CN112722007B - Deep learning-based train operation comprehensive early warning method - Google Patents
Deep learning-based train operation comprehensive early warning method Download PDFInfo
- Publication number
- CN112722007B CN112722007B CN202110002477.3A CN202110002477A CN112722007B CN 112722007 B CN112722007 B CN 112722007B CN 202110002477 A CN202110002477 A CN 202110002477A CN 112722007 B CN112722007 B CN 112722007B
- Authority
- CN
- China
- Prior art keywords
- safety
- early warning
- different types
- optimized
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 238000013135 deep learning Methods 0.000 title claims abstract description 35
- 239000002245 particle Substances 0.000 claims abstract description 74
- 238000013528 artificial neural network Methods 0.000 claims abstract description 64
- 238000012549 training Methods 0.000 claims abstract description 52
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 48
- 230000008569 process Effects 0.000 claims abstract description 27
- 230000007547 defect Effects 0.000 claims description 27
- 230000003137 locomotive effect Effects 0.000 claims description 23
- 230000006870 function Effects 0.000 claims description 14
- 230000005540 biological transmission Effects 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 10
- 238000003860 storage Methods 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 3
- 230000006872 improvement Effects 0.000 abstract description 4
- 230000005611 electricity Effects 0.000 description 15
- 238000005457 optimization Methods 0.000 description 15
- 238000007619 statistical method Methods 0.000 description 14
- 238000010586 diagram Methods 0.000 description 8
- 238000003066 decision tree Methods 0.000 description 6
- 238000001514 detection method Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 231100000279 safety data Toxicity 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 230000001502 supplementing effect Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L23/00—Control, warning or like safety means along the route or between vehicles or trains
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Mechanical Engineering (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention provides a deep learning-based comprehensive train operation safety early warning method, which comprises the following steps: collecting and calculating the number of different types of safety problems in the section where the train is located within the target time period until the current moment; if any one of the different types of safety problems exceeds the corresponding safety threshold, sending out a safety early warning; otherwise, inputting the quantity of different types of safety problems into a safety early warning model based on deep learning, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process. The method provided by the invention realizes the improvement of the accuracy and the effectiveness of the railway train operation safety early warning.
Description
Technical Field
The invention relates to the technical field of train operation early warning, in particular to a deep learning-based train operation comprehensive early warning method and device.
Background
With the continuous and rapid development of high-speed railways and general-speed railways and the higher and higher requirements of passengers and owners on transportation safety, the safety has a great significance in railway transportation production. As is known, railway transportation safety is a huge systematic engineering, and relates to a plurality of major and departments such as vehicle affairs, machine affairs, work, electricity affairs, vehicles, power supply and the like, and also relates to safety management, work equipment safety, electricity affair equipment safety, power supply equipment safety, vehicle equipment safety, railway personnel safety, external environment safety and the like.
The train operation safety problem is the most typical, most extensive and most widely influenced safety problem in railway accident faults, particularly the safety of high-speed railway and common-speed passenger trains is the heaviest of railway safety management, and the train operation safety is mainly closely related to the safety state of railway infrastructure equipment, the safety state of mobile equipment, the comprehensive safety capability of railway personnel and the safety of external environments. Specifically, the train operation safety is closely related to the number of work detection overrun, the number of electric service detection problems, the number of power supply detection defects, the number of locomotive equipment alarms, the number of vehicle equipment alarms, the comprehensive capability of railway drivers and the number of external environment alarms.
Safety comprehensive early warning of the position of the train is developed by analyzing safety risks, potential safety hazards, accident faults, work detection data, electricity detection data, power supply detection data, locomotive and vehicle alarm data, external environment safety data and driver safety assessment data of the position of the running train, safety management departments can know the running safety state of the train in time, safety potential risk points of the train drivers are reminded in time, the safety gateway of the running train moves forwards, the safety of the train in the running process is guaranteed, and safety of high-speed rails and passengers is guaranteed.
However, the existing train operation early warning method cannot thoroughly consider the main safety problem, so that the accurate railway train operation safety early warning cannot be made, and the safety early warning should be carried out for the safety problem of single serious problem; meanwhile, for the safety problems without single serious problem but with multiple safety problems, the safety early warning should be developed through deep analysis.
Therefore, how to avoid low accuracy of the railway train operation safety early warning and comprehensively consider historical equipment alarm and defect data, personnel safety data, safety management data, external environment safety data and the like is still a problem to be solved urgently by technical personnel in the field.
Disclosure of Invention
The invention provides a comprehensive train operation safety early warning method and device based on deep learning, which are used for solving the defects that the existing railway train operation safety early warning is low in accuracy rate and potential value information in historical data is not considered, and early warning judgment is made by combining a safety early warning model obtained based on historical data training with a preliminary threshold judgment method, so that the early warning accuracy is improved.
The invention provides a deep learning-based comprehensive train operation safety early warning method, which comprises the following steps:
collecting and calculating the quantity of different types of safety problems in the section where the train is located in the target time period until the current moment;
if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, safety early warning is sent out;
otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm;
the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process.
According to the comprehensive train operation safety early warning method based on deep learning, the safety problems of different types comprise risk number, hidden danger number, accident fault number, work overrun number, electricity alarm number, power supply defect number, locomotive alarm number, vehicle alarm number and external environment alarm number.
According to the comprehensive train operation safety early warning method based on deep learning, provided by the invention, the early warning indication label is used for judging and marking the number of different types of safety problems in a sample target time period manually.
According to the comprehensive train operation safety early warning method based on deep learning, parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process, and the comprehensive train operation safety early warning method specifically comprises the following steps:
optimizing parameters to be optimized in the neural network by using a particle swarm algorithm in the neural network training process of the safety early warning model;
deploying each particle swarm algorithm on a corresponding virtual machine; and each virtual machine is provided with a corresponding IP address, and the IP addresses are used for exchanging the global optimal position of the gbest between the virtual machines through data transmission to realize the update of the gbest.
According to the comprehensive train operation safety early warning method based on deep learning, when parameters to be optimized in a neural network of a safety early warning model are optimized based on a distributed particle swarm algorithm of multiple virtual machines, the fitness function of any particle is a mean square error function determined based on a label corresponding to a predicted value and an input sample value of the particle serving as the parameter to be optimized in the neural network.
According to the comprehensive train operation safety early warning method based on deep learning, provided by the invention, if any one of the different types of safety problems exceeds the corresponding safety threshold, safety early warning is sent out, and the method specifically comprises the following steps:
and if any one of the different types of safety problems exceeds the corresponding safety threshold value, sending out an alarm corresponding to any one type of safety problem.
The invention also provides a deep learning-based comprehensive safety early warning device for train operation, which comprises:
the system comprises a collecting unit, a judging unit and a judging unit, wherein the collecting unit is used for collecting and calculating the quantity of different types of safety problems in the section where the train is located in the target time period until the current moment;
the alarm unit is used for sending out safety early warning if any one of the safety problem quantities exceeds a corresponding safety threshold value;
otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm;
the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process.
According to the comprehensive safety early warning device for train operation based on deep learning, the safety problems of different types comprise risk number, hidden danger number, accident fault number, work overrun number, electric service alarm number, power supply defect number, locomotive alarm number, vehicle alarm number and external environment alarm number.
The invention also provides electronic equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the program to realize the steps of any one of the deep learning-based train operation comprehensive safety early warning methods.
The present invention also provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the deep learning-based train operation comprehensive safety precaution method as described in any of the above.
The comprehensive train operation safety early warning method based on deep learning provided by the invention has the advantages that the number of different types of safety problems in the section where the train is located in the target time period until the current moment is acquired and calculated; if any one of the different types of safety problems exceeds the corresponding safety threshold, sending out a safety early warning; otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process. The early warning judgment provided by the invention is divided into two steps, wherein the first step is equivalent to a decision tree, the aspect with serious problems is found out by using basic threshold screening for early warning, the quantity of different types of safety problems screened by the threshold is judged by using the safety early warning model of the second step, the safety early warning model is obtained by training based on the quantity of different types of safety problems of a large number of samples and corresponding early warning indication labels, the accuracy of judgment and prediction of the safety early warning model can be ensured, and the judgment of supplementing the safety early warning model on the basis of the threshold judgment can further prevent the missing report of the condition needing early warning. Therefore, the method provided by the invention realizes the improvement of the accuracy of the comprehensive safety early warning of the train operation.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a deep learning-based comprehensive train operation safety early warning method provided by the invention;
FIG. 2 is a schematic structural diagram of a deep learning-based comprehensive train operation safety early warning device provided by the invention;
FIG. 3 is a flow chart of the main steps of the comprehensive early warning of train operation safety provided by the present invention;
FIG. 4 is a frame diagram of a safety pre-warning method based on decision trees and neural networks according to the present invention;
fig. 5 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The existing train operation early warning method generally has the problems of low accuracy and incapability of extracting experience from early warning conditions of failure in historical data. The comprehensive train operation safety early warning method based on deep learning of the invention is described below with reference to fig. 1. Fig. 1 is a schematic flow diagram of a deep learning-based comprehensive train operation safety early warning method provided by the present invention, and as shown in fig. 1, the method includes:
and 110, collecting and calculating the quantity of different types of safety problems in the section where the train is located in the target time period until the current moment.
Specifically, to make an early warning of the comprehensive safety condition of the section where the train is located at the current time, the number of different types of safety problems in the latest period of time, that is, the target period of time up to the current time, needs to be collected and calculated. For example, if the time length of the target time period is 72 hours, the number of the different types of security questions collected and calculated is a time point obtained by pushing the current time by 72 hours, the current time is used as an end time, and the number of the different types of security questions collected and calculated between the start time and the end time is used as the start time. And the number of different types of security issues, including: the number of safety risks, the number of hidden dangers, the number of accident faults, the number of times of work trouble, the number of electricity service alarms, the number of power supply defects, the number of locomotive alarms, the number of vehicle alarms and the number of times of external environment safety alarms can be any combination of the parameters to form the number of different types of safety problems of the invention. In order to realize convenient collection and calculation of the number of the safety problems of the safety types, a safety risk library, a potential safety hazard library, an accident fault library, a work safety library, an electric safety library, a power supply safety library, a locomotive safety library, a vehicle safety library, an external environment safety library and a personnel safety library can be established in advance. Besides counting the number of the various types of safety problems, the various types of safety problems can be classified, which is equivalent to performing simple normalization, so that the subsequent threshold value judgment is more convenient. For example: aiming at a safety risk library, dividing risks into major risks, general risks and low risks, and carrying out statistical analysis according to risk types; aiming at a potential safety hazard library, dividing potential safety hazards into major potential safety hazards, prominent potential safety hazards and general potential safety hazards, and carrying out statistical analysis according to different potential safety hazard types; aiming at the engineering safety library, carrying out statistical analysis according to four-level overrun, three-level overrun, two-level overrun and one-level overrun; according to the electric service safety library, carrying out statistical analysis according to important alarm and general alarm; performing statistical analysis on a power supply safety library according to the primary defects and the secondary defects; carrying out statistical analysis according to important alarm and general alarm aiming at a locomotive safety library; aiming at a vehicle safety library, carrying out statistical analysis according to important alarm and general alarm; counting according to important safety problems and general safety problems aiming at an external environment safety library; counting according to key safety personnel and non-key safety personnel aiming at a train driver safety library; taking each kilometer as a unit, and carrying out statistics on safety risks, hidden dangers, accident faults, engineering diseases, electric service alarms, power supply defects, locomotive alarms, vehicle alarms, external environment safety and the like in line division, row division, type division and grade division.
Specifically, when any one of the different types of safety problems exceeds the corresponding safety threshold, namely when the most serious safety condition occurs to any one type of safety problem, an early warning is directly sent out; when the most serious condition does not occur, a safety early warning model based on a multilayer neural network is established, the safety early warning model is obtained after training is carried out on the basis of the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, for example, historical data of the line in the last year and whether an early warning result is taken as a training sample and a label or not can be collected, parameters to be optimized in the neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process, the parameters in the neural network are optimized by applying the distributed particle swarm algorithm, and the safety early warning model of the multilayer neural network is trained. The main idea of training the neural network by the distributed particle swarm optimization is as follows: and taking the minimum square difference of the standard values and the output values of all the samples as a target value, and taking all the weights and the thresholds in the neural network as optimization parameters, namely each particle represents a set of the weights and the thresholds of the neural network. The training time of the particle swarm algorithm is long, so that the training problem of the weight and the threshold of a neural network is solved by using the distributed particle swarm algorithm, the distributed particle swarm algorithm is a cooperative swarm particle swarm optimization method based on multiple virtual machines, each virtual machine executes one particle swarm algorithm, the position of each virtual machine can be different, and the transmission and sharing of the best optimal value are realized through cooperative work of the multiple virtual machines.
The comprehensive safety early warning method for train operation based on deep learning provided by the invention collects and calculates the number of different types of safety problems in the section where the train is located within the target time period until the current moment; if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, safety early warning is sent out; otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process. The early warning judgment provided by the invention is divided into two steps, wherein the first step is equivalent to a decision tree, the aspect with serious problems is found out by using basic threshold screening for early warning, the quantity of different types of safety problems screened by the threshold is judged by using the safety early warning model of the second step, the safety early warning model is obtained by training based on the quantity of different types of safety problems of a large number of samples and corresponding early warning indication labels, the accuracy of judgment and prediction of the safety early warning model can be ensured, and the judgment of supplementing the safety early warning model on the basis of the threshold judgment can further prevent the missing report of the condition needing early warning. Therefore, the method provided by the invention realizes the improvement of the accuracy of the comprehensive safety early warning of train operation.
On the basis of the above embodiment, the different types of safety problems include risk number, hidden danger number, accident number, work overrun number, electricity alarm number, power supply defect number, locomotive alarm number, vehicle alarm number and external environment alarm number.
Specifically, it is further limited herein that the different types of safety issue quantities specifically include 9 safety type safety issue quantities, and correspondingly, if any one of the different types of safety issue quantities exceeds its corresponding safety threshold, a safety warning is issued, specifically including: and if any one or any combination of the conditions that the risk number exceeds a corresponding risk number threshold value, the hidden danger number exceeds a corresponding hidden danger number threshold value, the accident fault number exceeds a corresponding accident fault number threshold value, the work service overrun number exceeds a corresponding work service overrun number threshold value, the electricity service alarm number exceeds a corresponding electricity service alarm number threshold value, the electricity supply defect number exceeds a corresponding electricity supply defect number threshold value, the locomotive alarm number exceeds a corresponding locomotive alarm number threshold value, the vehicle alarm number exceeds a corresponding vehicle alarm number threshold value and the external environment alarm number exceeds a corresponding external environment alarm number threshold value occurs, safety early warning is sent out.
On the basis of the embodiment, the early warning indication label is used for judging and labeling the quantity of different types of safety problems in a sample target time period manually.
Specifically, the construction of the number of different types of safety problems in the sample target time period is to collect data in a long period of time in history, then cut off the data into different sample target time periods, and then count the number of different types of safety problems in each sample target time period to obtain the final number of different types of safety problems in the sample. And the early warning indication labels corresponding to the quantity of different types of safety problems of each sample are judged and labeled by experienced service experts in the field of train safety early warning to the quantity of different types of safety problems in a sample target time period. Further, each early warning indication label is determined by voting by a plurality of service experts together for the number of different types of safety problems of the corresponding sample.
On the basis of the above embodiment, the optimization of the parameters to be optimized in the neural network of the safety early warning model by using a distributed particle swarm algorithm based on multiple virtual machines in the training process specifically includes:
optimizing parameters to be optimized in the neural network by using a particle swarm algorithm in the neural network training process of the safety early warning model;
deploying each particle swarm algorithm on a corresponding virtual machine; and each virtual machine is provided with a corresponding IP address, and the IP addresses are used for exchanging the global optimal position of the gbest between the virtual machines through data transmission to realize the update of the gbest.
Specifically, the input value of the neural network in the training process is a comprehensive hidden danger vector consisting of risk number, hidden danger number, accident fault number, work overrun number, electricity alarm number, power supply defect number, locomotive alarm number, vehicle alarm number and external environment alarm number, and the input value is output as an early warning result. For optimizing parameters to be optimized in the neural network by using a particle swarm optimization algorithm, the method specifically comprises the following steps: the dimensionality of each particle is set as the number of all parameters to be optimized in the neural network, wherein the parameters to be optimized comprise weight coefficients and threshold values of each layer of the network, and the process of adjusting the parameters to be optimized of the neural network each time is as follows:
initializing parameters of a particle swarm algorithm, the number of virtual machines and an IP address of the virtual machines, and automatically deploying the particle swarm algorithm on each virtual machine;
and secondly, executing a particle swarm algorithm on each virtual machine, and updating the pbest and the gbest when a single particle finds better pbest and the gbest. Meanwhile, transmitting the local gbest to different virtual machines through IP;
thirdly, when the local gbest is better than the gbest of other virtual machines, the local gbest is not updated, and the global optimal position information of the local gbest is fed back to other virtual machines; when the other virtual machines are better than the local gbest, the global optimal position information of the local gbest is updated;
step four, different virtual machines respectively execute respective particle swarm optimization, exchange the global optimal position of the gbest in a data transmission mode, and continuously search better position information until the maximum iteration times of the particle swarm optimization;
and fifthly, selecting the optimal position information gbest according to the gbest value of each virtual machine, wherein the optimal position information gbest is used as a parameter to be optimized in the next neural network training.
On the basis of the embodiment, when the parameters to be optimized in the neural network of the safety early warning model are optimized based on the distributed particle swarm optimization of the multiple virtual machines, the fitness function of any particle is a mean square error function determined based on a predicted value output by the particle as the parameters to be optimized in the neural network and a label corresponding to an input sample value.
Specifically, a fitness function when a distributed particle swarm algorithm based on multiple virtual machines optimizes a parameter to be optimized in a neural network of the safety early warning model is set as a mean square error function of a label corresponding to an output predicted value and an input sample value. Namely, when the optimal position of the particle is adjusted, the optimal pbest is found according to the mean square error between the output predicted value and the label value obtained by training the previous pbest used as the parameter to be optimized of the neural network, namely, the fitness function is set as the mean square error function of the label corresponding to the output predicted value and the input sample value.
On the basis of the above embodiment, if any one of the different types of security problem quantities exceeds the corresponding security threshold, then a security early warning is issued, specifically including:
and if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, sending out an alarm corresponding to any one type of safety problem.
Specifically, if any one of the different types of safety problem quantities exceeds the corresponding safety threshold, a safety early warning is issued, for example, the different types of safety problem quantities include a risk number, a hidden danger number, an accident fault number, a work overrun number, an electricity service alarm number, a power supply defect number, a locomotive alarm number, a vehicle alarm number and an external environment alarm number 9 types of safety problem quantities, and if the alarm condition is that the risk number exceeds the corresponding risk number threshold, the hidden danger number exceeds the corresponding hidden danger number threshold, the accident fault number exceeds the corresponding accident fault number threshold, the work service overrun number exceeds the corresponding work service overrun number threshold, the electricity service alarm number exceeds the corresponding electricity service alarm number threshold, the power supply defect number exceeds the corresponding power supply defect number threshold, the locomotive alarm number exceeds the corresponding locomotive alarm number threshold, the vehicle alarm number exceeds the corresponding vehicle alarm number threshold and the external environment alarm number exceeds the corresponding external environment alarm number threshold, any one of the above situations or any combination of the above occurs, the safety early warning is issued, specifically, which type of safety problem quantity exceeds the corresponding safety early warning is issued. For example: and if only the number of the power supply defects exceeds the threshold value in the number of the safety problems of different types, giving out early warning aiming at the power supply defects, wherein the early warning action comprises the steps of broadcasting the power supply defects by voice or displaying the power supply defects by character images on a platform of an early warning center, and informing a power supply department of the early warning.
The deep learning-based comprehensive train operation safety early warning device provided by the invention is described below, and the deep learning-based comprehensive train operation safety early warning device described below and the deep learning-based comprehensive train operation safety early warning method described above can be referred to correspondingly.
Fig. 2 is a schematic structural diagram of the deep learning-based comprehensive train operation safety early warning device provided in the present invention, and as shown in fig. 2, the device includes a collecting unit 210 and an alarm unit 220, wherein,
the acquisition unit 210 is configured to acquire and calculate the number of different types of safety problems in the section where the train is located within the target time period until the current time;
the alarm unit 220 is configured to send out a safety warning if any one of the different types of safety problems exceeds a corresponding safety threshold;
otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm;
the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process.
According to the comprehensive train operation safety early warning device based on deep learning, the number of different types of safety problems in the section where the train is located in the target time period until the current moment is acquired and calculated; if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, safety early warning is sent out; otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, giving an alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process. The early warning judgment provided by the invention is divided into two steps, wherein the first step is equivalent to a decision tree, the aspect with serious problems is found out by using basic threshold screening for early warning, the quantity of different types of safety problems screened by the threshold is judged by using the safety early warning model of the second step, the safety early warning model is obtained by training based on the quantity of different types of safety problems of a large number of samples and corresponding early warning indication labels, the accuracy of judgment and prediction of the safety early warning model can be ensured, and the judgment of supplementing the safety early warning model on the basis of the threshold judgment can further prevent the missing report of the condition needing early warning. Therefore, the device provided by the invention realizes the improvement of the accuracy of the comprehensive safety early warning of train operation.
On the basis of the above embodiment, in the device, the number of the different types of safety problems includes a risk number, a hidden danger number, an accident fault number, an engineering overrun number, an electric service alarm number, a power supply defect number, a locomotive alarm number, a vehicle alarm number and an external environment alarm number.
On the basis of the above embodiment, in the device, the early warning indication label is used for manually judging and labeling the number of different types of safety problems in the sample target time period.
On the basis of the above embodiment, in the device, the optimization of the parameters to be optimized in the neural network of the safety early warning model by using a distributed particle swarm algorithm based on multiple virtual machines in the training process specifically includes:
optimizing parameters to be optimized in the neural network by using a particle swarm algorithm in the neural network training process of the safety early warning model;
deploying each particle swarm algorithm on a corresponding virtual machine; and each virtual machine is provided with a corresponding IP address, and the IP addresses are used for exchanging the global optimal position of the gbest between the virtual machines through data transmission to realize the update of the gbest.
On the basis of the embodiment, in the device, when the distributed particle swarm optimization based on multiple virtual machines optimizes the parameters to be optimized in the neural network of the safety early warning model, the fitness function of any particle is a mean square error function determined based on a predicted value output by the particle as the parameters to be optimized in the neural network and a label corresponding to an input sample value.
On the basis of the above embodiment, in the device, if any one of the different types of safety problems exceeds the corresponding safety threshold, a safety warning is issued, specifically including:
and if any one of the different types of safety problems exceeds the corresponding safety threshold value, sending out an alarm corresponding to any one type of safety problem.
Based on the above embodiment, the present invention further provides a train safety comprehensive early warning method based on expert experience rules and a multilayer neural network, fig. 3 is a flow chart of main steps of the train operation safety comprehensive early warning provided by the present invention, fig. 4 is a frame diagram of the safety early warning method based on a decision tree and a neural network provided by the present invention, as shown in fig. 3 and fig. 4, the implementation steps of the train safety comprehensive early warning method based on the expert experience rules and the multilayer neural network provided by the present invention are as follows:
step 1, firstly, establishing a safety risk library, a potential safety hazard library, an accident fault library, a work safety library, an electric safety library, a power supply safety library, a locomotive safety library, a vehicle safety library, an external environment safety library and a personnel safety library.
And 2, aiming at the safety risk library, dividing the risk into major risk, general risk and low risk, and carrying out statistical analysis according to the risk types.
And 3, aiming at the potential safety hazard library, dividing the potential safety hazards into major potential safety hazards, prominent potential safety hazards and general potential safety hazards, and carrying out statistical analysis according to different types of the potential safety hazards.
And 4, carrying out statistical analysis according to four-level overrun, three-level overrun, two-level overrun and one-level overrun aiming at the engineering safety library.
And 5, performing statistical analysis on the electric service safety library according to important alarms and general alarms.
And 6, carrying out statistical analysis on the power supply safety library according to the primary defects and the secondary defects.
And 7, carrying out statistical analysis according to important alarm and general alarm aiming at the locomotive safety library.
And 8, carrying out statistical analysis according to important alarm and general alarm aiming at the vehicle safety library.
And 9, counting according to important safety problems and general safety problems aiming at an external environment safety library.
And step 10, counting according to key safety personnel and non-key safety personnel aiming at the train driver safety library.
Taking each kilometer as a unit, carrying out statistics and calculation on safety risks, hidden dangers, accident faults, work diseases, electric service alarms, power supply defects, locomotive alarms, vehicle alarms, external environment safety and the like in line, row, type and grade.
When the safety risk, the potential safety hazard, the accident fault, the work overrun, the electricity alarm, the power supply defect, the locomotive alarm, the vehicle alarm and the most serious safety condition of the external environment occur, the special safety item is alarmed.
When the most serious condition does not occur, a train safety early warning model based on the multilayer neural network is established, historical data of a certain line in the last year and whether an early warning result exists are collected, wherein the early warning result of each sample is determined by voting of multiple service experts together, parameters in the neural network are optimized by using a distributed particle swarm algorithm, and the safety early warning model of the multilayer neural network is trained.
The main idea of training the neural network by the distributed particle swarm optimization is as follows: and taking the minimum square difference of the standard values and the output values of all the samples as a target value, and taking all the weights and the thresholds in the neural network as optimization parameters, namely each particle represents a set of the weights and the thresholds of one neural network. The training time of the particle swarm algorithm is long, so that the training problem of the weight and the threshold of a neural network is solved by using the distributed particle swarm algorithm, the distributed particle swarm algorithm is a cooperative swarm particle swarm optimization method based on multiple virtual machines, each virtual machine executes one particle swarm algorithm, the position of each virtual machine can be different, and the transmission and sharing of the best optimal value are realized through cooperative work of the multiple virtual machines.
The first step is as follows: initializing parameters of the particle swarm algorithm, the number of virtual machines and the IP addresses of the virtual machines, and automatically deploying the particle swarm algorithm on each virtual machine.
The second step is that: and executing a particle swarm algorithm on each virtual machine, and updating the pbest and the gbest when a single particle finds the better pbest and the gbest. Meanwhile, the local gbest transmission is transmitted to different virtual machines through the IP.
The third step: when the local gbest is better than the gbest of other virtual machines, the local gbest is not updated, and the position information of the local gbest is fed back to other virtual machines; and when the other virtual machines are better than the local gbest, updating the local gbest position information.
The fourth step: different virtual machines respectively execute respective particle swarm optimization, and exchange the gbest position in a data transmission mode, and continuously search better position information until the maximum iteration times of the particle swarm optimization.
The fifth step: and finally, selecting the optimal position information according to the gbest value of each virtual machine.
After a safety early warning model based on a neural network is trained, according to the position information of a kilometer post where a train is located, the safety risk, hidden danger, accident fault, work safety, electric safety, power supply safety, locomotive safety, vehicle safety and external environment in the kilometer post where the train is located are counted, and whether serious conditions occur or not is calculated through a decision tree rule; if not, a multi-layer neural network model is used for classification, and safety early warning is carried out according to the classification result of the neural network.
Fig. 5 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 5: a processor (processor) 510, a communication Interface (Communications Interface) 520, a memory (memory) 530 and a communication bus 540, wherein the processor 510, the communication Interface 520 and the memory 530 communicate with each other via the communication bus 540. Processor 510 may invoke logic instructions in memory 530 to perform a deep learning based train operation integrated safety precaution method comprising: collecting and calculating the number of different types of safety problems in the section where the train is located within the target time period until the current moment; if any one of the different types of safety problems exceeds the corresponding safety threshold, sending out a safety early warning; otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, giving an alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process.
Furthermore, the logic instructions in the memory 530 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer being capable of executing the deep learning-based train operation comprehensive safety precaution method provided by the above methods, the method including: collecting and calculating the number of different types of safety problems in the section where the train is located within the target time period until the current moment; if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, safety early warning is sent out; otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, giving an alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process.
In yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to perform the deep learning-based train operation comprehensive safety precaution method provided in the above aspects, the method including: collecting and calculating the quantity of different types of safety problems in the section where the train is located in the target time period until the current moment; if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, safety early warning is sent out; otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, giving an alarm; the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment may be implemented by software plus a necessary general hardware platform, and may also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (6)
1. A comprehensive train operation safety early warning method based on deep learning is characterized by comprising the following steps:
collecting and calculating the number of different types of safety problems in the section where the train is located within the target time period until the current moment; the different types of safety problems comprise risk number, hidden danger number, accident fault number, engineering overrun number, electric service alarm number, power supply defect number, locomotive alarm number, vehicle alarm number and external environment alarm number;
if any one of the different types of safety problem quantity exceeds the corresponding safety threshold value, safety early warning is sent out;
otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, sending out a safety alarm;
the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process;
in the training process, a distributed particle swarm algorithm based on multiple virtual machines is adopted to optimize parameters to be optimized in a neural network of the safety early warning model, and the method specifically comprises the following steps:
optimizing parameters to be optimized in the neural network by using a particle swarm algorithm in the neural network training process of the safety early warning model; the fitness function of any particle is a mean square error function determined based on a predicted value output by the particle as a parameter to be optimized in the neural network and a label corresponding to an input sample value;
deploying each particle swarm algorithm on a corresponding virtual machine; and each virtual machine is provided with a corresponding IP address, and the IP addresses are used for exchanging the global optimal position of the gbest between the virtual machines through data transmission to realize the update of the gbest.
2. The deep learning-based comprehensive safety early warning method for train operation according to claim 1, wherein the early warning indication label is used for manually judging and labeling the number of different types of safety problems in a sample target time period.
3. The deep learning-based train operation comprehensive safety early warning method according to any one of claims 1-2, wherein if any one of the different types of safety problems exceeds its corresponding safety threshold, a safety early warning is issued, specifically including:
and if any one of the different types of safety problems exceeds the corresponding safety threshold value, sending out an alarm corresponding to any one type of safety problem.
4. The utility model provides a safety precaution device is synthesized in train operation based on degree of depth study which characterized in that includes:
the system comprises a collecting unit, a judging unit and a judging unit, wherein the collecting unit is used for collecting and calculating the quantity of different types of safety problems in the section where the train is located in the target time period until the current moment; the different types of safety problems comprise risk number, hidden danger number, accident fault number, engineering overrun number, electric service alarm number, power supply defect number, locomotive alarm number, vehicle alarm number and external environment alarm number;
the alarm unit is used for sending out safety early warning if the number of any one type of safety problems in the different types of safety problems exceeds the corresponding safety threshold value;
otherwise, inputting the quantity of the safety problems of different types into a safety early warning model, and if the safety early warning model outputs an early warning instruction, giving an alarm;
the safety early warning model is obtained after training based on the number of different types of safety problems in a sample target time period and corresponding early warning indication labels, and parameters to be optimized in a neural network of the safety early warning model are optimized by adopting a distributed particle swarm algorithm based on multiple virtual machines in the training process;
in the training process, a distributed particle swarm algorithm based on multiple virtual machines is adopted to optimize parameters to be optimized in a neural network of the safety early warning model, and the method specifically comprises the following steps:
in the process of training the neural network of the safety early warning model, optimizing parameters to be optimized in the neural network by using a particle swarm algorithm; the fitness function of any particle is a mean square error function determined based on a predicted value output by the particle serving as a parameter to be optimized in the neural network and a label corresponding to an input sample value;
deploying each particle swarm algorithm on a corresponding virtual machine; and each virtual machine is provided with a corresponding IP address, and the IP addresses are used for exchanging the global optimal position of the gbest between the virtual machines through data transmission to realize the update of the gbest.
5. An electronic device comprising a memory, a processor and a computer program stored on the memory and operable on the processor, wherein the processor when executing the program implements the steps of the deep learning based train operation comprehensive safety precaution method of any one of claims 1 to 3.
6. A non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program when executed by a processor implements the steps of the deep learning based train operation integrated safety precaution method of any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110002477.3A CN112722007B (en) | 2021-01-04 | 2021-01-04 | Deep learning-based train operation comprehensive early warning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110002477.3A CN112722007B (en) | 2021-01-04 | 2021-01-04 | Deep learning-based train operation comprehensive early warning method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112722007A CN112722007A (en) | 2021-04-30 |
| CN112722007B true CN112722007B (en) | 2022-10-25 |
Family
ID=75589544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110002477.3A Active CN112722007B (en) | 2021-01-04 | 2021-01-04 | Deep learning-based train operation comprehensive early warning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112722007B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104742939A (en) * | 2015-03-13 | 2015-07-01 | 南京理工大学 | Temperature information collection and early warning method and system for high speed railway |
| CN108357517A (en) * | 2018-05-21 | 2018-08-03 | 北京京天威科技发展有限公司 | Motor train unit train operation troubles information detecting method and system |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AUPR221900A0 (en) * | 2000-12-20 | 2001-01-25 | Central Queensland University | Vehicle dynamics prediction system and method |
| CN108288019B (en) * | 2017-12-06 | 2020-06-16 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Method and device for identifying preventive maintenance object of urban rail transit |
| KR20200137637A (en) * | 2019-05-31 | 2020-12-09 | 주식회사 케이티 | Training data generation method using virtual alarm, learning method of network failure analysis model, and network system implementing the same method |
| CN110440908A (en) * | 2019-07-22 | 2019-11-12 | 中国铁道科学研究院集团有限公司 | A kind of high-speed railway noise monitors automatically and data management system |
| CN110501160B (en) * | 2019-07-31 | 2021-07-02 | 中国神华能源股份有限公司神朔铁路分公司 | Train bearing fault early warning method, device, system and storage medium |
| CN111806516A (en) * | 2020-06-30 | 2020-10-23 | 苏州华启智能科技有限公司 | Health management device and method for intelligent train monitoring and operation and maintenance |
| CN111923966B (en) * | 2020-07-16 | 2022-07-05 | 北京交通大学 | Urban rail transit train operation control system for different intelligent levels |
| CN112002112B (en) * | 2020-08-25 | 2021-11-02 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Security and protection efficiency evaluation method and system for perimeter intrusion alarm system of high-speed railway |
-
2021
- 2021-01-04 CN CN202110002477.3A patent/CN112722007B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104742939A (en) * | 2015-03-13 | 2015-07-01 | 南京理工大学 | Temperature information collection and early warning method and system for high speed railway |
| CN108357517A (en) * | 2018-05-21 | 2018-08-03 | 北京京天威科技发展有限公司 | Motor train unit train operation troubles information detecting method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112722007A (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103760901B (en) | A kind of rail transit fault identification method based on Classification of Association Rules device | |
| CN112926666A (en) | Rail transit fault diagnosis method | |
| CN104077552A (en) | Rail traffic signal comprehensive operation and maintenance method and system based on cloud computing | |
| CN106202886A (en) | Track circuit red band Fault Locating Method based on fuzzy coarse central Yu decision tree | |
| CN111914687B (en) | Method for actively identifying accidents based on Internet of vehicles | |
| CN111401712A (en) | Urban rail transit risk assessment system and method | |
| CN113987001A (en) | Rail transit signal system fault analysis method and device and electronic equipment | |
| CN114707401A (en) | Fault early warning method and device for signal system equipment | |
| CN111126836A (en) | Security vulnerability risk assessment method and device for high-speed train operation control system | |
| CN108288019B (en) | Method and device for identifying preventive maintenance object of urban rail transit | |
| CN111806516A (en) | Health management device and method for intelligent train monitoring and operation and maintenance | |
| CN115909740A (en) | Vehicle-road cooperation realization method and device, computer equipment and storage medium | |
| CN112722007B (en) | Deep learning-based train operation comprehensive early warning method | |
| CN111709597B (en) | Power grid production domain operation monitoring system | |
| CN110913418B (en) | Method and system for track traffic communication fault early warning and positioning | |
| CN113850400A (en) | Inspection auxiliary judgment method and system | |
| CN113191631A (en) | Railway engineering equipment unit comprehensive portrait early warning method and system | |
| Yan et al. | Evaluating railway operation safety situation in China based on an improved TOPSIS method: A regional perspective | |
| CN116011813A (en) | Urban rail transit emergency monitoring method and device, electronic equipment and storage medium | |
| CN112446232A (en) | Continuous self-learning image identification method and system | |
| CN113781018B (en) | Automatic classification monitoring operation method and device of railway traffic safety monitoring system | |
| CN110135633A (en) | A kind of railway service Call failure prediction technique and device | |
| CN116522213A (en) | Service state level classification and classification model training method and electronic equipment | |
| CN111898895B (en) | Vehicle quality evaluation method and system based on big data fusion | |
| CN114912854A (en) | Subway train operation adjusting method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |