WO2020224445A1 - 一种车流路径分布信息的处理方法、装置及电子设备 - Google Patents

一种车流路径分布信息的处理方法、装置及电子设备 Download PDF

Info

Publication number
WO2020224445A1
WO2020224445A1 PCT/CN2020/086574 CN2020086574W WO2020224445A1 WO 2020224445 A1 WO2020224445 A1 WO 2020224445A1 CN 2020086574 W CN2020086574 W CN 2020086574W WO 2020224445 A1 WO2020224445 A1 WO 2020224445A1
Authority
WO
WIPO (PCT)
Prior art keywords
flow path
traffic flow
path distribution
distribution information
traffic
Prior art date
Application number
PCT/CN2020/086574
Other languages
English (en)
French (fr)
Inventor
张欣
茅嘉磊
杨磊
肖楠
贺亚静
Original Assignee
阿里巴巴集团控股有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阿里巴巴集团控股有限公司 filed Critical 阿里巴巴集团控股有限公司
Publication of WO2020224445A1 publication Critical patent/WO2020224445A1/zh

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • 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
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0133—Traffic data processing for classifying traffic situation

Definitions

  • This application relates to the field of traffic control technology, and more specifically, to a method for processing traffic flow path distribution information, a processing device for traffic flow path distribution information, an electronic device, and a computer-readable medium.
  • the future traffic demand of the traffic area can be obtained from the predicted traffic path distribution information.
  • the traffic flow path distribution information includes the path in the traffic area and the value of the traffic flow through each path. Among them, the path can be a sequence of road segments on the traffic road network.
  • the traffic path distribution in the target period may be related to the current traffic path distribution information in the traffic area, the traffic path distribution information around the traffic area, time, date, weather and other factors.
  • a regression model is usually established for each path in the traffic area, and the predicted traffic flow value of each path is obtained, and the predicted traffic flow value of each path is combined and superimposed to form a target time period for the traffic area. Prediction of traffic path distribution information.
  • this prediction method does not consider the overall factors of the preset area, which easily leads to the accumulation of errors, so that the combined traffic path distribution information may deviate greatly from the actual situation.
  • One purpose of the present application is to provide a new technical solution for predicting the distribution information of the traffic flow path in the preset area in the future.
  • a method for processing traffic flow path distribution information including:
  • the traffic flow path distribution information includes the path in the preset area and the value of the vehicle flow passing each path in the corresponding time period;
  • the traffic flow path distribution class includes a clustering result obtained by clustering the traffic flow path distribution information of at least one historical period;
  • the step of obtaining the predicted traffic flow path distribution information of the preset area in the target period includes:
  • the predicted traffic flow value of each path passing through the target time period obtain the predicted traffic flow path distribution information of the preset area in the target time period.
  • the step of obtaining the predicted traffic volume value passing the target path during the target time period includes:
  • the selected first feature vector includes a plurality of first features that affect the traffic flow value of the target path in the target period;
  • the plurality of first features include a first traffic feature and First environmental feature;
  • a predicted traffic volume value that passes through the target path in the target time period is obtained.
  • the first traffic feature includes at least one of traffic path distribution information in the preset area and traffic path distribution information around the preset area; and/or, the first environmental feature Including at least one of time, date, and weather.
  • the step of obtaining a first mapping function between the first feature vector and the value of the traffic volume passing through the target path includes:
  • the first mapping function is obtained by training according to the vector value of the first feature vector of the first training sample and the actual traffic flow value passing through the target path corresponding to the first training sample.
  • processing method further includes:
  • the method before acquiring the predicted traffic flow path distribution class corresponding to the target time period in the preset area, the method further includes:
  • the obtaining the predicted traffic flow path distribution class corresponding to the preset area in the target time period includes: obtaining the traffic flow path distribution class corresponding to the preset area in the target time period from the at least one vehicle flow path distribution class, as the Predict the distribution of traffic flow path.
  • the step of clustering multiple pieces of historical traffic flow path distribution information to obtain at least one of the traffic flow path distribution classes includes:
  • each historical vehicle flow path determines the value of the vehicle flow through each path in the corresponding historical period
  • clustering multiple pieces of historical traffic flow path distribution information to obtain at least one of the traffic flow path distribution classes.
  • the step of determining the distance between every two pieces of historical traffic flow path distribution information according to the number of road sections included in each path and the traffic flow value passing through the corresponding path in each historical period includes:
  • the road section flow sum corresponding to each historical traffic flow path distribution information; wherein, the road section flow sum is within the corresponding time period The sum of traffic flow values passing through each section;
  • the road flow difference corresponding to each two historical traffic path distribution information is determined; wherein, the road flow difference corresponds to two The sum of the difference in the value of the traffic volume passing through each road section during the time period;
  • the step of clustering multiple pieces of historical traffic flow path distribution information according to the distance between every two pieces of historical traffic flow path distribution information to obtain at least one of the traffic flow path distribution classes includes:
  • the step of using each vehicle flow path distribution information as a node, and constructing a relationship graph according to the distance between every two nodes includes:
  • each vehicle flow path distribution information as a node, respectively connecting each node and a set number of nodes with the closest distance to itself, to obtain the relationship graph.
  • the step of splitting the relationship graph into at least one subgraph according to the distance between every two nodes includes:
  • connection between two nodes whose distance exceeds a preset distance threshold is cut off to split the relationship graph into at least one subgraph.
  • the step of obtaining the predicted traffic flow path distribution class corresponding to the preset area in the target period includes:
  • the selected second feature vector includes a plurality of second features that affect the distribution of traffic flow paths corresponding to the preset area at the target time period; the plurality of second features include second Traffic characteristics and second environmental characteristics;
  • the traffic flow path distribution class corresponding to the preset area in the target period is obtained as the predicted traffic flow path distribution class.
  • the second traffic feature includes at least one of traffic flow path distribution information in the preset area and traffic flow path distribution information around the preset area; and/or, the second environment feature Including at least one of time, date, and weather.
  • the step of obtaining a second mapping function between the second feature vector and the traffic flow path distribution class includes:
  • the second mapping function is obtained by training according to the vector value of the second feature vector of the second training sample and the traffic path distribution class actually corresponding to the second training sample.
  • processing method further includes:
  • the step of modifying the predicted traffic flow path distribution information according to the predicted traffic flow path distribution class so that the revised predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution class includes:
  • Target vehicle flow path distribution information representing the cluster center of the predicted vehicle flow path distribution class
  • the predicted traffic flow path distribution information is corrected so that the corrected predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution category.
  • the predicted traffic flow path distribution information is corrected according to the distance between the predicted traffic flow path distribution information and the target traffic flow path distribution information, so that the corrected predicted traffic flow path distribution information belongs to the predicted
  • the steps of the traffic path distribution category include:
  • the step of determining target traffic flow path distribution information representing the cluster center of the predicted traffic flow path distribution class includes:
  • the target vehicle flow path distribution information is obtained.
  • processing method further includes:
  • traffic control is performed on the preset area.
  • an apparatus for processing traffic flow path distribution information including:
  • the distribution information prediction module is used to obtain the predicted traffic flow path distribution information of the preset area in the target time period;
  • the traffic flow path distribution information includes the path in the preset area and the flow of vehicles passing through each path in the corresponding time period value;
  • the distribution prediction module is used to obtain the predicted traffic flow path distribution category corresponding to the preset area in the target period; wherein the traffic flow path distribution category includes at least one historical period of traffic path distribution information;
  • the distribution information correction module is configured to correct the predicted traffic flow path distribution information according to the predicted traffic flow path distribution class, so that the corrected predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution class.
  • an electronic device including the processing device according to the second aspect of the present application; or, including a processor and a memory, the memory is used to store executable instructions, the instructions For controlling the processor to execute the processing method according to the first aspect of the present application.
  • a computer-readable storage medium having a computer program stored thereon, and the computer program, when executed by a processor, implements the processing method according to the first aspect of the present application.
  • FIG. 1 is a block diagram of an example of the hardware configuration of an electronic device that can be used to implement the embodiments of the present application.
  • FIG. 2 is a block diagram of another example of the hardware configuration of an electronic device that can be used to implement the embodiments of the present application;
  • Fig. 3 is a flowchart of a method for processing traffic flow path distribution information according to an embodiment of the present application
  • FIG. 4 is a schematic diagram of an example of a preset area according to an embodiment of the present application.
  • FIG. 5 is a schematic diagram of an example of traffic flow path distribution information in a first historical period according to an embodiment of the present application
  • FIG. 6 is a schematic diagram of an example of traffic flow path distribution information in a second historical period according to an embodiment of the present application
  • FIG. 7 is a schematic diagram of an example of traffic flow path distribution information in a third historical period according to an embodiment of the present application.
  • FIG. 8 is a schematic flowchart of an example of a method for processing traffic flow path distribution information according to an embodiment of the present application.
  • Fig. 9 is a schematic diagram of a device for processing traffic flow path distribution information according to an embodiment of the present application.
  • Fig. 10 is a functional block diagram of an electronic device provided according to a first embodiment of the present application.
  • Fig. 11 is a schematic diagram of the hardware structure of an electronic device according to a second embodiment of the present application.
  • FIG. 1 and 2 are block diagrams of the hardware configuration of an electronic device 1000 that can be used to implement the method for processing traffic flow path distribution information of any embodiment of the present application.
  • the electronic device may be a server 1100.
  • the server 1100 provides service points for processing, database, and communication facilities.
  • the server 1100 may be an integrated server or a distributed server that spans multiple computers or computer data centers.
  • the server can be of various types, such as, but not limited to, a web server, a news server, a mail server, a message server, an advertisement server, a file server, an application server, an interactive server, a database server, or a proxy server.
  • each server may include hardware, software, or embedded logic components or a combination of two or more such components for performing appropriate functions supported or implemented by the server.
  • the server may be a blade server, a cloud server, etc., or may be a server group composed of multiple servers, and may include one or more of the foregoing types of servers, and so on.
  • the server 1100 may be as shown in FIG. 1 and includes a processor 1110, a memory 1120, an interface device 1130, a communication device 1140, a display device 1150, and an input device 1160.
  • the server 1100 may also include a speaker, a microphone, etc., which are not limited herein.
  • the processor 1110 may be a dedicated server processor, or may be a desktop processor or a mobile processor that meets performance requirements, and is not limited herein.
  • the memory 1120 includes, for example, ROM (Read Only Memory), RAM (Random Access Memory), nonvolatile memory such as a hard disk, and the like.
  • the interface device 1130 includes, for example, various bus interfaces, such as a serial bus interface (including a USB interface), a parallel bus interface, and the like.
  • the communication device 1140 can perform wired or wireless communication, for example.
  • the display device 1150 is, for example, a liquid crystal display, an LED display touch screen, or the like.
  • the input device 1160 may include, for example, a touch screen, a keyboard, and the like.
  • the memory 1120 of the server 1100 is used to store instructions, which are used to control the processor 1110 to operate to at least execute the method for processing vehicle flow path distribution information according to any embodiment of the present application.
  • Technicians can design instructions according to the scheme disclosed in this application. How the instruction controls the processor to operate is well known in the art, so it will not be described in detail here.
  • the server 1100 only involves some of the devices.
  • the server 1100 only involves the memory 1120 and the processor 1110.
  • the electronic device may be a terminal device 1200 such as a PC or a notebook computer used by an operator, which is not limited herein.
  • the terminal device 1200 may include a processor 1210, a memory 1220, an interface device 1230, a communication device 1240, a display device 1250, an input device 1260, a speaker 1270, a microphone 1280, and so on.
  • the processor 1210 may be a mobile version processor.
  • the memory 1220 includes, for example, ROM (Read Only Memory), RAM (Random Access Memory), nonvolatile memory such as a hard disk, and the like.
  • the interface device 1230 includes, for example, a USB interface, a headphone interface, and the like.
  • the communication device 1240 can, for example, perform wired or wireless communication.
  • the communication device 1240 may include short-range communication devices, such as based on Hilink protocol, WiFi (IEEE 802.11 protocol), Mesh, Bluetooth, ZigBee, Thread, Z-Wave, NFC, UWB, Any device that performs short-range wireless communication with a short-range wireless communication protocol such as LiFi.
  • the communication device 1240 may also include a remote communication device, for example, any device that performs WLAN, GPRS, 2G/3G/4G/5G remote communication.
  • the display device 1250 is, for example, a liquid crystal display, a touch display, or the like.
  • the input device 1260 may include, for example, a touch screen, a keyboard, and the like. The user can input/output voice information through the speaker 1270 and the microphone 1280.
  • the memory 1220 of the terminal device 1200 is used to store instructions, which are used to control the processor 1210 to operate to at least execute the method for processing vehicle flow path distribution information according to any embodiment of the present application.
  • Technicians can design instructions according to the scheme disclosed in this application. How the instruction controls the processor to operate is well known in the art, so it will not be described in detail here.
  • the present application may only involve some of the devices.
  • the terminal device 1200 only involves the memory 1220, the processor 1210, and the display device 1250.
  • a method for processing traffic flow path distribution information is provided.
  • the processing method may be implemented by electronic equipment.
  • the electronic device may be the server 1100 shown in FIG. 1 or the terminal device 1200 shown in FIG. 2.
  • the method for processing traffic flow path distribution information of this embodiment may include the following steps S1000 to S3000:
  • Step S1000 Obtain the predicted traffic flow path distribution information of the preset area in the target period.
  • the target time period can be a time period in the future or a time period in the past.
  • the preset area in this embodiment may be a traffic area selected in the city according to application scenarios or specific needs.
  • the preset area may be as shown in the preset area in FIG. 4, or may be as shown in the preset area in FIGS. 5-7.
  • the predicted traffic flow path distribution information may include a path in a preset area and a predicted traffic flow value that passes through each path within a target time period.
  • the path in this embodiment may be a sequence of road sections on the road network, and the road section may refer to a traffic line in a driving direction between two adjacent intersections on the transportation network.
  • the paths in the preset area acquired through step S1000 may include: path A, path B, path C, path D, path E, and path F .
  • the step of obtaining the predicted traffic flow path distribution information of the preset area in the target period may include the following steps S1100 to S1300:
  • Step S1100 Obtain a path in a preset area.
  • Step S1200 Obtain the predicted traffic volume value of each path in the target time period.
  • Obtaining the predicted traffic volume value of each path in the target time period may be specifically: taking each path as the target path in turn, and obtaining the predicted traffic volume value of the target path in the target time period.
  • the time period may be set according to application scenarios or specific needs, and the duration of each time period (including the target time period, current time period, and historical time period described in this embodiment) is equal.
  • the period between every two adjacent whole hours can be regarded as a period.
  • 8-9 o'clock on the current date can be used as the current period
  • 9-10 o'clock can be used as the target period.
  • the step of obtaining the predicted traffic volume value passing the target route during the target time period may include the following steps S1210 to S1230:
  • Step S1210 Acquire the selected first vector feature.
  • the first feature vector may include a plurality of first features that affect the traffic flow value of the target path during the target period.
  • the plurality of first features includes a first traffic feature and a first environmental feature.
  • the first traffic feature may include at least one of traffic flow path distribution information in the preset area and traffic flow path distribution information around the preset area.
  • the first environmental feature may include at least one of time, date, and weather.
  • x j may be a first feature such as a first traffic feature, a first environmental feature, etc., which can affect the traffic flow value of the target path during the target period.
  • the first traffic characteristic may be traffic path distribution information in a preset area and traffic flow path distribution information around the preset area.
  • the first environmental characteristic may be time, date, and weather.
  • the feature vector X may also include other features related to the traffic flow value.
  • Step S1220 Obtain a first mapping function between the first feature vector and the value of the traffic volume passing through the target path.
  • the independent variable of the first mapping function F1(x) is the feature vector X
  • the dependent variable F1(x) is the predicted traffic flow value determined by the feature vector X.
  • obtaining the first mapping function between the first feature vector and the traffic flow value of the target path in step S1220 may further include the following steps S1221 to S1222:
  • Step S1221 Obtain a first training sample according to the historical traffic trajectory.
  • each first training sample includes a traffic trajectory matching the target path.
  • Step S1222 training to obtain a first mapping function according to the vector value of the first feature vector of the first training sample and the actual traffic volume value passing through the target path corresponding to the first training sample.
  • steps S1221 to S1222 of training the first mapping function may be performed according to a preset training period.
  • the training period can be set according to specific application scenarios or application requirements, for example, can be set to 1 day.
  • the first mapping function F(x) can be obtained by various fitting methods.
  • any multiple linear regression model can be used to obtain the first mapping function F1(x), which is not limited here.
  • the multiple linear regression model can be a polynomial function that simply reflects the first mapping function F1(x), where the coefficients of each order of the polynomial function are unknown, and the first feature vector of the first training sample The vector value of, and the actual vehicle flow value through the target path corresponding to the first training sample are substituted into the polynomial function to determine the coefficients of each order of the polynomial function, and then obtain the first mapping function F1(x).
  • various regression models can be used to use the first vector value of the first feature vector of the first training sample and the actual traffic volume value of the target path corresponding to the first training sample.
  • each round learns the residuals after the previous round of fitting, and iterates through T rounds to control the residuals to a very low value, so that the first mapping function F1( x) Has a very high accuracy.
  • the addition model is, for example, LightGBM, GBDT, XGBoost, etc., which are not limited here.
  • Step S1230 according to the first mapping function and the vector value of the first feature vector in the current period, obtain the predicted traffic volume value of the target path in the target period.
  • the vector value can be substituted into the first feature vector according to the vector value of the first feature vector in the current period.
  • Mapping function F(x) in order to obtain the value of the predicted traffic volume passing the target path during the target period.
  • the present application can obtain the predicted traffic flow value of the target path during the target period according to the first feature vector and the first mapping function. Since the first mapping function is trained based on a large number of training samples, the When a mapping function determines the predicted traffic flow value, the accuracy of the obtained predicted traffic flow value can be improved.
  • Step S1300 according to the predicted traffic flow value of each path in the next time period, obtain the predicted traffic flow path distribution information of the preset area in the next time period.
  • it may be to integrate the predicted traffic flow value of each path in the next time period to obtain the predicted traffic flow path distribution information of the preset area in the next time period.
  • Step S2000 Obtain the predicted traffic flow path distribution class corresponding to the preset area in the target period.
  • the traffic flow path distribution category includes clustering results obtained by clustering the traffic flow path distribution information of the preset area in at least one historical period.
  • clustering can be performed on the traffic flow path distribution information of at least one historical period in the preset area to obtain at least one traffic flow path distribution class.
  • Each traffic flow path distribution class can be the traffic flow path of all historical periods belonging to the cluster.
  • the collection of distribution information can also be the unique identifier of the corresponding cluster. Through the unique identification of the cluster, the traffic flow path distribution information of all historical periods belonging to the corresponding cluster can be uniquely determined. Wherein, the identification can be composed of at least one character.
  • the predicted traffic flow path distribution class corresponding to the target period may be one of at least one traffic flow path distribution class obtained by clustering.
  • the traffic flow path distribution information may include the paths in the preset area and the value of the flow of vehicles passing through each path in the corresponding time period.
  • FIGS. 5-7 are schematic diagrams of the traffic flow path distribution information of the preset area in FIG. 4 in different historical periods.
  • step S1000 and step S2000 may be executed simultaneously, and step S1000 may be executed first and then step S2000 may be executed, or step S2000 may be executed first and then step S1000 may be executed.
  • This application does not specifically limit the execution sequence of step S1000 and step S2000.
  • the processing method may further include the step of obtaining at least one vehicle flow path distribution class corresponding to the preset area, so as to obtain the vehicle flow path corresponding to the preset area in the target time period in step S2000
  • the distribution class is used as the distribution class of the predicted traffic flow path.
  • the predicted traffic flow path distribution class in this embodiment is one of the acquired at least one traffic flow path distribution class corresponding to the preset area.
  • the step of obtaining at least one vehicle flow path distribution category may further include the following steps S6100 to S6200:
  • Step S6100 Obtain the traffic flow path distribution information of the preset area in multiple historical time periods as historical traffic flow path distribution information.
  • FIG. 5 may be a schematic diagram of traffic flow path distribution information in a first historical period
  • FIG. 6 may be a schematic diagram of vehicle flow path distribution information in a second historical period
  • FIG. 7 may be a vehicle flow path distribution information in a third historical period. Schematic diagram.
  • the values of the traffic flow through path A, path B, path C, path D, path E, and path F can be obtained in the first historical period. It is 300, 168, 270, 156, 0, 0.
  • the values of the traffic flow through path A, path B, path C, path D, path E, and path F in the second historical period can be obtained respectively It is 340, 168, 270, 0, 227, 0.
  • the values of the traffic flow through path A, path B, path C, path D, path E, and path F can be obtained in the third historical period. It is 0, 168, 270, 0, 227, 100.
  • Step S6200 clustering multiple historical traffic flow path distribution information to obtain at least one traffic flow path distribution class.
  • clustering a plurality of historical traffic flow path distribution information to obtain at least one traffic flow path distribution category may further include steps S6210 to S6240 as shown below:
  • Step S6210 Determine the number of road sections included in each path.
  • the road section in this embodiment is a traffic line between two adjacent intersections in the corresponding path, and the number of road sections is the number of road sections included in the corresponding path.
  • the number of road segments included in path A, path B, and path F are all 5, and the number of road segments included in path C, path D, and path E are all 3. .
  • Step S6220 according to the distribution information of each historical traffic flow path, determine the traffic flow value of each path in the corresponding historical time period.
  • Step S6230 Determine the distance between every two pieces of historical traffic flow path distribution information according to the number of road sections included in each path and the traffic flow value passing through the corresponding path in each historical period.
  • the distance between every two pieces of historical traffic flow path distribution information can be used to characterize the degree of difference between corresponding two pieces of historical traffic flow path distribution information.
  • determining the distance between every two pieces of historical traffic flow path distribution information may include the following steps S6231 to S6233 according to the number of road sections included in each path and the traffic flow value passing through the corresponding path in each historical period.
  • Step S6231 according to the number of road sections included in each path and the value of the traffic flow passing through the corresponding path in each historical period, determine the road section flow sum corresponding to each historical traffic flow path distribution information.
  • the sum of the road section flow is the sum of the value of the traffic flow through each section in the corresponding time period.
  • the traffic flow path distribution information corresponding to the road section flow sum is f 1 , where:
  • the second historical period can be obtained.
  • the traffic flow path distribution information corresponding to the road section flow sum is f 2 , where:
  • the third historical period can be obtained.
  • the traffic flow path distribution information corresponding to the road section flow sum is f 3 , where:
  • Step S6232 according to the number of road sections included in each path and the traffic flow value of the corresponding path in each historical period, determine the road flow difference corresponding to each two historical traffic flow path distribution information.
  • the difference in road flow is the sum of the difference in the value of the flow of vehicles passing through each road section in the corresponding two time periods.
  • the sum of the difference of the traffic flow values corresponding to all sections of each path in the two time periods may be determined by first determining the difference in the traffic flow values corresponding to the path in the two time periods, and then passing through the path. This difference is obtained by multiplying the number of road sections included.
  • the first The value of the traffic flow through the corresponding path in the historical period and the value of the traffic flow through the corresponding path in the second historical period can determine the correspondence between the traffic flow path distribution information in the first historical period and the traffic flow path distribution information in the second historical period Poor flow in the road section.
  • the sum of the difference between the historical traffic flow of each section of the route A in the first historical period and the second historical period is (340-300)*5, the first historical period and the second historical period
  • the sum of the difference of the historical traffic flow of each section of route B in the historical period is (168-168)*5, the historical traffic flow of each section of route C in the first historical period and the second historical period
  • the sum of the difference of is (270-270)*3, the sum of the difference of the historical traffic flow of each section of path D in the first historical period and the second historical period is (156-0)*3
  • the sum of the difference between the historical traffic flow of each section of the path E in the first historical period and the second historical period is (227-0)*3, the path F in the first historical period and the second historical period
  • the sum of the difference of the historical traffic flow of each road section in is (0-0)*5.
  • the first The value of the traffic flow through the corresponding path in the historical period and the value of the traffic flow through the corresponding path in the third historical period can determine the correspondence between the traffic path distribution information in the first historical period and the traffic path distribution information in the third historical period. Poor flow in the road section.
  • the sum of the difference between the historical traffic flow of each road section in path A in the first historical period and the third historical period is (340-0)*5, and the difference between the first historical period and the third historical period
  • the sum of the difference between the historical traffic flow of each section of path B in the historical period is (168-168)*5, the historical traffic flow of each section of path C in the first historical period and the third historical period
  • the sum of the difference of is (270-270)*3, the sum of the difference of the historical traffic flow of each section of path D in the first historical period and the third historical period is (156-0)*3
  • the sum of the difference between the historical traffic flow of each section of the path E in the first historical period and the third historical period is (227-0)*3, the path F in the first historical period and the third historical period
  • the sum of the difference of the historical traffic flow of each road section in is (100-0)*5.
  • the second The value of the traffic flow through the corresponding path in the historical period and the value of the traffic flow through the corresponding path in the third historical period can determine the correspondence between the traffic flow path distribution information in the second historical period and the traffic flow path distribution information in the third historical period Poor flow in the road section.
  • the sum of the difference between the historical traffic flow of each road section in path A in the second historical period and the third historical period is (340-0)*5, and the second historical period and the third historical period
  • the sum of the difference of the historical traffic flow of each section of route B in the historical period is (168-168)*5, the historical traffic flow of each section of route C in the second historical period and the third historical period
  • the sum of the difference of is (270-270)*3, the sum of the difference of the historical traffic flow of each section of path D in the second historical period and the third historical period is (0-0)*3
  • the sum of the difference between the historical traffic flow of each section of the path E in the second historical period and the third historical period is (227-227)*3, the path F in the second historical period and the third historical period
  • the sum of the difference of the historical traffic flow of each road section in is (100-0)*5.
  • Step S6233 Determine the distance between every two pieces of historical traffic flow path distribution information according to the sum of the link flow corresponding to each piece of historical traffic flow path distribution information and the link flow difference corresponding to every two pieces of historical traffic flow path distribution information.
  • the method for determining the distance between every two pieces of historical traffic flow path distribution information may be: determining the geometric mean value of the sum of the section flows corresponding to the two historical traffic flow path distribution information, and then calculating the corresponding two historical traffic flow paths The ratio between the traffic difference of the road section corresponding to the distribution information and the geometric mean value is used as the distance between the corresponding two historical traffic flow path distribution information.
  • the vehicle flow path distribution in the first historical period can be determined first
  • the geometric mean value of the sum of the traffic flow corresponding to the traffic flow path distribution information in the second historical period is Then calculate the traffic flow path distribution information in the first historical period and the traffic flow path distribution information in the second historical period corresponding to the section flow difference ⁇ f 12 and the geometric mean value The ratio between, obtains the distance d 12 between the traffic flow path distribution information in the first historical period and the traffic flow path distribution information in the second historical period, where:
  • the vehicle flow path distribution in the first historical period can be determined first
  • the geometric mean of the sum of the traffic flow corresponding to the traffic flow path distribution information in the third historical period is Then calculate the traffic flow path distribution information in the first historical period and the traffic flow path distribution information in the third historical period corresponding to the section flow difference ⁇ f 13 and the geometric mean value The ratio between, obtains the distance d 13 between the traffic flow path distribution information in the first historical period and the traffic flow path distribution information in the third historical period, where:
  • the vehicle flow path distribution in the second historical period can be determined first
  • the geometric mean of the sum of the traffic flow corresponding to the traffic flow path distribution information in the third historical period is Then calculate the traffic flow path distribution information in the second historical period and the traffic flow path distribution information in the third historical period corresponding to the road flow difference ⁇ f 23 and the geometric mean value The ratio between, obtains the distance d 23 between the traffic flow path distribution information in the second historical period and the traffic flow path distribution information in the third historical period,
  • Step S6240 clustering multiple pieces of historical traffic flow path distribution information according to the distance between every two pieces of historical traffic flow path distribution information to obtain at least one traffic flow path distribution class.
  • the clustering method adopted in this embodiment may be any one or more of systematic clustering method, ordered sample clustering method, dynamic clustering method, fuzzy clustering method, and graph theory clustering method.
  • the specific clustering method is not limited.
  • the distance between the historical traffic flow path distribution information belonging to the same traffic flow path distribution category can be made smaller, and the historical traffic flow path distribution information belonging to different traffic flow path distribution categories The distance between them is far.
  • the step of clustering a plurality of historical traffic flow path distribution information according to the distance between every two historical traffic flow path distribution information to obtain at least one traffic flow path distribution category may include the steps S6241 to S6243 as shown below :
  • each traffic path distribution information is taken as a node, and a relationship graph is constructed according to the distance between every two nodes.
  • each vehicle flow path distribution information may be used as a node, and every two nodes may be connected to obtain the relationship graph.
  • each vehicle flow path distribution information may be used as a node, and each node and a set number of nodes with the closest distance to itself are respectively connected to obtain the relationship graph.
  • the set number can be set in advance according to the application scenario or specific needs.
  • the set number can be set to 5, so it can be the nearest 5 for each node and its distance.
  • the nodes are connected to get the relationship graph.
  • Step S6242 Split the relationship graph into at least one subgraph according to the distance between every two nodes.
  • connection between two nodes whose distance exceeds a preset distance threshold may be cut off, so as to split the relationship graph into at least one subgraph.
  • the graph segmentation method is applied to split the relationship graph into at least one subgraph, so that the sum of the distance between every two nodes in the same subgraph is the smallest, and every two nodes in different subgraphs The sum of the distances between nodes is the largest.
  • the graph segmentation method used in this example can be a minimum segmentation method, or a Normalized Cut, etc., which is not limited here.
  • step S6243 a one-to-one correspondence to each sub-graph is obtained, and the vehicle flow path distribution information corresponding to the nodes contained in each sub-graph is divided into corresponding vehicle flow path distribution classes.
  • the second subgraph corresponds to traffic path distribution category 2 one-to-one, and divides the traffic path distribution information corresponding to node 1 and node 2 into traffic path distribution category 1, and divides the traffic path distribution information corresponding to node 3 and node 4 Divided into traffic path distribution category 2.
  • one of the obtained at least one vehicle flow path distribution class can be used as the predicted vehicle flow path distribution class according to the obtained vehicle flow path distribution information in the target period of the preset area.
  • Obtaining the traffic flow path distribution class corresponding to the preset area in the target time period, as the step of predicting the traffic flow path distribution class, may further include steps S2100 to S2300 as shown below:
  • Step S2100 Obtain the selected second feature vector.
  • the second feature vector may include a plurality of second features that affect the distribution of the traffic flow path corresponding to the preset area in the target period.
  • the plurality of second features includes a second traffic feature and a second environmental feature.
  • the second traffic characteristic may include at least one of traffic flow path distribution information in the preset area and traffic flow path distribution information around the preset area.
  • the second environmental feature may include at least one of time, date, and weather.
  • y j may be a second feature such as a second traffic feature, a second environmental feature, etc., which can affect the distribution of the traffic flow path corresponding to the preset area at the target time period.
  • the second traffic characteristic may be traffic path distribution information in a preset area and traffic flow path distribution information around the preset area
  • the second environmental characteristic may be time, date, and weather.
  • the feature vector Y may also include other features related to the corresponding traffic flow path distribution class.
  • Step S2200 Obtain a second mapping function between the second feature vector and the vehicle flow path distribution class.
  • the independent variable of the second mapping function F2(y) is the feature vector Y
  • the dependent variable F2(y) is the predicted traffic flow path distribution class determined by the feature vector Y.
  • obtaining the second mapping function between the second feature vector and the vehicle flow path distribution class in step S2200 may further include the following steps S2210 to S2220:
  • Step S2210 Use the historical traffic trajectory as the second training sample.
  • Step S2220 training to obtain a second mapping function according to the vector value of the second feature vector of the second training sample and the traffic path distribution class actually corresponding to the second training sample.
  • the steps S2210 to S2220 of training the second mapping function may be performed according to a preset training period.
  • the training period can be set according to specific application scenarios or application requirements, for example, can be set to 1 day.
  • the second mapping function F2(y) can be obtained by various fitting methods.
  • any multiple linear regression model can be used to obtain the second mapping function F2(y), which is not limited here.
  • the multiple linear regression model can be a polynomial function that simply reflects the second mapping function F2(y), where the coefficients of each order of the polynomial function are unknown, and the second feature vector of the second training sample The vector value of and the traffic path distribution class actually corresponding to the second training sample are substituted into the polynomial function to determine the coefficients of each order of the polynomial function, thereby obtaining the second mapping function F2(y).
  • various regression models such as additive models, can be used, and the second vector value of the second feature vector of the second training sample and the traffic path distribution class actually corresponding to the second training sample can be used as accurate samples.
  • Multiple rounds of training each round learns the residuals after the previous round of fitting, iterating T rounds, the residuals can be controlled to a very low value, so that the final second mapping function F2(y) has a very High accuracy.
  • the addition model is, for example, LightGBM, GBDT, XGBoost, etc., which are not limited here.
  • Step S2300 according to the second mapping function and the vector value of the second feature vector in the current period, obtain the traffic flow path distribution class corresponding to the preset area in the target period as the predicted traffic flow path distribution class.
  • the vector value can be substituted into the second mapping function F according to the vector value of the second feature vector in the current period. (x) in order to obtain the predicted traffic flow path distribution class actually corresponding to the preset area in the target period.
  • the present application can obtain the predicted traffic flow path distribution class corresponding to the preset area in the target time period according to the second feature vector and the second mapping function. Since the second mapping function is obtained through training based on a large number of training samples, when the second mapping function is used to determine the predicted traffic flow path distribution class, the accuracy of the obtained predicted traffic flow path distribution class can be improved.
  • step S1000 and step S2000 continue to perform the following step S3000.
  • Step S3000 Correct the predicted traffic flow path distribution information according to the predicted traffic flow path distribution class, so that the corrected predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution class.
  • the predicted traffic flow path distribution class is a collection of traffic flow path distribution information of all historical periods belonging to the corresponding cluster
  • the predicted traffic flow path may be revised according to the traffic flow path distribution information of all historical periods belonging to the predicted traffic flow path distribution class Distribution information.
  • the predicted traffic flow path distribution class is the unique identifier of the corresponding cluster
  • the unique identification can be used to determine the traffic flow path distribution information of all historical periods belonging to the predicted traffic flow path distribution class, and then based on the predicted traffic flow path distribution class
  • the traffic flow path distribution information of all historical periods is used to modify the predicted traffic flow path distribution information.
  • correcting the predicted traffic flow path distribution information according to the predicted traffic flow path distribution class may include steps S3100 to S3300:
  • Step S3100 Determine target traffic flow path distribution information representing the cluster center of the predicted traffic flow path distribution class.
  • the target traffic flow path distribution information may not be the traffic flow path distribution information of any historical period included in the predicted traffic flow path distribution class, but is obtained based on the traffic flow path distribution information of all historical periods included in the predicted traffic flow path distribution class. It can represent the traffic path distribution information of the cluster center of the predicted traffic path distribution class.
  • the step of determining target traffic flow path distribution information may include steps S3110 to S3130:
  • Step S3110 Determine the optimization between the target vehicle flow value of each path and the index for measuring the cluster center according to the vehicle flow value of each path in the vehicle flow path distribution information contained in the predicted vehicle flow path distribution class. function.
  • the index for measuring cluster centers corresponding to each path may be the sum of the squares of the difference between the target traffic volume value passing through the corresponding path and the traffic volume value passing through each corresponding path.
  • step S3120 according to the optimization function corresponding to each path, the target vehicle flow value through each path when the index for measuring the cluster center is the smallest is determined.
  • the heuristic solver Louvain algorithm may be used to solve the optimization function corresponding to each path to obtain the target traffic volume value passing through each path.
  • Step S3130 Obtain target vehicle flow path distribution information according to the target vehicle flow value passing through each path.
  • the target vehicle flow value passing through each path may be integrated to obtain the target vehicle flow path distribution information.
  • Step S3200 Determine the distance between the predicted traffic flow path distribution information and the target traffic flow path distribution information.
  • Step S3300 Correct the predicted traffic flow path distribution information according to the distance between the predicted traffic flow path distribution information and the target traffic path distribution information, so that the corrected predicted traffic path distribution information belongs to the predicted traffic flow path distribution class.
  • the method of correcting the predicted traffic flow path distribution information according to the distance between the predicted traffic flow path distribution information and the target traffic flow path distribution information may make the distance between the corrected predicted traffic flow path distribution information and the target traffic path distribution information smaller.
  • the step of correcting the predicted vehicle flow path distribution information according to the distance between the predicted vehicle flow path distribution information and the target vehicle flow path distribution information, so that the corrected predicted vehicle flow path distribution information belongs to the predicted vehicle flow path distribution category may include The following steps S3310 ⁇ S3330:
  • Step S3310 Determine the distance between each traffic path distribution information included in the predicted traffic flow path distribution class and the target traffic path distribution information.
  • Step S3320 Determine the maximum value of the distance between each vehicle flow path distribution information and the target vehicle flow path distribution information.
  • step S3330 the predicted traffic flow path distribution information is corrected so that the distance between the corrected predicted traffic flow path distribution information and the target vehicle flow path distribution information is less than or equal to the maximum value.
  • the distance between the corrected predicted vehicle flow path distribution information and the target vehicle flow path distribution information is less than or equal to the maximum value, it can indicate that the corrected predicted vehicle flow path distribution information belongs to the predicted vehicle flow path distribution class.
  • the processing method may further include: determining whether the predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution category, if so, there is no need to modify the predicted traffic flow path distribution information according to the predicted traffic flow path distribution category; if not; , Step S3000 is executed to modify the predicted traffic flow path distribution information according to the predicted traffic flow path distribution class, so that the corrected predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution class.
  • the processing method of this embodiment to predict the traffic path distribution information of the preset area target period, the finally obtained revised predicted traffic path distribution information can be more accurate.
  • traffic control may be performed on the preset area according to the corrected predicted traffic path distribution information. Specifically, it can help the traffic manager to decide in advance the signal light optimization control plan based on the revised predicted traffic path distribution information. In this way, it can help traffic managers to specify or adjust traffic management plans more actively, and improve the quality of traffic management and decision-making efficiency.
  • the specific method of performing traffic control on the preset area may include: signal cycle duration of the signal lights in the preset area, green signal ratio of at least one phase, and phase difference of multiple intersections in at least one phase. At least one item is controlled accordingly.
  • the phase in this embodiment has a well-known meaning in the industry. For example, it may include, within a signal period, a sequence of signal states of one or several traffic flows with the same signal light color display is called a phase. The phase is divided according to the time sequence of the signal display obtained by the traffic flow. There are as many phases as there are different sequence arrangements. Each control state corresponds to a group of different lamp color combinations, which is called a phase. In short, a phase is also called a control state. For another example, for a group of traffic flows that do not conflict with each other and obtain the signal display state corresponding to the right of way at the same time, it can be called a phase. It can be seen that the phase is divided according to the change of the right of way of the intersection within a signal cycle.
  • the signal cycle time including the signal light change, the time required for the signal to run a cycle, is equal to the sum of the green, yellow, and red light time; also equal to the sum of the green light time and the yellow light time (usually fixed) required for all phases .
  • the green signal ratio refers to the proportion of time that a signal lamp can be used for vehicle traffic in a cycle. That is, the ratio of the green light time of a certain phase to the cycle time. Among them, the green light time can be the actual green light time or the effective green light time.
  • the actual green light time can be the time from the green light on to the green light off.
  • Effective green light time includes the actual vehicle travel time that is effectively used, which is equal to the sum of the green light time and the yellow light time minus the lost time.
  • the lost time includes two parts. One is the time when the green light signal is turned on and the vehicle starts; when the green light is off and the yellow light is turned on, only vehicles that have crossed the stop line can continue to pass, so there is also a part of the lost time, which is the actual green light time.
  • the end lag time is the effective part of the yellow light time.
  • the loss time of each phase is the difference between the start delay time and the end delay time.
  • Phase difference For two signal intersections, it refers to the difference between the start time of the green light (or red light) of the same phase at two adjacent intersections.
  • the way to control the traffic in the preset area within the target time period may include: setting the green signal ratio of the intersection at the corresponding phase of path A and path F to be greater than that of path C and path D The green signal ratio of the corresponding phase.
  • the method of controlling traffic in the preset area within the target time period may include: setting the phase difference of these intersections in the phases corresponding to the path A and the path F, so that the vehicle can enjoy the non-stop when driving along the path A or the path F The green wave effect of passing through these intersections continuously.
  • the processing method may further include steps S7110 to S7120:
  • Step S7110 Obtain the traffic trajectory matching each path in the future period as a new first training sample of the corresponding path.
  • Step S7120 according to the vector value of the first feature vector of the new first training sample of each path and the actual traffic volume value of the corresponding path in the target period corresponding to the new first training sample of each path, Correct the first mapping function of the corresponding path.
  • the actual traffic trajectory of the preset area that matches each path within the target time period can be obtained, as a new training sample of the corresponding path, to modify the first path of the corresponding path.
  • the mapping function is to add these new training samples and retrain the first mapping function of each path respectively, so that the prediction of the traffic flow value passing through each path becomes more and more accurate.
  • the processing method may further include steps S7210 to S7220:
  • Step S7210 Obtain the actual traffic trajectory in the target period as a new second training sample.
  • Step S7220 Correct the second mapping function according to the vector value of the second feature vector of the new second training sample and the traffic path distribution class actually corresponding to the new second training sample.
  • the actual traffic trajectory of the preset area in the target period can be obtained, and the actual traffic trajectory is used as a new training sample to modify the second mapping function, that is, to add these new Training samples and retraining the second mapping function to make the prediction of the corresponding traffic path distribution class more and more accurate.
  • FIG. 8 is an example of a method for processing traffic flow path distribution information. This example uses the preset areas shown in FIGS. 4 to 7 as an example to describe the method for processing traffic flow path distribution information.
  • the processing method may include the following steps S8001 to S8011:
  • Step S8001 Obtain a path in a preset area.
  • Step S8002 Acquire the selected first vector feature.
  • Step S8003 Obtain a first mapping function between the first feature vector and the traffic volume value passing through each path.
  • Step S8004 according to the first mapping function corresponding to each path and the vector value of the first feature vector in the current time period, respectively determine the predicted traffic volume value of the corresponding path in the target time period.
  • the target period can be a future period.
  • Step S8005 According to the predicted traffic flow value of each path in the target time period, obtain the predicted traffic flow path distribution information of the preset area in the target time period.
  • Step S8006 Obtain the traffic flow path distribution information of the preset area in multiple historical time periods as historical traffic flow path distribution information.
  • Step S8007 clustering multiple historical traffic flow path distribution information to obtain at least one traffic flow path distribution class.
  • Step S8008 Obtain the selected second feature vector.
  • Step S8009 Obtain a second mapping function between the second feature vector and the traffic flow path distribution class.
  • Step S8010 according to the second mapping function and the vector value of the second feature vector in the current period, obtain the traffic flow path distribution class corresponding to the preset area in the target period as the predicted traffic flow path distribution class.
  • Step S8011 Determine target traffic flow path distribution information representing the cluster center of the predicted traffic flow path distribution class.
  • Step S8012 Determine the maximum value of the distance between each vehicle flow path distribution information and the target vehicle flow path distribution information.
  • Step S8013 Determine the distance between the predicted traffic flow path distribution information and the target traffic flow path distribution information.
  • Step S8014 Determine the distance between each vehicle flow path distribution information contained in the predicted vehicle flow path distribution class and the target vehicle flow path distribution information.
  • Step S8015 Correct the predicted traffic flow path distribution information so that the distance between the corrected predicted traffic flow path distribution information and the target vehicle flow path distribution information is less than or equal to the maximum value.
  • the maximum value in this step is the maximum value obtained through step S8012.
  • a device 9000 for processing vehicle flow path distribution information includes a distribution information prediction module 9100, a distribution prediction module 9200, and a distribution information correction module 9300.
  • the distribution information prediction module 9100 is used to obtain the predicted traffic flow path distribution information of the preset area in the target time period; the traffic flow path distribution information includes the path in the preset area and the value of the flow of vehicles passing through each path in the corresponding time period; this distribution type
  • the prediction module 9200 is used to obtain the predicted traffic flow path distribution class corresponding to the preset area in the target time period; wherein the traffic flow path distribution class includes at least one historical period of traffic flow path distribution information;
  • the distribution information correction module 9300 is used for predicting the traffic flow path distribution Classes modify the predicted traffic flow path distribution information so that the revised predicted traffic flow path distribution information belongs to the predicted traffic flow path distribution category.
  • the distribution information prediction module 9100 can also be used to:
  • the predicted traffic flow value of each path in the target time period obtain the predicted traffic flow path distribution information of the preset area in the target time period.
  • obtaining the predicted traffic flow value of each path in the target period includes:
  • the selected first feature vector includes a plurality of first features that affect the traffic flow value of the target path in the target period;
  • the plurality of first features include a first traffic feature and a first environmental feature;
  • the predicted traffic volume value of the target path in the target period is obtained.
  • the first traffic feature includes at least one of traffic flow path distribution information in a preset area and traffic flow path distribution information around the preset area; and/or, the first environment feature includes time, date, and weather. At least one of.
  • obtaining the first mapping function between the first feature vector and the value of the traffic volume passing through the target path includes:
  • the first mapping function is obtained through training.
  • the device 9000 for processing vehicle flow path distribution information may further include:
  • the device 9000 for processing vehicle flow path distribution information may further include:
  • clustering a plurality of historical traffic flow path distribution information to obtain at least one traffic flow path distribution category includes:
  • each historical vehicle flow path determines the value of the vehicle flow through each path in the corresponding historical period
  • clustering multiple pieces of historical vehicle flow path distribution information to obtain at least one vehicle flow path distribution class According to the distance between every two pieces of historical vehicle flow path distribution information, clustering multiple pieces of historical vehicle flow path distribution information to obtain at least one vehicle flow path distribution class.
  • determining the distance between every two historical traffic flow path distribution information according to the number of road sections included in each path and the traffic flow value passing through the corresponding path in each historical period includes:
  • each road section According to the number of road sections included in each path and the value of the traffic volume passing through the corresponding path in each historical period, determine the road section flow sum corresponding to each historical traffic flow path distribution information; where the road section flow sum is each passing through the corresponding time period. The sum of the traffic flow value of each road section;
  • the road section flow difference corresponding to each two historical traffic flow path distribution information; where the road section flow difference corresponds to the two time periods The sum of the difference in the value of the traffic volume passing through each road section;
  • clustering multiple pieces of historical traffic flow path distribution information to obtain at least one traffic flow path distribution class includes:
  • taking each traffic path distribution information as a node, and constructing a relationship graph according to the distance between every two nodes includes:
  • each vehicle flow path distribution information as a node, respectively connecting each node and a set number of nodes with the closest distance to itself to obtain a relationship graph.
  • splitting the relationship graph into at least one subgraph according to the distance between every two nodes includes:
  • connection between the two nodes whose distance exceeds the preset distance threshold is cut off to split the relationship graph into at least one subgraph.
  • the distributed prediction module 9200 can also be used to:
  • the second feature vector includes a plurality of second features that affect the distribution of the traffic flow path corresponding to the preset area in the target period;
  • the plurality of second features include the second traffic feature and the second environment feature;
  • the traffic flow path distribution class corresponding to the preset area in the target period is obtained as the predicted traffic flow path distribution class.
  • the second traffic feature includes at least one of traffic flow path distribution information in a preset area and traffic flow path distribution information around the preset area; and/or, the second environment feature includes time, date, and weather. At least one of.
  • obtaining the second mapping function between the second feature vector and the vehicle flow path distribution class includes:
  • the second mapping function is obtained through training.
  • the processing device may further include:
  • the distribution information correction module 9300 can also be used to:
  • the predicted vehicle flow path distribution information is corrected so that the revised predicted vehicle flow path distribution information belongs to the predicted vehicle flow path distribution category.
  • the predicted vehicle flow path distribution information is corrected so that the revised predicted vehicle flow path distribution information belongs to the predicted vehicle flow path distribution category including:
  • determining the target vehicle flow path distribution information representing the cluster center of the predicted vehicle flow path distribution class includes:
  • the target vehicle flow path distribution information is obtained.
  • the device 9000 for processing vehicle flow path distribution information may further include:
  • the device 9000 for processing traffic flow path distribution information can be implemented in various ways.
  • the device 9000 for processing traffic flow path distribution information can be implemented by configuring the processor with instructions.
  • the instructions can be stored in the ROM, and when the device is started, the instructions are read from the ROM into the programmable device to realize the processing device 9000 of the traffic flow path distribution information.
  • the device 9000 for processing traffic flow path distribution information can be solidified into a dedicated device (for example, ASIC).
  • the device 9000 for processing vehicle flow distribution path information can be divided into mutually independent units, or they can be combined together for implementation.
  • the device 9000 for processing traffic flow path distribution information may be implemented by one of the foregoing various implementation manners, or may be implemented by a combination of two or more of the foregoing various implementation manners.
  • the device 9000 for processing vehicle flow path distribution information may have multiple implementation forms.
  • the device 9000 for processing vehicle flow path distribution information may be any software product or application that provides traffic path distribution information processing services.
  • an electronic device 7000 is also provided.
  • the electronic device 7000 may be the server 1100 shown in FIG. 1 or the terminal device 1200 shown in FIG. 2.
  • the electronic device 7000 may include the aforementioned vehicle flow path distribution information processing device 9000 for implementing the vehicle flow path distribution information processing method of any embodiment of the present application.
  • the electronic device 7000 may further include a processor 7100 and a memory 7200.
  • the memory 7200 is used to store executable instructions; the processor 7100 is used to execute the electronic device 7000 according to the control of the instructions.
  • a method for processing traffic flow path distribution information according to any embodiment of the present application.
  • a computer-readable storage medium on which a computer program is stored, and the computer program, when executed by a processor, implements the method for processing traffic flow path distribution information as in any embodiment of the present application.
  • the computer program product may include a computer-readable storage medium loaded with computer-readable program instructions for enabling a processor to implement various aspects of the present application.
  • the computer-readable storage medium may be a tangible device that can hold and store instructions used by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as a printer with instructions stored thereon
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory flash memory
  • SRAM static random access memory
  • CD-ROM compact disk read-only memory
  • DVD digital versatile disk
  • memory stick floppy disk
  • mechanical encoding device such as a printer with instructions stored thereon
  • the computer-readable storage medium used here is not interpreted as a transient signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (for example, light pulses through fiber optic cables), or through wires Transmission of electrical signals.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing/processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device .
  • the computer program instructions used to perform the operations of this application may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, status setting data, or in one or more programming languages
  • Source code or object code written in any combination the programming language includes object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as "C" language or similar programming languages.
  • Computer-readable program instructions can be executed entirely on the user's computer, partly on the user's computer, executed as a stand-alone software package, partly on the user's computer and partly executed on a remote computer, or entirely on the remote computer or server carried out.
  • the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to access the Internet connection).
  • LAN local area network
  • WAN wide area network
  • an electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), can be customized by using the status information of the computer-readable program instructions.
  • the computer-readable program instructions are executed to realize various aspects of the present application.
  • These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine such that when these instructions are executed by the processor of the computer or other programmable data processing device , A device that implements the functions/actions specified in one or more blocks in the flowchart and/or block diagram is produced. It is also possible to store these computer-readable program instructions in a computer-readable storage medium. These instructions make computers, programmable data processing apparatuses, and/or other devices work in a specific manner, so that the computer-readable medium storing instructions includes An article of manufacture, which includes instructions for implementing various aspects of the functions/actions specified in one or more blocks in the flowchart and/or block diagram.
  • each block in the flowchart or block diagram may represent a module, program segment, or part of an instruction, and the module, program segment, or part of an instruction contains one or more functions for implementing the specified logical function.
  • Executable instructions may also occur in a different order from the order marked in the drawings. For example, two consecutive blocks can be executed essentially in parallel, or they can sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart can be implemented by a dedicated hardware-based system that performs the specified functions or actions Or it can be realized by a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, implementation by software, and implementation by a combination of software and hardware are all equivalent.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种车流路径分布信息的处理方法、装置及电子设备，该处理方法包括：获取预设区域在下一时段的预测车流路径分布信息(S1000)；获取预设区域在下一时段对应的预测车流路径分布类(S2000)；根据预测车流路径分布类修正预测车流路径分布信息，以使修正后的预测车流路径分布信息属于预测车流路径分布类(S3000)。

Description

一种车流路径分布信息的处理方法、装置及电子设备
本申请要求2019年05月08日递交的申请号为201910381137.9、发明名称为“一种车流路径分布信息的处理方法、装置及电子设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及交通控制技术领域，更具体地，涉及一种车流路径分布信息的处理方法、一种车流路径分布信息的处理装置、一种电子设备、及一种计算机可读介质。
背景技术
交通区域在未来的交通需求可以是通过预测的车流路径分布信息中获得。车流路径分布信息包括该交通区域中的路径、及经过每一路径的车流量值。其中，路径可以是交通路网上的一个路段序列。
目标时段的车流路径分布可能与本交通区域中的当前车流路径分布信息、该交通区域周边的车流路径分布信息、时间、日期、天气等因素都有关系。
现有技术中通常是针对交通区域中的每个路径建立一个回归模型，得到每个路径的预测车流量值，并将每个路径的预测车流量值进行组合叠加，构成对交通区域在目标时段的车流路径分布信息的预测。但是，这种预测方式没有考虑预设区域的整体因素，容易导致误差的积累，使得组合得到的车流路径分布信息可能会与实际偏差很大。
发明内容
本申请的一个目的是提供一种预测预设区域在未来的车流路径分布信息的新技术方案。
根据本申请的第一方面，提供了一种车流路径分布信息的处理方法，包括：
获取预设区域在目标时段的预测车流路径分布信息；所述车流路径分布信息包括所述预设区域中路径、及在对应时段内经过各个路径的车流量值；
获取所述预设区域在所述目标时段对应的预测车流路径分布类；其中，所述车流路径分布类包括针对至少一个历史时段的车流路径分布信息进行聚类得到的聚类结果；
根据所述预测车流路径分布类修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类。
可选的，所述获取预设区域在目标时段的预测车流路径分布信息的步骤包括：
获取所述预设区域内的路径；
分别获取在目标时段经过每条路径的预测车流量值；
根据在所述目标时段经过每条路径的预测车流量值，获取预设区域在目标时段的预测车流路径分布信息。
可选的，将每条路径轮流作为目标路径，
获取在目标时段经过所述目标路径的预测车流量值的步骤包括：
获取选定的第一特征向量，其中，所述第一特征向量包括影响所述目标路径在目标时段的车流量值的多个第一特征；所述多个第一特征包括第一交通特征和第一环境特征；
获取所述第一特征向量与经过所述目标路径的车流量值之间的第一映射函数；
根据所述第一映射函数、及所述第一特征向量在当前时段的向量值，获得在目标时段经过所述目标路径的预测车流量值。
可选的，所述第一交通特征包括所述预设区域的车流路径分布信息、及所述预设区域周边的车流路径分布信息中的至少一项；和/或，所述第一环境特征包括时间、日期、天气中的至少一项。
可选的，所述获取所述第一特征向量与经过所述目标路径的车流量值之间的第一映射函数的步骤包括：
根据历史车流轨迹获取第一训练样本，其中，每个第一训练样本包括与所述目标路径匹配的历史车流轨迹；
根据所述第一训练样本的所述第一特征向量的向量值、与所述第一训练样本对应的经过所述目标路径的实际车流量值，训练得到所述第一映射函数。
可选的，所述处理方法还包括：
获取所述目标时段内与所述目标路径匹配的实际车流轨迹，作为新的第一训练样本；
根据所述新的第一训练样本的所述第一特征向量的向量值、及新的第一训练样本所对应的在所述目标时段内经过所述目标路径的实际车流量值，修正所述第一映射函数。
可选的，所述获取所述预设区域在所述目标时段对应的预测车流路径分布类之前还包括：
获取所述预设区域在多个历史时段内的车流路径分布信息，作为历史车流路径分布信息；
对多个所述历史车流路径分布信息进行聚类，得到至少一个车流路径分布类；
所述获取所述预设区域在目标时段对应的预测车流路径分布类包括：从所述至少一个车流路径分布类中，获取所述预设区域在目标时段对应的车流路径分布类，作为所述预测车流路径分布类。
可选的，所述对多个所述历史车流路径分布信息进行聚类，得到至少一个所述车流路径分布类的步骤包括：
确定每条路径所包含的路段数；
根据每个历史车流路径分布信息，确定对应历史时段内经过每条路径的车流量值；
根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离；
根据每两个历史车流路径分布信息之间的距离，对多个所述历史车流路径分布信息进行聚类，得到至少一个所述车流路径分布类。
可选的，所述根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离的步骤包括：
根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每个历史车流路径分布信息对应的路段流量和；其中，所述路段流量和为在对应时段内经过每个路段的车流量值的总和；
根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息对应的路段流量差；其中，所述路段流量差为对应两个时段内经过每个路段的车流量值的差异的总和；
根据每个历史车流路径分布信息对应的路段流量和、及每两个历史车流路径分布信息对应的路段流量差，确定每两个历史车流路径分布信息之间的距离。
可选的，所述根据每两个历史车流路径分布信息之间的距离，对多个所述历史车流路径分布信息进行聚类，得到至少一个所述车流路径分布类的步骤包括：
将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图；
根据每两个节点之间的距离，将所述关系图拆分为至少一个子图；
获得与每个子图一一对应的车流路径分布类，并将每个子图中包含的节点所对应的车流路径分布信息划分至对应车流路径分布类中。
可选的，所述将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图的步骤包括：
将每个车流路径分布信息作为一个节点，分别将每个节点、及与自身之间的距离最 近的设定数量个节点连接，得到所述关系图。
可选的，所述根据每两个节点之间的距离，将所述关系图拆分为至少一个子图的步骤包括：
截断距离超过预设的距离阈值的两个节点之间的连接，以将所述关系图拆分为至少一个子图。
可选的，所述获取所述预设区域在目标时段对应的预测车流路径分布类的步骤包括：
获取选定的第二特征向量，其中，所述第二特征向量包括影响所述预设区域在目标时段对应的车流路径分布类的多个第二特征；所述多个第二特征包括第二交通特征和第二环境特征；
获取所述第二特征向量与所述车流路径分布类之间的第二映射函数；
根据所述第二映射函数、及所述第二特征向量在当前时段的向量值，获得所述预设区域在目标时段对应的车流路径分布类，作为所述预测车流路径分布类。
可选的，所述第二交通特征包括所述预设区域的车流路径分布信息、及所述预设区域周边的车流路径分布信息中的至少一项；和/或，所述第二环境特征包括时间、日期、天气中的至少一项。
可选的，所述获取所述第二特征向量与所述车流路径分布类之间的第二映射函数的步骤包括：
将历史车流轨迹作为第二训练样本；
根据所述第二训练样本的所述第二特征向量的向量值、与所述第二训练样本实际对应的车流路径分布类，训练得到所述第二映射函数。
可选的，所述处理方法还包括：
获取所述目标时段内的实际车流轨迹，作为新的第二训练样本；
根据所述新的第二训练样本的第二特征向量的向量值、及所述新的第二训练样本实际对应的车流路径分布类，修正所述第二映射函数。
可选的，所述根据所述预测车流路径分布类修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类的步骤包括：
确定代表所述预测车流路径分布类的聚类中心的目标车流路径分布信息；
确定所述预测车流路径分布信息与所述目标车流路径分布信息之间的距离；
根据所述预测车流路径分布信息与所述目标车流路径分布信息之间的距离，修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径 分布类。
可选的，所述根据所述预测车流路径分布信息与所述目标车流路径分布信息之间的距离，修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类的步骤包括：
分别确定所述预测车流路径分布类中包含的每个车流路径分布信息与所述目标车流路径分布信息之间的距离；
确定每个车流路径分布信息与所述目标车流路径分布信息之间的距离的最大值；
修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息与所述目标车流路径分布信息之间的距离小于或等于所述最大值。
可选的，所述确定代表所述预测车流路径分布类的聚类中心的目标车流路径分布信息的步骤包括：
分别根据所述预测车流路径分布类中包含的车流路径分布信息中经过每条路径的车流量值，确定经过每条路径的目标车流量值与衡量聚类中心的指标之间的优化函数；
分别根据每条路径对应的优化函数，确定在衡量聚类中心的指标最小的情况下，经过每条路径的目标车流量值；
根据经过每条路径的目标车流量值，得到所述目标车流路径分布信息。
可选的，所述处理方法还包括：
根据修正后的预测车流路径分布信息，对所述预设区域进行交通控制。
根据本申请的第二方面，提供了一种车流路径分布信息的处理装置，包括：
分布信息预测模块，用于获取预设区域在目标时段的预测车流路径分布信息；所述车流路径分布信息包括所述预设区域中路径、及在对应时段内经过所述每条路径的车流量值；
分布类预测模块，用于获取所述预设区域在目标时段对应的预测车流路径分布类；其中，所述车流路径分布类包括至少一个历史时段的车流路径分布信息；
分布信息修正模块，用于根据所述预测车流路径分布类修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类。
根据本申请的第三方面，提供了一种电子设备，包括根据本申请第二方面所述的处理装置；或者，包括处理器和存储器，所述存储器用于存储可执行的指令，所述指令用于控制所述处理器执行根据本申请第一方面所述的处理方法。
根据本申请的第四方面，提供了一种计算机可读存储介质，其上存储有计算机程序， 所述计算机程序在被处理器执行时实现根据本申请第一方面所述的处理方法。
在本申请的实施例中，通过预先获取预设区域在目标时段的预测车流路径分布信息，再获取目标时段对应的预测车流路径分布类，并根据预测车流路径分布类修正预测车流路径分布信息，可以使得最终得到的修正后的预测车流路径分布信息更加精确。
通过以下参照附图对本申请的示例性实施例的详细描述，本申请的其它特征及其优点将会变得清楚。
附图说明
被结合在说明书中并构成说明书的一部分的附图示出了本申请的实施例，并且连同其说明一起用于解释本申请的原理。
图1是可用于实现本申请的实施例的电子设备的硬件配置的一个例子的框图。
图2是可用于实现本申请的实施例的电子设备的硬件配置的另一个例子的框图；
图3是根据本申请实施例的车流路径分布信息的处理方法的流程图；
图4是根据本申请实施例的预设区域的一个例子的示意图；
图5是根据本申请实施例的第一历史时段的车流路径分布信息的一个例子的示意图；
图6是根据本申请实施例的第二历史时段的车流路径分布信息的一个例子的示意图；
图7是根据本申请实施例的第三历史时段的车流路径分布信息的一个例子的示意图；
图8是根据本申请实施例的车流路径分布信息的处理方法的一个例子的流程示意图；
图9是根据本申请实施例的车流路径分布信息的处理装置的原理图；
图10是根据本申请第一个实施例提供的电子设备的原理框图；
图11是根据本申请第二个实施例提供的电子设备的硬件结构示意图。
具体实施方式
现在将参照附图来详细描述本申请的各种示例性实施例。应注意到：除非另外具体说明，否则在这些实施例中阐述的部件和步骤的相对布置、数字表达式和数值不限制本申请的范围。
以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本申请及其应用或使用的任何限制。
对于相关领域普通技术人员已知的技术、方法和设备可能不作详细讨论，但在适当情况下，所述技术、方法和设备应当被视为说明书的一部分。
在这里示出和讨论的所有例子中，任何具体值应被解释为仅仅是示例性的，而不是作为限制。因此，示例性实施例的其它例子可以具有不同的值。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步讨论。
<硬件配置>
图1和图2是可用于实现本申请任意实施例的车流路径分布信息的处理方法的电子设备1000的硬件配置的框图。
在一个实施例中，如图1所示，电子设备可以是服务器1100。
服务器1100提供处理、数据库、通讯设施的业务点。服务器1100可以是整体式服务器或是跨多计算机或计算机数据中心的分散式服务器。服务器可以是各种类型的，例如但不限于，网络服务器，新闻服务器，邮件服务器，消息服务器，广告服务器，文件服务器，应用服务器，交互服务器，数据库服务器，或代理服务器。在一些实施例中，每个服务器可以包括硬件，软件，或用于执行服务器所支持或实现的合适功能的内嵌逻辑组件或两个或多个此类组件的组合。例如，服务器例如刀片服务器、云端服务器等，或者可以是由多台服务器组成的服务器群组，可以包括上述类型的服务器中的一种或多种等等。
本实施例中，服务器1100可以如图1所示，包括处理器1110、存储器1120、接口装置1130、通信装置1140、显示装置1150、输入装置1160。
在该实施例中，服务器1100还可以包括扬声器、麦克风等等，在此不做限定。
处理器1110可以是专用的服务器处理器，也可以是满足性能要求的台式机处理器、移动版处理器等，在此不做限定。存储器1120例如包括ROM(只读存储器)、RAM(随机存取存储器)、诸如硬盘的非易失性存储器等。接口装置1130例如包括各种总线接口，例如串行总线接口(包括USB接口)、并行总线接口等。通信装置1140例如能够进行有线或无线通信。显示装置1150例如是液晶显示屏、LED显示屏触摸显示屏等。输入装置1160例如可以包括触摸屏、键盘等。
在该实施例中，服务器1100的存储器1120用于存储指令，该指令用于控制处理器1110进行操作以至少执行根据本申请任意实施例的车流路径分布信息的处理方法。技术人员可以根据本申请所公开方案设计指令。指令如何控制处理器进行操作，这是本领域公知，故在此不再详细描述。
尽管在图1中示出了服务器1100的多个装置，但是，本申请可以仅涉及其中的部分装置，例如，服务器1100只涉及存储器1120和处理器1110。
在一个实施例中，电子设备可以是操作人员使用的PC机、笔记本电脑等终端设备1200，在此不做限定。
本实施例中，参照图2所示，终端设备1200可以包括处理器1210、存储器1220、接口装置1230、通信装置1240、显示装置1250、输入装置1260、扬声器1270、麦克风1280等等。
处理器1210可以是移动版处理器。存储器1220例如包括ROM(只读存储器)、RAM(随机存取存储器)、诸如硬盘的非易失性存储器等。接口装置1230例如包括USB接口、耳机接口等。通信装置1240例如能够进行有线或无线通信，通信装置1240可以包括短距离通信装置，例如是基于Hilink协议、WiFi(IEEE 802.11协议)、Mesh、蓝牙、ZigBee、Thread、Z-Wave、NFC、UWB、LiFi等短距离无线通信协议进行短距离无线通信的任意装置，通信装置1240也可以包括远程通信装置，例如是进行WLAN、GPRS、2G/3G/4G/5G远程通信的任意装置。显示装置1250例如是液晶显示屏、触摸显示屏等。输入装置1260例如可以包括触摸屏、键盘等。用户可以通过扬声器1270和麦克风1280输入/输出语音信息。
在该实施例中，终端设备1200的存储器1220用于存储指令，该指令用于控制处理器1210进行操作以至少执行根据本申请任意实施例的车流路径分布信息的处理方法。技术人员可以根据本申请所公开方案设计指令。指令如何控制处理器进行操作，这是本领域公知，故在此不再详细描述。
尽管在图2中示出了终端设备1200的多个装置，但是，本申请可以仅涉及其中的部分装置，例如，终端设备1200只涉及存储器1220和处理器1210和显示装置1250。
<方法实施例>
在本实施例中，提供一种车流路径分布信息的处理方法。该处理方法可以是由电子设备实施。该电子设备可以是如图1所示的服务器1100，或者是如图2所示终端设备 1200。
根据图3所示，本实施例的车流路径分布信息的处理方法可以包括如下步骤S1000～S3000：
步骤S1000，获取预设区域在目标时段的预测车流路径分布信息。
其中，目标时段可以是未来的时段，也可以是过去的时段。
本实施例中的预设区域可以是根据应用场景或者具体需求在城市中选定的交通区域。例如，该预设区域可以是如图4中的预设区域所示，也可以是如图5～图7中的预设区域所示。
具体的，预测车流路径分布信息可以包含预设区域中的路径、及在目标时段内经过每条路径的预测车流量值。
本实施例中的路径可以是路网上的一个路段序列，路段可以是指交通网络上相邻两个路口之间在一个行驶方向上的交通线路。具体的，在如图4～图7中所示的实施例中，通过步骤S1000所获取的预设区域中的路径可以包括：路径A、路径B、路径C、路径D、路径E和路径F。
在一个实施例中，获取预设区域在目标时段的预测车流路径分布信息的步骤可以包括如下所示的步骤S1100～S1300：
步骤S1100，获取预设区域内的路径。
步骤S1200，分别获取在目标时段经过每条路径的预测车流量值。
分别获取在目标时段经过每条路径的预测车流量值具体可以为：将每条路径轮流作为目标路径，获取在目标时段经过目标路径的预测车流量值。
具体的，时段可以是根据应用场景或者是具体需求设置，且每个时段(包括本实施例中所述的目标时段、当前时段、历史时段)的时长相等。例如，可以是将每两个相邻整点之间的时段作为一个时段。那么，当前日期的8-9点可以作为当前时段，9-10点可以作为目标时段。
获取在目标时段经过目标路径的预测车流量值的步骤可以包括如下所示的步骤S1210～S1230：
步骤S1210，获取选定的第一向量特征。
第一特征向量可以包括影响目标路径在目标时段的车流量值的多个第一特征。多个第一特征包括第一交通特征和第一环境特征。
第一交通特征可以包括预设区域的车流路径分布信息、及预设区域周边的车流路径 分布信息中的至少一项。第一环境特征可以包括时间、日期、天气中的至少一项。
本实施例中，x j可以是第一交通特征、第一环境特征等能够影响目标路径在目标时段的车流量值的第一特征。例如，该第一交通特征可以为预设区域的车流路径分布信息和预设区域周边的车流路径分布信息，该第一环境特征可以为时间、日期和天气，在此，特征向量X可以具有5个特征，即n＝5，此时，可以将特征向量X表示为X＝(x 1,x 2,x 3,x 4,x 5)。当然，特征向量X中还可以包括与车流量值相关的其他特征。
步骤S1220，获取第一特征向量与经过目标路径的车流量值之间的第一映射函数。
该第一映射函数F1(x)的自变量即为特征向量X，因变量F1(x)即为由特征向量X决定的预测车流量值。
在本实施例中，该步骤S1220中获取第一特征向量与经过目标路径的车流量值之间的第一映射函数可以进一步包括如下步骤S1221～S1222：
步骤S1221，根据历史车流轨迹获取第一训练样本。
其中，每个第一训练样本中包括与目标路径匹配的车流轨迹。
步骤S1222，根据第一训练样本的第一特征向量的向量值、与第一训练样本对应的经过目标路径的实际车流量值，训练得到第一映射函数。
在一个实施例中，可以根据预设的训练周期，执行训练第一映射函数的步骤S1221～S1222。该训练周期可以根据具体应用场景或者应用需求设置，例如，可以设置为1天。
在本实施例中，基于训练样本的第一特征向量的向量值与训练样本对应的经过目标路径的实际车流量值，可以通过各种拟合手段获得第一映射函数F(x)。例如，可以利用任意的多元线性回归模型获得第一映射函数F1(x)，在此不做限定。
在一个例子中，该多元线性回归模型可以是简单的反映该第一映射函数F1(x)的多项式函数，其中，多项式函数的各阶系数未知，通过将该第一训练样本的第一特征向量的向量值、与第一训练样本对应的经过目标路径的实际车流量值代入该多项式函数，便可以确定多项式函数的各阶系数，进而获得第一映射函数F1(x)。
在另一个例子中，可以利用各种回归模型，例如加法模型，以该第一训练样本的第一特征向量的第一向量值、与第一训练样本对应的经过目标路径的实际车流量值作为准确样品进行多轮训练，每一轮都学习上一轮拟合后的残差，迭代T轮，即可将残差控制在很低的值，以使得最终得到的第一映射函数F1(x)具有非常高的精确度。该加法模型例如是LightGBM、GBDT、XGBoost等，在此不做限定。
步骤S1230，根据第一映射函数、及第一特征向量在当前时段的向量值，获得在目标时段经过目标路径的预测车流量值。
本实施例中，根据步骤S1220获得的第一特征向量与经过目标路径的车流量值之间的第一映射函数，根据第一特征向量在当前时段的向量值，便可将向量值代入第一映射函数F(x)中，以便获得在目标时段经过目标路径的预测车流量值。
根据本申请该实施例，其可以根据第一特征向量和第一映射函数获得在目标时段经过目标路径的预测车流量值，由于第一映射函数是根据大量的训练样本训练得到，从而利用该第一映射函数确定预测车流量值时，可以提高获得的预测车流量值的准确性。
步骤S1300，根据下一时段内经过每条路径的预测车流量值，得到预设区域在下一时段的预测车流路径分布信息。
具体的，可以是将下一时段内经过每条路径的预测车流量值进行整合，就可以得到预设区域在下一时段的预测车流路径分布信息。
步骤S2000，获取预设区域在目标时段对应的预测车流路径分布类。
其中，车流路径分布类包括针对预设区域在至少一个历史时段的车流路径分布信息进行聚类得到的聚类结果。
具体的，可以是针对预设区域在至少一个历史时段的车流路径分布信息进行聚类，得到至少一个车流路径分布类，每个车流路径分布类可以是属于该聚类的所有历史时段的车流路径分布信息的集合，也可以是对应聚类的唯一标识。通过聚类的唯一标识，可以唯一确定属于对应聚类的所有历史时段的车流路径分布信息。其中，该标识可以是由至少一个字符组成。
那么，目标时段对应的预测车流路径分布类可以是聚类得到的至少一个车流路径分布类中的一个。
车流路径分布信息可以是包括预设区域中的路径、及在对应时段内经过每条路径的车流量值。例如，图5～图7为图4中的预设区域在不同历史时段的车流路径分布信息的示意图。
在一个实施例中，步骤S1000和步骤S2000可以是同时执行的，可以是先执行步骤S1000再执行步骤S2000，还可以是先执行步骤S2000再执行步骤S1000。本申请并不对步骤S1000和步骤S2000的执行顺序做具体限定。
在一个实施例中，在执行步骤S2000之前，该处理方法还可以包括获取与预设区域对应的至少一个车流路径分布类的步骤，以供步骤S2000中获取预设区域在目标时段对 应的车流路径分布类，作为预测车流路径分布类。本实施例中的预测车流路径分布类为获取的与预设区域对应的至少一个车流路径分布类中的其中一个。
获取至少一个车流路径分布类的步骤可以进一步包括如下所示的步骤S6100～S6200：
步骤S6100，获取预设区域在多个历史时段内的车流路径分布信息，作为历史车流路径分布信息。
例如，图5可以为第一历史时段内的车流路径分布信息的示意图，图6可以为第二历史时段内的车流路径分布信息的示意图，图7可以为第三历史时段内的车流路径分布信息的示意图。
在如图5所示的第一历史时段内的车流路径分布信息中，可以得到在第一历史时段内经过路径A、路径B、路径C、路径D、路径E和路径F的车流量值分别为300、168、270、156、0、0。在如图6所示的第二历史时段内的车流路径分布信息中，可以得到在第二历史时段内经过路径A、路径B、路径C、路径D、路径E和路径F的车流量值分别为340、168、270、0、227、0。在如图7所示的第三历史时段内的车流路径分布信息中，可以得到在第三历史时段内经过路径A、路径B、路径C、路径D、路径E和路径F的车流量值分别为0、168、270、0、227、100。
步骤S6200，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类。
在一个例子中，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类可以进一步包括如下所示的步骤S6210～S6240：
步骤S6210，确定每条路径所包含的路段数。
本实施例中的路段是对应路径中相邻两个路口之间的交通线路，路段数是对应路径中所包含的路段的数量。
例如，在如图4～图7所示的预设区域中，路径A、路径B和路径F所包含的路段数均为5，路径C、路径D和路径E所包含的路段数均为3。
步骤S6220，根据每个历史车流路径分布信息，确定对应历史时段内经过每条路径的车流量值。
步骤S6230，根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离。
每两个历史车流路径分布信息之间的距离可以用于表征对应两个历史车流路径分布 信息之间的差异程度。
在一个例子中，根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离可以包括如下步骤S6231～S6233：
步骤S6231，根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每个历史车流路径分布信息对应的路段流量和。
其中，路段流量和为在对应时段内经过每个路段的车流量值的总和。
在如图5所示的第一历史时段内的车流路径分布信息中，根据每条路径所包含的路段数、及第一历史时段内经过对应路径的车流量值，可以得到第一历史时段内的车流路径分布信息对应的路段流量和为f 1，其中：
f 1＝300*5+168*5+270*3+156*3+0*5+0*5＝3618。
在如图6所示的第二历史时段内的车流路径分布信息中，根据每条路径所包含的路段数、及第二历史时段内经过对应路径的车流量值，可以得到第二历史时段内的车流路径分布信息对应的路段流量和为f 2，其中：
f 2＝340*5+168*5+270*3+0*3+227*5+0*5＝4031。
在如图7所示的第三历史时段内的车流路径分布信息中，根据每条路径所包含的路段数、及第三历史时段内经过对应路径的车流量值，可以得到第三历史时段内的车流路径分布信息对应的路段流量和为f 3，其中：
f 3＝0*5+168*5+270*3+0*3+227*5+100*5＝3285。
步骤S6232，根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息对应的路段流量差。
其中，路段流量差为对应两个时段内经过每个路段的车流量值的差值的总和。
具体的，可以是分别确定对应两个时段内经过每条路径中的所有路段的车流量值的差值的总和，再根据每条路径对应的车流量值的差值的总和求和，得到对应两个时段内经过每个路段的车流量值的差值的总和。其中，对应两个时段内经过其中一条路径中的所有路段的车流量值的差值的总和，可以是先确定对应两个时段内经过该路径的车流量值的差值，再通过该路径所包含的路段数乘以该差值得到。
对于如图5所示的第一历史时段内的车流路径分布信息、及如图6所示的第二历史时段内的车流路径分布信息而言，根据每条路径所包含的路段数、第一历史时段内经过 对应路径的车流量值、及第二历史时段内经过对应路径的车流量值，可以确定第一历史时段内的车流路径分布信息与第二历史时段内的车流路径分布信息对应的路段流量差。
具体的，可以是确定第一历史时段内和第二历史时段内经过路径A中每个路段的历史车流量的差值的和为(340-300)*5，第一历史时段内和第二历史时段内经过路径B中每个路段的历史车流量的差值的和为(168-168)*5，第一历史时段内和第二历史时段内经过路径C中每个路段的历史车流量的差值的和为(270-270)*3，第一历史时段内和第二历史时段内经过路径D中每个路段的历史车流量的差值的和为(156-0)*3，第一历史时段内和第二历史时段内经过路径E中每个路段的历史车流量的差值的和为(227-0)*3，第一历史时段内和第二历史时段内经过路径F中每个路段的历史车流量的差值的和为(0-0)*5。那么，第一历史时段内的车流路径分布信息与第二历史时段内的车流路径分布信息对应的路段流量差可以为Δf 12，其中，Δf 12＝(340-300)*5+(168-168)*5+(270-270)*3+(156-0)*3+(227-0)*3+(0-0)*5＝1349。
对于如图5所示的第一历史时段内的车流路径分布信息、及如图7所示的第三历史时段内的车流路径分布信息而言，根据每条路径所包含的路段数、第一历史时段内经过对应路径的车流量值、及第三历史时段内经过对应路径的车流量值，可以确定第一历史时段内的车流路径分布信息与第三历史时段内的车流路径分布信息对应的路段流量差。
具体的，可以是确定第一历史时段内和第三历史时段内经过路径A中每个路段的历史车流量的差值的和为(340-0)*5，第一历史时段内和第三历史时段内经过路径B中每个路段的历史车流量的差值的和为(168-168)*5，第一历史时段内和第三历史时段内经过路径C中每个路段的历史车流量的差值的和为(270-270)*3，第一历史时段内和第三历史时段内经过路径D中每个路段的历史车流量的差值的和为(156-0)*3，第一历史时段内和第三历史时段内经过路径E中每个路段的历史车流量的差值的和为(227-0)*3，第一历史时段内和第三历史时段内经过路径F中每个路段的历史车流量的差值的和为(100-0)*5。那么，第一历史时段内的车流路径分布信息与第三历史时段内的车流路径分布信息对应的路段流量差可以为Δf 13，Δf 13＝(340-0)*5+(168-168)*5+(270-270)*3+(156-0)*3+(227-0)*3+(100-0)*5＝3349。
对于如图6所示的第二历史时段内的车流路径分布信息、及如图7所示的第三历史时段内的车流路径分布信息而言，根据每条路径所包含的路段数、第二历史时段内经过对应路径的车流量值、及第三历史时段内经过对应路径的车流量值，可以确定第二历史 时段内的车流路径分布信息与第三历史时段内的车流路径分布信息对应的路段流量差。
具体的，可以是确定第二历史时段内和第三历史时段内经过路径A中每个路段的历史车流量的差值的和为(340-0)*5，第二历史时段内和第三历史时段内经过路径B中每个路段的历史车流量的差值的和为(168-168)*5，第二历史时段内和第三历史时段内经过路径C中每个路段的历史车流量的差值的和为(270-270)*3，第二历史时段内和第三历史时段内经过路径D中每个路段的历史车流量的差值的和为(0-0)*3，第二历史时段内和第三历史时段内经过路径E中每个路段的历史车流量的差值的和为(227-227)*3，第二历史时段内和第三历史时段内经过路径F中每个路段的历史车流量的差值的和为(100-0)*5。那么，第二历史时段内的车流路径分布信息与第三历史时段内的车流路径分布信息对应的路段流量差可以为Δf 23，Δf 23＝(340-0)*5+(168-168)*5+(270-270)*3+(0-0)*3+(227-227)*3+(100-0)*5＝2200。
步骤S6233，根据每个历史车流路径分布信息对应的路段流量和、及每两个历史车流路径分布信息对应的路段流量差，确定每两个历史车流路径分布信息之间的距离。
在一个例子中，确定每两个历史车流路径分布信息之间的距离的方式可以为：确定对应两个历史车流路径分布信息对应的路段流量和的几何平均值，再计算对应两个历史车流路径分布信息对应的路段流量差与该几何平均值之间的比值，作为对应两个历史车流路径分布信息之间的距离。
对于如图5所示的第一历史时段内的车流路径分布信息、及如图6所示的第二历史时段内的车流路径分布信息而言，可以先确定第一历史时段内的车流路径分布信息和第二历史时段内的车流路径分布信息对应的路段流量和的几何平均值为
Figure PCTCN2020086574-appb-000001
再计算第一历史时段内的车流路径分布信息和第二历史时段内的车流路径分布信息对应的路段流量差Δf 12与该几何平均值
Figure PCTCN2020086574-appb-000002
之间的比值，得到第一历史时段内的车流路径分布信息和第二历史时段内的车流路径分布信息之间的距离d 12，其中：
Figure PCTCN2020086574-appb-000003
对于如图5所示的第一历史时段内的车流路径分布信息、及如图7所示的第三历史时段内的车流路径分布信息而言，可以先确定第一历史时段内的车流路径分布信息和第三历史时段内的车流路径分布信息对应的路段流量和的几何平均值为
Figure PCTCN2020086574-appb-000004
再计算第一历史时段内的车流路径分布信息和第三历史时段内的车流路径分布信息对应的路段 流量差Δf 13与该几何平均值
Figure PCTCN2020086574-appb-000005
之间的比值，得到第一历史时段内的车流路径分布信息和第三历史时段内的车流路径分布信息之间的距离d 13，其中：
Figure PCTCN2020086574-appb-000006
对于如图6所示的第二历史时段内的车流路径分布信息、及如图7所示的第三历史时段内的车流路径分布信息而言，可以先确定第二历史时段内的车流路径分布信息和第三历史时段内的车流路径分布信息对应的路段流量和的几何平均值为
Figure PCTCN2020086574-appb-000007
再计算第二历史时段内的车流路径分布信息和第三历史时段内的车流路径分布信息对应的路段流量差Δf 23与该几何平均值
Figure PCTCN2020086574-appb-000008
之间的比值，得到第二历史时段内的车流路径分布信息和第三历史时段内的车流路径分布信息之间的距离d 23，
Figure PCTCN2020086574-appb-000009
步骤S6240，根据每两个历史车流路径分布信息之间的距离，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类。
本实施例中采用的聚类方式可以是系统聚类法、有序样品聚类法、动态聚类法、模糊聚类法、和图论聚类法中的任意一种或多种，在此不对具体聚类方式做限定。
通过对多个历史车流路径分布信息进行聚类，可以使得属于同一车流路径分布类中的历史车流路径分布信息之间的距离较小，而属于不同车流路径分布类中的历史车流路径分布信息之间的距离较远。
在一个例子中，根据每两个历史车流路径分布信息之间的距离，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类的步骤可以包括如下所示的步骤S6241～S6243：
步骤S6241，将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图。
在一个例子中，可以是将每个车流路径分布信息作为一个节点，将每两个节点连接，得到该关系图。
在另一个例子中，可以是将每个车流路径分布信息作为一个节点，分别将每个节点、及与自身之间的距离最近的设定数量个节点连接，得到该关系图。
在本例中，设定数量可以是预先根据应用场景或具体需求设定，例如，设定数量可 以设定为5，那么，可以是将每个节点、及与自身之间的距离最近的5个节点连接，得到该关系图。
步骤S6242，根据每两个节点之间的距离，将该关系图拆分为至少一个子图。
在一个例子中，可以是截断距离超过预设的距离阈值的两个节点之间的连接，以将该关系图拆分为至少一个子图。
在另一个例子中，应用图分割方法，将该关系图拆分为至少一个子图，使得位于同一子图内每两个节点之间的距离的总和最小、且位于不同子图内的每两个节点之间的距离的总和最大。
本例中采用的图分割方法可以是最小分割法、或者是Normalized Cut等，在此不做限定。
步骤S6243，获得与每个子图一一对应的车流路径分布类，并将每个子图中包含的节点所对应的车流路径分布信息划分至对应车流路径分布类中。
例如，在节点1和节点2位于第一子图中，节点3和节点4位于第二子图中的情况下，可以是获得与第一子图一一对应的车流路径分布类1、及与第二子图一一对应的车流路径分布类2，并将节点1和节点2所对应的车流路径分布信息划分至车流路径分布类1中，将节点3和节点4所对应的车流路径分布信息划分至车流路径分布类2中。
在得到至少一个车流路径分布类之后，可以根据获取预设区域在目标时段的车流路径分布信息属于得到的至少一个车流路径分布类中的一个，作为预测车流路径分布类。
获取所述预设区域在目标时段对应的车流路径分布类，作为预测车流路径分布类的步骤可以进一步包括如下所示的步骤S2100～S2300：
步骤S2100，获取选定的第二特征向量。
第二特征向量可以包括影响预设区域在目标时段对应的车流路径分布类的多个第二特征。多个第二特征包括第二交通特征和第二环境特征。
第二交通特征可以包括预设区域的车流路径分布信息、及预设区域周边的车流路径分布信息中的至少一项。第二环境特征可以包括时间、日期、天气中的至少一项。
本实施例中，y j可以是第二交通特征、第二环境特征等能够影响预设区域在目标时段对应的车流路径分布类的第二特征。例如，该第二交通特征可以为预设区域的车流路径分布信息和预设区域周边的车流路径分布信息，该第二环境特征可以为时间、日期和天气，在此，特征向量Y可以具有5个特征，即n＝5，此时，可以将特征向量Y表示为 Y＝(y 1,y 2,y 3,y 4,y 5)。当然，特征向量Y中还可以包括与对应的车流路径分布类相关的其他特征。
步骤S2200，获取第二特征向量与车流路径分布类之间的第二映射函数。
该第二映射函数F2(y)的自变量即为特征向量Y，因变量F2(y)即为由特征向量Y决定的预测车流路径分布类。
在本实施例中，该步骤S2200中获取第二特征向量与车流路径分布类之间的第二映射函数可以进一步包括如下步骤S2210～S2220：
步骤S2210，将历史车流轨迹作为第二训练样本。
步骤S2220，根据第二训练样本的第二特征向量的向量值、与第二训练样本实际对应的车流路径分布类，训练得到第二映射函数。
在一个实施例中，可以根据预设的训练周期，执行训练第二映射函数的步骤S2210～S2220。该训练周期可以根据具体应用场景或者应用需求设置，例如，可以设置为1天。
在本实施例中，基于训练样本的第二特征向量的向量值与训练样本实际对应的车流路径分布类，可以通过各种拟合手段获得第二映射函数F2(y)。例如，可以利用任意的多元线性回归模型获得第二映射函数F2(y)，在此不做限定。
在一个例子中，该多元线性回归模型可以是简单的反映该第二映射函数F2(y)的多项式函数，其中，多项式函数的各阶系数未知，通过将该第二训练样本的第二特征向量的向量值、与第二训练样本实际对应的车流路径分布类代入该多项式函数，便可以确定多项式函数的各阶系数，进而获得第二映射函数F2(y)。
在另一个例子中，可以利用各种回归模型，例如加法模型，以该第二训练样本的第二特征向量的第二向量值、与第二训练样本实际对应的车流路径分布类作为准确样品进行多轮训练，每一轮都学习上一轮拟合后的残差，迭代T轮，即可将残差控制在很低的值，以使得最终得到的第二映射函数F2(y)具有非常高的精确度。该加法模型例如是LightGBM、GBDT、XGBoost等，在此不做限定。
步骤S2300，根据第二映射函数、及第二特征向量在当前时段的向量值，获得预设区域在目标时段对应的车流路径分布类，作为预测车流路径分布类。
本实施例中，根据步骤S2200获得的第二特征向量与车流路径分布类之间的第二映射函数，根据第二特征向量在当前时段的向量值，便可将向量值代入第二映射函数F(x)中，以便获得预设区域在目标时段实际对应的预测车流路径分布类。
根据本申请该实施例，其可以根据第二特征向量和第二映射函数，获得预设区域在目标时段对应的预测车流路径分布类。由于第二映射函数是根据大量的训练样本训练得到，从而利用该第二映射函数确定预测车流路径分布类时，可以提高获得的预测车流路径分布类的准确性。
在执行完步骤S1000和步骤S2000之后，继续执行下述步骤S3000。
步骤S3000，根据预测车流路径分布类修正该预测车流路径分布信息，以使修正后的预测车流路径分布信息属于该预测车流路径分布类。
在预测车流路径分布类为属于对应聚类的所有历史时段的车流路径分布信息的集合的情况下，可以是根据属于预测车流路径分布类的所有历史时段的车流路径分布信息来修正该预测车流路径分布信息。在预测车流路径分布类为对应聚类的唯一标识的情况下，可以是通过该唯一标识，确定属于预测车流路径分布类的所有历史时段的车流路径分布信息，再根据属于预测车流路径分布类的所有历史时段的车流路径分布信息来修正该预测车流路径分布信息。
在一个例子中，根据预测车流路径分布类修正该预测车流路径分布信息可以包括步骤S3100～S3300：
步骤S3100，确定代表预测车流路径分布类的聚类中心的目标车流路径分布信息。
目标车流路径分布信息可以不是预测车流路径分布类中所包含的任一历史时段的车流路径分布信息，而是根据预测车流路径分布类中所包含的所有历史时段的车流路径分布信息所得到的、能够代表预测车流路径分布类的聚类中心的车流路径分布信息。
在一个例子中，确定目标车流路径分布信息的步骤可以包括步骤S3110～S3130：
步骤S3110，分别根据所述预测车流路径分布类中包含的车流路径分布信息中经过每条路径的车流量值，确定经过每条路径的目标车流量值与衡量聚类中心的指标之间的优化函数。
具体的，该对应每条路径的衡量聚类中心的指标，可以是经过对应路径的目标车流量值与每个经过对应路径的车流量值之间的差值的平方和。
步骤S3120，分别根据每条路径对应的优化函数，确定在衡量聚类中心的指标最小的情况下，经过每条路径的目标车流量值。
在一个实施例中，可以是利用启发式求解器Louvain算法对对应每条路径的优化函数进行求解，得到经过每条路径的目标车流量值。
步骤S3130，根据经过每条路径的目标车流量值，得到目标车流路径分布信息。
具体的，可以是对经过每条路径的目标车流量值进行整合，得到目标车流路径分布信息。
步骤S3200，确定预测车流路径分布信息与目标车流路径分布信息之间的距离。
确定预测车流路径分布信息与目标车流路径分布信息之间的距离的方式，具体可以参照前述的步骤S6231～S6233中确定每两个历史车流路径分布信息之间的距离的方式，在此不再赘述。
步骤S3300，根据预测车流路径分布信息与目标车流路径分布信息之间的距离修正预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类。
根据预测车流路径分布信息与目标车流路径分布信息之间的距离修正预测车流路径分布信息的方式，可以是使得修正后的预测车流路径分布信息与目标车流路径分布信息之间的距离更小。
在一个例子中，根据预测车流路径分布信息与目标车流路径分布信息之间的距离修正预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类的步骤可以包括如下步骤S3310～S3330：
步骤S3310，分别确定预测车流路径分布类中包含的每个车流路径分布信息与目标车流路径分布信息之间的距离。
确定预测车流路径分布类中包含的每个车流路径分布信息与目标车流路径分布信息之间的距离的方式，具体可以参照前述的步骤S6231～S6233中确定每两个历史车流路径分布信息之间的距离的方式，在此不再赘述。
步骤S3320，确定每个车流路径分布信息与目标车流路径分布信息之间的距离的最大值。
步骤S3330，修正预测车流路径分布信息，以使修正后的预测车流路径分布信息与目标车流路径分布信息之间的距离小于或等于该最大值。
如果修正后的预测车流路径分布信息与目标车流路径分布信息之间的距离小于或等于该最大值，则可以表明修正后的预测车流路径分布信息属于该预测车流路径分布类。
在一个例子中，该处理方法在执行步骤S3000之前还可以包括：确定预测车流路径分布信息是否属于预测车流路径分布类，如是，则无需根据预测车流路径分布类修正预测车流路径分布信息；如否，则执行步骤S3000，根据预测车流路径分布类修正该预测车流路径分布信息，以使修正后的预测车流路径分布信息属于该预测车流路径分布类。
这样，通过本实施例的处理方法来预测预设区域目标时段的车流路径分布信息，可以使得最终得到的修正后的预测车流路径分布信息更加精确。
在一个例子中，在通过步骤S3000修正预测车流路径分布信息之后，还可以根据该修正后的预测车流路径分布信息，对预设区域进行交通控制。具体的，可以帮助交通管理者根据该修正后的预测车流路径分布信息，提前决策信号灯优化调控方案。这样，可以帮助交通管理者更加主动地指定或调整交通管理方案，提高交通管理的质量和决策效率。
具体的，对预设区域进行交通控制的具体方式可以包括：对预设区域内的信号灯的信号周期时长、至少一个相位的绿信比、及多个路口的在至少一个相位的相位差中的至少一项进行相应的控制。
本实施例中的相位取业内公知的含义。例如,其可以包括，在一个信号周期内，具有相同的信号灯色显示的一股或几股交通流的信号状态序列称为一个相位。相位是按车流获得信号显示的时序来划分的，有多少种不同的时序排列，就有多少个相位。每一个控制状态，对应一组不同的灯色组合，称为一个相位。简而言之，一个相位也被称作一个控制状态。再例如，对于一组互不冲突的交通流同时获得通行权所对应的信号显示状态，可以将其称为相位。由此可见，相位是根据路口通行权在一个信号周期内的更迭来划分的。
信号周期时长，包括信号灯发生变化，信号运行一个循环所需的时间，等于绿、黄、红灯时间之和；也等于全部相位所需的绿灯时间和黄灯时间(一般是固定的)的总和。
绿信比是指信号灯一个周期内可用于车辆通行的比例时间。即某相位绿灯时间和周期时长的比值。其中，绿灯时间可以是实际绿灯时间，也可以是有效绿灯时间。
实际绿灯时间可以为绿灯开启至绿灯关闭所用的时间。有效绿灯时间：包括被有效利用的实际车辆通行时间，等于绿灯时间与黄灯时间之和减去损失时间。损失时间包括两部分，一是绿灯信号开启时，车辆启动时的时间；还有绿灯关闭、黄灯开启时，只有越过停止线的车辆才能继续通行，所以也有一部分损失时间，即为实际绿灯时间减去启动时间加速结束滞后时间。结束滞后时间是黄灯时间中有效利用的部分。每一相位的损失时间为启动延迟时间和结束滞后时间之差。
相位差：针对两个信号交叉路口而言，是指两个相邻交叉路口它们同一相位绿灯(或红灯)开始时间之差。
上述定义仅用于示例性描述本申请的具体实施方式，并不对发明保护范围进行限制 性解释。
例如，在根据如图4所示的预设区域得到的修正后的预测车流路径分布信息中，如果经过路径C和路径D的车流量值小于经过路径A和路径F的车流量值，那么，对于路径C与路径A的交叉路口，在目标时段内对预设区域进行交通控制的方式可以包括：设置该路口在与路径A和路径F对应相位的绿信比、大于在路径C和路径D对应相位的绿信比。
再例如，在根据如图4所示的预设区域得到的修正后的预测车流路径分布信息中，如果经过路径A和路径F的车流量值远大于经过其他路径的车流量值，那么，在目标时段内对预设区域进行交通控制的方式可以包括：设置这些路口在与路径A和路径F对应的相位的相位差，使车辆在沿着路径A或路径F行驶时，可以享受到不停车连续通过这些路口的绿波效果。
在一个例子中，在目标时段过后，该处理方法还可以包括步骤S7110～S7120：
步骤S7110，获取未来时段内与每条路径匹配的车流轨迹，作为对应路径的新的第一训练样本。
步骤S7120，根据每条路径的新的第一训练样本的第一特征向量的向量值、及每条路径的新的第一训练样本所对应的在目标时段内经过对应路径的实际车流量值，修正对应路径的第一映射函数。
根据本申请该实施例，其在目标时段过后，可以获得该预设区域在目标时段内与每条路径匹配的实际车流轨迹，作为对应路径的新的训练样本，分别去修正对应路径的第一映射函数，即增加这些新的训练样本，分别重新训练每条路径的第一映射函数，以使得对经过每条路径的车流量值的预测越来越准确。
在一个例子中，在目标时段过后，该处理方法还可以包括步骤S7210～S7220：
步骤S7210，获取目标时段内的实际车流轨迹，作为新的第二训练样本。
步骤S7220，根据新的第二训练样本的第二特征向量的向量值、及新的第二训练样本实际所对应的车流路径分布类，修正第二映射函数。
根据本申请该实施例，其在目标时段过后，可以获得该预设区域在目标时段的实际车流轨迹，将该实际车流轨迹作为新的训练样本，去修正第二映射函数，即增加这些新的训练样本，重新训练第二映射函数，以使得对对应的车流路径分布类的预测越来越准确。
<例子>
图8为一个例子的车流路径分布信息的处理方法，该例子以图4至图7所示的预设区域为例，对车流路径分布信息的处理方法进行描述。该处理方法可以包括如下步骤S8001～S8011：
步骤S8001，获取预设区域内的路径。
步骤S8002，获取选定的第一向量特征。
步骤S8003，获取第一特征向量与经过每条路径的车流量值之间的第一映射函数。
步骤S8004，根据与每条路径对应的第一映射函数、及第一特征向量在当前时段的向量值，分别确定在目标时段经过对应路径的预测车流量值。
在本例中，目标时段可以为未来的时段。
步骤S8005，根据目标时段内经过每条路径的预测车流量值，得到预设区域在目标时段的预测车流路径分布信息。
步骤S8006，获取预设区域在多个历史时段内的车流路径分布信息，作为历史车流路径分布信息。
步骤S8007，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类。
步骤S8008，获取选定的第二特征向量。
步骤S8009，获取第二特征向量与车流路径分布类之间的第二映射函数。
步骤S8010，根据第二映射函数、及第二特征向量在当前时段的向量值，获得预设区域在目标时段对应的车流路径分布类，作为预测车流路径分布类。
步骤S8011，确定代表预测车流路径分布类的聚类中心的目标车流路径分布信息。
步骤S8012，确定每个车流路径分布信息与目标车流路径分布信息之间的距离的最大值。
步骤S8013，确定预测车流路径分布信息与目标车流路径分布信息之间的距离。
步骤S8014，分别确定预测车流路径分布类中包含的每个车流路径分布信息与目标车流路径分布信息之间的距离。
步骤S8015，修正预测车流路径分布信息，以使修正后的预测车流路径分布信息与目标车流路径分布信息之间的距离小于或等于该最大值。
本步骤中的最大值为通过步骤S8012得到的最大值。
<装置实施例>
在本实施例中，提供一种车流路径分布信息的处理装置9000，如图9所示，包括分布信息预测模块9100、分布类预测模块9200和分布信息修正模块9300。该分布信息预测模块9100用于获取预设区域在目标时段的预测车流路径分布信息；车流路径分布信息包括预设区域中路径、及在对应时段内经过每条路径的车流量值；该分布类预测模块9200用于获取预设区域在目标时段对应的预测车流路径分布类；其中，车流路径分布类包括至少一个历史时段的车流路径分布信息；该分布信息修正模块9300用于根据预测车流路径分布类修正预测车流路径分布信息，以使修正后的预测车流路径分布信息属于预测车流路径分布类。
在一个例子中，该分布信息预测模块9100还可以用于：
用于获取预设区域内的路径；
分别获取在目标时段经过每条路径的预测车流量值；
根据在目标时段经过每条路径的预测车流量值，获取预设区域在目标时段的预测车流路径分布信息。
在一个例子中，分别获取在目标时段经过每条路径的预测车流量值包括：
将每条路径轮流作为目标路径；
获取选定的第一特征向量，其中，第一特征向量包括影响目标路径在目标时段的车流量值的多个第一特征；多个第一特征包括第一交通特征和第一环境特征；
获取第一特征向量与经过目标路径的车流量值之间的第一映射函数；
根据第一映射函数、及第一特征向量在当前时段的向量值，获得在目标时段经过目标路径的预测车流量值。
在一个例子中，第一交通特征包括预设区域的车流路径分布信息、及预设区域周边的车流路径分布信息中的至少一项；和/或，第一环境特征包括时间、日期、天气中的至少一项。
在一个例子中，获取第一特征向量与经过目标路径的车流量值之间的第一映射函数包括：
根据历史车流轨迹获取第一训练样本，其中，每个第一训练样本包括与目标路径匹配的历史车流轨迹；
根据第一训练样本的第一特征向量的向量值、与第一训练样本对应的经过目标路径的实际车流量值，训练得到第一映射函数。
在一个例子中，该车流路径分布信息的处理装置9000还可以包括：
用于获取目标时段内与目标路径匹配的实际车流轨迹，作为新的第一训练样本的模块；
用于根据新的第一训练样本的第一特征向量的向量值、及新的第一训练样本所对应的在目标时段内经过目标路径的实际车流量值，修正第一映射函数的模块。
在一个例子中，该车流路径分布信息的处理装置9000还可以包括：
用于获取预设区域在多个历史时段内的车流路径分布信息，作为历史车流路径分布信息的模块；
用于对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类的模块。
在一个例子中，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类包括：
确定每条路径所包含的路段数；
根据每个历史车流路径分布信息，确定对应历史时段内经过每条路径的车流量值；
根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离；
根据每两个历史车流路径分布信息之间的距离，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类。
在一个例子中，根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离包括：
根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每个历史车流路径分布信息对应的路段流量和；其中，路段流量和为在对应时段内经过每个路段的车流量值的总和；
根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息对应的路段流量差；其中，路段流量差为对应两个时段内经过每个路段的车流量值的差异的总和；
根据每个历史车流路径分布信息对应的路段流量和、及每两个历史车流路径分布信息对应的路段流量差，确定每两个历史车流路径分布信息之间的距离。
在一个例子中，根据每两个历史车流路径分布信息之间的距离，对多个历史车流路径分布信息进行聚类，得到至少一个车流路径分布类包括：
将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图；
根据每两个节点之间的距离，将关系图拆分为至少一个子图；
获得与每个子图一一对应的车流路径分布类，并将每个子图中包含的节点所对应的车流路径分布信息划分至对应车流路径分布类中。
在一个例子中，将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图包括：
将每个车流路径分布信息作为一个节点，分别将每个节点、及与自身之间的距离最近的设定数量个节点连接，得到关系图。
在一个例子中，根据每两个节点之间的距离，将关系图拆分为至少一个子图包括：
截断距离超过预设的距离阈值的两个节点之间的连接，以将关系图拆分为至少一个子图。
在一个例子中，分布类预测模块9200还可以用于：
获取选定的第二特征向量，其中，第二特征向量包括影响预设区域在目标时段对应的车流路径分布类的多个第二特征；多个第二特征包括第二交通特征和第二环境特征；
获取第二特征向量与车流路径分布类之间的第二映射函数；
根据第二映射函数、及第二特征向量在当前时段的向量值，获得预设区域在目标时段对应的车流路径分布类，作为预测车流路径分布类。
在一个例子中，第二交通特征包括预设区域的车流路径分布信息、及预设区域周边的车流路径分布信息中的至少一项；和/或，第二环境特征包括时间、日期、天气中的至少一项。
在一个例子中，获取第二特征向量与车流路径分布类之间的第二映射函数包括：
将历史车流轨迹作为第二训练样本；
根据第二训练样本的第二特征向量的向量值、与第二训练样本实际对应的车流路径分布类，训练得到第二映射函数。
在一个例子中，该处理装置还可以包括：
用于获取目标时段内的实际车流轨迹，作为新的第二训练样本的模块；
用于根据新的第二训练样本的第二特征向量的向量值、及新的第二训练样本实际所对应的车流路径分布类，修正第二映射函数的模块。
在一个例子中，分布信息修正模块9300还可以用于：
确定代表预测车流路径分布类的聚类中心的目标车流路径分布信息；
确定预测车流路径分布信息与目标车流路径分布信息之间的距离；
根据预测车流路径分布信息与目标车流路径分布信息之间的距离，修正预测车流路 径分布信息，以使修正后的预测车流路径分布信息属于预测车流路径分布类。
在一个例子中，根据预测车流路径分布信息与目标车流路径分布信息之间的距离，修正预测车流路径分布信息，以使修正后的预测车流路径分布信息属于预测车流路径分布类包括：
分别确定预测车流路径分布类中包含的每个车流路径分布信息与目标车流路径分布信息之间的距离；
确定每个车流路径分布信息与目标车流路径分布信息之间的距离的最大值；
修正预测车流路径分布信息，以使修正后的预测车流路径分布信息与目标车流路径分布信息之间的距离小于或等于最大值。
在一个例子中，确定代表预测车流路径分布类的聚类中心的目标车流路径分布信息包括：
分别根据预测车流路径分布类中包含的车流路径分布信息中经过每条路径的车流量值，确定经过每条路径的目标车流量值与衡量聚类中心的指标之间的优化函数；
分别根据每条路径对应的优化函数，确定在衡量聚类中心的指标最小的情况下，经过每条路径的目标车流量值；
根据经过每条路径的目标车流量值，得到目标车流路径分布信息。
在一个例子中，该车流路径分布信息的处理装置9000还可以包括：
用于根据修正后的预测车流路径分布信息，对预设区域进行交通控制的模块。
本领域技术人员应当明白，可以通过各种方式来实现车流路径分布信息的处理装置9000。例如，可以通过指令配置处理器来实现车流路径分布信息的处理装置9000。例如，可以将指令存储在ROM中，并且当启动设备时，将指令从ROM读取到可编程器件中来实现车流路径分布信息的处理装置9000。例如，可以将车流路径分布信息的处理装置9000固化到专用器件(例如ASIC)中。可以将车流分布路径信息的处理装置9000分成相互独立的单元，或者可以将它们合并在一起实现。车流路径分布信息的处理装置9000可以通过上述各种实现方式中的一种来实现，或者可以通过上述各种实现方式中的两种或更多种方式的组合来实现。
在本实施例中，车流路径分布信息的处理装置9000可以具有多种实现形式，例如，车流路径分布信息的处理装置9000可以是任何的提供车流路径分布信息处理服务的软件产品或者应用程序中运行的功能模块，或者是这些软件产品或者应用程序的外设嵌入件、插件、补丁件等，还可以是这些软件产品或者应用程序本身。
<电子设备>
在本实施例中，还提供一种电子设备7000。该电子设备7000可以是图1所示的服务器1100，也可以是如图2所示的终端设备1200。
在一方面，如图10所示，该电子设备7000可以包括前述的车流路径分布信息的处理装置9000，用于实施本申请任意实施例的车流路径分布信息的处理方法。
在另一方面，如图11所示，电子设备7000还可以包括处理器7100和存储器7200，该存储器7200用于存储可执行的指令；该处理器7100用于根据指令的控制运行电子设备7000执行根据本申请任意实施例的车流路径分布信息的处理方法。
<计算机可读存储介质>
在本实施例中，还提供一种计算机可读存储介质，其上存储有计算机程序，计算机程序在被处理器执行时实现如本申请任意实施例的车流路径分布信息的处理方法。
本申请可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本申请的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本申请操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、 机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本申请的各个方面。
这里参照根据本申请实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本申请的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本申请的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基 本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。对于本领域技术人员来说公知的是，通过硬件方式实现、通过软件方式实现以及通过软件和硬件结合的方式实现都是等价的。
以上已经描述了本申请的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。本申请的范围由所附权利要求来限定。

Claims (23)

  1. 一种车流路径分布信息的处理方法，其中，包括：
    获取预设区域在目标时段的预测车流路径分布信息；所述车流路径分布信息包括所述预设区域中路径、及在对应时段内经过各个路径的车流量值；
    获取所述预设区域在所述目标时段对应的预测车流路径分布类；其中，所述车流路径分布类包括针对至少一个历史时段的车流路径分布信息进行聚类得到的聚类结果；
    根据所述预测车流路径分布类修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类。
  2. 根据权利要求1所述的处理方法，其中，所述获取预设区域在目标时段的预测车流路径分布信息的步骤包括：
    获取所述预设区域内的路径；
    分别获取在目标时段经过每条路径的预测车流量值；
    根据在所述目标时段经过每条路径的预测车流量值，获取所述预设区域在目标时段的预测车流路径分布信息。
  3. 根据权利要求2所述的处理方法，其中，将所述预设区域内的每条路径轮流作为目标路径，
    获取在目标时段经过所述目标路径的预测车流量值的步骤包括：
    获取选定的第一特征向量，其中，所述第一特征向量包括影响所述目标路径在目标时段的车流量值的多个第一特征；所述多个第一特征包括第一交通特征和第一环境特征；
    获取所述第一特征向量与经过所述目标路径的车流量值之间的第一映射函数；
    根据所述第一映射函数、及所述第一特征向量在当前时段的向量值，获得在目标时段经过所述目标路径的预测车流量值。
  4. 根据权利要求3所述的处理方法，其中，所述第一交通特征包括所述预设区域的车流路径分布信息、及所述预设区域周边的车流路径分布信息中的至少一项；和/或，所述第一环境特征包括时间、日期、天气中的至少一项。
  5. 根据权利要求3所述的处理方法，其中，所述获取所述第一特征向量与经过所述目标路径的车流量值之间的第一映射函数的步骤包括：
    根据历史车流轨迹获取第一训练样本，其中，每个第一训练样本包括与所述目标路径匹配的历史车流轨迹；
    根据所述第一训练样本的所述第一特征向量的向量值、与所述第一训练样本对应的经过所述目标路径的实际车流量值，训练得到所述第一映射函数。
  6. 根据权利要求5所述的处理方法，其中，所述处理方法还包括：
    获取所述目标时段内与所述目标路径匹配的实际车流轨迹，作为新的第一训练样本；
    根据所述新的第一训练样本的所述第一特征向量的向量值、及新的第一训练样本所对应的在所述目标时段内经过所述目标路径的实际车流量值，修正所述第一映射函数。
  7. 根据权利要求1所述的处理方法，其中，所述获取所述预设区域在所述目标时段对应的预测车流路径分布类之前还包括：
    获取所述预设区域在多个历史时段内的车流路径分布信息，作为历史车流路径分布信息；
    对多个所述历史车流路径分布信息进行聚类，得到至少一个车流路径分布类；
    所述获取所述预设区域在目标时段对应的预测车流路径分布类包括：从所述至少一个车流路径分布类中，获取所述预设区域在目标时段对应的车流路径分布类，作为所述预测车流路径分布类。
  8. 根据权利要求7所述的处理方法，其中，所述对多个所述历史车流路径分布信息进行聚类，得到至少一个所述车流路径分布类的步骤包括：
    确定每条路径所包含的路段数；
    根据每个历史车流路径分布信息，确定对应历史时段内经过每条路径的车流量值；
    根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离；
    根据每两个历史车流路径分布信息之间的距离，对多个所述历史车流路径分布信息进行聚类，得到至少一个所述车流路径分布类。
  9. 根据权利要求8所述的处理方法，其中，所述根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每两个历史车流路径分布信息之间的距离的步骤包括：
    根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定每个历史车流路径分布信息对应的路段流量和；其中，所述路段流量和为在对应时段内经过每个路段的车流量值的总和；
    根据每条路径所包含的路段数、及每个历史时段内经过对应路径的车流量值，确定 每两个历史车流路径分布信息对应的路段流量差；其中，所述路段流量差为对应两个时段内经过每个路段的车流量值的差异的总和；
    根据每个历史车流路径分布信息对应的路段流量和、及每两个历史车流路径分布信息对应的路段流量差，确定每两个历史车流路径分布信息之间的距离。
  10. 根据权利要求8所述的处理方法，其中，所述根据每两个历史车流路径分布信息之间的距离，对多个所述历史车流路径分布信息进行聚类，得到至少一个所述车流路径分布类的步骤包括：
    将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图；
    根据每两个节点之间的距离，将所述关系图拆分为至少一个子图；
    获得与每个子图一一对应的车流路径分布类，并将每个子图中包含的节点所对应的车流路径分布信息划分至对应车流路径分布类中。
  11. 根据权利要求10所述的处理方法，其中，所述将每个车流路径分布信息作为一个节点，根据每两个节点之间的距离构建关系图的步骤包括：
    将每个车流路径分布信息作为一个节点，分别将每个节点、及与自身之间的距离最近的设定数量个节点连接，得到所述关系图。
  12. 根据权利要求11所述的处理方法，其中，所述根据每两个节点之间的距离，将所述关系图拆分为至少一个子图的步骤包括：
    截断距离超过预设的距离阈值的两个节点之间的连接，以将所述关系图拆分为至少一个子图。
  13. 根据权利要求1所述的处理方法，其中，所述获取所述预设区域在所述目标时段对应的预测车流路径分布类的步骤包括：
    获取选定的第二特征向量，其中，所述第二特征向量包括影响所述预设区域在目标时段对应的车流路径分布类的多个第二特征；所述多个第二特征包括第二交通特征和第二环境特征；
    获取所述第二特征向量与所述车流路径分布类之间的第二映射函数；
    根据所述第二映射函数、及所述第二特征向量在当前时段的向量值，获得所述预设区域在所述目标时段对应的车流路径分布类，作为所述预测车流路径分布类。
  14. 根据权利要求13所述的处理方法，其中，所述第二交通特征包括所述预设区域的车流路径分布信息、及所述预设区域周边的车流路径分布信息中的至少一项；和/或，所述第二环境特征包括时间、日期、天气中的至少一项。
  15. 根据权利要求13所述的处理方法，其中，所述获取所述第二特征向量与所述车流路径分布类之间的第二映射函数的步骤包括：
    将历史车流轨迹作为第二训练样本；
    根据所述第二训练样本的所述第二特征向量的向量值、与所述第二训练样本实际对应的车流路径分布类，训练得到所述第二映射函数。
  16. 根据权利要求15所述的处理方法，其中，所述处理方法还包括：
    获取所述目标时段内的实际车流轨迹，作为新的第二训练样本；
    根据所述新的第二训练样本的第二特征向量的向量值、及所述新的第二训练样本实际对应的车流路径分布类，修正所述第二映射函数。
  17. 根据权利要求1所述的处理方法，其中，所述根据所述预测车流路径分布类修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类的步骤包括：
    确定代表所述预测车流路径分布类的聚类中心的目标车流路径分布信息；
    确定所述预测车流路径分布信息与所述目标车流路径分布信息之间的距离；
    根据所述预测车流路径分布信息与所述目标车流路径分布信息之间的距离，修正所述预测车流路径分布信息，以使所述修正后的预测车流路径分布信息属于所述预测车流路径分布类。
  18. 根据权利要求17所述的处理方法，其中，所述根据所述预测车流路径分布信息与所述目标车流路径分布信息之间的距离，修正所述预测车流路径分布信息，以使所述修正后的预测车流路径分布信息属于所述预测车流路径分布类的步骤包括：
    分别确定所述预测车流路径分布类中包含的每个车流路径分布信息与所述目标车流路径分布信息之间的距离；
    确定每个车流路径分布信息与所述目标车流路径分布信息之间的距离的最大值；
    修正所述预测车流路径分布信息，以使所述修正后的预测车流路径分布信息与所述目标车流路径分布信息之间的距离小于或等于所述最大值。
  19. 根据权利要求17所述的处理方法，其中，所述确定代表所述预测车流路径分布类的聚类中心的目标车流路径分布信息的步骤包括：
    分别根据所述预测车流路径分布类中包含的车流路径分布信息中经过每条路径的车流量值，确定经过每条路径的目标车流量值与衡量聚类中心的指标之间的优化函数；
    分别根据每条路径对应的优化函数，确定在衡量聚类中心的指标最小的情况下，经 过每条路径的目标车流量值；
    根据经过每条路径的目标车流量值，得到所述目标车流路径分布信息。
  20. 根据权利要求1-19中任一项所述的处理方法，其中，所述处理方法还包括：
    根据所述修正后的预测车流路径分布信息，对所述预设区域进行交通控制。
  21. 一种车流路径分布信息的处理装置，其中，包括：
    分布信息预测模块，用于获取预设区域在目标时段的预测车流路径分布信息；所述车流路径分布信息包括所述预设区域中路径、及在对应时段内经过各个路径的车流量值；
    分布类预测模块，用于获取所述预设区域在所述目标时段对应的预测车流路径分布类；其中，所述车流路径分布类包括针对至少一个历史时段的车流路径分布信息进行聚类得到的聚类结果；
    分布信息修正模块，用于根据所述预测车流路径分布类修正所述预测车流路径分布信息，以使修正后的预测车流路径分布信息属于所述预测车流路径分布类。
  22. 一种电子设备，其中，包括根据权利要求21所述的处理装置；或者，包括处理器和存储器，所述存储器用于存储可执行的指令，所述指令用于控制所述处理器执行根据权利要求1至20中任一项所述的处理方法。
  23. 一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序在被处理器执行时实现如权利要求1至20中任一项所述的处理方法。
PCT/CN2020/086574 2019-05-08 2020-04-24 一种车流路径分布信息的处理方法、装置及电子设备 WO2020224445A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910381137.9A CN111915877A (zh) 2019-05-08 2019-05-08 一种车流路径分布信息的处理方法、装置及电子设备
CN201910381137.9 2019-05-08

Publications (1)

Publication Number Publication Date
WO2020224445A1 true WO2020224445A1 (zh) 2020-11-12

Family

ID=73050578

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/086574 WO2020224445A1 (zh) 2019-05-08 2020-04-24 一种车流路径分布信息的处理方法、装置及电子设备

Country Status (2)

Country Link
CN (1) CN111915877A (zh)
WO (1) WO2020224445A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114463969A (zh) * 2021-12-22 2022-05-10 高德软件有限公司 红绿灯周期时长的挖掘方法、电子设备及计算机程序产品
CN114973659A (zh) * 2022-05-12 2022-08-30 山东高速集团有限公司创新研究院 一种高速公路的间接事件检测的方法、装置及系统
CN115148033A (zh) * 2021-03-31 2022-10-04 丰田自动车株式会社 交通控制系统、交通控制方法和存储介质
CN115206092A (zh) * 2022-06-10 2022-10-18 南京工程学院 一种基于注意力机制的BiLSTM和LightGBM模型的交通预测方法
CN116311941A (zh) * 2023-03-24 2023-06-23 阿波罗智联(北京)科技有限公司 主要车流路径提取方法、装置、设备及介质
CN116453333A (zh) * 2023-03-24 2023-07-18 阿波罗智联(北京)科技有限公司 用于预测主要车流路径的方法及模型训练方法

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117612379B (zh) * 2024-01-24 2024-03-29 山东华夏高科信息股份有限公司 一种智慧交通车流量预测方法及系统

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101789176A (zh) * 2010-01-22 2010-07-28 天津市市政工程设计研究院 预约集港模式下的港区短时交通流预测方法
JP4743523B2 (ja) * 2006-03-22 2011-08-10 住友電気工業株式会社 交差点における車両挙動の予測システム
CN102693633A (zh) * 2012-06-07 2012-09-26 浙江大学 一种短时交通流加权组合预测方法
CN103839409A (zh) * 2014-02-27 2014-06-04 南京大学 基于多断面视觉传感聚类分析的交通流状态判别方法
CN105185115A (zh) * 2015-09-07 2015-12-23 华迪计算机集团有限公司 一种交通预测方法和预测系统
WO2016095708A1 (zh) * 2014-12-16 2016-06-23 高德软件有限公司 一种交通流量预测方法、预测模型生成方法及装置
CN106530762A (zh) * 2016-12-26 2017-03-22 东软集团股份有限公司 交通信号控制方法和装置
CN107862864A (zh) * 2017-10-18 2018-03-30 南京航空航天大学 基于驾驶习惯和交通路况的行驶工况智能预测估计方法
CN108492568A (zh) * 2018-04-25 2018-09-04 南京邮电大学 一种基于时空特性分析的短时交通流预测方法
CN108629973A (zh) * 2018-05-11 2018-10-09 四川九洲视讯科技有限责任公司 基于固定检测设备的路段交通拥堵指数计算方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103632542A (zh) * 2012-08-27 2014-03-12 国际商业机器公司 交通信息处理方法、装置和相应设备
KR102200121B1 (ko) * 2014-02-17 2021-01-08 삼성전자 주식회사 차량 흐름 예측 방법 및 장치
CN108960431A (zh) * 2017-05-25 2018-12-07 北京嘀嘀无限科技发展有限公司 指标的预测、模型的训练方法及装置
CN109598936B (zh) * 2018-12-18 2020-01-10 中国科学院地理科学与资源研究所 一种基于动态stknn模型的短时交通预测方法

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4743523B2 (ja) * 2006-03-22 2011-08-10 住友電気工業株式会社 交差点における車両挙動の予測システム
CN101789176A (zh) * 2010-01-22 2010-07-28 天津市市政工程设计研究院 预约集港模式下的港区短时交通流预测方法
CN102693633A (zh) * 2012-06-07 2012-09-26 浙江大学 一种短时交通流加权组合预测方法
CN103839409A (zh) * 2014-02-27 2014-06-04 南京大学 基于多断面视觉传感聚类分析的交通流状态判别方法
WO2016095708A1 (zh) * 2014-12-16 2016-06-23 高德软件有限公司 一种交通流量预测方法、预测模型生成方法及装置
CN105185115A (zh) * 2015-09-07 2015-12-23 华迪计算机集团有限公司 一种交通预测方法和预测系统
CN106530762A (zh) * 2016-12-26 2017-03-22 东软集团股份有限公司 交通信号控制方法和装置
CN107862864A (zh) * 2017-10-18 2018-03-30 南京航空航天大学 基于驾驶习惯和交通路况的行驶工况智能预测估计方法
CN108492568A (zh) * 2018-04-25 2018-09-04 南京邮电大学 一种基于时空特性分析的短时交通流预测方法
CN108629973A (zh) * 2018-05-11 2018-10-09 四川九洲视讯科技有限责任公司 基于固定检测设备的路段交通拥堵指数计算方法

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115148033A (zh) * 2021-03-31 2022-10-04 丰田自动车株式会社 交通控制系统、交通控制方法和存储介质
CN114463969A (zh) * 2021-12-22 2022-05-10 高德软件有限公司 红绿灯周期时长的挖掘方法、电子设备及计算机程序产品
CN114463969B (zh) * 2021-12-22 2023-05-16 高德软件有限公司 红绿灯周期时长的挖掘方法、电子设备及计算机程序产品
CN114973659A (zh) * 2022-05-12 2022-08-30 山东高速集团有限公司创新研究院 一种高速公路的间接事件检测的方法、装置及系统
CN115206092A (zh) * 2022-06-10 2022-10-18 南京工程学院 一种基于注意力机制的BiLSTM和LightGBM模型的交通预测方法
CN115206092B (zh) * 2022-06-10 2023-09-19 南京工程学院 一种基于注意力机制的BiLSTM和LightGBM模型的交通预测方法
CN116311941A (zh) * 2023-03-24 2023-06-23 阿波罗智联(北京)科技有限公司 主要车流路径提取方法、装置、设备及介质
CN116453333A (zh) * 2023-03-24 2023-07-18 阿波罗智联(北京)科技有限公司 用于预测主要车流路径的方法及模型训练方法
CN116311941B (zh) * 2023-03-24 2023-12-26 阿波罗智联(北京)科技有限公司 主要车流路径提取方法、装置、设备及介质
CN116453333B (zh) * 2023-03-24 2024-04-16 阿波罗智联(北京)科技有限公司 用于预测主要车流路径的方法及模型训练方法

Also Published As

Publication number Publication date
CN111915877A (zh) 2020-11-10

Similar Documents

Publication Publication Date Title
WO2020224445A1 (zh) 一种车流路径分布信息的处理方法、装置及电子设备
US10846052B2 (en) Community discovery method, device, server and computer storage medium
Bedogni et al. The Bologna ringway dataset: Improving road network conversion in SUMO and validating urban mobility via navigation services
US9207090B2 (en) System and method for dynamic path optimization
US10608879B2 (en) Validation using natural language processing
TW201901116A (zh) 用於數位路線規劃的系統以及方法
US11574254B2 (en) Adaptive asynchronous federated learning
US11748305B2 (en) Suggesting a destination folder for a file to be saved
US20230128318A1 (en) Automated Parameterized Modeling And Scoring Intelligence System
CN111462489A (zh) 交通拥堵区域预测方法及装置
CN114357105A (zh) 地理预训练模型的预训练方法及模型微调方法
Hasan et al. Graphettes: Constant-time determination of graphlet and orbit identity including (possibly disconnected) graphlets up to size 8
Ren et al. A measure of identifying influential waypoints in air route networks
US10169508B2 (en) Efficient deployment of table lookup (TLU) in an enterprise-level scalable circuit simulation architecture
US11782918B2 (en) Selecting access flow path in complex queries
US20220284023A1 (en) Estimating computational cost for database queries
Golasowski et al. Alternative paths reordering using probabilistic time-dependent routing
CN111915875A (zh) 一种车流路径分布信息的处理方法、装置及电子设备
US9835460B2 (en) GPS map-matching based on space map-matching
KR102608343B1 (ko) 패턴 매칭을 이용하여 도로 정체 전파를 예측하는 방법, 컴퓨터 장치, 및 컴퓨터 프로그램
US20240085205A1 (en) Method and apparatus for machine learning-based prediction of an estimated time of arrival
CN112652161B (zh) 一种车流路径分布信息的处理方法、装置、及电子设备
US11914593B2 (en) Generate digital signature of a query execution plan using similarity hashing
Trinh et al. Combining multi-agent deep deterministic policy gradient and rerouting technique to improve traffic network performance under mixed traffic conditions
US20200167641A1 (en) Contrastive explanations for interpreting deep neural networks

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20802163

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20802163

Country of ref document: EP

Kind code of ref document: A1