CN112530157A - Road traffic congestion propagation prediction method based on knowledge graph and Conv1D-LSTM-D - Google Patents
Road traffic congestion propagation prediction method based on knowledge graph and Conv1D-LSTM-D Download PDFInfo
- Publication number
- CN112530157A CN112530157A CN202011107018.3A CN202011107018A CN112530157A CN 112530157 A CN112530157 A CN 112530157A CN 202011107018 A CN202011107018 A CN 202011107018A CN 112530157 A CN112530157 A CN 112530157A
- Authority
- CN
- China
- Prior art keywords
- traffic
- lstm
- time
- lane
- 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
Images
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
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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
-
- 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
- G08G1/0133—Traffic data processing for classifying traffic situation
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
Abstract
A road traffic jam propagation prediction method based on a knowledge graph and Conv1D-LSTM-D comprises the following steps: 1) aiming at traffic flow data of each lane in a data set, fitting a flow-density scatter diagram of the lane by using a function based on the principle of a macroscopic basic diagram to obtain a critical traffic flow state threshold value of each lane; 2) defining a quadruple set of a time knowledge graph, and constructing a traffic jam propagation time knowledge graph according to the geographical position information, the traffic flow data set and the critical traffic flow state threshold of each lane; 3) and realizing road traffic jam propagation prediction based on the time knowledge graph and Conv 1D-LSTM-D. The method fully excavates and standardizes the complex relation between the road traffic flow data based on the knowledge map, realizes the propagation prediction of the road traffic jam based on Conv1D-LSTM-D, and improves the accuracy of the road traffic jam prediction.
Description
Technical Field
The invention belongs to the field of traffic prediction, and relates to a road traffic jam propagation prediction method based on a knowledge graph and Conv 1D-LSTM-D.
Background
In recent years, the increase in road traffic demand has caused serious congestion, delays and environmental problems, especially in large cities. Traffic congestion not only affects various traffic users in the society and reduces social and economic efficiency, but also has many other negative effects. With the continuous progress of science and technology, the tendency of treating traffic jam in the future is to delay the jam in the intelligence. The continuously improved traffic information acquisition quality and precision provide guarantee for accurately acquiring traffic data from a traffic network in real time and provide basic support for real-time traffic analysis and prediction research. Therefore, a scientific, effective and reasonable intelligent method is used for constructing a model for the traffic road network data, and the method has important research significance for relieving urban road congestion.
The traditional traffic jam prediction scheme is difficult to solve the problem that the traffic flow data result is huge and complex, so a new traffic jam solution is urgently needed. The rise of knowledge-graphs provides the possibility to identify, discover and infer relationships from a large amount of complex data. However, due to the particularity of the data in the traffic field, the application of the knowledge graph to traffic is rare at present.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a road traffic jam propagation prediction method based on a knowledge graph and Conv1D-LSTM-D, the method fully excavates and normatively expresses the complex relation between road traffic flow data based on the knowledge graph, and the propagation prediction of the road traffic jam is realized based on the Conv1D-LSTM-D, so that the accuracy of the road traffic jam prediction is improved.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a road traffic jam propagation prediction method based on a knowledge graph and Conv1D-LSTM-D comprises the following steps:
1) acquiring a lane critical traffic flow state threshold according to a road traffic flow data set: aiming at traffic flow data of each lane in a data set, fitting a flow-density scatter diagram of the lane by using a function based on the principle of a macroscopic basic diagram to obtain a critical traffic flow state threshold value of each lane;
2) constructing a traffic jam propagation time knowledge graph: defining a quadruple set of a time knowledge graph, and constructing a traffic jam propagation time knowledge graph according to the geographical position information, the traffic flow data set and the critical traffic flow state threshold of each lane;
3) and (3) realizing road traffic congestion propagation prediction based on the time knowledge graph and Conv 1D-LSTM-D: converting according to a time knowledge graph to obtain a traffic state vector of a time sequence, and extracting low-dimensional features of the input traffic state vector through a one-dimensional convolutional layer Conv1D and a global maximum pooling layer Globalmaxboosting; the time sequence characteristics are learned and the traffic state at the next moment is predicted by taking the time sequence characteristics as the input of an LSTM layer; and finally, the predicted low-dimensional feature vector is re-mapped back to a high-dimensional space through a decoding layer, so that the road traffic jam propagation prediction is realized.
Further, the process of the step 1) is as follows:
according to the flow (q) -density (o) model in the road macroscopic basic map, qmIs a critical flow threshold value, omFor the critical density threshold, a quadratic function relationship between flow and density is obtained:
wherein v isfFor free flow velocity, ojIs the blocking density;
respectively acquiring the same traffic flow data from the road traffic flow data set for each laneTraffic flow data and traffic flow density data of a lane at different moments, drawing a flow-density scatter diagram of the lane, and fitting by using a parabola to obtain a critical traffic flow state threshold (q) of each lanem,om)。
Further, in the step 2), the process of constructing the traffic jam propagation time knowledge graph is as follows:
2.1) defining a set of time-knowledge-graph quadruplets
Defining a set of temporal knowledge graph quadruplets, described as follows:
a quadruple I: intersection-adjacent-intersection, which describes whether two intersections are adjacent at time t;
quadruple II/quadruple III: the method comprises the following steps of (1) lane-congestion/non-congestion-intersection, wherein whether a certain lane at a certain intersection is congested at the moment t is described, and the congestion state is judged through a threshold value of a critical traffic flow state;
2.2) constructing a traffic jam propagation time knowledge graph
Judging the connection relation between intersections according to the geographical position information, and constructing a quadruple I at each moment t; comparing traffic flow data of each lane at different moments with a critical traffic flow state threshold value, judging congestion states of the lanes at each moment t, constructing a quadruple II and a quadruple III, and when q is equal to q>qmOr o>omWhen the road is in the congested state, the road is considered to be in the uncongested state, otherwise, the road is not congested; and integrating the quadruple according to the time t to realize the construction of the traffic jam propagation time knowledge graph.
Further, in the step 3), the road traffic congestion propagation prediction process based on the time knowledge graph and Conv1D-LSTM-D is implemented as follows:
3.1) extracting low-dimensional features of the time-series traffic state vector based on Conv1D and a GlobalmaxPoint layer;
acquiring time series traffic state vectors according to a traffic jam propagation time knowledge graph, wherein each vector represents the state of all lanes at one moment, 1 is used for indicating the smoothness of the lanes, 0 is used for indicating the traffic jam of the lanes, and then the time series traffic state vectors are input into a 1-layer Conv1D and a 1-layer GlobalMaxPoint to extract low-dimensional features;
3.2) learning temporal features of time series traffic state vectors based on the LSTM layer
The low-dimensional features extracted in the last step are used as input of an LSTM layer to learn timing sequence features and used for predicting the traffic state of the next moment, 4 feedforward neural networks are actually arranged in the LSTM to form a forgetting gate, an input gate and an output gate structure of an LSTM memory unit, the forgetting gate determines whether the memory unit forgets historical information or not, the input gate determines the influence of current input data on the state of the memory unit, the output gate determines the output information of the memory unit, and if x is assumed, the output gate determines the output information of the memory unittRepresenting the input vector at time t, ht-1And when the output at the time t-1 is represented, the LSTM memory module performs the processes of state updating and information output as follows:
first, the forgetting gate determines the degree f of keeping the past history informationt:
ft=σ(Wf·[ht-1,xt]+bf) (2)
Then, by input gate and ftStorage information C of common pair memory celltUpdating:
it=σ(Wi·[ht-1,xt]+bi) (3)
finally, the output h of the current time t is determined by an output gatet:
ot=σ(Wo[ht-1,xt]+bo) (6)
ht=ot*tanh(Ct) (7)
Wherein f ist,it,otRespectively represent the current t timeThe output states of the forgetting gate, the input gate and the output gate are carved, sigma represents an activation function, tanh () represents a hyperbolic activation function, Wf,Wi,WC,WoWeights representing internal feedforward neural networks, bf,bi,bC,boRepresenting a bias vector;
3.3) realizing the road traffic jam propagation prediction based on the Decoder
And inputting the low-dimensional feature vector output by the prediction of the LSTM layer into a decoding layer for decoding, and remapping the low-dimensional feature vector into a high-dimensional space to obtain congestion prediction information at the next moment, thereby realizing the propagation prediction of the road traffic congestion.
The method is based on the knowledge graph concept, models and excavates traffic data, and then the deep learning model Conv1D-LSTM-D is used for carrying out congestion prediction, so that a novel solution is provided for predicting traffic congestion.
The invention has the beneficial effects that: by combining knowledge of the knowledge map and the Conv1D-LSTM-D model, a complete road traffic jam prediction solution is provided. According to the method, the time knowledge graph is constructed, deep excavation and standard expression of road traffic flow data are achieved, low-dimensional characteristics and time characteristics of time sequence traffic state vectors are extracted for learning based on the Conv1D-LSTM-D model, and accuracy of road traffic jam propagation prediction is further improved.
The road traffic jam propagation prediction is an important component of intelligent traffic, so that the method can effectively and reasonably establish a model for traffic network data to a certain extent, and has important research significance for relieving urban road jam.
Drawings
FIG. 1 is a schematic flow (q) -density (o) model of a macroscopic base map;
FIG. 2 is a set of quadruplets of a temporal knowledge-graph;
FIG. 3 is a diagram of a Conv1D-LSTM-D deep learning model architecture;
FIG. 4 is a diagram of the LSTM model composition structure.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 to 4, a method for predicting road traffic congestion propagation based on a knowledge graph and Conv1D-LSTM-D includes the following steps:
1) acquiring a lane critical traffic flow state threshold according to a road traffic flow data set, wherein the process is as follows:
according to the flow (q) -density (o) model in the road macroscopic basic map (as shown in fig. 1), qmIs a critical flow threshold value, omFor the critical density threshold, a quadratic function relationship between flow and density is obtained:
wherein v isfFor free flow velocity, ojIs the blocking density;
aiming at each lane, traffic flow data and traffic flow density data of the same lane at different moments are respectively obtained from the road traffic flow data set, a traffic flow-density scatter diagram of the lane is drawn, and a parabola is used for fitting to obtain a critical traffic flow state threshold (qm, o) of each lanem);
2) A traffic jam propagation time knowledge graph is constructed, and the process is as follows:
2.1) defining a set of time-knowledge-graph quadruplets
Defining a set of temporal knowledge-graph quadruplets (as shown in FIG. 2), described as follows:
a quadruple I: intersection-adjacent-intersection, which describes whether two intersections are adjacent at time t; if there is a link between intersection 61 and intersection 62, both may be represented in the form of (intersection 61, adjacent, intersection 62, t);
and a quadruple II: the method comprises the following steps that a lane-jam-intersection describes whether a certain lane at a certain intersection is jammed at the moment t, and the jam state is judged through a threshold value of a critical traffic flow state; if the lane 1 of the intersection 61 is in the traffic jam state at a certain time, it can be represented as (lane 1, traffic jam, intersection 61, t). Conversely, if lane 1 is not congested, it is represented in the form of a quadruple iii: (lane 1, no congestion, intersection 61, t);
2.2) constructing a traffic jam propagation time knowledge graph
Judging the connection relation between intersections according to the geographical position information, and constructing a quadruple I at each moment t; comparing the traffic flow data of each lane at different moments with the critical traffic flow state threshold value, judging the congestion state of the lane at each moment t, and when q is greater>qmOr o>omWhen the traffic congestion propagation time knowledge graph is constructed, a road can be considered to be in a congestion state, otherwise, the road is not congested, a quadruple II and a quadruple III are constructed, and the quadruple is integrated according to a time t to realize construction of the traffic congestion propagation time knowledge graph;
3) the road traffic jam propagation prediction is realized based on the time knowledge graph and Conv1D-LSTM-D, and the process is as follows:
3.1) extracting low-dimensional features of the time-series traffic state vector based on Conv1D and a GlobalmaxPoint layer;
acquiring time series traffic state vectors according to a traffic jam propagation time knowledge graph, wherein each vector represents the state of all lanes at one moment, 1 is used for indicating the smoothness of the lanes, 0 is used for indicating the traffic jam of the lanes, and then the time series traffic state vectors are input into a 1-layer Conv1D and a 1-layer GlobalMaxPoint to extract low-dimensional features;
3.2) learning temporal features of time series traffic state vectors based on the LSTM layer
Taking the low-dimensional features extracted in the previous step as input of an LSTM layer to learn timing sequence features so as to predict the traffic state at the next moment, wherein 4 feedforward neural networks are actually arranged in the LSTM to form a forgetting gate, an input gate and an output gate structure (shown in figure 4) of an LSTM memory unit, the forgetting gate determines whether the memory unit forgets history information or not, the input gate determines the influence of current input data on the state of the memory unit, the output gate determines the output information of the memory unit, and if x is assumed, the output information of the memory unit is determinedtRepresenting the input vector at time t, ht-1And when the output at the time t-1 is represented, the LSTM memory module performs the processes of state updating and information output as follows:
first, the forgetting gate determines the degree f of keeping the past history informationt:
ft=σ(Wf·[ht-1,xt]+bf) (2)
Then, by input gate and ftStorage information C of common pair memory celltUpdating:
it=σ(Wi·[ht-1,xt]+bi) (3)
finally, the output h of the current time t is determined by an output gatet:
ot=σ(Wo[ht-1,xt]+bo) (6)
ht=ot*tanh(Ct) (7)
Wherein f ist,it,otRespectively representing the output states of the forgetting gate, the input gate and the output gate at the current time t, sigma represents an activation function, tanh () represents a hyperbolic activation function, Wf,Wi,WC,WoWeights representing internal feedforward neural networks, bf,bi,bC,boRepresenting a bias vector;
3.3) realizing the road traffic jam propagation prediction based on the Decoder
And inputting the low-dimensional feature vector output by the prediction of the LSTM layer into a decoding layer for decoding, and remapping the low-dimensional feature vector into a high-dimensional space to obtain congestion prediction information at the next moment, thereby realizing the propagation prediction of the road traffic congestion.
The data in the actual experiment are as follows:
(1) selecting experimental data
The original traffic flow data set is SCATS coil data of 30 days in 2017 of Hangzhou city in 6 months, the sampling interval T is 15min, data of 648 lanes in total at 66 intersections are selected from the research area, and 2880 traffic flow data and traffic flow density data are respectively arranged in each lane.
2880 traffic jam propagation time knowledge map conversion is carried out to obtain 2880 time sequence traffic state vectors with 648 dimensions, the first 80% of the traffic state vectors are used as training verification set data, model parameter training and hyper-parameter selection are carried out, the remaining 20% of the traffic state vectors are used as test set data, and algorithm verification is carried out.
(2) Parameter determination
This experimental result all realizes under the based on Tensorflow environment, uses keras to accomplish the frame of whole experimental model and builds.
In the Conv1D-LSTM-D model, 1 layer of one-dimensional convolution layer is arranged, the number of filters of the convolution layer is 20, and the size of the convolution layer is 10; setting 1 layer of global maximum pooling layer (Globalmax pooling); setting 2 layers of LSTM, wherein the output size of each layer is 20, and adopting a tanh activation function; setting 2 decoding layers, wherein the output sizes are 100 and 300 respectively, and adopting a ReLU activation function; and finally, 1 layer of full connection layer is arranged, the number of output units is 648, and a Sigmoid activation function is adopted. The number of model training iterations is 500, the batch size during each iteration training is 128, and Adam is adopted to optimize the training model parameters.
(3) Results of the experiment
The purpose of the experiment of the invention is to predict the traffic jam of the lane, because the traffic clear time in the experimental data is far more than the imbalance caused by the jam time, when judging whether the predicted output is jammed or not, the threshold value which is larger than 0.5 is adopted to carry out the comparison experiment, and the error which gives the true value of 0 is used as the weight which is larger than the error which gives the true value of 0As a loss function, selecting an Accuracy (Accuracy) as an evaluation index of the road traffic jam propagation prediction precision, wherein the calculation formulas are respectively as follows:
wherein, yi=1Representing true values of 1, n1Number of true values 1, yi=0Representing true values of 0, n2The number of true values of 0 is,represents the predicted value and weight represents the weight given to the error of 0 for the true value. m is the number of correct predictions, and n is the total number. The Accuracy index is divided into two cases of a true value of 1 and a true value of 0, which are respectively marked as Accuracy _1 and Accuray _ 0.
The statistics of the results of all road traffic congestion propagation prediction experiments are shown in table 1, where table 1 shows the results of congestion prediction using the Conv1D-LSTM-D model under different weights and threshold values:
table 1.
The embodiments described in this specification are merely illustrative of implementations of the inventive concepts, which are intended for purposes of illustration only. The scope of the present invention should not be construed as being limited to the particular forms set forth in the examples, but rather as being defined by the claims and the equivalents thereof which can occur to those skilled in the art upon consideration of the present inventive concept.
Claims (4)
1. A road traffic congestion propagation prediction method based on a knowledge graph and Conv1D-LSTM-D is characterized by comprising the following steps:
1) acquiring a lane critical traffic flow state threshold according to a road traffic flow data set: aiming at traffic flow data of each lane in a data set, fitting a flow-density scatter diagram of the lane by using a function based on the principle of a macroscopic basic diagram to obtain a critical traffic flow state threshold value of each lane;
2) constructing a traffic jam propagation time knowledge graph: defining a quadruple set of a time knowledge graph, and constructing a traffic jam propagation time knowledge graph according to the geographical position information, the traffic flow data set and the critical traffic flow state threshold of each lane;
3) and (3) realizing road traffic congestion propagation prediction based on the time knowledge graph and Conv 1D-LSTM-D: converting according to a time knowledge graph to obtain a traffic state vector of a time sequence, and extracting low-dimensional features of the input traffic state vector through a one-dimensional convolutional layer Conv1D and a global maximum pooling layer Globalmaxboosting; the time sequence characteristics are learned and the traffic state at the next moment is predicted by taking the time sequence characteristics as the input of an LSTM layer; and finally, the predicted low-dimensional feature vector is re-mapped back to a high-dimensional space through a decoding layer, so that the road traffic jam propagation prediction is realized.
2. The method for predicting road traffic congestion propagation based on knowledge-graph and Conv1D-LSTM-D as claimed in claim 1, wherein the process of step 1) is:
according to the flow (q) -density (o) model in the road macroscopic basic map, qmIs a critical flow threshold value, omFor the critical density threshold, a quadratic function relationship between flow and density is obtained:
wherein v isfFor free flow velocity, ojIs the blocking density;
aiming at each lane, traffic flow data and traffic flow density data of the same lane at different moments are respectively obtained from a road traffic flow data set, a traffic flow-density scatter diagram of the lane is drawn, and a parabola is used for fitting to obtain a critical traffic flow state threshold (q) of each lanem,om)。
3. The method for predicting road traffic congestion propagation based on knowledge-graph and Conv1D-LSTM-D as claimed in claim 1 or 2, wherein in the step 2), the process of constructing the knowledge-graph of traffic congestion propagation time is as follows:
2.1) defining a set of time-knowledge-graph quadruplets
Defining a set of temporal knowledge graph quadruplets, described as follows:
a quadruple I: intersection-adjacent-intersection, which describes whether two intersections are adjacent at time t;
quadruple II/quadruple III: the method comprises the following steps of (1) lane-congestion/non-congestion-intersection, wherein whether a certain lane at a certain intersection is congested at the moment t is described, and the congestion state is judged through a threshold value of a critical traffic flow state;
2.2) constructing a traffic jam propagation time knowledge graph
Judging the connection relation between intersections according to the geographical position information, and constructing a quadruple I at each moment t; comparing the traffic flow data of each lane at different moments with the critical traffic flow state threshold value, judging the congestion state of the lane at each moment t, and when q is greater>qmOr o>omAnd (3) constructing a quadruple II and a quadruple III by considering that the road is in a congestion state, otherwise, the road is not congested, and integrating the quadruple according to the time t to realize the construction of the traffic congestion propagation time knowledge graph.
4. The method for predicting the propagation of road traffic congestion based on the knowledge-graph and Conv1D-LSTM-D as claimed in claim 1 or 2, wherein in the step 3), the prediction process of the propagation of road traffic congestion based on the time knowledge-graph and Conv1D-LSTM-D is implemented as follows:
3.1) extracting low-dimensional features of the time-series traffic state vector based on Conv1D and a GlobalmaxPoint layer;
acquiring time series traffic state vectors according to a traffic jam propagation time knowledge graph, wherein each vector represents the state of all lanes at one moment, 1 is used for indicating the smoothness of the lanes, 0 is used for indicating the traffic jam of the lanes, and then the time series traffic state vectors are input into a 1-layer Conv1D and a 1-layer GlobalMaxPoint to extract low-dimensional features;
3.2) learning temporal features of time series traffic state vectors based on the LSTM layer
The low-dimensional features extracted in the last step are used as input of an LSTM layer to learn timing sequence features and used for predicting the traffic state of the next moment, 4 feedforward neural networks are actually arranged in the LSTM to form a forgetting gate, an input gate and an output gate structure of an LSTM memory unit, the forgetting gate determines whether the memory unit forgets historical information or not, the input gate determines the influence of current input data on the state of the memory unit, the output gate determines the output information of the memory unit, and if x is assumed, the output gate determines the output information of the memory unittRepresenting the input vector at time t, ht-1And when the output at the time t-1 is represented, the LSTM memory module performs the processes of state updating and information output as follows:
first, the forgetting gate determines the degree f of keeping the past history informationt:
ft=σ(Wf·[ht-1,xt]+bf) (2)
Then, by input gate and ftStorage information C of common pair memory celltUpdating:
it=σ(Wi·[ht-1,xt]+bi) (3)
finally, the output h of the current time t is determined by an output gatet:
ot=σ(Wo[ht-1,xt]+bo) (6)
ht=ot*tanh(Ct) (7)
Wherein f ist,it,otRespectively representing the output states of the forgetting gate, the input gate and the output gate at the current time t, sigma represents an activation function, tanh () represents a hyperbolic activation function, Wf,Wi,WC,WoWeights representing internal feedforward neural networks, bf,bi,bC,boRepresenting a bias vector;
3.3) realizing the road traffic jam propagation prediction based on the Decoder
And inputting the low-dimensional feature vector output by the prediction of the LSTM layer into a decoding layer for decoding, and remapping the low-dimensional feature vector into a high-dimensional space to obtain congestion prediction information at the next moment, thereby realizing the propagation prediction of the road traffic congestion.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011107018.3A CN112530157B (en) | 2020-10-16 | 2020-10-16 | Road traffic congestion propagation prediction method based on knowledge graph and Conv1D-LSTM-D |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011107018.3A CN112530157B (en) | 2020-10-16 | 2020-10-16 | Road traffic congestion propagation prediction method based on knowledge graph and Conv1D-LSTM-D |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112530157A true CN112530157A (en) | 2021-03-19 |
CN112530157B CN112530157B (en) | 2022-05-03 |
Family
ID=74979218
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011107018.3A Active CN112530157B (en) | 2020-10-16 | 2020-10-16 | Road traffic congestion propagation prediction method based on knowledge graph and Conv1D-LSTM-D |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112530157B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113763712A (en) * | 2021-10-19 | 2021-12-07 | 西南交通大学 | Regional traffic jam tracing method based on travel event knowledge graph |
CN116543561A (en) * | 2023-07-06 | 2023-08-04 | 之江实验室 | Knowledge and data double-drive-based traffic congestion propagation prediction method |
CN117275236A (en) * | 2023-10-11 | 2023-12-22 | 宁波宁工交通工程设计咨询有限公司 | Traffic jam management method and system based on multi-target recognition |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1548407A1 (en) * | 2003-12-26 | 2005-06-29 | Aisin Aw Co., Ltd. | System and method of displaying predicted traffic information |
US20140309976A1 (en) * | 2013-04-16 | 2014-10-16 | International Business Machines Corporation | Performing-time-series based predictions with projection thresholds using secondary time-series-based information stream |
CN110070715A (en) * | 2019-04-29 | 2019-07-30 | 浙江工业大学 | A kind of road traffic flow prediction method based on Conv1D-NLSTMs neural network structure |
CN110164129A (en) * | 2019-04-25 | 2019-08-23 | 浙江工业大学 | Single Intersection multi-lane traffic flow amount prediction technique based on GERNN |
CN110442731A (en) * | 2019-07-24 | 2019-11-12 | 中电科新型智慧城市研究院有限公司 | A kind of traffic operation system based on traffic administration knowledge mapping |
CN110599766A (en) * | 2019-08-22 | 2019-12-20 | 浙江工业大学 | Road traffic jam propagation prediction method based on SAE-LSTM-SAD |
CN111160753A (en) * | 2019-12-25 | 2020-05-15 | 大连理工大学 | Knowledge graph-based road network node importance evaluation method |
CN111462485A (en) * | 2020-03-31 | 2020-07-28 | 电子科技大学 | Traffic intersection congestion prediction method based on machine learning |
CN111768618A (en) * | 2020-06-04 | 2020-10-13 | 北京航空航天大学 | Traffic jam state propagation prediction and early warning system and method based on city portrait |
-
2020
- 2020-10-16 CN CN202011107018.3A patent/CN112530157B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1548407A1 (en) * | 2003-12-26 | 2005-06-29 | Aisin Aw Co., Ltd. | System and method of displaying predicted traffic information |
US20140309976A1 (en) * | 2013-04-16 | 2014-10-16 | International Business Machines Corporation | Performing-time-series based predictions with projection thresholds using secondary time-series-based information stream |
CN110164129A (en) * | 2019-04-25 | 2019-08-23 | 浙江工业大学 | Single Intersection multi-lane traffic flow amount prediction technique based on GERNN |
CN110070715A (en) * | 2019-04-29 | 2019-07-30 | 浙江工业大学 | A kind of road traffic flow prediction method based on Conv1D-NLSTMs neural network structure |
CN110442731A (en) * | 2019-07-24 | 2019-11-12 | 中电科新型智慧城市研究院有限公司 | A kind of traffic operation system based on traffic administration knowledge mapping |
CN110599766A (en) * | 2019-08-22 | 2019-12-20 | 浙江工业大学 | Road traffic jam propagation prediction method based on SAE-LSTM-SAD |
CN111160753A (en) * | 2019-12-25 | 2020-05-15 | 大连理工大学 | Knowledge graph-based road network node importance evaluation method |
CN111462485A (en) * | 2020-03-31 | 2020-07-28 | 电子科技大学 | Traffic intersection congestion prediction method based on machine learning |
CN111768618A (en) * | 2020-06-04 | 2020-10-13 | 北京航空航天大学 | Traffic jam state propagation prediction and early warning system and method based on city portrait |
Non-Patent Citations (2)
Title |
---|
JINGYUAN 等: "DrivingwithKnowledgefromthePhysicalWorld", 《PROCEEDINGSOFTHE17THSIGKDDCONFERENCEONKNOWLEDGEDISCOVERYAND DATAMININ》 * |
周光临: "组合知识图谱和深度学习的城市交通拥堵区域预测研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113763712A (en) * | 2021-10-19 | 2021-12-07 | 西南交通大学 | Regional traffic jam tracing method based on travel event knowledge graph |
CN113763712B (en) * | 2021-10-19 | 2022-05-10 | 西南交通大学 | Regional traffic jam tracing method based on travel event knowledge graph |
CN116543561A (en) * | 2023-07-06 | 2023-08-04 | 之江实验室 | Knowledge and data double-drive-based traffic congestion propagation prediction method |
CN116543561B (en) * | 2023-07-06 | 2023-10-31 | 之江实验室 | Knowledge and data double-drive-based traffic congestion propagation prediction method |
CN117275236A (en) * | 2023-10-11 | 2023-12-22 | 宁波宁工交通工程设计咨询有限公司 | Traffic jam management method and system based on multi-target recognition |
CN117275236B (en) * | 2023-10-11 | 2024-04-05 | 宁波宁工交通工程设计咨询有限公司 | Traffic jam management method and system based on multi-target recognition |
Also Published As
Publication number | Publication date |
---|---|
CN112530157B (en) | 2022-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112530157B (en) | Road traffic congestion propagation prediction method based on knowledge graph and Conv1D-LSTM-D | |
CN109887282B (en) | Road network traffic flow prediction method based on hierarchical timing diagram convolutional network | |
CN111210633B (en) | Short-term traffic flow prediction method based on deep learning | |
CN111860951B (en) | Rail transit passenger flow prediction method based on dynamic hypergraph convolutional network | |
CN109142171B (en) | Urban PM10 concentration prediction method based on feature expansion and fusing with neural network | |
CN111292525B (en) | Traffic flow prediction method based on neural network | |
CN115240425B (en) | Traffic prediction method based on multi-scale space-time fusion graph network | |
CN110942637B (en) | SCATS system road traffic flow prediction method based on airspace map convolutional neural network | |
CN109272157A (en) | A kind of freeway traffic flow parameter prediction method and system based on gate neural network | |
CN109920248B (en) | Bus arrival time prediction method based on GRU neural network | |
CN112489426B (en) | Urban traffic flow space-time prediction scheme based on graph convolution neural network | |
CN113538910B (en) | Self-adaptive full-chain urban area network signal control optimization method | |
CN110164129B (en) | Single-intersection multi-lane traffic flow prediction method based on GERNN | |
Sun et al. | Passenger flow prediction of subway transfer stations based on nonparametric regression model | |
Parsa et al. | Pabo: Pseudo agent-based multi-objective bayesian hyperparameter optimization for efficient neural accelerator design | |
CN110929958A (en) | Short-term traffic flow prediction method based on deep learning parameter optimization | |
CN111341109A (en) | City-level signal recommendation system based on space-time similarity | |
CN115496257A (en) | Short-term vehicle speed prediction based on space-time fusion | |
Wu et al. | Traffic signal networks control optimize with PSO algorithm | |
Liu et al. | An intelligence optimized rolling grey forecasting model fitting to small economic dataset | |
CN116631203A (en) | Self-adaptive traffic signal control method combining traffic flow prediction and reinforcement learning | |
CN116256468A (en) | Air quality prediction optimization method based on dynamic self-adaptive particle swarm | |
CN112766240A (en) | Residual multi-graph convolution crowd distribution prediction method and system based on space-time relationship | |
CN113947238A (en) | Urban pollutant concentration prediction method based on multi-site space-time correlation | |
Huang et al. | A long short-term memory based wind power prediction method |
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 |