CN108335487A - Road traffic state prediction system based on traffic state time sequence - Google Patents
Road traffic state prediction system based on traffic state time sequence Download PDFInfo
- Publication number
- CN108335487A CN108335487A CN201810232199.9A CN201810232199A CN108335487A CN 108335487 A CN108335487 A CN 108335487A CN 201810232199 A CN201810232199 A CN 201810232199A CN 108335487 A CN108335487 A CN 108335487A
- Authority
- CN
- China
- Prior art keywords
- data
- matrix
- prediction
- traffic
- state
- 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
- 238000004891 communication Methods 0.000 claims abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 12
- 230000002159 abnormal effect Effects 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 50
- 238000012360 testing method Methods 0.000 claims description 13
- 210000002569 neuron Anatomy 0.000 claims description 11
- 238000000034 method Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 8
- 230000007423 decrease Effects 0.000 claims description 5
- 238000013210 evaluation model Methods 0.000 claims description 4
- 230000001537 neural effect Effects 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 2
- 238000005259 measurement Methods 0.000 claims description 2
- 238000012986 modification Methods 0.000 claims description 2
- 230000004048 modification Effects 0.000 claims description 2
- 230000008054 signal transmission Effects 0.000 claims description 2
- 230000000007 visual effect Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000004140 cleaning Methods 0.000 abstract 1
- 238000005070 sampling Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 6
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 210000005036 nerve Anatomy 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000009514 concussion Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000003973 paint Substances 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
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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"
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Chemical & Material Sciences (AREA)
- Development Economics (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a road traffic state prediction system based on traffic state time sequence. The system comprises a data acquisition module, a data processing module, a prediction module and a network communication module. Firstly, collecting traffic flow and speed information of a certain road section traffic information detection device for a plurality of days; cleaning original data by judging data abnormal points and outliers, readjusting data sampling frequency and other data processing modes; inputting the processed data into a prediction module for iterative training of the model; and finally, forecasting data at a plurality of future moments is obtained through the forecasting of the traffic state information, and the forecasted traffic state data can be transmitted to a traffic management department through a network communication module. The invention comprehensively considers the front and back time sequence characteristics of the traffic state, accurately grasps the internal association between time sequences, achieves the aim of predicting the traffic state, finally transmits data to relevant traffic management departments, and provides a new management basis for the traffic management departments.
Description
Technical field
The invention belongs to technical field of intelligent traffic, and in particular to a kind of urban road friendship considering traffic behavior timing
Logical status predication system, good administration base is provided for traffic department.
Background technology
It includes timing controlled to be currently being widely used road control mainly, induction controls and self adaptive control, often
Through generate certain time congestion when, just start control optimization, that is to say, that these control methods can not to the following block status into
Row prediction, then after having obtained the status information in following several moment sections, can arrange emergency measure in time and execute in advance
Control method.Traffic status prediction is that vehicle supervision department takes traffic guidance measure Main Basiss, is that intelligent transportation system is ground
The key problem studied carefully.So traffic situation prediction apparatus is applied to section, have for alleviation urban traffic blocking very heavy
Want meaning.
Invention content
Pass through the digging to a large amount of traffic behavior historical datas for deficiency, the present invention existing for present road traffic control
Pick, in conjunction with prediction module, it can be achieved that prediction to road traffic state, while by network communication module, by the data of prediction
It is sent to vehicle supervision department.The technical solution specifically used is as follows：
System includes data acquisition module, by being interacted with intersection annunciator, reads semaphore communication protocol, obtains section
Traffic behavior historical data information；Data after acquisition are divided into working day and nonworkdays two by data processing module first
Class, then to data carry out riding Quality Analysis, if data time sequence is unstable, handled by missing values and abnormal point and from
Group's point differentiates to clean data；If data time sequence is steady, to data visual description and it is normalized；Prediction
The data that data processing module exports are divided into working day and nonworkdays data by module, working day and nonworkdays data
90% is used as training data, is stored in matrix TrainDi [] and TrainUDi [] respectively, by working day and nonworkdays data
Remaining 10% is used as test data, is stored in matrix TestDi [] and TestUDi [] respectively, and above four matrixes are divided
At several MINI-Batch matrixes；Training data is input in LSTM-EF prediction models and is trained, determines that model respectively saves
Point weight parameter W matrixes, then carry out model measurement with test data, finally obtain prediction result；Network communication module is used
Signal transmission between each module of system and communication interaction.
Preferably, the prediction technique of the LSTM-EF prediction models is：
(1) it establishes and forgets door output matrix FGt, the neuron state S of last moment is controlled by Sigmoid functionst-1It protects
It is left to the neuron state S at this momenttIn quantity：
FGt=Sigmoid (Wf*[yt-1,xt]+Bf)
Wherein, WfIt is to forget door weight matrix, BfIt is to forget door bias matrix, yt-1It is the output of previous moment, xtIt is to work as
The input at preceding moment；Sigmoid function value ranges are that [0,1] forgets all letters when Sigmoid (x) is infinitely close to 0
Breath；When Sigmoid (x) is infinitely close to 1, information is not forgotten；When 0<Sigmoid(x)<1, forget partial information；
(2) input gate matrix IG is establishedt, the state matrix of last moment by with forget door output matrix FGtIt is multiplied, really
The information for the last moment forgotten is needed calmly, while plus the information for needing to remember at this timeObtain the shape of Current neural member
State St：
IGt=Sigmoid (Wi*[yt-1,xt]+Bi)
Wherein, WiIt is input gate weight matrix, BiIt is input gate bias matrix, yt-1It is the output of previous moment, xtIt is to work as
The input at preceding moment,It is the state for indicating currently to input, Ws、BsThe weight matrix and biasing square of current state are indicated respectively
Battle array；
(3) output gate matrix OG is establishedt, obtain output valve yt：
OGt=Sigmoid (Wo*[yt-1,xt]+Bo)
yt=OGt*Tanh(St)
WoIt is input gate weight matrix, BoIt is input gate bias matrix, yt-1It is the output of previous moment, xtIt is current time
Input, Tanh () value range of function is [- 1,1]；
(4) as output valve ytWhen appearance is infinitely close to 0, forgetting compensation door was introduced, forgetting state was avoided the occurrence of：
ξ=0.01
(5) by way of gradient decline, back-propagation Grad, after iteration several times, convergence obtains Wf、Bf、
Wi、Bi、Ws、Bs、Wo、BoDeng opposite optimized parameter matrix.
Preferably, by Analysis of Mean Square Error error prediction model and evaluation model precision of prediction, while modification mould is returned
Type training parameter, iterations, again optimal prediction model
Wherein, N:Sampled point number in one test sample；yi:Sampled point actual numerical value in one test sample；
prediction:Model predication value.
The present invention has following beneficial technique effect：
(1) function is advanced, can accurately obtain road traffic state predicted value.
(2) time series sample data feature representation degree is can be improved into data cleansing using data processing means, utilized
LSTM-EF (Long Short-Term Memory-Excessive Forgotten) prediction technique can depth excavation traffic behavior
The feature essence of sequential, improves precision of prediction.
Description of the drawings
Fig. 1 is data processing module, prediction module calculation flow chart.
Fig. 2 is LSTM-EF (Long Short-Term Memory-Excessive Forgotten) model schematic.
Fig. 3 is LSTM-EF (Long Short-Term Memory-Excessive Forgotten) hidden layer neuron
Door figure.
Fig. 4 is present system structure chart.
Specific implementation mode
Fig. 1 is mainly the data processing module of entire traffic situation prediction apparatus and the calculation flow chart of prediction module, master
It is divided into the parts such as data acquisition, data processing module, prediction module, error analysis.Before data preparation link is model training
Vital link is related to speed and the accuracy of model training, and most important data preparation link is exactly traffic behavior
The determination of stability of time series data is handled by missing values, outlier processing and auto-correlation coefficient are analyzed, raising time series data
Stability.The prediction module of device is mainly made of LSTM-EF models, since LSTM-EF is the one of neural network algorithm
Kind, there are three types of the modes of data input model：Batch input (Batch Input), random odd number value input (Stochastic
Input), small lot input (MINI-Batch Input)；The selection of input mode determines that BP gradients decline Optimal Parameters
Speed and globally optimal solution as a result, batch input (Batch Input) although Local Phase can be can be obtained to optimal solution,
The convergence rate for being gradient decline can be very slow；Random odd number value (Stochastic Input) although input fast convergence rate,
It is to be easy generating concussion near optimal solution, and influence the accuracy of final result；And small lot inputs (MINI-Batch
Input it) both can guarantee convergent speed, while opposite optimal solution can also be found, so selection small lot inputs (MINI-
Batch Input) mode input LSTM-EF (Long Short-Term Memory-Excessive Forgotten) mould
Type.LSTM-EF (Long Short-Term Memory-Excessive Forgotten) model part can be in the part Fig. 2, Fig. 3
It illustrates.MSE error evaluations are the effective means of evaluation model error rate, and the property of model is calculated by MSE assessments
Energy index, while adjustment parameter again, continue Optimized model.It calculates, export final predicted value and paints eventually by renormalization
Prediction data processed.
(1) data after acquisition are divided into two class of working day and nonworkdays, since traffic state data is time series
Data, so first having to carry out the riding Quality Analysis of data；Then it is handled by missing values, abnormal point and outlier differentiation etc.
Numerical analysis mode cleans data.
(2) it using 90% data of treated in step (1) working day and nonworkdays as training data, is stored in respectively
In matrix TrainDi [] and TrainUDi [], using working day and nonworkdays remaining 10% data as test data, divide
It Cun Ru not be in matrix TestDi [] and TestUDi [].And above four matrixes are divided into several MINI-Batch matrixes, so as to
Data input model.
(3) training data is input to LSTM-EF (Long Short-Term Memory-Excessive
Forgotten) prediction model is trained, and determines each node weights parameter W matrixes of model.Then mould is carried out with test data
Type is tested.
(4) by MSE (mean square error) analysis models error, evaluation model precision of prediction, while return to step (3) is repaiied
Change model training parameter, iterations etc., optimizes LSTM-EF (Long Short-Term Memory-Excessive again
Forgotten) prediction model.
N:Sampled point number in one test sample；yi:Sampled point actual numerical value in one test sample；
prediction:Model predication value；
Fig. 2, Fig. 3 are the macro of LSTM-EF (Long Short-Term Memory-Excessive Forgotten) respectively
See the logical construction of structure and hidden layer intrinsic nerve member, LSTM-EF (Long Short-Term Memory-
Excessive Forgotten) be by a large amount of time series sample data training pattern so that model can find and remember sample
Context between this sequence of values can predict following timing variations in input test data.LSTM-EF(Long
Short-Term Memory-Excessive Forgotten) it is to be made of three-layer network：Input layer, hidden layer, output layer structure
At.The traffic behavior in this section is assessed by section history vehicle speed data v,For the average speed value of the t moment in section 1, in advance
The surveying traffic behavior of the task is exactly to utilize certain a road section history average speed sequenceIt goes to predict lower a period of time
The average speed numerical value of t+1 is carved, as shown in Figure 2.Since LSTM-EF models have the characteristics of logic gate structure, solves tradition
Gradient explosion that RNN is brought when BP gradients decline Optimal Parameters since time series is long and the problems such as gradient dispersion.Such as figure
Shown in 3, each neuron has the matrix St for indicating current time state, forgets about unwanted lengthy and jumbled sequential by forgeing door
Information；The information at current time is inputted by input gate, and updates state matrix St；Since output quantity will pass through Sigmoid
() function, so will appear forgetting phenomenon, that is, output result yt is approximately 0, so forgeing compensation by being added
The logic judgment of door and out gate decides whether to compensate the output yt-1 of last moment to current time, to solve
Forgetting problem.Finally the speed data Vt+1 predicted is sent to traffic control department by network communication module.
Fig. 4 is prediction meanss module structure schematic diagram, and the module of prediction meanss includes mainly：Semaphore communication module, number
According to processing and prediction module, network communication module.By with intersection annunciator communication module, read semaphore communication protocol, into
And 200 days traffic behavior big data information of history in section traffic information detection device is obtained, data information is car speed v,
Speed is more than 30km/h, the coast is clear；Speed is between 20km/h-30km/h, low running speed；Speed is in 10km/h-20km/
H, blocked state；Speed is in 10km/h hereinafter, heavy congestion.Network communication module supports the wireless communication transmissions skills such as GPRS, 4G
Art also supports the wire communications transmission technologys such as optical fiber.In turn, obtained prediction data can be transmitted by network communication module
To traffic control department.
It is former based on the method for LSTM-EF (Long Short-Term Memory-Excessive Forgotten) predictions
Reason illustrates and steps are as follows：
LSTM-EF (Long Short-Term Memory-Excessive Forgotten) prediction technique is cycle nerve
The deformation algorithm of network (RNN), is mainly made of input layer, output layer, hidden layer, the serial memorization machine of hidden layer neuron
System is controlled by four doors, has stronger memory capability to the longer data of sequence.
(* hereinafter represents the multiplication of matrix corresponding element)
(1) prediction technique is forward by LSTM-EF (Long Short-Term Memory-Excessive Forgotten)
It calculates
(1) forget door (Forget Gate)
FGt=Sigmoid (Wf*[yt-1,xt]+Bf)
WfIt is to forget door weight matrix, BfIt is to forget door bias matrix, yt-1It is the output of previous moment, xtIt is current time
Input.The effect for forgeing door is that the neuron state S of last moment is controlled by Sigmoid functionst-1How many can retain
To the neuron state S at this momenttIn.
Sigmoid function value ranges are [0,1], when Sigmoid (x) is infinitely close to 0, indicate to forget all letters
Breath；When Sigmoid (x) is infinitely close to 1, information is not forgotten in expression；When 0<Sigmoid(x)<1, portion can be forgotten by forgeing door
Divide information.
(2) input gate (Input Gate)
IGt=Sigmoid (Wi*[yt-1,xt]+Bi)
WiIt is input gate weight matrix, BiIt is input gate bias matrix, yt-1It is the output of previous moment, xtIt is current time
Input.It is the state for indicating currently to input, Ws、BsThe weight matrix and bias matrix of current state are indicated respectively.St-1It is
The state matrix of last moment, by with forget door output matrix FGtIt is multiplied, determines the information for the last moment for needing to forget,
Simultaneously plus the information for needing to remember at this timeThe state S of Current neural member is just obtainedt。
(3) out gate (Output Gate)
OGt=Sigmoid (Wo*[yt-1,xt]+Bo)
yt=OGt*Tanh(St)
WoIt is input gate weight matrix, BoIt is input gate bias matrix, yt-1It is the output of previous moment, xtIt is current time
Input.OGtDetermine which part of neuron state matrix is exported, Tanh (St) by StValue be compressed between [- 1,1],
Finally obtain output valve yt。
(4) it crosses and forgets compensation door (EF Gate)
Due to output matrix ytBy Sigmoid () function influences, it is possible to output result occur to be infinitely close to
0, when there are this situation, the y of next neuron can be causedt-1Input component is close to 0, to forgetting phenomenon occurred,
So needing to introduce forgetting compensation door.
ξ=0.01
(2)：LSTM-EF (Long Short-Term Memory-Excessive Forgotten) prediction technique passes through
The mode that gradient declines, back-propagation Grad, after iteration several times, convergence obtains Wf、Bf、Wi、Bi、Ws、Bs、Wo、 Bo
Deng opposite optimized parameter matrix.
Claims (3)
1. a kind of road traffic state forecasting system based on traffic behavior timing, which is characterized in that the system includes
Data acquisition module reads semaphore communication protocol, obtains road section traffic volume state history by being interacted with intersection annunciator
Data information；
Data after acquisition are divided into two class of working day and nonworkdays by data processing module first, then are carried out to data steady
Property analysis handled by missing values if data time sequence is unstable and abnormal point and outlier differentiate to clean data；
If data time sequence is steady, to data visual description and it is normalized；
The data that data processing module exports are divided into working day and nonworkdays data, working day and inoperative by prediction module
The 90% of day data is used as training data, is stored in matrix TrainDi [] and TrainUDi [] respectively, by working day and non-work
Make day data remaining 10% and be used as test data, is stored in matrix TestDi [] and TestUDi [] respectively, and by above four
A matrix is divided into several MINI-Batch matrixes；Training data is input in LSTM-EF prediction models and is trained, determines mould
Each node weights parameter W matrixes of type, then carry out model measurement with test data, finally obtain prediction result；
Network communication module, for the signal transmission and communication interaction between each module of system.
2. the road traffic state forecasting system based on traffic behavior timing as claimed in claim 2, which is characterized in that
The prediction technique of the LSTM-EF prediction models is：
(1) it establishes and forgets door output matrix FGt, the neuron state S of last moment is controlled by Sigmoid functionst-1It remains into
The neuron state S at this momenttIn quantity：
FGt=Sigmoid (Wf*[yt-1,xt]+Bf)
Wherein, WfIt is to forget door weight matrix, BfIt is to forget door bias matrix, yt-1It is the output of previous moment, xtWhen being current
The input at quarter；Sigmoid function value ranges are that [0,1] forgets all information when Sigmoid (x) is infinitely close to 0；When
When Sigmoid (x) is infinitely close to 1, information is not forgotten；When 0<Sigmoid(x)<1, forget partial information；
(2) input gate matrix IG is establishedt, the state matrix of last moment by with forget door output matrix FGtIt is multiplied, determining needs
The information of the last moment to be forgotten, while plus the information for needing to remember at this timeObtain the state S of Current neural membert：
IGt=Sigmoid (Wi*[yt-1,xt]+Bi)
Wherein, WiIt is input gate weight matrix, BiIt is input gate bias matrix, yt-1It is the output of previous moment, xtWhen being current
The input at quarter,It is the state for indicating currently to input, Ws、BsThe weight matrix and bias matrix of current state are indicated respectively；
(3) output gate matrix OG is establishedt, obtain output valve yt：
OGt=Sigmoid (Wo*[yt-1,xt]+Bo)
yt=OGt*Tanh(St)
WoIt is input gate weight matrix, BoIt is input gate bias matrix, yt-1It is the output of previous moment, xtIt is the defeated of current time
Enter, Tanh () value range of function is [- 1,1]；
(4) as output valve ytWhen appearance is infinitely close to 0, forgetting compensation door was introduced, forgetting state was avoided the occurrence of：
ξ=0.01
(5) by way of gradient decline, back-propagation Grad, after iteration several times, convergence obtains Wf、Bf、Wi、Bi、
Ws、Bs、Wo、BoDeng opposite optimized parameter matrix.
3. the road traffic state forecasting system based on traffic behavior timing as claimed in claim 2, which is characterized in that
By Analysis of Mean Square Error error prediction model and evaluation model precision of prediction, at the same return modification model training parameter,
Iterations, again optimal prediction model
Wherein, N:Sampled point number in one test sample；yi:Sampled point actual numerical value in one test sample；
prediction:Model predication value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810232199.9A CN108335487B (en) | 2018-03-20 | 2018-03-20 | Road traffic state prediction system based on traffic state time sequence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810232199.9A CN108335487B (en) | 2018-03-20 | 2018-03-20 | Road traffic state prediction system based on traffic state time sequence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108335487A true CN108335487A (en) | 2018-07-27 |
| CN108335487B CN108335487B (en) | 2020-04-10 |
Family
ID=62932184
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810232199.9A Expired - Fee Related CN108335487B (en) | 2018-03-20 | 2018-03-20 | Road traffic state prediction system based on traffic state time sequence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108335487B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109684310A (en) * | 2018-11-22 | 2019-04-26 | 安徽继远软件有限公司 | A kind of information system performance Situation Awareness method based on big data analysis |
| CN109800908A (en) * | 2018-12-27 | 2019-05-24 | 北京交通大学 | Signalized intersections operating status prediction technique and system based on LSTM model |
| CN109886387A (en) * | 2019-01-07 | 2019-06-14 | 北京大学 | It is a kind of that the traffic time sequence forecasting method returned is promoted based on gating network and gradient |
| CN112784906A (en) * | 2021-01-26 | 2021-05-11 | 中国科学院半导体研究所 | Agricultural machinery monitoring data cleaning method and device based on multi-condition time sequence |
| CN113436433A (en) * | 2021-06-24 | 2021-09-24 | 福建师范大学 | Efficient urban traffic outlier detection method |
| CN114387782A (en) * | 2022-01-12 | 2022-04-22 | 智道网联科技(北京)有限公司 | Method and device for predicting traffic state and electronic equipment |
| CN114418071A (en) * | 2022-01-24 | 2022-04-29 | 中国光大银行股份有限公司 | Cyclic neural network training method |
| CN117854279A (en) * | 2024-01-09 | 2024-04-09 | 南京清正源信息技术有限公司 | Road condition prediction method and system based on edge calculation |
| CN118097968A (en) * | 2024-04-22 | 2024-05-28 | 哈尔滨学院 | Road traffic safety assessment method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106157616A (en) * | 2016-06-29 | 2016-11-23 | 肖锐 | A kind of magnitude of traffic flow short-term prediction device |
| KR20170048888A (en) * | 2015-10-27 | 2017-05-10 | 현대자동차주식회사 | A system for predicting a volume of traffic, a displaying apparatus for a vehicle, a vehicle and a method for predicting a volume of traffic |
| CN106781489A (en) * | 2016-12-29 | 2017-05-31 | 北京航空航天大学 | A kind of road network trend prediction method based on recurrent neural network |
| WO2018035058A1 (en) * | 2016-08-15 | 2018-02-22 | University Of Southern California | Latent space model for road networks to predict time-varying traffic |
-
2018
- 2018-03-20 CN CN201810232199.9A patent/CN108335487B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170048888A (en) * | 2015-10-27 | 2017-05-10 | 현대자동차주식회사 | A system for predicting a volume of traffic, a displaying apparatus for a vehicle, a vehicle and a method for predicting a volume of traffic |
| CN106157616A (en) * | 2016-06-29 | 2016-11-23 | 肖锐 | A kind of magnitude of traffic flow short-term prediction device |
| WO2018035058A1 (en) * | 2016-08-15 | 2018-02-22 | University Of Southern California | Latent space model for road networks to predict time-varying traffic |
| CN106781489A (en) * | 2016-12-29 | 2017-05-31 | 北京航空航天大学 | A kind of road network trend prediction method based on recurrent neural network |
Non-Patent Citations (2)
| Title |
|---|
| RUI FU等: "Using LSTM and GRU neural network methods for traffic flow prediction", 《2016 31ST YOUTH ACADEMIC ANNUAL CONFERENCE OF CHINESE ASSOCIATION OF AUTOMATION》 * |
| 王祥雪等: "基于深度学习的短时交通流预测研究", 《交通运输系统工程与信息》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109684310A (en) * | 2018-11-22 | 2019-04-26 | 安徽继远软件有限公司 | A kind of information system performance Situation Awareness method based on big data analysis |
| CN109800908A (en) * | 2018-12-27 | 2019-05-24 | 北京交通大学 | Signalized intersections operating status prediction technique and system based on LSTM model |
| CN109886387A (en) * | 2019-01-07 | 2019-06-14 | 北京大学 | It is a kind of that the traffic time sequence forecasting method returned is promoted based on gating network and gradient |
| CN112784906A (en) * | 2021-01-26 | 2021-05-11 | 中国科学院半导体研究所 | Agricultural machinery monitoring data cleaning method and device based on multi-condition time sequence |
| CN112784906B (en) * | 2021-01-26 | 2024-02-02 | 中国科学院半导体研究所 | Agricultural machinery monitoring data cleaning method and device based on multi-condition time sequence |
| CN113436433A (en) * | 2021-06-24 | 2021-09-24 | 福建师范大学 | Efficient urban traffic outlier detection method |
| CN113436433B (en) * | 2021-06-24 | 2022-06-21 | 福建师范大学 | Efficient urban traffic outlier detection method |
| CN114387782A (en) * | 2022-01-12 | 2022-04-22 | 智道网联科技(北京)有限公司 | Method and device for predicting traffic state and electronic equipment |
| CN114418071A (en) * | 2022-01-24 | 2022-04-29 | 中国光大银行股份有限公司 | Cyclic neural network training method |
| CN117854279A (en) * | 2024-01-09 | 2024-04-09 | 南京清正源信息技术有限公司 | Road condition prediction method and system based on edge calculation |
| CN117854279B (en) * | 2024-01-09 | 2024-07-26 | 南京清正源信息技术有限公司 | Road condition prediction method and system based on edge calculation |
| CN118097968A (en) * | 2024-04-22 | 2024-05-28 | 哈尔滨学院 | Road traffic safety assessment method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108335487B (en) | 2020-04-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108335487A (en) | Road traffic state prediction system based on traffic state time sequence | |
| CN109285346B (en) | Urban road network traffic state prediction method based on key road sections | |
| CN110223517B (en) | Short-term traffic flow prediction method based on space-time correlation | |
| CN108846517B (en) | Integration method for predicating quantile probabilistic short-term power load | |
| CN107992536B (en) | Urban traffic missing data filling method based on tensor decomposition | |
| CN111292525B (en) | Traffic flow prediction method based on neural network | |
| CN106781489B (en) | A kind of road network trend prediction method based on recurrent neural network | |
| CN108764540B (en) | Water supply network pressure prediction method based on parallel LSTM series DNN | |
| CN108197739A (en) | A kind of urban track traffic ridership Forecasting Methodology | |
| CN111210621B (en) | Signal green wave coordination route optimization control method and system based on real-time road condition | |
| CN110188936B (en) | Short-term traffic flow prediction method based on multi-factor space selection deep learning algorithm | |
| CN112466117A (en) | Road network short-term traffic flow prediction method based on deep space-time residual error network | |
| CN109637128B (en) | Markov-based gray Verhulst short-time traffic flow prediction method and system | |
| CN108595803A (en) | Shale gas well liquid loading pressure prediction method based on recurrent neural network | |
| CN112633584B (en) | River sudden water pollution accident water quality prediction method based on improved LSTM-seq2seq model | |
| CN107862877A (en) | A kind of urban traffic signal fuzzy control method | |
| CN111967308A (en) | Online road surface unevenness identification method and system | |
| CN111754034A (en) | Time sequence prediction method based on chaos optimization neural network model | |
| CN105303835B (en) | A kind of Forecasting Approach for Short-term of road traffic stream mode | |
| CN111723516B (en) | Seawater intrusion simulation-optimization method based on adaptive deep neural network substitution model | |
| CN114692983B (en) | Automatic gear shifting prediction method and system for special vehicle | |
| CN113051811A (en) | Multi-mode short-term traffic jam prediction method based on GRU network | |
| CN108873706A (en) | Evaluation of trap intelligent expert recommended method based on deep neural network | |
| CN112215421A (en) | Deep learning water quality index prediction method based on generation countermeasure network | |
| CN115829150A (en) | Accumulated water prediction system |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200410 Termination date: 20210320 |