CN112183926B

CN112183926B - Traffic running condition evaluation method, device, equipment and medium based on graph database

Info

Publication number: CN112183926B
Application number: CN202010871780.2A
Authority: CN
Inventors: 季青原; 徐甲; 陈乾; 万雨茜; 林文霞; 吴占宁; 温晓岳
Original assignee: Yinjiang Technology Co ltd
Current assignee: Yinjiang Technology Co ltd
Priority date: 2020-08-26
Filing date: 2020-08-26
Publication date: 2023-03-31
Anticipated expiration: 2040-08-26
Also published as: CN112183926A

Abstract

The invention discloses a traffic running condition evaluation method and device based on a graph database, electronic equipment and a computer storage medium, relates to the field of intelligent traffic, and aims to realize comprehensive evaluation of urban traffic running conditions. Wherein the method comprises the steps of: storing a knowledge graph of traffic operation data through a graph database; obtaining an index from the graph database: alarm and timing data of the target intersection, and an entrance road section and road section speed data of the target intersection; calculating regulation and control efficiency according to the alarm and timing data; calculating a periodic correlation coefficient of the target intersection according to the road section of the entrance and the speed data of the road section; calculating average regulation efficiency and average cycle correlation coefficient of the target area under different time resolutions; and on the basis of the average regulation efficiency and the average period correlation coefficient, calculating the comprehensive operation index of the target area under different time resolutions through a predefined index importance ranking rule.

Description

Traffic operation condition evaluation method, device, equipment and medium based on graph database

Technical Field

The invention relates to the field of intelligent traffic, in particular to a traffic running condition evaluation method, device, equipment and medium based on a graph database.

Background

In recent years, with the rapid development of big data technology and artificial intelligence technology, the concept of "smart traffic" is gradually proposed to manage urban traffic congestion, and urban traffic control is also being changed into digitalization and intellectualization. At a data storage end, a traditional traffic control system depends on a relational database to store traffic data in a table, and the storage mode is simple and convenient when data are written.

However, traffic control often involves multi-source heterogeneous and correlated entities and traffic running state data, and the relational database cannot effectively express the complex association relationship, and meanwhile, due to low efficiency of multi-table correlated query operation, in an actual application scenario, multi-dimensional data is acquired from the relational database to realize multi-dimensional evaluation of the running condition of urban traffic, which is difficult and time-consuming.

Disclosure of Invention

In order to overcome the defects of the prior art, one of the purposes of the invention is to provide a graph database-based traffic operation condition evaluation method, which effectively expresses the semantic association relationship between entities and obtains multidimensional data with higher query efficiency so as to realize comprehensive evaluation of urban traffic operation conditions.

One of the purposes of the invention is realized by adopting the following technical scheme:

a traffic operation condition evaluation method based on a graph database comprises the following steps:

storing a knowledge graph of traffic operation data through a graph database, wherein the knowledge graph of the traffic operation data comprises entities, attributes and relationships among the entities of the traffic operation data;

acquiring alarm and timing data of a target intersection, an entrance road section of the target intersection and road section speed data in a preset time period from the map database;

calculating the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection;

calculating a cycle correlation coefficient of the target intersection according to the entrance road section and the road section speed data of the target intersection;

calculating average regulation efficiency and average cycle correlation coefficient of a target area under different time resolutions according to the regulation efficiency of the target intersection and the cycle correlation coefficient of the target intersection;

and based on the average regulation and control efficiency and the average cycle correlation coefficient, calculating a comprehensive operation index of the target area under different time resolutions through a predefined importance ranking rule.

Further, the map database stores a knowledge map of traffic operation data, and comprises:

establishing a knowledge graph of static road network data according to entities, attributes and relationships among the entities of the static road network data, and storing the knowledge graph of the static road network data in the graph database, wherein the static road network data comprises different hardware components of a road network and incidence relationships among the different hardware components;

establishing a knowledge graph of the dynamic road network data according to the entities, attributes and relationships among the entities of the dynamic road network data, and storing the knowledge graph of the dynamic road network data in the graph database, wherein the dynamic road network data comprises timing data of signal machines and traffic speed of road sections;

establishing a knowledge graph of the traffic event data according to the entities, attributes and relationships among the entities of the traffic event data, and storing the knowledge graph of the traffic event data in the graph database, wherein the traffic event data comprises traffic events of intersections, coordination groups and areas.

Further, the calculating the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection comprises:

calculating the duration of continuous alarm according to the alarm data of the target intersection;

respectively calculating the time length of the target intersection with the regulation alarm and the time length of the target intersection without the regulation alarm according to the timing data of the target intersection;

according to the time length of the target intersection with the regulation alarm and the time length of the target intersection without the regulation alarm, calculating the regulation efficiency of the target intersection through the following formula:

wherein d is _n Indicates that the target intersection has n times of unregulated alarm and the time length of each time is d, f _m The time length of each time is f, and the avg () is a mean value function.

Further, the calculating a periodic correlation coefficient of the target intersection according to the entrance section and the section speed data of the target intersection includes:

calculating the congestion coefficient of the target intersection by the following formula:

v ^inst (P _i ,T _d,k )＝max{v ^nostop (L _j1,i ,T _d,k ),…,v ^nostop (L _jn,i ,T _d,k )}，

wherein L is _j,i Representing a target intersection P _i Traffic flow direction P _j N denotes the number of entry links, { L _j1,i ,L _j2,i ,…,L _jn,i Denotes a set of inlet road sections, v ^free (P _i ,T _d,k ) Representing a target intersection P _i At T _d,k Free speed of the vehicle, v ^nostop Indicating non-stopping speed, v ^inst (P _i ,T _d,k ) Representing a target intersection P _i At T _d,k Instantaneous vehicle speed of time, s (P) _i ,T _d,k ) Representing a target intersection P _i At T _d,k A congestion coefficient of time;

calculating a congestion time sequence curve of the target intersection in a specific time period according to the congestion coefficient:

S(P _i ，T _d )＝{s(P _i ，T _d,k ) H, where k =1,2, …, NUM;

obtaining annunciator XH _i A cycle timing curve of the specific period: ZQ (XH) _i ,XT _d )＝{zq(XH _i ,XT _d,l ) I =1,2, …, NM;

extracting the congestion time sequence curve and the period time sequence curve corresponding to time, and calculating a period correlation coefficient a (P) of the target intersection _i ，T _d ) Wherein

Further, based on the average regulation efficiency and the average cycle correlation coefficient, a comprehensive operation index of the target area under different time resolutions is obtained through a predefined importance ranking rule, and the method comprises the following steps:

and according to a predefined importance ranking rule, carrying out normalization, hierarchical analysis and weighted summation processing on the average regulation and control index and the average period correlation coefficient to obtain a comprehensive operation index of the target area under different space-time resolutions.

Further, the method further comprises: screening of the target intersections:

the query task is received and the query task is received,

analyzing the query task to extract a search keyword of the query task;

extracting entities and attributes matched with the search keywords from the knowledge graph according to the search keywords, extracting features and generating a feature graph;

and based on the characteristic graph, performing similarity calculation with the entities in the knowledge graph to obtain entities with similar characteristics and returning.

Further, the method further comprises:

according to the intersection P _i Congestion coefficient S (P) _i ，T _d ) And a predefined congestion level determination rule for said intersection P _i Matching a corresponding congestion level S ^deg (P _i ,T _d,k )；

Calculating the intersection P _i At T _d,k Sequence of intensity of congestion propagation Q of a time _i,j 。

The invention also aims to provide a traffic operation condition evaluation device based on a graph database, which aims to effectively express semantic association relation between entities and acquire multi-dimensional data with higher query efficiency so as to realize comprehensive evaluation of urban traffic operation conditions.

A graph database-based traffic behavior assessment apparatus comprising:

the data storage module is used for storing the knowledge graph of the traffic operation data through the graph database, and the knowledge graph of the traffic operation data comprises entities, attributes and relationships among the entities of the traffic operation data;

the data extraction module is used for acquiring alarm and timing data of the target intersection, an entrance road section of the target intersection and road section speed data in a preset time period from the map database;

the data calculation module is used for calculating the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection; calculating a cycle correlation coefficient of the target intersection according to the entrance road section and the road section speed data of the target intersection; calculating average regulation efficiency and average cycle correlation coefficient of a target area under different time resolutions according to the regulation efficiency of the target intersection and the cycle correlation coefficient of the target intersection;

and the comprehensive evaluation module is used for solving a comprehensive operation index of the target area under different time resolutions through a predefined importance ranking rule based on the average regulation and control efficiency and the average period correlation coefficient.

It is a further object of the present invention to provide an electronic device for performing one of the above objects, comprising a processor, a storage medium and a computer program, the computer program being stored in the storage medium, the computer program, when executed by the processor, implementing the above method for estimating traffic conditions based on a graph database.

It is a fourth object of the present invention to provide a computer-readable storage medium storing one of the objects of the invention, having a computer program stored thereon, which, when executed by a processor, implements the above-described map database-based traffic behavior assessment method.

Compared with the prior art, the invention has the beneficial effects that:

storing entities, semantics, and attributes within the traffic management system based on the knowledge-graph and the graph database. The knowledge graph stores entities in the traffic control system with different space-time scales, relates to a plurality of dimensional data of traffic control, and stores the knowledge graph by using the graph database, so that the association relation between the entities is effectively stored, and the multidimensional data can be acquired with higher query efficiency.

Based on the graph database, relevant data stored in the knowledge graph can be read, corresponding traffic operation indexes are calculated, and multi-dimensional and all-around evaluation on traffic operation conditions can be achieved.

Drawings

FIG. 1 is a flow chart of a graph database based traffic operating condition assessment method of the present invention;

FIG. 2 is a schematic diagram showing static road network data entities and relationships;

FIG. 3 is a schematic diagram showing dynamic road network data entities, relationships, and attributes;

FIG. 4 is a schematic diagram showing entities, relationships, and attributes relating to a traffic event;

FIG. 5 is a block diagram showing a configuration of a data updating apparatus according to the distributed database according to embodiment 3;

fig. 6 is a schematic structural diagram of an electronic device according to embodiment 4 of the present invention.

Detailed Description

The present invention will now be described in more detail with reference to the accompanying drawings, in which the description of the invention is given by way of illustration and not of limitation. The various embodiments may be combined with each other to form other embodiments not shown in the following description.

Example 1

The embodiment provides a traffic running condition evaluation method based on a graph database, which aims to effectively express semantic association relations between entities and acquire multi-dimensional data with higher query efficiency so as to realize comprehensive evaluation of urban traffic running conditions.

Fig. 1 is a flowchart of a traffic operation condition evaluation method based on a graph database according to the present invention, as shown in fig. 1, the traffic operation condition evaluation method based on the graph database of the present embodiment specifically includes the following steps:

and S110, storing a knowledge graph of the traffic operation data through a graph database, wherein the knowledge graph of the traffic operation data comprises entities, attributes and relationships among the entities of the traffic operation data.

In this embodiment, the knowledge graph storing the traffic operation data is used for storing traffic control entities with different space-time scales, and simultaneously stores a plurality of dimensional data of traffic control from a single lane, intersection, road section to the whole area. The traffic operation data comprises static road network data, dynamic road network data and traffic events, and entities, semantic relations and attributes of the traffic operation data are stored through a knowledge graph respectively to form a traffic operation knowledge graph. The traffic operation knowledge graph is stored through the graph database, the responsible incidence relation among the traffic control related entities can be conveniently expressed, higher query efficiency compared with a relational data path can be provided, the service requirement and the processing efficiency requirement of intelligent traffic are met, and data support is provided for the evaluation of traffic operation conditions.

S120, alarm and timing data of the target intersection, an entrance road section of the target intersection and road section speed data in a preset time period are obtained from the map database.

In the embodiment, by using the knowledge map calculation and mining technology, data such as entities, relations and attributes can be inquired from a map database, and a traffic control system is evaluated to evaluate the pressure faced by the traffic control entity and the corresponding regulation capacity. In the embodiment, an application scenario of multidimensional traffic operation condition evaluation is mainly discussed, and the preset time period may be 1 month.

From different dimensions, the operating conditions of the traffic control entity can be evaluated. In this embodiment, the following 2 indexes are used to evaluate the traffic operation condition, and the 2 indexes are: the regulation and control efficiency of the intersection and the periodic correlation coefficient of the intersection. Of course, in other embodiments, other indexes may be selected according to actual conditions to evaluate the traffic operation conditions.

The target intersection may be all intersections in the road network, may be an intersection in a certain area, or may be an intersection selected by the user.

And S130, calculating the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection.

When an alarm occurs at an intersection, the traffic abnormality at the intersection is indicated, so that the traffic abnormality needs to be regulated and controlled, the traffic abnormality can be recovered to be normal, the regulation and control efficiency can be effectively evaluated, and the regulation and control influence on shortening of the alarm time at the intersection is avoided. The higher the regulation efficiency is, the larger the regulation amplitude is, so that the alarm time of the intersection is shortened.

And S140, calculating the cycle correlation coefficient of the target intersection according to the entrance road section and the road section speed data of the target intersection.

When the intersection is congested, the traffic speed of the intersection will drop significantly, and corresponding measures should be taken to the intersection, such as increasing the period of an intersection signal machine, so that the period correlation coefficient of the intersection can be calculated through the entrance road section and the road section speed data of the intersection to reflect the congestion condition of the intersection. Specifically, when the cycle correlation coefficient is higher, the intersection has better operation conditions for congestion (when the intersection is more congested, the larger the signal period is, the more helpful for relieving congestion is). Otherwise, if the cycle correlation coefficient is low, the intersection has poor operation conditions for congestion.

S150, calculating average regulation efficiency and average cycle correlation coefficient of the target area under different time resolutions according to the regulation efficiency of the target intersection and the cycle correlation coefficient of each intersection.

The time resolution refers to that different time resolutions are selected in a specific time period every day or every month, and the average regulation efficiency and the cycle correlation coefficient of a certain region under the specific time resolution can be calculated according to the regulation efficiency and the cycle correlation coefficient of all intersections in the region.

And S160, based on the average regulation efficiency and the average cycle correlation coefficient, calculating the comprehensive operation index of the target area under different time resolutions through a predefined importance ranking rule.

In the embodiment, an importance ranking rule of the regulation efficiency and the period correlation coefficient is predefined, and according to the rule, a comprehensive operation index of the target area under different time resolutions can be obtained through a weighted summation mode so as to comprehensively evaluate the traffic operation condition. Of course, in this embodiment, only 2 indexes are taken as an example, and the traffic operation condition is comprehensively evaluated from 2 dimensions of the intersection alarm and the traffic operation state. In an actual application scene, each dimension often has a plurality of indexes, and the comprehensive operation index can be obtained by adopting a plurality of indexes of the plurality of dimensions.

The embodiment obtains multi-dimensional data based on the graph database and the knowledge graph, and can realize multi-dimensional comprehensive evaluation on traffic operation conditions.

Preferably, the map database stores a knowledge map of traffic operation data, including:

establishing a knowledge graph of the traffic event data according to the entities, attributes and relationships among the entities of the traffic event data, and storing the knowledge graph of the traffic event data in the graph database, wherein the traffic event data comprises traffic events of intersections, coordination groups and regions.

The static road network data relates to different hardware components of the road network, such as detectors, semaphores, intersections, road segments and the like, and the incidence relation among the hardware components, such as one semaphores serving a certain intersection, and one road segment being an exit road segment and an entrance road segment of a certain intersection.

The dynamic road network data comprises signal timing sequence data and road section passing speed sequence data, and the two data are used as decision support for traffic running state evaluation.

The traffic events comprise traffic abnormity, traffic regulation and control and operation state switching, the traffic abnormity data comprise data such as congestion, accidents and alarms, the traffic regulation and control data comprise signals, namely data such as period change and phase sequence change, and the operation state switching comprises information such as intersection change from unobstructed to congested. Traffic events may describe and assess traffic operating pressures and subsequent pressure regulation capabilities.

Fig. 2 is a schematic diagram showing entities and relationships of static road network data. In fig. 2 the entities are represented by circles and the relations are represented by arrows. The attributes of the static road network data in this embodiment are shown in the following table:

/>

after static road network data is represented by a knowledge graph, the static road network data is stored in a map database, and the selection of the specific map database is not limited herein.

Fig. 3 is a schematic diagram showing entities, relationships and attributes of dynamic road network data, and as shown in fig. 3, in the dynamic road network data, the timing data of the traffic signal includes a timing cycle and a phase sequence of the traffic signal, and the traffic speed of the road section includes a parking speed and a non-parking speed of the road section. And after the dynamic road network data is expressed by a knowledge graph, storing the dynamic road network data in a database.

Traffic events may occur at traffic management entities in different spatial dimensions, including intersections. Coordinating groups and regions. Taking the intersection as an example, the traffic abnormal data of the intersection comprises alarm data, congestion data and accident data. The traffic regulation and control data of the intersection comprises cycle increasing information, cycle decreasing information and phase sequence change information, the operation state switching data of the intersection comprises (from smooth to jammed) and switching type B (from jammed to smooth), and the traffic event data of the coordination group and the area is similar to the intersection. FIG. 4 is a schematic diagram showing entities, relationships, and attributes relating to a traffic event. It should be noted that all arrows in fig. 4 represent a relationship, represent "occurrence" of a traffic event, and the occurrence time is stored in the relationship in an attribute manner. The traffic event data is represented by a knowledge graph and then stored in a graph database.

In the present embodiment, each data is represented by the following notation: road network S comprises I intersections P, P _i Indicates the ith intersection, the range of i [1,I]Section L of road _i,j Indicating an intersection P _i Traffic flow direction crossing P _j Road section L _i,j At intersection P _i Intersection P _j Are connected to each other, L _j,i Indicating intersection P _j Traffic flow direction crossing P _i Road section of (1), hop _i,j Indicating vehicle departure from intersection P _i Travel to the intersection P _j The shortest number of passing road sections, one intersection corresponding to one signaler, XH _i Representing the ith signal machine;

T _d,k denotes the k-th velocity data acquisition time on day d, the range of k [1, NUM ]],v ^stop (L _i,j ,T _d,k ) Representing a section of road L _i,j At T _d,k Time-derived parking speed, v ^nostop (L _i,j ,T _d,k ) Representing a section of road L _i,j At T _d,k Time-collected no-stop speed;

XT _d,l denotes the l-th time-to-date data acquisition time on day d, the range of l [1, NM]，zq(XH _i ,XT _d,l ) Representing annunciators XH _i At XT _d,l Period of time acquisition, xx (XH) _i ,XT _d,l ) Representing annunciators XH _i At XT _d,l Phase sequence of time acquisition, { zq (XH) _i ,XT _d,l )，xx(XH _i ,XT _d,l ) The timing data is composed;

for intersection P _i It is said to be at T _d,k The traffic event at which time is located may be represented as: j (P) _i ,T _d,k )。

Preferably, the calculating the control efficiency of the target intersection according to the alarm and timing data of the target intersection includes:

respectively calculating the time length of the target intersection with regulation and control alarm and the time length of the target intersection without regulation and control alarm according to the timing data of the target intersection;

according to the time length of the regulation alarm and the time length of the non-regulation alarm of the target intersection, calculating the regulation efficiency of the target intersection by the following formula:

wherein d is _n Indicates that the target intersection has n times of unregulated alarm and the time length of each time is d, f _m The time length of each time is f, and avg () is a mean function.

In this embodiment, a certain continuous alarm duration is named as "no-regulation alarm duration", that is, no regulation occurs at the intersection; and the alarm time length is regulated and controlled, namely the regulation and control of the intersection are carried out. According to a preset time period T _L That is, given a time period, in which the preset time period is one month, for the intersection P _i Suppose that no regulation alarm { d) occurs n times at the intersection ₁ ,d ₂ ,…,d _n And the time length of each time is d, supposing that m times of regulation and control alarm occur at the intersection { f ₁ ,f ₂ ,…,f _m And each time duration is f, and avg () is specified as a mean value function, then the regulation efficiency E of the intersection can be obtained _i The higher the regulation efficiency is, the better the effect of regulating and controlling the alarm duration of the intersection is shortened. By adopting the regulation efficiency calculation method, the regulation efficiency of each intersection can be sequentially and circularly calculated.

Preferably, in an actual application scenario, an intersection set with the lowest regulation and control efficiency can be determined according to the calculated regulation and control efficiency of each intersection, and targeted regulation and control scheme optimization can be performed for intersections in the intersection set, so that the regulation and control efficiency is improved.

Preferably, the calculating a periodic correlation coefficient of the target intersection according to the entry road section and the road section speed data of the target intersection includes:

wherein L is _j,i Representing a target intersection P _i Traffic flow direction P _j N denotes the number of entry links, { L { _j1,i ,L _j2,i ,…,L _jn,i Denotes a set of inlet road sections, v ^free (P _i ,T _d,k ) Representing a target intersection P _i At T _d,k Free speed of the vehicle, v ^nostop Indicating non-stopping speed, v ^inst (P _i ,T _d,k ) Representing a target intersection P _i At T _d,k Instantaneous vehicle speed of time, s (P) _i ,T _d,k ) Representing a target intersection P _i At T _d,k A congestion coefficient of time;

S(P _i ，T _d )＝{s(P _i ，T _d,k ) Where k =1,2, …, NUM;

obtaining annunciator XH _i A cycle timing curve of the specific period: ZQ (XH) _i ,XT _d )＝{zq(XH _i ,XT _d,l ) Where i =1,2, …, NM;

extracting the congestion time sequence curve and the periodic time sequence curve corresponding to time, and calculating a periodic correlation coefficient a (P) of the target intersection _i ，T _d ) Wherein

In this embodiment, the specific time interval is one day, the congestion time sequence curve and the periodic time sequence curve are fitted respectively, then time sampling is performed, and S (P) corresponding to time is extracted _i ，T _d ) And ZQ (XH) _i ,XT _d ) And calculating, specifically, calculating a periodic correlation coefficient by using DTW, wherein the periodic correlation coefficient can represent the congestion condition of the intersection. The periodic correlation coefficient of each intersection can be sequentially and circularly calculated by adopting the periodic correlation coefficient calculating method.

Preferably, in an actual application scenario, an intersection set with the lowest periodic correlation coefficient may be determined according to the calculated periodic correlation coefficient of each intersection, and a targeted solution may be formulated for intersections in the intersection set to alleviate the congestion condition.

Preferably, the step of obtaining the comprehensive operation index of the target region under different time resolutions by a predefined importance ranking rule based on the average regulation efficiency and the average period correlation coefficient includes:

In this embodiment, the hierarchical analysis uses an AHP hierarchical analysis method, and averages the regulation efficiency and the cycle correlation coefficient of each intersection in the target area at a specific time resolution to obtain an average regulation efficiency and a cycle correlation coefficient of the target area, and sequentially performs normalization processing, hierarchical analysis processing, and weighted summation processing on the average regulation efficiency and the cycle correlation coefficient according to a predefined importance ranking rule, thereby obtaining a comprehensive operation index of the target area at a specific time-space resolution to complete comprehensive evaluation of the traffic operation condition. By the graph database-based traffic operation condition evaluation method, comprehensive operation indexes of different areas under different spatial resolutions can be realized, so that traffic operation conditions can be comprehensively evaluated.

Example 2

The embodiment is different from the above embodiments in that after the knowledge graph of the traffic operation data is stored by the graph database, the knowledge graph is applied to entity semantic search and knowledge matching to construct a question-answering system.

Because users often do not have the query experience of graph databases and do not concern the specific mode of searching graph data, when users initiate query tasks through terminals, the users often ask questions in a natural language mode and need to return answers by using knowledge graphs.

For example, in the query task initiated by the user, a question is provided as "3/20/2020, which of the ten intersections most similar to the congestion propagation rule of the intersection a is, and what the comprehensive operation index of the intersection a and the ten similar intersections is".

The problem solution can be realized based on the knowledge graph of the traffic operation data stored in the graph database.

Therefore, a traffic operation condition evaluation method based on a graph database further comprises the following steps:

screening of the target intersection specifically comprises:

the query task is received and the query task is received,

analyzing the query task to extract search keywords of the query task;

extracting entities and attributes matched with the search keywords from the knowledge graph according to the search keywords, extracting the characteristics of the entities matched with the keywords, and generating a characteristic graph;

and based on the characteristic map, performing similarity calculation with the entities in the knowledge map to obtain entities with similar characteristics, and returning.

And analyzing the query task, namely converting the problems which are provided by the user in natural language in the query task into a search task related to the traffic control entity in the knowledge graph. Mainly based on the conventional natural semantic processing technology, the analysis steps are as follows:

performing word segmentation and semantic analysis (semantic analysis) on the problems in the query task to obtain an analysis result;

the purpose of word segmentation is to divide the Chinese natural language into words as units, and the main technical means used comprises word segmentation based on a dictionary and word segmentation based on statistics. In the embodiment, the knowledge graph is applied to the professional field of the traffic control field, and the word segmentation based on the field dictionary can obtain higher accuracy, so that the word segmentation method based on the field dictionary is adopted. The semantic analysis aims to enable a computer to understand the meaning of a sentence semantically, for example, "3/20/2020 early peak hours, which are the ten intersections most similar to the congestion propagation rule of the intersection 1", the sentence means that intersections similar to the congestion propagation rule of the intersection are found, the intersections are sorted in a descending order according to the similarity, and the intersection with the similarity top10 is taken as an answer. Semantic analysis, which mainly adopts a word embedding technology and a deep neural network model, such as a BERT model of ***, can extract semantic information in natural language, and is a conventional processing method in the field, and the process is not described herein.

And returning the analysis result as a keyword, wherein the keyword comprises keywords such as time, space, entity, characteristic, limitation and the like. Taking the above-mentioned problems as an example, the keywords include keywords such as 3/20/2020 early peak hours (6.

After the search keywords are specified, the search keywords are matched with the entities and attributes of the knowledge-graph of the traffic operation data, such as time keywords (3/20/2020 early peak) indicating that dynamic road network data under this time period (6.

In the embodiment, the congestion propagation law needs to be extracted, which depends on the construction of the feature map. First, a series of hyper-parameters need to be defined: calculating the interval length x, the time window W and the probability threshold PP by the congestion coefficient _t The hop distance threshold DD.

The specific construction process is as follows:

1) To the intersection P _i Determining its congestion time interval [ T _d,k-x ,T _d,k ]。

2) To crossing P _i Search for satisfaction of hop _i,j <Intersection P of DD _j A total of m, forming a set of neighboring intersections { P _j1 ,P _j2 ,…,P _jm }. Next, only P will be calculated _i At time [ T _d,k-x ,T _d,k ]Next, for any one P _j Congestion propagation coefficient of (PP) _i,j 。

3)PP _i,j The specific calculation method is as follows:

assuming that the intersection Pi is congested, the congestion time set { T is formed by X +1 congestion times _d,k-x ,T _d,k-x+1 ,…,T _d,k }. The number n of instants d satisfying the condition is calculated such that at each instant T a corresponding (closed interval) time interval [ T, T + w ] is provided]Lower, crossing P _j An overcrowding occurs. Calculating congestion propagation coefficient PP _i,j = n/(x + 1). If PP _i,j >PP _t Then the congestion is considered to exist at the secondary crossing P _i To P _j 。

PP is described below as an example _i,j Wherein the hyperparameter x is 8,W is 10 minutes, PP _t At 50%, calculate PP from the congestion records at intersection A and intersection B _A，B And assume that hop has been satisfied _A,B <DD：

As shown in the above table, for intersection P _A Of the 5 th congestion record 8, corresponding to a 10 minute interval of [8]Crossing P in the section _B Congestion of (a) is recorded as 1: 8:18. Thus consider intersection P _B Congestion occurs within this interval (as long as P is found) _B Is determined to be P, then at least 1 congestion record of _B Congestion occurs). By analogy with P _A Congestion time 8 _B Congestion also occurs in the 10-minute interval corresponding to this time. Thus in this example m is 5 and n is 9. Thus PP _A,B Is 5/9.

4) Marking the congestion propagation relation according to the congestion propagation coefficient if PP _i,j >PP _t Sign of congestion propagation relationship Tip _i,j =1, otherwise Tip _i,j ＝0。

In practical application, the road network S includes I intersections P _i I =1,2, …, I, at T _d,k Temporal congestion propagation coefficient PP _i,j Congestion propagation relation mark Tip _i,j . Based on the following table (representing T) _d,k Congestion propagation coefficient PP at time _i,j Marking Tip in relation to congestion propagation _i,j ) The congestion propagation relationship in the method can construct a feature map, and the feature map comprises various relationships such as a congestion propagation relationship, a static road network relationship and the like, and is applied to subsequent steps.

P ₁

P ₂

…

P _i

…

P _I

P ₁

——

PP _1,2 |Tip _1,2 ＝1

PP _1,i |Tip _1,i ＝0

PP _1,I |Tip _1,I ＝1

P ₂

PP _2,1 |Tip _2,1 ＝1

——

PP _2,i |Tip _2,i ＝1

PP _2,I |Tip _2,I ＝0

…

——

P _i

PP _i,1 |Tip _i,1 ＝0

PP _i,2 |Tip _i,2 ＝1

——

PP _i,I |Tip _i,I ＝1

…

——

P _I

PP _I,1 |Tip _I,1 ＝1

PP _I,2 |Tip _I,2 ＝0

PP _I,i |Tip _I,i ＝1

——

Based on the feature map, similarity calculation is carried out on the entity in the knowledge map, and an entity with similar features is found out:

in the present embodiment, taking intersection as an example, intersection P is aimed at _i According to the static road network data, the dynamic road network data and the calculated congestion propagation coefficient, the following parameters or characteristics can be obtained:

timing data including period data and phase sequence data { zq (XH) _i ,XT _d,l )，xx(XH _i ,XT _d,l ) }, intersection P _i At T _d,k Free stream vehicle speed v of time ^free (P _i ) At the intersection P _i At T _d,k Instantaneous vehicle speed v ^inst (P _i ,T _d,k ) At the intersection P _i At T _d,k Congestion coefficient s (P) of time _i ,T _d,k ) At the intersection P _i The cyclic correlation coefficient on day d is a (P) _i ,T _d ) At the intersection P _i At T _d,k Temporal congestion propagation relation PP _i,j And marking Tip of congestion propagation relation _i,j At the intersection P _i Attribute and relationship integration in static road networkAs a feature, LK (P) _i )。

Aiming at the congestion propagation relation, the characteristics can be further extracted, namely the intersection P _i At T _d,k In-degree, out-degree: out-degree (P) _i ,T _d,k ) In the congestion propagation relation table, intersection P _i Spread to the intersection P _j Tip of _i,j Number of =1, in-degree (P) _i ,T _d,k ) In the congestion propagation relation table, intersection P _j Spread to the intersection P _i Tip of _j,i Number of = 1.

Based on the characteristics, intersection similarity can be calculated, and the main method comprises the steps of processing static road network data by using a TransE model, extracting digital characteristics, classifying by using KNN (K-Nearest neighbor) and other algorithms together with other characteristics, and searching for approximate entities.

The above mentioned TransE model is one of the classic methods for representing and learning a knowledge graph, and the processing process thereof is not described herein. The core idea is to re-project the entities and attributes in the knowledge graph to a dense low-dimensional Euclidean vector space, so that irregular graph data is converted into regular data. For example, the static road network map KG can be constructed by using the data of the static road network ^static Then the KG is made by means of the TransE model ^static Each crossing P in _i Having its static feature representation (x) in two-dimensional Euclidean space _i1 ,x _i2 ). These features are then stitched together with other features to form a vector representation (x) _i1 ,x _i2 ,x _i3 ,…,x _in ) And then searching similar entities by using a KNN algorithm to obtain similar intersections.

In this embodiment, the obtained similar entities are sorted in a descending order, and the entities of the top10 are taken for returning, specifically, the intersection corresponding to the entities is returned.

In the above example: and returning to the intersection which is most similar to the congestion propagation rule of the intersection X in the early peak period of 3 months and 20 days, and searching the congestion propagation path by using the knowledge graph of the traffic operation data. However, for the city traffic management department, the city traffic management department not only wants to know whether the congestion is propagated, but also wants to know whether the congestion is increased (the congestion becomes serious) or dissipated (the congestion becomes light) in the propagation process, and also wants to know the evolution law of the congestion or the dissipation on the city road network, so that a targeted persuasion or countermeasure can be taken.

Therefore, in this embodiment, a congestion propagation intensity sequence is further obtained to show features, that is, to indicate an evolution rule of congestion in an urban road network, and specifically, the method further includes:

according to the intersection P _i Congestion coefficient S (P) _i ，T _d ) And a predefined congestion level determination rule for said intersection P _i Matching the corresponding congestion level S ^deg (P _i ,T _d,k )；

Calculating the intersection P _i At T _d,k Time of day congestion propagation intensity sequence Q _i,j 。

In order to realize the calculation of the congestion propagation strength sequence, different traffic running states are depicted by utilizing the rich semantic expression capability of a knowledge graph, the transformation rule of the running states is extracted based on dynamic road network data (time sequence data), and finally the spatio-temporal evolution process of the congestion propagation relation is obtained.

Aiming at the problems: "extract the space-time evolution law of urban road network traffic jam and dissipation during the early peak period of 20 days in 3 months (6-9).

At the moment, the congestion coefficient s (P) of the intersection can be used _i ,T _d,k ) The congestion level s of the intersection can be defined ^deg (P _i ,T _d,k ). The congestion level determination rule predefined in the present embodiment is as follows: the congestion coefficient [1.0,1.5 ], the congestion level is 1, and the corresponding semantic meaning is light congestion; the congestion coefficient [1.5,2.0) ] is the congestion level 2, the corresponding semantic is moderate congestion, the congestion coefficient [2.0,2.5) ] is the congestion level 3, and the corresponding semantic is severe congestion; congestion coefficient [2.5, + ∞), the congestion level is 4, corresponding to a semantic of full congestion.

Calculating intersection P _i At T _d,k Intensity of congestion propagation in time, i.e. congestion propagation sequence Q _i,j The specific determination process is as follows:

congestion propagation relation label Tip _i,j =1, explain intersection P _i Spread to the intersection P _j ；

T _d,h Time crossing P _i Achieving congestion in a congestion spreading time interval range T _d,h ,T _d,h +W]Inner, T _d,hh Time crossing P _j Congestion is reached, at which time s (P) _j ,T _d,hh ) Congestion level s ^deg (P _j ,T _d,hh )；

In [ T ] _d,h ,T _d,h +W]Inside, there are multiple conforming intersections P _j T to reach congestion _d,hh And a plurality of corresponding congestion levels s ^deg (P _j ,T _d,hh )，{s ^deg (P _j ,T _d,hh1 )，s ^deg (P _j ,T _d,hh2 ) …, the congestion level with the largest number of times is taken out as the congestion propagation intensity q _i,j (T _d,h )；

In [ T ] _d,k-x ,T _d,k ]In which there are multiple coincidences of T _d,h Time, and a plurality of corresponding q _i,j (T _d,h ) Obtaining at T _d,k Time crossing P _i Spread to the intersection P _j Congestion propagation intensity sequence Q of _i,j ＝{q _i,j (T _d,h1 ),q _i,j (T _d,h2 ),…}。

Will cross P _i At T _d,k Time of day congestion propagation intensity sequence Q _i,j As a new feature.

Then at T _d,k For each crossing P _i And Q corresponding to the intersection _i,j And a congestion propagation intensity knowledge graph can be obtained. By adopting the TransE model, each intersection P can be divided into a plurality of intersections _i Re-projecting to 2D European space, each crossing P _i With vector representation (c) _i1 ,c _i2 ). This vector representation can be spliced together with the intersection vector representation found above to yield (x) _i1 ,x _i2 ,x _i3 ,…,x _in ,c _i1 ,c _i2 ). Then, the KNN algorithm can be used for crossing at T _d,k The similarity classification of (3). And with T _d,k Change of direction of each intersectionThe quantity representation will also change, resulting in the outcome of intersection similarity classification, which will also change over time. The congestion propagation strength, the law of change over time, can be determined from this.

In other embodiments, the method of the embodiment can construct a semantic search and knowledge matching framework of any traffic entity, including a task analysis module, a feature extraction module, a representation learning module, and a knowledge inference module; the task analysis module is used for performing natural language word segmentation, natural language semantic extraction, entity search task analysis and search keyword generation; the characteristic extraction module is used for matching space-time search keywords, matching entity attribute keywords, calculating and generating characteristics and generating a characteristic map; the representation learning module is used for embedding a knowledge map, calculating entity similarity, matching knowledge and searching entities and returning similar entities; the knowledge reasoning module is used for carrying out time sequence map reasoning and knowledge space-time evolution analysis.

Example 3

This embodiment of the present invention discloses a device corresponding to the method for updating data in a distributed graph database according to embodiment 1, which is a virtual structure device, and fig. 5 is a block diagram of a data updating device in a distributed graph database according to embodiment 3, as shown in fig. 5,

a graph database-based traffic behavior assessment apparatus comprising:

the data storage module 510 is used for storing a knowledge graph of the traffic operation data through a graph database, wherein the knowledge graph of the traffic operation data comprises entities, attributes and relationships among the entities of the traffic operation data;

a data extraction module 520, configured to obtain alarm and timing data of the target intersection, an entry road section of the target intersection, and road section speed data within a preset time period from the map database;

a data calculating module 530, configured to calculate the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection; calculating a cycle correlation coefficient of the target intersection according to the entrance road section and the road section speed data of the target intersection; calculating average regulation efficiency and average cycle correlation coefficient of a target area under different time resolutions according to the regulation efficiency of the target intersection and the cycle correlation coefficient of the target intersection;

and the comprehensive evaluation module 540 is configured to calculate a comprehensive operation index of the target area under different time resolutions according to a predefined importance ranking rule based on the average regulation efficiency and the average cycle correlation coefficient.

Example 4

Fig. 6 is a schematic structural diagram of an electronic device according to embodiment 4 of the present invention, and as shown in fig. 6, an electronic device is provided, where the electronic device may be a server, and its internal structural diagram may be as shown in fig. 6. The electronic device comprises a processor, a memory, an input device and an output device; wherein the number of processors in the electronic device may be one or more, and one processor is taken as an example in fig. 6; the processor, memory, input devices and output devices in the electronic apparatus may be connected by a bus or other means, and fig. 6 illustrates the connection by a bus as an example.

The memory, which is a computer-readable storage medium, may include a high-speed random access memory, a non-volatile memory, etc., which may be used to store an operating system, software programs, computer-executable programs, and databases, and may also include a memory, which may be used to provide a running environment for the operating system and the computer programs. The processor is used for providing calculation and control capability, and executing various functional applications and data processing of the electronic device by running computer-executable programs, software programs, instructions and modules stored in the memory, namely, implementing the map database-based traffic operation condition evaluation method of the embodiment 1-2.

The output device of the electronic equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on a shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

The electronic device may further include a network interface for communicating with an external terminal through a network connection. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

Those skilled in the art will appreciate that the configuration shown in fig. 6 is a block diagram of only a portion of the configuration associated with the present application, and does not constitute a limitation on the electronic device to which the present application is applied, and a particular electronic device may include more or less components than those shown in the drawings, or may combine certain components, or have a different arrangement of components.

It will be understood by those skilled in the art that all or part of the processes of the methods of the above embodiments 1-2 can be implemented by a computer program, which can be stored in a non-volatile computer readable storage medium, and can include the processes of the above embodiments of the methods when executed. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

Example 5

Embodiment 5 of the present invention also provides a storage medium containing computer-executable instructions, which when executed by a computer processor, are configured to implement a method for estimating traffic operating conditions based on a graph database, the method including:

and on the basis of the average regulation efficiency and the average cycle correlation coefficient, calculating the comprehensive operation index of the target area under different time resolutions through a predefined importance ranking rule.

Of course, the storage medium provided by the embodiment of the present invention contains computer-executable instructions, and the computer-executable instructions are not limited to the operations of the method described above, and may also perform related operations in the traffic operation condition assessment method based on the database provided by any embodiment of the present invention.

From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes instructions for enabling an electronic device (which may be a mobile phone, a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.

It should be noted that, in the embodiment of the method or the device for evaluating traffic operating conditions based on the graph database, the included units and modules are only divided according to the functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.

Various other modifications and changes may be made by those skilled in the art based on the above-described technical solutions and concepts, and all such modifications and changes should fall within the scope of the claims of the present invention.

Claims

1. A traffic running condition evaluation method based on a graph database is characterized by comprising the following steps:

calculating the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection comprises the following steps:

calculating the continuous alarm time length according to the alarm data of the target intersection, and respectively calculating the regulation alarm time length and the non-regulation alarm time length of the target intersection according to the timing data of the target intersection and the continuous alarm time length;

obtaining the regulation and control efficiency of the target intersection according to the ratio of the non-regulation and control alarm duration to the regulation and control alarm duration in a preset time period;

calculating the cycle correlation coefficient of the target intersection according to the inlet road section and the road section speed data of the target intersection comprises the following steps:

acquiring the free vehicle speed and the instantaneous vehicle speed of the target intersection at a preset moment according to the inlet road section and the road section speed data of the target intersection, and acquiring the congestion coefficient of the target intersection according to the ratio of the free vehicle speed and the instantaneous vehicle speed of the target intersection at the preset moment;

acquiring a congestion time sequence curve of the target intersection in a specific time period according to the congestion coefficient, and acquiring a periodic time sequence curve of the signal machine in the specific time period;

carrying out DTW _ distance calculation on the congestion time sequence curve and the periodic time sequence curve to obtain a periodic correlation coefficient of the target intersection;

2. The method for graph-based traffic behavior assessment according to claim 1, wherein said storing a knowledge-graph of traffic behavior data by a graph database comprises:

3. The graph database-based traffic operating condition assessment method according to claim 1, wherein said calculating a regulation efficiency of said target intersection based on said alarm and timing data of said target intersection comprises:

4. The graph database-based traffic behavior assessment method according to claim 3, wherein said calculating a periodic correlation coefficient for said target intersection based on said target intersection's entry road segment and road segment speed data comprises:

wherein L is _i,j Representing a target intersection P _i Flow direction P of traffic _j N denotes the number of entry links, { L _j1,i ,L _j2,i ,…,L _jn,i Denotes a set of inlet road sections, v ^free (P _i ,T _d,k ) Representing a target intersection P _i At T _d,k Free speed of the vehicle, v ^nostop Indicating non-stopping speed, v ^inst (P _i ,T _d,k ) Representing a target intersection P _i At T _d,k Instantaneous vehicle speed of time, s (P) _i ,T _d,k ) Representing a target intersection P _i At T _d,k Congestion coefficient of time, T _d,k Represents the kth velocity data acquisition time on day d;

S(P _i ，T _d )＝{s(P _i ，T _d,k ) Where k =1,2, …, NUM, T _d Represents the set of all speed data acquisition times on day d;

obtaining annunciator XH _i A cycle timing curve of the specific period: ZQ (XH) _i ,XT _d )＝{zq(XH _i ,XT _d,l ) Where i =1,2, …, NM, XT _d,l Denotes the l-th timing data acquisition time, XT, of day d _d Representing the set of all timing data acquisition times on day d, ZQ (XH) _i ,XT _d ) Representing annunciators XH _i A periodic time series curve of the specific period, zq (XH) _i ,XT _d,l ) Representing annunciators XH _i At XT _d,l A period of time acquisition;

extracting the congestion time sequence curve and the period time sequence curve corresponding to the time, and calculating the position of the target intersectionPeriodic correlation coefficient a (P) _i ，T _d ) Wherein

Wherein DTW _ distance (S (P)) _i ,T _d ),ZQ(XH _i ,XT _d ) Respectively fitting a congestion time sequence curve and a periodic time sequence curve, then performing time sampling, and extracting S (P) corresponding to time _i ，T _d ) And ZQ (XH) _i ,XT _d ) And (6) performing calculation.

5. The graph database-based traffic operating condition evaluation method according to claim 1 or 4, wherein the step of obtaining the comprehensive operating index of the target area at different time resolutions according to the predefined importance ranking rule based on the average regulation efficiency and the average cycle correlation coefficient comprises:

and according to a predefined importance ranking rule, carrying out normalization, hierarchical analysis and weighted summation processing on the average regulation efficiency and the average period correlation coefficient to obtain a comprehensive operation index of the target area under different space-time resolutions.

6. The method for graph database based assessment of traffic conditions according to claim 1, further comprising: screening of the target intersections:

the query task is received and the query task is received,

analyzing the query task to extract a search keyword of the query task;

extracting entities and attributes matched with the search keywords from the knowledge graph according to the search keywords, extracting features, and generating a feature graph;

7. The graph database-based traffic behavior assessment method according to claim 6, further comprising:

8. A graph database-based traffic behavior assessment apparatus, comprising:

the data calculation module is used for calculating the regulation and control efficiency of the target intersection according to the alarm and timing data of the target intersection, and comprises the following steps:

obtaining the regulation and control efficiency of the target intersection according to the ratio of the non-regulation and control alarm time length to the regulation and control alarm time length in a preset time period;

acquiring a congestion time sequence curve of the target intersection at a specific time period according to the congestion coefficient, and acquiring a periodic time sequence curve of the signaler at the specific time period;

carrying out DTW _ distance calculation on the congestion time sequence curve and the period time sequence curve to obtain a period correlation coefficient of the target intersection;

calculating the average regulation and control efficiency and the average cycle correlation coefficient of a target area under different time resolutions according to the regulation and control efficiency of the target intersection and the cycle correlation coefficient of the target intersection;

9. An electronic device comprising a processor, a storage medium, and a computer program stored in the storage medium, wherein the computer program, when executed by the processor, implements the graph database-based traffic behavior assessment method according to any one of claims 1 to 7.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method for graph-database-based assessment of traffic conditions according to any one of claims 1 to 7.