CN113344408B - Processing method for multi-scale situation awareness process of civil aviation traffic control operation - Google Patents
Processing method for multi-scale situation awareness process of civil aviation traffic control operation Download PDFInfo
- Publication number
- CN113344408B CN113344408B CN202110685656.1A CN202110685656A CN113344408B CN 113344408 B CN113344408 B CN 113344408B CN 202110685656 A CN202110685656 A CN 202110685656A CN 113344408 B CN113344408 B CN 113344408B
- Authority
- CN
- China
- Prior art keywords
- situation
- control operation
- civil aviation
- safety
- data
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Abstract
The embodiment of the invention discloses a processing method for a multi-scale situation awareness flow of civil aviation traffic control operation, which comprises the following steps: air traffic control operation data safety protection and fusion, civil aviation control operation safety situation representation, civil aviation control operation multi-scale safety situation assessment and civil aviation control operation safety situation intelligent prediction. By implementing the method provided by the embodiment of the invention, a multi-scale intelligent situation perception system for civil aviation control operation safety can be formed. The method can intelligently analyze and process information and intelligently sense safety risks, changes the current situation of the traditional technology which seriously depends on manpower, and can realize accurate sensing, early warning and effective monitoring of control operation quality of civil aviation control operation safety situation.
Description
Technical Field
The invention relates to the technical field of civil aviation traffic control, in particular to a processing method for a multi-scale intelligent situation awareness flow of civil aviation traffic control operation.
Background
Safety is a permanent theme of civil aviation, and the development of the air traffic transportation industry of the civil aviation must be based on the safety as a premise. The civil aviation air traffic control is used as a command center of the civil aviation transportation industry. In recent years, with the rapid increase of air traffic flow and the continuous enlargement of control scale, the control operation environment becomes increasingly complex, short board effects such as insufficient safety guarantee capability, insufficient technical support and low refinement degree are continuously shown in the control operation process, and the outstanding contradiction and deep problems influencing and restricting the high-quality development of the air traffic pipe are exposed, so that the safety situation is worried.
At present, the control operation safety management mainly adopts a control front-line command system such as air traffic control automation, flow management, a cooperative decision system, scene monitoring and the like, depends on manual experience judgment too much, and is lack of support of multi-scale intelligent risk identification, evaluation and prediction technical means.
The existing part of traditional civil aviation traffic control operation safety perception technology:
1) and the secondary radar responder is utilized to realize the conversion from program control to radar control, and target position information is formed through radar reflected waves, so that the target situation is analyzed.
2) The automatic relevant monitoring system can receive data messages of the aircraft, obtain accurate positioning information of the target, can be used for monitoring the target in an airway, a terminal area and a control airspace of a tower, and realizes effective management of air traffic.
3) Flow prediction and track target monitoring techniques rely on a single 4D track prediction and monitoring of aircraft position information.
4) Conflict detection, alarm and release in the air traffic control automation system are important decision support tools for air traffic control, and are main means for improving air traffic flight safety.
The above scheme has the following disadvantages: the existing civil aviation traffic control operation is mainly carried out by people and assisted by machines, and the problems of air traffic control information acquisition and processing are mainly solved. The perception data independently run in an isolated island form, effective information sharing and data integration are lacked among systems, and the data integration degree are not enough, so that the control running cooperativity is poor, and the control running quality is not high. Situation judgment and early warning cannot be carried out on the future air traffic control operation risk, and short board effects such as insufficient safety guarantee capability, insufficient technical support and low refinement degree are continuously shown in the control operation process.
The reasons for this are: the existing fusion integration of key structured data of the air traffic control operation center does not consider heterogeneous fusion of unstructured data, so that the operation perception data type is single, and a control operation knowledge graph cannot be formed. Meanwhile, intelligent analysis means of relationships and characteristics among entities are lacked, and multi-scale situation perception, trend prediction and whole empty pipe operation capacity evaluation of the air traffic flow are not carried out on the basis of a knowledge graph framework.
Disclosure of Invention
The embodiment of the invention aims to focus on fine safety management of civil aviation control operation, optimize air traffic control operation and control, provide a processing method of a multi-scale intelligent situation perception flow of civil aviation traffic control operation, and realize accurate perception, early warning and effective monitoring of control operation quality of the safety situation of the civil aviation control operation.
Based on the above purpose, the technical problems to be solved by the embodiments of the present invention are: aiming at a complex civil aviation control operation environment and massive air traffic control operation data, comprehensiveness, accuracy and high efficiency of control operation safety situation perception are guaranteed by utilizing data multiscale, knowledge multiscale and perception multiscale. Meanwhile, based on the artificial intelligence technologies such as knowledge graph, deep learning and reinforcement learning, the intelligentization and visualization of the sensing process are realized, the high efficiency and reliability of situation sensing are ensured, and the situation sensing and dynamic evolution capabilities are improved. Finally, refined civil aviation control operation safety management is achieved through a multi-scale intelligent situation perception technology, control service operation and control are optimized, and the refined civil aviation control operation safety management is supported.
In order to achieve the above object, an embodiment of the present invention provides a processing method for a multi-scale situation awareness process of civil aviation traffic control operation, including:
the safety protection and fusion of the air traffic control operation data comprises the following steps: acquiring empty pipe operation data, and performing data security and fusion processing on the empty pipe operation data based on a block chain technology, a missing empty pipe operation data filling algorithm and a space-time aligned empty pipe operation data fusion algorithm to obtain first data;
the method comprises the following steps of civil aviation control operation safety situation representation: based on the first data, the characterization of the safety situation of civil aviation control operation is realized by adopting feature dimension reduction and extraction, an embedded learning technology and a deep reinforcement learning technology;
the method comprises the following steps of civil aviation control operation multi-scale safety situation assessment: performing multi-scale safety situation assessment by adopting a collaborative learning and fusion algorithm;
the method comprises the following steps of intelligent prediction of civil aviation control operation safety situation: adopting a heuristic deep recursion frame to carry out intelligent prediction on the safety situation of civil aviation control operation;
the method specifically comprises the following steps of evaluating the multi-scale safety situation of civil aviation control operation:
constructing a civil aviation control operation safety situation index system: refining a first-level index and a second-level index which influence the safety situation of the civil aviation control operation by analyzing the requirements of the civil aviation control operation and the safety management service; designing quantitative calibration of indexes of the safety situation of civil aviation control operation, designing index weights according to the influence degree of each index on the safety situation and the actual situation, and dividing a safety situation index grade interval;
and (4) safety situation assessment: inputting the representation data of the civil aviation control operation safety situation into the collaborative multi-core situation assessment model, and realizing multi-scale safety situation assessment of the civil aviation control operation by combining with the safety situation index grade interval; the cooperative multi-core situation assessment model adopts a softmax function and a multi-core learning theory.
As a specific implementation manner of the application, the characterization steps of the safety situation of civil aviation control operation are specifically as follows:
performing feature dimensionality reduction and extraction on the air traffic control operation data to generate a preliminary knowledge map;
adopting an embedded learning technology to carry out integrity supplement on the knowledge graph;
and performing intelligent extraction of implicit relation link on the constructed knowledge graph by adopting a deep reinforcement learning mechanism, thereby realizing the representation of the civil aviation control operation safety situation based on the knowledge graph library.
Further, as a specific embodiment of the present application, the safety situation assessment specifically includes:
inputting civil aviation control operation safety situation representation data, wherein each datum is used as a situation sample;
mapping the situation samples to four situation kernel spaces; the four state potential nuclear spaces are respectively a traffic flow dynamic sub-situation space, a control operation safety sub-situation space, a control operation efficiency sub-situation space and a control work load sub-situation space;
respectively training a softmax function in the four situation nuclear spaces, and carrying out loss weighting;
introducing a regular term as a collaborative learning constraint to obtain a final loss function;
and adopting the final loss function to realize the safety situation evaluation and outputting a situation evaluation result.
As a specific implementation manner of the application, the intelligent prediction step of the civil aviation control operation safety situation specifically comprises the following steps:
improving the architecture of a situation prediction network model by adopting a heuristic deep recursion frame;
a risk condition set is formed by combining a long-term historical situation early warning sequence, and the safety risk index and tolerance evaluation under short-term situation prediction are performed, so that the intelligent analysis capability of a situation prediction network model is improved;
and the situation prediction network model is adopted to realize the intelligent prediction of the civil aviation control operation safety situation.
Further, as a preferred embodiment of the present application, the method further includes training the collaborative multi-core situation assessment model by using a K-fold cross validation method without repeated sampling, specifically:
and evenly dividing the N situation samples into K groups, training a model by using the K-1 group, then evaluating the next group of situations by using the trained model, calculating evaluation errors, and calculating the error average value of K times to obtain the cross validation errors.
By implementing the method provided by the embodiment of the invention, a multi-scale intelligent situation perception system for civil aviation control operation safety can be formed. The method can intelligently analyze and process information and intelligently sense safety risks, changes the current situation of the traditional technology which seriously depends on manpower, and can realize accurate sensing, early warning and effective monitoring of the control operation quality of the civil aviation control operation safety situation.
Drawings
In order to more clearly illustrate the detailed description of the invention or the technical solutions in the prior art, the drawings that are needed in the detailed description of the invention or the prior art will be briefly described below.
Fig. 1 is a schematic flow chart of a processing method for a multi-scale situation awareness flow of civil aviation traffic control operation according to an embodiment of the present invention;
FIG. 2 is a flow chart of multi-scale security situation assessment for civil aviation control operation;
FIG. 3 is a safety level diagram;
FIG. 4 is a flow diagram of a collaborative multi-core regulatory operation situation assessment model;
FIG. 5 is a chart of the safety situation space of civil aviation control operation.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, 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.
Referring to fig. 1, the inventive concept of the present invention is: by taking 'safety protection and fusion of civil aviation air traffic control operation data → characterization of civil aviation control operation safety situation → assessment of multi-scale safety situation of civil aviation control operation → intelligent prediction of safety situation of civil aviation control operation' as a main line, a multi-scale intelligent perception flow scheme of the safety situation of civil aviation control operation is established, the problems of perception and early warning of the safety situation of civil aviation control operation, monitoring of control operation quality and the like are overcome, the control workload is reduced, the safety guarantee capability of civil aviation control operation is improved, and fine safety management of civil aviation control operation is supported.
Based on the inventive concept, the processing method for the multi-scale situation awareness process of civil aviation traffic control operation provided by the embodiment of the invention mainly comprises the following steps:
(1) air traffic control operation data safety protection and fusion
The method comprises the steps of obtaining empty pipe operation data, and performing data security and fusion processing on the empty pipe operation data based on a block chain technology, a missing empty pipe operation data filling algorithm and a space-time aligned empty pipe operation data fusion algorithm to obtain first data.
Specifically, the civil aviation air traffic control operation data (such as monitoring data, meteorological data, flow control information, control voice and the like) lack an effective security encryption verification means in the acquisition process, and have the risk of being easily tampered; secondly, data loss may be caused by data-level threats, communication faults, system faults and system maintenance; and the multi-source heterogeneous data of the empty pipe operation independently exist in each control operation system, so that the consistency and the cooperativity of the data resources are poor. Aiming at the problems of information isolated island and data safety in the existing air traffic control system, the patent researches from two parts of data safety and data fusion. In the data security part, aiming at the problems of information isolated islands, data verification uncertainty and data abnormal loss existing among internal data of the existing air traffic control system, the block chain technology is utilized to realize the reality and reliability of the multi-source heterogeneous air traffic control operation data universe node information and ensure the authenticity of the data; analyzing the existing missing data filling method, and ensuring the data integrity by combining the characteristics of an empty pipe system; in the data fusion part, the multi-scale feature fusion extraction of massive air traffic control data is emphasized, and an air traffic control data fusion algorithm considering space-time alignment is performed from multiple dimensions such as time and space, so that the data consistency is ensured.
(2) Civil aviation control operation safety situation representation
And based on the first data, the characterization of the safety situation of civil aviation control operation is realized by adopting feature dimension reduction and extraction, an embedded learning technology and a deep reinforcement learning technology.
Specifically, firstly, the air traffic control running multi-source heterogeneous data containing rich information is independently and dispersedly stored in a database or an unstructured file, so that the association depth among the data is insufficient; secondly, the data has structural difference and diversity characteristics; finally, the traditional control operation knowledge reasoning method is not high in intelligent degree and cannot adapt to a large amount of control operation knowledge graph. Aiming at the characteristics of insufficient correlation depth, insufficient knowledge fusion degree, difficulty in finding implicit knowledge and the like among multi-source heterogeneous civil aviation air traffic control operation data, the method carries out situation characterization research from three parts of air traffic control operation data explicit relation extraction, explicit relation embedding fusion and implicit relation embedding. Firstly, explicit relation extraction is focused on the problem of massive multi-scale of civil aviation air traffic control operation data, and feature dimensionality reduction and extraction are carried out on the civil aviation air traffic control operation data to generate a preliminary knowledge map; secondly, the integrity of the civil aviation air traffic control operation knowledge map is supplemented by an embedded learning technology aiming at the fact that the integrity of the civil aviation air traffic control knowledge is not high. Finally, the implicit relation embedding key point aims at the problem that implicit knowledge in large-volume civil aviation management operation data is difficult to extract, and intelligent extraction of implicit relation links is carried out on the basis of the established knowledge map by utilizing a deep reinforcement learning mechanism. And realizing the control operation safety situation representation based on the knowledge graph library on the basis of the research of the three parts.
(3) Multi-scale safety situation assessment for civil aviation control operation
a. Constructing a civil aviation control operation safety situation index system: refining a first-level index and a second-level index which influence the safety situation of the civil aviation control operation by analyzing the requirements of the civil aviation control operation and the safety management service; designing quantitative calibration of indexes of the safety situation of civil aviation control operation, designing index weights according to the influence degree of each index on the safety situation and the actual situation, and dividing a safety situation index grade interval;
b. and (4) safety situation assessment: and inputting the safety situation representation data of civil aviation control operation into the collaborative multi-core situation assessment model, and realizing multi-scale safety situation assessment of civil aviation control operation by combining the safety situation index grade interval.
The existing part of civil aviation control operation indexes have qualitative and non-quantitative characteristics, fuzzy problems exist in the description of the control safety condition, and in addition, a comprehensive multi-scale civil aviation control operation safety situation index system is formed in China, so that the safety risk insight is insufficient; secondly, the traditional evaluation algorithm does not have the capabilities of continuous evaluation, intelligent evaluation and the like, so that the evaluation result is low in accuracy and reliability. Aiming at the problems that the evaluation index of the civil aviation control operation safety situation is single, the evaluation method does not have multiple scales and the like, the method starts with two parts of construction and evaluation model of the control operation safety situation index system. Firstly, in the aspect of construction of a situation index system, aiming at the problem that indexes lack uniformity and comprehensiveness, the method analyzes the macroscopic safety property of the civil aviation control operation safety situation and the requirements of control operation safety management services, constructs multi-stage key indexes on the basis, and designs index weights; secondly, in the aspect of an evaluation model, aiming at the multi-scale problem existing in the safety situation of civil aviation control operation, mapping of situation data samples to a sub-situation space is realized, a classification method based on a learning theory is used for realizing segmentation representation of the samples, and meanwhile, a fusion learning method is used for completing multi-scale evaluation of the situation.
Referring to fig. 2, the main research idea of safety situation assessment is: the left part in the figure aims at constructing an air traffic control safety situation index system, constructing and quantizing multi-scale civil aviation control operation safety situation indexes, and carrying out weight design and situation index grade interval division according to the actual situation of each control area so as to provide a basis for next accurate evaluation; the right part in the diagram is a collaborative multi-core situation evaluation model, the model maps samples to four situation core spaces, and combines a softmax function and a multi-core learning theory to enrich the final characteristic expression of the operation situation, so that the problem of multi-scale situation safety intelligent evaluation is solved.
1) Hierarchical index construction
According to the characteristics of safe operation data of civil aviation control operation and the characteristics of situation assessment, the assessment idea is provided as follows:
firstly, primary and secondary indexes influencing the safety situation of civil aviation control operation are extracted by analyzing the requirements of the civil aviation control operation and the safety management service.
Secondly, quantitative calibration of indexes of the safety situation of civil aviation control operation is designed, and index weights are designed according to the influence degree of each index on the safety situation and the actual situation, so that the safety situation index grade interval is divided. And constructing a control operation safety situation index system with universality and completeness based on measurability and quantifiability, and providing an index basis for a situation evaluation model.
In this embodiment, the current civil aviation control operation safety situation is divided into four primary indexes, namely, traffic flow dynamics, control operation safety, control operation efficiency and controller workload, covering part of the secondary indexes, and mainly comprising: flight quantity index, navigation mileage, taxi time, sector saturation, excess capacity, navigation data change, flight interval index, traffic flow conflict, alarm, continuous flight approach/departure efficiency, temporary airline usage, controller call load, controller land-air call efficiency index and other 15 secondary indexes. The above-described index is merely an example, and the index is not limited to the above-described index in practical applications.
Each index can continuously enrich and improve a control operation safety situation index system through longitudinal and transverse expansion, the situation is sensed in a multi-scale mode, the prediction accuracy is improved, and unsafe events caused by control errors are reduced.
Further, the index quantification is to design the weight parameter alpha according to the actual situation of each control operationi，βiAnd (4) constructing a multi-scale safety assessment index system by carrying out various combinations of parameters in the concrete implementation of the distribution values. In order not to lose generality, the weight index should be defined to satisfy the condition:
u, R respectively indicate the number of first and second level indicators. The safety assessment is carried out aiming at the potential safety hazard caused by the transient change of the quantized data, so that the possible empty pipe operation risk is identified, and the position and the occurrence scene of the dangerous source are confirmed. The situation evaluation result can be displayed more visually, and the situation index is increased gradually.
Further, five types of intervals of the situation index A, B, C, D, E are planned to be divided according to the actual situation of each controlled area, and the corresponding situation index intervals are qualitatively described as five grades of excellent, good, medium, poor and dangerous, as shown in fig. 3. In the figure, the optimal representation has the lowest situation index and good operation situation; the danger indicates that the situation index is highest, the safety quality is poor, and the problem exists and needs to be paid attention to. The good and good are 'safe' and the medium, bad and dangerous are 'warning'.
2) Collaborative multi-core control operation situation evaluation model
The multi-scale situation assessment model for the civil aviation control operation safety is used for carrying out correlation combination on a plurality of quantitative index data from a knowledge graph from a multi-scale safety angle, so that the multi-scale safety situation and threat under the civil aviation control operation scene are completely assessed. The civil aviation control operation safety situation assessment can be classified into a multi-classification problem, the problem generally needs to manufacture a two-classifier for each class or data of any two classes separately, then the two-classifier is aggregated to form a multi-classifier model, and the method of expanding the two-classification model into the multi-classification model in the category faces problems of sample imbalance, decision boundary repeatability, calculation complexity and the like. In this embodiment, a softmax function and a multi-core Learning theory (Multiple Kernel Learning) are combined to enrich the final feature expression of the situation data, and the intelligent classification capability of the softmax function is improved, so that the problem of multi-classification evaluation in the situation perception of civil aviation control operation is solved, and the flow is shown in fig. 4.
Assume existing situation data { x1φ1},{x2φ2}，…，{xNφNTherein of
The label (number of situations) corresponding to each sample is phiiE {1, 2, …, K }, and the symmetric positive semidefinite kernel matrix is defined as K ═ kerij]N×NAnd kerij＝Φ(xi)·Φ(xj). Assuming that the rank order of K is r, the kernel matrix K can be decomposed into:
wherein Λr×rIs a diagonal matrix whose elements consist of r positive eigenvalues of K, QN×rAre eigenvectors corresponding to the r eigenvalues. To reflect the visualization form in the kernel space, a mapping function is defined as
For the original feature space
Can be mapped to kernel space by e
And the calculation formula of e (x) is as follows:
a method of combining multiple kernels is proposed to improve the accuracy of the classification and regression tasks. In this patent, four kernel mapping matrices W are to be employedl1, …, 4, with the first-level index one-to-one in the civil aviation control operation index system, along with further establishment and the optimization of civil aviation control operation safety situation index system, this patent will carry out more reasonable, meticulous division to the situation space to richen the expression of situation kernel function. The situation space representation is shown in the following table, and for the current situation sample set, the expressions in the four kernel spaces are defined as
As shown in fig. 5.
In a high-dimensional feature space of the safety situation of civil aviation control operation, if only one kernel function is selected to realize the mapping operation of the sample feature vector, a better classification effect cannot be determined. According to the multi-core learning rule, the method maps samples to four state potential nuclear spaces (namely a traffic flow dynamic sub-situation space, a control operation safety sub-situation space, a control operation efficiency sub-situation space and a control work load sub-situation space); the softmax function is then trained separately in the four subspaces to intelligently classify the samples into respective classes in each kernel space.
Training set for situation sequences { x1φ1}，{x2φ2}，…，{xnφnThe softmax loss function (in matrix form) is as follows:
wherein, x'i＝[xi，1](return 1 if the tags match, otherwise return xi)，
Representing the weight and bias of all classes.
Is a row vector whose elements are all equal to 1. y isiIs a one-dimensional vector representation of the label. The potential implication of the loss function J is that if xiBelongs to class j, then x is requirediwj+bjIs as large as possible, thereby increasing the sample xiPossibility of belonging to class j.
In general, the accuracy of the situation sequence training samples increases as the value of the loss function decreases. Loss R of softmax function in the l-th situation kernel spaceemp(fl) The detailed description is as follows:
wherein
yiIs a single heat vector representation of the tag. In each situation kernel space, the loss function of the softmax function must be as good as possibleThe energy is small. In order to combine the loss functions in each situation core space, the loss function weight to be designed based on the primary weight of the civil aviation control operation safety situation evaluation index is used for realizing the softmax intelligent classification under the cooperative multi-core learning framework.
In order to make the learning process in each kernel space cooperate with each other, the patent introduces a regularization term RUAs a collaborative learning constraint, the relations of different situation nuclear spaces are considered, so that the softmax function classifiers in the different situation nuclear spaces can learn each other. In particular, assuming that the number of potential subspaces is equal to 4, in order to obtain supplementary information in the kernel space of different potentials, the regularization term RUThe samples in different potential kernel spaces are required to have consistent output, so that the softmax functions in the different potential kernel spaces can cooperate with each other. Regularization term RUThe formula of (c) is defined as:
where λ is the regular term RUThe weight of (c). The loss function of the complete state potential kernel space can be expressed as:
a situation assessment model training process based on multi-core cooperation adopts a K-fold cross validation method without repeated sampling, and aims to effectively improve the generalization capability of the model. Specifically, N situation samples are equally divided into K groups, K-1 groups of the K groups are used for training a model, then the model obtained through training is used for evaluating the next group of situations, and evaluation errors are calculated. Since there are K choices to select the K-1 group from the K groups, there will be a training set (K-1 group of data) test set. And calculating the error average value of K times to obtain the cross validation error.
(4) Intelligent prediction of civil aviation control operation safety situation
And (3) intelligently predicting the safety situation of civil aviation control operation by adopting a heuristic deep recursion frame.
Specifically, the traditional civil aviation control operation safety situation prediction mainly depends on the experience judgment of control personnel, and the existing prediction method has the defects of insufficient dynamic evolution capability, weak robustness, incapability of avoiding prediction false alarm and the like. Although the technology for predicting the safety situation of the controlled operation is developing towards intellectualization, the theoretical system and the method are still imperfect. The method aims at solving the problem that the civil aviation control operation safety situation cannot be modeled by a linear model due to instability caused by diversity and strong time sequence.
Firstly, a civil aviation control operation safety situation prediction model is perfected, a recursive network framework is adopted to realize the architecture improvement of the situation prediction network model so as to model situation prediction under nonlinear dynamics, so that the situation prediction model has the capability of being separated from a local optimum value, the dynamic evolution capability of the model is increased, and the robustness of the model is improved; and secondly, a risk condition set is formed by combining a long-term historical situation early warning sequence, and the safety risk index and tolerance evaluation under short-term situation prediction are combined, so that the intelligent analysis capability of a prediction model is improved, and the air traffic control safety situation prediction false alarm is avoided.
From the above description, it can be seen that the following points are mainly used to implement the embodiments of the present invention:
(1) aiming at the problem that the reality, the integrity and the consistency of the civil aviation air traffic control operation data are not enough, a novel safety protection and fusion process of the civil aviation air traffic control operation data is innovatively provided, the intellectualization, the intensification and the sharing of the air traffic control operation data are realized, a systematic and collaborative shared multi-source data asset is formed, and the system has better coverage, universality, multi-scale property and high efficiency in an upper-layer framework.
(2) Aiming at the characteristics of insufficient correlation depth, insufficient knowledge fusion degree, difficulty in finding implicit knowledge and the like among civil aviation air traffic control operation data, a multi-scale civil aviation control operation situation representation framework based on a knowledge graph is creatively constructed, the civil aviation control operation knowledge graph with strong multi-scale entity correlation and high knowledge integrity is generated, and the representation learning of the civil aviation control operation multi-scale situation can be realized.
(3) Aiming at the problems of fuzzy safety situation indexes, single evaluation angle and the like of civil aviation control operation, a situation evaluation system formed by control index weight and a safety situation index grade interval is creatively established, a quantitative and associated combination method of multi-scale civil aviation control operation safety situation indexes is provided, and a multi-scale civil aviation control operation safety situation cooperation evaluation framework can be realized. The patent situation classifier has intelligent, multi-scale and sustainable evaluation capacity; the model emphasizes that different situation spaces have consistent output requirements, so that situation characteristics in different kernel spaces can be combined for learning, and a K-fold cross validation method is adopted in the training process to improve the generalization of the classification model. And then, combining with a civil aviation control operation safety situation index system, and accurately evaluating the safety risk in the civil aviation control operation.
(4) In order to overcome the defects that the situation prediction model is insufficient in dynamic evolution capability, poor in robustness and incapable of reducing false alarms and the like, the recursive situation prediction model is constructed in the situation prediction scene of civil aviation control operation safety situations, and a civil aviation control safety situation prediction system is innovatively optimized, so that the prediction accuracy is improved, and the risk of predicting false alarms is reduced.
Based on the above procedures, a multi-scale intelligent situation perception system oriented to civil aviation control operation safety can be formed. The technical process can intelligently analyze and process information and intelligently sense safety risks, and changes the traditional technical situation that the traditional technical situation seriously depends on manual work.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (5)
1. A processing method for a multi-scale situation awareness process of civil aviation traffic control operation is characterized by comprising the following steps:
the safety protection and fusion of the air traffic control operation data comprises the following steps: acquiring empty pipe operation data, and performing data security and fusion processing on the empty pipe operation data based on a block chain technology, a missing empty pipe operation data filling algorithm and a space-time aligned empty pipe operation data fusion algorithm to obtain first data;
the method comprises the following steps of civil aviation control operation safety situation representation: based on the first data, the characterization of the safety situation of civil aviation control operation is realized by adopting feature dimension reduction and extraction, an embedded learning technology and a deep reinforcement learning technology;
the method comprises the following steps of civil aviation control operation multi-scale safety situation assessment: carrying out multi-scale security situation evaluation by adopting a cooperative learning and fusion algorithm;
the method comprises the following steps of intelligent prediction of civil aviation control operation safety situation: adopting a heuristic deep recursion frame to carry out intelligent prediction on the safety situation of civil aviation control operation;
the method specifically comprises the following steps of evaluating the multi-scale safety situation of civil aviation control operation:
constructing an index system of civil aviation control operation safety situation: refining a first-level index and a second-level index which influence the safety situation of the civil aviation control operation by analyzing the requirements of the civil aviation control operation and the safety management service; designing quantitative calibration of indexes of the safety situation of civil aviation control operation, designing index weights according to the influence degree of each index on the safety situation and the actual situation, and dividing a safety situation index grade interval;
and (4) safety situation assessment: inputting the safety situation representation data of civil aviation control operation into the collaborative multi-core situation assessment model, and realizing multi-scale safety situation assessment of civil aviation control operation by combining with the safety situation index grade interval; the cooperative multi-core situation assessment model adopts a softmax function and a multi-core learning theory;
the safety situation assessment specifically comprises:
inputting civil aviation control operation safety situation representation data, wherein each datum is used as a situation sample;
mapping the situation samples to four situation nuclear spaces; the four state potential nuclear spaces are respectively a traffic flow dynamic sub-situation space, a control operation safety sub-situation space, a control operation efficiency sub-situation space and a control work load sub-situation space;
respectively training a softmax function in the four situation nuclear spaces, and carrying out loss weighting;
introducing a regular term as a collaborative learning constraint to obtain a final loss function;
and adopting the final loss function to realize the safety situation evaluation and outputting a situation evaluation result.
2. The method according to claim 1, characterized in that the step of characterizing the safety situation of civil aviation control operation is specifically:
performing feature dimension reduction and extraction on the air traffic control operation data to generate a preliminary knowledge map;
adopting an embedded learning technology to carry out integrity supplement on the knowledge graph;
and carrying out intelligent extraction of implicit relation link on the constructed knowledge graph by adopting a deep reinforcement learning mechanism, thereby realizing the representation of the civil aviation control operation security situation based on the knowledge graph library.
3. The method according to claim 1 or 2, characterized in that the intelligent prediction step of the safety situation of civil aviation control operation is specifically as follows:
improving the architecture of a situation prediction network model by adopting a heuristic deep recursion frame;
a risk condition set is formed by combining a long-term historical situation early warning sequence, and the safety risk index and tolerance evaluation under short-term situation prediction are performed, so that the intelligent analysis capability of a situation prediction network model is improved;
and the situation prediction network model is adopted to realize the intelligent prediction of the civil aviation control operation safety situation.
4. The method according to claim 3, further comprising training the collaborative multi-core situation assessment model using a no-oversampling K-fold cross-validation method, in particular:
equally dividing N situation samples into K groups, training a model by using the K-1 group, then evaluating the next group of situations by using the trained model, calculating evaluation errors, and calculating the error average value of K times to obtain cross validation errors.
5. The method of claim 1, wherein the air traffic control operational data includes meteorological data, flight plans, flow control data, control voice, and surveillance data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110685656.1A CN113344408B (en) | 2021-06-21 | 2021-06-21 | Processing method for multi-scale situation awareness process of civil aviation traffic control operation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110685656.1A CN113344408B (en) | 2021-06-21 | 2021-06-21 | Processing method for multi-scale situation awareness process of civil aviation traffic control operation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113344408A CN113344408A (en) | 2021-09-03 |
| CN113344408B true CN113344408B (en) | 2022-07-05 |
Family
ID=77478316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110685656.1A Active CN113344408B (en) | 2021-06-21 | 2021-06-21 | Processing method for multi-scale situation awareness process of civil aviation traffic control operation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113344408B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114613037B (en) * | 2022-02-15 | 2023-07-18 | 中国电子科技集团公司第十研究所 | Prompt searching method and device for airborne fusion information guide sensor |
| CN114694420B (en) * | 2022-03-23 | 2024-01-12 | 合肥置顶信息技术有限公司 | Civil aviation weather forecast making and publishing system and method capable of intelligently correcting errors |
| CN114898600B (en) * | 2022-07-13 | 2022-10-21 | 安胜（天津）飞行模拟系统有限公司 | UAM four-dimensional track sharing and management method based on block chain technology architecture |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111508280A (en) * | 2020-04-10 | 2020-08-07 | 聂党民 | Mobile time slot aviation control method based on 4D track |
| CN112183605A (en) * | 2020-09-23 | 2021-01-05 | 南京航空航天大学 | Civil aviation control sector classification method based on operation characteristics |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107332698A (en) * | 2017-06-19 | 2017-11-07 | 西北大学 | A kind of Security Situation Awareness Systems and method towards bright Great Wall intelligent perception system |
| CN107944472B (en) * | 2017-11-03 | 2019-05-28 | 北京航空航天大学 | A kind of airspace operation situation calculation method based on transfer learning |
| CN112488066A (en) * | 2020-12-18 | 2021-03-12 | 航天时代飞鸿技术有限公司 | Real-time target detection method under unmanned aerial vehicle multi-machine cooperative reconnaissance |
-
2021
- 2021-06-21 CN CN202110685656.1A patent/CN113344408B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111508280A (en) * | 2020-04-10 | 2020-08-07 | 聂党民 | Mobile time slot aviation control method based on 4D track |
| CN112183605A (en) * | 2020-09-23 | 2021-01-05 | 南京航空航天大学 | Civil aviation control sector classification method based on operation characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113344408A (en) | 2021-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113344408B (en) | Processing method for multi-scale situation awareness process of civil aviation traffic control operation | |
| Yu et al. | Prediction of highway tunnel pavement performance based on digital twin and multiple time series stacking | |
| Guo et al. | Complex power system status monitoring and evaluation using big data platform and machine learning algorithms: a review and a case study | |
| Li et al. | Improved Bayesian network-based risk model and its application in disaster risk assessment | |
| CN112270445B (en) | Flight delay wave and comprehensive evaluation method based on statistical analysis and classification prediction | |
| CN112257914B (en) | Aviation safety causal prediction method based on random forest | |
| CN112530559A (en) | Intelligent medical material allocation system for sudden public health event | |
| CN114331041A (en) | Smart power plant maturity assessment method and system based on information physical system architecture | |
| CN116187640B (en) | Power distribution network planning method and device based on grid multi-attribute image system | |
| CN115734274A (en) | Cellular network fault diagnosis method based on deep learning and knowledge graph | |
| Lin et al. | Analysis and prediction of overloaded extra-heavy vehicles for highway safety using machine learning | |
| Liang et al. | A statistical analysis model of big data for precise poverty alleviation based on multisource data fusion | |
| Zhao et al. | An intelligent evaluation method to analyze the competitiveness of airlines | |
| Xin | Application of optimized support vector machine model in tax forecasting system | |
| Du et al. | Airport capacity prediction with multisource features: A temporal deep learning approach | |
| CN113449936A (en) | Urban space evolution simulation prediction method, device, electronic equipment and storage medium | |
| CN113469504A (en) | Dynamic testing method, medium and system for water traffic risk | |
| Qian et al. | Research on multi-source data fusion in the field of atmospheric environmental monitoring | |
| Li et al. | Real-time prediction of highway equipment faults based on GCN and GRU algorithms | |
| Wang | SVR short-term traffic flow forecasting model based on spatial-temporal feature selection | |
| CN117436708B (en) | Risk assessment method for territorial space planning | |
| LU502895B1 (en) | A collaborative forecasting method for the air traffic state in the context of long-range operation | |
| CN112508276B (en) | Power grid rapid diagnosis and optimization system and optimization method | |
| Yang et al. | Design and Application of Big Data Technology Management for the Analysis System of High Speed Railway Operation Safety Rules | |
| Weng et al. | Performance Evaluation of High-tech Innovation Enterprises Based on Grey Chaotic Neural Network |
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 |