Traffic flow prediction device, traffic flow prediction method, and program

Abstract

This enables more accurate prediction of traffic flow. [Solution] A traffic flow prediction device is provided, comprising: a traffic volume prediction unit that predicts a second inflow of traffic into a range corresponding to congestion data at a second time after the first time based on a first inflow of traffic that entered the range at a first time, and predicts a second outflow of traffic that leaves the range at a second time based on a first outflow of traffic that left the range at a first time, and calculates a cumulative inflow of traffic by accumulating the first inflow of traffic and the second inflow of traffic, and a cumulative outflow of traffic by accumulating the first outflow of traffic and the second outflow of traffic; a congestion traffic volume calculation unit that calculates the difference between the cumulative inflow of traffic and the cumulative outflow of traffic as the congestion traffic volume; a congestion area prediction unit that predicts a congestion area at a second time based on the congestion traffic volume; and a traffic density prediction unit that predicts the traffic density of a target location at a second time based on the congestion area.

Description

[Technical Field]

[0001] The present invention relates to a traffic flow prediction device, a traffic flow prediction method, and a program. [Background technology]

[0002] In recent years, various methods for predicting traffic flow have been developed. For example, one method involves using a first model that utilizes deep learning to predict the amount of traffic flowing into a given continuous section from the upstream side as the upstream inflow traffic volume, and then using a second model that utilizes deep learning to predict the time required for vehicles to travel through that continuous section based on the upstream inflow traffic volume.

[0003] Furthermore, Patent Document 1 discloses a method for learning the relationship between the traffic density of a predetermined range based on a target location at a past time and the traffic density of the target location at the time immediately following that past time. In this method, the traffic density of the target location at times immediately following the target time is predicted based on the learned relationship and the traffic density of that range at the target time. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-86647 [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] However, methods for predicting the time required for a vehicle to travel a predetermined continuous section of road use deep learning models. Therefore, a large amount of data is needed for training, not only on congestion caused by the concentration of traffic volume due to increased traffic demand (hereinafter also referred to as "natural congestion"), but also on special types of congestion caused by sudden events. Consequently, methods using deep learning models are prone to insufficient training, which can easily lead to a decrease in prediction accuracy.

[0006] Furthermore, methods for predicting the time required for a vehicle to travel a predetermined continuous section require data on upstream traffic volume and travel time for each combination of the upstream and downstream ends of the continuous section. Consequently, as the number of continuous sections for which travel time needs to be predicted increases, the amount of data required for training becomes enormous. For example, as road networks become more complex, the number of continuous sections for which travel time needs to be predicted increases, and the amount of data required for training tends to become enormous.

[0007] Furthermore, in methods that use learned relationships to predict traffic density at a target location at subsequent time points, the predicted traffic density is used sequentially to predict traffic density that has not yet been predicted. As a result, the prediction error tends to increase as traffic density predictions progress. Consequently, the predicted traffic volume corresponding to a certain time and location may not match the traffic volume derived from the law of conservation of traffic volume, and the predicted traffic volume may become an abnormal value.

[0008] Therefore, it is desirable to have technologies that enable more accurate prediction of traffic flow. [Means for solving the problem]

[0009] To solve the above problems, according to one aspect of the present invention, a range corresponding to congestion data is identified, which includes the congestion time, which is the time when congestion occurs, and the congestion location, which is the location where congestion occurs at the congestion time. Based on the first inflow traffic volume that flows into the range at the first time, a second inflow traffic volume that flows into the range at a second time after the first time is predicted. Based on the first outflow traffic volume that flows out of the range at the first time, a second outflow traffic volume that flows out of the range at the second time is predicted. A traffic flow prediction device is provided, comprising: a traffic volume prediction unit that calculates a cumulative inflow traffic volume obtained by accumulating the first inflow traffic volume and the second inflow traffic volume, and a cumulative outflow traffic volume obtained by accumulating the first outflow traffic volume and the second outflow traffic volume; a congested traffic volume calculation unit that calculates the difference between the cumulative inflow traffic volume and the cumulative outflow traffic volume as the congested traffic volume; a congested basin prediction unit that predicts the congested basin at the second time based on the congested traffic volume; and a traffic density prediction unit that predicts the traffic density at the target location at the second time based on the congested basin.

[0010] The congestion area prediction unit may predict the length of the congestion according to the amount of traffic that is currently held, and predict the congestion area at the second time based on the length of the congestion.

[0011] The congestion area prediction unit may predict the congestion area at the second time based on the downstream position and the congestion length in the congestion data.

[0012] The congestion area prediction unit may predict the congestion area at the second time as extending from the lowest downstream position in the congestion data to a position upstream by the length of the congestion relative to the lowest downstream position.

[0013] The traffic density prediction unit may obtain a congestion determination result by determining whether the target location at the second time point belongs to the congested watershed, and based on the congestion determination result, predict the traffic density of the target location at the second time point.

[0014] The traffic flow prediction device includes a learning unit that stores a combination of the traffic density at a predetermined location at a third time and a previous time traffic density pattern, which is the traffic density within a range corresponding to the predetermined location at a fourth time that is prior to the third time, as a traffic density pattern. The traffic density prediction unit may predict the traffic density at the target location at the second time based on the traffic density pattern and the congestion determination result.

[0015] The traffic density prediction unit may predict the traffic density of the predetermined location at the third time as the traffic density of the target location at the second time if the target location at the second time belongs to the congested area, the predetermined location is a congested location at the third time, and the traffic density pattern of the previous time, which is the traffic density of the range corresponding to the target location at the first time, is similar to the traffic density pattern of the previous time, which is the traffic density of the range corresponding to the predetermined location at the fourth time.

[0016] The traffic density prediction unit may predict the traffic density of the predetermined location at the third time as the traffic density of the target location at the second time if the target location at the second time does not belong to the congested area, the predetermined location is not a congested location at the third time, and the traffic density pattern of the previous time, which is the traffic density of the range corresponding to the target location at the first time, is similar to the traffic density pattern of the previous time, which is the traffic density of the range corresponding to the predetermined location at the fourth time.

[0017] The traffic volume prediction unit may specify the range as the vehicle detection position located downstream of the furthest downstream position in the congestion data, and the vehicle detection position located upstream of the furthest upstream position in the congestion data.

[0018] The traffic volume prediction unit may determine, based on the vehicle speed statistics, whether the statistics are below a threshold, and if it determines that the statistics are below the threshold, it may extract the position corresponding to the vehicle speed as the congestion location and the time corresponding to the vehicle speed as the congestion time.

[0019] Furthermore, in order to solve the above problems, according to another aspect of the present invention, a range corresponding to congestion data is identified, which includes the congestion time, which is the time when congestion occurs, and the congestion location, which is the location where congestion occurs at the congestion time. Based on the first inflow traffic volume that flows into the range at the first time, a second inflow traffic volume that flows into the range at a second time after the first time is predicted, and based on the first outflow traffic volume that flows out of the range at the first time, a second outflow traffic volume that flows out of the range at the second time is predicted, and the first inflow A computer-based traffic flow prediction method is provided, comprising: a traffic volume prediction unit that calculates a cumulative inflow traffic volume obtained by accumulating the traffic volume and the second inflow traffic volume, and a cumulative outflow traffic volume obtained by accumulating the first outflow traffic volume and the second outflow traffic volume; a congested traffic volume calculation unit that calculates the difference between the cumulative inflow traffic volume and the cumulative outflow traffic volume as the congested traffic volume; a congested basin prediction unit that predicts the congested basin at the second time based on the congested traffic volume; and a traffic density prediction unit that predicts the traffic density of a target location at the second time based on the congested basin.

[0020] Furthermore, in order to solve the above problems, according to another aspect of the present invention, the computer identifies a range corresponding to congestion data that includes a congestion time, which is the time when congestion occurs, and a congestion location, which is the location where congestion occurs at the congestion time; predicts a second inflow traffic volume that will flow into the range at a second time after the first time, based on a first inflow traffic volume that flowed into the range at a first time; and predicts a second outflow traffic volume that will flow out of the range at a second time, based on a first outflow traffic volume that flowed out of the range at the first time. A program is provided that functions as: a traffic volume prediction unit that predicts and calculates a cumulative inflow traffic volume by accumulating the first inflow traffic volume and the second inflow traffic volume, and a cumulative outflow traffic volume by accumulating the first outflow traffic volume and the second outflow traffic volume; a congested traffic volume calculation unit that calculates the difference between the cumulative inflow traffic volume and the cumulative outflow traffic volume as the congested traffic volume; a congested basin prediction unit that predicts the congested basin at the second time based on the congested traffic volume; and a traffic density prediction unit that predicts the traffic density at the target location at the second time based on the congested basin. [Effects of the Invention]

[0021] As described above, the present invention provides a technology that enables more accurate prediction of traffic flow. [Brief explanation of the drawing]

[0022] [Figure 1] This figure shows an example of the functional configuration of a traffic flow prediction device 1 according to an embodiment of the present invention. [Figure 2] This figure shows a detailed configuration example of the learning unit 141. [Figure 3] This figure shows a detailed configuration example of the inference unit 142. [Figure 4] This figure shows an example of master data stored by the master data storage unit 126. [Figure 5] This figure shows an example of learning parameters stored by the learning parameter storage unit 127. [Figure 6]This figure shows an example of free-flow data stored by the free-flow data storage unit 123. [Figure 7] This figure shows an example of probe data stored by the probe data storage unit 122. [Figure 8] This figure shows the relationship between traffic density (K) and speed (V), and the KV relationship. [Figure 9] This figure shows an example of the configuration of KV parameters stored in the learning parameter storage unit 127. [Figure 10] This flowchart shows an example of the KV parameter creation process performed by the KV parameter creation unit 1411. [Figure 11] This diagram illustrates an example of calculating the passage time at a measurement point for a vehicle that has reached an estimated point. [Figure 12] This diagram illustrates an example of calculating the traffic volume at an estimated location from the traffic volume at a measurement point. [Figure 13] This flowchart shows an example of the traffic density calculation process performed by the traffic density calculation unit 1412. [Figure 14] This diagram schematically shows the mesh data used for training. [Figure 15] This is a diagram illustrating the overview of traffic density pattern learning. [Figure 16] This figure shows an example of a traffic density pattern. [Figure 17] This flowchart shows an example of how the traffic density pattern learning operation is performed by the traffic density pattern learning unit 1413. [Figure 18] This diagram schematically shows the predicted traffic density and speed. [Figure 19] This diagram shows an example of speed corresponding to a combination of time and cell. [Figure 20] This is a diagram showing traffic congestion data D1. [Figure 21] This diagram illustrates the traffic volume prediction performed by the traffic volume prediction unit 1421. [Figure 22]This flowchart shows an example of the operation of the traffic volume prediction unit 1421 and the traffic congestion volume calculation unit 1422. [Figure 23] This is a diagram to explain congested watersheds. [Figure 24] This flowchart shows an example of the operation of the congestion basin prediction unit 1423. [Figure 25] This flowchart shows an example of the operation of the traffic density prediction unit 1424. [Figure 26] This figure shows the prediction results from traffic flow prediction related to the comparative example. [Figure 27] This figure shows the prediction results obtained by traffic flow prediction according to an embodiment of the present invention. [Figure 28] This figure shows the hardware configuration of an information processing device 900 as an example of a traffic flow prediction device 1 according to an embodiment of the present invention. [Modes for carrying out the invention]

[0023] Preferred embodiments of the present invention will be described in detail below with reference to the attached drawings. In this specification and drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant descriptions will be omitted.

[0024] (1. Details of the Embodiment) The following describes in detail the embodiments of the present invention.

[0025] (1-1. Configuration of the traffic flow prediction device) First, an example of the configuration of the traffic flow prediction device 1 according to an embodiment of the present invention will be described. Figure 1 is a diagram showing an example of the functional configuration of the traffic flow prediction device 1 according to an embodiment of the present invention. The traffic flow prediction device 1 is a device that predicts traffic flow. Traffic flow is data related to traffic and may be a concept that includes, for example, traffic volume, traffic density, and vehicle speed. The traffic flow prediction device 1 may be an example of an information processing device.

[0026] Referring to Figure 1, vehicles M1 to M3 are shown as examples of vehicles traveling on a road. Furthermore, referring to Figure 1, vehicles M1 and M2 are traveling in the far lane (vehicle M1 is following vehicle M2), and vehicle M3 is traveling in the near lane, in the opposite direction to vehicles M1 and M2 in the far lane. Thus, the embodiment of the present invention mainly assumes a road composed of multiple lanes, but the road may also consist of a single lane. Each of vehicles M1 to M3 is equipped with an on-board device. Note that the term "lane" can also be used as "lane."

[0027] The traffic flow prediction device 1 includes a driving history data storage unit 121, a probe data storage unit 122, a free flow data storage unit 123, a traffic volume data storage unit 124, a traffic density data storage unit 125, a master data storage unit 126, a learning parameter storage unit 127, a prediction result storage unit 128, and a mesh data storage unit 129.

[0028] Furthermore, the traffic flow prediction device 1 comprises a statistical processing unit 131, a traffic volume calculation unit 133, a processing unit 140, and a mesh data generation unit 135. The processing unit 140 comprises a learning unit 141 and an inference unit 142.

[0029] Figure 2 shows a detailed configuration example of the learning unit 141. As shown in Figure 2, the learning unit 141 comprises a KV parameter creation unit 1411, a traffic density calculation unit 1412, and a traffic density pattern learning unit 1413. These blocks of the learning unit 141 primarily operate during the learning phase. The detailed functions of these blocks of the learning unit 141 will be described later.

[0030] Figure 3 shows a detailed configuration example of the inference unit 142. As shown in Figure 3, the inference unit 142 comprises a traffic volume prediction unit 1421, a traffic congestion volume calculation unit 1422, a congestion basin prediction unit 1423, a traffic density prediction unit 1424, and a supply unit 1425. These blocks of the inference unit 142 primarily operate during the inference phase. The detailed functions of these blocks of the inference unit 142 will be described later.

[0031] A probe antenna 112 is connected to the driving history data storage unit 121, and a free-flow antenna 114 is connected to the free-flow data storage unit 123.

[0032] The statistical processing unit 131, the traffic volume calculation unit 133, the mesh data generation unit 135, and the processing unit 140 include a computing device such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), and their functions can be realized by a program stored in ROM (Read Only Memory) being loaded into RAM by the computing device and executed. In this case, a computer-readable recording medium on which the program is stored may also be provided.

[0033] Alternatively, the statistical processing unit 131, the traffic volume calculation unit 133, the mesh data generation unit 135, and the processing unit 140 may be composed of dedicated hardware or a combination of multiple hardware components. The data necessary for calculations by the arithmetic unit is appropriately stored in a storage unit (not shown).

[0034] The driving history data storage unit 121, the probe data storage unit 122, the free flow data storage unit 123, the traffic volume data storage unit 124, the traffic density data storage unit 125, the master data storage unit 126, the learning parameter storage unit 127, the prediction result storage unit 128, and the mesh data storage unit 129 are implemented by storage units (not shown). Such storage units may consist of memory such as RAM (Random Access Memory), a hard disk drive, or flash memory.

[0035] (Master data storage unit 126) The master data storage unit 126 stores various types of master data in advance. Master data is data that is input to the traffic flow prediction device 1 by the user and managed by the traffic flow prediction device 1. Here, an example of master data will be explained with reference to Figure 4.

[0036] Figure 4 shows an example of master data stored by the master data storage unit 126. As shown in Figure 4, the master data storage unit 126 stores the driving history time width 611, driving history section width 612, simulation execution time interval 621, simulation execution cell width 622, congestion flow determination speed 641, congestion wave propagation speed 645, and congestion flow determination speed 649 as examples of master data. In the following explanation, we will mainly assume the case where the congestion flow determination speed 641 and the congestion flow determination speed 649 are provided separately. However, the congestion flow determination speed 641 and the congestion flow determination speed 649 may be the same value or may be common to each other. Details of this master data will be explained later.

[0037] (Learning parameter storage unit 127) The learning parameter storage unit 127 can store various learning parameters. Learning parameters are parameters obtained through learning and are stored in the learning parameter storage unit 127 during the learning phase. Furthermore, learning parameters are retrieved from the learning parameter storage unit 127 during the inference phase. Now, with reference to Figure 5, an example of learning parameters stored in the learning parameter storage unit 127 will be explained.

[0038] Figure 5 shows an example of learning parameters stored by the learning parameter storage unit 127. As shown in Figure 5, the learning parameter storage unit 127 can store KV parameters 710 and traffic density patterns 720 as examples of learning parameters. Details of these learning parameters will be described later.

[0039] (Free-flow antenna 114) The free-flow antenna 114 functions as an example of a vehicle detection unit that detects vehicles traveling at various locations on the road. In other words, the embodiments of the present invention primarily assume that the vehicle detection unit includes the free-flow antenna 114. This allows the ETC free-flow antenna of an already constructed ETC (Electronic Toll Collection) system to be used as the vehicle detection unit, eliminating the need to provide a new vehicle detection unit. However, other vehicle detection units (e.g., vehicle detectors, infrared sensors, or ultrasonic sensors) may be used instead of the free-flow antenna 114.

[0040] More specifically, the free-flow antenna 114 detects vehicles in real time by receiving vehicle identification information (vehicle ID) from the vehicle through communication with an on-board unit installed in the vehicle. In the embodiments of the present invention, the case where an ETC on-board unit is used as an example of an on-board unit is mainly assumed. It is assumed that the free-flow antenna 114 is compatible with multiple versions of ETC on-board units, while the probe antenna 112, which will be described later, is assumed to be compatible with only a specific version of ETC on-board unit. In other words, the free-flow antenna 114 can detect more vehicles than the probe antenna 112.

[0041] The "locations" on the road where a vehicle is detected by the free-flow antenna 114 are not particularly limited, as long as they are locations where a vehicle can be detected by the vehicle detection unit. Hereinafter, the locations on the road where a vehicle can be detected by the free-flow antenna 114 may simply be referred to as "vehicle detection locations."

[0042] The free-flow antenna 114 detects a vehicle in real time by receiving a vehicle ID through communication with an in-vehicle device mounted on the vehicle, and outputs the vehicle detection result (hereinafter also referred to as "free-flow data") to the free-flow data storage unit 123 in real time. The vehicle detection result output from the free-flow antenna 114 to the free-flow data storage unit 123 in real time can then be used in real time by the traffic volume calculation unit 133.

[0043] Here, "real-time" can refer to any short period of time between the time a vehicle reaches the vehicle detection location and before the traffic conditions on the road change. If a vehicle is detected at this time, and the vehicle detection result is used by the traffic volume calculation unit 133 at this time, some action can be taken according to the traffic conditions at the time the vehicle was detected, before the traffic conditions on the road change. On the other hand, the driving history data acquired by the probe antenna 112, which will be described later, is historical data related to the vehicle's driving. Therefore, the timing at which a vehicle is detected by the probe antenna 112 may be later than the timing at which a vehicle is detected by the free-flow antenna 114.

[0044] (Free-flow data storage unit 123) Figure 6 shows an example of free-flow data stored by the free-flow data storage unit 123. The free-flow data corresponds to the detection result of a vehicle traveling at the vehicle detection location. As shown in Figure 6, the free-flow data is associated with a "vehicle ID" and a "passing time". The "vehicle ID" is the vehicle identification information received from the vehicle by the free-flow antenna 114 through communication between the free-flow antenna 114 and an in-vehicle device mounted on the vehicle. The "passing time" is the time when the vehicle ID was received from the vehicle by the free-flow antenna 114, and may correspond to the time when the vehicle passed the vehicle detection location.

[0045] (Traffic volume calculation unit 133) The traffic volume calculation unit 133 acquires free flow data from the free flow data storage unit 123. Then, based on the free flow data (vehicle ID and passing time), the traffic volume calculation unit 133 calculates the number of vehicles that passed the vehicle detection location per unit time (in other words, the number of vehicle IDs detected per unit time by the free flow antenna 114) as the traffic volume at the vehicle detection location. The traffic volume calculation by the traffic volume calculation unit 133 can be repeated every unit time.

[0046] Furthermore, not all vehicles that pass through a vehicle detection location are equipped with an on-board device capable of communicating with the free-flow antenna 114. Therefore, not all vehicles that pass through a vehicle detection location are detected by the free-flow antenna 114. Accordingly, if the ratio of vehicles equipped with an on-board device capable of communicating with the free-flow antenna 114 to the total number of vehicles (including vehicles not equipped with an on-board device capable of communicating with the free-flow antenna 114) is set in advance as the "on-board device installation ratio," and the number of vehicles detected by the free-flow antenna 114 is set as the "number of detected vehicles," then it is desirable for the traffic volume calculation unit 133 to estimate the traffic volume at the vehicle detection location with higher accuracy using the following formula (1).

[0047] (Estimated traffic volume) = (Number of detected vehicles) ÷ (Percentage of vehicles equipped with onboard devices) ... (1)

[0048] However, such estimation may be omitted if in-vehicle devices capable of communicating with the free-flow antenna 114 are widely available. The traffic volume calculation unit 133 outputs data to the traffic volume data storage unit 124 each time it estimates traffic volume, which is data that associates the estimated traffic volume, the measurement time which is the end of the time period corresponding to that traffic volume, and the vehicle detection position corresponding to that traffic volume.

[0049] (Traffic volume data storage unit 124) The traffic volume data storage unit 124 stores the traffic volume data output from the traffic volume calculation unit 133.

[0050] (Probe antenna 112) The probe antenna 112 functions as an example of a driving history data acquisition unit that acquires vehicle driving history data. That is, in the embodiments of the present invention, it is mainly assumed that the driving history data acquisition unit includes the probe antenna 112. As a result, the ETC probe antenna of an already constructed ETC system can be used as the driving history data acquisition unit, so there is no need to set up a new driving history data acquisition unit. However, other driving history data acquisition units (for example, a mobile phone base station) may be used instead of the probe antenna 112.

[0051] More specifically, when the probe antenna 112 acquires driving history data from the vehicle through communication with an in-vehicle device mounted on the vehicle, it outputs the acquired driving history data to the driving history data storage unit 121.

[0052] (Driving history data storage unit 121) The driving history data storage unit 121 stores the driving history data output from the probe antenna 112. The driving history data stored in the driving history data storage unit 121 can be used by the statistical processing unit 131.

[0053] The driving history data includes the speed of each vehicle that traveled for each section (between kilometer posts) on the road defined by a driving history section width of 612 (e.g., 100m), and for each driving history time width of 611 (e.g., 1 minute), as well as the collection time, which is the time when the speed was collected by the probe antenna 112.

[0054] (Statistical Processing Unit 131) The statistical processing unit 131 acquires driving history data from the driving history data storage unit 121, performs statistical processing on the driving history data, and outputs the statistically processed driving history data as probe data to the probe data storage unit 122. The probe data output from the statistical processing unit 131 to the probe data storage unit 122 can then be mainly used by the mesh data generation unit 135, the learning unit 141, and the inference unit 142.

[0055] For example, the statistical processing unit 131 applies a predetermined statistical process (e.g., averaging) to the speed of one or more vehicles that have traveled for each section (between kilometer posts) on the road and for each travel history time interval 611. This provides statistical data of vehicle speed for each travel history time interval 611 in each section of the road. An example of averaging is taking the harmonic mean.

[0056] (Probe data storage unit 122) Figure 7 shows an example of probe data stored by the probe data storage unit 122. As shown in Figure 7, the probe data is associated with "time of passage", "kilometer post", "speed", and "collection time".

[0057] "Passing time" is the end time of each 611-second interval in the travel history timeline. For example, "Passing time: January 1, 2019, 10:24 AM" is the end time of the time interval that spans 611 seconds, from January 1, 2019, 10:23 AM to January 1, 2019, 10:24 AM. "Kilometer post" is a distance marker from the starting point of the road.

[0058] "Speed" is a statistical value (e.g., average speed) of the speed of one or more vehicles traveling through each section (between kilometer posts) on the road for every 611 time intervals (between passing times). For example, "Speed: 80 km / h" is a statistical value of the speed of one or more vehicles traveling from "Kilometer Post: 100" to "Kilometer Post: 100.1" between 10:23 AM on January 1, 2019 and 10:24 AM on January 1, 2019.

[0059] The "collection time" is the time when the statistical processing unit 131 obtained the "speed," which is a statistical value of the vehicle's speed.

[0060] (KV parameter creation unit 1411) The KV parameter creation unit 1411 acquires speed and passage time for each vehicle detection location for a predetermined period from the probe data storage unit 122 at predetermined intervals. Furthermore, the KV parameter creation unit 1411 acquires traffic volume data (traffic volume and measurement time) for each vehicle detection location for a predetermined period from the traffic volume data storage unit 124 at predetermined intervals.

[0061] The KV parameter creation unit 1411 creates parameters for each vehicle detection location that show the correspondence between traffic density (K) and speed (V) (in the example below, an approximate formula) based on the traffic volume and measurement time for a predetermined period, and the speed and passing time for the same predetermined period, using machine learning.

[0062] More specifically, the KV parameter creation unit 1411 associates traffic volume (Q) and speed (V) corresponding to the same time and the same vehicle detection location. Then, the KV parameter creation unit 1411 calculates the traffic density (K) for each associated traffic volume (Q) and speed (V) using the following equation (2).

[0063] K = Q × 1 / (V × 60) ... (2)

[0064] However, K represents traffic density (vehicles / km), Q represents traffic volume (vehicles / minute), and V represents speed (km / hour). The KV parameter creation unit 1411 approximates the corresponding pairs of traffic density (K) and speed (V) in a predetermined relationship, and creates the parameters of the approximate formula (KV relation) obtained by the approximation as KV parameters.

[0065] Here, we assume that the KV parameter creation unit 1411 creates parameters for an approximation formula corresponding to congested flow. More specifically, the KV parameter creation unit 1411 creates parameters for the KV relation formula corresponding to congested flow by approximating a pair of speeds (V) below the congestion flow determination speed 641 (for example, 55 km / h on expressways) and the traffic density (K) corresponding to that speed (V) in a predetermined relationship. For example, the KV relation formula corresponding to congested flow may be an exponential function expressed by the following equation (3).

[0066] V = a × exp(-b × K) ... (3)

[0067] However, K represents traffic density (vehicles / km), V represents speed (km / hour), and a and b are KV parameters. In this specification, congested flow is used as a concept encompassed within jammed flow, and can refer to a state within jammed flow where the speed is greater than a certain value.

[0068] Figure 8 shows a diagram illustrating pairs of traffic density (K) and speed (V), along with the KV relationship. In the example shown in Figure 8, the horizontal axis represents traffic density (K), and the vertical axis represents speed (V). In this KV diagram, each pair of traffic density (K) and speed (V) is plotted as a point. Referring to Figure 8, an exponential function representing the KV relationship obtained by approximating these pairs is also shown.

[0069] The KV parameter creation unit 1411 stores the created KV parameters as KV parameters 710 in the learning parameter storage unit 127. Typically, the predetermined period may be one month. However, the predetermined period is not limited to one month. For example, the predetermined period may be one day.

[0070] Figure 9 shows an example of the configuration of KV parameters stored in the learning parameter storage unit 127. As shown in Figure 9, the KV parameters include "parameter a" and "parameter b". Furthermore, since "parameter a" and "parameter b" are expected to differ depending on the road shape and the number of lanes, it is desirable that they be created for each vehicle detection position.

[0071] Therefore, as shown in Figure 9, it is desirable that "parameter a" and "parameter b" be associated with the "route code" and "vehicle detection location ID," respectively. The "route code" is a code used to identify the route (the road from the starting point to the ending point). The vehicle detection location ID is a code used to identify the vehicle detection location.

[0072] Furthermore, the learning parameter storage unit 127 may also store standard KV parameters corresponding to the number of lanes. For example, when the number of lanes at a vehicle detection position is obtained, standard KV parameters corresponding to that number of lanes may be used as the initial value of the KV parameter corresponding to that vehicle detection position.

[0073] Figure 10 is a flowchart showing an example of the KV parameter creation process performed by the KV parameter creation unit 1411. First, the KV parameter creation unit 1411 obtains the speed and passing time for a predetermined period for each vehicle detection position from the probe data stored in the probe data storage unit 122 (S11). The KV parameter creation unit 1411 also obtains traffic volume data (traffic volume and measurement time) for a predetermined period for each vehicle detection position from the traffic volume data storage unit 124 (S12).

[0074] The KV parameter creation unit 1411 associates traffic volume (Q) and speed (V) corresponding to the same time and the same vehicle detection location. Then, the KV parameter creation unit 1411 calculates the traffic density (K) for each associated traffic volume (Q) and speed (V) (S13).

[0075] The KV parameter creation unit 1411 creates parameters for the KV relational expression, which shows the relationship between traffic density (K) and speed (V), using machine learning (S14). More specifically, the KV parameter creation unit 1411 approximates the corresponding pairs of traffic density (K) and speed (V) with a predetermined relationship (e.g., an exponential function), and creates the parameters of the KV relational expression obtained by the approximation as KV parameters. Here, we assume that the KV parameter creation unit 1411 creates parameters for an approximation formula corresponding to congested flow.

[0076] The KV parameter creation unit 1411 saves the created KV parameters as KV parameters 710 in the learning parameter storage unit 127 (S15).

[0077] (Traffic density calculation unit 1412) The traffic density calculation unit 1412 acquires the speed and passage time for each section from the probe data storage unit 122, which is stored at a time later than the previous acquisition time of the probe data used to calculate the traffic density, at a pre-set simulation execution time interval 621 (for example, every 5 minutes). The traffic density calculation unit 1412 also acquires the traffic volume data (traffic volume and measurement time) for each vehicle detection position from the traffic volume data storage unit 124, which is stored at a time later than the previous acquisition time of the traffic volume data used to calculate the traffic density, at a simulation execution time interval 621.

[0078] The traffic density calculation unit 1412 calculates the traffic density for each section, at least based on speed. Here, the traffic density in each section may be calculated in any specific way.

[0079] For example, the traffic density calculation unit 1412 may obtain KV parameters from the learning parameter storage unit 127 and calculate the traffic density in a target section of the road based on the speed in that target section and a KV relation defined by the KV parameters.

[0080] More specifically, the traffic density calculation unit 1412 may calculate the traffic density in a section with congested flow, that is, in a target section where the speed is less than or equal to the congestion flow determination speed 641 (threshold), based on the speed in that target section and the KV relationship.

[0081] Furthermore, the traffic density calculation unit 1412 may use a point on the road corresponding to the target section as an estimation point, and calculate the traffic density in the target section based on the traffic volume at the estimation point and the speed in the target section. For example, the estimation point may be the starting point of the target section.

[0082] More specifically, the traffic density calculation unit 1412 may calculate the traffic density in the target section where the speed exceeds the congestion flow determination speed 641 based on the traffic volume at the estimated point and the speed in the target section. For example, the traffic density calculation unit 1412 may calculate the traffic volume at the estimated point based on the traffic volume at the measurement point calculated by the traffic volume calculation unit 133 and the speed for each section from the measurement point to the target section. For example, the measurement point may be a vehicle detection position upstream from the target section (for example, the nearest upstream vehicle detection position from the target section).

[0083] Referring to Figures 11 and 12, we will explain an example of a method for calculating traffic density in a target section based on traffic volume at the estimated point and speed in the target section.

[0084] Figure 11 illustrates an example of calculating the passage time at a measurement point for a vehicle that has reached an estimated point. In the example shown in Figure 11, the horizontal axis represents distance on the road (downward is the direction of downstream movement), and the vertical axis represents time (downward is the direction of passage of time). The speed corresponding to the interval (between kilometer posts) and unit time (i.e., the time from the passage time to the passage time immediately preceding that passage time) in the probe data is indicated by the intensity of the color within the rectangle (spatiotemporal range) corresponding to that interval and unit time. In this example, the darker the color within the rectangle, the lower the speed.

[0085] The traffic density calculation unit 1412 calculates the trajectory of vehicles from the estimated point and the time of arrival at the estimated point to the measurement point, looking upstream and into the past, based on the speed corresponding to each section and unit time. More specifically, the traffic density calculation unit 1412 calculates the trajectory by moving within a rectangle using a straight line with a slope corresponding to the speed corresponding to each section and unit time, and at the boundaries of the rectangle, moving between adjacent rectangles.

[0086] This allows the traffic density calculation unit 1412 to obtain the time at the measurement point reached by the trajectory as the time of passage at the measurement point. Referring to Figure 11, the trajectories of two vehicles are drawn. Here, the traffic volume between the arrival times of the two vehicles at the estimated point and the traffic volume between the passage times of the two vehicles at the measurement point can be considered to be the same.

[0087] Figure 12 illustrates an example of calculating the traffic volume at an estimated location from the traffic volume at a measurement point. Referring to Figure 12, the traffic volume Q during the time of arrival at the estimated location and the traffic volume Q' during the time of passage at the measurement point are shown. Here, the traffic volume for each time of passage at the vehicle detection location that coincides with the measurement point is stored in the traffic volume data storage unit 124. Therefore, the traffic density calculation unit 1412 can obtain the traffic volume Q' during the time of passage at the measurement point from the traffic volume data storage unit 124.

[0088] The traffic density calculation unit 1412 can calculate the traffic volume Q at the estimated location from the traffic volume Q' between passing times at the measurement point, using the ratio of the difference in arrival time at the estimated location to the difference in passing time at the measurement point. For example, if the difference in passing time at the measurement point is t1 and the difference in arrival time at the estimated location is t2, the traffic density calculation unit 1412 can calculate the traffic volume Q at the estimated location by multiplying the traffic volume Q' between passing times at the measurement point by t1 / t2. Hereinafter, this method of calculating the traffic volume Q at the estimated location from the traffic volume Q' at the measurement point will also be referred to as the "vehicle tracking" traffic volume calculation method.

[0089] Figure 13 is a flowchart showing an example of the traffic density calculation process performed by the traffic density calculation unit 1412. First, the traffic density calculation unit 1412 acquires probe data (speed and passage time for each section) stored at a time later than the previous acquisition time of the probe data used to calculate the traffic density from the probe data storage unit 122 at a pre-set simulation execution time interval 621 (S21).

[0090] Furthermore, the traffic density calculation unit 1412 acquires traffic volume data (traffic volume and measurement time) for each vehicle detection location from the traffic volume data storage unit 124 at a simulation execution time interval 621, which is stored at a time later than the previous acquisition time of the traffic volume data used to calculate the traffic density (S22). The traffic density calculation unit 1412 sequentially sets each section from the start to the end of the road as a target section and determines whether the speed in the target section is less than or equal to the speed of congested flow, that is, the speed of congestion flow determination 641 (S23).

[0091] The traffic density calculation unit 1412 calculates the traffic density in the target section based on the speed in that section and the KV relation expression if the speed in the target section is the speed of congested flow (i.e., a speed that is less than or equal to the congestion flow determination speed 641) (YES in S23) (S24).

[0092] On the other hand, if the speed in the target section is the speed of free flow (i.e., a speed exceeding the congestion flow determination speed 641) (NO in S23), the traffic density calculation unit 1412 calculates the traffic volume at the measurement point by tracking vehicles, and calculates the traffic density in the target section based on the calculated traffic volume at the measurement point and the speed for each section from the measurement point to the target section (S25). The traffic density calculation unit 1412 outputs traffic density data, which associates the measurement time of traffic volume, section, and traffic density, to the traffic density data storage unit 125.

[0093] (Traffic density data storage unit 125) The traffic density data storage unit 125 stores the traffic density data output from the traffic density calculation unit 1412.

[0094] (Mesh data generation unit 135) The mesh data generation unit 135 acquires probe data from the probe data storage unit 122. As described above, the probe data is composed of data associated with passage time, kilometer post, speed, and collection time. On the other hand, in the following learning process, data (hereinafter also referred to as "mesh data") consisting of multiple data (hereinafter also referred to as "mesh") associated with time, cell, speed, and traffic density is used. Therefore, the mesh data generation unit 135 generates mesh data based on the probe data. Note that a cell can also be described as a location on the road or a point on the road.

[0095] Here, the width of the section on the road (between kilometer posts) is defined by the travel history section width of 612. On the other hand, the width of the cell is defined by the simulation execution cell width of 622.

[0096] The width of a cell may be the same as the width of a section, or it may be different from the width of a section. For example, if the width of a section is 100m, the width of a cell may be 500m. In cases where the width of a cell differs from the width of a section, as in this example, the mesh data generation unit 135 may identify multiple sections belonging to the cell, perform statistical processing on the speed corresponding to each of those sections, and obtain the speed after statistical processing as the speed corresponding to the cell. Furthermore, the mesh data generation unit 135 may perform statistical processing on the traffic density corresponding to each of those sections, and obtain the traffic density after statistical processing as the traffic density corresponding to the cell.

[0097] The time intervals corresponding to sections on the road are defined in the driving history time interval 611. On the other hand, the time intervals corresponding to cells are defined in the simulation execution time interval 621.

[0098] The simulation execution time interval 621 (the time interval corresponding to the cell) may be the same as the driving history time width 611 (the time interval corresponding to the section), or it may be different from the driving history time width 611 (the time interval corresponding to the section). For example, if the driving history time width 611 (the time interval corresponding to the section) is 1 minute, the simulation execution time interval 621 (the time interval corresponding to the cell) may be 5 minutes.

[0099] As in this example, if the simulation execution time interval 621 (time interval corresponding to a cell) differs from the travel history time width 611 (time interval corresponding to a section), the mesh data generation unit 135 may identify the section-corresponding time that belongs to the time corresponding to the cell, perform statistical processing on the speed corresponding to the identified time, and obtain the speed after statistical processing as the speed for the time corresponding to the cell. Furthermore, the mesh data generation unit 135 may perform statistical processing on the traffic density corresponding to the identified time, and obtain the traffic density after statistical processing as the traffic density corresponding to the cell.

[0100] Figure 14 schematically shows the mesh data used for training. Referring to Figure 14, cells 1 to n are shown from the upstream side to the downstream side of the road. The upstream end of cell 1 is the starting point of the road, and the downstream end of cell n is the ending point of the road. i is any integer satisfying 1 ≤ i ≤ N. Δt is the time interval corresponding to the cell. h is an integer greater than or equal to 1.

[0101] Referring to Figure 14, the velocity V1(t) ~ V n (t) (km / hour) is shown. Below, we will mainly explain the case where time t is the current time, but time t is not limited to the current time. The velocity V1(t-hΔt)~V corresponding to time t-hΔt n From (t-hΔt), the velocity V1(t)~V corresponds to time t (current time). n Up to (t), it is mainly used by the traffic density pattern learning unit 1413.

[0102] Referring to Figure 14, traffic density K1(t) ~ K n(t) (vehicles / km) is shown. From the traffic density K1(t - hΔt) to K n (t - hΔt) corresponding to the time t - hΔt to the traffic density K1(t) to K n (t) corresponding to the time t (current time) is mainly used by the traffic density pattern learning unit 1413.

[0103] The mesh data generation unit 135 is from the speed V1(t - hΔt) to V n (t - hΔt) corresponding to the time t - hΔt to the speed V1(t) to V n (t) corresponding to the time t (current time), and from the traffic density K1(t - hΔt) to K n (t - hΔt) corresponding to the time t - hΔt to the traffic density K1(t) to K n (t) corresponding to the time t (current time) are output as mesh data to the mesh data storage unit 129.

[0104] For example, the mesh data generation unit 135 updates the time t (current time) every time interval Δt, and the speed V1(t) to speed V n (t) and the traffic density K1(t) to K n (t) corresponding to the updated time t (current time) may be output to the mesh data storage unit 129.

[0105] (Mesh data storage unit 129) The mesh data storage unit 129 stores the mesh data output from the mesh data generation unit 135.

[0106] (Traffic density pattern learning unit 1413) The traffic density pattern learning unit 1413 performs traffic density pattern learning. Here, the outline of the traffic density pattern learning will be described while referring to FIGS. 15 to 16.

[0107] FIG. 15 is a diagram for explaining the outline of the traffic density pattern learning. Referring to FIG. 15, the speed corresponding to the time and the distance from the starting point of the cell is shown by the shade of the color of the rectangular area corresponding to the time and the distance from the starting point of the cell.

[0108] For example, according to traffic density pattern learning, a traffic density pattern H1 is stored in the learning parameter storage unit 127, which associates a previous time traffic density pattern, consisting of multiple traffic densities from the traffic density h11 of a cell two cells downstream of a given cell at a given time to the traffic density h17 of a cell four cells upstream of that cell at that time, with the traffic density h23 of that cell at the next time.

[0109] Similarly, according to traffic density pattern learning, traffic density patterns H2 to H4 are stored in the learning parameter storage unit 127, which associate a previous time traffic density pattern, consisting of multiple traffic densities from the traffic density of two cells downstream of another cell at another time to the traffic density of four cells upstream of that other cell at the same time, with the traffic density of that other cell at the next time.

[0110] In the example shown in Figure 15, the previous time traffic density pattern (traffic densities h11-h17) is composed of traffic densities corresponding to seven cells. However, the number of traffic densities included in the previous time traffic density pattern is not limited to seven.

[0111] Figure 16 shows an example of a traffic density pattern. As shown in Figure 16, the traffic density pattern stored in the learning parameter storage unit 127 is configured by associating the next time-series traffic density of a cell with the previous time-series traffic density pattern (multiple traffic densities from the traffic density of the cell two cells downstream to the traffic density of the cell four cells upstream of the cell). Details of this traffic density pattern learning will be explained with reference to Figure 17.

[0112] Figure 17 is a flowchart illustrating an example of the operation of traffic density pattern learning by the traffic density pattern learning unit 1413. An example of the operation of traffic density pattern learning by the traffic density pattern learning unit 1413 will be explained with reference to Figure 17.

[0113] First, when the current time reaches a preset learning time, the traffic density pattern learning unit 1413 acquires traffic density data for a predetermined target time period based on the current time from the traffic density data storage unit 125 as new data (S41). The new data is traffic density data that has not yet been used for learning. The traffic density pattern learning unit 1413 then acquires the traffic density from the new data.

[0114] Next, the traffic density pattern learning unit 1413 identifies the upstream cell of the target cell (S42) for each time period and for each cell, based on the acquired traffic density. More specifically, the traffic density pattern learning unit 1413 sets the target cells in order from cells 1 to n, calculates the travel time assuming that vehicles travel at the speed corresponding to each cell and time, starting from cells located upstream of the target cell toward the target cell, calculates the total travel time, and identifies the cell located upstream when the total first reaches a predetermined time range as the upstream cell. At this time, an upper limit (for example, 1 km) may be set for the distance from the target cell to the cell located upstream.

[0115] Furthermore, the traffic density pattern learning unit 1413 identifies downstream cells based on a pre-set congestion wave propagation speed 645 (e.g., 18 km / h) (S43). More specifically, the traffic density pattern learning unit 1413 calculates the distance traveled by a vehicle assuming that the vehicle traveled for a pre-set simulation execution time interval 621 at a congestion wave propagation speed 645, adds a pre-set predicted pattern downstream margin (e.g., 200 m) to the calculated distance, and identifies the cell that is furthest from the target cell and whose distance from the target cell is within the distance of the added result as the downstream cell.

[0116] The traffic density pattern learning unit 1413 generates a relationship between the previous time traffic density pattern and the next time traffic density in the target cell by associating the previous time traffic density pattern, which includes the traffic density from upstream to downstream cells of the target cell, with the next time traffic density in the target cell. The traffic density pattern learning unit 1413 then stores this relationship as a traffic density pattern 720 in the learning parameter storage unit 127 (S44).

[0117] (Predicted traffic density and speed) As described above, a traffic density pattern 720 is generated by traffic density pattern learning. The inference unit 142 can predict the traffic density and speed at a time later than the current time by performing inference using the traffic density pattern 720 thus generated.

[0118] Figure 18 schematically shows the traffic density and speed to be predicted. Referring to Figure 18, similar to the schematic diagram of traffic density data used for training (Figure 14), the traffic densities K1(t)~K of cells 1~n at time t are shown. n (t) (units / km) is shown.

[0119] Referring to Figure 18, similarly, the traffic density K1(t+Δt)~K of cells 1~n at time t+Δt. n (t+Δt) is shown, and the traffic density K1(t+MΔt)~K of cells 1~n at time t+MΔt is shown. n (t+MΔt) is shown, where M is an integer greater than or equal to 1.

[0120] In the inference performed by the inference unit 142, traffic flow characteristics corresponding to time t+Δt~t+MΔt and cells 1~N are predicted. The following describes how such traffic flow characteristics are predicted. Examples of traffic flow characteristics include at least one of traffic density K and velocity V.

[0121] (Traffic volume forecasting unit 1421) The traffic volume prediction unit 1421 acquires mesh data from the mesh data storage unit 129. Then, based on the mesh data, the traffic volume prediction unit 1421 performs the following processes in order: "extraction of congestion data" and "calculation of cumulative inflow and cumulative outflow traffic volume." These processes will be explained in order below.

[0122] (Extraction of traffic congestion data) The traffic volume prediction unit 1421 extracts the times when congestion occurs (hereinafter also referred to as "congestion times") based on the mesh data acquired from the mesh data storage unit 129, and also extracts the cells where congestion occurs at those times (hereinafter also referred to as "congestion cells").

[0123] The traffic volume forecasting unit 1421 extracts combinations of congestion times and congestion cells as congestion data. If the traffic volume forecasting unit 1421 extracts multiple combinations of congestion times and congestion cells, it only needs to extract combinations that are consecutive in time or space from among those multiple combinations as congestion data.

[0124] Figure 19 shows an example of velocity corresponding to a combination of time and cell. Referring to Figure 19, the combinations of time t (current time) and cells 20-40, as well as the corresponding velocities in their surrounding areas, are shown. In Figure 19, the vertical axis represents time, and the horizontal axis represents locations on the road. Locations on the road can be represented by cells. The velocity corresponding to time and cell is indicated by the intensity of the hatching.

[0125] Referring to Figure 19, congestion data D1 extracted from mesh data is shown. For example, the traffic volume prediction unit 1421 may determine whether the speed is less than or equal to the congestion flow determination speed 641 based on the speed corresponding to each combination of time and cell. If the traffic volume prediction unit 1421 determines that the speed is less than or equal to the congestion flow determination speed 641, it may extract that cell as a congestion cell and extract that time as the congestion time.

[0126] The congestion flow determination speed 641 is a threshold value stored in the master data storage unit 126 beforehand. For example, if the prediction target is the traffic flow on an expressway, the congestion flow determination speed 641 may be set to 40 km / h in accordance with the definition of congestion set by the Ministry of Land, Infrastructure, Transport and Tourism. Congestion data D1 is a combination of multiple combinations of congestion time and congestion cell that are consecutive in time or space. In the following explanation, we mainly assume that one congestion data D1 is extracted by the traffic volume prediction unit 1421, but multiple congestion data may be extracted by the traffic volume prediction unit 1421.

[0127] (Calculation of cumulative inflow and outflow traffic) The traffic volume prediction unit 1421 detects the start time of congestion based on the congestion data D1. Furthermore, based on the extraction of the congestion data D1, the traffic volume prediction unit 1421 identifies a range corresponding to the congestion data D1. The range identified by the traffic volume prediction unit 1421 corresponds to the measurement range for the traffic congestion volume, which will be calculated later. Therefore, in the following explanation, the range identified by the traffic volume prediction unit 1421 will also be referred to as the "traffic congestion volume measurement range."

[0128] For example, since there is more vehicle movement outside of the congestion data D1 than inside the congestion data D1, it is possible that the traffic volume can be measured more accurately. Therefore, the traffic volume prediction unit 1421 identifies the area from the vehicle detection position located downstream of the cell at the furthest downstream position in the congestion data D1 (hereinafter also referred to as the "congestion leading cell") (hereinafter also referred to as the "downstream vehicle detection position P2") to the vehicle detection position located upstream of the cell at the furthest upstream position in the congestion data D1 (hereinafter also referred to as the "congestion trailing cell") (hereinafter also referred to as the "upstream vehicle detection position P1") as the congestion volume measurement range.

[0129] Figure 20 shows the traffic congestion data D1. In the traffic congestion data D1 shown in Figure 20, the hatching indicating speed has been removed from within the traffic congestion data D1. In the traffic congestion data D1, the time when the congestion started is time t-5Δt. The period from time t-5Δt (time when the congestion started) to time t (current time) may correspond to the first time. Also, in the traffic congestion data D1, cell 51, which is at the furthest downstream position, may correspond to the leading cell of the congestion. Also, in the traffic congestion data D1, cell 17, which is at the furthest upstream position, may correspond to the trailing cell of the congestion.

[0130] In this case, the traffic volume prediction unit 1421 may identify the area from the downstream vehicle detection position P2, located downstream of cell 51, which is the leading cell of the congestion, to the upstream vehicle detection position P1, located upstream of cell 17, which is the trailing cell of the congestion, as the congestion traffic volume measurement range (P1~P2). The traffic volume prediction unit 1421 then acquires the inflow traffic volume that flowed into the congestion traffic volume measurement range (P1~P2) from the traffic volume data storage unit 124 between time t-5Δt (congestion start time) and time t (current time), and acquires the outflow traffic volume that flowed out of the congestion traffic volume measurement range (P1~P2) from the traffic volume data storage unit 124 between time t-5Δt (congestion start time) and time t (current time).

[0131] Figure 21 is a diagram illustrating the traffic volume prediction by the traffic volume prediction unit 1421. Referring to Figure 21, the traffic volume Q at the upstream vehicle detection position P1 at time t-5Δt (time of congestion start) is shown. P1 (t-5Δt) is shown. Furthermore, the traffic volume Q at the downstream vehicle detection position P2 at time t-5Δt (time of congestion start) is shown. P2 (t-5Δt) is shown.

[0132] The traffic volume prediction unit 1421 calculates the traffic volume Q P1 (t-5Δt) is the incoming traffic volume Q that entered the traffic volume measurement area (P1~P2) at time t-5Δt (time of congestion start). P1 It can be obtained as (t-5Δt). In addition, the traffic volume prediction unit 1421 obtains the traffic volume Q P2 (t-5Δt) is the outflow traffic volume Q that left the congestion measurement area (P1~P2) at time t-5Δt (time of congestion start).P2 It can be obtained as (t-5Δt).

[0133] Similarly, the traffic volume prediction unit 1421 calculates the incoming traffic volume Q that entered the traffic volume measurement range (P1~P2) from time t-4Δt to time t (current time). P1 (t-4Δt)~Q P1 (t) can be obtained. Inflow traffic volume Q P1 (t-5Δt)~Q P1 (t) may correspond to the first inflow traffic volume. Furthermore, the traffic volume prediction unit 1421 calculates the outflow traffic volume Q that has flowed out of the congestion traffic volume measurement range (P1~P2) from time t-4Δt to time t (current time). P2 (t-4Δt)~Q P2 (t) can be obtained. Outflow traffic volume Q P2 (t-5Δt)~Q P2 (t) may correspond to the first outflow traffic volume.

[0134] The traffic volume prediction unit 1421 calculates the incoming traffic volume Q. P1 (t-5Δt)~Q P1 Based on (t), the incoming traffic volume Q is the amount of traffic that flows into the traffic volume measurement range (P1~P2) at each time from time t+Δt to time t+MΔt. P1 (t+Δt)~Q P1 Predict (t+MΔt). Note that time t+Δt may correspond to a second time. Also, the incoming traffic volume Q P1 (t+Δt) may represent the second inflow of traffic.

[0135] Note: Inflow traffic volume Q P1 (t+Δt)~Q P1 Any prediction method can be used to predict (t+MΔt), such as time series prediction methods like the Autoregressive Integrated Moving Average (ARIMA) model or the Seasonal Autoregressive Integrated Moving Average (SARIMA) model, or spatiotemporal prediction methods like graph neural networks (GNNs).

[0136] Furthermore, the traffic volume prediction unit 1421 calculates the outflow traffic volume Q P2 (t-5Δt)~Q P2 Based on (t), the outflow traffic volume Q is the amount of traffic flowing out of the congestion traffic volume measurement range (P1~P2) at each time from time t+Δt to time t+MΔt. P2 (t+Δt)~Q P2 Predict (t+MΔt). Note that the outflow traffic volume Q P2 (t+Δt) may correspond to the second outflow traffic volume.

[0137] Note: Outflow traffic volume Q P2 (t-5Δt)~Q P2 (t) is predicted to be a nearly constant value, limited by the traffic capacity, which is the maximum number of vehicles per unit time that can pass through a congested cell. Therefore, the traffic volume prediction unit 1421 predicts the outflow traffic volume Q P2 (t-5Δt)~Q P2 The outflow traffic volume Q is obtained by linearly extrapolating (t) P2 (t+Δt)~Q P2 (t+MΔt) can be predicted, or the outflow traffic volume Q can be predicted using a method that predicts traffic capacity. P2 (t+Δt)~Q P2 We may also predict (t+MΔt).

[0138] The traffic volume prediction unit 1421 calculates the cumulative inflow traffic volume R, which is the cumulative value of inflow traffic volume from time t-5Δt (time of congestion start) to time t+Δt. P1 (t+Δt)=Q P1 (t-5Δt)+···+Q P1 (t+Δt) is calculated. Similarly, the traffic volume prediction unit 1421 calculates the cumulative inflow traffic volume R P1 (t+2Δt)=Q P1 (t-5Δt)+···+Q P1 From (t+2Δt), cumulative inflow traffic R P1 (t+MΔt)=Q P1 (t-5Δt)+···+Q P1 Calculate up to (t+MΔt).

[0139] Furthermore, the traffic volume prediction unit 1421 calculates the cumulative outflow traffic volume R, which is the cumulative value of outflow traffic volume from time t-5Δt (time of congestion start) to time t+Δt. P2 (t+Δt)=Q P2 (t-5Δt)+···+Q P2 (t+Δt) is calculated. Similarly, the traffic volume prediction unit 1421 calculates the cumulative outflow traffic volume R P2 (t+2Δt)=Q P2 (t-5Δt)+···+Q P2 From (t+2Δt), cumulative inflow traffic R P2 (t+MΔt)=Q P2 (t-5Δt)+···+Q P2 Calculate up to (t+MΔt).

[0140] (Stayed traffic calculation unit 1422) The traffic volume calculation unit 1422 calculates the cumulative inflow traffic volume R P1 (t+Δt) and cumulative outflow traffic R P2 The difference from (t+Δt) is the traffic congestion R diff (t+Δt)=R P1 (t+Δt)-R P2 It is calculated as (t+Δt). Similarly, the traffic congestion volume calculation unit 1422 calculates the traffic congestion volume R diff (t+2Δt)=R P1 (t+2Δt)-R P2 From (t+2Δt), the amount of stagnant traffic R diff (t+MΔt)=R P1 (t+MΔt)-R P2 Calculate up to (t+MΔt).

[0141] Furthermore, traffic congestion volumes that take negative values ​​are thought to be caused by observation errors or prediction errors. Therefore, the traffic congestion volume calculation unit 1422 calculates the traffic congestion volume R diff (t+Δt) to stagnant traffic R diff The traffic congestion R takes on a negative value by (t+MΔt). diff If such a traffic volume R takes a negative value, diff It may be corrected to 0 (zero). Furthermore, if congestion data D1 is not extracted, the traffic congestion calculation unit 1422 may set the traffic congestion to 0 (zero).

[0142] (Operations of Traffic Volume Prediction Unit 1421 and Standing Traffic Volume Calculation Unit 1422) FIG. 22 is a flowchart showing an example of the operations of traffic volume prediction unit 1421 and standing traffic volume calculation unit 1422. With reference to FIG. 22, the operations of traffic volume prediction unit 1421 and standing traffic volume calculation unit 1422 will be sorted out.

[0143] The traffic volume prediction unit 1421 acquires mesh data from the mesh data storage unit 129 (S501). Then, the traffic volume prediction unit 1421 attempts to extract congestion data D1 based on the mesh data. If the congestion data D1 is not extracted (S502: NO), the traffic volume prediction unit 1421 sets 0 (zero) for the standing traffic volume (S503) and ends the operation. On the other hand, if the congestion data D1 is extracted (S502: YES), the traffic volume prediction unit 1421 detects the congestion start time (S504).

[0144] The traffic volume prediction unit 1421 specifies the standing traffic volume measurement range (P1 to P2) corresponding to the congestion data D1. Then, the traffic volume prediction unit 1421 determines the inflow traffic volume Q P1 (t - 5Δt) to Q P1 Based on (t), the inflow traffic volume Q P1 (t + Δt) to Q P1 (t + MΔt) flowing into the standing traffic volume measurement range (P1 to P2) at each time from time t + Δt to time t + MΔt is predicted (S505).

[0145] The traffic volume prediction unit 1421 calculates the cumulative inflow traffic volume R P1 (t + Δt) = Q P1 (t - 5Δt) + ··· + Q P1 (t + Δt) to calculate the cumulative inflow traffic volume R P1 (t + MΔt) = Q P1 (t - 5Δt) + ··· + Q P1 (t + MΔt) (S507).

[0146] Furthermore, the traffic volume prediction unit 1421 calculates the outflow traffic volume Q P2 (t - 5Δt) to QP2 Based on (t), the outflow traffic volume Q flowing out from the traffic volume measurement range (P1 to P2) at each of the times from time t+Δt to time t+MΔt P2 (t+Δt)~Q P2 (t+MΔt) is predicted (S506).

[0147] The traffic volume prediction unit 1421 calculates the cumulative outflow traffic volume R P2 (t+Δt)=Q P2 (t-5Δt)+···+Q P2 From the cumulative outflow traffic volume R P2 (t+MΔt)=Q P2 (t-5Δt)+···+Q P2 Up to (t+MΔt) is calculated (S508).

[0148] The traffic volume calculation unit 1422 calculates the traffic volume R diff (t+Δt)=R P1 (t+Δt)-R P2 From the traffic volume R diff (t+MΔt)=R P1 (t+MΔt)-R P2 Up to (t+MΔt) is calculated (S509).

[0149] The traffic volume calculation unit 1422 calculates the traffic volume R diff From the traffic volume R diff Up to (t+MΔt), if there is a traffic volume R that takes a negative value diff The traffic volume R that takes the negative value diff May be corrected to 0 (zero) (S510).

[0150] (Congestion basin prediction unit 1423) The congestion basin prediction unit 1423 obtains the traffic volume R diff (t+Δt)~R diff (t+MΔt) from the traffic volume calculation unit 1422. Then, the congestion basin prediction unit 1423 calculates the traffic volume R diff (t+Δt)~R diffBased on (t+MΔt), the congested watershed is predicted for each time point from t+Δt to t+MΔt. Note that the congested watershed may be a continuous range of congested cells at each time point.

[0151] Figure 23 is a diagram illustrating a congested watershed. As shown in Figure 23, the congestion watershed prediction unit 1423 predicts the amount of traffic congestion R diff The congestion length L(t+Δt) is predicted according to (t+Δt). Here, the amount of traffic congestion R diff The method for predicting the congestion length L(t+Δt) from (t+Δt) is not limited. For example, the congestion basin prediction unit 1423 uses the amount of traffic congestion R diff The congestion length L(t+Δt) can be predicted such that the congestion length L(t+Δt) increases as (t+Δt) increases.

[0152] For example, the congestion basin prediction unit 1423 calculates the amount of traffic congestion R diff Assuming that (t+Δt) and the congestion length L(t+Δt) are proportional, the amount of traffic congestion R diff The congestion length L(t+Δt) can be predicted from (t+Δt). Alternatively, at a specific location (for example, a location where the number of lanes decreases), the traffic volume R can be predicted. diff It is possible that (t+Δt) and the congestion length L(t+Δt) are not proportional.

[0153] In such cases, the congestion basin prediction unit 1423 may calculate the number of vehicles in the congestion basin at time t by multiplying the traffic density (vehicles / km) of the congestion basin at time t by the length of the congestion basin (km). Then, the congestion basin prediction unit 1423 considers the relationship between the number of vehicles in the congestion basin at time t and the traffic volume (vehicles / second) at time t, and the traffic volume R at time t+Δt. diffThe congestion area prediction unit 1423 may calculate the predicted number of vehicles in the congested area at time t+Δt based on (t+Δt)(vehicles / second). The congestion area prediction unit 1423 may calculate the predicted number of vehicles per lane in the congested area at time t+Δt by dividing the predicted number of vehicles in the congested area at time t+Δt by the number of lanes. The congestion area prediction unit 1423 may then calculate the congestion length L(t+Δt) by assuming that the predicted number of vehicles per lane in the congested area at time t+Δt is proportional to the congestion length L(t+Δt).

[0154] The congestion basin prediction unit 1423 predicts the congestion basin E1(t+Δt) at time t+Δt based on the congestion length L(t+Δt). More specifically, it is considered highly likely that the congestion front will not change even as time passes. Therefore, the congestion basin prediction unit 1423 predicts the congestion basin E1(t+Δt) at time t+Δt based on cell 51, which is the congestion front cell in the congestion data D1, and the congestion length L(t+Δt). Even more specifically, the congestion basin prediction unit 1423 predicts the congestion basin E1(t+Δt) at time t+Δt as the area from cell 51, which is the congestion front cell, upstream by a distance of congestion length L(t+Δt) relative to cell 51, which is the congestion front cell.

[0155] In the example shown in Figure 23, the congestion length L(t+Δt) at time t+Δt is equivalent to the width of 32 cells. Therefore, the congestion basin prediction unit 1423 may predict the congestion basin E1(t+Δt) at time t+Δt as extending from cell 51, which is the leading cell of the congestion, to cell 19, which is 32 cells upstream from cell 51, the leading cell of the congestion.

[0156] Although Figure 23 only shows the congested watershed from E1(t+Δt) to E1(t+5Δt), the congestion watershed prediction unit 1423 may similarly predict the congested watershed E1(t+2Δt) to E1(t+MΔt) at time t+2Δt to t+MΔt, starting from cell 51, which is the leading cell of the congestion, up to a cell upstream of L(t+2Δt) to L(t+MΔt) relative to cell 51, which is the leading cell of the congestion.

[0157] (Operation of the congestion basin prediction unit 1423) Figure 24 is a flowchart illustrating an example of the operation of the congestion basin prediction unit 1423. The operation of the congestion basin prediction unit 1423 will be summarized with reference to Figure 24.

[0158] The congestion basin prediction unit 1423 calculates the traffic congestion volume R calculated by the traffic congestion volume calculation unit 1422. diff (t+Δt)~R diff (t+MΔt) is obtained (S601). Then, the congestion basin prediction unit 1423 repeats S602 to S603 for each time interval from time t+Δt to time t+MΔt.

[0159] First, the congestion basin prediction unit 1423 calculates the amount of traffic congestion R diff Based on this, the congestion length L is predicted (S602). Subsequently, the congestion area prediction unit 1423 predicts the congestion area based on the congestion length L (S603). The congestion area prediction unit 1423 repeats steps S602 to S603 for each time interval from time t+Δt to time t+MΔt, and then terminates its operation.

[0160] (Traffic density prediction unit 1424) The traffic density prediction unit 1424 calculates the traffic density K1(t+Δt) ~ K based on the congestion area E1(t+Δt) at time t+Δt, which was predicted by the congestion area prediction unit 1423. n The traffic density prediction unit 1424 sequentially sets the target cell j from cells 1 to n, and obtains a congestion determination result by determining whether the target cell j at time t+Δt belongs to the congested watershed E1(t+Δt). Based on the congestion determination result, it predicts the traffic density K of the target cell j at time t+Δt. j Predict (t+Δt).

[0161] More specifically, the learning parameter storage unit 127 stores a traffic density pattern (Figure 16). The traffic density pattern stored in the learning parameter storage unit 127 is configured by associating the next time zone traffic density of a cell with the previous time zone traffic density pattern. The cell may correspond to a predetermined location. The next time zone may correspond to a third time zone. Furthermore, the previous time zone may correspond to a fourth time zone.

[0162] The traffic density prediction unit 1424 calculates the traffic density K of the target cell j at time t+Δt based on the traffic density pattern (Figure 16) stored in the learning parameter storage unit 127 and the congestion determination result. j Predict (t+Δt).

[0163] For example, if the target cell j at time t+Δt belongs to the congested watershed E1(t+Δt), the traffic density prediction unit 1424 obtains from the learning parameter storage unit 127 a traffic density pattern in which cell i is a congested cell at time u+Δt (hereinafter also referred to as the "congested flow pattern"). On the other hand, if the target cell j at time t+Δt does not belong to the congested watershed E1(t+Δt), the traffic density prediction unit 1424 obtains from the learning parameter storage unit 127 a traffic density pattern in which cell i is not a congested cell at time u+Δt (hereinafter also referred to as the "free flow pattern").

[0164] The traffic density prediction unit 1424 then uses the acquired congestion flow pattern or free flow pattern and the traffic density K at time t+Δt. j (t+Δt) and the traffic density K, which is the traffic density pattern at time t prior to time t. j-4 (t)~K j+2 A pattern matching prediction is performed based on (t).

[0165] For example, the traffic density prediction unit 1424 predicts the traffic density K, which is the traffic density pattern at time t. j-4 (t)~K j+2 If (t) is similar to the traffic density pattern at the previous time in the acquired congestion flow pattern or free flow pattern, then the traffic density K i (u+Δt) is the traffic density Kj The prediction is made as (t+Δt). The traffic density prediction unit 1424 may determine whether the traffic density patterns of previous time points are similar by checking whether the error calculated by the mean squared error between the traffic density patterns of previous time points is smaller than a predetermined error.

[0166] Figure 23 shows a traffic density pattern that associates the traffic density pattern from the previous time, which is composed of multiple traffic densities from the traffic density h31 of cell 23 two positions downstream of cell 21 at time t+Δt, to the traffic density h37 of cell 17 four positions upstream of cell 21 at time t+Δt, with the traffic density h43 of cell 21 at the next time t+Δt. In this traffic density pattern, since cell 21 at time t+Δt belongs to the congested watershed E1(t+Δt), the traffic density prediction unit 1424 acquires the congestion flow pattern, and pattern matching prediction is performed based on the congestion flow pattern.

[0167] Note that the traffic density at time t+Δt is K1(t+Δt)~K n If (t+Δt) is predicted, then the traffic density K1(t+Δt)~K at time t+Δt can be predicted. n Based on (t+Δt), the traffic density at time t+2Δt is K1(t+Δt)~K n (t+2Δt) can be similarly predicted. Similarly, the traffic density K1(t+3Δt)~K at time t+3Δt~t+MΔt n (t+3Δt)~K1(t+MΔt)~K n (t+MΔt) can also be predicted in a chain reaction. However, different traffic density patterns are used for prediction depending on whether the cell with the predicted traffic density belongs to a congested watershed at the time of prediction. Therefore, traffic density can be predicted with greater accuracy.

[0168] (Operation of the traffic density prediction unit 1424) Figure 25 is a flowchart illustrating an example of the operation of the traffic density prediction unit 1424. The operation of the traffic density prediction unit 1424 will be summarized with reference to Figure 25.

[0169] The traffic density prediction unit 1424 acquires mesh data from the mesh data storage unit 129 (S701). Then, the traffic density prediction unit 1424 acquires the congested watershed for each time t+Δt to t+MΔt from the congested watershed prediction unit 1423 (S702). Then, the traffic density prediction unit 1424 repeats steps S703 to S706 for each mesh.

[0170] First, the traffic density prediction unit 1424 determines whether the cell corresponding to the mesh belongs to a congested watershed at the time corresponding to the mesh (S703). If the cell corresponding to the mesh belongs to a congested watershed at the time corresponding to the mesh (S703: YES), the traffic density prediction unit 1424 obtains a congested flow pattern from the learning parameter storage unit 127. On the other hand, if the cell corresponding to the mesh does not belong to a congested watershed at the time corresponding to the mesh (S703: NO), the traffic density prediction unit 1424 obtains a free flow pattern from the learning parameter storage unit 127 (S705).

[0171] The traffic density prediction unit 1424 then performs a pattern matching prediction based on the acquired congestion flow pattern or free flow pattern and the traffic density pattern corresponding to the mesh (S706). The traffic density prediction unit 1424 repeats S703 to S706 for each mesh and then terminates its operation.

[0172] (Provider 1425) The providing unit 1425 may provide prediction results. For example, the providing unit 1425 outputs the prediction results to the prediction result storage unit 128. For example, the providing unit 1425 outputs the traffic density K1(t+Δt)~K predicted by the traffic density prediction unit 1424. n (t+Δt)~K1(t+MΔt)~K n (t+MΔt) may be included in the prediction result. Furthermore, the providing unit 1425 may include the congestion length L(t+Δt)~L(t+MΔt) in the prediction result.

[0173] Furthermore, the supply unit 1425 provides speed V1(t+Δt)~V n (t+Δt)~V1(t+MΔt)~Vn (t+MΔt) is predicted, and the predicted velocity V1(t+Δt)~V n (t+Δt)~V1(t+MΔt)~V n (t+MΔt) may be included in the prediction result. For example, the supply unit 1425 may calculate the speed corresponding to the traffic density using the KV relation shown by equation (3) above.

[0174] Alternatively, if the predicted speed of a cell whose speed is predicted at the next time (target cell) is greater than the congestion flow determination speed 641, the supply unit 1425 may use the current speed of the cell closest to the target cell among the cells located upstream of the target cell, whose speed at the current time is greater than the congestion flow determination speed 641, as the predicted speed of the target cell at the next time.

[0175] Furthermore, the supply unit 1425 provides speed V1(t+Δt)~V n (t+Δt)~V1(t+MΔt)~V n Based on (t+MΔt), the time required for a vehicle to travel from the starting point to the ending point at the predicted speed corresponding to each cell at each time step may be calculated, and this calculated time may be included in the prediction result.

[0176] (Prediction result storage unit 128) The prediction result storage unit 128 stores the prediction results output from the inference unit 142.

[0177] The embodiments of the present invention have been described in detail above.

[0178] (1-2. Effects) As described above, according to the embodiment of the present invention, the difference between the cumulative inflow traffic volume into the area corresponding to the congestion data and the cumulative outflow traffic volume from the area corresponding to the congestion data is calculated as the stagnant traffic volume, and a congested watershed corresponding to the stagnant traffic volume is predicted. Then, different traffic density patterns are used to predict traffic density depending on whether the cell of the predicted traffic density belongs to a congested watershed at the time of the predicted traffic density. This makes it possible to predict traffic density with greater accuracy.

[0179] Figure 26 shows the prediction results from traffic flow prediction according to a comparative example. Figure 27 shows the prediction results from traffic flow prediction according to an embodiment of the present invention. The prediction results include traffic density corresponding to time and location.

[0180] As can be seen from Figures 26 and 27, in the comparative example of traffic flow prediction, the difference between the predicted result and the measured value is large in all cases: prediction (10 minutes after start), prediction (20 minutes after start), and prediction (50 minutes after start). On the other hand, in the embodiment of the present invention of traffic flow prediction, the difference between the predicted result and the measured value is small in all cases: prediction (10 minutes after start), prediction (20 minutes after start), and prediction (50 minutes after start).

[0181] The effects of the traffic flow prediction device 1 according to the embodiment of the present invention have been described above.

[0182] (2. Hardware Configuration Example) Next, an example of the hardware configuration of the traffic flow prediction device 1 according to an embodiment of the present invention will be described.

[0183] In the following, an example of the hardware configuration of the information processing device 900 will be described as an example of the hardware configuration of the traffic flow prediction device 1 according to an embodiment of the present invention. Note that the example of the hardware configuration of the information processing device 900 described below is merely one example of the hardware configuration of the traffic flow prediction device 1. Therefore, the hardware configuration of the traffic flow prediction device 1 may be modified by removing unnecessary components from the hardware configuration of the information processing device 900 described below, or by adding new components.

[0184] Figure 28 shows the hardware configuration of an information processing device 900 as an example of a traffic flow prediction device 1 according to an embodiment of the present invention. The information processing device 900 includes a CPU (Central Processing Unit) 901, a ROM (Read Only Memory) 902, a RAM (Random Access Memory) 903, a host bus 904, a bridge 905, an external bus 906, an interface 907, an input device 908, an output device 909, a storage device 910, and a communication device 911.

[0185] The CPU 901 functions as both an arithmetic processing unit and a control unit, controlling the overall operation of the information processing unit 900 according to various programs. The CPU 901 may also be a microprocessor. The ROM 902 stores programs and arithmetic parameters used by the CPU 901. The RAM 903 temporarily stores programs used in the execution of the CPU 901 and parameters that change as needed during its execution. These are interconnected by a host bus 904, which consists of a CPU bus and other components.

[0186] The host bus 904 is connected to an external bus 906, such as a PCI (Peripheral Component Interconnect / Interface) bus, via a bridge 905. It is not always necessary to configure the host bus 904, bridge 905, and external bus 906 separately; these functions may be implemented on a single bus.

[0187] The input device 908 consists of input means for the user to input information, such as a mouse, keyboard, touch panel, buttons, microphone, switches, and levers, and an input control circuit that generates input signals based on the user's input and outputs them to the CPU 901. The user operating the information processing device 900 can input various types of data to the information processing device 900 or instruct it to perform processing operations by operating this input device 908.

[0188] The output device 909 includes, for example, display devices such as CRT (Cathode Ray Tube) display devices, liquid crystal display (LCD) devices, OLED (Organic Light Emitting Diode) devices, lamps, and audio output devices such as speakers.

[0189] The storage device 910 is a device for storing data. The storage device 910 may include a storage medium, a recording device for recording data on the storage medium, a reading device for reading data from the storage medium, and a deletion device for deleting data recorded on the storage medium. The storage device 910 is composed of, for example, an HDD (Hard Disk Drive). This storage device 910 drives the hard disk and stores programs executed by the CPU 901 and various data.

[0190] The communication device 911 is a communication interface composed of, for example, a communication device for connecting to a network. The communication device 911 may support either wireless or wired communication.

[0191] The above describes an example of the hardware configuration of the traffic flow prediction device 1 according to an embodiment of the present invention.

[0192] (3. Summary) Although preferred embodiments of the present invention have been described in detail above with reference to the attached drawings, the present invention is not limited to these examples. It is clear to any person with ordinary skill in the art to which the present invention belongs that various modifications or alterations can be conceived within the scope of the technical idea described in the claims, and these are also understood to fall within the technical scope of the present invention. [Explanation of symbols]

[0193] 1. Traffic flow prediction device 112 Probe Antenna 114 Free-flow antenna 121 Driving history data storage unit 122 Probe data storage unit 123 Free-flow data storage unit 124 Traffic volume data storage unit 125 Traffic density data storage unit 126 Master Data Storage Unit 127 Learning parameter storage unit 128 Prediction Result Storage Unit 129 Mesh data storage unit 131 Statistics Processing Department 133 Traffic Volume Calculation Department 135 Mesh Data Generation Unit 140 Processing Unit 141 Learning Department 1411 KV Parameter Creation Section 1412 Traffic density calculation section 1413 Traffic density pattern learning unit 142 Reasoning part 1421 Traffic Volume Forecasting Department 1422 Stagnant traffic calculation department 1423 Congestion Area Prediction Section 1424 Traffic Density Prediction Section 1425 Providing Department

Claims

1. A traffic volume prediction unit identifies a range corresponding to traffic congestion data that includes a congestion time, which is the time when congestion occurs, and a congestion location, which is the location where congestion occurs at the congestion time; predicts a second inflow traffic volume that will flow into the range at a second time, which is after the first time, based on a first inflow traffic volume that flows into the range at a first time; predicts a second outflow traffic volume that will flow out of the range at a second time, based on a first outflow traffic volume that flows out of the range at a first time; and calculates a cumulative inflow traffic volume by accumulating the first inflow traffic volume and the second inflow traffic volume, and a cumulative outflow traffic volume by accumulating the first outflow traffic volume and the second outflow traffic volume. A traffic congestion volume calculation unit calculates the difference between the cumulative inflow traffic volume and the cumulative outflow traffic volume as the traffic congestion volume, A congestion area prediction unit predicts the congestion area at the second time based on the aforementioned traffic volume, A traffic density prediction unit that predicts the traffic density of the target location at the second time based on the aforementioned congested watershed, A traffic flow prediction device equipped with the following features.

2. The aforementioned congestion basin prediction unit is: The length of the traffic congestion is predicted according to the traffic volume, and the congested area at the second time is predicted based on the length of the traffic congestion. The traffic flow prediction device according to claim 1.

3. The aforementioned congestion basin prediction unit is: Based on the downstream location and the congestion length in the aforementioned congestion data, the congested watershed at the second time is predicted. The traffic flow prediction device according to claim 2.

4. The aforementioned congestion basin prediction unit is: The congestion area at the second time is predicted to extend from the lowest point in the aforementioned congestion data to a point upstream by the length of the congestion relative to the lowest point. The traffic flow prediction device according to claim 3.

5. The traffic density prediction unit obtains a congestion determination result by determining whether the target location at the second time point belongs to the congested waterway, and predicts the traffic density of the target location at the second time point based on the congestion determination result. The traffic flow prediction device according to claim 1.

6. The traffic flow prediction device includes a learning unit that stores a combination of the traffic density at a predetermined location at a third time and a previous time traffic density pattern, which is the traffic density within a range corresponding to the predetermined location at a fourth time that is prior to the third time, as a traffic density pattern. The traffic density prediction unit predicts the traffic density at the target location at the second time based on the traffic density pattern and the congestion determination result. The traffic flow prediction device according to claim 5.

7. The traffic density prediction unit predicts the traffic density of the predetermined location at the third time as the traffic density of the target location at the second time if the target location at the second time is located in the congested area, the predetermined location is located in the congested area at the third time, and the traffic density pattern of the previous time, which is the traffic density of the range corresponding to the target location at the first time, is similar to the traffic density pattern of the previous time, which is the traffic density of the range corresponding to the predetermined location at the fourth time. The traffic flow prediction device according to claim 6.

8. The traffic density prediction unit predicts the traffic density of the predetermined location at the third time as the traffic density of the target location at the second time if the target location at the second time does not belong to the congested area, the predetermined location at the third time is not a congested location, and the traffic density pattern at the previous time, which is the traffic density for the range corresponding to the target location at the first time, is similar to the traffic density pattern at the previous time, which is the traffic density for the range corresponding to the predetermined location at the fourth time. The traffic flow prediction device according to claim 6.

9. The traffic volume prediction unit identifies the range as the area from a vehicle detection position located downstream of the lowest downstream position in the congestion data to a vehicle detection position located upstream of the highest upstream position in the congestion data. The traffic flow prediction device according to claim 1.

10. The traffic volume prediction unit determines, based on the vehicle speed statistics, whether the statistics are below a threshold, and if it determines that the statistics are below the threshold, it extracts the position corresponding to the vehicle speed as the congestion location and extracts the time corresponding to the vehicle speed as the congestion time. The traffic flow prediction device according to claim 1.

11. A traffic volume prediction unit identifies a range corresponding to traffic congestion data that includes a congestion time, which is the time when congestion occurs, and a congestion location, which is the location where congestion occurs at the congestion time; predicts a second inflow traffic volume that will flow into the range at a second time, which is after the first time, based on a first inflow traffic volume that flows into the range at a first time; predicts a second outflow traffic volume that will flow out of the range at a second time, based on a first outflow traffic volume that flows out of the range at a first time; and calculates a cumulative inflow traffic volume by accumulating the first inflow traffic volume and the second inflow traffic volume, and a cumulative outflow traffic volume by accumulating the first outflow traffic volume and the second outflow traffic volume. A traffic congestion volume calculation unit calculates the difference between the cumulative inflow traffic volume and the cumulative outflow traffic volume as the traffic congestion volume, A congestion area prediction unit predicts the congestion area at the second time based on the aforementioned traffic volume, A traffic density prediction unit that predicts the traffic density of the target location at the second time based on the aforementioned congested watershed, A computer-based traffic flow prediction method comprising the following features.

12. Computers, A traffic volume prediction unit identifies a range corresponding to traffic congestion data that includes a congestion time, which is the time when congestion occurs, and a congestion location, which is the location where congestion occurs at the congestion time; predicts a second inflow traffic volume that will flow into the range at a second time, which is after the first time, based on a first inflow traffic volume that flows into the range at a first time; predicts a second outflow traffic volume that will flow out of the range at a second time, based on a first outflow traffic volume that flows out of the range at a first time; and calculates a cumulative inflow traffic volume by accumulating the first inflow traffic volume and the second inflow traffic volume, and a cumulative outflow traffic volume by accumulating the first outflow traffic volume and the second outflow traffic volume. A traffic congestion volume calculation unit calculates the difference between the cumulative inflow traffic volume and the cumulative outflow traffic volume as the traffic congestion volume, A congestion area prediction unit predicts the congestion area at the second time based on the aforementioned traffic volume, A traffic density prediction unit that predicts the traffic density of the target location at the second time based on the aforementioned congested watershed, A program that makes it function as such.

Images