CN116978218A - Urban traffic flow prediction method, system and equipment based on generation countermeasure network - Google Patents

Urban traffic flow prediction method, system and equipment based on generation countermeasure network Download PDF

Info

Publication number
CN116978218A
CN116978218A CN202310681033.6A CN202310681033A CN116978218A CN 116978218 A CN116978218 A CN 116978218A CN 202310681033 A CN202310681033 A CN 202310681033A CN 116978218 A CN116978218 A CN 116978218A
Authority
CN
China
Prior art keywords
traffic flow
city
countermeasure network
research
research city
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.)
Pending
Application number
CN202310681033.6A
Other languages
Chinese (zh)
Inventor
魏志成
张韬毅
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hebei Normal University
Original Assignee
Hebei Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hebei Normal University filed Critical Hebei Normal University
Priority to CN202310681033.6A priority Critical patent/CN116978218A/en
Publication of CN116978218A publication Critical patent/CN116978218A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • 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
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • 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
    • G06N3/0475—Generative 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • 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
    • G06N3/094—Adversarial learning
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A30/00—Adapting or protecting infrastructure or their operation
    • Y02A30/60—Planning or developing urban green infrastructure

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a city traffic flow prediction method, a system and equipment based on a generated countermeasure network, and relates to the field of city traffic flow prediction, wherein the method comprises the following steps: determining an embedding matrix of the current stage research city by adopting a Node2Vec graph embedding algorithm according to travel demand data of each traffic flow area in the current stage research city; and taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into a trained generator, and predicting the traffic flow of the research city at the future stage. The invention can capture the basic mode of how the traffic flow evolves along with the change of travel demands, thereby realizing accurate prediction of the traffic flow.

Description

Urban traffic flow prediction method, system and equipment based on generation countermeasure network
Technical Field
The invention relates to the field of urban traffic flow prediction, in particular to an urban traffic flow prediction method, an urban traffic flow prediction system and urban traffic flow prediction equipment based on a generated countermeasure network.
Background
Along with the continuous promotion of the urban process, an Intelligent Transportation System (ITS) plays an important role in facilitating urban travel and promoting urban economy and cultural development. Accurate traffic flow prediction is crucial to the development of ITS, and can provide valuable insight for traffic planning, resident trips, and operational decisions of traffic departments. Therefore, more and more traffic flow prediction models are proposed by researchers.
There are two primary traffic flow prediction methods initially: statistical models and models based on machine learning. The former includes methods such as Historical Average (HA) models, autoregressive integrated moving average (ARIMA) models, and various variants thereof. The latter includes K-nearest neighbor (KNN), random forest, support Vector Machine (SVM) models, and the like.
In recent years, the rapid development of deep learning has shown significant advantages in solving the complex nonlinear regression problem. The deep learning model can effectively solve the inherent defects of the multi-module model (the realization of the local optimal performance can not guarantee the global optimal performance) by utilizing an end-to-end learning method. Currently, traffic flow prediction models are mainly proposed based on deep learning technologies, such as Convolutional Neural Networks (CNNs), recurrent Neural Networks (RNNs), graph roll-up networks (GCNs), graph annotation force networks (GATs), generation countermeasure networks (GANs), and the like.
Some researchers have presented challenges in solving traffic flow predictions using deep learning methods. For example, using CNNs to capture spatial dependencies and RNNs to capture temporal dependencies. Zhang et al proposed ST-ResNet to predict crowd flow. Convolution operations are commonly used to process euclidean structured data. However, traffic data is non-euclidean structured. Thus, conventional convolution operations may not be effective in capturing complex spatial relationships and patterns inherent in traffic data. To address this challenge, researchers have developed GCNs. Some studies utilize GCN to extract spatial and temporal features in traffic prediction. Zhang et al propose STGCN to solve the problem of time series prediction in the traffic domain. Lu et al propose a spatio-temporal adaptive gated convolutional network to predict traffic after several time steps. Peng et al designed a dynamic graph circular convolution network for urban traffic passenger flow prediction. Zhang et al propose a dynamic graph rolling network based on spatiotemporal data embedding to more accurately and stably predict traffic flow. Lv et al propose T-MGCN, employing a multi-graph convolution network and a recurrent neural network for traffic flow prediction. Wang et al introduced graph convolution neural networks based on hypergraph structures to capture the spatiotemporal features of traffic data, while Li et al combined gating mechanisms with graph convolution neural networks for traffic prediction. Ni et al propose a one-dimensional convolutional neural network and channel attention mechanism for extracting temporal features, and a multi-graph convolutional network framework and gating mechanism for capturing spatial features. However, the traffic pattern has inherent complexity and dynamic variability, so that the inherent spatiotemporal dependency characteristics between traffic flow areas cannot be obtained according to a certain determination state, and the above model for traffic prediction using GCN ignores the change situation of the traffic pattern in the captured history data.
Disclosure of Invention
The invention aims to provide a city traffic flow prediction method, system and equipment based on a generated countermeasure network, which can capture a basic mode of how traffic flow evolves along with the change of travel demands, so as to realize accurate prediction of traffic flow.
In order to achieve the above object, the present invention provides the following solutions:
in a first aspect, the present invention provides a method for urban traffic flow prediction based on generation of an countermeasure network, comprising:
acquiring travel demand data of each traffic flow area in a research city at the current stage;
determining an embedding matrix of the current stage research city by adopting a Node2Vec graph embedding algorithm according to travel demand data of each traffic flow area in the current stage research city;
and taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into a trained generator, and predicting the traffic flow of the research city at the future stage.
In a second aspect, the present invention provides an urban traffic flow prediction system based on generation of an countermeasure network, comprising:
the travel demand data acquisition module is used for acquiring travel demand data of each traffic flow area in the research city at the current stage;
the embedded matrix calculation module is used for determining an embedded matrix of the current stage research city by adopting a Node2Vec graph embedded algorithm according to travel demand data of each traffic flow area in the current stage research city;
and the traffic flow prediction module is used for taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into the trained generator, and predicting the traffic flow of the research city at the future stage.
In a third aspect, the present invention provides an electronic device comprising a memory for storing a computer program and a processor that runs the computer program to cause the electronic device to perform a method of urban traffic flow prediction based on generating an countermeasure network according to the first aspect.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention introduces a method for constructing a traffic embedded graph based on a Node2Vec graph embedding algorithm, which converts the Node relation of the graph into a two-dimensional embedded matrix so that traffic data is suitable for convolutional neural network processing, and specifically comprises the following steps: the convolutional neural network based on the Node2Vec graph embedding algorithm is adopted to capture the deep space-time dependency relationship, and the basic mode of how the traffic flow evolves along with the change of travel demands is captured by generating an countermeasure network frame, so that the accurate prediction of the traffic flow is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a city traffic flow prediction method based on a generated countermeasure network according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a city traffic flow prediction model (TGAN) based on generation of an countermeasure network according to an embodiment of the present invention; fig. 2 (a) is a schematic structural diagram of a generator according to an embodiment of the present invention; fig. 2 (b) is a schematic structural diagram of a discriminator according to the embodiment of the invention;
fig. 3 is a heat map of traffic flow distribution and real traffic flow distribution generated by different models provided by an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
Example 1
With the rapid development of intelligent traffic systems, accurate traffic flow prediction has become an important aspect of traffic planning and management. However, existing methods mostly ignore changes in traffic patterns in captured history data. In order to solve the problem, the embodiment provides a deep learning model which combines a convolutional neural network based on a Node2Vec graph embedding algorithm and generates an countermeasure network frame, and the smart combination can effectively capture a basic mode of deep space-time dependency relationship and evolution of traffic flow along with travel demand change. Experimental results performed on two real world traffic data sets indicate that the proposed model performs better than other baseline models, showing its potential as a tool for traffic management and city planning.
As shown in fig. 1, the urban traffic flow prediction method based on the generation of the countermeasure network provided in this embodiment includes:
step 100: and acquiring travel demand data of each traffic flow area in the research city at the current stage.
Step 200: and determining an embedding matrix of the current stage research city by adopting a Node2Vec graph embedding algorithm according to travel demand data of each traffic flow area in the current stage research city.
Step 300: and taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into a trained generator, and predicting the traffic flow of the research city at the future stage.
In this embodiment, step 200 specifically includes:
firstly, constructing a current-stage research city traffic map according to travel demand data of each traffic flow area in the current-stage research city; and secondly, determining an embedding matrix of the city research at the current stage according to the city traffic map research at the current stage and the Node2Vec map embedding algorithm.
The detailed process is as follows: defining a traffic flow network as a graph G (V, E, A), wherein V represents a node set and actually represents a traffic flow area obtained by dividing a research city; e represents an edge set, and actually represents travel requirements among traffic flow areas; a represents an adjacency matrix, which is actually a weight of the degree of connection between the traffic flow areas. Is provided withThe characteristic value of the node i at the time interval t is represented, and in practice, the flow value of the ith traffic flow area at the time interval t is represented; />The characteristic value of all nodes at time interval t is represented, and in practice, the flow value of the study city at time interval t is represented. H t ＝{X t -P+1 ,X t-P+2 ,…,X t }∈R N×P Representing the eigenvalues of all nodes in the past P time interval, the actualThe flow value of the study city in the past P time interval. Traffic flow prediction is to predict future traffic flow according to historical traffic data, and the concrete expression is shown in a formula (1), wherein F () can represent some deep neural network models, and Y represents model output results.
Y＝F(H t ) (1)。
A s ∈R N×N For the graph G, an embedding matrix is obtained according to the Node2Vec algorithm, and N represents the number of nodes in the Node set. Node2Vec is a graph embedding algorithm that aims to learn the mapping of nodes from high-dimensional space to low-dimensional space while maximizing the feature representation of neighboring nodes. Node2Vec walk is an offset random walk that combines depth-first sampling and breadth-first sampling strategies. Breadth-first sampling requires that the sampling node be a direct neighbor of the source node, whereas depth-first sampling is a continuous sampling method, gradually increasing the distance of the sampling node from the source node.
Table 1 Node2Vec diagram embedding algorithm
In this embodiment, the trained generator is a generator in a model for urban traffic flow prediction (TGAN) based on generation of an countermeasure network, the structure of which is shown in fig. 2.
The determination process based on the urban traffic flow prediction model for generating the countermeasure network is as follows:
(1) Constructing and generating an countermeasure network frame; the generating an countermeasure network framework includes a generator and a arbiter.
(2) Constructing sample data; the sample data comprises an embedding matrix of a first historical stage research city and a real traffic flow of a second historical stage research city; the second history phase is a future phase of the first history phase.
(3) And taking the embedded matrix of the first historical stage research city and random noise conforming to Gaussian distribution as input values, and inputting the input values into a generator in a generating countermeasure network frame to obtain the predicted traffic flow of the second historical stage research city.
(4) And taking the real traffic flow of the second historical stage research city and the predicted traffic flow of the second historical stage research city as input values, inputting the input values into a discriminator in the generated countermeasure network frame to obtain a discrimination result, adjusting the network parameters of the generator in the generated countermeasure network frame according to the discrimination result until the loss values of the generator and the discriminator meet the set requirements, and further obtaining the urban traffic flow prediction model based on the generated countermeasure network.
The generator consists of four convolution layers (Node 2 Vec-convolution layers) based on the Node2Vec graph embedding algorithm and one full connection layer, and the discriminator consists of two Node2 Vec-convolution layers and one full connection layer.
Further, the goal of the Node2 Vec-convolution layer is to learn a function that aims to extract spatio-temporal features. The representation of this function is as follows:
H l+1 ＝f(H l ,A S )＝ReLU(A S H l W l ) (2)。
A S ＝Node2Vec(G(V，E，A)) (3)。
wherein H is l Representing input data of layer I, A S Representing an embedding matrix obtained by a Node2Vec graph embedding algorithm, W i A matrix of leachable parameters representing layer l, reLU () representing the activation function, and H l+1 The output of the first layer is indicated.
The activation function of the Node2 Vec-convolutional layer in the generator is ReLU (), and layer-to-layer normalization is performed using batch Batch Normalization. The activation function of the Node2 Vec-convolution layer in the discriminator is LeakyReLU (), and Batch Normalization is adopted for layer normalization between layers.
Generating an countermeasure network framework is a generation model that utilizes generators and discriminators to learn the distribution of training data. In TGAN, the generator takes random noise as input, outputs traffic data, and aims to make the generated data very close to the real historical traffic data distribution. The discriminator takes as input the true traffic data and the traffic data generated by the generator, outputs a probability value between 0 and 1, indicating the likelihood that the input data is true. By optimizing the generator and the arbiter in the countermeasure mode, the TGAN can generate a result very close to the real historical traffic data distribution.
Specifically, according to the historical travel demand of each traffic flow area in the research city, an embedding matrix A is obtained by adopting a Node2Vec graph embedding algorithm s ∈R N×N N represents the number of traffic flow areas. Noise data conforming to Gaussian distribution is embedded in matrix A s As input of a generator, obtaining predicted traffic flow of the researched city; wherein the format of the Noise data is Noise E R B×N×1 B represents the batch training size. And taking the predicted traffic flow obtained by the generator and the real traffic flow in the training set as inputs of the discriminator. The output of the discriminator: the output is a probability value, which is the probability of judging whether the input data is the predicted traffic flow or the real traffic flow of the generator.
Table 2 training process based on urban traffic flow prediction model to generate countermeasure network
The loss functions of the generator and the arbiter are as shown in formulas (4) to (6):
V(D)＝E z～P(Z) [log(D(z))]+E M～P(G) [log(1-D(M))] (4)。
V(G)＝E M～P(G) [log(1-D(M))]+μMAELoss (5)。
wherein, the liquid crystal display device comprises a liquid crystal display device,representing the output of the generator, y i Representing the real traffic flow, μ is a constant that controls the proportion of MAELoss in the generator loss function.
The above technical scheme is described below through a specific experiment.
The present experiment, in combination with the TGAN model provided in this example, uses the data set of about 11 months in 2016 on a vehicle in the metropolitan area network to predict traffic flow. The capital study was conducted with areas of longitude 104.042E to 104.130E and latitude 30.652N to 30.728N, divided into 81 alternating current flow areas of 1km by 1km in total. The traffic flow distribution reflects the traffic conditions of the traffic flow area, quantified in this experiment by the inflowing and outflowing traffic flow. Since it is difficult to obtain the total flow of all traffic patterns within each traffic flow area or grid, the present experiment represents the flow value of the traffic flow area or grid as the inflow and outflow of taxis or net-bound vehicles, which has been previously proven to be effective. In this experiment, the day was divided into 144 10 minute intervals, the total flow of each traffic flow zone was represented by the number of vehicles arriving and leaving during each interval, and the resulting Data set was in the form of Data ε R 4320×81×1 . Since the present experiment aims at capturing the change situation of the traffic pattern in the history data, unlike the conventional traffic prediction method, the traffic flow values of several time intervals in the future are not predicted from the traffic flow values of several time intervals in the past, but are learned from all the history data. The experiment divided the dataset into training (70%) and testing (30%) sets, the training set being in the form of data_train ε R 3024×81×1 The test set is in the form of data_test εR 1296×81×1 。
The initial learning rate of the discriminators and generators was set to 0.001 for the adult network about vehicle dataset, and was reduced to 0.0001 after 10 rounds. The batch size was set to 156 and the total training run was 200. The number of neurons in the Node2 Vec-convolutional layer of the arbiter is set to 32 and the generator is set to 64. Adam was chosen as the optimizer and a constant μ controlling the ratio of MAELoss in the generator loss function was set to 0.75.
The experiment uses three evaluation indexes, namely Mean Absolute Error (MAE), root Mean Square Error (RMSE) and definition D 1 To measure the difference between the generated traffic flow distribution and the actual traffic flow distribution as shown in equations (7) to (9), respectively.
Where N represents the total number of samples, y i Representing the true value of the code,representing the predicted value output by the model. D (D) 1 Refers to the euclidean distance between the generated traffic flow profile and the actual traffic flow profile.
The experiment used the metropolitan network about vehicle dataset to compare TGAN with TrafficGAN, CGAN, WGAN and DCGAN and obtain performance comparisons of TGAN with various baseline methods, as shown in table 3.
Table 3 baseline model comparison (Chengdu City network about vehicle data set)
To better understand the predicted performance of TGAN, the true values and predicted results are visualized as shown in fig. 3. As can be seen from the generated heat map of traffic flow distribution and real traffic flow distribution, the fitting degree of the TGAN and the real value is highest, and the traffic flow area with dense traffic flow can be accurately captured. This illustrates that TGAN can effectively learn the distribution of raw data and generate accurate data from the learned distribution.
The invention provides a deep learning model TGAN, which combines a convolution neural network based on a Node2Vec graph embedding algorithm and a generation countermeasure network framework to solve the problem that most of existing prediction methods ignore the limitation of traffic mode change in captured historical data. Experimental results in the adult Du-net about vehicle dataset indicate that MAE, RMSE and D of TGAN 1 Are the minimum values in the proposed baseline model, TGAN vs. MAE, RMSE and D of CGAN 1 Reduced by 37.15%, 31.81% and 31.80%, respectively; MAE, RMSE and D compared to WGAN 1 Reduced by 73.51%, 57.46% and 57.45%, respectively; MAE, RMSE and D compared to DCGAN 1 Reduced by 41.53%, 36.51% and 36.49%, respectively; MAE, RMSE and D compared with TrafficGAN 1 Reduced by 41.60%, 36.17% and 36.17%, respectively.
DCGAN, WGAN and CGAN use conventional convolutional neural networks to extract spatiotemporal features from traffic data, while TrafficGAN uses graph convolution neural networks to extract spatiotemporal features. TGAN uses Node2Vec convolution layers to capture deep space-time dependencies. Based on the comparison of evaluation indexes and the visualization of thermodynamic diagrams between TGAN and DCGAN, WGAN, CGAN and TrafficGAN, a conclusion can be drawn that the TGAN introduces a Node2Vec diagram embedding algorithm to construct a traffic embedding diagram, so that the convolutional neural network can better process traffic data, and the change of a flow pattern in the capture historical data of the countering network frame is introduced, thereby improving the prediction precision and providing a valuable tool for traffic management and urban planning.
Example two
In order to perform a corresponding method of the above embodiment to achieve the corresponding functions and technical effects, an urban traffic flow prediction system based on generating an countermeasure network is provided below.
The urban traffic flow prediction system based on the generation countermeasure network provided in this embodiment includes:
and the travel demand data acquisition module is used for acquiring travel demand data of each traffic flow area in the research city at the current stage.
And the embedded matrix calculation module is used for determining an embedded matrix of the current stage research city by adopting a Node2Vec graph embedded algorithm according to travel demand data of each traffic flow area in the current stage research city.
And the traffic flow prediction module is used for taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into the trained generator, and predicting the traffic flow of the research city at the future stage.
Example III
An embodiment of the present invention provides an electronic device including a memory for storing a computer program and a processor that runs the computer program to cause the electronic device to execute the urban traffic flow prediction method based on generating an countermeasure network of the first embodiment.
Alternatively, the electronic device may be a server.
In addition, the embodiment of the invention also provides a computer readable storage medium storing a computer program, which when executed by a processor, implements the urban traffic flow prediction method based on the generation countermeasure network of the first embodiment.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.

Claims (9)

1. A method for urban traffic flow prediction based on generation of an countermeasure network, comprising:
acquiring travel demand data of each traffic flow area in a research city at the current stage;
determining an embedding matrix of the current stage research city by adopting a Node2Vec graph embedding algorithm according to travel demand data of each traffic flow area in the current stage research city;
and taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into a trained generator, and predicting the traffic flow of the research city at the future stage.
2. The urban traffic flow prediction method based on the generated countermeasure network according to claim 1, wherein the method for determining the embedded matrix of the current stage research city by adopting the Node2Vec graph embedded algorithm according to the travel demand data of each traffic flow region in the current stage research city comprises the following steps:
constructing a current-stage research city traffic map according to travel demand data of each traffic flow area in the current-stage research city;
and determining an embedding matrix of the city research at the current stage according to the city traffic map research at the current stage and the Node2Vec map embedding algorithm.
3. The urban traffic flow prediction method based on generation of an countermeasure network according to claim 1, wherein the trained generator is a generator in an urban traffic flow prediction model based on generation of an countermeasure network;
the determination process based on the urban traffic flow prediction model for generating the countermeasure network is as follows:
constructing and generating an countermeasure network frame; the generating an countermeasure network frame comprises a generator and a discriminator;
constructing sample data; the sample data comprises an embedding matrix of a first historical stage research city and a real traffic flow of a second historical stage research city; the second history phase is a future phase of the first history phase;
taking the embedded matrix of the first historical stage research city and random noise conforming to Gaussian distribution as input values, and inputting the input values into a generator in a generating countermeasure network frame to obtain predicted traffic flow of the second historical stage research city;
and taking the real traffic flow of the second historical stage research city and the predicted traffic flow of the second historical stage research city as input values, inputting the input values into a discriminator in the generated countermeasure network frame to obtain a discrimination result, and adjusting the network parameters of a generator in the generated countermeasure network frame according to the discrimination result until the loss values of the generator and the discriminator meet the set requirements to obtain the city traffic flow prediction model based on the generated countermeasure network.
4. A method of urban traffic flow prediction based on generation of an countermeasure network according to claim 3, wherein the generator consists of four Node2 Vec-convolution layers and a fully connected layer; the Node2 Vec-convolution layer is a convolution layer based on the Node2Vec graph embedding algorithm.
5. The urban traffic flow prediction method according to claim 4, wherein the Node2 Vec-convolutional layer activation function in the generator is ReLU (), and layer normalization is performed between layers using a batch Batch Normalization.
6. The urban traffic flow prediction method according to claim 4, wherein the discriminator consists of two Node2 Vec-convolution layers and a fully-connected layer.
7. The urban traffic flow prediction method according to claim 6, wherein the activation function of the Node2 Vec-convolution layer in the arbiter is LeakyReLU (), and layer-to-layer normalization is performed by Batch Normalization.
8. An urban traffic flow prediction system based on a generation countermeasure network, comprising:
the travel demand data acquisition module is used for acquiring travel demand data of each traffic flow area in the research city at the current stage;
the embedded matrix calculation module is used for determining an embedded matrix of the current stage research city by adopting a Node2Vec graph embedded algorithm according to travel demand data of each traffic flow area in the current stage research city;
and the traffic flow prediction module is used for taking the embedded matrix of the research city at the current stage and random noise conforming to Gaussian distribution as input values, inputting the input values into the trained generator, and predicting the traffic flow of the research city at the future stage.
9. An electronic device comprising a memory for storing a computer program and a processor that runs the computer program to cause the electronic device to perform a method of urban traffic flow prediction based on generating an countermeasure network according to any one of claims 1 to 7.
CN202310681033.6A 2023-06-09 2023-06-09 Urban traffic flow prediction method, system and equipment based on generation countermeasure network Pending CN116978218A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310681033.6A CN116978218A (en) 2023-06-09 2023-06-09 Urban traffic flow prediction method, system and equipment based on generation countermeasure network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310681033.6A CN116978218A (en) 2023-06-09 2023-06-09 Urban traffic flow prediction method, system and equipment based on generation countermeasure network

Publications (1)

Publication Number Publication Date
CN116978218A true CN116978218A (en) 2023-10-31

Family

ID=88470172

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310681033.6A Pending CN116978218A (en) 2023-06-09 2023-06-09 Urban traffic flow prediction method, system and equipment based on generation countermeasure network

Country Status (1)

Country Link
CN (1) CN116978218A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117576918A (en) * 2024-01-17 2024-02-20 四川国蓝中天环境科技集团有限公司 Urban road flow universe prediction method based on multi-source data

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117576918A (en) * 2024-01-17 2024-02-20 四川国蓝中天环境科技集团有限公司 Urban road flow universe prediction method based on multi-source data
CN117576918B (en) * 2024-01-17 2024-04-02 四川国蓝中天环境科技集团有限公司 Urban road flow universe prediction method based on multi-source data

Similar Documents

Publication Publication Date Title
Luo et al. An algorithm for traffic flow prediction based on improved SARIMA and GA
Yin et al. Multi-stage attention spatial-temporal graph networks for traffic prediction
Xu et al. Deep belief network-based support vector regression method for traffic flow forecasting
Xu et al. Accurate and interpretable bayesian mars for traffic flow prediction
Yan et al. Spatial-temporal chebyshev graph neural network for traffic flow prediction in iot-based its
Ma et al. Short-term traffic flow forecasting by selecting appropriate predictions based on pattern matching
CN108985965A (en) A kind of photovoltaic power interval prediction method of combination neural network and parameter Estimation
Jin et al. A GAN-based short-term link traffic prediction approach for urban road networks under a parallel learning framework
CN112396218B (en) Crowd flow prediction method based on urban area multi-mode fusion
CN115376317B (en) Traffic flow prediction method based on dynamic graph convolution and time sequence convolution network
Shao et al. The Traffic Flow Prediction Method Using the Incremental Learning‐Based CNN‐LTSM Model: The Solution of Mobile Application
CN113159403B (en) Intersection pedestrian track prediction method and device
CN116978218A (en) Urban traffic flow prediction method, system and equipment based on generation countermeasure network
Chen et al. A review on traffic prediction methods for intelligent transportation system in smart cities
CN117977568A (en) Power load prediction method based on nested LSTM and quantile calculation
CN110210656A (en) Shared bicycle method for predicting and system based on website behavioural analysis
Zhao et al. Short-term traffic flow intensity prediction based on CHS-LSTM
Shen et al. An attention-based digraph convolution network enabled framework for congestion recognition in three-dimensional road networks
CN116311921A (en) Traffic speed prediction method based on multi-spatial scale space-time converter
CN108073978A (en) A kind of constructive method of the ultra-deep learning model of artificial intelligence
Xiao et al. Parking prediction in smart cities: A survey
Tian et al. RETRACTED ARTICLE: Research on emergency vehicle routing planning based on short-term traffic flow prediction
He et al. A hybrid deep learning model for link dynamic vehicle count forecasting with Bayesian optimization
CN117116045A (en) Traffic flow prediction method and device based on space-time sequence deep learning
Pan et al. Traffic speed prediction based on time classification in combination with spatial graph convolutional network

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