CN108898838A - A kind of aerodrome traffic congestion prediction technique and device based on LSTM model - Google Patents
A kind of aerodrome traffic congestion prediction technique and device based on LSTM model Download PDFInfo
- Publication number
- CN108898838A CN108898838A CN201810878081.3A CN201810878081A CN108898838A CN 108898838 A CN108898838 A CN 108898838A CN 201810878081 A CN201810878081 A CN 201810878081A CN 108898838 A CN108898838 A CN 108898838A
- Authority
- CN
- China
- Prior art keywords
- road
- congestion
- prediction
- information
- lstm model
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000000694 effects Effects 0.000 claims description 43
- 239000013598 vector Substances 0.000 claims description 26
- 230000004913 activation Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 claims description 8
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 claims description 8
- 238000000714 time series forecasting Methods 0.000 claims description 7
- 241001269238 Data Species 0.000 claims description 5
- 230000001550 time effect Effects 0.000 abstract description 4
- 230000002123 temporal effect Effects 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 20
- 238000010586 diagram Methods 0.000 description 13
- 241001123248 Arma Species 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 230000003111 delayed effect Effects 0.000 description 5
- 230000001419 dependent effect Effects 0.000 description 5
- 238000013459 approach Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000007667 floating Methods 0.000 description 2
- 239000003595 mist Substances 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000010006 flight Effects 0.000 description 1
- 239000000295 fuel oil Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 239000003921 oil Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000032258 transport Effects 0.000 description 1
- 238000001744 unit root test Methods 0.000 description 1
- 230000017260 vegetative to reproductive phase transition of meristem Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Educational Administration (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Analytical Chemistry (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Traffic Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of aerodrome traffic congestion prediction technique and device based on LSTM model, this method includes：In real time obtain airport periphery preset range in road traffic related information and air station flight landing information and airport periphery preset range in aeronautical meteorology information；By the traffic related information, flight landing information and aeronautical meteorology information, LSTM model is inputted；The output of the LSTM model is obtained as a result, the output result is to predict the congestion index of road in the preset range of future time period airport periphery.This method is based on the considerations of room and time effect, and increase aeronautical meteorology information, the system that traffic in region is regarded as a temporal and spatial correlations obtains prediction result based on LSTM model, can further promote the accuracy of congestion in road exponential forecasting in the preset range of airport periphery.
Description
Technical field
The present invention relates to computer science and field of intelligent transportation technology, in particular to a kind of airport based on LSTM model
Traffic congestion prediction technique and device.
Background technique
Quickly propelled with Urbanization in China, an important factor for urban congestion has become puzzlement urban development it
One, how to improve urban transportation operational efficiency, alleviation congestion pressure has become city and realizes that sustainable and healthy development must solve
Certainly the problem of.In face of challenge, each big city and navigation companies issue " traffic congestion delay index " in succession, as traffic administration
With the important means of guidance.Such as, " traffic index " and " traffic circulation analysis report " of Beijing Communication committee publication, Gao De
" the Chinese main cities congestion ranking " of map publishing etc..
It can effectively improve urban traffic control level using these index informations, but traffic index is only capable of reflecting current friendship
Logical situation lacks foresight to the variation of future traffic condition.And traffic congestion can not only bring out the reduction of line efficiency, go out
Row cost, accident rate increase, and can because oil consumption increase caused by imperfect combustion with fuel oil energy waste and
Atmosphere pollution.How to avoid and alleviates traffic congestion as people's concern.
On the other hand, with the arrival of the application of car networking technology and big data era, people can use multiple means
Traffic congestion is monitored, such as Andrea etc. [1] using GPS tracker and smart phone identify real-time traffic congestion and
Accident, Kong etc. is identified and is predicted to urban traffic blocking by Floating Car track data, provides the interconnection of road conditions service
Net company is also based primarily upon floating vehicle model, and Bauza and Gozalvez are based between vehicle and vehicle, vehicle and infrastructure node
Information exchange monitors road traffic congestion.
In recent decades, scholar has been developed that a variety of models with regard to forecasting traffic flow problem, is generally divided into following a few classes：
(1) using Time Series Method as representative, to models such as traffic congestion delay index construction ARIMA, following a period of time is predicted (such as
It, hour, minute etc.) traffic condition, such as Voort has studied application of the ARIMA model on forecasting traffic flow,
Williams etc. carries out the short-term prediction of traffic flow using seasonality ARIMA model；(2) k nearest neighbor, support vector machines, kernel function
Nonparametric techniques, such as the Smith comparative studies prediction effect of non parametric regression and seasonality ARIMA model such as return, in
Shore etc. is based on k nearest neighbor nonparametric technique and carries out short-term prediction to traffic flow, establishes traffic flow with non parametric regression up to celebrating east etc.
Flow and rate pattern；(3) with bayes methods such as Gaussian processes, traffic condition is considered as random process and models and predicts, such as
Health military affairs etc.；(4) application of Kalman filter method on forecasting traffic flow, such as XIEYuan-chang, XUDong-wei etc.
The research of people.(5) research of the short-time traffic flow forecast technology based on artificial neural network, such as Yao Zhihong, condition love force et al..
These methods obtain good application effect under its specific scene, but still have much room for improvement.
Summary of the invention
In view of the above problems, the present invention provides a kind of aerodrome traffic congestion prediction techniques and dress based on LSTM model
It sets, this method increases aeronautical meteorology information based on the considerations of room and time effect, and traffic in region is regarded as a space-time
Relevant system obtains prediction result based on LSTM model, can further promote congestion in road in the preset range of airport periphery and refer to
The accuracy of number prediction.
In a first aspect, the embodiment of the present invention provides a kind of aerodrome traffic congestion prediction technique based on LSTM model, including：
The traffic related information and air station flight landing information, Yi Jiji of road in the preset range of airport periphery are obtained in real time
Aeronautical meteorology information in the preset range of field periphery；
By the traffic related information, flight landing information and aeronautical meteorology information, LSTM model is inputted；
The output of the LSTM model is obtained as a result, the output result is prediction future time period airport periphery preset range
The congestion index of interior road.
In one embodiment, the LSTM model training process is as follows：
The traffic related information and air station flight landing information, Yi Jiji of road in the preset range of airport periphery are obtained in real time
Aeronautical meteorology information in the preset range of field periphery；
By the traffic related information, flight landing information and aeronautical meteorology information, it is associated according to the time；
Congestion in road index, aeronautical data and the meteorological data in per hour are obtained after association, merge festivals or holidays, working day
With other activity time sequence datas, data set is constituted；
Hysteresis effect is increased to the aeronautical data and meteorological data, generates the LSTM model.
In one embodiment, hysteresis effect is increased to the aeronautical data, meteorological data, generates the LSTM model,
Including：
LSTM model is used for time series forecasting, obtains output vector htAfterwards, a full articulamentum is connected, thus finally
Obtain predicted value：
Wherein, σ indicates activation primitive；WoutTo connect layer matrix entirely, dimension is 1 × dh；htIndicate output vector；boutTable
Show intercept item；For the predicted value of the s articles time series t+1 phase.
In one embodiment, the loss function of the LSTM model, including：MAE and MAPE loss function；
Forward prediction n step loss function formula be：
In formula (8), MAEsIt is the index name for measuring prediction effect；S indicates road；T indicates the time；L indicates the time
Sequence；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction；
In formula (9), MAPEsIt is the index name that another measures prediction effect；S indicates road；T indicates the time；L table
Show time series；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction.
In one embodiment, the traffic related information of road in the preset range of airport periphery is obtained in real time, including：
It obtains the vehicle flowrate and travel speed of road in the preset range of airport periphery in real time by data-interface, calculates road
Congestion index.
Second aspect, the embodiment of the present invention provide a kind of aerodrome traffic congestion prediction meanss based on LSTM model, including：
Module is obtained, is risen for obtaining the traffic related information of road and air station flight in the preset range of airport periphery in real time
Aeronautical meteorology information in information and airport periphery preset range drops；
Input module, for inputting LSTM mould for the traffic related information, flight landing information and aeronautical meteorology information
Type；
Prediction module, for obtaining the output of the LSTM model as a result, the output result is prediction future time period machine
The congestion index of road in the preset range of field periphery.
In one embodiment, the LSTM model includes：
Acquisition submodule, for obtaining the traffic related information and air station flight of road in the preset range of airport periphery in real time
Aeronautical meteorology information in landing information and airport periphery preset range；
Input submodule is used for by the traffic related information, flight landing information and aeronautical meteorology information, according to the time
It is associated；
Submodule is constituted, for obtaining congestion in road index, aviation number in per hour after being associated with according to input submodule
According to and meteorological data, merge festivals or holidays, working day and other activity time sequence datas, constitute data set；
Submodule is generated, for increasing hysteresis effect to the aeronautical data and meteorological data, generates the LSTM model.
In one embodiment, the generation submodule is obtained specifically for LSTM model is used for time series forecasting
Output vector htAfterwards, a full articulamentum is connected, to finally obtain predicted value：
Wherein, σ indicates activation primitive；WoutTo connect layer matrix entirely, dimension is 1 × dh；htIndicate output vector；boutTable
Show intercept item；For the predicted value of the s articles time series t+1 phase.
In one embodiment, the loss function of the LSTM model, including：MAE and MAPE loss function；
Forward prediction n step loss function formula be：
In formula (8), MAEsIt is the index name for measuring prediction effect；S indicates road；T indicates the time；L indicates the time
Sequence；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction；
In formula (9), MAPEsIt is the index name that another measures prediction effect；S indicates road；T indicates the time；L table
Show time series；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction.
In one embodiment, the traffic shape of road in the preset range of airport periphery is obtained in the acquisition submodule in real time
Condition information, including：
It obtains the vehicle flowrate and travel speed of road in the preset range of airport periphery in real time by data-interface, calculates road
Congestion index.
The beneficial effect of above-mentioned technical proposal provided in an embodiment of the present invention includes at least：
A kind of aerodrome traffic congestion prediction technique based on LSTM model provided in an embodiment of the present invention, obtains airport in real time
Aviation in the traffic related information of road and air station flight landing information and airport periphery preset range in the preset range of periphery
Weather information；By the traffic related information, flight landing information and aeronautical meteorology information, LSTM model is inputted；Described in acquisition
The output of LSTM model is as a result, the output result is to predict that the congestion of road in the preset range of future time period airport periphery refers to
Number.This method increases aeronautical meteorology information based on the considerations of room and time effect, and traffic in region is regarded as a space-time
Relevant system obtains prediction result based on LSTM model, can further promote congestion in road in the preset range of airport periphery and refer to
The accuracy of number prediction.
Other features and advantages of the present invention will be illustrated in the following description, also, partly becomes from specification
It obtains it is clear that understand through the implementation of the invention.The objectives and other advantages of the invention can be by written explanation
Specifically noted structure is achieved and obtained in book, claims and attached drawing.
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
Detailed description of the invention
Attached drawing is used to provide further understanding of the present invention, and constitutes part of specification, with reality of the invention
It applies example to be used to explain the present invention together, not be construed as limiting the invention.In the accompanying drawings：
Fig. 1 is the flow chart of the aerodrome traffic congestion prediction technique provided in an embodiment of the present invention based on LSTM model；
Fig. 2 is RNN model structure schematic diagram；
Fig. 3 is the structure chart of classics RNN model (left side) and LSTM model (right side)；
Fig. 4 is LSTM model module structure chart；
Fig. 5 is the flow chart of LSTM model training process provided in an embodiment of the present invention；
Fig. 6 is the schematic diagram of in July, 2017 Capital Airport high speed congestion delay index；
Fig. 7 is airport expressway congestion delay index and air passenger flow magnitude relation schematic diagram；
Fig. 8 is the aerodrome traffic congestion prediction meanss block diagram provided in an embodiment of the present invention based on LSTM model；
Fig. 9 is the block diagram of LSTM model provided in an embodiment of the present invention.
Specific embodiment
Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
It is fully disclosed to those skilled in the art.
Shown in referring to Fig.1, the aerodrome traffic congestion prediction technique provided in an embodiment of the present invention based on LSTM model, packet
It includes：S101~S103；
S101, in real time obtain airport periphery preset range in road traffic related information and air station flight landing information,
And aeronautical meteorology information in the preset range of airport periphery；
S102, by the traffic related information, flight landing information and aeronautical meteorology information, input LSTM model；
S103, the output of the LSTM model is obtained as a result, the output result is that prediction future time period airport periphery is pre-
If the congestion index of road in range.
In the present embodiment, above-mentioned steps S101 obtains the traffic related information of road in the preset range of airport periphery in real time,
For example be vehicle flowrate, the automobile driving speed on road, congestion in road index can be calculated in turn.
By taking Beijing Capital Airport as an example, 58, airport periphery road traffic and traveling speed are obtained in real time by data-interface
Degree, and then congestion in road index is calculated, calculation formula is highway layout speed per hour divided by real-time average overall travel speed；It is also possible to lead to
Cross other existing calculated congestion in road indexes of calculation.Such as：If current vehicle flow is 0, congestion index 1
(indicating not congestion completely), the bigger expression road more congestion of congestion index finally obtains following data structure：
Beijing Capital Airport flight landing information is obtained in real time, including the practical country reaches sortie, the practical country to intelligent
Number, the practical worlds reach sortie, practical country's arrival number, the practical country set out sortie, the practical country set out number, practical state
Border set out sortie, the practical world set out number, the plan country reach sortie, plan domestic arrival number, the plan world reaches frame
It is secondary, plan domestic arrival number, the plan country set out number, the plan world of sortie, the plan country of setting out is set out sortie, plan state
Border is set out number.
By airport data-interface, Capital Airport flight data is obtained in real time, it is real particularly when there is air traffic control
Border aeronautical data is 0, but planning data is not still 0.Wherein, obtaining Flight Information is specific flight detail, such as " 21: 03
Point, fly to the flight in Shanghai from Beijing ", the above data that obtain are summarized on each hour, i.e., by specific flight
Information summarizes for data, and the data meaning after summarizing is that " country reaches the sum of sortie between when at 20170809 days 0 to 1
And the sum of arrival number etc. ".And by data organization for such as flowering structure：
The aviations such as airport periphery aeronautical meteorology information, including wind speed, wind direction, temperature, precipitation, thunder and lightning, mist, dirt are obtained in real time
Influence factor；
Real-time aeronautical meteorology message data, including customary weather message (METAR) and spy are obtained by meteorological data interface
Different weather message (SPECI) extracts the weather information in message by packet parsing module, and special meteorological data is every half
The message data of hour takes meteorology of the meteorological condition nearest from the integral point moment as the integral point moment in modeling, such as meteorological report
The literary time be 21 points 05 minute, modeling when we as the meteorological data of 21 integral points, be as follows by aeronautical data tissue
Structure：
By above-mentioned traffic related information, flight landing information and aeronautical meteorology information, LSTM model is inputted；Including mould
Type dependent variable and model independent variable；Model dependent variable is such as the hour level congestion index of 58, airport periphery road, is amounted to
58 dependent variables.Model independent variable includes that meteorological data, flight data and congestion index lag item.Wherein meteorological variables packet
It includes：Wind speed, wind direction, temperature, precipitation, thunder and lightning, mist, dirt amount to 7 variables；Aviation variable includes：The practical country reaches sortie, reality
The border country reaches that number, the practical world reach sortie, set out sortie, the practical country of practical country's arrival number, the practical country goes out
Hair number, the practical world set out number, the plan country of sortie, the practical world of setting out reach sortie, plan domestic arrival number, meter
Draw the world reach sortie, plan domestic arrival number, the plan country set out number, the plan world of sortie, the plan country of setting out goes out
Hair sortie, plan the world set out number amount to 16 variables；The second-order lag item of congestion index such as predicts the congestion when morning 9
Index, second-order lag are 7 points of congestion index of 8 a.m. and the morning.
The output of LSTM model is as a result, i.e.：For the congestion index of road in prediction future time period airport periphery preset range.
When predicting future traffic condition, using aeronautical data, meteorological data and family's holiday data as input, it is brought into model and transports
It calculates, obtains the congestion index predicted value of 58, airport periphery road.This method is increased based on the considerations of room and time effect
Aeronautical meteorology information, the system that traffic in region is regarded as a temporal and spatial correlations obtain prediction result based on LSTM model, can be into
One step promotes the accuracy of congestion in road exponential forecasting in the preset range of airport periphery.
LSTM model was proposed by SeppHochreiter and J ü rgenSchmidhuber in 1997 earliest, for solving
A kind of certain variations that gradient disappears in RNN model or gradient expands.By introducing multiple thresholdings in RNN, so that in model
In the case that parameter is fixed, different moments integral be can change, to avoid gradient disappearance or expansion issues.LSTM model is in sequence
Columns is it is predicted that aspect achieves thrilling effect.
LSTM model is one of most widely used widest model in RNN model, and the feature of RNN model maximum is in the hidden of it
Hiding layer output not only connects output layer, is also connected with the hidden layer of subsequent time.Fig. 2 is RNN model structure schematic diagram, equal sign right side
It is the RNN network structure after expansion, including input layer X, hidden layer modules A and output layer h, wherein replicated blocks A is in a network
Any time structure is all the same.Unlike typical neural network, the output of modules A includes identical two parts, portion transmitting
To output layer, portion is transferred to next hidden module.Since modules A only passes through a tanh activation in classical RNN model
Function converts input variable, causes RNN model in processing sequence data, especially long sequence when, be frequently run onto gradient
The problem of disappearing with gradient expansion, eventually leads to model and does not restrain.
Referring to shown in Fig. 3, LSTM model is based on RNN model development, and difference is that intermediate replicated blocks are different.?
It include that (wherein three are by the activation primitive of four dependent interactions in LSTM model duplication model (also known as storage unit)
Sigmoid function, one is tanh function) three kinds of threshold structures constituting, it is to forget door, input gate and out gate respectively, leads to
It crosses this threshold structure and carrys out outputting and inputting for analog switch control information, realize that training error can be propagated with forward and reverse,
To reach training pattern to convergent purpose.
It is now assumed that there is raw time series variable in kM exogenous variableIt enablesSo present target is to establish LSTM model prediction future endogenous variable
S=1 ..., k.In order to enable LSTM model can keep remembering to sequence, lag item x need to be used when constructing samplet-1, xt-2, because
This when LSTM model is used for time series, can be moved back for this angle in the case where activation primitive is linear transformation
VAR model is turned to, therefore VAR model is actually a special case of LSTM model.
Referring to shown in Fig. 4, LSTM model core is three kinds of threshold structures in storage unit, and storage unit finally obtains shape
State vector CtWith output vector ht, specifically include following three processes：
1) it is made of formula 1 and forgets thresholding (forget gate), by the way that connection matrix W is arrangedfTo determine from last moment
Output vector ht-1Which information middle removal retains
ft=σ (Wf·[ht-1, xt]+bf) formula (1)
Wherein, xtFor dk+mDimension is originally inputted, ht-1For dhDimension output, hidden layer dimension are dc(usual dh=dc), [ht-1,
xt] indicate two vectors splicing longer vector, dimension dh+dk+m, WfIt is d to forget door coefficient matrix dimensionc×(dh+
dk+m), bfFor bias vector, dimension dc, f is obtained after activation primitive σ is converted point by pointt, dimension dc。
2) input gate (input gate) carrys out determining means state vector CtHow this updates, specific as follows：
it=σ (Wf·[ht-1, xt]+bi) formula (2)
Vector i is obtained by formula (2)t, dimension dc, transitory state vector is obtained secondly by formula (3)Finally
By upper phase state vector Ct-1With interim shape vectorWeighted sum updates state vector Ct。
3) formula (5) and formula (6) constitute out gate (output gate), are obtaining most recent units state vector CtAfterwards, it utilizes
Tanh activation primitive obtains output vector ht
ot=σ (Wo·[ht-1, xt]+bo) formula (5)
ht=ot*tanh(Ct) formula (6)
Structure is identical between module in LSTM model, and weight matrix W is shared between disparate modulesf、Wi、WcAnd Wo,
This controls model parameter scale.
When LSTM model is used for time series forecasting, output vector h is obtainedtAfterwards, a full articulamentum can be also connected, from
And predicted value is finally obtained, as shown in formula (7)
Wherein, σ indicates activation primitive；WoutTo connect layer matrix entirely, dimension is 1 × dh；htIndicate output vector；boutTable
Show intercept item, also referred to as amount of bias is a parameter to be estimated；For the predicted value of the s articles time series t+1 phase.
MAE (mean absolute error, mean absolute error) and MAPE (mean absolute are selected herein
Percentage error, average absolute percentage error) loss function as LSTM model, mainly there is two o'clock consideration：First,
Time series forecasting is substantially regression problem, thus is not suitable for the cross entropy generally used in selection sort problem as loss
Function；Second, MAE and MAPE have better robustness compared to MSE (mean square error, mean square error).It is pre- forward
Shown in loss function such as formula (8)-(9) for surveying n step
One in expression road s, such as 58 roads, the true congestion index in l moment (when such as morning 9), and
Indicate the congestion index of prediction, n indicates the step-length (n hours future of prediction) of forward prediction, MAEsIt is the finger for measuring prediction effect
Entitling claims, and full name in English is Mean Absolute Error.
One in expression road s, such as 58 roads, the true congestion index in l moment (when such as morning 9), and
Indicate the congestion index of prediction, n indicates the step-length (n hours future of prediction) of forward prediction, MAPEsIt is that another measures prediction effect
The index name of fruit, full name in English are Mean AbsolutePerccent Error.
The method for solving LSTM model is BPTT (Back Propagation Through Time) algorithm, is based on
BPTT algorithm can conduct gradient in heterogeneous networks layer and different moments, and then update model parameter.
In one embodiment, above-mentioned LSTM model training process is as follows：Referring to Figure 5, including：S501~S504；
S501, in real time obtain airport periphery preset range in road traffic related information and air station flight landing information,
And aeronautical meteorology information in the preset range of airport periphery；
S502, by the traffic related information, flight landing information and aeronautical meteorology information, be associated according to the time；
Obtained after S503, association per hour in congestion in road index, aeronautical data and meteorological data, merge festivals or holidays,
Working day and other activity time sequence datas constitute data set；
S504, hysteresis effect is increased to the aeronautical data and meteorological data, generates the LSTM model.
In step S503, to obtaining and all data of parsing are associated, it is associated according to the time, it can be with after association
Obtain congestion in road index, aeronautical data and meteorological data hourly, at the same will according to information such as festivals or holidays association in time,
Constitute complete data set.Wherein, congestion in road index is dependent variable, other all data are independent variable.
The step of training pattern, specifically includes：Increase aeronautical data, meteorological data increases hysteresis effect.For example, setting current
When moment is the morning 10 on the 9th of August in 2017, when being further added by the morning 9 on the 9th of August in 2017 in the independent variable of prediction, 2017 8
When month morning 8 on the 9th, the aviation and meteorological data at these three moment when the morning 7 on the 9th of August in 2017, the reason is that the friendship at current time
Logical situation can be influenced by factors such as the aviations at preceding several moment, after reaching such as flight, passenger need certain time take luggage or
Ground traffic tools are taken, this past backward apparent time lag.
Carry out the accuracy of proof analysis prediction technique of the present invention below by specific embodiment.
1 data explanation and descriptive analysis
For example the present invention is on August 1,00 2016 using data:00:00 up to 31 days 23 July in 2017:00:00 Beijing
Traffic congestion delay index (being provided by high moral) of local 58 roads of Captical International Airport and flight entering and leaving port data are (in
Economic Growth of Civil Aviation Transportation scientific and technical research institute provides), time granularity is hour grade, amounts to 8756 time points.Thus it arranges and derives
Variable be shown in Table 1.Wherein, congestion delay index refers between city dweller's average primary trip practical traveling time and under freestream conditions
Ratio, the evaluation index as congestion level.
Table 1：Variable declaration
By data descriptive analysis, following rule is found：
(1) different road congestion conditions differences are obvious
58 congestion in road delay index modeling and forecasting local to Beijing Capital International Airport, uses y1..., y58It indicates
The congestion delay exponential time sequence data of 58 roads.The jam situation of every road is there are larger difference, wherein airport goods
Transporting road, Airport Expwy and Airport Expwy bypass is three roads the busiest, and average congestion is delayed index 2 1.3 or more, and
Seldom there is congestion in remaining road.Wherein, airport shipping road congestion delay index highest, average congestion delay index reached 1.9, but
The reason is that taxi is waited in line, rather than real traffic congestion.The main roads on three connection areas of Beijing and airport are gathered around
For stifled situation there are larger difference, Airport Expwy is connection airport and the shortest Freeway in areas of Beijing, and congestion is delayed index most
Height, followed by Airport Expwy bypass, although and airport shipping road plays shunting function to airport passenger flow, it is even sooner or later high
The peak period, this road is also without there is apparent traffic congestion.The characteristics of differing greatly for every congestion in road situation, this
Text models every road respectively, and when modeling considers that may be present between road influence each other.
(2) congestion in road periodic feature is obvious
Fig. 6 is indicated 2017 7MonthCapital machine?High speed congestion delay index, can from Fig. 6See, the congestion of Capital Airport high speed
The index that is delayed is presented significantly periodically, and working day, Saturday, all daily variations are obvious, and workaday peak period typically occurs in
7:00~11:00, congestion is delayed index 2 or more, and Saturday substantially whole day is not in congestion, and the congestion on Sunday usually goes out
Afternoon 13 now:00~17:00.Therefore, when modeling introduce the period, week and whether workaday time factor.It can from Fig. 6
To observe that congestion delay index and period have similar to sinusoidal relationship, consideration is introduced into frequency analysisKnowTwo, to indicate period effect, similarly consider to introduce
Four to indicate week and month effect, wherein t indicates the period, and w indicates week, and m indicates month.Compared to the introducing period, star
The dummy variable of phase, month, handle is advantageous in that the number that can not only reduce parameter to be estimated in this way, but also considers the period
Property.
(3) traffic above-ground and aviation factor are closely related
Referring to shown in Fig. 7, into (out) ETA estimated time of arrival and plan into (out) ETA estimated time of arrival there may be inconsistent since flight is practical,
Arrival & departure flights information is processed herein, is obtained per hour by corporation plan time and practical disengaging two dimensions of ETA estimated time of arrival
Entering and leaving port volume of the flow of passengers data.Rule of thumb judge, people would generally be shifted to an earlier date 1-2 hours by the plan Departure airport of flight and be reached
Airport, and flight is practical approach after leave airport within 1-2 hours.Fig. 7 show lag period plan departure from port number and it is practical into
Hongkong people's number and Airport Expwy congestion delay index timing diagram, field high speed congestion be delayed index and the lag period plan departure from port number and
There are stronger correlativities for the number that actually approaches, and related coefficient is calculated 0.6 or so.Therefore, 1 phase of lag is introduced herein
It with the plan of 2 phases of lag departure from port number and actually approaches number 4 variables relevant to aviation factor, for predicting traffic above-ground
Congestion delay index.
3.2 dataset construction
Modeling herein includes following variable：58 roads and its 2 ranks lag item add up to 174 variables, according to plan the time from
Hongkong people's number and add up to 4 variables and 7 time variables by 1 rank of real time to Hongkong people's number and 2 ranks lag item, amounts to 185
A variable, sample size 8754.(the plan departure from port number of 3 phases of lag and the related coefficient of congestion in road delay index only have
0.23, the number that actually approaches and congestion in road delay index also only have 0.15, therefore only consider the aviation visitor of 1 phase of lag and 2 phases
Data on flows.Due to considering 2 ranks lag, sample size loss 2.)
To construct steady prediction model, over-fitting and comparison model prediction effect are prevented, herein divides data set
For training set, verifying collects and three parts of test set, wherein preceding 80% sequence data is as training set (August 1 day 00 in 2016:
00:00 to 2017 on May 19,23:00:00), rear 20% data are as cross validation collection (20 days 00 May in 2017:00:00
To 24 days 23 July in 2017:00:00), last weekly data of data set (25 days 00 July in 2017:00:In July, 00 to 2017
31 days 23:00:00) it is used as test set.Training set for establishing model and estimating model parameter, join for preference pattern by verifying collection
Model over-fitting is counted and prevents, test set is for comparing final prediction result.MAE and MAPE is chosen herein to imitate as model prediction
The evaluation index of fruit, the two value is smaller to show that forecast result of model is better.
3.3 comparison model
Linear model ARMA and VAR model is chosen herein as comparison model.To all time serieses involved in research
Unit root test is carried out, refuses unit root null hypothesis under 0.05 significance, shows that time series is steady, Ke Yijian
Vertical ARMA and VAR model.Arma modeling and VAR model are shown in formula (10)-(11).
Wherein,Indicate that the congestion delay index of t phase kth article road, DS indicate the departure from port number of time according to plan, AR
It indicates to arrive Hongkong people's number by the real time, WE indicates whether workaday dummy variable.
Different from LSTM model, arma modeling and VAR model are linear model.Arma modeling only considers every road certainly
The influence of body lag period, and VAR model not only considers the influence of itself lag period, is also added into the shadow of other roads lag period
It rings, that is, considers space correlation that may be present.
4 model results and analysis
Such as choose 58 roads high-speed bidirectional road in the Capital Airport therein, i.e., from the Capital Airport to Dongzhimen direction and
From Dongzhimen to Capital Airport direction, the prediction effect for comparison model.(being shown in Table 2-3)
2 Capital Airport of table high speed (Capital Airport to Dongzhimen direction) congestion delay exponential forecasting compares
3 Capital Airport of table high speed (Dongzhimen to Capital Airport direction) congestion delay exponential forecasting compares
In terms of prediction effect, LSTM model is almost unanimously better than ARMA and VAR model in all time span of forecasts.For airport height
The fast Capital Airport to Dongzhimen direction, in time span of forecast, the MAPE mean value of LSTM model is only 8.6%, the prediction evaluated with MAE
Effect ratio ARMA promotes 42%, promotes 22% than VAR.Compared with the LSTM model for not considering aviation factor, aviation factor is considered
The prediction effect of LSTM model improve 13%, it was confirmed that the particularity of airport roadway traffic forecast, i.e., by aviation factor shadow
Sound is larger.
For Airport Expwy Dongzhimen to Capital Airport direction, in time span of forecast, the MAPE mean value of LSTM model is only
9.0%, prediction effect also has promotion by a relatively large margin compared to ARMA and VAR, improves 20% and 10% respectively, and does not consider
The LSTM model of aviation factor is compared, and prediction effect also has promotion, and the amplitude that promoted is 9%.
From the point of view of the stability of two indexs of MAE and MAPE, the two shows there is no significantly increasing as time goes by
Model prediction ability has good robustness.
Above-described embodiment is based on Beijing Capital International Airport Air-Ground data and predicts airport roadway congestion delay index.
Confirm that the promotion of the air passenger flow amount road pavement traffic congestion prediction effect of lag period has a significant impact, it is left that MAE promotes 10%
It is right；Secondly, the prediction effect of the LSTM model based on deep learning algorithm is unanimously better than linear model ARMA and VAR model, in advance
Precision is surveyed to be obviously improved.It can be concluded that, when predicting traffic congestion, different roads are widely different due to road attribute in the embodiment of the present invention
Different, the factor for influencing traffic condition is also different, if after it sufficiently can analyze the reason of leading to congestion in road, prediction model
Effect, which has, to be obviously improved.
Method provided in an embodiment of the present invention is conducive to the prediction effect for promoting airport road traffic congestion delay index,
To which the work offer to regulatory authorities is more accurately instructed, such as deployment of traffic-police's police strength etc..Future studies are also
It is further contemplated that the factor of traffic condition, such as weather conditions, festivals or holidays factor and traffic accident are more influenced, into
One step promotes the accuracy of airport roadway prediction.
Based on the same inventive concept, the embodiment of the invention also provides a kind of aerodrome traffic congestion based on LSTM model is pre-
Device is surveyed, by principle and the aforementioned aerodrome traffic congestion prediction technique phase based on LSTM model of the solved problem of the device
Seemingly, therefore the implementation of the device may refer to the implementation of preceding method, and overlaps will not be repeated.
The embodiment of the invention also provides a kind of the aerodrome traffic congestion prediction meanss based on LSTM model, the device reference
Shown in Fig. 8,
Module 81 is obtained, for obtaining the traffic related information and air station flight of road in the preset range of airport periphery in real time
Aeronautical meteorology information in landing information and airport periphery preset range；
Input module 82, for inputting LSTM for the traffic related information, flight landing information and aeronautical meteorology information
Model；
Prediction module 83, for obtaining the output of the LSTM model as a result, the output result is prediction future time period
The congestion index of road in the preset range of airport periphery.
In one embodiment, the LSTM model includes：
Acquisition submodule 91, for obtaining the traffic related information of road and airport boat in the preset range of airport periphery in real time
Aeronautical meteorology information in class's landing information and airport periphery preset range；
Input submodule 92, for by the traffic related information, flight landing information and aeronautical meteorology information, according to when
Between be associated；
Submodule 93 is constituted, for obtaining the congestion in road index in per hour, aviation after being associated with according to input submodule
Data and meteorological data merge festivals or holidays, working day and other activity time sequence datas, constitute data set；
Submodule 94 is generated, for increasing hysteresis effect to the aeronautical data and meteorological data, generates the LSTM mould
Type.
In one embodiment, the generation submodule 94 is obtained specifically for LSTM model is used for time series forecasting
To output vector htAfterwards, a full articulamentum is connected, to finally obtain predicted value：
Wherein, σ indicates activation primitive；WoutTo connect layer matrix entirely, dimension is 1 × dh；htIndicate output vector；boutTable
Show intercept item；For the predicted value of the s articles time series t+1 phase.
In one embodiment, the loss function of the LSTM model, including：MAE and MAPE loss function；
Forward prediction n step loss function formula be：
In formula (8), MAEsIt is the index name for measuring prediction effect；T indicates the time；L indicates congestion index；N is indicated
The step-length of forward prediction；Indicate road s；Indicate the congestion index of prediction；
In formula (9), MAPEsIt is the index name that another measures prediction effect；S indicates road；T indicates the time；L table
Show time series；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction.
In one embodiment, the traffic of road in the preset range of airport periphery is obtained in the acquisition submodule 91 in real time
Condition information, including：
It obtains the vehicle flowrate and travel speed of road in the preset range of airport periphery in real time by data-interface, calculates road
Congestion index.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The shape for the computer program product implemented in usable storage medium (including but not limited to magnetic disk storage and optical memory etc.)
Formula.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (10)
1. a kind of aerodrome traffic congestion prediction technique based on LSTM model, which is characterized in that including：
The traffic related information of road and air station flight landing information and airport week in the preset range of airport periphery are obtained in real time
Aeronautical meteorology information in the preset range of side；
By the traffic related information, flight landing information and aeronautical meteorology information, LSTM model is inputted；
The output of the LSTM model is obtained as a result, the output result is road in prediction future time period airport periphery preset range
The congestion index on road.
2. the method as described in claim 1, which is characterized in that the LSTM model training process is as follows：
The traffic related information of road and air station flight landing information and airport week in the preset range of airport periphery are obtained in real time
Aeronautical meteorology information in the preset range of side；
By the traffic related information, flight landing information and aeronautical meteorology information, it is associated according to the time；
Obtained after association per hour in congestion in road index, aeronautical data and meteorological data, merge festivals or holidays, working day and its
His activity time sequence data constitutes data set；
Hysteresis effect is increased to the aeronautical data and meteorological data, generates the LSTM model.
3. method according to claim 2, which is characterized in that hysteresis effect is increased to the aeronautical data, meteorological data, it is raw
At the LSTM model, including：
LSTM model is used for time series forecasting, obtains output vector htAfterwards, a full articulamentum is connected, to finally obtain
Predicted value：
Wherein, σ indicates activation primitive；WoutTo connect layer matrix entirely, dimension is 1 × dh；htIndicate output vector；boutIt indicates to cut
Away from item；For the predicted value of the s articles time series t+1 phase.
4. method according to claim 2, which is characterized in that the loss function of the LSTM model, including：MAE and MAPE
Loss function；
Forward prediction n step loss function formula be：
In formula (8), MAEsIt is the index name for measuring prediction effect；S indicates road；T indicates the time；L indicates time series；n
Indicate the step-length of forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction；
In formula (9), MAPEsIt is the index name that another measures prediction effect；S indicates road；T indicates the time；When l is indicated
Between sequence；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction.
5. method according to any of claims 1-4, which is characterized in that obtain road in the preset range of airport periphery in real time
Traffic related information, including：
It obtains the vehicle flowrate and travel speed of road in the preset range of airport periphery in real time by data-interface, calculates congestion in road
Index.
6. a kind of aerodrome traffic congestion prediction meanss based on LSTM model, which is characterized in that including：
Module is obtained, for obtaining the traffic related information of road and air station flight landing letter in the preset range of airport periphery in real time
Aeronautical meteorology information in breath and airport periphery preset range；
Input module, for inputting LSTM model for the traffic related information, flight landing information and aeronautical meteorology information；
Prediction module, for obtaining the output of the LSTM model as a result, the output result is prediction future time period airport week
The congestion index of road in the preset range of side.
7. device as claimed in claim 6, which is characterized in that the LSTM model includes：
Acquisition submodule, for obtaining the traffic related information of road and air station flight landing in the preset range of airport periphery in real time
Aeronautical meteorology information in information and airport periphery preset range；
Input submodule, for being carried out according to the time by the traffic related information, flight landing information and aeronautical meteorology information
Association；
Constitute submodule, for obtained after being associated with according to input submodule per hour in congestion in road index, aeronautical data and
Meteorological data merges festivals or holidays, working day and other activity time sequence datas, constitutes data set；
Submodule is generated, for increasing hysteresis effect to the aeronautical data and meteorological data, generates the LSTM model.
8. device as claimed in claim 7, which is characterized in that the generation submodule, specifically for LSTM model is used for
Time series forecasting obtains output vector htAfterwards, a full articulamentum is connected, to finally obtain predicted value：
Wherein, σ indicates activation primitive；WoutTo connect layer matrix entirely, dimension is 1 × dh；htIndicate output vector；boutIt indicates to cut
Away from item；For the predicted value of the s articles time series t+1 phase.
9. device as claimed in claim 7, which is characterized in that the loss function of the LSTM model, including：MAE and MAPE
Loss function；
Forward prediction n step loss function formula be：
In formula (8), MAEsIt is the index name for measuring prediction effect；S indicates road；T indicates the time；L indicates time series；n
Indicate the step-length of forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction；
In formula (9), MAPEsIt is the index name that another measures prediction effect；S indicates road；T indicates the time；When l is indicated
Between sequence；The step-length of n expression forward prediction；Indicate practical congestion index；Indicate the congestion index of prediction.
10. device as claim in any one of claims 6-9, which is characterized in that obtain airport in real time in the acquisition submodule
The traffic related information of road in the preset range of periphery, including：
It obtains the vehicle flowrate and travel speed of road in the preset range of airport periphery in real time by data-interface, calculates congestion in road
Index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810878081.3A CN108898838B (en) | 2018-08-03 | 2018-08-03 | Method and device for predicting airport traffic jam based on LSTM model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810878081.3A CN108898838B (en) | 2018-08-03 | 2018-08-03 | Method and device for predicting airport traffic jam based on LSTM model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108898838A true CN108898838A (en) | 2018-11-27 |
| CN108898838B CN108898838B (en) | 2020-08-28 |
Family
ID=64353052
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810878081.3A Expired - Fee Related CN108898838B (en) | 2018-08-03 | 2018-08-03 | Method and device for predicting airport traffic jam based on LSTM model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108898838B (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190948A (en) * | 2018-08-20 | 2019-01-11 | 北京航空航天大学 | A kind of association analysis method of large aerospace hinge operation and urban traffic blocking |
| CN109559512A (en) * | 2018-12-05 | 2019-04-02 | 北京掌行通信息技术有限公司 | A kind of regional traffic flow prediction technique and device |
| CN109635246A (en) * | 2018-12-06 | 2019-04-16 | 西南交通大学 | A kind of multiattribute data modeling method based on deep learning |
| CN109887272A (en) * | 2018-12-26 | 2019-06-14 | 阿里巴巴集团控股有限公司 | A kind of prediction technique and device of traffic flow of the people |
| CN110223517A (en) * | 2019-06-20 | 2019-09-10 | 青岛科技大学 | Short-term traffic flow forecast method based on temporal correlation |
| CN110428613A (en) * | 2019-07-09 | 2019-11-08 | 中山大学 | A kind of intelligent transportation trend prediction method of machine learning |
| CN110991698A (en) * | 2019-11-07 | 2020-04-10 | 南通大学 | Seasonal traffic flow grey prediction method based on mixed processing |
| CN110991913A (en) * | 2019-12-09 | 2020-04-10 | 南京航空航天大学 | Busy airport peak time congestion risk analysis method |
| CN111080110A (en) * | 2019-12-09 | 2020-04-28 | 南京航空航天大学 | Airport congestion risk analysis system |
| CN111259537A (en) * | 2020-01-14 | 2020-06-09 | 中交第二公路勘察设计研究院有限公司 | Road surface performance prediction method based on VAR (variable-offset-ratio) multivariate time sequence |
| CN111340068A (en) * | 2020-02-10 | 2020-06-26 | 北京中旖新能源有限公司 | Road traffic state prediction method and device |
| CN112712695A (en) * | 2020-12-30 | 2021-04-27 | 桂林电子科技大学 | Traffic flow prediction method, device and storage medium |
| CN112836905A (en) * | 2021-04-22 | 2021-05-25 | 中航信移动科技有限公司 | Flight event occurrence time prediction method, device, equipment and storage medium |
| CN112967013A (en) * | 2021-02-20 | 2021-06-15 | 海南太美航空股份有限公司 | Method and system for determining takeoff time of pre-opened flight and electronic equipment |
| CN114063150A (en) * | 2021-11-16 | 2022-02-18 | 西南科技大学 | Seismic source-station velocity model selection method based on ML-KNN algorithm |
| CN114997506A (en) * | 2022-06-17 | 2022-09-02 | 西北师范大学 | Atmospheric pollution propagation path prediction method based on link prediction |
| CN115346401A (en) * | 2022-08-01 | 2022-11-15 | 南京航空航天大学 | Low-altitude unmanned aerial vehicle monitoring and track prediction method |
| CN115620526B (en) * | 2022-12-21 | 2023-03-07 | 中国民航大学 | Airport land side traffic management system based on big data analysis and optimization method |
| CN116631196A (en) * | 2023-07-25 | 2023-08-22 | 南京农业大学 | Traffic road condition prediction method and device based on big data |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE3855163D1 (en) * | 1988-05-11 | 1996-05-02 | Alex Frauchiger | METHOD FOR RESOLVING OR PREVENTING UNWANTED TRAFFIC JAMS |
| JP2007140745A (en) * | 2005-11-16 | 2007-06-07 | Sumitomo Electric Ind Ltd | Traffic congestion prediction system, traffic congestion factor estimation system, traffic congestion prediction method and traffic congestion factor estimation method |
| CN101154317A (en) * | 2006-09-27 | 2008-04-02 | 株式会社查纳位资讯情报 | Traffic state predicting apparatus |
| CN101438334A (en) * | 2006-03-03 | 2009-05-20 | 因瑞克斯有限公司 | Dynamic time series prediction of future traffic conditions |
| CN102346964A (en) * | 2010-08-05 | 2012-02-08 | 王学鹰 | Real-time jam prediction and intelligent management system for road traffic network area |
| JP5214759B2 (en) * | 2011-03-16 | 2013-06-19 | 三菱電機インフォメーションシステムズ株式会社 | Congestion degree estimation device and congestion degree estimation program |
| CN104157139A (en) * | 2014-08-05 | 2014-11-19 | 中山大学 | Prediction method and visualization method of traffic jam |
| CN204791180U (en) * | 2015-07-17 | 2015-11-18 | 四川精工伟达智能技术股份有限公司 | Wisdom tourism management system |
| CN105389980A (en) * | 2015-11-09 | 2016-03-09 | 上海交通大学 | Short-time traffic flow prediction method based on long-time and short-time memory recurrent neural network |
| CN106355879A (en) * | 2016-09-30 | 2017-01-25 | 西安翔迅科技有限责任公司 | Time-space correlation-based urban traffic flow prediction method |
| CN107230351A (en) * | 2017-07-18 | 2017-10-03 | 福州大学 | A kind of Short-time Traffic Flow Forecasting Methods based on deep learning |
| CN108229724A (en) * | 2017-12-06 | 2018-06-29 | 华南理工大学 | A kind of transport data stream Forecasting Methodology in short-term based on Spatial-temporal Information Fusion |
-
2018
- 2018-08-03 CN CN201810878081.3A patent/CN108898838B/en not_active Expired - Fee Related
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE3855163D1 (en) * | 1988-05-11 | 1996-05-02 | Alex Frauchiger | METHOD FOR RESOLVING OR PREVENTING UNWANTED TRAFFIC JAMS |
| JP2007140745A (en) * | 2005-11-16 | 2007-06-07 | Sumitomo Electric Ind Ltd | Traffic congestion prediction system, traffic congestion factor estimation system, traffic congestion prediction method and traffic congestion factor estimation method |
| CN101438334A (en) * | 2006-03-03 | 2009-05-20 | 因瑞克斯有限公司 | Dynamic time series prediction of future traffic conditions |
| CN101154317A (en) * | 2006-09-27 | 2008-04-02 | 株式会社查纳位资讯情报 | Traffic state predicting apparatus |
| CN102346964A (en) * | 2010-08-05 | 2012-02-08 | 王学鹰 | Real-time jam prediction and intelligent management system for road traffic network area |
| JP5214759B2 (en) * | 2011-03-16 | 2013-06-19 | 三菱電機インフォメーションシステムズ株式会社 | Congestion degree estimation device and congestion degree estimation program |
| CN104157139A (en) * | 2014-08-05 | 2014-11-19 | 中山大学 | Prediction method and visualization method of traffic jam |
| CN204791180U (en) * | 2015-07-17 | 2015-11-18 | 四川精工伟达智能技术股份有限公司 | Wisdom tourism management system |
| CN105389980A (en) * | 2015-11-09 | 2016-03-09 | 上海交通大学 | Short-time traffic flow prediction method based on long-time and short-time memory recurrent neural network |
| CN106355879A (en) * | 2016-09-30 | 2017-01-25 | 西安翔迅科技有限责任公司 | Time-space correlation-based urban traffic flow prediction method |
| CN107230351A (en) * | 2017-07-18 | 2017-10-03 | 福州大学 | A kind of Short-time Traffic Flow Forecasting Methods based on deep learning |
| CN108229724A (en) * | 2017-12-06 | 2018-06-29 | 华南理工大学 | A kind of transport data stream Forecasting Methodology in short-term based on Spatial-temporal Information Fusion |
Non-Patent Citations (9)
| Title |
|---|
| ARUNAS MARMA ET AL.: "Parking Traffic Jam Forecast System", 《IEEE》 * |
| C. MAHABIR ET AL.: "Transferability of a neuro-fuzzy river ice jam flood forecasting model", 《EL SEVIER》 * |
| 崔玮: "高速路网交通状态判别与预测的研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
| 张海鹏 等: "基于公交车ＧＰＳ数据的短时交通流预测研究", 《内蒙古工业大学学报》 * |
| 杨磊 等: "大型繁忙机场场面离场交通流拥堵特征分析", 《航空学报》 * |
| 查尔斯·W·奥斯特罗姆 等: "《时间序列分析回归技术 第2版》", 30 July 2017 * |
| 王俊友: "《智能交通***及应用》", 31 August 2017 * |
| 王斌会: "《计量经济学模型及R语言应用》", 31 May 2015 * |
| 高晓波: "基于深度学习的短时交通拥堵预测模型", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190948A (en) * | 2018-08-20 | 2019-01-11 | 北京航空航天大学 | A kind of association analysis method of large aerospace hinge operation and urban traffic blocking |
| CN109559512A (en) * | 2018-12-05 | 2019-04-02 | 北京掌行通信息技术有限公司 | A kind of regional traffic flow prediction technique and device |
| CN109635246A (en) * | 2018-12-06 | 2019-04-16 | 西南交通大学 | A kind of multiattribute data modeling method based on deep learning |
| CN109887272B (en) * | 2018-12-26 | 2021-08-13 | 创新先进技术有限公司 | Traffic pedestrian flow prediction method and device |
| CN109887272A (en) * | 2018-12-26 | 2019-06-14 | 阿里巴巴集团控股有限公司 | A kind of prediction technique and device of traffic flow of the people |
| CN110223517A (en) * | 2019-06-20 | 2019-09-10 | 青岛科技大学 | Short-term traffic flow forecast method based on temporal correlation |
| CN110428613A (en) * | 2019-07-09 | 2019-11-08 | 中山大学 | A kind of intelligent transportation trend prediction method of machine learning |
| CN110991698A (en) * | 2019-11-07 | 2020-04-10 | 南通大学 | Seasonal traffic flow grey prediction method based on mixed processing |
| CN110991698B (en) * | 2019-11-07 | 2023-09-22 | 南通大学 | Seasonal traffic flow grey prediction method based on mixed processing |
| CN110991913A (en) * | 2019-12-09 | 2020-04-10 | 南京航空航天大学 | Busy airport peak time congestion risk analysis method |
| CN111080110A (en) * | 2019-12-09 | 2020-04-28 | 南京航空航天大学 | Airport congestion risk analysis system |
| CN110991913B (en) * | 2019-12-09 | 2023-04-07 | 南京航空航天大学 | Busy airport peak time congestion risk analysis method |
| CN111259537B (en) * | 2020-01-14 | 2023-03-31 | 中交第二公路勘察设计研究院有限公司 | Road surface performance prediction method based on VAR (variable-value response) multivariate time sequence |
| CN111259537A (en) * | 2020-01-14 | 2020-06-09 | 中交第二公路勘察设计研究院有限公司 | Road surface performance prediction method based on VAR (variable-offset-ratio) multivariate time sequence |
| CN111340068A (en) * | 2020-02-10 | 2020-06-26 | 北京中旖新能源有限公司 | Road traffic state prediction method and device |
| CN112712695A (en) * | 2020-12-30 | 2021-04-27 | 桂林电子科技大学 | Traffic flow prediction method, device and storage medium |
| CN112712695B (en) * | 2020-12-30 | 2021-11-26 | 桂林电子科技大学 | Traffic flow prediction method, device and storage medium |
| CN112967013A (en) * | 2021-02-20 | 2021-06-15 | 海南太美航空股份有限公司 | Method and system for determining takeoff time of pre-opened flight and electronic equipment |
| CN112967013B (en) * | 2021-02-20 | 2024-03-19 | 海南太美航空股份有限公司 | Method, system and electronic device for determining take-off time of pre-opened flight |
| CN112836905B (en) * | 2021-04-22 | 2021-08-03 | 中航信移动科技有限公司 | Flight event occurrence time prediction method, device, equipment and storage medium |
| CN112836905A (en) * | 2021-04-22 | 2021-05-25 | 中航信移动科技有限公司 | Flight event occurrence time prediction method, device, equipment and storage medium |
| CN114063150A (en) * | 2021-11-16 | 2022-02-18 | 西南科技大学 | Seismic source-station velocity model selection method based on ML-KNN algorithm |
| CN114063150B (en) * | 2021-11-16 | 2023-05-23 | 西南科技大学 | ML-KNN algorithm-based 'seismic source-station' speed model selection method |
| CN114997506A (en) * | 2022-06-17 | 2022-09-02 | 西北师范大学 | Atmospheric pollution propagation path prediction method based on link prediction |
| CN114997506B (en) * | 2022-06-17 | 2024-05-14 | 西北师范大学 | Atmospheric pollution propagation path prediction method based on link prediction |
| CN115346401B (en) * | 2022-08-01 | 2023-09-12 | 南京航空航天大学 | Low-altitude unmanned aerial vehicle monitoring and track prediction method |
| CN115346401A (en) * | 2022-08-01 | 2022-11-15 | 南京航空航天大学 | Low-altitude unmanned aerial vehicle monitoring and track prediction method |
| CN115620526B (en) * | 2022-12-21 | 2023-03-07 | 中国民航大学 | Airport land side traffic management system based on big data analysis and optimization method |
| CN116631196A (en) * | 2023-07-25 | 2023-08-22 | 南京农业大学 | Traffic road condition prediction method and device based on big data |
| CN116631196B (en) * | 2023-07-25 | 2023-11-28 | 南京农业大学 | Traffic road condition prediction method and device based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108898838B (en) | 2020-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108898838A (en) | A kind of aerodrome traffic congestion prediction technique and device based on LSTM model | |
| Chen et al. | Design of limited-stop bus service with capacity constraint and stochastic travel time | |
| CN105070042B (en) | A kind of modeling method of traffic forecast | |
| WO2018214675A1 (en) | Quantified analysis method of influence on road travel time from urban built-up environment | |
| CN109493599A (en) | A kind of Short-time Traffic Flow Forecasting Methods based on production confrontation network | |
| CN105679102B (en) | A kind of national flight flow spatial and temporal distributions prediction deduction system and method | |
| CN111275965B (en) | Real-time traffic simulation analysis system and method based on internet big data | |
| CN110517492A (en) | Based on the traffic route recommended method of parallel integrated study, system, device | |
| CN109558985A (en) | A kind of bus passenger flow amount prediction technique based on BP neural network | |
| CN107919014B (en) | Taxi running route optimization method for multiple passenger mileage | |
| CN109410577A (en) | Adaptive traffic control sub-area division method based on Spatial Data Mining | |
| CN113065074B (en) | Track destination prediction method based on knowledge graph and self-attention mechanism | |
| CN110400015A (en) | A kind of Time Estimation Method and its device, equipment | |
| CN109542876A (en) | Extracting method based on Hadoop data mining aircraft experience locus model key factor | |
| CN111160622A (en) | Scenic spot passenger flow prediction method and device based on hybrid neural network model | |
| Chen et al. | Planning aircraft taxiing trajectories via a multi-ojective immune optimisation | |
| Muresan et al. | Adaptive traffic signal control with deep reinforcement learning an exploratory investigation | |
| CN107220724A (en) | Passenger flow forecast method and device | |
| Seliverstov et al. | Modeling of megalopolis traffic flows with the introduction of a new line of water intercity passenger transport | |
| Han et al. | Development and evaluation of adaptive transit signal priority control with updated transit delay model | |
| Lv et al. | Virtual geographic environment based coach passenger flow forecasting | |
| CN110363358A (en) | Public transportation mode share prediction technique based on multi-agent simulation | |
| Taparia et al. | Bus journey and arrival time prediction based on archived avl/gps data using machine learning | |
| Sheng et al. | Deep spatial-temporal travel time prediction model based on trajectory feature | |
| Wu et al. | A flood-discharge-based spatio-temporal diffusion method for multi-target traffic hotness construction from trajectory data |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200828 Termination date: 20210803 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |