety of passengers and the normal operation of elevator equipment to the greatest extent. The full process automation from data acquisition, pre-identification, danger analysis to rescue plan generation greatly shortens the response time, can respond to elevator dangers in a timely manner, reduce the time passengers are trapped due to elevator failures, reduce the losses and impacts caused by accidents, and improve user experience and social satisfaction.

[0028] The visualization terminal 400 is respectively connected to the data acquisition terminal 100, the cloud processing platform 200, and the response feedback module 300 for visually presenting the pre-identification results of the elevator monitoring data, the risk factors of the dangerous situation, and the intelligent rescue plan.

[0029] It should be noted that the visualization terminal 400 can be a terminal with communication function such as a smart phone or a tablet computer, and the visualization terminal 400 is connected to the data acquisition terminal 100, the cloud processing platform 200, and the response feedback module 300 via a local area network, 5G or Bluetooth communication.

[0030] In an embodiment of the present invention, the data acquisition terminal 100 pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, thereby realizing distributed data pre-identification processing, which is suitable for elevator monitoring scenarios with multi-source heterogeneity and may involve data from different subjects, and improves the efficiency and accuracy of the data processing of the danger recognition model. The cloud processing platform 200 identifies and analyzes the monitoring information set based on the danger recognition model, thereby ensuring the active identification response speed and comprehensive evaluation accuracy of the elevator danger, and realizing distributed data pre-identification processing.

[0031] The embodiment of the present invention provides a method for pre-identifying elevator monitoring data based on a federated learning algorithm combined with a relational attention mechanism. Figure 2 The schematic diagram of the implementation process of the method for pre-identifying elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism is shown. The method for pre-identifying elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism includes: Step S101: Loading and integrating the normalized elevator monitoring data, using principal component analysis to assign weights to the data types of the elevator monitoring data, and decomposing the elevator monitoring data based on a sliding window method, wherein the sliding window width is 1.5 seconds and the sliding step length is 0.5 seconds; In this embodiment, the elevator monitoring data includes but is not limited to basic operation data, operation direction, speed, position, floor information, load, fault code, fault time, temperature, humidity, maintenance time / content, top and bottom fault data, trapped data, and real-time monitoring data. Through integration and normalization processing, the multi-source heterogeneous elevator monitoring data is converted into a unified format and dimension. This helps to eliminate the differences between different data sources, improve the consistency and comparability of data, and provide a more reliable data basis for subsequent analysis and processing. The principal component analysis method is used to assign weights to data types, and different weights can be assigned according to the importance and relevance of the data. This helps to highlight key data features, reduce the interference of noise data, and improve the efficiency and accuracy of data processing.

[0032] Step S102: performing spectrum analysis on the elevator monitoring data in each sliding window to determine the position, quantity, amplitude and bandwidth of the noise in the signal; It should be noted that by setting the appropriate sliding window width and step size (such as 1.5 second window width and 0.5 second step size), real-time monitoring and dynamic analysis of the elevator operation status can be achieved, and abnormal conditions can be discovered in time. Spectral analysis of the data in each sliding window can accurately determine the position, quantity, amplitude and bandwidth of the noise in the signal. This helps to identify and separate noise signals and provide a basis for subsequent data filtering.

[0033] Step S103: determining whether the weight value of the data type corresponding to the signal exceeds a preset filtering threshold, where the preset filtering threshold may be 0.2-0.5, and determining a data filtering algorithm based on the weight value of the data type corresponding to the signal; In the embodiment of the present invention, when pre-identifying elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, the filtering algorithm is dynamically selected according to the weight value of the data type, and the most suitable filtering method can be used for data of different importance. This not only improves the filtering effect, but also avoids the limitations of a single filtering algorithm. By flexibly adjusting the filtering strategy, it ensures more refined processing of key data, while efficiently filtering secondary data, thereby improving the efficiency of overall data processing.

[0034] Step S104: if the weight value of the data type corresponding to the signal does not exceed the preset filtering threshold, the elevator monitoring data is decomposed by discrete wavelet using the Daubechies filtering algorithm of DB6 order; When the weight value of the signal corresponding to the data type does not exceed the preset filtering threshold, the Daubechies filtering algorithm with DB6 order is used to perform discrete wavelet decomposition on the elevator monitoring data. Daubechies wavelet has good compact support and orthogonality, and can effectively remove noise while better retaining the local characteristics and time-frequency information of the data. The choice of DB6 order is a balance between computational complexity and filtering effect, which can not only ensure effective suppression of noise, but also avoid the loss of important data features due to excessive processing, thereby providing a high-quality data foundation for subsequent feature extraction and analysis.

[0035] Step S105: if the weight value of the data type corresponding to the signal exceeds the preset filtering threshold, the elevator monitoring data is decomposed and reconstructed based on the biorthogonal wavelet basis algorithm; When the weight value of the signal corresponding to the data type exceeds the preset filtering threshold, the elevator monitoring data is decomposed and reconstructed based on the biorthogonal wavelet basis algorithm. The CDF5 / 3 wavelet basis has the advantages of simple calculation, fast speed and symmetry. It can improve the calculation efficiency while ensuring the filtering effect, and is especially suitable for elevator monitoring scenarios with high real-time requirements.

[0036] Among them, when decomposing and reconstructing the elevator monitoring data based on the biorthogonal wavelet basis algorithm, the CDF5 / 3 wavelet basis and the number of decomposition layers are selected to decompose the elevator monitoring data, and the CDF5 / 3 wavelet basis is used to perform convolution operations on two groups of filters, and upsampling processing is performed. The elevator monitoring data of each decomposition layer is decomposed into approximate coefficients and detail coefficients in turn, and at least one group of detail coefficients is integrated. The shrinkage threshold of the detail coefficient is determined based on the cross-validation criterion combined with the random undersampling algorithm, and the detail coefficient whose coefficient amplitude is less than the shrinkage threshold is set to 0, and the reconstruction of the elevator monitoring data is completed; The contraction threshold is determined by the following formula: (1); (2); in, represents the contraction threshold; Indicates the corresponding data type The weight value of is the maximum value of detail coefficient; is the mean value of detail coefficient; is the standard deviation of detail coefficient; is the amount of noise in the signal; is the number of decomposition layers, which can be 2-5; To shrink the initial threshold, it can be preset to 0.1-0.5; Indicates detail correction value; is the undersampling factor of the noise quantity.

[0037] Step S106: Integrate the elevator monitoring data after filtering, and perform feature type clustering on the elevator monitoring data based on the federated learning algorithm combined with fuzzy clustering iteration to obtain at least one group of sub-cluster centers; Among them, when the feature type clustering of elevator monitoring data is performed based on the federated learning algorithm combined with fuzzy clustering iteration, the optimization objective function is expressed as: (3); in, represents the optimization objective function of feature type clustering, Monitoring data for elevators With sub-cluster centers The membership matrix of The matrix of order, Sub-cluster center The number of Monitoring data for elevators With sub-cluster centers The Euclidean distance of is the fuzzy index, which can be set to 2-4 in this embodiment; Membership Matrix The constraints are: (4); In the embodiment of the present invention, fuzzy clustering can handle the uncertainty and fuzziness of data, so that the clustering results are more in line with the actual situation. In this way, at least one group of sub-cluster centers is obtained, which can more accurately reflect the characteristic distribution of elevator monitoring data and provide a more reliable basis for subsequent abnormal data identification.

[0038] Step S107: Calculate elevator monitoring data based on relational attention mechanism With sub-cluster centers The closeness of the elevator monitoring data is judged based on the preset closeness threshold Whether it belongs to the sub-cluster center ; Among them, elevator monitoring data With sub-cluster centers The closeness is calculated by the following formula: (5); in, Indicates closeness, Indicates elevator monitoring data Pair cluster center The membership matrix of Indicates elevator monitoring data For data types The membership matrix of Step S108: Calculate sub-cluster centers Internal elevator monitoring data The cluster entropy value of With sub-cluster centers The cluster difference judgment threshold is used to extract the sub-cluster center Pre-identify abnormal data and integrate sub-cluster centers Elevator monitoring data , pre-identify abnormal data and obtain the monitoring information set.

[0039] In the embodiment of the present invention, by calculating the cluster entropy value of the data in the sub-cluster center and using the cluster entropy mean as the judgment threshold, the pre-identified abnormal data can be effectively identified. This judgment method based on cluster differences can improve the accuracy of abnormal data extraction. By calculating the cluster entropy value, the distribution characteristics of the data can be quantified, so as to more accurately identify abnormal data that is significantly different from normal data. Integrating the data in the sub-cluster center and the abnormal data to form a monitoring information set provides more comprehensive and accurate data support for subsequent crisis identification and analysis.

[0040] On the other hand, when pre-identifying elevator monitoring data, the introduction of the relational attention mechanism can dynamically adjust the weights between data features, highlight key features, and reduce noise interference. In elevator monitoring, this means that features related to elevator safety, such as elevator running speed, acceleration, vibration signals, etc., can be more accurately identified. In combination with federated learning and the relational attention mechanism, multi-source heterogeneous data can be effectively extracted and fused without the need to concentrate all data on a central node for processing. This distributed processing method makes full use of the computing resources of each node, realizes parallel processing of data, greatly shortens the time for data pre-identification, and improves processing efficiency. At the same time, when facing large-scale elevator monitoring data, distributed processing can effectively avoid single point failures and data transmission bottlenecks, ensure the stability and reliability of the system, and by integrating a variety of advanced technologies such as federated learning algorithms, relational attention mechanisms, and biorthogonal wavelet basis algorithms, the model can mine personalized features of elevator monitoring data from different angles. The federated learning algorithm can learn the distributed features of different elevator data, the relational attention mechanism can focus on key features, and the biorthogonal wavelet basis algorithm can refine the data. The synergy of these technologies enables the model to understand the operating status of the elevator more comprehensively and deeply, providing strong support for personalized pre-identification.

[0041] In this embodiment, the cloud processing platform 200 includes: A database management unit 210, used to identify and store monitoring information sets, and update the elevator operation database based on the monitoring information sets; A model building unit 220, used to build and train a danger recognition model based on an elevator operation database; The danger value determination unit 230 is used to load the danger recognition model, identify and analyze the monitoring information set based on the danger recognition model, output the danger risk factor, and calculate the elevator danger value based on the danger risk factor.

[0042] It should be noted that the database management unit 210, the model building unit 220, and the danger value determination unit 230 are connected via Bluetooth or local area network communication, and the elevator operation database stores elevator maintenance logs, operation logs, monitoring information, and intelligent rescue plans.

[0043] The embodiment of the present invention provides a method for constructing and training a danger recognition model based on an elevator operation database. Figure 3 The schematic diagram of the implementation process of the method for constructing and training a dangerous situation identification model based on an elevator operation database is shown. The method for constructing and training a dangerous situation identification model based on an elevator operation database includes: Step S201: loading a pre-built crisis recognition model, presetting the connection weights between the input layer, improved Transformer architecture, and output layer in the crisis recognition model, as well as the number of neurons, activation function, and loss function; Step S202: traverse the elevator operation database, import training samples from the elevator operation database, and divide the training samples into a training set and a test set. The ratio of the training set to the test set may be 4:1. Step S203: preprocess the training set using a feature optimization strategy, input the preprocessed training set into the Transformer architecture, iteratively train the Transformer architecture, calculate the risk factor recognition accuracy and the critical situation value accuracy of the Transformer architecture under different activation functions, and select the optimal risk factor recognition accuracy and the critical situation value accuracy as the optimal activation function of the Transformer architecture; Step S204: Based on the optimal activation function of the Transformer architecture as a precondition, set the initial learning rate and training rounds. In this embodiment, the optimal activation function of the Transformer architecture can be a hard threshold function, the initial learning rate is set to 0.01, and the training rounds are set to 100-150 times. Select the Adam optimizer to adaptively adjust the Transformer architecture hyperparameters and output a converged danger recognition model. This optimization strategy can dynamically adjust the learning rate and hyperparameters according to the training situation of the model to improve the convergence speed and stability of the model. The Adam optimizer combines the advantages of Adagrad and Adadelta, and can adaptively adjust the learning rate of each parameter, avoiding the tediousness and difficulty of manually adjusting the learning rate. At the same time, it can also converge to the optimal solution faster, output a converged danger recognition model, and improve the practicality and reliability of the model; Step S205: obtaining a test set, using the test set to test the crisis recognition model, and outputting the test results; Step S206: determining whether the test result meets a preset test threshold, where the test threshold may be 0.85-0.9; Step S207: If the preset test threshold is met, a converged danger identification model is output.

[0044] If it is determined that the test result does not meet the preset test threshold, return to step S203 and continue to iteratively train the model. This test and evaluation mechanism can timely discover problems and deficiencies in the model and ensure that the performance of the model meets the expected requirements. If the test result does not meet the preset threshold, return to step S203 to continue iteratively training the model. This iterative optimization process can continuously improve and perfect the model, gradually improve the accuracy and generalization ability of the model, until the preset test threshold is met, thereby obtaining a high-performance crisis recognition model.

[0045] In this embodiment, when the model construction unit 220 constructs and trains the danger recognition model based on the elevator operation database, the extreme learning machine model is used as the initial model. The initial model includes an input layer, a hidden layer, and an output layer. The hidden layer is frozen and the hidden layer is replaced by an improved Transformer architecture. The improved Transformer architecture includes an autoencoder, a feature extraction layer, a feature fusion layer, and a decoder. The autoencoder and the decoder are composed of a three-layer self-attention mechanism, a feedforward neural network, a global pooling layer, and an average pooling layer. The autoencoder and the decoder are composed of a three-layer self-attention mechanism, a feedforward neural network, a global pooling layer, and an average pooling layer. The self-attention mechanism can automatically focus on the important parts of the data, capture the long sequence dependencies of the data, and improve the model's understanding and processing capabilities of complex data. The feedforward neural network can perform nonlinear transformations on the data to enhance the expressive power of the model. The global pooling layer and the average pooling layer can reduce the dimension and extract features of the data, reduce the amount of data, and improve the computational efficiency and generalization ability of the model. The entropy weight method is introduced in the autoencoder to obtain the objective weight of the features. The game theory weighting method is introduced in the feature fusion layer to calculate the comprehensive weight of each indicator. The multi-layer perception mechanism and the neighbor propagation clustering algorithm are introduced in the decoder to calculate the elevator crisis value. The feature extraction layer includes 1 BiLSTM layer, 3 attention layers and 1 CFCs layer. The feature fusion layer is a federated learning architecture including a softmax layer.

[0046] In the embodiment of the present invention, a danger identification model and a training method are provided. The extreme learning machine model is used as the initial model, and its advantages of fast learning and high efficiency are utilized to replace the hidden layer with an improved Transformer architecture. The extreme learning machine has advantages in processing large-scale data and complex nonlinear problems, while the Transformer architecture performs well in processing sequence data and capturing long sequence dependencies. This combination method gives full play to the advantages of both, so that the model can not only quickly learn data features, but also effectively process the time series information in the elevator operation data, improve the accuracy of danger identification, introduce the entropy weight method in the autoencoder to obtain the objective weight of the feature, and the entropy weight method can determine the weight of the feature according to the information entropy of the data, avoiding the influence of subjective factors, and making the feature weight more objective and accurate. In the feature fusion layer, the game theory weighting method is introduced to calculate the comprehensive weight of each indicator. The game theory weighting method can take into account the relationship and influence between different indicators, and determine the optimal weight of each indicator through the game method, so as to more comprehensively and accurately reflect the characteristics and risk factors of the elevator operation data.

[0047] The embodiment of the present invention provides a method for identifying and analyzing a monitoring information set using a crisis recognition model. Figure 4 The schematic diagram of the implementation process of the method for identifying and analyzing monitoring information sets using a crisis recognition model is shown. The method for identifying and analyzing monitoring information sets using a crisis recognition model includes: Step S301: Load the monitoring information set, extract the pre-identified abnormal data in the monitoring information set at the input layer, and assign objective weights to the characteristic factors of the pre-identified abnormal data based on the entropy weight method to obtain the pre-identified abnormal data after the abnormal characteristic factors are assigned weights; Step S302: The pre-identified abnormal data weighted by the abnormal feature factor is propagated through the BiLSTM layer, the attention layer and the CFCs layer in the feature extraction layer. The BiLSTM layer and the attention layer extract the feature vector of the pre-identified abnormal data, and the CFCs are used to reduce the feature extraction calculation amount and speed up the inference speed, and output the feature vector set; In this embodiment, the feature extraction layer includes 1 BiLSTM layer, 3 attention layers and 1 CFCs layer. The BiLSTM layer can process bidirectional sequence data and capture the time dependency and context information in the elevator operation data. The attention layer can automatically focus on the important parts of the data and improve the model's attention to key features. The CFCs layer can further enhance the model's feature extraction capability, extract richer and more representative features, and provide strong support for subsequent emergency identification.

[0048] Step S303: Load the feature vector set, the feature fusion layer calculates the comprehensive weight of each abnormal feature factor based on the game theory weighting method, the federated learning architecture globally and dynamically aggregates the abnormal feature factors based on the comprehensive weight of the abnormal feature factors, and extracts the critical situation risk factors through the softmax layer, and outputs the feature fusion set containing the critical situation risk factors. The critical situation risk factors are extracted through the softmax layer, and the information after feature fusion can be converted into specific critical situation risk indicators, providing a quantitative basis for subsequent critical situation assessment; Step S304: Obtain a feature fusion set, input the feature fusion set into a decoder, and the decoder calculates the elevator danger value based on a multi-layer perception mechanism and a neighbor propagation clustering algorithm.

[0049] It should be noted that the multi-layer perception mechanism can extract and represent multi-level features of data, enhancing the learning and generalization capabilities of the model. The neighbor propagation clustering algorithm can cluster data based on the similarity between data, thereby more accurately identifying the dangerous state of the elevator and providing strong support for subsequent rescue and decision-making. In this embodiment, the elevator danger value is calculated by the following formula: (6); (7); in, is the elevator danger value, represents the activation function of the decoder's multi-layer perception mechanism, Indicates the current risk factor The input representation is, Indicates the current risk factor The comprehensive weight of Indicates the current risk factor The bias term, is the number of risk factors for critical situations, Represents the current crisis risk factor based on the neighbor propagation clustering algorithm The cumulative distribution function of is the input mean of the critical risk factor, represents the damping coefficient of the neighbor propagation clustering algorithm, Indicates the maximum number of iterations of the neighbor propagation clustering algorithm. In this embodiment, the damping coefficient may be 0.05-0.2, and the maximum number of iterations may be 5-8.

[0050] In the embodiment of the present invention, the multi-layer perception mechanism in the decoder can perceive and analyze the operating status of the elevator from multiple levels, further improving the ability to identify elevator dangers. The neighbor propagation clustering algorithm can cluster and classify the data according to the similarity between them, so that the model can accurately judge the similarity between the current elevator state and the known danger pattern, thereby calculating a more accurate elevator danger value. The elevator danger value calculated in this way can intuitively reflect the safety status of the elevator and provide an important reference for the maintenance and management of the elevator.

[0051] In summary, the present invention provides an elevator emergency active identification and intelligent rescue system. In an embodiment of the present invention, the data acquisition terminal 100 pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, thereby realizing data distributed pre-identification processing, which is suitable for elevator monitoring scenarios with multi-source heterogeneity and may involve data from different subjects, and improves the efficiency and accuracy of emergency identification model data processing. The cloud processing platform 200 identifies and analyzes the monitoring information set based on the emergency identification model, thereby ensuring the active identification response speed and comprehensive evaluation accuracy of the elevator emergency, and realizing data distributed pre-identification processing.

[0052] It should be noted that, for the above-mentioned embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, some steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0053] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in this field can still combine, add, delete or make other adjustments to the features in the various embodiments of the present invention according to the circumstances without conflict, without making creative work, so as to obtain different other technical solutions that do not deviate from the concept of the present invention in essence, and these technical solutions also belong to the scope of protection of the present invention.

Claims

1. An elevator emergency active identification and intelligent rescue system, characterized in that: The system comprises: The data acquisition end is used to obtain multi-source heterogeneous elevator monitoring data in real time, integrate and normalize the elevator monitoring data, pre-identify the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, and upload the monitoring information set after data pre-identification to the cloud processing platform; A cloud processing platform, the cloud processing platform is used to identify and store monitoring information sets, update an elevator operation database based on the monitoring information sets, build and train a danger recognition model based on the elevator operation database, the danger recognition model recognizes and analyzes the monitoring information sets, outputs a danger risk factor, and calculates an elevator danger value based on the danger risk factor; A response feedback module, wherein the response feedback module responds to the elevator danger value, indexes the danger risk factor in the monitoring information set based on the elevator danger value, triggers an alarm feedback signal based on the danger risk factor, and traverses the elevator operation database to generate an intelligent rescue plan corresponding to the alarm feedback signal; The visualization terminal is respectively connected to the data acquisition end, the cloud processing platform, and the response feedback module, and is used to visualize the pre-identification results of elevator monitoring data, crisis risk factors, and intelligent rescue plans.

2. The elevator emergency active identification and intelligent rescue system according to claim 1, characterized in that: The data acquisition end includes: An elevator sensor group, wherein the elevator sensor group is distributedly deployed in the elevator operation environment, and the elevator sensor group is used to obtain elevator mechanical operation data, electrical operation data, and environmental monitoring data in real time; A video monitoring terminal, which is used to obtain real-time monitoring data inside and outside the elevator; A front-end controller is connected to the elevator sensor group and the video monitoring terminal for communication. The front-end controller is used to collect elevator mechanical operation data, electrical operation data, environmental monitoring data, and internal and external real-time monitoring data and process them into elevator monitoring data, and integrate and normalize the elevator monitoring data; The multi-layer perception module pre-identifies the elevator monitoring data based on the federated learning algorithm combined with the relational attention mechanism, and uploads the monitoring information set after pre-identification to the cloud processing platform.

3. The elevator emergency active identification and intelligent rescue system according to claim 2, characterized in that: The method for pre-identifying elevator monitoring data based on a federated learning algorithm combined with a relational attention mechanism includes: Load and integrate the normalized elevator monitoring data, use the principal component analysis method to assign weights to the data types of the elevator monitoring data, and decompose the elevator monitoring data based on the sliding window method, where the sliding window width is 1.5 seconds and the sliding step is 0.5 seconds; Perform spectrum analysis on the elevator monitoring data in each sliding window to determine the position, quantity, amplitude and bandwidth of the noise in the signal, judge whether the weight value of the data type corresponding to the signal exceeds the preset filtering threshold, and determine the data filtering algorithm based on the weight value of the data type corresponding to the signal; If the weight value of the signal corresponding to the data type does not exceed the preset filtering threshold, the Daubechies filtering algorithm of DB6 order is used to perform discrete wavelet decomposition on the elevator monitoring data; If the weight value of the signal corresponding to the data type exceeds the preset filtering threshold, the elevator monitoring data is decomposed and reconstructed based on the biorthogonal wavelet basis algorithm; Integrate the elevator monitoring data after filtering, and perform feature type clustering on the elevator monitoring data based on the federated learning algorithm combined with fuzzy clustering iteration to obtain at least one set of sub-cluster centers; Among them, when the feature type clustering of elevator monitoring data is performed based on the federated learning algorithm combined with fuzzy clustering iteration, the optimization objective function is expressed as: （3）； in, represents the optimization objective function of feature type clustering, Monitoring data for elevators With sub-cluster centers The membership matrix of The matrix of order, Sub-cluster center The number of Monitoring data for elevators With sub-cluster centers The Euclidean distance of is the fuzzy index; Membership Matrix The constraints are: （4）； Computing elevator monitoring data based on relational attention mechanism With sub-cluster centers The closeness of the elevator monitoring data is judged based on the preset closeness threshold Whether it belongs to the sub-cluster center ; Among them, elevator monitoring data With sub-cluster centers The closeness is calculated by the following formula: （5）； in, Indicates closeness, Indicates elevator monitoring data Pair cluster center The membership matrix of Indicates elevator monitoring data For data types The membership matrix of Calculate subcluster centers Internal elevator monitoring data The cluster entropy value of With sub-cluster centers The cluster difference judgment threshold is used to extract the sub-cluster center Pre-identify abnormal data and integrate sub-cluster centers Elevator monitoring data , pre-identify abnormal data and obtain the monitoring information set.

4. The elevator emergency active identification and intelligent rescue system according to claim 3, characterized in that: When decomposing and reconstructing elevator monitoring data based on the biorthogonal wavelet basis algorithm, the CDF5 / 3 wavelet basis and the number of decomposition layers are selected to decompose the elevator monitoring data, and the CDF5 / 3 wavelet basis is used to perform convolution operations on two groups of filters, and upsampling processing is performed. The elevator monitoring data of each decomposition layer is decomposed into approximate coefficients and detail coefficients in turn, and at least one group of detail coefficients is integrated. The shrinkage threshold of the detail coefficient is determined based on the cross-validation criterion combined with the random undersampling algorithm, and the detail coefficient with a coefficient amplitude less than the shrinkage threshold is set to 0 to complete the reconstruction of the elevator monitoring data. The contraction threshold is determined by the following formula: （1）； （2）； in, represents the shrinkage threshold, Indicates the corresponding data type The weight value of is the maximum value of detail coefficient, is the mean value of detail coefficient, is the standard deviation of detail coefficient, is the amount of noise in the signal, is the number of decomposition layers, is the initial threshold for contraction, Indicates the detail correction value, is the undersampling factor of the noise quantity.

5. The elevator emergency active identification and intelligent rescue system according to claim 1, characterized in that: The cloud processing platform includes: A database management unit, used for identifying and storing monitoring information sets, and updating an elevator operation database based on the monitoring information sets; A model building unit, used to build and train a danger recognition model based on an elevator operation database; The danger value determination unit is used to load the danger recognition model, identify and analyze the monitoring information set based on the danger recognition model, output the danger risk factor, and calculate the elevator danger value based on the danger risk factor.

6. The elevator emergency active identification and intelligent rescue system according to claim 5, characterized in that: When the model construction unit constructs and trains the crisis recognition model based on the elevator operation database, the extreme learning machine model is used as the initial model. The initial model includes an input layer, a hidden layer, and an output layer. The hidden layer is frozen and the hidden layer is replaced by an improved Transformer architecture. The improved Transformer architecture includes an autoencoder, a feature extraction layer, a feature fusion layer, and a decoder. The autoencoder and the decoder are both composed of a three-layer self-attention mechanism, a feedforward neural network, a global pooling layer, and an average pooling layer. The entropy weight method is introduced in the autoencoder to obtain the objective weight of the feature, and the game theory weighting method is introduced in the feature fusion layer to calculate the comprehensive weight of each indicator. The multi-layer perception mechanism and the neighbor propagation clustering algorithm are introduced in the decoder to calculate the elevator crisis value. The feature extraction layer includes 1 BiLSTM layer, 3 attention layers and 1 CFCs layer, and the feature fusion layer is a federated learning architecture including a softmax layer.

7. The elevator emergency active identification and intelligent rescue system according to claim 6, characterized in that: The method of building and training a crisis recognition model based on the elevator operation database includes: Load the pre-built crisis recognition model, preset the connection weights between the input layer, improved Transformer architecture, and output layer in the crisis recognition model, as well as the number of neurons, activation function, and loss function; Traverse the elevator operation database, import training samples from the elevator operation database, and divide the training samples into a training set and a test set; The training set is preprocessed using a feature optimization strategy, and the preprocessed training set is input into the Transformer architecture. The Transformer architecture is iteratively trained, and the risk factor recognition accuracy and the critical situation value accuracy of the Transformer architecture under different activation functions are calculated. The optimal risk factor recognition accuracy and the critical situation value accuracy are selected as the optimal activation function of the Transformer architecture. Based on the optimal activation function of the Transformer architecture as a precondition, set the initial learning rate and training rounds, select the adam optimizer to adaptively adjust the Transformer architecture hyperparameters, and output a converged crisis recognition model; Obtain a test set, use the test set to test the crisis recognition model, output the test results, and determine whether the test results meet the preset test threshold. If they meet the preset test threshold, output a converged crisis recognition model.

8. The elevator emergency active identification and intelligent rescue system according to claim 7, characterized in that: The method for identifying and analyzing a monitoring information set using a crisis recognition model comprises: The monitoring information set is loaded, and the input layer extracts the pre-identified abnormal data in the monitoring information set. The autoencoder assigns objective weights to the feature factors of the pre-identified abnormal data based on the entropy weight method to obtain the pre-identified abnormal data after the abnormal feature factors are assigned weights. The pre-identified abnormal data weighted by the abnormal feature factor is propagated through the BiLSTM layer, the attention layer and the CFCs layer in the feature extraction layer. The BiLSTM layer and the attention layer extract the feature vector of the pre-identified abnormal data and output the feature vector set. Load the feature vector set. The feature fusion layer calculates the comprehensive weight of each abnormal feature factor based on the game theory weighting method. The federated learning architecture dynamically aggregates the abnormal feature factors globally based on the comprehensive weight of the abnormal feature factors, extracts the critical risk factors through the softmax layer, and outputs the feature fusion set containing the critical risk factors. A feature fusion set is obtained and input into a decoder. The decoder calculates the elevator danger value based on a multi-layer perception mechanism and a neighbor propagation clustering algorithm.

9. The elevator emergency active identification and intelligent rescue system according to claim 8, characterized in that: The elevator danger value is calculated by the following formula: （6）； （7）； in, is the elevator danger value, represents the activation function of the decoder's multi-layer perception mechanism, Indicates the current risk factor The input representation is, Indicates the current risk factor The comprehensive weight of Indicates the current risk factor The bias term, is the number of risk factors for critical situations, Represents the current crisis risk factor based on the neighbor propagation clustering algorithm The cumulative distribution function of is the input mean of the critical risk factor, represents the damping coefficient of the neighbor propagation clustering algorithm, Indicates the maximum number of iterations of the Affinity Propagation clustering algorithm.

Images