CN113256980A - Road network state determination method, device, equipment and storage medium - Google Patents
Road network state determination method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113256980A CN113256980A CN202110592888.2A CN202110592888A CN113256980A CN 113256980 A CN113256980 A CN 113256980A CN 202110592888 A CN202110592888 A CN 202110592888A CN 113256980 A CN113256980 A CN 113256980A
- Authority
- CN
- China
- Prior art keywords
- road
- determining
- traffic
- intersection
- network
- 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
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
-
- 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
- G08G1/0129—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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for determining a road network state, wherein the method comprises the following steps: determining an incidence relation between intersections and road sections in a road network; constructing a graph neural network model of a road network based on the incidence relation; training vertexes representing the road sections in the graph neural network model to obtain characteristic vectors of the vertexes; under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector; and determining intersections and road sections in the road network, wherein the intersections and the road sections are associated with each other in traffic modes and traffic states according to the similarity. According to the method, the characteristic vectors of the road sections are extracted from the graph neural network model, the attribute information of the road sections is combined with the characteristic vectors, the similarity among the road sections in the road network is calculated, the state of the road network is evaluated according to the similarity, the state of the road network is evaluated from two dimensions of space and time, the attribute information of the road sections is further integrated into the characteristic evaluation of the whole road network, and the state of the road network can be analyzed more comprehensively.
Description
Technical Field
The embodiment of the invention relates to a road network traffic analysis technology, in particular to a road network state determination method, a road network state determination device, road network state determination equipment and a storage medium.
Background
At present, in a road system (road network) of a city, information of traffic elements such as people, vehicles, roads and the like can be acquired in real time, the source of urban traffic big data is increasingly rich, the analysis of the state of a large-scale urban road network by utilizing various traffic elements is beneficial to effective management, control and operation of smart traffic, and the evaluation of the existing traffic control strategy and the deduction evaluation of a future traffic control strategy are beneficial.
The existing method for analyzing the state of the urban road network mainly starts from the point of data statistics, namely, time factors influencing the state of the current road network are analyzed by using historical traffic patterns and traffic state data, the traffic patterns and the states of the current road section are analyzed by simply using geographical position information (for example, the traffic patterns and the states of adjacent road sections are analyzed), and then the overall state of the urban road network is analyzed by using the time factors and the geographical information. However, the above method can only consider the dynamic change characteristics of the traffic states of the road segments with association in the geographic location, and can only analyze the traffic mode change of each road segment in the road network from the dimension of time change, and lacks of considering the similarity of the traffic states and the traffic modes that may exist between two road segments without geographic location association in space, and lacks of analyzing the influence factors of the traffic road conditions between a plurality of road segments with space location isolation on the traffic states and the traffic modes of the whole road network, so that the existing analysis method for the road network states is not comprehensive enough, poor in expandability, low in reliability, and unable to effectively provide a constructive scheme for the management and control of intelligent traffic.
Disclosure of Invention
The embodiment of the invention provides a method, a device, equipment and a storage medium for determining a road network state, and aims to solve the problems that the analysis of the road network state is not comprehensive and has low reliability in the existing method.
In a first aspect, an embodiment of the present invention provides a method for determining a road network state, where the method includes:
determining an incidence relation between intersections and road sections in a road network;
constructing a graph neural network model of the road network based on the incidence relation;
training a vertex representing the road section in the graph neural network model to obtain a feature vector of the vertex;
under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector;
and determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
In a second aspect, an embodiment of the present invention further provides a device for determining a road network state, where the device includes:
the incidence relation determining module is used for determining the incidence relation between intersections and road sections in the road network;
the graph neural network model building module is used for building a graph neural network model of the road network based on the incidence relation;
the characteristic vector determining module is used for training a vertex representing the road section in the graph neural network model to obtain a characteristic vector of the vertex;
the similarity calculation module is used for calculating the similarity between the vertexes based on the feature vector under the constraint of the attribute information of the road section;
and the state determining module is used for determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
In a third aspect, an embodiment of the present invention further provides a computer device, where the computer device includes:
one or more processors;
a memory for storing one or more programs;
when the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining a road network state according to the first aspect.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for determining a road network state according to the first aspect is implemented.
The method comprises the steps of determining the incidence relation between intersections and road sections in a road network; constructing a graph neural network model of a road network based on the incidence relation; training vertexes representing the road sections in the graph neural network model to obtain characteristic vectors of the vertexes; under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector; and determining intersections and road sections in the road network, wherein the intersections and the road sections are associated with each other in traffic modes and traffic states according to the similarity. The method constructs a graph neural network model of the road network according to the association relationship, trains vertexes of the expressed road segments in the graph neural network model, extracts characteristic vectors of the road segments, acquires attribute information of the road segments, combines the attribute information and the characteristic vectors to calculate the similarity between the road segments in the road network, evaluates the state of the road network according to the similarity, evaluates the traffic state of intersections and road segments in the road network from two dimensions of space and time, integrates the attribute information of the road segments into the characteristic evaluation of the whole road network, finds intersections and road segments with mutual association of traffic modes and traffic states according to the similarity, can analyze the traffic state and the traffic mode of the road network more comprehensively, enables the reliability of an analysis result to be higher, enables the attribute information of the road segments to be changed along with actual road conditions, and has strong data expandability, the method is more beneficial to providing effective suggestions for the management and control strategy of the intelligent traffic.
Drawings
Fig. 1 is a flowchart of a road network state determining method according to an embodiment of the present invention;
fig. 2 is a flowchart of a road network state determining method according to a second embodiment of the present invention;
fig. 3 is a schematic view of an entering and exiting direction between a road section and an intersection according to a second embodiment of the present invention;
fig. 4 is a schematic view of a construction process of a neural network model according to a second embodiment of the present invention;
fig. 5 is a schematic structural diagram of a road network state determining device according to a third embodiment of the present invention.
Fig. 6 is a schematic structural diagram of a computer device according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
It should be noted that: in the description of the embodiments of the present invention, the terms "first," "second," "third," and the like are used solely to distinguish one from another and are not intended to indicate or imply relative importance.
Example one
Fig. 1 is a flowchart of a method for determining a road network state according to an embodiment of the present invention, where the method is applicable to analyzing traffic states and traffic patterns of road junctions and road segments in a road network and determining a traffic state of the entire road network, and the method may be executed by a road network state determining device, where the road network state determining device may be implemented by software and/or hardware and may be configured in a computer device, such as a server, a workstation, a personal computer, and the like, and the method specifically includes the following steps:
and S110, determining the association relation between the intersections and the road sections in the road network.
The road network refers to a road system formed by interconnecting and interlacing various roads in a certain area, and is distributed in a net shape. The road network in the present embodiment may be a road network or other road systems such as an urban road network, and is not limited thereto.
In this embodiment, the road network includes intersections and road segments, where the road segments are roads divided by a fixed length in the road system, and the intersections are road intersections intersected by two or more roads; in a road network, one intersection may intersect with multiple road segments, that is, the number of road segments merging into a single intersection and the number of road segments flowing out from a single intersection can be determined according to the flow direction of vehicles running in the road segments, and the association relationship between intersections and road segments referred to in this embodiment may include the adjacency relationship between intersections and road segments, the connection manner between adjacent intersections, and the like; the expression form of the association relationship may also be various, for example, the association relationship is expressed by an adjacency matrix, the association relationship is expressed by an association matrix, the association relationship is expressed by a tree diagram, the association relationship is expressed by a topological diagram, and the like; this embodiment is not limited to this.
And S120, constructing a graph neural network model of the road network based on the incidence relation.
In this embodiment, the graph neural network model is a model in which graph data and a neural network are combined and end-to-end calculation is performed on the graph data. The graph data is composed of vertices and edges connecting the vertices, and may be represented as G ═ V, E, where V denotes a vertex in the graph, and E denotes an edge connecting the vertices and the vertices, and the purpose of the graph neural network model is to convert the entire graph G, each vertex V, and each edge E into a dense vector. More specifically, graph (G) is defined as (V, E), and is denoted as G ═ V, E. Wherein: v is a non-empty finite set of vertices (Vertex), denoted as V (G); e is a subset of the unordered set V & V, denoted as E (G). A graph whose vertices are collected as empty is referred to as an empty graph. There are many basic types of graphs, such as a directed graph and an undirected graph (whether an edge has a direction), an unweighted graph and a weighted graph (whether an edge has a weight), a connected graph and an unconnected graph (whether an isolated vertex exists), and a bipartite graph (any edge belongs to 2 sub-graphs thereof), and this embodiment is not particularly limited.
After determining the association relationship between intersections and road segments in the road network, the road segments can be used as vertexes, the association relationship is converted into edges connecting the vertexes, graph data representing the road network is constructed, and the graph data is combined with a neural network to construct a graph neural network model.
And S130, training the vertexes representing the road sections in the graph neural network model to obtain the feature vectors of the vertexes.
It can be understood that the graph neural network model functions like a feature extractor, except that its object is graph data, and the graph neural network model subtly designs a method for extracting features from the graph data, so that the features can be used for node classification (node classification), graph classification (graph classification), edge prediction (link prediction) and graph embedding (graph embedding) of the graph data, and the graph neural network model has wide application.
The specific calculation process of the graph neural network model can be generally understood as aggregating neighbors, that is, each vertex receives information of neighbor vertices, and in order to more comprehensively depict each vertex, besides attribute information of the vertex itself, more comprehensive structural information is needed, so that neighbor vertices in the neighborhood range of the current vertex and even neighbors including the neighbors need to be aggregated.
In this embodiment, the training of the vertices representing the road segments in the graph neural network model to obtain the feature vectors of the vertices may specifically include sampling neighbor vertices for each vertex representing the road segment, aggregating attribute information of the neighbor vertices to generate the feature vectors of the vertices, or may classify the vertices representing the road segments, input feature matrices represented by vertex sets belonging to different categories into the graph neural network model to train, select an activation function (e.g., a ReLU function) to activate the feature matrices in the training process, and finally obtain the feature vectors of the vertices. The embodiment does not limit the specific operation of training the vertices representing the road segments in the graph neural network model.
And S140, under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector.
In this embodiment, the attribute information of the link refers to information that can reflect the traffic conditions of the link, and may include: the congestion time of the road section, the saturated traffic flow of the road section, the traffic density of the road section, the maximum driving speed of vehicles of the road section, the vehicle queuing time of the road section, and the like; the attribute information of the road section can be used for providing a reference value for the feature vector, or the attribute information of the road section is fused into the feature vector to obtain an adjusted feature vector, and the similarity between vertexes is calculated by the adjusted feature vector through a cosine similarity calculation formula or an Euclidean distance calculation formula. The embodiment does not limit the specific similarity calculation method.
And S150, determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
In this embodiment, because the calculation of the similarity integrates the attribute information of the links and the features in the feature vectors representing the vertices of the links, the similarity can characterize the links in the road network having association relationship between traffic patterns and traffic states, and the connection relationship between the links is closely related to intersections, so that intersections and links in the road network having the traffic patterns and traffic states associated with each other can be determined according to the similarity, specifically, the similarity of the vertices representing the links can be used to assemble a state similarity matrix, the traffic pattern and traffic state of the whole road network can be characterized by the state similarity matrix, the road patterns and traffic states of each element (intersection and link) in the road network can be analyzed, the traffic condition of each element can be examined and monitored, and the method can be applied to traffic pattern change and cause analysis, traffic network state evaluation, traffic pattern analysis, traffic state evaluation, traffic state analysis, and traffic state analysis of the road network, Traffic emergency management and control, traffic management and control strategy optimization, active traffic management, control, operation and other application scenarios.
The embodiment of the invention determines the incidence relation between the intersections and the road sections in the road network; constructing a graph neural network model of a road network based on the incidence relation; training vertexes representing the road sections in the graph neural network model to obtain characteristic vectors of the vertexes; under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector; and determining intersections and road sections in the road network, wherein the intersections and the road sections are associated with each other in traffic modes and traffic states according to the similarity. The method constructs a graph neural network model of the road network according to the association relationship, trains vertexes of the expressed road segments in the graph neural network model, extracts characteristic vectors of the road segments, acquires attribute information of the road segments, combines the attribute information and the characteristic vectors to calculate the similarity between the road segments in the road network, evaluates the state of the road network according to the similarity, evaluates the traffic state of intersections and road segments in the road network from two dimensions of space and time, integrates the attribute information of the road segments into the characteristic evaluation of the whole road network, finds intersections and road segments with mutual association of traffic modes and traffic states according to the similarity, can analyze the traffic state and the traffic mode of the road network more comprehensively, enables the reliability of an analysis result to be higher, enables the attribute information of the road segments to be changed along with actual road conditions, and has strong data expandability, the method is more beneficial to providing effective suggestions for the management and control strategy of the intelligent traffic.
Example two
Fig. 2 is a flowchart of a road network state determination method provided in a second embodiment of the present invention, which is further detailed based on the foregoing embodiments, and the method may be implemented by a road network state determination device, where the road network state determination device may be implemented by software and/or hardware, and may be configured in a computer device, such as a server, a workstation, a personal computer, and the like, and the method specifically includes the following steps:
s210, acquiring geographic information of road sections in a road network.
In the present embodiment, the geographic information of the road segment includes geographic coordinates of the road segment on the road network map, the length of the road segment, the number of lanes of the road segment, the traffic direction marked in the road segment, and the like. Geographic information of road sections in a road network can be acquired from a live-action map; the geographic information of the road section can also be provided through data acquired by infrastructure with a traffic control function in the road network, for example, video data acquired by a traffic camera in the road network is acquired, and the geographic information of the road section is extracted from the video data; the geographic information of the road segments in the road network can be obtained through a remote sensing map, and the embodiment is not limited.
And S220, determining a road intersection where the road sections intersect as an intersection based on the geographic information.
And S230, constructing an adjacency matrix by using the adjacency relation between the intersections and the road sections as the incidence relation between the intersections and the road sections.
In a specific implementation manner, S230 may include the following specific steps:
s2301, determining the entrance and exit directions between the road sections in the neighborhood range of each intersection and the intersections.
The entering and exiting direction refers to the passing direction of a road section with intersection with the intersection.
For example, as shown in fig. 3, a road segment 301, a road segment 302, a road segment 303, and a road segment 304 exist in the neighborhood of the intersection 300, for the intersection 300, the preset passing direction in the road segment 301 and the road segment 304 is the direction of entering the intersection 300, the preset passing direction in the road segment 302 and the road segment 303 is the direction of exiting from the intersection 300, that is, the entering and exiting direction between the road segment 301 and the intersection 300 is the input direction of the road segment 301 pointing to the intersection 300, the entering and exiting direction between the road segment 302 and the intersection 300 is the output direction of the intersection 300 pointing to the road segment 302, the entering and exiting direction between the road segment 303 and the intersection 300 is the output direction of the intersection 300 pointing to the road segment 303, and the entering and exiting direction between the road segment 304 and the intersection 300 is the input direction of the road segment 304 pointing to the intersection 300.
S2302, determining a road section of the intersection into which the traffic flow is input as an incident road section and a road section of the intersection from which the traffic flow is output as an exit road section according to the entering and exiting directions.
As can be seen from fig. 3, the preset passing direction in the road segments 301, 302, 303, and 304 is the passing direction of the traffic flow in each road segment, in this embodiment, the road segment where the traffic flow is input to the intersection can be determined as the incident road segment and the road segment where the traffic flow is output from the intersection can be determined as the exit road segment according to the entering and exiting directions between each road segment and the intersection, for example, for the intersection 300, the road segments 301 and 304 are the incident road segment of the intersection 300, and the road segments 302 and 303 are the exit road segment of the intersection 300.
S2303, constructing an adjacent matrix of the intersection by the incident road section and the emergent road section, and taking the adjacent matrix as the incidence relation between the intersection and the road section.
In a specific implementation, a directed graph can be drawn according to each intersection in a road network and an incident road section and an emergent road section of the intersection, the nodes of the directed graph are each intersection, the edges of the directed graph are each road section, the direction of the edges is determined by a judgment rule that the road sections are the incident road section or the emergent road section, namely the edges of the directed graph are directed edges, and the nodes and the edges of the directed graph are stored in a matrix according to the drawn directed graph to obtain an adjacent matrix; the number of the ith row of non-zero elements in the adjacency matrix is the out degree of the ith node, the number of the ith column of non-zero elements is the in degree of the ith node, the degree of the ith node is the sum of the numbers of the ith row and the ith column of non-zero elements, wherein the degree of the node represents the number of edges associated with the node, the out degree represents the number of edges which point to other nodes from the node, and the in degree represents the number of edges which converge to the node from other nodes. The present embodiment can be expressed as an adjacency matrix as a relationship between intersections and links in the road network.
And S240, constructing a graph neural network model of the road network based on the incidence relation.
In a specific implementation manner of this embodiment, the association relationship may be expressed as an adjacency matrix, and the road segments in the road network are classified into a plurality of road segment subsets according to the adjacency matrix, where each road segment subset includes all road segments belonging to the same intersection; determining a sub-graph region composed of each road section subset; and connecting the plurality of sub-graph areas together based on the intersection of the road section subsets to form a graph neural network model of the road network.
The step of determining the sub-graph region formed by each road segment subset specifically includes: and in each road section subset, constructing a sub-graph region of the road section subset by taking each road section as a vertex and taking a connection relation between any two vertexes as an edge.
In another specific implementation manner of this embodiment, referring to the flowchart shown in fig. 4, connection relationships between multiple intersections in a road network may be determined, an adjacency matrix of each intersection is determined, the adjacency matrices are aggregated into an association matrix based on the connection relationships, a topological graph composed of the intersections and road segments in the road network is obtained, each road segment is used as a vertex, edges connecting the vertices are determined according to the association matrix and the topological graph, and a graph neural network model of the road network is constructed by using all the vertices and the edges connecting the vertices. It should be noted that the present embodiment does not limit the specific manner of constructing the neural network model.
After the Graph neural network model of the road network is constructed, feature extraction can be performed on Graph data (vertexes and edges) in the Graph neural network model by using a Graph embedding algorithm to obtain features for performing high-dimensional vectorization on the vertexes and features for performing high-dimensional vectorization on the opposite sides, for example, the features of the vertexes can be learned and trained by using a Graph Sample and aggregation model (Graph sampling and clustering) commonly used in the Graph embedding algorithm to further calculate node embedding vectors under different features, or the Graph neural network model can be solved by using a deep walk algorithm, deep walk is the first algorithm for learning node embedding in an unsupervised manner, and is very similar to word embedding in terms of a training process. It should be noted that, since GraphSAGE is a framework of inductive learning capable of efficiently generating unknown vertex embedding (embedding) by using the attribute information of a vertex, the model can identify the structural attributes of the neighborhood of a vertex, and express the local role and the position of the vertex in the whole graph. Therefore, preferably, the embodiment may use the GraphSAGE model to solve the graph neural network model.
And S250, in the graph neural network model, aiming at each vertex of the representation road section, sampling a preset number of vertexes in a neighborhood range of the vertexes as neighbor vertexes.
For reasons of computational efficiency, a certain number of neighbor vertices may be sampled for each vertex as vertices for information to be aggregated. For example, if the number of vertex neighbors is less than k, the sampling method with put back is adopted until k vertices are sampled. If the vertex neighbor number is greater than k, then samples without a put back are used. If the calculation efficiency is not considered, information aggregation can be completely carried out on each vertex by utilizing all the neighbor vertices, and therefore information is lossless.
And S260, determining an aggregation function.
In order to achieve information aggregation on neighbor vertices, an aggregation function needs to be determined in advance.
Since the neighbors of the vertices in the graph are naturally disordered, in order to make the constructed aggregation function symmetric (i.e. change the order of the inputs, the output result of the function is not changed), and have high expression capability, the aggregation function in this embodiment may include any one of a Mean aggregation function (Mean aggregator), a Pooling aggregation function (Pooling aggregator), and a time-loop aggregation function (LSTM aggregator).
S270, aiming at each vertex, performing information aggregation on the neighbor vertex of the vertex according to the aggregation function to obtain the feature vector of the vertex.
Specifically, the vertices in the graph neural network may be traversed in a breadth-first traversal (broad first traversal) or a depth-first traversal (deep first traversal), and for each vertex, information aggregation is performed on neighboring vertices of the vertex according to an aggregation function to obtain a feature vector of the vertex.
In this embodiment, the feature vector of each vertex may be obtained by calculating a weighted sum of features of all neighboring vertices of the current vertex, or may be obtained by calculating a weighted sum of the features of the current vertex and features of all other neighboring vertices; this embodiment is not limited thereto.
As an example, in the graph neural network model, the number of network layers K of the model is determined, where K also represents the number of neighbor vertices that can be aggregated for each vertex representing a link, for example, when K is 2, it indicates that each vertex can learn its own embedded vector expression according to the information of at most 2 neighbor vertices of the vertex, in the cycle K (K is 1 … K) of each layer in the graph neural network model, for each vertex, the aggregated information feature of the K-th layer of the neighbor vertex is first generated using the embedded vector of the K-1 layer of the neighbor vertex of the vertex, and then the aggregated information feature is spliced with the embedded vector of the vertex at the K-1 layer, and the embedded vector of the K-th layer of the vertex is generated through a nonlinear change, and the embedded vector can be used as the feature vector of the vertex in this embodiment.
And S280, calculating the similarity between the vertexes based on the feature vectors under the constraint of the attribute information of the road section.
In the embodiment, the traffic road condition of the road section changing in space at any time in the road network can be determined, and the attribute information of the road section is obtained; the feature vector is adjusted using the attribute information, and the similarity between the vertices is calculated based on the adjusted feature vector, and the adjustment may be performed by fusing the attribute information and the feature vector to obtain a new feature vector, and calculating the similarity between the vertices using a calculation formula for similarity based on the new feature vector, or may be performed by configuring different weights for elements in the feature vector using the attribute information as a reference of the feature vector, and obtaining the adjusted feature vector, and calculating the similarity between the vertices.
In one example of this embodiment, the road network is constructed as a graph neural network model, where the road segments are vertices (nodes) and the connection relations between the road segments are edges (edges). The multi-source heterogeneous traffic state time series data of the road section can be used as a vertex feature vector, for example, one expression form of the vertex feature vector is as follows:
wherein, γnFor the length range class of the road segment,
indicates the number of intermediate lanes of the road segment,
indicating the number of left-side lanes of the road segment,
indicating the number of lanes on the right side of the road segment,
and
respectively correspond to one dayAverage vehicle speed, minimum vehicle speed, and maximum vehicle speed for the inner road segment.
Obtaining the adjusted feature vector by using the attribute information of the road segment as the constraint of the feature vector, wherein the feature vector may further include: the strength of association of adjacent road segments, the direction of association of adjacent road segments, the range of lengths of adjacent road segments, the direction of each road segment of adjacent road segments (left turn, straight run, right turn), the number of lanes on a road segment, the average vehicle speed per day for each road segment of adjacent road segments, the minimum vehicle speed for a road segment, and the maximum vehicle speed for a road segment, and so forth.
In a specific implementation, determining a traffic road condition of a road segment changing in a road network at any time and in space to obtain attribute information of the road segment may include the following specific steps:
acquiring historical traffic state data of a road network;
determining a traffic flow model representing traffic flow in a road network based on the traffic status data;
calculating the association time length, the association strength, the association direction and the association transitivity of the road section in the traffic flow model to obtain the attribute information of the road section;
the link association time length represents the influence time of the adjacent link on the traffic state of the link, the link association strength represents the ratio of the traffic flow of the link to the total traffic flow of the adjacent link, the link association direction represents the direction of the traffic flow of the link, and the link association transmissibility represents the transmission direction of the traffic wave of the link.
In this embodiment, the association time length of the link, the association strength of the link, the association direction of the link, and the association transitivity of the link may be respectively used as upper-level analysis of the traffic network mode and state of the road network, that is, as the attribute information of the link, constraints are provided for the similarity calculation of the vertices.
The characteristic vector of the road section is calculated by using the graph neural network model, the related space-time data required by road network state analysis can be dynamically selected according to the result of the correlation time length and the correlation direction of the road section, the similarity of different road sections can be finely adjusted by using the correlation strength of the road section, and the correlation transitivity of the road section defines the indirectly connected intersection series required to be considered for constructing the graph neural network model.
In a specific implementation, the association time length of a road segment determines the influence time of an adjacent road segment, and can represent that the traffic state of a certain target road segment is influenced by the adjacent road segment, and the influence is related to the spatial average speed of the road segment and the adjacent road segment. The correlation time length can be calculated by the following formula:
wherein tau, L, VsThe time length of the link, the length of the link and the spatial average speed of the link are respectively associated, the spatial average speed is an average value of all vehicle speed distribution in the link with a certain length on the road, and the value of the spatial average speed is a harmonic average value of the average speed of each vehicle in the link length.
In this embodiment, the correlation strength of the adjacent road sections is related to the traffic flow thereof, and the greater the traffic flow of the road section, the greater the correlation strength, and the correlation strength can be calculated by the following formula:
wherein sigmaiIndex of strength of association, Q, representing a link iiRepresenting the traffic flow of a link I, I representing a set of adjacent links to the link I, QjIs the traffic flow for the road segment in the set of adjacent road segments that is adjacent to road segment i.
According to the traffic flow theory, the target link is influenced by the transmission of traffic waves due to the change in the density of the traffic flow (the number of vehicles in the link) both upstream and downstream of the target link. The upstream of the target road section refers to an adjacent road section of which the traffic direction points to the target road section, the downstream of the target road section refers to an adjacent road section of which the traffic direction is sent out from the target road section, the traffic wave transmission direction comprises transmission to the upstream of the road section and transmission to the downstream of the road section, the transmission speed can be used for representing the associated transmissibility score of the road section, and the calculation can be carried out by the following formula:
wherein u iswFor the speed of transmission of traffic waves, qdAnd q isuRespectively the downstream road section flow of the target road section and the upstream road section flow of the target road section, kdAnd kuRespectively, a downstream vehicle density of the target road segment and an upstream vehicle density of the target road segment.
When analyzing the relation between the target road section and the direction of the downstream road section, if uwIf the value is more than 0, the traffic wave direction moves from the target road section to the downstream, the influence of the traffic state of the downstream on the target road section does not need to be considered, and otherwise if u iswIf the traffic wave direction is less than 0, the traffic wave direction moves from the downstream of the target road section to the target road section, and the traffic state of the downstream needs to be considered in the traffic state analysis of the target road section. Although there is no direct correlation between non-adjacent road segments, the traffic state can be transferred to non-adjacent road segments by using adjacent road segments as a medium.
In one example, using the feature vectors to compute the similarity of vertices may be obtained by the following formula:
wherein h iscAnd hc′All representing feature vectors of different vertices, fsim(hc,hc′) Representing the similarity between two vertices.
And S290, determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
According to the similarity between vertexes obtained by calculation in the graph neural network model, the connection relation between road segments can be determined, the traffic mode and the state similarity matrix of the road network can be automatically established for the traffic state of the whole road network according to the similarity of the vertexes representing the road segments, the traffic mode and the state similarity matrix are used for representing the correlation information influencing the road network, the highly nonlinear, random and dynamic coupling relation among the road segments in the road network is analyzed, and the effective suggestion for the intelligent traffic control strategy is provided.
EXAMPLE III
Fig. 5 is a schematic structural diagram of a road network state determining device provided in the third embodiment of the present invention, where the device may specifically include the following modules:
an association relation determining module 501, configured to determine an association relation between an intersection and a road segment in a road network;
a graph neural network model building module 502, configured to build a graph neural network model of the road network based on the incidence relation;
a feature vector determining module 503, configured to train a vertex representing the road segment in the graph neural network model to obtain a feature vector of the vertex;
a similarity calculation module 504, configured to calculate, based on the feature vectors, similarities between vertices under the constraint of the attribute information of the road segment;
and a state determining module 505, configured to determine intersections and road segments in the road network, where the traffic modes and the traffic states are related to each other, according to the similarity.
In an embodiment of the present invention, the association relation determining module 501 includes:
the geographic information acquisition submodule is used for acquiring geographic information of road sections in a road network;
the intersection determining submodule is used for determining a road intersection where the road sections intersect as an intersection based on the geographic information;
and the adjacency relation determining submodule is used for constructing an adjacency matrix by utilizing the adjacency relation between the intersection and the road section as the incidence relation between the intersection and the road section.
In one embodiment of the present invention, the adjacency relation determination sub-module includes:
the access direction determining unit is used for determining the access direction between the road section in the neighborhood range of each intersection and the intersection;
a road section determining unit for determining a road section in which traffic flow is input to the intersection as an incident road section and a road section in which traffic flow is output from the intersection as an exit road section according to the access direction;
and the incidence road section and the emergence road section are used for constructing an adjacent matrix of the intersection as an incidence relation between the intersection and the road section.
In one embodiment of the invention, the association is represented as an adjacency matrix; the graph neural network model building module 502 includes:
a road segment classification submodule, configured to classify the road segments in the road network into a plurality of road segment subsets according to the adjacency matrix, where each road segment subset includes all the road segments belonging to the same intersection;
a sub-image region determining sub-module for determining a sub-image region composed of each road segment subset;
and the graph neural network model building submodule is used for connecting a plurality of sub-graph regions together based on the intersection of all the road section subsets to form the graph neural network model of the road network.
In one embodiment of the present invention, the sub-picture region determination sub-module includes:
and the sub-image region determining unit is used for constructing the sub-image regions of the road segment subsets by taking each road segment as a vertex and taking a connection relation between any two vertexes as an edge in each road segment subset.
In one embodiment of the present invention, the feature vector determination module 503 includes:
the neighbor vertex sampling sub-module is used for sampling a preset number of vertexes in a neighborhood range of the vertexes as neighbor vertexes aiming at each vertex representing the road section in the graph neural network model;
the aggregation function determining submodule is used for determining an aggregation function;
and the feature vector calculation submodule is used for performing information aggregation on the neighbor vertex of the vertex according to the aggregation function aiming at each vertex to obtain the feature vector of the vertex.
In one embodiment of the present invention, the similarity calculation module 504 includes:
the attribute information determining submodule is used for determining the traffic road condition of the road section in the road network along with the time-space change to obtain the attribute information of the road section;
and the similarity calculation submodule is used for adjusting the feature vectors by using the attribute information and calculating the similarity between the vertexes based on the adjusted feature vectors.
In an embodiment of the present invention, the attribute information determination sub-module includes:
a traffic state data acquisition unit, configured to acquire historical traffic state data of the road network;
a traffic flow model determination unit for determining a traffic flow model representing a traffic flow in the road network based on the traffic state data;
the attribute information determining unit is used for calculating the association time length, the association strength, the association direction and the association transitivity of the road section in the traffic flow model to obtain the attribute information of the road section;
wherein the associated time length of the link represents the influence time of the adjacent link on the traffic state of the link, the associated strength of the link represents the ratio of the traffic flow of the link to the total traffic flow of the adjacent link, the associated direction of the link represents the direction of the traffic flow of the link, and the associated transmissibility of the link represents the transmission direction of the traffic wave of the link.
The road network state determination device provided by the embodiment of the invention can execute the road network state determination method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example four
Fig. 6 is a schematic structural diagram of a computer device according to a fourth embodiment of the present invention. FIG. 6 illustrates a block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention. The computer device 12 shown in FIG. 6 is only an example and should not bring any limitations to the functionality or scope of use of embodiments of the present invention.
As shown in FIG. 6, computer device 12 is in the form of a general purpose computing device. The components of computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16.
The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)30 and/or cache memory 32. Computer device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, and commonly referred to as a "hard drive"). Although not shown in FIG. 6, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.
A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of the described embodiments of the invention.
The processing unit 16 executes various functional applications and data processing by running a program stored in the system memory 28, for example, to implement the road network state determination method provided in any embodiment of the present invention.
EXAMPLE five
Fifth, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for determining a road network state according to any one of the above embodiments is implemented.
The road network state determining method comprises the following steps:
determining an incidence relation between intersections and road sections in a road network;
constructing a graph neural network model of the road network based on the incidence relation;
training a vertex representing the road section in the graph neural network model to obtain a feature vector of the vertex;
under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector;
and determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (11)
1. A method for determining a road network state, comprising:
determining an incidence relation between intersections and road sections in a road network;
constructing a graph neural network model of the road network based on the incidence relation;
training a vertex representing the road section in the graph neural network model to obtain a feature vector of the vertex;
under the constraint of the attribute information of the road section, calculating the similarity between the vertexes based on the feature vector;
and determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
2. The method of claim 1, wherein said determining the association between intersections and road segments in the road network comprises:
acquiring geographic information of road sections in a road network;
determining a road intersection where a plurality of the road segments intersect as an intersection based on the geographic information;
and constructing an adjacency matrix by utilizing the adjacency relation between the intersection and the road section as the incidence relation between the intersection and the road section.
3. The method according to claim 2, wherein the constructing an adjacency matrix using the adjacency relationship between the intersection and the road segment as the association relationship between the intersection and the road segment comprises:
determining the entering and exiting direction between the road section in the neighborhood range of each intersection and the intersection;
determining a road section of which the traffic flow is input into the intersection as an incident road section and a road section of which the traffic flow is output from the intersection as an emergent road section according to the access direction;
and constructing an adjacent matrix of the intersection by using the incident road section and the emergent road section as an incidence relation between the intersection and the road section.
4. The method of claim 1, wherein the association is represented as a adjacency matrix;
the constructing of the graph neural network model of the road network based on the incidence relation comprises the following steps:
classifying said road segments in said road network into a plurality of road segment subsets according to said adjacency matrix, each of said road segment subsets comprising all said road segments belonging to the same intersection;
determining a sub-graph region composed of each of the road segment subsets;
and connecting a plurality of the sub-graph regions together based on the intersection of the sub-graph regions to form a graph neural network model of the road network.
5. The method of claim 4, wherein determining the sub-graph region comprised of each of the subset of road segments comprises:
and in each road segment subset, constructing a sub-graph region of the road segment subset by taking each road segment as a vertex and taking a connection relation between any two vertexes as an edge.
6. The method of claim 1, wherein the training vertices representing the segments in the graph neural network model to obtain feature vectors of the vertices comprises:
in the graph neural network model, aiming at each vertex representing the road section, sampling a preset number of vertexes in a neighborhood range of the vertexes as neighbor vertexes;
determining an aggregation function;
and for each vertex, performing information aggregation on the neighbor vertex of the vertex according to the aggregation function to obtain the feature vector of the vertex.
7. The method according to any one of claims 1 to 6, wherein the calculating the similarity between the vertices based on the feature vectors under the constraint of the attribute information of the road segment includes:
determining the traffic road condition of the road section along with the time-space change in the road network to obtain the attribute information of the road section;
and adjusting the feature vectors by using the attribute information, and calculating the similarity between the vertexes based on the adjusted feature vectors.
8. The method according to claim 7, wherein said determining the traffic road condition of said road segment in said road network varying with time and space to obtain the attribute information of said road segment comprises:
acquiring historical traffic state data of the road network;
determining a traffic flow model representing traffic flow in the road network based on the traffic status data;
calculating the association time length, the association strength, the association direction and the association transitivity of the road section in the traffic flow model to obtain the attribute information of the road section;
wherein the associated time length of the link represents the influence time of the adjacent link on the traffic state of the link, the associated strength of the link represents the ratio of the traffic flow of the link to the total traffic flow of the adjacent link, the associated direction of the link represents the direction of the traffic flow of the link, and the associated transmissibility of the link represents the transmission direction of the traffic wave of the link.
9. A road network state determination device, comprising:
the incidence relation determining module is used for determining the incidence relation between intersections and road sections in the road network;
the graph neural network model building module is used for building a graph neural network model of the road network based on the incidence relation;
the characteristic vector determining module is used for training a vertex representing the road section in the graph neural network model to obtain a characteristic vector of the vertex;
the similarity calculation module is used for calculating the similarity between the vertexes based on the feature vector under the constraint of the attribute information of the road section;
and the state determining module is used for determining intersections and road sections in the road network, wherein the intersections and the road sections are related to each other in traffic mode and traffic state according to the similarity.
10. A computer device, characterized in that the computer device comprises:
one or more processors;
a memory for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the method for determining road network status according to any one of claims 1-8.
11. A computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when executed by a processor, the computer program implements the road network state determination method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110592888.2A CN113256980A (en) | 2021-05-28 | 2021-05-28 | Road network state determination method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110592888.2A CN113256980A (en) | 2021-05-28 | 2021-05-28 | Road network state determination method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113256980A true CN113256980A (en) | 2021-08-13 |
Family
ID=77185137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110592888.2A Pending CN113256980A (en) | 2021-05-28 | 2021-05-28 | Road network state determination method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113256980A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114187341A (en) * | 2021-11-16 | 2022-03-15 | 泰瑞数创科技(北京)有限公司 | Artificial neural network road texture mapping method and system based on mobile following recognition |
| CN114418468A (en) * | 2022-03-29 | 2022-04-29 | 成都秦川物联网科技股份有限公司 | Smart city traffic scheduling strategy control method and Internet of things system |
| CN114822019A (en) * | 2022-03-31 | 2022-07-29 | 阿里巴巴（中国）有限公司 | Traffic information processing method and device |
| CN116303835A (en) * | 2023-05-24 | 2023-06-23 | 北京建筑大学 | Mapping data management method for road planning |
| CN116680423A (en) * | 2023-08-03 | 2023-09-01 | 国网浙江浙电招标咨询有限公司 | Management method, device, equipment and medium for multi-source heterogeneous data of power supply chain |
| US12033501B2 (en) | 2022-03-29 | 2024-07-09 | Chengdu Qinchuan Iot Technology Co., Ltd. | Methods for traffic scheduling at intersections in smart cities and Internet of Things (IoT) systems thereof |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111540198A (en) * | 2020-04-17 | 2020-08-14 | 浙江工业大学 | Urban traffic situation recognition method based on directed graph convolution neural network |
| CN111754019A (en) * | 2020-05-08 | 2020-10-09 | 中山大学 | Road section feature representation learning algorithm based on space-time diagram information maximization model |
| CN112382082A (en) * | 2020-09-30 | 2021-02-19 | 银江股份有限公司 | Method and system for predicting traffic running state in congested area |
| CN112669594A (en) * | 2020-12-11 | 2021-04-16 | 国汽(北京)智能网联汽车研究院有限公司 | Method, device, equipment and storage medium for predicting traffic road conditions |
-
2021
- 2021-05-28 CN CN202110592888.2A patent/CN113256980A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111540198A (en) * | 2020-04-17 | 2020-08-14 | 浙江工业大学 | Urban traffic situation recognition method based on directed graph convolution neural network |
| CN111754019A (en) * | 2020-05-08 | 2020-10-09 | 中山大学 | Road section feature representation learning algorithm based on space-time diagram information maximization model |
| CN112382082A (en) * | 2020-09-30 | 2021-02-19 | 银江股份有限公司 | Method and system for predicting traffic running state in congested area |
| CN112669594A (en) * | 2020-12-11 | 2021-04-16 | 国汽(北京)智能网联汽车研究院有限公司 | Method, device, equipment and storage medium for predicting traffic road conditions |
Non-Patent Citations (2)
| Title |
|---|
| 周明等: "基于时空信息的交通流预测", 《工业控制计算机》 * |
| 张舒等: "空间位置的关联分析及其向量化表示方法", 《计算机***应用》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114187341A (en) * | 2021-11-16 | 2022-03-15 | 泰瑞数创科技(北京)有限公司 | Artificial neural network road texture mapping method and system based on mobile following recognition |
| CN114187341B (en) * | 2021-11-16 | 2022-09-06 | 泰瑞数创科技(北京)股份有限公司 | Artificial neural network road texture mapping method and system based on mobile following recognition |
| CN114418468A (en) * | 2022-03-29 | 2022-04-29 | 成都秦川物联网科技股份有限公司 | Smart city traffic scheduling strategy control method and Internet of things system |
| CN114418468B (en) * | 2022-03-29 | 2022-07-05 | 成都秦川物联网科技股份有限公司 | Smart city traffic scheduling strategy control method and Internet of things system |
| US11587436B1 (en) | 2022-03-29 | 2023-02-21 | Chengdu Qinchuan Iot Technology Co., Ltd. | Methods for controlling traffic scheduling strategies in smart cities and Internet of Things (IoT) systems thereof |
| US12033501B2 (en) | 2022-03-29 | 2024-07-09 | Chengdu Qinchuan Iot Technology Co., Ltd. | Methods for traffic scheduling at intersections in smart cities and Internet of Things (IoT) systems thereof |
| CN114822019A (en) * | 2022-03-31 | 2022-07-29 | 阿里巴巴（中国）有限公司 | Traffic information processing method and device |
| CN116303835A (en) * | 2023-05-24 | 2023-06-23 | 北京建筑大学 | Mapping data management method for road planning |
| CN116303835B (en) * | 2023-05-24 | 2023-07-21 | 北京建筑大学 | Mapping data management method for road planning |
| CN116680423A (en) * | 2023-08-03 | 2023-09-01 | 国网浙江浙电招标咨询有限公司 | Management method, device, equipment and medium for multi-source heterogeneous data of power supply chain |
| CN116680423B (en) * | 2023-08-03 | 2023-10-20 | 国网浙江浙电招标咨询有限公司 | Management method, device, equipment and medium for multi-source heterogeneous data of power supply chain |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ali et al. | Exploiting dynamic spatio-temporal graph convolutional neural networks for citywide traffic flows prediction | |
| CN113256980A (en) | Road network state determination method, device, equipment and storage medium | |
| Zhang et al. | Trafficgan: Network-scale deep traffic prediction with generative adversarial nets | |
| Miglani et al. | Deep learning models for traffic flow prediction in autonomous vehicles: A review, solutions, and challenges | |
| Fang et al. | FTPG: A fine-grained traffic prediction method with graph attention network using big trace data | |
| CN112216108A (en) | Traffic prediction method based on attribute-enhanced space-time graph convolution model | |
| Rahmani et al. | Graph neural networks for intelligent transportation systems: A survey | |
| CN114330868A (en) | Passenger flow prediction method based on self-attention personalized enhanced graph convolution network | |
| Faro et al. | Evaluation of the traffic parameters in a metropolitan area by fusing visual perceptions and CNN processing of webcam images | |
| CN113283581B (en) | Multi-fusion graph network collaborative multi-channel attention model and application method thereof | |
| Liu et al. | Spatial-temporal conv-sequence learning with accident encoding for traffic flow prediction | |
| Shen et al. | An attention-based digraph convolution network enabled framework for congestion recognition in three-dimensional road networks | |
| James | Citywide estimation of travel time distributions with Bayesian deep graph learning | |
| Yu et al. | Passenger flow prediction for new line using region dividing and fuzzy boundary processing | |
| CN116662468A (en) | Urban functional area identification method and system based on geographic object space mode characteristics | |
| Qaddoura et al. | Temporal prediction of traffic characteristics on real road scenarios in Amman | |
| Prontri et al. | Traffic signal control using fuzzy logic | |
| Liao et al. | Taxi demand forecasting based on the temporal multimodal information fusion graph neural network | |
| Ke et al. | AutoSTG+: An automatic framework to discover the optimal network for spatio-temporal graph prediction | |
| He et al. | Multi-feature urban traffic prediction based on unconstrained graph attention network | |
| Zheng et al. | A deep spatiotemporal network for forecasting the risk of traffic accidents in low-risk regions | |
| Zhang et al. | Off-deployment traffic estimation—a traffic generative adversarial networks approach | |
| Menguc et al. | A data driven approach to forecasting traffic speed classes using extreme gradient boosting algorithm and graph theory | |
| CN113159371A (en) | Unknown target feature modeling and demand prediction method based on cross-modal data fusion | |
| Bhaumik et al. | STLGRU: Spatio-temporal lightweight graph GRU for traffic flow prediction |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210813 |