CN114999154B - Road service level prediction method based on graph rolling network - Google Patents
Road service level prediction method based on graph rolling network Download PDFInfo
- Publication number
- CN114999154B CN114999154B CN202210589674.4A CN202210589674A CN114999154B CN 114999154 B CN114999154 B CN 114999154B CN 202210589674 A CN202210589674 A CN 202210589674A CN 114999154 B CN114999154 B CN 114999154B
- Authority
- CN
- China
- Prior art keywords
- road
- traffic
- service level
- traffic flow
- flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000005096 rolling process Methods 0.000 title claims abstract description 9
- 238000013507 mapping Methods 0.000 claims abstract description 4
- 238000012545 processing Methods 0.000 claims abstract description 3
- 239000011159 matrix material Substances 0.000 claims description 22
- 239000000523 sample Substances 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 5
- 238000012067 mathematical method Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 230000004913 activation Effects 0.000 claims description 2
- 230000000903 blocking effect Effects 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 claims description 2
- 238000012887 quadratic function Methods 0.000 claims description 2
- 238000013473 artificial intelligence Methods 0.000 abstract description 3
- 230000000295 complement effect Effects 0.000 abstract 1
- 238000003062 neural network model Methods 0.000 abstract 1
- 238000010276 construction Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process 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
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- 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
- 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
- G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
A road service level prediction method based on a graph rolling network comprises the following steps: 1) Acquiring and processing traffic flow data, and performing outlier rejection and default value complement on the data acquired by the equipment; 2) Establishing a road network, and establishing a road network topological graph from the correlation between road sections and the similarity between the road sections; 3) Fitting the flow and density data by a least square method to obtain a maximum traffic flow value of the road; 4) Traffic flow prediction, namely inputting historical flow and speed data into a neural network model to obtain predicted flow and speed data; 5) And predicting the road service level, calculating to obtain the road density and saturation rate at the predicted moment according to the data predicted in the previous step, and obtaining the score of the road service level through a mapping function by combining the density and the saturation rate. The invention combines the artificial intelligence method to predict the future traffic condition of the road, and greatly improves the existing road service level.
Description
Technical Field
The invention relates to the application fields of intelligent traffic and artificial intelligence, in particular to a road service level prediction method based on a graph rolling network.
Background
Expressways are used as the most basic traffic infrastructure, are important struts for connecting other traffic modes and comprehensively playing the whole efficiency of a traffic network, and play an irreplaceable role in a comprehensive traffic system. At present, the road construction in China develops rapidly, but has obvious defects such as low road quality, road construction lag and the like, so that the overall service level of the road is low, and the traffic efficiency is low.
Traffic flow prediction is one of the key problems of an intelligent traffic system, and is a process for analyzing traffic conditions (such as speed, flow, density and the like) of a road network, excavating traffic modes and predicting future traffic conditions of the road network. Traffic flow prediction may enable various intelligent applications. For example, it may help private drivers to route and schedule departure times, helping traffic managers to improve traffic efficiency and safety.
The road service level refers to a measure of the quality of service perceived by drivers and passengers in a traffic stream, i.e. the level of operational service provided by a road under certain traffic conditions. Reasonable prediction is carried out on the road service level, so that the future road service quality can be known, and scientific basis is provided for improving road conditions and improving traffic management level.
In this case, by introducing the neural network, future traffic conditions of the road can be predicted, and the existing road service level can be greatly improved.
Disclosure of Invention
Aiming at the problems that the existing road service level prediction scheme cannot fully utilize traffic flow data and cannot timely and effectively represent the traffic condition occurring in the future, the invention designs a road service level prediction method based on the neural network by utilizing traffic flow information and according to the road saturation rate and density, and provides a basis for road managers to accurately grasp the actual condition of road traffic and formulate reasonable traffic policy measures.
The technical scheme adopted for solving the technical problems is as follows:
a road service level prediction method based on a graph rolling network comprises the following steps:
1) Acquiring traffic data
The method comprises the steps of obtaining lane flow, driving speed and geographic position of a road at corresponding moments of each road section through equipment such as monitoring equipment, probes and the like on the road, setting a threshold value to remove abnormal values in traffic flow data, and complementing the abnormal values by using a cube interpolation method;
2) Road network establishment
Construction of road network topology maps is a key step, and if the generated maps cannot well encode the correlation between roads, model learning is not facilitated, and prediction performance may even be reduced. The road network topological graph is constructed from two aspects of road network topological structure and traffic mode correlation, and the road of the road network topological graph is represented by equipment for acquiring traffic flow information on the road section. The road network topology structure considers the relativity between roads, if the road sections are communicated, the two road sections are related, and the relativity degree is determined by the hop count between the two road sections (the hop count refers to the number of nodes from one road node to another road node); the traffic mode correlation is to consider the similarity between the historical traffic conditions of roads, take the historical traffic flow sequences with the same length of different roads, and use cosine similarity to obtain the similarity between the two sequences;
3) Obtaining the maximum traffic flow of the road
And (3) calculating the lane density according to the historical lane flow and driving speed information obtained in the step (1) and the relation among the flow, the speed and the density. Fitting by using a least square method to obtain a flow-density curve, and calculating the maximum value of the curve through the obtained curve parameter by using a mathematical method, namely, the maximum traffic flow of the corresponding road;
4) Traffic flow prediction
Given a traffic condition (vehicle speed and traffic flow data) sample I t ，I t Can be regarded as an r×h matrix, where R is the number of roads and H is the number of historical traffic condition signals; processing I using a sliding window of window size h and step size d t Then we will get a J-stretch sequence (i.e) Each being an R x h matrix;
inputting the obtained traffic flow sequence, the adjacency matrix of the road network directed graph and the adjacency matrix of the historical traffic mode into a graph rolling network (GCN) to capture the spatial dependence of the road network; the output of the GCN is then input to a long-short-term memory (LSTM) to capture dynamic changes in the time of road traffic data. Finally, the predicted traffic flow information (vehicle speed and vehicle flow data) is obtained through the full connection layer;
5) Road service level prediction
And (3) calculating the density of the corresponding lane according to the lane speed and flow data obtained in the step (4). And then obtaining the saturation rate through the calculation of the maximum traffic flow obtained in the step 3). And obtaining the score of the road service level through a mapping function according to the service level grading table of the expressway basic road section.
The beneficial effects of the invention are as follows: and the method is combined with an artificial intelligence method to predict the future traffic condition of the road, so that the existing road service level is greatly improved.
Drawings
FIG. 1 shows a schematic flow diagram of an embodiment of the present invention;
fig. 2 shows a predictive model diagram of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 and 2, a road service level prediction method based on a graph rolling network includes the steps of:
1) Acquiring traffic data
The method comprises the steps of obtaining lane flow, driving speed and geographical coordinates of a road at corresponding moments of each road section through equipment such as monitoring equipment, probes and the like on the road, setting a threshold value to remove abnormal values in traffic flow data, and complementing the abnormal values by using a cube interpolation method, wherein the formula is as follows:
v＝((v 3 -v 2 )-(v 0 -v 1 ))t 3 +(2(v 0 -v 1 )-(v 3 -v 2 ))t 2 +(v 2 -v 0 )t+v 1 (1)
wherein v is a default value and represents traffic flow data acquired by road equipment at a certain moment, v 0 、v 1 、v 2 、v 3 Data representing 4 successive moments, t being [0,1]Will produce a segment of the connection v 1 And v 2 Is a curve of (2);
2) Road network establishment
The construction of the road network topological graph is a key step, if the generated graph cannot well encode the correlation between roads, model learning is not facilitated, prediction performance is possibly even reduced, and the road network topological graph is constructed from two aspects of road network topological structure (considering the correlation between roads) and traffic mode correlation (considering the similarity between historical traffic conditions of roads). The road is represented by a device on the road section that obtains traffic flow information;
constructing road network directed graph G according to hop count between road sections d ＝(V,E,W d ) Node V i Representing road section, edge e ij Representing road segment V i And road section V j Correlation between, edge e ij Weight W of (2) ij Representing road segment V i And road section V j The strength of the correlation between the two is defined by W ij Forming a topology adjacency matrix W d 。W ij The definition is as follows:
if it is unable to run from road section V i To road section V j W is then ij ＝0，
Constructing road network traffic pattern graph G according to historical traffic conditions among road sections h = (V ', E', wh), node V i Representing road section, edge e ij ' represent road segment V i And road section V j Similarity between historical traffic conditions, edge e ij Weight W of ij ' represent road segment V i And road section V j The similarity between the two is defined by W ij ' composition topology adjacency matrix W h Taking traffic flow sequences of two nodes, P i ＝(x 1 ,x 2 ,...,x n )、P j ＝(y 1 ,y 2 ,...,y n )，W ij ' is defined as:
3) Obtaining the maximum traffic flow of the road
Based on the lane flow Q (vehicle/h) and the driving speed information V (Km/h) obtained in step 1), the formula for calculating the lane density K (vehicle/Km) is as follows:
Q＝K·V (4)
the relationship between speed and density is analyzed by green Hill's algorithm based on the statistical law of traffic flow:
wherein V is f Is the theoretical highest speed, i.e. the free flow speed; k (K) j Is the blocking density; v is the traffic speed when the traffic density is K;
the combination of formulas (4) and (5) results in:
the flow and density obtained by the formula (6) are quadratic function relationship, the Q-K curve is obtained by least square fitting,calculating the maximum value of the curve through the obtained curve parameters by a mathematical method, namely, corresponding to the maximum traffic flow Q of the road max ；
4) Traffic flow prediction
Given a traffic condition (vehicle speed and traffic flow data) sample I t ，I t Can be regarded as an RxH matrix, wherein R is the number of roads, H is the number of historical traffic signals, and I is processed by using a sliding window with window size H and step size d t Then we will get a J-stretch sequence (i.e) Each being an R x h matrix;
inputting the obtained traffic flow sequence, the adjacency matrix of the road network directed graph and the adjacency matrix of the historical traffic mode into a graph rolling network (GCN) to capture the spatial dependence of the road network, wherein the formula is as follows:
wherein X is an adjacency matrix, the form is as follows:
the form of the adjacency matrix being self-connecting is as follows:
for a diagonal matrix, the formula is as follows:
H (0) is traffic flow data input into GCN, and ReLU is an activation function;
the output of two different GCNs is fused and then input into a long-short-term memory LSTM to capture the dynamic change of road traffic data in time, and finally the predicted traffic flow information (vehicle speed and vehicle flow data) is obtained through a full-connection layer, and the model structure is shown in the figure 2;
5) Road service level prediction
The predicted lane speed V according to step 4) predict Flow rate Q predict Calculating the density K of the corresponding lane according to the formula (4), and calculating the saturation ratio S through the maximum traffic flow obtained in the step 3):
according to the basic road section service level grading table of the expressway, the obtained density and saturation rate are subjected to a mapping function f (.) to obtain a score of the road service level:
y road service level score ＝f(S,K)。
Table 1 is a basic highway section service level grading table.
TABLE 1
Finally, it should be noted that: the above examples are only specific embodiments of the present invention, and are not intended to limit the scope of the present invention, but it should be understood by those skilled in the art that the present invention is not limited thereto, and that the present invention is described in detail with reference to the foregoing examples: any person skilled in the art may modify or easily conceive of the technical solution described in the foregoing embodiments, or perform equivalent substitution of some of the technical features, while remaining within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (1)
1. A method for predicting a road service level based on a graph roll-up network, the method comprising the steps of:
1) Acquisition and processing of traffic data
The monitoring on the road, the probe equipment acquires the traffic flow, the driving speed and the geographic coordinates of the road at the corresponding moment of each road section, the threshold value is set to eliminate abnormal values in the traffic flow data, and the default values are complemented by using a cube interpolation method, wherein the formula is as follows:
v＝((v 3 -v 2 )-(v 0 -v 1 ))t 3 +(2(v 0 -v 1 )-(v 3 -v 2 ))t 2 +(v 2 -v 0 )t+v 1 (1)
wherein v is a default value and represents traffic flow data acquired by road equipment at a certain moment, v 0 、v 1 、v 2 、v 3 Data representing 4 successive moments, t being [0,1]Will produce a segment of the connection v 1 And v 2 Is a curve of (2);
2) Road network establishment
Constructing road network directed graph G according to hop count between road sections d ＝(V,E,W d ) Node V i Representing road section, edge e ij Representing road segment V i And road section V j Correlation between, edge e ij Weight W of (2) ij Representing road segment V i And road section V j The strength of the correlation between the two is defined by W ij Composition of topology mapAdjacency matrix W d ，W ij The definition is as follows:
if it is unable to run from road section V i To road section V j W is then ij ＝0；
Constructing road network traffic pattern graph G according to historical traffic conditions among road sections h ＝(V′,E′,W h ) Node V i Representing road section, edge e ij ' represent road segment V i And road section V j Similarity between historical traffic conditions, edge e ij Weight W of ij ' represent road segment V i And road section V j The similarity between the two is defined by W ij ' composition topology adjacency matrix W h Taking traffic flow sequences of two nodes, P i ＝(x 1 ,x 2 ,...,x n )、P j ＝(y 1 ,y 2 ,...,y n )，W ij ' is defined as:
3) Obtaining the maximum traffic flow of the road
According to the lane flow Q and the driving speed information V obtained in the step 1), the formula for calculating the lane density K is as follows:
Q＝K·V (4)
the relationship between speed and density is analyzed by green Hill's algorithm based on the statistical law of traffic flow:
wherein V is f Is the theoretical highest speed, i.e. the free flow speed; k (K) j Is the blocking density; v is the traffic speed when the traffic density is K;
the combination of formulas (4) and (5) results in:
obtaining a Q-K curve by using a least square fitting method and adopting the flow and density obtained in the formula (6) as a quadratic function relation, and calculating the maximum value of the curve through the obtained curve parameters by using a mathematical method, namely, obtaining the maximum traffic flow Q of the corresponding road max ；
4) Traffic flow prediction
Given a traffic condition sample I t ，I t Can be regarded as an RxH matrix, wherein R is the number of roads, H is the number of historical traffic signals, and I is processed by using a sliding window with window size H and step size d t Then, a J-segment sequence is obtained, i.eEach being an R x h matrix;
inputting the obtained traffic flow sequence, the adjacency matrix of the road network directed graph and the adjacency matrix of the historical traffic mode into a graph rolling network GCN to capture the spatial dependence of the road network, wherein the formula is as follows:
wherein X is an adjacency matrix, the form is as follows:
the form of the adjacency matrix being self-connecting is as follows:
for a diagonal matrix, the formula is as follows:
H (0) is traffic flow data input into GCN, and ReLU is an activation function;
the output of two different GCNs is fused and then input to a long-short-term memory LSTM to capture the dynamic change of road traffic data in time, and finally the predicted traffic flow information is obtained through a full-connection layer;
5) Road service level prediction
The predicted lane speed V according to step 4) predict Flow rate Q predict Calculating the density K of the corresponding lane according to the formula (4), and calculating the saturation ratio S through the maximum traffic flow obtained in the step 3):
obtaining the road service level by passing the obtained density and saturation through a mapping function f (-) according to the basic road section service level grading table of the expressway:
y road service level score ＝f(S,K)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210589674.4A CN114999154B (en) | 2022-05-26 | 2022-05-26 | Road service level prediction method based on graph rolling network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210589674.4A CN114999154B (en) | 2022-05-26 | 2022-05-26 | Road service level prediction method based on graph rolling network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114999154A CN114999154A (en) | 2022-09-02 |
| CN114999154B true CN114999154B (en) | 2024-02-02 |
Family
ID=83028433
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210589674.4A Active CN114999154B (en) | 2022-05-26 | 2022-05-26 | Road service level prediction method based on graph rolling network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114999154B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115359662B (en) * | 2022-10-18 | 2023-01-10 | 智道网联科技(北京)有限公司 | Lane congestion prediction method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021109318A1 (en) * | 2019-12-03 | 2021-06-10 | 东南大学 | Method for estimating and predicting short-term traffic circulation state of urban road network |
| CN113053115A (en) * | 2021-03-17 | 2021-06-29 | 北京工商大学 | Traffic prediction method based on multi-scale graph convolution network model |
| CN113112819A (en) * | 2021-03-26 | 2021-07-13 | 华南理工大学 | Improved LSTM-based graph convolution traffic speed prediction method |
| WO2021212866A1 (en) * | 2020-04-21 | 2021-10-28 | 长安大学 | Vehicle travel volume prediction model construction method, and prediction method and system |
-
2022
- 2022-05-26 CN CN202210589674.4A patent/CN114999154B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021109318A1 (en) * | 2019-12-03 | 2021-06-10 | 东南大学 | Method for estimating and predicting short-term traffic circulation state of urban road network |
| WO2021212866A1 (en) * | 2020-04-21 | 2021-10-28 | 长安大学 | Vehicle travel volume prediction model construction method, and prediction method and system |
| CN113053115A (en) * | 2021-03-17 | 2021-06-29 | 北京工商大学 | Traffic prediction method based on multi-scale graph convolution network model |
| CN113112819A (en) * | 2021-03-26 | 2021-07-13 | 华南理工大学 | Improved LSTM-based graph convolution traffic speed prediction method |
Non-Patent Citations (4)
| Title |
|---|
| LSGCN: Long Short-Term Traffic Prediction with Graph Convolutional Networks;Rongzhou Huang , et al;Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence (IJCAI-20);全文 * |
| Traffic prediction based on GCN-LSTM model;Zhizhu Wu et al.;Journal of Physics: Conference Series;全文 * |
| 基于图卷积神经网络的交通流量预测;朱凯利;朱海龙;刘靖宇;石晔琼;王欢;;智能计算机与应用(第06期);全文 * |
| 基于图卷积网络的路网短时交通流预测研究;陈喜群等;交通运输***工程与信息;第第20卷卷(第第4期期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114999154A (en) | 2022-09-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109754605B (en) | Traffic prediction method based on attention temporal graph convolution network | |
| CN112561146B (en) | Large-scale real-time traffic flow prediction method based on fuzzy logic and depth LSTM | |
| CN113313947B (en) | Road condition evaluation method of short-term traffic prediction graph convolution network | |
| CN112489426B (en) | Urban traffic flow space-time prediction scheme based on graph convolution neural network | |
| CN112561191B (en) | Prediction model training method, prediction device, prediction apparatus, prediction program, and program | |
| GB2599765A (en) | Vehicle traffic flow prediction method with missing data | |
| CN109584552A (en) | A kind of public transport arrival time prediction technique based on network vector autoregression model | |
| CN114999154B (en) | Road service level prediction method based on graph rolling network | |
| CN114881356A (en) | Urban traffic carbon emission prediction method based on particle swarm optimization BP neural network optimization | |
| Zhang et al. | TrafficGAN: Off-deployment traffic estimation with traffic generative adversarial networks | |
| CN112542049A (en) | Comprehensive management and control platform for intelligent traffic | |
| CN114495507A (en) | Traffic flow prediction method integrating space-time attention neural network and traffic model | |
| CN114757389A (en) | Federal learning-based urban traffic flow space-time prediction method | |
| Thakurzx et al. | Modeling and characterization of urban vehicular mobility using web cameras | |
| CN116227362A (en) | Municipal drainage pipe network state prediction method based on graph and deep learning | |
| Li et al. | Deep reinforcement learning for optimal rescue path planning in uncertain and complex urban pluvial flood scenarios | |
| Zhang et al. | Off-deployment traffic estimation—a traffic generative adversarial networks approach | |
| Jiang et al. | The study on the application of fuzzy clustering analysis in the dynamic identification of road traffic state | |
| Ramhormozi et al. | Multi-task graph neural network for truck speed prediction under extreme weather conditions | |
| CN113554221A (en) | Simulation and prediction method for town development boundary under 'flow space' view angle | |
| Wang et al. | Identifying the key nodes and sections of urban roadway network based on GPS trajectory data | |
| CN112330150A (en) | Early-stage judging and identifying method for expandability of large-area debris flow disasters | |
| Cheng et al. | Method for predicting travel time of motor vehicle based on stack noise reduction self-encoder | |
| Han | Framework for real-time traffic management with case studies | |
| CN117576918B (en) | Urban road flow universe prediction method based on multi-source 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 |