CN113140134B - Intelligent flight delay prediction frame for intelligent air traffic control system - Google Patents

Intelligent flight delay prediction frame for intelligent air traffic control system Download PDF

Info

Publication number
CN113140134B
CN113140134B CN202110267382.4A CN202110267382A CN113140134B CN 113140134 B CN113140134 B CN 113140134B CN 202110267382 A CN202110267382 A CN 202110267382A CN 113140134 B CN113140134 B CN 113140134B
Authority
CN
China
Prior art keywords
air traffic
layer
airport
knowledge graph
graph
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
Application number
CN202110267382.4A
Other languages
Chinese (zh)
Other versions
CN113140134A (en
Inventor
蔡开泉
朱衍波
李悦
王慧
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beihang University
Original Assignee
Beihang University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beihang University filed Critical Beihang University
Priority to CN202110267382.4A priority Critical patent/CN113140134B/en
Publication of CN113140134A publication Critical patent/CN113140134A/en
Application granted granted Critical
Publication of CN113140134B publication Critical patent/CN113140134B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
    • G08G5/0073—Surveillance aids
    • 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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]
    • H04L67/025—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP] for remote control or remote monitoring of applications
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Traffic Control Systems (AREA)

Abstract

The application provides a flight delay intelligent prediction frame towards wisdom air traffic control system. The prediction framework includes: perception layer, transmission layer, processing layer, service layer and application layer. The sensing layer is used for collecting data related to air traffic operation; the transmission layer is used for transmitting data related to air traffic operation to the processing layer; the processing layer is used for filling the data related to the air traffic operation into the knowledge graph facing the air traffic management operation field as the attribute of the entity to obtain the filled knowledge graph, and carrying out knowledge reasoning on the filled knowledge graph to obtain initial decision information corresponding to the data related to the air traffic operation; the service layer provides intelligent air management service for the air traffic department based on the initial decision information when the application layer receives a service request from a client of the air traffic department; and the application layer is used for outputting the instruction to be executed corresponding to the intelligent air traffic control service to the client. Through the prediction framework, reliable interaction of information and knowledge is realized.

Description

Intelligent flight delay prediction frame for intelligent air traffic control system
Technical Field
The application relates to the field of intelligent civil aviation, in particular to a flight delay intelligent prediction framework for an intelligent air traffic control system.
Background
The air traffic management (air traffic management for short) system is used for maintaining air traffic safety and air traffic order and ensuring air traffic smoothness. According to a Communication Navigation monitoring/Air Traffic Management (CNS/ATM) architecture, an airspace partition Management is used as a basis of an Air Traffic Management system. On the basis, all air traffic departments cooperate in division of labor to realize air traffic management. With the increasing of the number of flights and the increasing of the pressure for ensuring the air traffic safety, the limitations of the existing air traffic management system gradually appear: (1) the information sharing between the airplane and the ground facilities (such as communication facilities and the like) and the air traffic department (such as air traffic control department, airline company and airport) is insufficient, so that the actual flight information of the airplane is inconsistent with the information of the ground facilities, even the deviation is large, and further the ground facilities are difficult to accurately predict the flight condition of the airplane. (2) When the air traffic management system based on the airspace partition management is used for air traffic management, partition control is carried out on the whole flight phase, and further the global consideration is insufficient. Moreover, there is a lack of methods and capabilities for collaborative decision-making between air traffic sectors (e.g., air traffic sector, airline, airport), which in turn makes it difficult for air traffic to achieve better operating conditions.
The intelligent air management system is provided for solving the problems that information sharing among the airplane, ground facilities and air traffic departments is insufficient, and global consideration on flight is insufficient. The intelligent air traffic control system can promote operation flows such as air traffic service, flow management and airspace management through technologies such as the Internet of things, cloud computing, mobile internet and big data so as to realize credible interaction of information and knowledge, and then improve monitoring and early warning capacity and emergency response speed of air traffic control operation, and achieve the technical effect of improving air traffic control operation efficiency.
However, the architecture of the intelligent air management system is not yet clear.
Disclosure of Invention
The application provides an intelligent flight delay prediction framework for an intelligent air traffic control system, so as to solve the technical problem that the architecture of the intelligent air traffic control system is not clear yet.
The application provides a flight delay intelligent prediction frame towards wisdom air traffic control system, the prediction frame includes: the system comprises a perception layer, a transmission layer, a processing layer, a service layer and an application layer;
the sensing layer is used for collecting data related to air traffic operation; the data related to the air traffic operation comprises data of at least two attributes, and the data of different attributes come from different kinds of sensing equipment in the sensing layer;
the transmission layer is used for transmitting the data related to the air traffic operation to the processing layer;
the processing layer is used for filling the data related to the air traffic operation into a knowledge graph facing the air traffic management operation field as the attribute of an entity to obtain a filled knowledge graph, and carrying out knowledge reasoning on the filled knowledge graph to obtain initial decision information corresponding to the data related to the air traffic operation;
the service layer is used for providing intelligent air traffic control service for the air traffic department based on the initial decision information when the application layer receives a service request from a client of the air traffic department;
and the application layer is used for outputting the instruction to be executed corresponding to the intelligent air traffic control service to the client.
Optionally, the intelligent air traffic control service includes at least one of the following:
aviation flow prediction, aviation monitoring information, meteorological forecast, aviation information and airport operation situation.
Optionally, the processing layer is specifically configured to obtain a distributed representation of an entity and a relationship of the filled knowledge graph, and learn the knowledge representation of the filled knowledge graph based on the distributed representation of the entity and the relationship to obtain the initial decision information.
Optionally, the service layer is specifically configured to obtain, based on the service request, the initial decision information, and a preset negotiation policy, an instruction to be executed by the air traffic department based on the service request; wherein the negotiating policy comprises: and initiating a mapping relation between the decision information and the instruction.
Optionally, the processing layer is further configured to, before filling the data related to air traffic operation as an attribute of an entity into a knowledge graph facing the air traffic management operation field to obtain the filled knowledge graph, construct an ontology base of the knowledge graph based on a data source in the air traffic management operation field, and obtain a relationship between terms in the ontology base; generating the knowledge graph facing the air traffic management operation field by using the terms in the ontology library and the relations among the terms in the ontology library;
wherein the ontology library comprises: at least one term of the air traffic management operation domain; the terms are entities of the knowledge graph facing the air traffic management operation field, and the relationship among the terms is the edge of the knowledge graph facing the air traffic management operation field.
Optionally, the processing layer is specifically configured to extract terms from the data source by using a preset extraction template; and removing the terms with the confidence coefficient lower than a first preset threshold value from the extracted terms to obtain the ontology base of the knowledge graph.
Optionally, the processing layer is further configured to, after the knowledge graph for the air traffic management operation field is generated, compare the new term with an entity in the knowledge graph for the air traffic management operation field when the new term is extracted from the data source, so as to obtain a similarity of the new term; and when the similarity of the new terms is lower than a second preset threshold, updating the knowledge graph facing the air traffic management operation field by using the new terms.
Optionally, the data related to air traffic operation includes data of at least two dimensions:
time dimension data, space dimension data, business dimension data.
Optionally, the service layer is an air traffic control cloud platform.
Optionally, the application layer is configured to implement at least one of the following functions: wide area traffic collaborative management, busy airport operation, four-dimensional track operation, flexible use of complex airspace, and regional traffic collaborative decision-making.
The application provides a flight delay intelligent prediction frame towards wisdom air traffic control system gathers and air traffic operation relevant data through the perception layer, and the transmission layer will with air traffic operation relevant data transmission extremely the processing layer, the conversion of processing layer realization based on information to knowledge, the air traffic management service of diversification and individuation is realized to the service layer, and the application layer realizes alternately with the user, has realized the credible interaction of information and knowledge, has reached the technological effect that improves air traffic control operating efficiency.
Drawings
In order to more clearly illustrate the technical solutions in the present application or the prior art, the following briefly introduces the drawings needed to be used in the description of the embodiments or the prior art, and obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without inventive labor.
FIG. 1 is a schematic diagram of a CNS/ATM framework;
FIG. 2 is a schematic diagram of an intelligent air traffic control system according to the present application;
FIG. 3 is a flow chart illustrating a method for implementing a process layer function provided herein;
fig. 4 is a flowchart illustrating a method for generating a knowledge graph for the air traffic management operation domain according to the present application.
With the above figures, there are shown specific embodiments of the present application, which will be described in more detail below. These drawings and written description are not intended to limit the scope of the inventive concepts in any manner, but rather to illustrate the inventive concepts to those skilled in the art by reference to specific embodiments.
Detailed Description
To make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely with reference to the accompanying drawings in the present application, and it is obvious that the described embodiments are some, but not all embodiments of the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.
The intelligent air management system is a system built by adopting a CNS/ATM framework. Fig. 1 is a schematic diagram of a CNS/ATM framework, as shown in fig. 1, under which the intelligent air management system can implement the following functions: communications, navigation, surveillance, and Air Traffic Management (ATM).
The ATM may include: air Space Management (ASM), Air Traffic Service (ATS), and Air Traffic Flow Management (ATFM). The ATS may include Flight Information Service (FIS), Aeronautical Weather Service (AWS), World Wide Navigation Warning Service (WWNWS), Search and Rescue (SAR), and the like.
Based on this framework, fig. 2 is a schematic architecture diagram of an intelligent air traffic control system provided in the present application. As shown in fig. 2, according to the functions required to be implemented by the air traffic management system, the architecture of the intelligent air traffic management system can be divided into the following layers:
the sensing layer may include, for example, communication facilities, navigation facilities, surveillance facilities, weather detection equipment, intelligence gathering equipment, and other sensors, etc. For ease of description, the following is referred to collectively as the sensing layer device. The sensing layer equipment can be dispersedly deployed in the space range of airplane flying, landing, taking off and the like, and can be specifically set according to actual requirements. Through the perception layer, the intelligent air traffic control system can realize the functions of communication, navigation and monitoring and can also acquire data related to air traffic operation.
The data related to air traffic operation may include data in at least two dimensions: the data of time dimension, the data of space dimension, the data of business dimension to improve the intelligence air traffic control system to above-mentioned data analysis's accuracy. Wherein, the data of different dimensions come from different sensing devices in the sensing layer. For example, weather data from weather detection devices, communication data from communication facilities, and navigation data from navigation facilities, etc.
That is, the data related to the air traffic operation acquired by the sensing layer can cover data with different dimensions of time, space, service and the like and different attributes.
The transport layer may include, for example, a professional communication Network in an air traffic control system such as an Aviation Fixed Telecommunication Network (AFTN), a data link Network, a monitoring Network, an Aviation Telecommunication Network (ATN), a satellite Network, or other public networks. The intelligent air management system can transmit data related to air traffic operation to the processing layer through the various networks.
After the data related to the air traffic operation is acquired, the processing layer may implement, for example, the functions shown in fig. 1, such as data storage, data cleaning, data modeling, data mining, data interconnection, data evolution analysis, and data maintenance, for the acquired data. As a possible implementation manner, the processing layer can use a knowledge graph facing the air traffic management operation field to carry out reasoning analysis, obtain initial decision information corresponding to the data related to the air traffic operation, and realize the conversion from information-based (namely, the data related to the air traffic operation) to knowledge.
And the service layer is used for providing intelligent air traffic control service for the air traffic department based on the initial decision information obtained by the processing layer when the application layer receives a service request from a client of the air traffic department. By the conversion from knowledge to intelligence, diversified and personalized intelligent air traffic control service is realized. For example, dedicated services such as traffic prediction services, monitoring Information services, weather forecast services, and aviation Information services, and general services such as System Wide Information Management (SWIM), cloud platform, visualization and simulation, virtual reality, and intelligent decision making.
And the application layer is used as an input/output interface of the intelligent air traffic control system, realizes interaction with the user and provides intelligent air traffic control service for the user. That is, departments such as an airline company, an air traffic control department, and an airport can output a to-be-executed instruction corresponding to the intelligent air traffic control service to the client by using the intelligent air traffic control system through the corresponding client interface (realized by an application layer, for example, the application layer can receive a service request from the client of the air traffic control department). The application layer as an operation intelligence of the system architecture of the intelligent air traffic control system can, for example, implement at least one of the following applications: wide area traffic coordination management, busy airport operation, four-dimensional track operation, complex airspace use, and regional traffic coordination decision-making, etc.
The following explains the concepts involved in the above-mentioned "wide area traffic coordination management, busy airport operation, four-dimensional track operation, complex airspace usage, and, regional traffic coordination decision".
Wide area traffic cooperative management: by establishing a multi-party cooperative decision-making mechanism of an area and airport multi-level flight flow management system, an air management department, an airport, an airspace user and the like, the dynamic balance of airspace capacity and flight requirements is realized. The intelligent air traffic control reconstructs a Flight Information Exchange Model (FIXM) by means of a knowledge map technology, and a semantic network among aviation data is constructed based on a data mining technology, so that the formation of a Flight flow Information collaborative environment is promoted. In addition, track flow hot spot areas and bottleneck points are identified through prediction technologies such as an artificial neural network, and resource allocation, planned tracks, airspace organization and the like before flight are dynamically adjusted by combining meteorological information and utilizing a cooperative decision mechanism, so that real-time capacity demand balance is met.
And (3) operation of a busy airport: as an integral part of the air traffic management system, airports must provide the necessary ground facilities for aircraft operation to improve airport throughput and operational efficiency under all weather conditions. The intelligent air traffic control enhances the dynamic real-time information perception capability of flight operation time nodes by means of technologies such as knowledge graph, voice recognition and natural language processing, tracks and analyzes the occurrence time of key milestone events in real time, and then actively warns abnormal events. In addition, the inbound and outbound flights are sequenced through a multi-objective optimization method, the takeoff/landing time is reasonably formulated, the sliding time and the ground waiting are reduced, the comprehensive coordination of the inbound and outbound of all aircrafts is realized, the congestion phenomenon of an airport and a peripheral terminal area is improved, the sliding path of the airport surface aircraft is optimized by means of algorithms such as a graph theory, a graph convolution neural network and the like, and the ground vehicles and the aircrafts are automatically guided to slide, so that the comprehensive running capacity of the airport surface is enhanced.
Four-dimensional track operation: the air management department, the airline company and the aircraft realize the integration of the whole process of flight planning and flight implementation by sharing, negotiating and managing dynamic flight paths. The intelligent air traffic control improves the integration capability of all flight information through technologies such as situation awareness and data mining, and obtains more accurate flight tracks. In addition, the intelligent air traffic control cooperative sharing capability can further promote the synchronous flight trajectory of the aircraft before takeoff, and each aircraft can receive detailed and accurate constraint information, so that the performance of the whole network is improved.
The complex space domain is flexibly used: the method can be mainly divided into spatial domain organization and management, and civil aviation and other types of aviation jointly operate. The airspace organization and management meet the requirements of different types of flight activities, traffic volumes and different levels of service by dividing different airspace structures, and an airspace management mechanism is formulated to meet the requirements of all related parties. The intelligent air traffic management excavates the requirements of various airspace users by means of technologies such as data mining and expert systems, assists in making an airspace use plan and distributes and uses airspace resources in real time. In addition, hot sectors and air routes are identified by adopting technologies such as graph theory, complex network and the like, the air route/sector operation efficiency is optimized based on an intelligent optimization algorithm, congestion is reduced, and dynamic planning and configuration of the sectors are further realized. The civil aviation and other types of aviation jointly run, the airspace resource utilization efficiency is improved through a flexible and efficient management system, and the interconnection and intercommunication of the civil aviation and other types of aviation infrastructures are realized through unified planning, respective implementation and overall improvement. The intelligent air management system integrates the operation networks of the air management systems of the civil aviation and the other types of aviation within a safe and controllable range, and the interconnection and the intercommunication of the civil aviation and the other types of aviation cooperatively operated physical infrastructures are realized. In addition, a knowledge graph technology is adopted to construct an information exchange model for consistent cognition of civil aviation and other types of aviation, the same semantics of information transmission among the civil aviation and other types of aviation are realized, and further the interoperation of the civil aviation and other types of aviation information systems is achieved.
And (3) area flow cooperative decision: based on a Collaborative decision-making (CDM), the collaboration between regulatory agencies, airports, and airlines is enhanced. Through the cooperation among all the related parties, a fair and efficient traffic management mode is realized. The intelligent air traffic control system collects all structured data of control departments, airports and airlines by means of a knowledge map technology, and realizes interconnection and intercommunication of information systems among all related parties. In addition, based on the multi-objective optimization technology, the independent decision making situation of an airline company is broken through, all related parties negotiate flight information at the same time, and therefore the flight decision making rationality and the resource utilization efficiency are improved.
It should be understood that the processing layer and the service layer may be disposed on the same hardware entity, or may be implemented by different hardware entities. For example, the processing layer may be composed of at least one processing node, which may be, for example, a server. The service layer can be an air traffic control cloud platform, and the intelligent air traffic control system is abstracted into one service through the air traffic control cloud platform, so that users of air traffic control related departments can obtain data required by the users by using corresponding clients of the users, or the intelligent air traffic control service is provided for the air traffic control departments.
The following describes in detail how to implement air traffic management by using the intelligent air traffic management system according to the specific embodiment based on the intelligent air traffic management system shown in fig. 2. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
Taking the "wide area traffic cooperative management" involved in the application layer as an example, the intelligent air traffic control system first uses the "sensing layer" to collect data such as flight, weather, Broadcast Automatic Dependent Surveillance (ADS-B) and the like from communication facilities, navigation facilities, monitoring facilities, weather detection equipment and other facilities. Next, the data is transmitted to a "processing layer" through a network such as an AFTN, an ATN, and a data link network in the "transport layer", and the processing layer in the intelligent air traffic control system fills the flight data into a knowledge map for the air traffic control operation field as an attribute of an entity, integrates data from different sensing devices, and restores the association between the data (which may also be referred to as data activation). And then, carrying out knowledge reasoning on the filled knowledge graph to accurately and comprehensively analyze data related to the air traffic operation so as to obtain initial decision information aiming at the data, thereby forming intelligent air traffic control service. The method is corresponding to special services such as flow prediction service, weather forecast service and aviation information service in a service layer and general services such as SWIM, cloud platform and intelligent decision. Based on the general and special services, wide area traffic cooperative management in an application layer is finally realized.
Fig. 3 is a schematic flowchart of a method for implementing a processing layer function provided in the present application, and as shown in fig. 3, the method includes the following steps:
and S101, acquiring data related to air traffic operation.
The processing layer can acquire data related to air traffic operation through a private communication network in the intelligent air traffic control system. For example, the processing layer may obtain data from the communication facilities in the awareness layer through an AFTN, ATN, data link network, or the like. For example, the data may be voice communication data of the pilot with the controller, or the like.
And S102, filling data related to air traffic operation into the knowledge graph facing the air traffic management operation field by taking the data related to the air traffic operation as attributes of entities to obtain the filled knowledge graph.
After the processing layer acquires the data related to the air traffic operation, the data related to the air traffic operation can be activated in a mode of filling the data related to the air traffic operation into a knowledge graph facing the air traffic management operation field by taking the data as the attribute of an entity.
For example, the processing layer may construct an airport network knowledge graph by using flight data obtained by the sensing layer, implement information-to-knowledge conversion, and capture the time correlation and spatial dynamics of flight delays by using the airport network knowledge graph.
The airport network knowledge graph may be described as a directed weighted network. For example, at time t, the directed weighting network may be represented as G(t)＝(V，E(t)，W(t)). Wherein V ═ { V ═ V1，v2，…，vnDenotes the set of all airport nodes in the airport network. E(t)Representing the set of edges between nodes in the set of airport nodes.
Figure GDA0003453121070000081
Represents G(t)The weighted adjacency matrix of (2). Elements thereof
Figure GDA0003453121070000082
Shows the airport v during (t-1, t)iAnd airport vjNumber of flights in between. In addition to this, the present invention is,
Figure GDA0003453121070000091
for a delay vector representing n airports at time t. Wherein each element
Figure GDA0003453121070000092
Record the airport v during (t-1, t)iThe delay time of (c). Wherein the content of the first and second substances,
Figure GDA0003453121070000093
can be expressed as the following formula (1):
Figure GDA0003453121070000094
wherein the content of the first and second substances,
Figure GDA0003453121070000095
indicating a departure from an airport v during (t-1, t)iThe total delay time for flights leaving the port,
Figure GDA0003453121070000096
and
Figure GDA0003453121070000097
respectively representing airport v during (t-1, t)iThe number of cancelled flights and the number of flights scheduled to take off, p represents the equivalent delay value for cancelled flights. Illustratively, ρ may take a value of 180, for example.
S103, carrying out knowledge reasoning on the filled knowledge graph to obtain initial decision information corresponding to the data related to air traffic operation.
As a possible implementation manner, the knowledge inference process can be implemented by the following method:
the processing layer can obtain the distributed representation of the entity and the relation of the filled knowledge graph, and then the processing layer learns the knowledge representation of the filled knowledge graph based on the distributed representation of the entity and the relation to obtain initial decision information.
Illustratively, the processing layer may use a network structure-based deep-embedded representation learning approach to obtain a distributed representation of entities and relationships of the populated knowledge-graph. The intelligent air management system takes the filled knowledge graph as the input of deep embedding representation learning based on the network structure, and the distributed representation of the entity and the relation of the filled knowledge graph is output and obtained through the method.
For example, the processing layer may use a neural network algorithm based on random walk of meta-paths to learn the knowledge representation of the populated knowledge graph with the distributed representation of the entity and the relationship as input, so as to realize the knowledge inference of the populated knowledge graph and obtain the initial decision information. It should be understood that the knowledge inference process may also be implemented by using other existing manners of implementing knowledge inference based on distributed representation of entities and relationships, and the method of knowledge inference described above is not limited in this application.
It should be appreciated that the populated knowledgegraph of air traffic operation-related data does not correspond to the initial decision information one-to-one, e.g., the processing layer may infer multiple pieces of initial decision information from one populated knowledgegraph.
For example, in the time dimension, considering the airport network structure with time variation, the application will correspond the adjacency matrix W at the time ttAnd historical delay time series X(t)Defined as a graph snapshot. Each graph snapshot (directed graph) is a multiple relationship graph with input and output relationships. Thus, the processing layer may first capture spatial correlations in a single Graph snapshot using a correlation-Graph Convolutional neural network (R-GCN). Then, the processing layer mines the flight delay time dynamics between two adjacent graph snapshots according to the Markov property. Finally, the processing layer may obtain a time-graph volume block that captures flight delay time-varying features.
For a single graph snapshot, with an adjacency matrix W(t)And historical delay time series X(t)As input, a spectrogram convolution operator based on a single graph snapshot
Figure GDA0003453121070000101
As shown in the following equation (2):
Figure GDA0003453121070000102
wherein, thetaFRepresenting the set of parameters used in a single graph snapshot. σ () represents an activation function.
Figure GDA0003453121070000103
A set of relationship types is represented as a set,
Figure GDA0003453121070000104
representing the input relationship of the network node,
Figure GDA0003453121070000105
is a diagonal matrix, and
Figure GDA0003453121070000106
Figure GDA0003453121070000107
representing the network node output relationship. Min,MoutWeight matrices, M, representing input and output relationships, respectively0Representing a self-join matrix.
Further, the processing layer may employ a linear combination using input and output relationships to simplify the self-join matrix to avoid over-fitting the R-GCN model. That is, the above equation (2) may be replaced with the following equation (3):
Figure GDA0003453121070000108
wherein, the first and the second end of the pipe are connected with each other,
Figure GDA0003453121070000109
Inrepresenting an n-dimensional identity matrix.
Furthermore, given the large number of closely-related aviation resources (e.g., aircraft, crew, passengers, and infrastructure), the operation of a single airport at time t is closely related to other airports at the previous time t-1, and thus, the processing layer may use the Markov property to model the temporal correlation of flight delays. The graph convolution operations of two adjacent graph snapshots are taken as a combination of the current state and the accumulated historical graph snapshots. Considering two adjacent graph snapshots at time t-1 and time t, the processed adjacency matrix
Figure GDA00034531210700001010
And historical delay time series X(t-1),X(t)As input, then populates the graph volume calculator
Figure GDA00034531210700001011
Two adjacent graph snapshots are processed. The processing procedure can be shown as the following formula (4):
Figure GDA00034531210700001012
wherein, thetaBRepresenting the use of two adjacent graph snapshotsThe set of parameters is then set to,
Figure GDA00034531210700001013
represents the graph convolution operation at time t-1, and
Figure GDA00034531210700001014
and X(t-1)As an input. Wherein the parameter set ΘHDoes not change with time.
Finally, the processing layer may implement the memory-accumulated history map snapshot with hidden states and generate new states using a combination of hidden states and current inputs. This new state is shown by the following equation (5):
Figure GDA00034531210700001015
wherein Θ isBComprises thetaHAnd ΘF. Referring to the foregoing method, the processing layer may apply equation (5) by loop-through to process a time-varying airport network graph (i.e., a series of graph snapshots).
In the spatial dimension, considering the unknown emergency requirements of temporary flight paths, the application provides an adaptive volume block which can capture the unknown spatial interaction hidden in the airport network. Specifically, as shown in the following formula (6): the processing layer may hide the last hidden state of the temporal volume block
Figure GDA0003453121070000111
Corresponding adjacency matrix, and, finally, hidden state
Figure GDA0003453121070000112
As input to an adaptive volume block to obtain a volume operator
Figure GDA0003453121070000113
Figure GDA0003453121070000114
Wherein the content of the first and second substances,
Figure GDA0003453121070000115
representing the last hidden state of the previous temporal convolution block
Figure GDA0003453121070000116
A corresponding adjacency matrix. θ represents a parameter set in spatial volume block modeling. I.C. AnRepresenting an n-dimensional identity matrix.
Figure GDA0003453121070000117
Is a diagonal matrix, and Qii＝∑jAij。Q-1Representing the transposed matrix of Q.
Further, it is contemplated that in a practical situation, when an airport experiences a severe flight delay or emergency, the airport's neighbors (airports within 300 kilometers of the airport) may help absorb reserve flights even without planned airlines. Thus, if one airport experiences a severe delay, its neighbor airports may also incur high delays due to the take-up of the reserve flight. Thus, the present application proposes an adaptive volume block to capture complex dynamic associations between airports.
Specifically, based on the planned route structure of an airport network, the processing layer may use the parameterized importance matrix Z and the similarity matrix S to obtain the graph volume calculator according to equation (7) below
Figure GDA0003453121070000118
Figure GDA0003453121070000119
Where M represents the planned route structure of the airport network.
Figure GDA00034531210700001110
Representing the importance of each airport in an airport networkSex (its function is the same as the attention mechanism). In addition, the present application does not limit the value of each element in Z. The processing layer may be trained with other parameters through input data to update elements in Z.
Figure GDA00034531210700001111
Representing the similarity of each airport in the airport network.
The processing layer carries out delay prediction on each airport in the airport network through a deep learning model formed by the time map volume block and the self-adaptive map volume block, and the prediction result can monitor the flight delay abnormal condition of the whole airport network and early warn delay propagation in the airport network.
And S103, providing intelligent air traffic control service for the air traffic department based on the initial decision information.
After obtaining the initial decision information, the service layer of the intelligent air traffic control system can provide intelligent air traffic control service for the air traffic department based on the initial decision information.
As a possible implementation manner, the service layer may obtain, based on the initial decision information and the mapping relationship between the initial decision information and the instructions, the instructions corresponding to the initial decision information, and send the instructions to the clients of the corresponding departments, so that the departments execute the corresponding instructions, thereby implementing the intelligent air traffic control service.
As another possible implementation manner, the service layer may obtain the instruction to be executed of the at least one air traffic department based on the initial decision information and a preset negotiation policy, and then send the corresponding instruction to be executed to the client of the at least one air traffic department, thereby implementing the intelligent air traffic control service.
Wherein the negotiation policy may include: and initiating a mapping relation between the decision information and the instruction. And the service layer obtains the instruction corresponding to the initial decision information according to the mapping relation between the initial decision information and the instruction. Through the negotiation strategy, the cooperative decision of each service unit aiming at the initial decision information can be realized. The negotiation policy may be stored in the service layer after being negotiated in advance among the departments performing air traffic management.
As described above, the processing layer may obtain the filled-in knowledgegraph by filling the data related to the air traffic operation as the attribute of the entity into the knowledgegraph facing the air traffic management operation field. The knowledge graph facing the air traffic management operation field can be generated by an actual airport network structure, and can also be generated by a processing layer in the following mode, specifically:
fig. 4 is a flowchart illustrating a method for generating a knowledge graph for the air traffic management operation domain according to the present application. As shown in fig. 4, the method comprises the steps of:
s201, constructing an ontology base of the knowledge graph based on a data source in the air traffic management operation field.
The data source can be an encyclopedia website, a vertical website and the like related to the air traffic control system.
As a possible implementation, the processing layer may extract terms from the data source by using a preset extraction template. Taking H site as an example, the terms in H site are extracted by a pattern matching method, for example. Illustratively, the above term may be: tower controllers, access controllers, regulatory directives, airlines, airports, and like terms. Of course, the term may be obtained from the data source by other existing methods (e.g., web crawler, etc.) for extracting the term, which is not limited herein.
Optionally, the processing layer may use the obtained terms as an ontology library for constructing a knowledge graph facing the air traffic management operation field. Or, after obtaining the above terms, the processing layer may verify the above terms according to an existing general term set in the air traffic management field to obtain a confidence of the above terms. The confidence may characterize the accuracy of the extracted term. The processing layer then removes terms with confidence levels below a first predetermined threshold from the above terms to eliminate terms that may be in error. The final term set is used as an ontology base for constructing the knowledge graph. By the method, the accuracy of the constructed knowledge graph facing the air traffic management operation field can be improved. The setting method and the value of the first preset threshold are not limited in the present application. For example, the first preset threshold may be calculated according to the confidence value of the extracted term, or may be set according to the actual requirement of the user.
S202, obtaining the relation among the terms in the ontology library.
After the processing layer obtains the ontology library, terms in the ontology library are independent from each other, so that the processing layer needs to acquire the relationship between the terms in the ontology library. The method of obtaining the relationship between the above terms is not limited in this application. For example, the processing layer may obtain the relationship between the above terms by a minimum spanning tree method.
S203, generating the knowledge graph facing the air traffic management operation field by using the terms in the ontology library and the relationship among the terms in the ontology library.
And the processing layer takes the terms in the ontology base as entities of the knowledge graph and takes the relationship among the terms as edges of the knowledge graph, so that the knowledge graph facing the air traffic management operation field is generated.
As a possible implementation manner, the processing layer may also update the knowledge graph after generating the knowledge graph facing the air traffic management operation field. For example, the processing layer may generate a new knowledge-graph in the manner shown in FIG. 4 after using the knowledge-graph for a period of time. For another example, the processing layer may also update the knowledge graph by using an incremental update method. For example, the processing layer may extract terms from the data sources, and for the extracted new terms, the new terms may be compared with entities already existing in the knowledge graph to obtain similarity of the new terms. If the similarity of the new term is lower than a second preset threshold value, the knowledge graph does not contain the term, the knowledge graph facing the air traffic management operation field is updated by using the new term, and the new term is added into the knowledge graph as a new entity; and if the similarity of the new term is higher than a second preset threshold value, not updating the knowledge graph. The setting method and the value of the second preset threshold are not limited in the present application. For example, the second preset threshold may be calculated according to the similarity value of the extracted terms, or may be set according to the actual requirement of the user.
The processing layer adopts a knowledge map increment updating method, and adds a new entity in the knowledge map facing the air traffic management operation field, so that not only can frequent reconstruction of the knowledge map facing the air traffic management operation field caused by randomness in air traffic operation be avoided, but also the accuracy of the knowledge map can be further improved by adding the new entity.
In this embodiment, an ontology base of the knowledge graph is constructed through a data source in the air traffic management operation field, the knowledge graph facing the air traffic management operation field is constructed based on the relationship among terms in the ontology base, and then the knowledge graph can be updated incrementally, so that the accuracy of the knowledge graph facing the air traffic management operation field is improved, and further, the processing layer obtains more accurate initial decision information corresponding to the data related to the air traffic operation.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application 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 or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present application.

Claims (8)

1. An intelligent flight delay prediction framework oriented to an intelligent air traffic control system, the prediction framework comprising: the system comprises a perception layer, a transmission layer, a processing layer, a service layer and an application layer;
the sensing layer is used for collecting data related to air traffic operation; the data related to the air traffic operation comprises data of at least two attributes, the data of different attributes come from different kinds of sensing devices in the sensing layer, and the data related to the air traffic operation comprises data of at least two dimensions: time dimension data, space dimension data, business dimension data;
the transmission layer is used for transmitting the data related to the air traffic operation to the processing layer;
the processing layer is used for filling the data related to the air traffic operation into a knowledge graph facing to the air traffic management operation field as the attribute of an entity to obtain the filled knowledge graph, and performing knowledge reasoning on the filled knowledge graph to obtain initial decision information corresponding to the data related to the air traffic operation;
the service layer is used for providing intelligent air traffic control service for the air traffic department based on the initial decision information when the application layer receives a service request from a client of the air traffic department;
the application layer is used for outputting a to-be-executed instruction corresponding to the intelligent air traffic control service to the client;
the service layer is specifically used for acquiring a to-be-executed instruction of the air traffic department based on the service request, the initial decision information and a preset negotiation strategy; wherein the negotiating policy comprises: initiating a mapping relationship between the decision information and the instruction;
the processing layer utilizes flight data obtained by the sensing layer to construct an airport network knowledge graph, the airport network knowledge graph is a directed weighting network, and the directed weighting network is represented as G(t)＝(V，E(t)，W(t)) Where t represents time, V ═ V1，v2，…，vnDenotes the set of all airport nodes in the airport network, E(t)Representing a set of edges between nodes in the set of airport nodes,
Figure FDA0003606694350000011
represents G(t)Weighted adjacency matrix of (2) elements
Figure FDA0003606694350000012
Shows the airport v during (t-1, t)iAnd airports vjNumber of flights in between;
Figure FDA0003606694350000013
delay vector for representing n airports at time t, where each element
Figure FDA0003606694350000014
Record the airport v during (t-1, t)iThe delay time of (a), wherein,
Figure FDA0003606694350000015
is shown as
Figure FDA0003606694350000016
Wherein, the first and the second end of the pipe are connected with each other,
Figure FDA0003606694350000021
indicating a departure from an airport v during (t-1, t)iThe total delay time for flights leaving the port,
Figure FDA0003606694350000022
and
Figure FDA0003606694350000023
respectively representing airport v during (t-1, t)iThe number of cancelled flights and the number of flights scheduled to take off, wherein rho represents the equivalent delay value of the cancelled flights;
the processing layer deduces a plurality of pieces of initial decision information from a filled knowledge graph;
in the time dimensionDegree, the adjacent matrix W corresponding to the time t(t)And historical delay time series X(t)Defined as a graph snapshot, for a single graph snapshot, with a adjacency matrix W(t)And historical delay time series X(t)
Figure FDA0003606694350000024
Wherein, thetaFRepresents the set of parameters used in a single graph snapshot, σ () represents the activation function,
Figure FDA0003606694350000025
a set of relationship types is represented as a set,
Figure FDA0003606694350000026
representing the input relationship of the network node,
Figure FDA0003606694350000027
is a diagonal matrix, an
Figure FDA0003606694350000028
Figure FDA0003606694350000029
Representing network node output relationships, Min，MoutWeight matrices, M, representing input and output relationships, respectively0Representing a self-join matrix;
taking the time map convolution operation of two adjacent map snapshots as a combination of the current state and the accumulated historical map snapshot; considering two adjacent graph snapshots at time t-1 and time t, the processed adjacency matrix
Figure FDA00036066943500000210
And historical delay time series X(t-1)，X(t)As input, then generalize the time chart volume calculator
Figure FDA00036066943500000211
Two adjacent graph snapshots are processed, and the processing procedure is shown as the following formula:
Figure FDA00036066943500000212
wherein, thetaBA parameter set representing the use of two adjacent graph snapshots,
Figure FDA00036066943500000213
representing a time-graph convolution operation at time t-1 with
Figure FDA00036066943500000214
And X(t-1)As input, among others, the parameter set ΘHDoes not change with time;
the processing layer adopts the hidden state to realize the history map snapshot of memory accumulation, and uses the combination of the hidden state and the current input to generate a new state, wherein the new state is shown as the following formula:
Figure FDA00036066943500000215
wherein, thetaBComprises thetaHAnd ΘFThe processing layer processes the time-varying airport network map through the formula;
in the spatial dimension, the processing layer integrates the last hidden state of the temporal map volume block
Figure FDA00036066943500000216
Corresponding adjacency matrix, and, finally, hidden state
Figure FDA00036066943500000217
As input to a spatial graph convolution operation to obtain a graph convolution operator
Figure FDA00036066943500000218
Figure FDA00036066943500000219
Wherein, the first and the second end of the pipe are connected with each other,
Figure FDA00036066943500000220
representing the last hidden state of the time map volume block
Figure FDA00036066943500000221
Corresponding adjacency matrix, theta denotes a parameter set in the spatial map volume block modeling, InRepresents an n-dimensional identity matrix of the cell,
Figure FDA0003606694350000031
is a diagonal matrix, and Qii＝∑jAij，Q-1A transposed matrix representing Q;
based on the planned route structure of an airport network, the processing layer uses a parameterized importance matrix Z and a similarity matrix S to obtain an adaptive graph volume calculator according to the following formula
Figure FDA0003606694350000032
Figure FDA0003606694350000033
Wherein M represents a planned route structure of the airport network,
Figure FDA0003606694350000034
representing the importance of each airport in the airport network, the processing layer is trained with other parameters, via input data, to update elements in Z,
Figure FDA0003606694350000035
representing the similarity of each airport in the airport network;
the processing layer carries out delay prediction on each airport in the airport network through a deep learning model formed by a time map volume block and a self-adaptive map volume block, the prediction result monitors the flight delay abnormal condition of the whole airport network, and early warning is carried out on delay propagation in the airport network.
2. The predictive framework of claim 1, wherein the intelligent air management services include at least one of:
aviation flow prediction, aviation monitoring information, meteorological forecast, aviation information and airport operation situation.
3. The prediction framework of claim 1, wherein the processing layer is specifically configured to obtain a distributed representation of entities and relationships of the populated knowledgegraph, and learn the knowledgerepresentation of the populated knowledgegraph based on the distributed representation of entities and relationships to obtain the initial decision information.
4. The prediction framework of any one of claims 1 to 3, wherein the processing layer is further configured to, before the data related to the air traffic operation is filled into the knowledge graph facing the air traffic management operation field as the attribute of the entity to obtain the filled knowledge graph, construct an ontology base of the knowledge graph based on the data source of the air traffic management operation field, and obtain the relationship between the terms in the ontology base; generating the knowledge graph facing the air traffic management operation field by using the terms in the ontology library and the relations among the terms in the ontology library;
wherein the ontology library comprises: at least one term in the air traffic management operations domain; the terms are entities of the knowledge graph facing the air traffic management operation field, and the relationship among the terms is the edge of the knowledge graph facing the air traffic management operation field.
5. The prediction framework of claim 4, characterized in that the processing layer is specifically configured to extract terms from the data source using a preset extraction template; and removing the terms with the confidence coefficient lower than a first preset threshold value from the extracted terms to obtain the ontology base of the knowledge graph.
6. The prediction framework of claim 4, wherein the processing layer is further configured to, after generating the air traffic management operation domain-oriented knowledge graph, compare the new term with an entity in the air traffic management operation domain-oriented knowledge graph when the new term is extracted from the data source to obtain a similarity of the new term; and when the similarity of the new terms is lower than a second preset threshold, updating the knowledge graph facing the air traffic management operation field by using the new terms.
7. The prediction framework of any of claims 1-3, wherein the service layer is an air traffic cloud platform.
8. The prediction framework of any of claims 1-3, wherein the application layer is configured to implement at least one of the following functions: wide area traffic collaborative management, busy airport operation, four-dimensional track operation, flexible use of complex airspace, and regional traffic collaborative decision-making.
CN202110267382.4A 2021-03-12 2021-03-12 Intelligent flight delay prediction frame for intelligent air traffic control system Active CN113140134B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110267382.4A CN113140134B (en) 2021-03-12 2021-03-12 Intelligent flight delay prediction frame for intelligent air traffic control system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110267382.4A CN113140134B (en) 2021-03-12 2021-03-12 Intelligent flight delay prediction frame for intelligent air traffic control system

Publications (2)

Publication Number Publication Date
CN113140134A CN113140134A (en) 2021-07-20
CN113140134B true CN113140134B (en) 2022-07-08

Family

ID=76811013

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110267382.4A Active CN113140134B (en) 2021-03-12 2021-03-12 Intelligent flight delay prediction frame for intelligent air traffic control system

Country Status (1)

Country Link
CN (1) CN113140134B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113808440A (en) * 2021-09-09 2021-12-17 许兵 Air traffic management system based on resource operation
CN114493206A (en) * 2022-01-13 2022-05-13 湖北国际物流机场有限公司 BIM-based global airport management system and method
CN115240475B (en) 2022-09-23 2022-12-13 四川大学 Aircraft approach planning method and device fusing flight data and radar image

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB0703508D0 (en) * 2007-02-22 2007-04-04 Iordanova Blaga N Global control mechanisms for air traffic and environmentally sustainable airports
CN103413462A (en) * 2013-07-18 2013-11-27 北京航空航天大学 Air traffic network flow optimizing method fully taking airspace jam and flight delay into consideration
CA3020024A1 (en) * 2017-10-06 2019-04-06 Tata Consultancy Services Limited System and method for flight delay prediction
CN111738425A (en) * 2020-08-07 2020-10-02 北京航空航天大学 Airport delay reason analysis method based on graph mining
WO2020240177A1 (en) * 2019-05-28 2020-12-03 Sita Information Networking Computing Uk Limited System and method for flight arrival time prediction

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107783973B (en) * 2016-08-24 2022-02-25 慧科讯业有限公司 Method, device and system for monitoring internet media event based on industry knowledge map database
CN110222127B (en) * 2019-06-06 2021-07-09 中国电子科技集团公司第二十八研究所 Knowledge graph-based information aggregation method, device and equipment
CN110569369A (en) * 2019-09-16 2019-12-13 神州数码融信软件有限公司 Generation method and device, application method and device of knowledge graph of bank financial system
CN111583723A (en) * 2020-04-30 2020-08-25 中国航空无线电电子研究所 Air traffic information management system architecture based on big data and cloud computing technology
CN112307217B (en) * 2020-09-16 2021-12-14 北京中兵数字科技集团有限公司 Knowledge graph model construction method and device, and storage medium
CN112182241A (en) * 2020-09-24 2021-01-05 四川大学 Automatic construction method of knowledge graph in field of air traffic control
CN112216151B (en) * 2020-10-15 2021-12-28 北京航空航天大学 Air traffic four-dimensional track regulation and control decision method
CN112241401A (en) * 2020-10-16 2021-01-19 中国民用航空华东地区空中交通管理局 Knowledge graph-based digital quality management system and method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB0703508D0 (en) * 2007-02-22 2007-04-04 Iordanova Blaga N Global control mechanisms for air traffic and environmentally sustainable airports
CN103413462A (en) * 2013-07-18 2013-11-27 北京航空航天大学 Air traffic network flow optimizing method fully taking airspace jam and flight delay into consideration
CA3020024A1 (en) * 2017-10-06 2019-04-06 Tata Consultancy Services Limited System and method for flight delay prediction
WO2020240177A1 (en) * 2019-05-28 2020-12-03 Sita Information Networking Computing Uk Limited System and method for flight arrival time prediction
CN111738425A (en) * 2020-08-07 2020-10-02 北京航空航天大学 Airport delay reason analysis method based on graph mining

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
邢志伟等.基于贝叶斯网的离港航班滑行时间动态估计.《计算机工程与应用》.2018,(第24期), *

Also Published As

Publication number Publication date
CN113140134A (en) 2021-07-20

Similar Documents

Publication Publication Date Title
CN113140134B (en) Intelligent flight delay prediction frame for intelligent air traffic control system
Chavhan et al. IoT-based context-aware intelligent public transport system in a metropolitan area
Belcastro et al. Using scalable data mining for predicting flight delays
Bennell et al. Airport runway scheduling
Sridhar et al. Modeling and optimization in traffic flow management
Cai et al. A deep learning approach for flight delay prediction through time-evolving graphs
Lujak et al. Centrality measures for evacuation: Finding agile evacuation routes
Choi et al. Artificial neural network models for airport capacity prediction
CN113299120B (en) Intelligent sensing system for air traffic situation supported by edge cloud in cooperation
US20210358313A1 (en) Airport capacity prediction system
Sun et al. A dual decomposition method for sector capacity constrained traffic flow optimization
Jun et al. Towards a greener Extended-Arrival Manager in air traffic control: A heuristic approach for dynamic speed control using machine-learned delay prediction model
Cao et al. Link transmission model for air traffic flow management
US10762458B1 (en) Satellite scheduling system
Misra et al. S-NAV: Safety-aware IoT navigation tool for avoiding COVID-19 hotspots
Pettet et al. Hierarchical planning for resource allocation in emergency response systems
Zhang et al. A hierarchical heuristic approach for solving air traffic scheduling and routing problem with a novel air traffic model
US20220147664A1 (en) Systems and Methods for Multi-Modal Transportation Simulation Verification
Sandamali et al. Two-stage scalable air traffic flow management model under uncertainty
Yang et al. Origin-destination matrix prediction via hexagon-based generated graph
Sekine et al. Multi-objective takeoff time optimization using cellular automaton-based simulator
Sridhar et al. Lessons Learned in the Application of Machine Learning Techniques to Air Traffic Management
Wang et al. Efficient resource allocation in multi-UAV assisted vehicular networks with security constraint and attention mechanism
Lundberg et al. Human-in-the-loop AI: Requirements on future (unified) air traffic management systems
Haselein et al. Multiple machine learning modeling on near mid-air collisions: An approach towards probabilistic reasoning

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