Traffic congestion information creation device

Abstract

To improve accuracy of a congestion location.SOLUTION: Provided is a congestion information creation device 100 comprising: a communication unit 21 which receives location information and data of time from each mobile; and a stagnation point identification unit 24 which clusters the data in which a vehicle speed is lower than a threshold, in the data for which the location information is converted into a location on a road, and identifies a congestion end or a stagnation point on the road on the basis of the grouped data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a traffic congestion information creating device that creates traffic congestion information regarding traffic congestion of mobile objects. [Background technology]

[0002] A well-known technique for understanding road congestion is to use probe data transmitted from vehicles known as probe cars. Probe data is information that can be acquired by vehicles, and includes repeatedly transmitted location information (latitude and longitude) and vehicle speed, etc.

[0003] A technology for creating congestion information using probe data is known (see, for example, Patent Document 1). Patent Document 1 discloses a server system that divides a map into mesh areas and calculates the flow rate of vehicles currently traveling in each mesh area based on probe data. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-126123 Summary of the Invention [Problem to be solved by the invention]

[0005] However, conventional technologies have a problem in that congestion information is created for each mesh, i.e., for each fixed area, resulting in low accuracy of congestion location. Here, low accuracy of congestion location means that the range of locations where congestion starts or ends is wide, making it difficult to identify where the congestion starts or ends.

[0006] In view of the above-mentioned problems, the present invention aims to provide a technique that can improve the accuracy of determining congestion locations. [Means for solving the problem]

[0007] In view of the above-mentioned problems, the present invention provides a communication unit that receives position information, time, and vehicle speed data from each moving object, converts the position information into a position on a road, generates a graph in which data points corresponding to the speed of each moving object are associated with an axis corresponding to the position on the road and an axis corresponding to the time, performs clustering on the data points in the graph where the vehicle speed is less than a threshold, and groups the data points, and selects a representative point of the data points that is upstream in the traveling direction of the graph from the grouped data points. By curve fitting the representative points for each time, End of traffic jam of A representative point of the data point downstream in the direction of travel is predicted. By curve fitting the representative points for each time, Retention point of and a congestion information creation device having a stay point identification unit that predicts congestion. [Effects of the Invention]

[0008] According to the present invention, it is possible to provide a technology that can improve the accuracy of determining congestion locations. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is an example of a system configuration diagram of a traffic congestion information creation system. [Figure 2] 10 is a flowchart illustrating an example of a procedure in which a stay point identifying unit identifies the end or head of a traffic jam. [Figure 3] FIG. 10 is a diagram illustrating a distance-time-vehicle speed graph. [Figure 4] FIG. 10 is a diagram showing data points in a distance-time-vehicle speed graph where the vehicle speed is less than a threshold value. [Figure 5] FIG. 5 shows the clustering results of the data points of FIG. 4. [Figure 6] FIG. 10 shows upper and lower representative points of grouped data points. [Figure 7] 10A and 10B are diagrams illustrating an example of output in which an approximation curve is converted into a stagnation occurrence coordinate and the stagnation occurrence coordinate is output. [Figure 8]10 is a flowchart illustrating an example of a procedure in which a stay point identifying unit identifies the end or head of a traffic jam (modification); [Figure 9] FIG. 10 is a diagram for schematically explaining data points where the slope is less than a threshold value. [Figure 10] FIG. 10 is a diagram showing a system configuration diagram of a traffic congestion information creation system (modification). DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, a traffic congestion information creation device and a traffic congestion information creation method performed by the traffic congestion information creation device will be described as an example of an embodiment of the present invention.

[0011] <System configuration example> The traffic congestion information creation device of this embodiment acquires the position and time on the road as probe data from each vehicle and extracts probe data where the vehicle speed is below a threshold.The traffic congestion information creation device then clusters the extracted probe data and identifies the end or beginning of the traffic congestion for the vehicle based on the clustering results.This improves the accuracy of identifying the end or beginning of the traffic congestion.

[0012] In this embodiment, the end of the traffic jam refers to the tail of the traffic jam in the direction of travel of the probe car, and the head of the traffic jam refers to the head of the traffic jam in the direction of travel of the probe car. The end of the traffic jam is on the upstream side, and the tail of the traffic jam is on the downstream side. Note that the head of the traffic jam can also be called a stagnation point, since it is not necessarily the case that following vehicles are congested.

[0013] FIG. 1 shows a system configuration diagram of a traffic congestion information creation system 100. As shown in FIG. 1, the traffic congestion information creation system 100 has a vehicle 10 and a traffic congestion information creation device 20. The traffic congestion information created by the traffic congestion information creation device 20 is provided to a road management business operator 40 and the vehicle 10. The vehicle 10 is equipped with a communication device that communicates with the traffic congestion information creation device 20 via a communication network such as a mobile phone network. This communication device is connected to a base station (not shown). The base station is also connected to the traffic congestion information creation device 20 mainly via a wired communication network or a gateway. The traffic congestion information creation device 20 and the road management business operator 40 are connected to be able to communicate with each other via a communication network such as the Internet or a LAN. The traffic congestion information creation device 20 and the road management business operator 40 may also be integrated.

[0014] The vehicle 10 is a so-called probe car, which collects probe data while traveling and transmits it to the traffic congestion information creation device 20 in nearly real time. "Real time" preferably means that processing is performed nearly simultaneously with input, and in a broader sense means that processing is completed within the maximum delay time. A probe car is a vehicle 10 that is considered to be a moving monitoring device while traveling (system on), and transmits location information, vehicle behavior, weather information, and the like to a server. In Figure 1, there is one vehicle 10, but many vehicles 10 repeatedly transmit probe data.

[0015] Probe data is information generated based on data acquired by sensors mounted on the vehicle 10 while the vehicle 10 is moving (including when the system is on and the vehicle is stopped). Probe data is mainly data related to the driving conditions and the vehicle state. Data related to the driving conditions includes engine speed, accelerator / brake operation status, vehicle speed, acceleration, shift position, mileage, and location information (latitude, longitude) of the vehicle 10. Data related to the vehicle state includes warning light display information, wiper operation status, and diagnostic information. Any data flowing through an in-vehicle LAN such as a Controller Area Network (CAN), a Local Interconnect Network (LIN), or a FlexRay can be considered probe data.

[0016] The vehicle 10 may be equipped with a navigation device, a display audio, or the like. A display audio is a device that has mainly AV functions and communication functions but does not have a navigation function. The display audio provides navigation functions through communication with a terminal device such as a smartphone. In this case, an application installed on the smartphone creates a navigation screen, which the display audio acquires via communication and displays on the vehicle's display.

[0017] A navigation device or the like installed in the vehicle 10 displays the congestion information received from the congestion information creation device 20, either based on a user request or automatically. The congestion information in this embodiment is the end point of the congestion as viewed from the direction of travel, as will be described later, or a congestion point where other vehicles are stuck (hereinafter referred to as the head of the congestion).

[0018] The traffic congestion information creation device 20 receives probe data from multiple vehicles 10 and creates traffic congestion information. In this embodiment, the traffic congestion information creation device 20 is described as a single server, but the functions of the traffic congestion information creation device 20 may be distributed across multiple servers. The traffic congestion information creation device 20 may also be compatible with cloud computing. Cloud computing refers to a usage mode in which resources on a network are used without being aware of specific hardware resources.

[0019] The traffic congestion information creation device 20 is an information processing device equipped with the functions of a general computer, such as a CPU, RAM, ROM, input / output units, etc. The traffic congestion information creation device 20 can provide the created traffic congestion information to the vehicle 10 either via the road management company 40 or directly. A computer or software that performs the function of providing information or processing results in response to requests from a client such as the vehicle 10 is called a server.

[0020] The road management company 40 collects information on road conditions such as congestion, fallen objects, and potholes, and distributes this information via a communication network to any vehicle, public broadcasting organization, and information processing terminal 41. In addition, the road management company 40 operates and manages toll roads, repairs roads, implements traffic safety measures, and formulates improvement plans.

[0021] The road management company 40 can provide congestion information to an information processing terminal 41 other than the vehicle 10. The information processing terminal 41 may be a mobile phone, smartphone, tablet terminal, PDA (Personal Digital / Data Assistance), personal computer, digital signage, or the like, as long as it has a display and communication functions.

[0022] <Functions of the traffic congestion information creation device> 1, the traffic congestion information creation device 20 has a communication unit 21, a vehicle speed DB 22, a congestion detection unit 23, a congestion point identification unit 24, and a congestion point DB 25. The traffic congestion information creation device 20 also has a position matching information storage unit 27, a standardization data storage unit 28, and a past probe data storage unit 26, which are used for batch processing. Each of these functions of the traffic congestion information creation device 20 is a function or means realized by a CPU that executes a program expanded from the auxiliary storage device to the RAM, controlling the general hardware (CPU, RAM, auxiliary storage device, input / output I / F, display, keyboard, touch panel, communication device, etc.) that a computer has.

[0023] First, the location matching information and standardization data that are prepared in advance by batch processing for generating congestion information will be described. As shown in FIG. 1, past probe data is stored in the past probe data storage unit 26. If the batch processing is performed once a month, one month's worth of probe data is stored. If the batch processing is performed once a week, one week's worth of probe data is stored. If the batch processing is performed once a day, one day's worth of probe data is stored. These periods are merely examples. The probe data stored in the past probe data storage unit 26 is processed by batch processing into location matching information in the location matching information storage unit 27 and standardization data in the standardization data storage unit 28.

[0024] The position matching information stored in the position matching information storage unit 27 is matching information for converting the position information (latitude, longitude) transmitted in real time by the vehicle 10 into "distance." This distance, which will be described later, is the distance along the road from any given point (for example, a highway entrance, a major node, an interchange, a toll booth, etc.). Past probe data also includes vehicle speed, position information (latitude, longitude), and time.

[0025] Briefly, the location matching information is prepared as follows. The traffic congestion information generating device 20 The times t1 and t2 when the vehicle 10 as a probe car transmits two pieces of location information (latitude and longitude) to the traffic congestion information creation device 20 Vehicle speeds v1 and v2 at two locations (latitude and longitude) Using "(t2-t1) × average of v1, v2" The distance along the road until passing through two pieces of position information (latitude, longitude) is calculated. For example, if consecutive probe data are used as these two pieces of position information (latitude, longitude), the distance can be determined for each piece of position information (latitude, longitude) included in the past probe data. Since the past probe data storage unit 26 stores probe data from many different vehicles 10, the traffic congestion information creation device 20 may average the distances calculated using the probe data from each vehicle 10.

[0026] The position matching information stores the distance from a certain point calculated in this manner in association with the position information (latitude, longitude). The distance from a certain point and the position information (latitude, longitude) may be associated in a table format, or may be associated with each other using a formula that relates them. Furthermore, since this position matching information is used to calculate the distance along a road, it is associated with the links that make up the road in the road information (i.e., road names, etc.).

[0027] The standardization data stored in the standardization data storage unit 28 is used to standardize the vehicle speed transmitted by the vehicle 10 in real time. Standardization means converting a group of data expressed in a normal distribution into a standard normal distribution with a mean of 0 and a variance of 1. For example, at junctions and intersections, a steady decrease in vehicle speed occurs, but this cannot be said to be a traffic jam. In order to express that "vehicle speed is slower than usual" excluding such locations, the traffic congestion information creation device 20 standardizes the vehicle speed. Since the vehicle speed of past probe data follows a normal distribution, the mean μ and variance σ can be known. The traffic congestion information creation device 20 standardizes the vehicle speed transmitted by the vehicle 10 in real time using the following formula. Normalized vehicle speed = (original vehicle speed - μ) / σ Therefore, the standardization data is the average μ and variance σ calculated for the vehicle speed transmitted together with the position information (latitude, longitude) as past probe data. In order for the traffic congestion information creation device 20 to calculate the average μ and variance σ, multiple vehicle speeds are required. For this reason, a certain range of position information (latitude, longitude) is used before and after the location (for example, at a certain distance on the road) where the average μ and variance σ are calculated. Also, in order to distinguish whether the certain range of position information (latitude, longitude) belongs to the inbound or outbound lane, vehicle speeds in the same direction of movement based on the position information (latitude, longitude) are used to calculate the standardization data.

[0028] The mean μ and variance σ of the vehicle speed vary greatly depending on the day of the week and the time of day, so it is preferable to prepare standardization data for each day of the week and each time of day.

[0029] Next, with reference to Fig. 1, the function of the traffic congestion information creation device 20 that processes probe data transmitted in real time by the vehicles 10 will be described. First, the communication unit 21 receives the probe data transmitted by each vehicle 10 and stores it in the vehicle speed DB 22. The vehicle speed DB 22 stores the probe data transmitted by the vehicle 10. In this embodiment, of the probe data, mainly vehicle speed and position information (latitude, longitude) are used.

[0030] The congestion detection unit 23 detects the possibility of a traffic jam based on the vehicle speed. For example, the congestion detection unit 23 detects the possibility of a traffic jam when a certain number or more of probe data in which the vehicle speed is less than a threshold is detected within a certain range in the same or a nearby location.

[0031] The stay point identifying unit 24 performs clustering on vehicle speeds below a threshold, and identifies the end or beginning of the congestion for the vehicle 10 based on the clustering results. Details will be explained using a flowchart or the like.

[0032] The stay point DB 25 stores the end or head of a traffic jam identified by the stay point identifying unit 24. The end or head of a traffic jam is represented by coordinates (latitude, longitude).

[0033] <Determining the end and beginning of a traffic jam> A method for identifying the end or head of a traffic jam will be described using Figs. 2 to 5. First, Fig. 2 is a flowchart showing the procedure for the stay point identification unit 24 to identify the end or head of a traffic jam. Note that the processing in Fig. 2 only needs to be performed when there is a possibility that a traffic jam has occurred, and may be executed when it is detected that the vehicle speed is below a threshold. In Fig. 2, the determination of whether the vehicle speed is below a threshold is made within the flow. Each step in Fig. 2 will be described below.

[0034] (S1) The stay point identifying unit 24 standardizes the vehicle speed transmitted from the vehicle 10 using the standardization data (μ, σ).

[0035] (S2) Next, the stay point identification unit 24 uses the position matching information to convert the position information (latitude, longitude) transmitted from the vehicle 10 into a distance along the road from a certain point (base point). Because the position matching information converts the position information (latitude, longitude) into a distance from a certain point, the stay point identification unit 24 can simply find the distance associated with the position information (latitude, longitude) from a table or a formula. Since the distance from a certain point is known, the position of the vehicle 10 on the road (specified by the distance from the certain point) can also be identified.

[0036] (S3) Next, the stay point identifying unit 24 creates a distance-time-vehicle speed graph by associating the distance in step S2 with the time and vehicle speed of the probe data.

[0037] FIG. 3 is a diagram illustrating a distance-time-vehicle speed graph 101. As shown in an image 102 of the distance-time-vehicle speed graph 101, the distance-time-vehicle speed graph 101 has time on the horizontal axis and distance on the vertical axis. The time on the horizontal axis is the time transmitted from the vehicle 10 together with the location information (latitude, longitude). The distance is converted in step S2. Vehicle speed may be represented as an axis perpendicular to the drawing, or may be represented by different colors. FIG. 3 is monochrome for ease of drawing, but data points are represented in colors according to vehicle speed.

[0038] Although the axis for distance points downward, the values ​​on the vertical axis of the distance-time-vehicle speed graph 101 increase upward. This is because on roads with both inbound and outbound lanes, the distance is measured in the opposite direction to the traveling direction of the vehicle 10. The distance follows the direction of the axis, with the upper side of the vertical axis representing the upstream side of the road.

[0039] In this way, the distance-time-vehicle speed graph 101 shows one data point corresponding to distance, time, and vehicle speed for each piece of probe data. As shown in the simplified diagram 103, each vehicle moves in a downward and rightward direction. There is a proportional relationship between time and distance, but as the vehicle speed increases, the relationship between time and distance approaches a vertical line. The stay point identification unit 24 creates a distance-time-vehicle speed graph for each road on which the vehicle 10 travels. When creating a nationwide graph, the stay point identification unit 24 creates a distance-time-vehicle speed graph for each route.

[0040] (S4) Next, the stay point identification unit 24 extracts data points where the vehicle speed is less than a threshold value from the distance-time-vehicle speed graph 101. This threshold value is a value for extracting data points corresponding to congestion, and may be, for example, a few km / h to 20 km / h. This threshold value is also standardized.

[0041] 4 shows data points in the distance-time-vehicle speed graph 101 where the vehicle speed is below the threshold. In this way, the distance from a certain point to where a vehicle with a speed below the threshold exists is associated with time. In other words, it is possible to see how the state of where vehicles are staying on the road changes over time.

[0042] (S5) Next, the stay point identification unit 24 performs clustering on the data points of step S4. Clustering is a data processing method for grouping data based on the similarity between the data. In clustering of data points arranged in a two-dimensional direction as shown in FIG. 4, multiple data points that are close to each other (island-like) are grouped into the same group. Therefore, data points of stays that occur in close locations on the road and change over time are grouped together.

[0043] Clustering methods include DBSCAN, k-means, EM algorithm, group average method, Ward's method, etc., and any clustering method is applicable in this embodiment. Clustering can eliminate extremely large or small data points, thereby removing noise unrelated to retention.

[0044] Figure 5 shows the clustering results of the data points in Figure 4. In Figure 5, the data are grouped into two groups, 1 and 2, but the number of groups may be one or three or more. For example, Group 1 indicates that vehicles are staying at a distance of approximately 1,000 to 5,000 m around 16:14. Group 2 indicates that vehicles are staying at a distance of approximately 1,000 to 7,000 m from 16:15 to 16:18.

[0045] In the data points of Figs. 3 to 5, the upper side of the vertical axis is the upstream side, so the upper data points among the grouped data points are the end of the congestion at each time, and the lower data points are the beginning of the congestion.

[0046] (S6) Next, the stagnation point identification unit 24 determines upper and lower representative points of the grouped data points. FIG. 6 shows upper and lower representative points 111, 112 of the grouped data points. Note that the data in FIG. 6 differs from the data in FIGS. 4 and 5 for ease of explanation. The upper representative point 111 indicates the end of congestion at each time, and the lower representative point 112 indicates the beginning of congestion. The stagnation point identification unit 24 may determine the representative points 111, 112 using an alpha shape algorithm implemented in statistical software or the like, or may determine the data point with the greatest distance (distance on the vertical axis) as representative point 111 and the data point with the smallest distance as representative point 112 for each time of the grouped data points.

[0047] (S7) Next, the stagnation point identification unit 24 determines the end of the congestion by curve fitting a representative point 111 upstream in the direction of travel, and determines the head of the congestion by curve fitting a representative point 112 farther away in the direction of travel.

[0048] FIG. 6 shows an approximate curve 110 obtained by curve fitting a representative point 111. The approximate curve 110 represents the end of the traffic jam at each time. By performing curve fitting in this way, it is possible to reduce the variation in the end of the traffic jam even if the number of probe data (vehicles) is small. Note that if there is a sufficient number of probe data (vehicles) so that the variation in the end of the traffic jam does not occur, curve fitting is not necessarily required.

[0049] The model formula used for curve fitting of the representative point 111 may be, for example, a quadratic formula, a cubic formula, or the like. A model formula suited to the shape of the representative point 111 is prepared in advance. In addition, several model formulas suited to the shape of the representative point 111 may be prepared in advance, and a model formula may be determined in order of priority so that the correlation coefficient satisfies a standard. For example, in the case of a quadratic formula, the model formula is as follows: y=ax 2 +bx+c ……(1) Here, y is the distance and x is the time.

[0050] The range of the x-axis (time) direction for curve fitting is not particularly restricted, and may include all representative points grouped by clustering. However, the stagnation point identification unit 24 may use only data points up to a certain time before the current time for curve fitting. This may improve the accuracy of identifying the end and beginning of the congestion at the current time.

[0051] 6, curve fitting is not performed on the representative point 112 at the beginning of the congestion, but the stay point identification unit 24 can perform curve fitting on the representative point 112 at the end of the congestion in the same way as on the representative point 111 at the end of the congestion. The model formula used for curve fitting of the representative point 112 at the beginning of the congestion may be different from the model formula for the representative point 111, as long as it is suitable for the shape of the representative point 112.

[0052] As a result, it is possible to accurately determine the start and end points of multiple vehicles currently stuck on the road. In addition, since the approximation curve 110 can be externally fitted, it is also possible to predict the end of future congestion.

[0053] <Example of traffic congestion information output> Next, we will explain how to output congestion occurrence coordinates (latitude, longitude) based on the approximate curve obtained by the process in Figure 2. Figure 7 is a diagram that explains an output example in which the approximate curve is converted into congestion occurrence coordinates and the congestion occurrence coordinates are output. Figure 7(a) is the same as Figure 6, and Figure 7(b) shows an output example of congestion occurrence coordinates.

[0054] The congestion point identification unit 24 inputs the time for which congestion information is desired into x in equation (1). Since y in equation (1) is distance, the distance from a certain point at this time is obtained. If the current time is 6:00 p.m., the probe data up to 6:00 p.m. is used to calculate the end and beginning of the congestion. The output y is the latest (current time) end or beginning of the congestion on the approximation curve.

[0055] The distance from a certain point is not a coordinate as it is, and therefore the congestion information creation device 20 cannot display it on an electronic map. For this reason, the stay point identification unit 24 converts the distance from a certain point into coordinates (latitude, longitude). As a conversion method, the stay point identification unit 24 traces the links in the road information DB from the certain point along the road to the distance. That is, the stay point identification unit 24 accumulates the link lengths until the distance from the certain point on the road is calculated using equation (1). The destination of the link tracing to the distance is the stay occurrence coordinates. Note that this road along the road is the road on which the data points were collected in FIG. 2. Alternatively, the stay point identification unit 24 may use position matching information to convert the distance from a certain point into coordinates (latitude, longitude). In this case, contrary to step S2, position information (latitude, longitude) associated with the distance is acquired.

[0056] The right side of Figure 7(b) is an enlarged electronic map of the left side. On the electronic map, congestion occurrence coordinates 120, which are represented by dots, are shown. The congestion information creation device 20 may map match the congestion occurrence coordinates 120. The viewer can check on the map where the current congestion is located and how far it extends.

[0057] <Modification> Next, a modified example of the method for identifying the end or head of a traffic jam described in Fig. 2 will be described with reference to Fig. 8 and Fig. 9. Fig. 8 is a flowchart (modified example) showing the procedure by which the stay point identification unit 24 identifies the end or head of a traffic jam. Note that the explanation of Fig. 8 will mainly focus on the differences from Fig. 2.

[0058] First, in Figure 8, step S1 in Figure 2 is omitted. This is because vehicle speed is not used as a data point (data points are not extracted based on vehicle speed itself). The processing in the next step S21 can be the same as step S2 in Figure 2.

[0059] Next, in step S22, the stay point identifying unit 24 creates a distance-time graph, which does not include vehicle speed.

[0060] Next, in step S23, the stay point identifying unit 24 extracts data points where the slope is less than a threshold (see FIG. 9). Because the slope of the distance-time graph correlates with the vehicle speed, step S23 extracts only data points where the vehicle speed is less than the threshold.

[0061] The processes in the following steps S24 to S26 may be the same as steps S5 to S7 in FIG.

[0062] 9 is a diagram illustrating data points where the slope is less than the threshold. For example, suppose vehicles A to E transmit probe data to the traffic congestion information creation device 20. As shown in FIG. 9(a), the slope of the graph of time versus distance in the probe data of vehicles A and B is large. On the other hand, the slope of the graph of time versus distance in the probe data of vehicles C to E includes a portion 130 where the slope is less than the threshold.

[0063] Figure 9(b) shows data points that were extracted because the slope was less than the threshold. This allows us to extract data points that may have been stuck without comparing the vehicle speed itself with the threshold.

[0064] <Devices other than vehicles that transmit location information and time> 10, this embodiment can be applied to location information and time transmitted by a moving body that can transmit location information (latitude, longitude) and time, not limited to the vehicle 10. The moving speed may or may not be transmitted.

[0065] Fig. 10 shows a system configuration diagram of a traffic congestion information creation system 100 (modified example). In the explanation of Fig. 10, components with the same reference numerals as those in Fig. 1 perform similar functions, so in some cases only the main components of this embodiment will be mainly explained.

[0066] 10, a smartphone 140 serving as a mobile object transmits location information (latitude, longitude) to the traffic congestion information creation device 20. The smartphone 140 is not mounted on the vehicle 10 but is carried by the user when traveling. For example, traffic congestion information for walking or cycling is useful when traveling to a tourist spot, an event venue, etc.

[0067] <Major Effects> The effects of the congestion information generation system 100 of this embodiment described above will be described.

[0068] The traffic congestion information creation device 20 of this embodiment uses the position information (latitude, longitude) of the vehicle 10 to determine traffic congestion, so the resolution of traffic congestion locations is high (high location accuracy). In contrast, the conventional technology converts probe data into mesh areas and aggregates them, so the resolution of traffic congestion locations is low. In the conventional technology, if you simply output "locations where 100 vehicles are experiencing a speed drop," you cannot identify where the traffic congestion starts and ends. Furthermore, you cannot determine whether the traffic congestion is the same or different.

[0069] The congestion information creation device 20 of this embodiment links probe data to roads and sets a starting point along the road (along the link), making it possible to create congestion information by distinguishing between inbound and outbound lanes. In contrast, conventional technology aggregates data without distinguishing between inbound and outbound lanes. The conventional technology calculates the congestion direction by averaging the direction of vehicle travel within a mesh, which is similar to the concept of inbound and outbound lanes. However, there is a possibility that the congestion direction may become unclear if there is a cancellation or if there are many vehicles with zero speed.

[0070] The traffic congestion information creation device 20 of this embodiment can create traffic congestion information in real time by sequentially clustering probe data from the vehicle 10 and performing curve fitting. Conventional technologies perform aggregation processing every 15 minutes, making real-time processing difficult.

[0071] The traffic congestion information creation device 20 of this embodiment uses probe data from the past few hours, making it possible to estimate the end of a traffic congestion even when the number of vehicles in a certain time span is small.

[0072] The traffic congestion information creation device 20 of this embodiment utilizes the characteristics of clustering (data that does not belong to any cluster is treated as noise) to improve robustness against noise. In conventional technology, when there are few vehicles that can upload probe data, noise (such as a coincidental speed drop) becomes a cause of error.

[0073] The traffic congestion information creation device 20 of this embodiment standardizes vehicle speeds in advance using past probe data, making it possible to express whether a vehicle is "slower than usual." Conventional technologies set vehicle speed thresholds for each road type, which can lead to overdetection of traffic congestion in places where speed drops occur frequently, such as before interchanges or junctions, even on the same expressway.

[0074] Furthermore, the traffic congestion information creation device 20 of this embodiment clusters data points and performs curve fitting, making it impossible to trace information on individual vehicles from the end of a traffic jam, thereby protecting privacy.

[0075] Furthermore, the traffic congestion information creation device 20 of this embodiment can estimate the end of traffic congestion by clustering and remove noise by using points where speeds are low. A possible method would be to extract points where the absolute value of speed change (acceleration) is greater than a threshold as speed drop points, but in that case, the data points would be sparse, making clustering impossible and making it difficult to determine whether a point is congested. There is also concern that this method is susceptible to noise (a random drop in speed). [Explanation of symbols]

[0076] 10 vehicles 20 Traffic congestion information creation device 100 Traffic Information Creation System

Claims

[Claim 1] a communication unit that receives location information, time, and vehicle speed data from each moving object; converting the location information into a location on a road; generating a graph in which data points corresponding to the speed of each moving object are associated with an axis corresponding to the position on the road and an axis corresponding to the time; clustering the data points in the graph where the vehicle speed is less than a threshold value; a stagnation point identification unit that predicts the end of congestion at each time by curve fitting a representative point of the data points that is on the upstream side of the traveling direction in the graph among the grouped data points and that is a representative point for each time, and predicts a stagnation point at each time by curve fitting a representative point of the data points that is on the downstream side of the traveling direction and that is a representative point for each time; A traffic congestion information creation device having the above.

Images