Information processing method and information processing device
The method efficiently determines the impact of events on traffic conditions by setting a target link and expanding to adjacent links based on travel time histogram comparisons and threshold corrections, improving accuracy in traffic condition assessments.
Patent Information
- Application Number
- JP2021023892
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-02-18
- Publication Date
- 2024-11-28
- Estimated Expiration
- 2041-02-18
AI Technical Summary
Existing methods for determining the impact of events on traffic conditions are inefficient due to the need to analyze a large number of roads within divided areas, leading to low processing efficiency.
An information processing method that sets a target link related to an event location, determines its influence, and expands the determination to adjacent links based on a comparison of travel time histograms before and after the event, using threshold corrections for accuracy.
This approach allows for efficient and accurate determination of the impact of events on traffic conditions by expanding the target area dynamically and using threshold adjustments, enhancing the precision of traffic condition assessments.
Smart Images

Figure 0007591422000007 
Figure 0007591422000008 
Figure 0007591422000009
Abstract
Description
[Technical field]
[0001] The present invention relates to an information processing method and an information processing device. [Background technology]
[0002] Conventionally, methods have been proposed for determining the impact of events such as sporting events and special events on traffic conditions. For example, Patent Document 1 discloses a method for dividing an area into a plurality of areas in a grid pattern based on latitude and longitude, and predicting congestion using the average congestion level on weekdays and the average congestion level on holidays in the prediction target area. Patent Document 1 also discloses a method for selecting prediction target areas, such as selecting areas with many residents as prediction target areas, selecting areas with POIs as prediction target areas, and selecting areas based on the past congestion level for each area. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2016-110360 A Summary of the Invention [Problem to be solved by the invention]
[0004] However, in the process of determining traffic conditions for each area obtained by dividing a region, it is necessary to determine the traffic conditions on a large number of roads included in the area, which poses a problem of low efficiency. The present invention has been made in view of the above background, and has an object to provide a method capable of efficiently performing a process for determining the effect of an event on traffic conditions. [Means for solving the problem]
[0005] As a first aspect for achieving the above object, an information processing method for determining an influence of an event on a traffic situation includes a first link setting step of setting a link related to a position where the event occurs as a target link, a determination step of determining a degree of influence of the event on the traffic situation of the target link, and a second link setting step of setting a link adjacent to the target link as the new target link based on a determination result of the determination step. The process of the determination step is performed on the target link identified in the second link setting step. and a step of: generating a first histogram showing a distribution of travel times of the target link in a first time period and a second histogram showing a distribution of travel times of the target link in a second time period that is further away from the occurrence time of the event than the first time period; calculating a degree of difference between the first histogram and the second histogram; and comparing the degree of difference with a predetermined threshold value to determine a degree of influence of the event on the target link. An information processing method is also provided. The determining step may be configured to calculate the degree of difference using any one of the following formulas (1), (2), (3), (4), and (5).
[0006] In the above information processing method, the second link setting step may not be executed if it is determined in the determining step that the degree of influence of the event on the traffic condition of the target link is low.
[0007] In the above information processing method, in the determining step, a degree of influence of the event on a traffic situation of the target link may be determined for each travel direction on the target link.
[0008] In the above information processing method, in the determining step, if the target link is a large road, the degree of influence of the event on traffic conditions of the target link may be determined for each travel direction on the target link.
[0011] In the above information processing method, in the determining step, the threshold value may be corrected based on a difference in spread of distribution between the first histogram and the second histogram, and the corrected threshold value may be used to perform the determination.
[0012] In the above information processing method, in the determination step, when the target link is a road on which traffic can be traveled in a first direction and a second direction opposite to the first direction, the threshold value used to determine the degree of impact of the event on traffic conditions in the first direction of the target link may be corrected based on traffic conditions in the second direction of the target link.
[0013] In the above information processing method, in the determination step, if the target link is not a large road, the threshold value used when determining the degree of impact of the event on traffic conditions in the first direction of the target link may be corrected based on traffic conditions in the second direction of the target link.
[0014] In the above information processing method, the first link setting step may be configured to set, as the target link, a link close to a facility of a type associated with a type of the event among facilities associated with the occurrence position.
[0015] As a second aspect for achieving the above object, there is provided an information processing device for determining an influence of an event on a traffic situation based on road data related to nodes and links, the information processing device setting a link related to a position where the event occurs as a target link, and executing a determination process for determining an influence degree of the event on the traffic situation of the target link, Based on the result of the determination process, a link adjacent to the target link is set as the new target link, and the determination process is performed on the identified target link. As the determination process, a first histogram showing a distribution of travel times of the target link in a first time period and a second histogram showing a distribution of travel times of the target link in a second time period that is further away from the occurrence time of the event than the first time period are created, a difference between the first histogram and the second histogram is calculated, and the difference is compared with a predetermined threshold value, thereby executing a process of determining an impact degree of the event on the target link. An example of such an apparatus is an information processing apparatus. Effect of the Invention
[0016] According to the above configuration, by setting a new target link based on the results of judging the degree of impact of the event on the traffic conditions of the target link, the scope of judgment is expanded, so that the process of judging the impact of the event on the traffic conditions can be performed efficiently. [Brief description of the drawings]
[0017] [Figure 1] FIG. 1 is a schematic configuration diagram of an information processing device. [Diagram 2]FIG. 4 is a schematic diagram showing an example of the configuration of start position setting data. [Diagram 3] 4 is a flowchart showing the operation of the information processing device. [Figure 4] 4 is a flowchart showing the operation of the information processing device. [Diagram 5] 4 is a flowchart showing the operation of the information processing device. [Figure 6] FIG. 13 is an explanatory diagram of a link searching operation. [Figure 7] FIG. 13 is an explanatory diagram of a link searching operation. [Figure 8] 11A and 11B are diagrams showing examples of histograms created by the information processing device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0018] [1. Configuration of information processing device] 1 is a schematic diagram of an information processing device 1 according to an embodiment of the present invention. The information processing device 1 is a computer that processes data related to road traffic conditions. The information processing device 1 is connected to a traffic database 50 via a communication network 2.
[0019] The information processing device 1 of this embodiment generates prediction data 55 for predicting the effect of an event on traffic conditions by determining, for each link, the effect that the event will have on traffic conditions when the event is held.
[0020] In the following description, an event refers to an occurrence including an exhibition, a sporting event, an entertainment event, a commercial event, a political or non-political gathering, and other artificially organized occasions. An event may include an unintentionally occurring artificial event such as an accident, and an unintentional event such as a natural phenomenon including a disaster. An event may also include an event that occurs due to the above various events.
[0021] The information processing device 1 includes a control unit 10, a communication unit 31, an input unit 32, and an output unit 33. The control unit 10 includes a processor 11 and a memory 21. The processor 11 is configured with a CPU (Central Processing Unit) or a microcontroller. The memory 21 stores programs and data executed by the processor 11. The memory 21 may be a non-volatile storage device that stores programs and data in a non-volatile manner. The memory 21 may also be a volatile storage device that forms a work area for the processor 11.
[0022] The processor 11 includes an information acquisition unit 12 and a processing unit 13 . The information acquisition unit 12 controls the communication unit 31 to acquire data from the traffic database 50 via the communication network 2 . The processing unit 13 processes the data acquired by the information acquisition unit 12. The processing unit 13 transmits the processed data to the traffic database 50 via the communication unit 31.
[0023] The communication unit 31 is a communication interface device having a connector for connecting to the communication network 2, a transmitting / receiving circuit, an encoder, a decoder, etc. The communication unit 31 executes data communication with the traffic database 50 under the control of the control unit 10. The input unit 32 is an input interface that accepts operations by an operator of the information processing device 1. For example, input devices such as a keyboard, a mouse, and a track pad are connected to the input unit 32. The input unit 32 acquires an operation signal input from the input device and outputs data indicating the operation content to the control unit 10. The output unit 33 outputs information under the control of the control unit 10. The output unit 33 may be connected to a display, for example, and cause an image to be displayed on the display. The output unit 33 may also be connected to a printer, and cause the printer to execute printing.
[0024] The information processing device 1 may be configured to be connected to a terminal device (not shown) via the communication unit 31 and to operate by accepting remote access from the terminal device. Although Fig. 1 shows an example in which the information processing device 1 is configured as a separate entity from the traffic database 50, the information processing device 1 may be the same server device as the traffic database 50. The information processing device 1 and the traffic database 50 may also be configured as a system in which multiple server devices perform distributed processing.
[0025] The traffic database 50 stores road data 51 , facility data 52 , traffic data 53 , start position setting data 54 , and prediction data 55 .
[0026] The road data 51 is geographic data related to roads, and includes information on nodes and links. For example, the road data 51 includes, for nodes, node numbers, position coordinates, altitude, node types, number of connecting links, connecting node numbers, intersection names, and the like. For links, the road data 51 includes, for example, link numbers, road types, route numbers, information on heavily used routes, link lengths, service status, road width classifications, number of lanes, roadway width, center lane width, position coordinates of interpolation points, altitude of interpolation points, expressway numbering, and the like. The link numbers may be node numbers at the start and end points of the link. For links, the road data 51 may also include attributes such as bridges, overpasses, tunnels, caverns, railroad crossings, pedestrian bridges, and underpasses.
[0027] The facility data 52 includes data on the locations, names, etc. of facilities such as local government offices, service areas, parking areas, roadside stations, ferry terminals, train stations, airports, etc.
[0028] The traffic data 53 includes data on traffic conditions of links included in the road data 51. Specifically, the traffic data 53 includes traffic volume and link travel time. The traffic data 53 includes traffic volume and link travel time in association with a link, a traffic direction on the link, and a date / time segment including a date and a time zone. The traffic data 53 includes a plurality of data corresponding to one link, a traffic direction, and a date / time segment. The traffic data 53 may be composed of data on vehicle traffic, but may also include data on a plurality of types of moving objects. For example, the traffic data 53 may include vehicle traffic volume and link travel time, and pedestrian traffic volume and link travel time. Travel time refers to the time required for a vehicle to pass through a link.
[0029] The traffic data 53 is data obtained by observing and tabulating traffic conditions of links included in the road data 51. The traffic data 53 includes, for example, data observed on the date and time when an event occurred in the past, the time period when the event was held, and other time periods. For example, the traffic data 53 includes past data observed by a system that observes traffic conditions at regular time intervals for links included in the road data 51.
[0030] The start position setting data 54 is data for specifying the type of facility that is to be the initial position for determination in the process of determining the effect of an event on the traffic conditions of a link.
[0031] FIG. 2 is a schematic diagram showing an example of the configuration of the start position setting data 54. As shown in FIG. As shown in FIG. 2, the start position setting data 54 associates the type of facility that is the initial position for determination with the type of event. For example, when the type of event is an academic conference and an exhibition, a station is associated with the facility that is the initial position for determination, i.e., the start position. An academic conference includes, for example, an academic conference or other conference. Also, for example, when the type of event is a sporting event such as baseball, a parking lot is associated with the facility that is the start position.
[0032] The traffic database 50 may have start position setting data 54 for each region. For example, the traffic database 50 may have start position setting data 54 applied to urban areas and start position setting data 54 applied to suburban areas.
[0033] [2. Operation of information processing device] Figures 3, 4, and 5 are flowcharts showing the operation of the information processing device 1. Figures 6 and 7 are explanatory diagrams of the operation of the information processing device 1 searching for links. Figure 8 is a diagram showing an example of a histogram created by the information processing device 1. Hereinafter, the operation of the information processing device 1 will be described with reference to these figures.
[0034] [2-1. Overall sequence] The control unit 10 determines a start node or start link related to the position where the event occurred (step S11). Step S11 will be described in detail. The control unit 10 refers to the start position setting data 54 and identifies the type of facility associated with the type of event. The control unit 10 sets the start node or start link based on a facility at the position where the event occurred (referred to as an "event occurrence facility") and / or a facility associated with the position where the event occurred and of a type specified in the start position setting data 54 (referred to as an "associated type facility"). The facility associated with the position where the event occurred may be a facility closest to the position where the event occurred or a facility within a specified range from the position where the event occurred. When the control unit 10 sets a start node as a start position of the process, the control unit 10 sets the event occurrence facility, the node closest to the event occurrence facility, or the node associated with the facility, among the nodes included in the road data 51, as the start node. Alternatively or in addition, when the control unit 10 sets a start node as a start position of the process, the control unit 10 sets the node closest to the related type facility, or the node associated with the facility, among the nodes included in the road data 51, as the start node. Here, examples of the "node associated with the facility" include an intersection with the name or abbreviation of the facility, and an intersection connected to a road link adjacent to the facility. Furthermore, when the control unit 10 sets a start link, the control unit 10 sets a road link adjacent to the event occurrence facility, or a road link associated with the facility, among the links included in the road data 51. Alternatively or in addition, when the control unit 10 sets a start link as a start link, the control unit 10 sets a road link adjacent to the related type facility, or a road link associated with the facility, among the links included in the road data 51. Here, the "link associated with a facility" may be, for example, a link that connects to the entrance or exit of the facility or a facility adjacent to the facility, a road link that has the name or abbreviation of the facility, etc. These are collectively referred to as links close to the facility where the event occurs.
[0035] The control unit 10 sets a search range centered on the start node or the start link (step S12). In step S12, the control unit 10 sets the search range to, for example, links connected to the start node, links adjacent to the start node, or a range including the start link. As a result, a link related to the occurrence position of the event is set as the target link.
[0036] The control unit 10 acquires traffic data 53 related to links within the search range (step S13). The control unit 10 classifies the traffic data 53 acquired in step S13 into data when an event has occurred and data when an event has not occurred (step S14). In step S14, the control unit 10 classifies the data by date and time, for example. In detail, the control unit 10 classifies the data into traffic data 53 including the date and time or time period when an event has occurred, or traffic data 53 for a time period close to the date and time when an event has occurred, and traffic data 53 for other time periods.
[0037] The control unit 10 selects one link to be subjected to the determination from among the links within the search range, and sets the link as the subject link (step S15). The control unit 10 executes a determination process (step S16). The determination process is a process for determining the degree of influence that the event has had on the traffic conditions of the target link. For example, the determination process determines whether or not the traffic conditions of the target link have been influenced by the occurrence of the event. Alternatively, the control unit 10 uses the determination process to calculate an index of the degree of influence that the event has had on the traffic conditions of the target link. The determination process will be described later with reference to FIG. 4.
[0038] The control unit 10 temporarily holds the result of the determination process in association with the target link (step S17). For example, the control unit 10 stores the result of the determination process and information indicating the determined link in the memory 21 in association with each other.
[0039] The control unit 10 judges whether or not there is a link within the search range for which the judgment process has not been performed (step S18). If there is a link for which the judgment process has not been performed (step S18; YES), the control unit 10 returns to step S15 and selects the next target link (step S15).
[0040] If there is no link for which the determination process has not been performed (step S18; NO), the control unit 10 determines whether or not there is a link that is a candidate for determination outside the search range based on the determination result held in step S17 (step S19). The candidate link is specifically a link that is highly influenced by the event, or a link adjacent to a link determined to be influenced by the event, and is outside the search range. The determination in step S19 is made based on the determination result held in step S17. If there is a candidate link (step S19; YES), the control unit 10 changes (expands) the search range to include the candidate link (step S20) and returns to step S13. If there are multiple links that are determined to be highly influenced by the event, and if there are multiple links adjacent to a link determined to be highly influenced by the event, the search range is changed (expanded) to include all of these multiple links.
[0041] Changing (expanding) the search range will be described with reference to Figures 6 and 7. In the following description, this will simply be referred to as "expanding." Fig. 6 shows how a search range is set to a range including nodes N1 to N14 and links L1 to L17. The example in Fig. 6 is an example in which the control unit 10 sets node N1 as the start node. In this example, the control unit 10 sets links L1, L2, and L3 connected to node N1 as the search range.
[0042] The control unit 10 executes a determination process for the links L1, L2, and L3. If it is determined that the link L1 is affected by the event, the links adjacent to the link L1 become candidate links. In the example of Fig. 6, the links adjacent to the link L1 are the links L4 and L5 that connect to the node N2, which is the end point of the link L1.
[0043] Here, an example is shown in which links L1 and L2 are determined to be affected by the event, and link L3 is determined to be not affected by the event. In other words, links L4 and L5 adjacent to link L1 and links L6 and L7 adjacent to link L2 are candidate links, and links L8, L9, and L10 adjacent to link L3 are not candidate links. Based on these determination results, the control unit 10 expands the search range.
[0044] FIG. 7 shows the state after the search range is expanded from the state shown in FIG. 7, the search range is expanded to include links L4, L5, L6, and L7. After this, the control unit 10 performs a determination on the links L4, L5, L6, and L7 that are newly included in the search range due to the expansion of the search range.
[0045] Returning to FIG. 3, if there is no candidate link (step S19; NO), the control unit 10 generates impact estimation data based on the determination result held in step S17 (step S21). The impact estimation data includes data indicating the link affected by the traffic situation due to the occurrence of the event. The impact estimation data may include information such as the node adjacent to the link affected by the traffic situation due to the occurrence of the event, the time period when the traffic situation of the link is affected, and the type of event. The control unit 10 generates a prediction model based on the impact estimation data (step S22) and ends the operation. The prediction model is obtained by a learning model of artificial intelligence (AI) performing machine learning using the impact estimation data as learning data. It is possible to evaluate the impact on the traffic situation when a virtual event occurs using the prediction model. Step S22 is a step of making the learning model perform learning, and may be performed after the impact estimation data is accumulated. For example, the control unit 10 may accumulate and store the impact estimation data generated in step S21 in the traffic database 50. In this case, the process of making the machine learning model learn the impact estimation data may be performed by a device different from the information processing device 1. Furthermore, the control unit 10 or the traffic database 50 may organize and consolidate data items of the impact estimation data to generate new learning data. The learning performed by the learning model may be supervised learning, or other learning methods may be adopted. Furthermore, the newly generated impact estimation data may be used to cause the trained prediction model to perform further learning.
[0046] In the above operation, steps S12 to S15 correspond to an example of a first link setting step, step S16 corresponds to an example of a determination step, and step S19 corresponds to an example of a second link setting step.
[0047] [2-2. Judgment process] FIG. 4 shows in detail the determination process shown in step S16 of FIG. The control unit 10 classifies the traffic data 53 of the target link by the travel direction and the time period (step S31). For example, if the target link is a road extending in the north-south direction, the control unit 10 classifies the traffic data 53 into southbound data and northbound data.
[0048] The control unit 10 further classifies the traffic data 53 by time period. For example, the control unit 10 classifies the traffic data 53 into data for a first time period and data for a second time period. The first time period is a time period during which the link to be judged is estimated to be affected by an event. For example, the first time period is a time period that includes the date and time when the event occurred or the time period when the event occurred. The first time period may also be a time period close to the date and time when the event occurred or the time period when the event occurred. If the target link is far away from the location where the event occurred, it is estimated that there is a large time difference between the date and time when the event occurred and the timing when the target link is affected. In such a case, the control unit 10 may set the first time period to a time period that does not include the date and time when the event occurred and the time period when the event occurred. The second time period is a time period that does not include the date and time when the event occurred and the time period when the event occurred. band The second time period is a time period that is farther from both the date and time of the event occurrence and the time period in which the event occurred than the first time period. The second time period is a time period in which the event is estimated not to have affected the target link.
[0049] The control unit 10 creates a histogram of the first time period and the second time period for each travel direction (step S32). In step S32, the control unit 10 classifies the travel time of the target link into classes in predetermined time units, and creates a histogram with the classes on the horizontal axis and the frequency on the vertical axis.
[0050] Specifically, in step S32, the control unit 10 creates a first histogram for one traffic direction based on the traffic data 53 for a first time period. Furthermore, for this traffic direction, the control unit 10 creates a second histogram for the same traffic direction based on the traffic data 53 for a second time period.
[0051] Fig. 8 is an example of a histogram created by the control unit 10, and shows two histograms created by the control unit 10 for one traffic direction of one target link. Histogram H1 corresponds to the first histogram, and histogram H2 corresponds to the second histogram. For ease of understanding, Fig. 8 illustrates approximate curves showing the frequency distribution of traffic data 53 as histograms H1 and H2.
[0052] In the example of Figure 8, the histograms differ due to the effect of an event. Histogram H1 has wider tails than histogram H2, and contains many data in classes farther from the median. For example, if the maximum value of the data in histogram H1 in Fig. 8 is Va, and the maximum value of the data in histogram H2 is Vb, then histogram H1 contains many data between Va and Vb. Data with high class values indicates that the travel time of the link is long, so histogram H1 in Fig. 8 indicates that the event caused traffic congestion or congestion on the target link.
[0053] In the figure, the average value of histogram H1 is shown as M1, and the average value of histogram H2 is shown as M2. The average values M1 and M2 are strongly influenced by the high frequency data of the data of histograms H1 and H2, so the difference in the low frequency classes is not easily reflected. In the example of FIG. 8, the difference between the average value M1 and the average value M2 is also clearly smaller than the difference between the class value Va and the class value Vb. In this way, when the average value is used as an index showing the difference between the histograms, the difference between the histograms H1 and H2 is evaluated to be small, making it difficult to accurately determine the impact of the event.
[0054] 4, the control unit 10 calculates the difference between the distributions of the first histogram in the first time period and the second histogram in the second time period (step S33). For example, the control unit 10 calculates a distance index between the first histogram and the second histogram as an index indicating the difference between the distributions of the histograms.
[0055] As a distance index, for example, the KL divergence shown in the following formula (1) can be used. Also, the JS divergence shown in the following formula (2) may be used. In the following formulas (1), (2), and (3), Xn is a vector representing the data of the first histogram, Xe is a vector representing the data of the second histogram, and both Xn and Xe are vectors of the same size. Also, i is a class index.
[0056]
number
[0057] As a more preferable distance index, distance index D shown in formula (3) below may be used. Distance index D emphasizes and reflects the difference in low frequency classes between the first and second histograms. Therefore, by using distance index D in formula (3) below, it is possible to obtain the difference between the first and second histograms while reflecting the influence of data with low frequency. In formula (3) below, the root sign of a vector means a vector obtained by applying a root sign to each element of the vector. Furthermore, T is a symbol that represents the transposition of a vector.
[0058]
number
[0059] Similarly, as a preferable distance index, for p obtained by the following formula (4), an estimate of p may be obtained by the following formula (5) and used as the distance index. In the following formula (4), P(a) is a symbol representing the probability of a.
[0060]
number
[0061] Furthermore, an index indicating the spread of the tail of the histogram may be used as an index indicating the difference between the first and second histograms. As this index, the difference A between the B percentile value of the second histogram and the maximum value of the data included in the first histogram is calculated, and this difference can be used as the index. B is set arbitrarily from natural numbers. In the example of FIG. 8, the difference A between the 90th percentile value Vx of the histogram H2 and the maximum value Va of the data included in the histogram H1 can be used as the index.
[0062] If the target link is a road having a plurality of travel directions, the control unit 10 calculates the distribution difference of the histogram for each travel direction in step S33. The control unit 10 sets a threshold value for determining the difference between the distributions of the histograms (step S34). The threshold value is a preset value, and is stored in the memory 21 or the traffic database 50.
[0063] The control unit 10 executes a threshold correction process to correct the threshold in accordance with the scale of the road that is the target link (step S35). The details of the threshold correction process will be described later. The control unit 10 compares the distribution difference calculated in step S33 with a threshold value to determine the degree of influence of the event on the target link (step S36).
[0064] In step S36, for example, if the value of the distribution difference calculated in step S33 is equal to or greater than a threshold, the control unit 10 determines that the event has affected the target link, and if the value of the distribution difference is less than the threshold, the control unit 10 determines that the event has not affected the target link. If the target link is a road with multiple traffic directions, in step S36, the control unit 10 compares the distribution difference of the histogram for each traffic direction with each threshold. In this way, the control unit 10 determines the degree of impact of the event for each traffic direction. Furthermore, the control unit 10 combines the determination results for each traffic direction to determine the determination result for the target link.
[0065] For example, if it is determined that the event has affected any of the travel directions of the target link, it is determined that the event has affected the target link. If it is determined that the event has not affected any of the travel directions, it is determined that the event has not affected the target link. After obtaining the determination result in step S36, the control unit 10 proceeds to step S17.
[0066] In the process of Fig. 4, the control unit 10 determines the degree of influence of the event on the traffic situation of the target link for each traffic direction of the target link. The control unit 10 performs the above operation when the target link is a large-scale road, and may synthesize traffic data 53 for each traffic direction of the target link when the target link is not a large-scale road. In this case, a histogram is created for the target link without limiting the traffic direction, i.e., without classifying the traffic direction, and the degree of influence of the event is determined for the traffic situation combining all traffic directions.
[0067] In this way, when the target link is a large road, the control unit 10 determines the degree of impact of the event on the traffic conditions of the target link for each direction of travel on the target link, and when the target link is not a large road, the control unit 10 determines the degree of impact of the event on the traffic conditions of the target link without classifying the direction of travel on the target link. When the road is small, traffic in one direction is easily affected by traffic in the other direction, so for small roads, it may be preferable to determine the impact of an event without distinguishing between traffic directions.
[0068] Furthermore, when the road is small in scale, the number of vehicles passing through it is not high, so the number of data items in the traffic data 53 tends to be small. When determining the impact of an event on such a road, if the average value of travel time or the like is used as in conventional methods, the accuracy of the determination may decrease due to the small amount of data. For example, in Patent Document 1, congestion prediction is performed using the average congestion degree on weekdays and the average congestion degree on holidays in the area to be predicted, but since the average of the data is used, the accuracy decreases when the number of data items is insufficient or when the data includes outliers. In contrast, in this embodiment, by calculating the difference between the distributions of the histograms of the target link, it is possible to make a judgment with higher accuracy than when using the average of the data related to the traffic conditions. Furthermore, by synthesizing the traffic data 53 by traffic direction for roads that are not large, it is possible to increase the apparent number of data used to create the histogram, and the judgment accuracy can be further improved.
[0069] [2-3. Threshold correction processing] FIG. 5 shows in detail the threshold correction process shown in step S35 of FIG. The control unit 10 corrects the threshold value if the target link is not a large-scale road. A large-scale road is one that has a median strip and / or a width that exceeds a set value. The control unit 10 judges whether the target link is a large-scale road or not in steps S51 to S53. In the following description, a road that does not have a median strip and has a width equal to or less than a set value is called a "small-scale road" as an example of a road that is not a large-scale road.
[0070] The control unit 10 judges whether or not the target link is a road having a median strip (step S51). If the target link is a road having a median strip (step S51; YES), the control unit 10 proceeds to step S36.
[0071] If the target link is not a road with a median strip (step S51; NO), the control unit 10 judges whether the target link is a one-way road (step S52). If the target link is a one-way road (step S52; YES), the control unit 10 proceeds to step S36.
[0072] If the target link is not a one-way road (step S52; NO), the control unit 10 judges whether the width of the target link is equal to or smaller than a set value (step S52). If the width is greater than the set value (step S53; NO), the control unit 10 proceeds to step S36.
[0073] If the width of the target link is equal to or smaller than the set value (step S53; YES), the control unit 10 corrects the threshold value (step S54). S36 transition to. In step S54, the control unit 10 generates different threshold values for each travel direction by correcting the threshold values set in step S34 (FIG. 4).
[0074] As described above, it is difficult to make an accurate judgment for a narrow road because the number of data items in the traffic data 53 is small. In this embodiment, the accuracy of the judgment is improved by correcting the threshold value using the traffic data of the oncoming lane.
[0075] Here, the travel direction of the target link is defined as a first direction and a second direction opposite to the first direction. The control unit 10 calculates the feature amount T R1 , and the tail feature of the histogram in the second direction T R2 Calculate. Feature quantity T of the tail of the histogram R1 , T R2 is an index showing the difference between the first histogram and the second histogram in a class with a small number of data (frequency). For example, the difference A between the B percentile value of the second histogram and the maximum value of the data included in the first histogram, or a value calculated from the difference A, is used as the feature quantity T R1 , T R2 B can be arbitrarily set from natural numbers.
[0076] In addition, the control unit 10 determines the threshold T R1,base , and the threshold value T in the second direction R2,base Obtain the threshold T R1,base , T R2,base is a basic threshold value in the first direction and the second direction, and is set in step S34. R1,base , T R2,base may be the same value. The control unit 10 calculates the threshold value in the first direction of the target link by the following formula (6), and calculates the threshold value in the second direction by the following formula (7).
[0077] Thr R1 =T R1,base +αT R2 (6) Thr R2 =T R2,base +αT R1 (7) Here, α is a predetermined constant.
[0078] In this way, in the threshold correction process, if the target link is not a large-scale road, the threshold used to determine the degree of influence of the event on the traffic conditions in the first direction of the target link is corrected based on the traffic conditions in the second direction of the target link. Also, if the target link is a large-scale road, the process of correcting the threshold used to determine the degree of influence of the event on the traffic conditions in the first direction of the target link based on the traffic conditions in the second direction is not performed. This makes it possible to reflect the traffic conditions of the oncoming lane in the process of determining the degree of influence of the event on a small road. On a narrow road that allows traffic in both directions, it is possible that congestion or jams will occur in the other direction due to the influence of congestion or jams in one direction. Therefore, by reflecting the traffic conditions of the oncoming lane, it is possible to determine with high accuracy whether traffic congestion or jams have occurred due to the influence of the event.
[0079] 3. Other embodiments The above embodiment shows a specific example to which the present invention is applied, and does not limit the form to which the invention is applied.
[0080] In the threshold correction process shown in FIG. 5, an example is shown in which the threshold is corrected for a small road, but the threshold may also be corrected for a target link that is not a small road. For example, in step S34, the control unit 10 may correct the threshold value stored in the memory 21 or the traffic database 50 based on factors related to the target link. The factors related to the target link include, for example, the geographical relationship between the target link and the start node or the event occurrence facility, the consistency of traffic conditions between the target link and adjacent links, and the time period during which the impact of the traffic conditions is evaluated. The geographical relationship between the target link and the start node or the event occurrence facility is, for example, the distance from the start node or the event occurrence facility to the target link. The consistency of traffic conditions refers to whether or not the distribution difference between the histograms in the target link and the distribution difference between the histograms in the links adjacent to the target link show the same tendency. The time period during which the impact of the traffic conditions is evaluated refers to whether or not the time difference between the first and second time periods related to the creation of the histograms and the date and time of the event is equal to or greater than a threshold. By correcting the threshold value based on at least one of these factors, the accuracy of the determination can be improved. For example, it is possible to expect effects such as eliminating the effects of traffic congestion and jams caused by factors unrelated to the event, and suppressing the effects of variations in the occurrence of traffic congestion and jams between links.
[0081] Also, for example, in step S34, the control unit 10 may correct the threshold stored in the memory 21 or the traffic database 50 based on the histogram of the target link. In this case, the control unit 10 may correct the threshold according to the difference in the spread of the tails between the first histogram and the second histogram. For example, the control unit 10 may be configured to loosely correct the threshold when the above-mentioned difference A is equal to or greater than a set value, making it easier to determine that the event has had an effect, and to strictly correct the threshold when the difference A is less than the set value, making it harder to determine that the event has had an effect. In this case, the change in the class with a small amount of data in the histogram can be strongly reflected in the determination result, and the impact of the event can be determined more accurately.
[0082] In the above embodiment, the control unit 10 performs the determination in step S16, and determines whether or not there is a candidate link based on the result of the determination. In this example, the determination of the impact of the event on the target link and the determination of whether or not to expand the search range are performed using substantially the same criteria.
[0083] When determining the presence or absence of a candidate link in step S19, the control unit 10 may perform the determination based on a criterion different from that of the determination process in step S16. That is, the determination may be performed using a plurality of thresholds in the determination step. For example, the control unit 10 may perform a second determination to identify the presence or absence of a candidate link using a different threshold value after performing the determination on the target link in step S36. In this second determination, a threshold value that is looser than the determination in step S36 may be used. Specifically, a threshold value for the second determination is set so that for a target link determined in step S36 to have no influence of the event, a link adjacent to the target link is determined to be a candidate link. In this case, the control unit 10 may set a threshold value for the second determination for each travel direction after the determination in step S36. In this case, in the process of expanding the search range, all links that may be influenced by the event can be included in the search range, and the influence of the event can be determined more accurately.
[0084] [4. Configuration supported by the above embodiment] The above embodiment is a specific example of the following configuration.
[0085] (Clause 1) An information processing method for determining the impact of an event on traffic conditions, comprising: a first link setting step of setting a link related to the location where the event occurred as a target link; a determination step of determining the degree of impact the event has on the traffic conditions of the target link; and a second link setting step of setting a link adjacent to the target link as the new target link based on the determination result of the determination step, wherein the determination step is performed on the target link identified in the second link setting step. According to the information processing method of paragraph 1, the scope of judgment is expanded by setting a new target link based on the result of judging the degree of impact of the event on the traffic conditions of the target link, so that the process of judging the impact of the event on the traffic conditions can be performed efficiently.
[0086] (2) The information processing method according to claim 1, wherein, when it is determined in the determining step that the impact of the event on the traffic conditions of the target link is low, the second link setting step is not executed. According to the information processing method of the second aspect, links that are unlikely to be affected by the event are not subject to the judgment, so that the process of judging the impact of the event on the traffic conditions can be performed more efficiently.
[0087] (Item 3) The information processing method according to item 1 or 2, wherein in the determining step, a degree of influence of the event on a traffic situation of the target link is determined for each direction of travel on the target link. According to the information processing method of the third aspect, the effect of an event on traffic conditions can be determined with higher accuracy.
[0088] (4) An information processing method according to claim 3, wherein in the determination step, if the target link is a large road, the degree of influence of the event on the traffic conditions of the target link is determined for each direction of travel on the target link. According to the information processing method of paragraph 4, taking into consideration that there may be less data on traffic conditions on roads that are not large, the impact of an event on traffic conditions can be determined with greater accuracy.
[0089] (Clause 5) An information processing method described in any one of clauses 1 to 4, wherein in the determination step, a first histogram showing a distribution of travel times for the target link in a first time period and a second histogram showing a distribution of travel times for the target link in a second time period that is further away from the occurrence time of the event than the first time period are created, and the impact of the event on the traffic conditions of the target link is determined based on the difference between the distributions of the first histogram and the second histogram. According to the information processing method of paragraph 5, multiple histograms are created for each time period, and a judgment is made using the difference between the distributions of the histograms. This reduces the influence of outliers in the data and the influence of a small amount of data, and makes it possible to judge the impact of an event on traffic conditions with a higher degree of accuracy.
[0090] (Item 6) An information processing method as described in Item 5, in which, in the judgment step, a distance index between the first histogram and the second histogram is calculated, and the distance index is compared with a threshold value to judge whether or not the event has an impact on the traffic conditions of the target link. According to the information processing method of the sixth aspect, by using the distance indices of multiple histograms, the difference in histograms appearing on a target link can be accurately evaluated. This makes it possible to determine the impact of an event on traffic conditions with a higher degree of accuracy.
[0091] (Item 7) The information processing method according to item 6, wherein in the judgment step, the threshold is corrected based on a difference in distribution spread between the first histogram and the second histogram, and judgment is made using the corrected threshold. According to the information processing method of paragraph 7, the threshold is corrected based on the difference in spread between a plurality of histograms, thereby making it possible to determine with higher accuracy the effect of an event on traffic conditions.
[0092] (Clause 8) An information processing method described in clause 6 or 7, wherein in the judgment step, when the target link is a road on which traffic can be traveled in a first direction and a second direction opposite to the first direction, the threshold value used when judging the degree of impact of the event on traffic conditions in the first direction of the target link is corrected based on traffic conditions in the second direction of the target link. According to the information processing method of paragraph 8, the traffic conditions on a target link can be evaluated and judged taking into account the traffic conditions in the opposite direction, thereby making it possible to judge the effect of an event on the traffic conditions with a higher degree of accuracy.
[0093] (Clause 9) An information processing method as described in Clause 8, wherein in the judgment step, if the target link is not a large road, the threshold value used when judging the degree of impact of the event on the traffic conditions in the first direction of the target link is corrected based on the traffic conditions in the second direction of the target link. According to the information processing method of paragraph 9, when the target link is a small road, the traffic conditions on the target link are evaluated taking into account the traffic conditions in the opposite direction. This makes it possible to determine the effect of an event on the traffic conditions with a higher degree of accuracy.
[0094] (Clause 10) An information processing method described in any one of clauses 1 to 9, wherein in the first link setting step, a link close to a facility of a type associated with the type of event among facilities associated with the occurrence location is set as the target link. According to the information processing method of the tenth aspect, the target link that is the starting point of the search can be appropriately set according to the type of event, thereby making it possible to determine the effect of the event on the traffic situation with a higher degree of accuracy.
[0095] (Item 11) An information processing device for determining the impact of an event on traffic conditions, which sets a link related to the location where the event occurs as a target link, executes a determination process for determining the degree of impact the event has on traffic conditions of the target link, sets a link adjacent to the target link as a new target link based on the results of the determination process, and performs the determination process on the identified target link. According to the information processing device of paragraph 11, a new target link is set based on the result of judging the degree of influence of an event on the traffic conditions of a target link, thereby expanding the scope of judgment, thereby making it possible to efficiently perform the process of judging the influence of an event on traffic conditions. [Explanation of symbols]
[0096] 1...information processing device, 10...control unit, 11...processor, 12...information acquisition unit, 13...processing unit, 21...memory, 31...communication unit, 32...input unit, 33...output unit, 50...traffic database, 51...road data, 52...facility data, 53...traffic data, 54...start position setting data, 55...prediction data.
Claims
1. 1. An information processing method for determining an impact of an event on a traffic situation, comprising: a first link setting step of setting a link related to the occurrence position of the event as a target link; a determination step of determining an impact of the event on a traffic condition of the target link; a second link setting step of setting a link adjacent to the target link as a new target link based on a result of the determination step; performing a process of the determining step on the target link identified in the second link setting step; The determining step includes: creating a first histogram showing a distribution of travel times of the target link in a first time period and a second histogram showing a distribution of travel times of the target link in a second time period that is farther away from the occurrence time of the event than the first time period; Calculating a degree of difference between the first histogram and the second histogram; The degree of difference is compared with a predetermined threshold to determine the degree of impact of the event on the target link. Information processing methods.
2. The determination step calculates the degree of difference using any one of the following formulas (1), (2), (3), (4), and (5). [0010] where D KL (X n ∥X e ): degree of difference between the first and second histograms due to KL divergence, D JS : degree of difference between the first and second histograms due to JS divergence, X n : vector representing the data of the first histogram, X e : vector representing the data of the second histogram (X n and X e are vectors of the same size), i: index of the class that classifies the travel time of the target link in specified time units. [0025] Here, D is the degree of difference between the first histogram and the second histogram, the root sign of a vector means a vector obtained by applying the root sign to each element of the vector, and T is a symbol representing the transposition of a vector. [0030] Here, p^ is the degree of difference between the first histogram and the second histogram, and P(a) is a symbol indicating the probability of a.
2. The information processing method according to claim 1.
3. when it is determined in the determining step that the degree of influence of the event on the traffic condition of the target link is low, the second link setting step is not executed.
3. The information processing method according to claim 1.
4. In the determining step, a degree of influence of the event on a traffic situation of the target link is determined for each moving direction on the target link. The information processing method according to any one of claims 1 to 3.
5. in the determining step, when the target link is a large road, a degree of influence of the event on a traffic situation of the target link is determined for each moving direction of the target link; The information processing method according to claim 4.
6. In the determination step, the threshold value is corrected based on a difference in distribution spread between the first histogram and the second histogram, and the determination is made using the corrected threshold value. The information processing method according to any one of claims 1 to 5.
7. In the determining step, When the target link is a road on which travel is possible in a first direction and a second direction opposite to the first direction, correcting the threshold value when determining the degree of influence of the event on the traffic condition of the target link in the first direction, based on the traffic condition of the target link in the second direction; The information processing method according to any one of claims 1 to 6.
8. In the determining step, correcting the threshold value for determining the degree of influence of the event on the traffic condition of the target link in the first direction based on the traffic condition of the target link in the second direction when the target link is not a large road. The information processing method according to claim 7.
9. In the first link setting step, a link close to a facility of a type associated with a type of the event among facilities associated with the occurrence location is set as the target link. The information processing method according to any one of claims 1 to 8.
10. 1. An information processing device for determining an impact of an event on a traffic situation based on road data relating to nodes and links, comprising: A link related to the occurrence position of the event is set as a target link; execute a determination process for determining the degree of influence of the event on a traffic condition of the target link; Based on a result of the determination process, a link adjacent to the target link is set as a new target link; performing the determination process on the identified target link; As the determination process, creating a first histogram showing a distribution of travel times of the target link in a first time period and a second histogram showing a distribution of travel times of the target link in a second time period that is farther away from the occurrence time of the event than the first time period; Calculating a degree of difference between the first histogram and the second histogram; A process of determining an impact of the event on the target link is performed by comparing the degree of difference with a predetermined threshold. Information processing device.
Citation Information
Patent Citations
Congestion prediction device, congestion prediction system and congestion prediction method
JP2016110360A
Presentation device, presentation method, and presentation program
JP2016173270A
Systems and methods for monitoring traffic congestion
US20200160695A1
