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

Abstract

According to the present invention, an acquisition unit (110) acquires speed information and position information of a moving body, and a prediction unit (130) predicts a cause of traffic congestion in each section on the basis of temporal change in the speed information for each section identified by the position information.

Description

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

[0001] The disclosed technology relates to a traffic congestion prediction device, a traffic congestion prediction method, and a traffic congestion prediction program.

[0002] In recent years, changes have been occurring in traffic congestion, a social issue. Traffic demand has changed significantly due to various factors, including the COVID-19 pandemic, the opening of large commercial facilities, travel restrictions due to large-scale events, and dynamic pricing. As a result, the volume of vehicles and people traveling in various locations has been affected, resulting in problems such as traffic congestion and congestion occurring in previously unanticipated locations. In addition, traffic congestion and congestion caused by traffic concentration, traffic accidents, broken-down vehicles, etc. continue to occur on roads nationwide.

[0003] In response to these problems, vehicle drivers have traditionally responded by making behavioral changes, such as changing departure times or taking detours, to avoid sections experiencing congestion or congestion, based on existing congestion information services.

[0004] Another approach that drivers can take is to use a congestion prediction service to identify sections and time periods where congestion or congestion is likely to occur, and then consider changing their behavior to avoid those sections and time periods.

[0005] In addition to changes in behavior based on the driver's own judgment, information provided by a traffic congestion information service and a traffic congestion prediction service may be output to a car navigation system, causing the car navigation system to search for a route that bypasses a congested area.

[0006] Japan Road Traffic Information Center (JARTIC), [online], [searched November 8, 2023], Internet <URL: https: / / www.jartic.or.jp> "Traffic Congestion Forecast Calendar," NEXCO Central Japan, [online], [searched November 8, 2023], Internet <URL: https: / / dc.c-nexco.co.jp / jam / cal / >

[0007] However, there is a time lag between the occurrence of a traffic jam and its provision to the user when the traffic jam information service provides the information on the traffic jam section (the coordinate set of the start and end points of the traffic jam), so it is assumed that the situation will have changed by the time the user actually arrives at the target section. This is because the traffic jam or congestion information provided by the traffic jam information service is calculated by statistical processing based on location information from in-vehicle devices, the driver's smartphone, and cross-sectional traffic volume and vehicle speed measured by cameras or sensors installed in various locations. In addition, on expressways, a traffic jam is defined as "a state in which a line of vehicles traveling at a low speed of 40 km / h or less or repeatedly stopping and starting continues for 1 km or more and for 15 minutes or more." The fact that the duration of the traffic jam is taken into consideration is also a factor in the time lag before the information is provided.

[0008] Furthermore, as mentioned above, the causes of congestion or jams in the jammed section information are diverse, and in cases where the information is planned and publicized in advance, such as a reduction in the number of lanes due to construction, the information is provided before the congestion or jam occurs, and users can recognize it early. In contrast, in cases where the congestion or jams tend to occur periodically, such as morning and evening rush hours or drive-through congestion or jams at lunchtime, drivers must make their own judgment, such as "this is a regular traffic jam."

[0009] In addition, for unexpected incidents such as traffic accidents and broken-down vehicles, which are impossible to predict when and where they will occur, the information on the section where the accident occurred mentions the cause, such as "accident vehicle." However, because this information is based on reports from individual drivers and visual confirmation by the police, there is an even greater time lag before the cause of the congestion is posted in the information on the section where the accident occurred.

[0010] Furthermore, if an accident or other incident occurs at a time and place where periodic congestion tends to occur, even if the congestion is actually due to a sudden factor such as an accident, the driver may mistakenly believe that "this is just a regular traffic congestion" due to the above-mentioned time lag. In this case, it may be impossible to respond by making the correct behavioral changes.

[0011] Furthermore, the traffic congestion forecast information provided by traffic congestion forecasting services is currently limited to predicting periodic traffic congestion, such as weekday morning and evening rush hours, highway congestion during long holidays, etc. Therefore, it is difficult to predict the occurrence of traffic congestion caused by sudden events such as those mentioned above, where it is impossible to predict when and where it will occur.

[0012] One method for recognizing the real-time occurrence of traffic congestion and its causes is to use fixed cameras to capture images of the relevant sections. However, this method is difficult to use to comprehensively recognize all expressways and ordinary roads nationwide.

[0013] Furthermore, as a means of detecting the occurrence of congestion more precisely and quickly, there are also methods that use probe information such as location information and vehicle speed collected from vehicles traveling in various locations in addition to cross-sectional traffic volume and its passing speed measured by fixed sensors in various locations. Attempts are being made to use these methods to more accurately and quickly estimate the occurrence of congestion and its sections. However, as mentioned above, since the definition of congestion includes duration, there is still a time lag before the information is shared with drivers.

[0014] The disclosed technology has been made in consideration of the above points, and aims to predict the cause of congestion early.

[0015] A first aspect of the present disclosure is a traffic congestion prediction device that includes an acquisition unit that acquires speed information and position information of a moving body, and a prediction unit that predicts the cause of traffic congestion in a section based on changes in the speed information over time for each section identified by the position information.

[0016] A second aspect of the present disclosure is a congestion prediction method, in which an acquisition unit acquires speed information and position information of a moving body, and a prediction unit predicts the cause of congestion in the section based on the change in the speed information over time for each section identified by the position information.

[0017] A third aspect of the present disclosure is a traffic congestion prediction program that causes a computer to function as each part of the traffic congestion prediction device.

[0018] According to the disclosed technology, the cause of congestion can be predicted early.

[0019] 1 is a block diagram showing a hardware configuration of a traffic congestion prediction device; FIG. 2 is a functional block diagram of a traffic congestion prediction device according to a first embodiment; FIG. 3 is a diagram showing an example of a section information DB; FIG. 4 is a diagram showing an example of an image of information stored in the section information DB; FIG. 5 is a diagram for explaining detection of the occurrence of traffic congestion; FIG. 6 is a diagram showing an example of transition of traffic congestion caused by traffic concentration; FIG. 7 is a diagram showing an example of transition of traffic congestion caused by an accident or a broken-down vehicle; FIG. 8 is a diagram showing a more specific example of transition of traffic congestion caused by an accident or a broken-down vehicle; FIG. 9 is a diagram showing a process for detecting the occurrence of traffic congestion caused by traffic concentration; FIG. 10 is a diagram showing a process for detecting the occurrence of traffic congestion caused by an accident or a broken-down vehicle; FIG. 11 is a diagram showing a comparison result of extension trends; FIG. 12 is a diagram for explaining recovery of traffic capacity; FIG. 13 is a diagram showing a comparative example of recovery trends of average speeds; FIG. 14 is a flowchart showing an example of traffic congestion prediction processing according to the first embodiment; FIG. 15 is a functional block diagram of a traffic congestion prediction device according to a second embodiment; FIG. 16 is a flowchart showing an example of traffic congestion prediction processing according to the second embodiment; FIG. 17 is a functional block diagram of a traffic congestion prediction device according to a third embodiment; FIG. 18 is a flowchart showing an example of traffic congestion prediction processing according to the third embodiment; FIG. 19 is a functional block diagram of a traffic congestion prediction device according to a fourth embodiment; FIG. 19 is a diagram for explaining detection of the resolution of traffic congestion; FIG. 19 is a flowchart showing an example of traffic congestion prediction processing according to the fourth embodiment; FIG. 19 is a diagram for explaining the occurrence of traffic congestion on a road including a branch section; FIG. 19 is a diagram for explaining another example of acquisition of speed information.

[0020] An example of an embodiment of the disclosed technology will be described below with reference to the drawings. In each of the following embodiments, a case where the "mobile body" of the present disclosure is a vehicle will be described as an example. Note that the same reference numerals are used to designate identical or equivalent components and parts in each drawing. Also, the dimensional proportions in the drawings have been exaggerated for the sake of explanation and may differ from the actual proportions.

[0021] 1 is a block diagram showing the hardware configuration of a traffic congestion prediction device 100 according to a first embodiment. As shown in Fig. 1, the traffic congestion prediction device 100 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage 14, an input unit 15, a display unit 16, and a communication I / F (Interface) 17. Each component is connected to each other via a bus 19 so as to be able to communicate with each other.

[0022] The CPU 11 is a central processing unit that executes various programs and controls each part. That is, the CPU 11 reads programs from the ROM 12 or the storage 14 and executes the programs using the RAM 13 as a work area. The CPU 11 controls the above components and performs various arithmetic processing in accordance with the programs stored in the ROM 12 or the storage 14. In this embodiment, the ROM 12 or the storage 14 stores a traffic congestion prediction program, which will be described later.

[0023] The ROM 12 stores various programs and various data. The RAM 13 temporarily stores programs or data as a working area. The storage 14 is configured by a storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive), and stores various programs including an operating system and various data.

[0024] The input unit 15 includes a pointing device such as a mouse and a keyboard, and is used to input various types of information. The display unit 16 is, for example, a liquid crystal display, and displays various types of information. The display unit 16 may also function as the input unit 15 by employing a touch panel system.

[0025] The communication I / F 17 is an interface for communicating with other devices such as an in-vehicle device, etc. For this communication, a wireless communication standard such as 4G, 5G, or Wi-Fi (registered trademark) is used.

[0026] Next, the functional configuration of the traffic congestion prediction device 100 will be described. FIG. 2 is a block diagram showing an example of the functional configuration of the traffic congestion prediction device 100. As shown in FIG. 2, the traffic congestion prediction device 100 includes, as its functional configuration, an acquisition unit 110, a calculation unit 120, a prediction unit 130, and an output unit 140. The prediction unit 130 further includes an occurrence detection unit 131 and a cause prediction unit 132. In addition, a predetermined storage area of ​​the traffic congestion prediction device 100 stores an acquired information DB (DataBase) 101, a road NW (Network) 102, and a section information DB 103. Each functional configuration is realized when the CPU 11 reads out a traffic congestion prediction program stored in the ROM 12 or storage 14, expands it into the RAM 13, and executes it.

[0027] The acquisition unit 110 acquires speed information and position information of each vehicle from each of a plurality of vehicles. The speed information and position information are measured, for example, by a speed sensor, a GPS (Global Positioning System), or the like in the vehicle, and are collected from CAN data by an OBD2 or the like and transmitted from the in-vehicle device at predetermined time intervals (for example, every second). The acquisition unit 110 acquires the speed information and position information of each vehicle transmitted from the in-vehicle device of that vehicle. The acquisition unit 110 stores the acquired information in the acquired information DB 101.

[0028] The road network 102 is network data that represents a road network using links and nodes, and contains attribute information about the roads, such as the latitude and longitude of the road, the road type (e.g., expressway, national highway, or prefectural road), the number of lanes, the presence or absence of shoulders, and the legal speed limit.

[0029] The calculation unit 120 calculates an average speed by averaging the speed information of each of multiple vehicles for each section identified by the position information and for each time period of a predetermined time, based on the speed information and position information stored in the acquired information DB 101. Specifically, the calculation unit 120 uses the road network 102 to divide each road into sections of a predetermined distance (e.g., 125 m) and assigns a section ID, which is identification information for the section, to each section. The calculation unit 120 calculates the average value of the speed information stored in the acquired information DB 101 for each section and for each time period (e.g., 5-minute time periods), and sets this average value as the average speed.

[0030] The calculation unit 120 associates the calculated average speed with the section ID of the corresponding section and the time period, and stores the average speed in the section information DB 103. FIG. 3 shows an example of the section information DB 103. In the example of FIG. 3, the time period "0:00 to 0:05" is represented as "0:00". FIG. 4 shows an example of an image of the information stored in the section information DB 103. In the example of FIG. 4, blocks indicating the sections are superimposed on a map, and the color (shade) of each block varies depending on the average speed of the section corresponding to that block.

[0031] The occurrence detection unit 131 uses information stored in the section information DB 103 to detect that congestion has occurred in a section and in a time period when the average speed for each section and each time period is less than a predetermined first threshold (e.g., 40 km / h). Hereinafter, a section where congestion has been detected is referred to as a "congested section." The occurrence detection unit 131 also identifies multiple consecutive congested sections as congested sections. In this case, if there are a certain number of non-congested sections (e.g., one section) between two congested sections, the occurrence detection unit 131 may identify the consecutive congested sections, including the certain number of sections, as congested sections. Such a certain number of sections is considered to be due to a temporary gap, and the congestion is considered to be continuing.

[0032] For example, as shown in Figure 5, in the time slot of 14:55, there is no section where the average speed is less than the first threshold, and congestion is not detected. In the time slot of 15:00, a section where the average speed is less than the first threshold occurs, and this section is detected as a congested section. Then, in the time slot of 15:05, a section with multiple consecutive congested sections is identified as a congested section (the dashed line in Figure 5).

[0033] The occurrence detection unit 131 passes to the cause prediction unit 132 information about the identified congested section, including information such as the time when the congested section was identified, i.e., the time when the congestion occurred, the congestion length which is the length of the congested section, and the average speed of each section within the congested section during the time period when the congestion occurs.

[0034] The cause prediction unit 132 predicts the cause of congestion in a congested section based on the time change of the speed information. Here, traffic congestion and accidents or broken-down vehicles are considered as causes of congestion.

[0035] Figure 6 shows an example of the transition of congestion caused by traffic concentration. Under normal circumstances, the ratio of inflow and outflow volumes for a section of road is the same, and this section can handle all the inflowing vehicles without congestion. As the inflow volume increases, it begins to put pressure on the road's traffic capacity, making it difficult to maintain a safe distance between vehicles, gradually reducing the average speed and causing congestion. Note that traffic capacity is the number of vehicles that can travel on that road per unit time. Furthermore, if the state in which the inflowing vehicles cannot be handled continues for a long time, the average speed will fall below the first threshold, and the congested section will extend further back.

[0036] Figure 7 shows an example of the transition of congestion caused by an accident or a broken-down vehicle. When an accident or a broken-down vehicle occurs in a certain section, even if the inflow volume to that section does not change, the accident makes the lane unusable, reducing traffic capacity, and congestion occurs because the inflow vehicles cannot be handled. Then, when the accident vehicle or broken-down vehicle is removed and traffic capacity is restored, the inflow vehicles can be handled, and the congestion is resolved over time. Note that even if an accident or broken-down vehicle occurs, there may be cases where traffic capacity does not decrease and congestion does not occur, for example, if the vehicle is quickly evacuated to the shoulder on its own.

[0037] Figure 8 shows a more specific example of the transition of traffic congestion caused by an accident or broken-down vehicle. (1) When an accident occurs, (2) average speed decreases and traffic congestion occurs, and (3) over time, the traffic congestion extends backward. (4) While the traffic congestion continues to extend backward, when the cause of the traffic congestion is resolved, for example by removing the accident vehicle, (5) average speed improves near the front of the traffic congestion, and (6) the front of the congested section moves near the end, and the traffic congestion begins to resolve.

[0038] FIG. 9 shows the process of detecting the occurrence of a traffic jam caused by traffic concentration. In the case of a traffic jam caused by traffic concentration, vehicles gradually decrease in speed and the distance between vehicles decreases as a result of concentrating on the road. Then, in the section indicated by A in FIG. 9, the average speed falls below the first threshold, and a traffic jam is detected. Note that, under the general definition of a traffic jam on a highway, if this state continues for a certain period of time, it is recognized as a traffic jam. Furthermore, in a traffic jam caused by traffic concentration, the coordinates of the beginning of the traffic jam section may fluctuate.

[0039] FIG. 10 shows the process of detecting a traffic jam caused by an accident or a broken-down vehicle. In the case of a traffic jam caused by an accident or a broken-down vehicle, the moment the accident occurs, the following vehicles take measures such as braking, causing a sudden drop in vehicle speed and a decrease in the distance between vehicles. Then, in the section indicated by A in FIG. 10, the average speed falls below the first threshold, and a traffic jam is detected. As with the above, under the general definition of a traffic jam on a highway, a traffic jam is recognized if this state continues for a certain period of time. Furthermore, in a traffic jam caused by an accident or a broken-down vehicle, the coordinates of the beginning of the traffic jam section do not move much from around the point where the accident occurred.

[0040] As described above, the change in average speed over time differs depending on the cause of the congestion. Therefore, the cause prediction unit 132 predicts the cause of the congestion from the change in average speed over time. Specifically, the cause prediction unit 132 predicts the cause of the congestion based on at least one of the following information regarding the congested section notified by the occurrence detection unit 131: (1) Initial congestion length at the time of congestion occurrence (2) Change in congestion length after the congestion occurrence (tendency to extend) (3) Average speed after the congestion occurrence (4) Recovery tendency of the average speed in the section ahead of the beginning of the congested section (5) Coordinates of the beginning of the congested section

[0041] Regarding (1) and (2) above, in the case of congestion caused by traffic congestion, there may be a wide section where the average speed drops to the level where congestion is just detected (the distance between vehicles is close), and the initial congestion length tends to be long. On the other hand, in the case of congestion caused by an accident or a broken-down vehicle, the following vehicle suddenly reduces its speed by braking for safety reasons, and this causes a chain reaction that further affects the following vehicles, so the initial congestion length is shorter than in the case of traffic congestion, but the amount of congestion length tends to increase.

[0042] Figure 11 (columns 3 to 5) shows the results of a comparison of the lengthening trends confirmed using actual data between congestion caused by traffic concentration and congestion caused by accidents or broken-down vehicles. The target data was traffic congestion data on expressways, and confirmed the lengthening trends for 10 minutes from the start of congestion for congestion caused by accidents or broken-down vehicles and congestion caused by traffic concentration. As shown in Figure 11, at the initial stage (0 minutes after the start of congestion), the length of congestion caused by traffic concentration is longer. Furthermore, when comparing the length of congestion 5 and 10 minutes later, the length of congestion caused by accidents or broken-down vehicles is greater.

[0043] Regarding (3) above, in the case of congestion caused by traffic concentration, the average speed initially drops only to near the threshold at which the road is judged to be congested or crowded. On the other hand, in the case of congestion caused by an accident or a broken-down vehicle, braking occurs in a chain reaction, causing the average speed to drop sharply below the threshold. In other words, when comparing the average speeds in the same section and the sections before and after the congestion just before and after the congestion occurs, the average speed tends to drop sharply in the case of congestion caused by an accident or a broken-down vehicle.

[0044] Figure 11 (columns 6-8) shows the results of a comparison of average speeds over 10 minutes from the start of congestion caused by accidents or broken-down vehicles and congestion caused by traffic congestion, using the same actual data as above. As shown in Figure 11, at the initial stage (0 minutes after the start of congestion), the average speed drops significantly in congestion caused by accidents or broken-down vehicles. Furthermore, when comparing the average speeds 5 and 10 minutes later, the drop in average speed is greater in congestion caused by accidents or broken-down vehicles.

[0045] Regarding (4) above, in the case of a traffic jam caused by traffic concentration, even if the driver has passed the front of the traffic jam, it tends to take a certain distance for the driver's speed to return to the legal speed limit in the next section. This is because the number of vehicles entering the jam is extremely large, and the traffic capacity remains the same as normal, and there are no clear features that cause the jam, making it difficult for the driver to recognize that he or she has passed the jam. On the other hand, in the case of a traffic jam caused by an accident or a broken-down vehicle, as shown in Figure 12, the previously unavailable lanes become usable and traffic capacity is restored. Furthermore, because the driver can clearly recognize that "this is the front and cause of the jam," the driver tends to recover speed quickly and all at once.

[0046] Figure 13 shows a comparative example of the recovery trend of average speed. Figure 13A shows an example of traffic congestion caused by an accident or a broken-down vehicle. In this case, before the congested section where the average vehicle speed is less than 40 km / h, there are only two sections where the average speed is between 40 km / h and 80 km / h, and the average speed quickly recovers to 80 km / h. On the other hand, Figure 13B shows an example of traffic congestion caused by traffic concentration. In this case, before the congested section, there is a long stretch of sections where the average speed is between 40 km / h and 80 km / h.

[0047] Regarding (5) above, there are cases where the head of a vehicle is almost fixed, such as at an intersection with a main road, but also cases where the head coordinates change, such as when some vehicles unintentionally slow down due to a sag.

[0048] Based on the above-mentioned trends, the cause prediction unit 132 sets a threshold value for each cause for each of (1) to (5) above, and predicts the cause by comparing the congestion section information passed from the occurrence detection unit 131 with the threshold value.

[0049] For example, the cause prediction unit 132 (1) predicts that the initial length of a traffic jam at the time of the occurrence of the traffic jam is the result of traffic congestion if the initial length of the traffic jam is equal to or greater than a threshold value THA, and that the initial length of the traffic jam is the result of an accident or a broken-down vehicle if the initial length of the traffic jam is less than THA.

[0050] Furthermore, for example, the cause prediction unit 132 (2) calculates the amount of extension from the change in congestion length over time regarding the change in congestion length after the occurrence of congestion, and predicts that if the amount of extension is equal to or greater than the threshold value THB, the congestion is caused by an accident or a broken-down vehicle, and if it is less than THB, the congestion is caused by traffic concentration.

[0051] Furthermore, for example, the cause prediction unit 132 calculates the average value of the average speed of each section included in the congested section after the occurrence of congestion (3) for the average speed after the occurrence of congestion. Then, if the average value of the average speed is equal to or greater than the threshold value THC, the cause prediction unit 132 predicts that the congestion is caused by traffic congestion, and if it is less than THC, the congestion is caused by an accident or a broken-down vehicle.

[0052] Furthermore, for example, (4) regarding the recovery tendency of the average speed in sections ahead of the beginning of the congested section, the cause prediction unit 132 counts the number of consecutive sections ahead of the congested section where the average speed is within a predetermined range (for example, 40 km / h to 80 km / h).The cause prediction unit 132 then predicts that if the number of sections is equal to or greater than the threshold THD, the congestion is due to traffic congestion, and if it is less than THD, the congestion is due to an accident or a broken-down vehicle.

[0053] Furthermore, for example, (5) the cause prediction unit 132 calculates the amount of change in the leading coordinate of the congested section from the time change of the leading coordinate of the congested section since the onset of the congestion. If the amount of change in the leading coordinate is equal to or greater than a threshold THE, the cause prediction unit 132 predicts that the congestion is caused by traffic congestion, and if it is less than THE, the congestion is caused by an accident or a broken-down vehicle.

[0054] The cause prediction unit 132 may predict the cause using any one of the above (1) to (5), or may predict the cause using a combination of two or more of (1) to (5). The prediction method is not limited to the above examples, as long as it reflects differences in the characteristics of the time change in average speed due to the cause of the congestion. For example, instead of determining the cause using a binary value of "above or below the threshold," the cause may be predicted using a probability, and an output may be given such as a 20% probability that the congestion is due to traffic concentration, and an 80% probability that the congestion is due to an accident or a broken-down vehicle.

[0055] Furthermore, since road attributes vary depending on the location, such as traffic volume, number of lanes, whether damage is likely to be severe in the event of an accident, road type, legal speed limit, and the presence or absence of a shoulder, the above-mentioned trends may also differ depending on the location. Therefore, each threshold may be set according to the characteristics of the location. For example, on local roads, average speed decreases when vehicles stop at traffic lights or when waiting to turn right or left, so each threshold may be changed taking into account the ratio between the green and red light periods of traffic lights installed in the direction of travel.

[0056] Furthermore, depending on the location or time of day, traffic congestion caused by traffic concentration and traffic congestion caused by accidents or broken-down vehicles may show similar trends. For example, on roads with shoulders, broken-down vehicles may be able to coast for a while and may voluntarily move to the shoulder to avoid blocking a lane. The change in average speed over time in this case may be similar to traffic congestion caused by traffic concentration. Therefore, thresholds may be set by learning past trends in various units, such as by section or road link, depending on the presence or absence of detours, the presence or absence of shoulders, the legal speed limit, the number of lanes, etc.

[0057] The cause prediction unit 132 passes the congested section information passed from the occurrence detection unit 131 and the congestion information including the predicted cause to the output unit 140 .

[0058] The output unit 140 outputs the congestion information received from the cause prediction unit 132 to an information processing device used by a user of the vehicle, such as a car navigation system installed in the vehicle or an information processing terminal such as a smartphone held by the user. This makes it possible to provide the user with the congestion information via a web application on the information processing terminal.

[0059] Next, the operation of the traffic congestion prediction device 100 will be described. Fig. 14 is a flowchart showing the flow of traffic congestion prediction processing by the traffic congestion prediction device 100. The CPU 11 reads a traffic congestion prediction program from the ROM 12 or the storage 14, expands it into the RAM 13, and executes it, thereby performing the traffic congestion prediction processing. Note that the traffic congestion prediction processing is an example of a traffic congestion prediction method disclosed herein.

[0060] In step S101 , the CPU 11 functions as the acquisition unit 110 to acquire speed information and position information of each of a plurality of vehicles from each of the vehicles, and stores the acquired information in the acquired information DB 101 .

[0061] Next, in step S102, the CPU 11, as the calculation unit 120, calculates an average speed by averaging the speed information of each of the multiple vehicles for each section and time period specified by the position information, based on the speed information and position information stored in the acquired information DB 101. The CPU 11, as the calculation unit 120, stores the calculated average speed in the section information DB 103 in association with the section ID of the corresponding section and the time period.

[0062] Next, in step S103, the CPU 11, as the occurrence detection unit 131, detects a congested section where the average speed for each section and for each time period is less than a predetermined first threshold, using information stored in the section information DB 103. Furthermore, the CPU 11, as the occurrence detection unit 131, identifies a plurality of consecutive congested sections as a congestion section.

[0063] Next, in step S104, the CPU 11, functioning as the occurrence detection unit 131, determines whether or not a congested section has been identified in step S103, thereby determining whether or not the occurrence of congestion has been detected. If the occurrence of congestion has been detected, the CPU 11, functioning as the occurrence detection unit 131, passes the congested section information to the cause prediction unit 132 and proceeds to step S105, but if congestion has not been detected, returns to step S101.

[0064] In step S105, the CPU 11, functioning as the cause prediction unit 132, predicts the cause of congestion in the congested section based on the change in average speed over time. Next, in step S106, the CPU 11, functioning as the cause prediction unit 132, passes the congested section information and congestion information including the predicted cause to the output unit 140. Then, the CPU 11, functioning as the output unit 140, outputs the congestion information passed from the cause prediction unit 132 to an information processing device used by the user of the vehicle, and the process returns to step S101.

[0065] As described above, the traffic congestion prediction device according to the first embodiment acquires vehicle speed information and location information, and predicts the cause of traffic congestion in a section based on the time variation of the speed information for each section identified by the location information. In this way, the cause of traffic congestion can be predicted early, without taking into account the duration of the traffic congestion, by focusing on the fact that the characteristics of the time variation of the average vehicle speed differ depending on the cause of the traffic congestion. This also allows the user to more quickly recognize the cause of the traffic congestion and determine the need for behavioral changes early.

[0066] Specifically, according to this embodiment, by detecting when a traffic jam or congestion is "starting" or "just starting," the user can recognize the occurrence of the traffic jam or congestion earlier than conventional methods. Furthermore, the cause of the traffic jam or congestion can be predicted from the time change in the average speed in the section where the traffic jam or congestion is "starting" or "just starting." Furthermore, by taking the predicted cause into account, the user can infer the time required for the traffic jam or congestion to clear. For example, suppose that a user normally tends to infer that a traffic jam that usually occurs during the morning and evening rush hour will clear around 9:00 a.m. when the traffic congestion clears. Here, if information is provided that the cause of the traffic jam is due to an accident or a broken-down vehicle, the user can infer that the traffic jam will take longer than usual to clear and may continue until around 10:00 a.m.

[0067] Second Embodiment Next, a second embodiment will be described. In the traffic congestion prediction device according to the second embodiment, components similar to those in the traffic congestion prediction device 100 according to the first embodiment will be assigned the same reference numerals, and descriptions thereof will be omitted. Furthermore, components that share some functions with functional units in each embodiment will be assigned reference numerals with the same last two digits, and detailed descriptions thereof will be omitted.

[0068] Fig. 15 is a block diagram showing an example of the functional configuration of the traffic congestion prediction device 200. As shown in Fig. 15, the traffic congestion prediction device 200 includes, as its functional configuration, an acquisition unit 210, a calculation unit 220, a prediction unit 230, and an output unit 140. The prediction unit 230 further includes an occurrence detection unit 131 and a cause prediction unit 232. In addition, an acquisition information DB 201, a road NW 102, and a section information DB 203 are stored in a predetermined storage area of ​​the traffic congestion prediction device 200. Each functional configuration is realized by the CPU 11 reading out a traffic congestion prediction program stored in the ROM 12 or the storage 14, expanding it in the RAM 13, and executing it.

[0069] The acquisition unit 210 acquires operation information in addition to the speed information and position information of each of the multiple vehicles acquired by the acquisition unit 110 of the first embodiment. The operation information is information indicating the amount of operation related to a change in the speed of the vehicle. For example, the acquisition unit 210 may acquire, as the operation information, the accelerator amount, the number of brakes, the brake amount, the hazard light illumination rate, etc. The accelerator amount is an amount that indicates how much the accelerator is depressed, and similarly, the brake amount is an amount that indicates how much the brake is depressed. The acquisition unit 210 stores the acquired speed information, position information, and operation information in the acquired information DB 201.

[0070] The acquired information DB 201 differs from the acquired information DB 101 of the first embodiment in that the acquired information DB 201 stores the above-mentioned operation information in addition to the speed information and position information of each of a plurality of vehicles.

[0071] Similar to the calculation unit 120 of the first embodiment, the calculation unit 220 calculates an average speed for each section and time period specified by the location information based on the speed information and location information stored in the acquired information DB 201. Similarly, the calculation unit 220 calculates an average value of the operation amount (hereinafter referred to as the "average operation amount") for each section and time period for the operation information. The calculation unit 220 stores the calculated average speed and the average operation amount for each piece of operation information in the section information DB 203.

[0072] Furthermore, the calculation unit 220 calculates the reliability of the average speed and the average operation amount. Specifically, the calculation unit 220 may calculate the number of vehicles used when calculating the average speed and the average operation amount as the reliability. For example, the calculation unit 220 may determine that the reliability is low when the number of vehicles used in calculating the average is less than three, and that the reliability is high when the number of vehicles used is three or more.

[0073] Furthermore, if the reliability of the average speed and the average operation amount calculated for the target section is lower than a predetermined second threshold, the calculation unit 220 corrects the calculated average speed and the average operation amount. For example, the calculation unit 220 sets the average of the average speeds and the average operation amounts of the target section and the sections before and after the target section as the average speed and the average operation amount of the target section, respectively.

[0074] The section information DB 203 differs from the section information DB 103 of the first embodiment in that it stores the average operation amount of each piece of operation information in addition to the speed information and position information of each of the plurality of vehicles.

[0075] The cause prediction unit 232 predicts the cause of congestion based on the change over time in the average speed in the congested section and the sections before and after the congested section, similar to the cause prediction unit 132 in the first embodiment. In addition, the cause prediction unit 232 also makes predictions using the average operation amount.

[0076] For example, the number of braking times in the entire congested section tends to be higher in congestion caused by accidents or broken-down vehicles. Also, the rate at which hazard lights are turned on tends to be higher at the end of the congested section compared to other times around the beginning or middle of the congested section, and this tendency tends to be higher in congestion caused by traffic concentration. Based on these trends, the cause prediction unit 232 sets a threshold value for each cause for the average operation amount of each piece of operation information, and predicts the cause by comparing the congested section information passed from the occurrence detection unit 131 with the threshold value.

[0077] The hardware configuration of the traffic congestion prediction device 200 according to the second embodiment is similar to the hardware configuration of the traffic congestion prediction device 100 according to the first embodiment shown in FIG. 1, and therefore a description thereof will be omitted.

[0078] Next, a description will be given of the operation of the traffic congestion prediction device 200. Fig. 16 is a flowchart showing the flow of traffic congestion prediction processing by the traffic congestion prediction device 200. The CPU 11 reads a traffic congestion prediction program from the ROM 12 or the storage 14, expands it into the RAM 13, and executes it, thereby performing the traffic congestion prediction processing.

[0079] In step S201 , the CPU 11 functions as the acquisition unit 210 to acquire speed information, position information, and operation information of each vehicle from each of a plurality of vehicles, and stores the acquired information in the acquired information DB 201 .

[0080] Next, in step S202, the CPU 11, as the calculation unit 220, calculates the average speed and the average operation amount for each section and for each time period based on the information stored in the acquired information DB 201. Furthermore, the CPU 11, as the calculation unit 220, stores the calculated average speed and average operation amount in the section information DB 103 in association with the section ID of the corresponding section and the time period.

[0081] Next, in step S203, the CPU 11, functioning as the calculation unit 220, calculates the number of vehicles used when calculating the average speed and the average operation amount as the reliability. Then, the CPU 11, functioning as the calculation unit 220, corrects the average speed and the average operation amount whose reliability is lower than a predetermined second threshold value by averaging them with the average speeds and average operation amounts of the preceding and following sections, for example.

[0082] Next, in step S204, the CPU 11, functioning as the occurrence detection unit 131, detects the occurrence of congestion using information stored in the section information DB 203. Next, in step S205, the CPU 11, functioning as the occurrence detection unit 131, determines whether or not the occurrence of congestion was detected in step S204. If the occurrence of congestion is detected, the CPU 11, functioning as the occurrence detection unit 131, notifies the cause prediction unit 232 of the congested section information and proceeds to step S205; if the occurrence of congestion is not detected, the process returns to step S201.

[0083] In step S206, the CPU 11, as the cause prediction unit 232, predicts the cause of congestion in the congested section based on the time change of the average speed and the average operation amount. Next, in step S207, the CPU 11, as the output unit 140, outputs congestion information including the congested section information and the predicted cause to the information processing device used by the user in the vehicle, and then returns to step S201.

[0084] As described above, the traffic congestion prediction device according to the second embodiment acquires operation information in addition to vehicle speed information and position information, and predicts the cause of traffic congestion using the average operation amount for each section as well as the change in average speed over time for each section, thereby enabling more accurate prediction of the cause of traffic congestion.

[0085] <Third embodiment> Next, a third embodiment will be described. In the traffic congestion prediction device according to the third embodiment, components similar to those of the traffic congestion prediction device 100 according to the first embodiment or the traffic congestion prediction device 200 according to the second embodiment will be assigned the same reference numerals, and descriptions thereof will be omitted. Furthermore, components having some functions in common with functional units in each embodiment will be assigned reference numerals having the same last two digits, and detailed descriptions thereof will be omitted.

[0086] Fig. 17 is a block diagram showing an example of the functional configuration of the traffic congestion prediction device 300. As shown in Fig. 17, the traffic congestion prediction device 300 includes, as its functional configuration, an acquisition unit 210, a calculation unit 220, a prediction unit 330, and an output unit 140. The prediction unit 330 further includes an occurrence detection unit 331 and a cause prediction unit 332. In addition, an acquisition information DB 201, a road NW 102, and a section information DB 203 are stored in a predetermined storage area of ​​the traffic congestion prediction device 300. Each functional configuration is realized when the CPU 11 reads out a traffic congestion prediction program stored in the ROM 12 or the storage 14, expands it in the RAM 13, and executes it.

[0087] The traffic congestion prediction device 300 is also connected to a machine learning device 600 via a network. The machine learning device 600 includes a learning unit 601 as a functional configuration.

[0088] The learning unit 601 assigns a correct answer label to information collected in advance and stored in the section information DB 203, i.e., the average speed and average operation amount for each section and each time period, indicating whether the section is included in a congested section and, if included in a congested section, the cause of the congestion. The correct answer label may be annotated by a user or may be assigned according to a predetermined rule. The learning unit 601 performs machine learning of the learning model 611 using the average speed and average operation amount with the correct answer label and data on the corresponding section on the road NW 102 as training data.

[0089] The learning model 611 is configured with a deep neural network or the like. In addition, in order to capture the characteristics of time-varying changes in average speed, it is preferable that the learning model 611 be capable of handling time-series data. For example, when the average speed and average operation amount for a certain section for a predetermined time period and data on the corresponding section on the road NW 102 are input, the learning model 611 outputs the probability that the section is a congested section and the probability that the cause of the congestion is the cause corresponding to each label. The trained learning model 611 is stored in a predetermined storage area of ​​the machine learning device 600.

[0090] The occurrence detection unit 331 uses the learning model 611 to detect the occurrence of congestion from the information stored in the section information DB 203. Specifically, the occurrence detection unit 331 inputs the average speed and average operation amount for each section and each time period stored in the section information DB 203, as well as data on the corresponding section in the road NW 102, into the learning model 611. Then, the occurrence detection unit 331 detects the occurrence of congestion based on the probability that the section is a congested section output by the learning model 611.

[0091] The cause prediction unit 332 predicts the cause of congestion from the information stored in the section information DB 203 using the learning model 611. Specifically, for a section identified as a congested section by the occurrence detection unit 331, the cause prediction unit 332 predicts the cause of congestion from the probability of each label included in the output of the learning model 611 obtained by the processing of the occurrence detection unit 331.

[0092] The hardware configuration of the traffic congestion prediction device 300 according to the third embodiment is similar to the hardware configuration of the traffic congestion prediction device 100 according to the first embodiment shown in FIG. 1, and therefore a description thereof will be omitted.

[0093] Next, the operation of the traffic congestion prediction device 300 will be described.

[0094] First, the learning unit 601 of the machine learning device 600 performs machine learning of the learning model 611 using, as training data, the average speed and average operation amount with correct labels indicating whether a section is congested or not and the cause of the congestion, and data on the corresponding section on the road NW 102. With the trained learning model 611 stored in a predetermined storage area, the congestion prediction device 300 executes a congestion prediction process.

[0095] 18 is a flowchart showing the flow of traffic congestion prediction processing by the traffic congestion prediction device 300. The CPU 11 reads out a traffic congestion prediction program from the ROM 12 or the storage 14, and deploys and executes it in the RAM 13, thereby performing the traffic congestion prediction processing. Note that in the traffic congestion prediction processing according to the third embodiment, processing that is the same as the traffic congestion prediction processing according to the second embodiment (FIG. 16) is assigned the same step number, and detailed description thereof will be omitted.

[0096] After steps S201 to S203, in the next step S304, the CPU 11, as the occurrence detection unit 331, inputs the average speed and average operation amount for each section and each time period stored in the section information DB 203, and data on the corresponding section in the road NW 102, to the learning model 611. Then, the CPU 11, as the occurrence detection unit 331, detects the occurrence of congestion based on the probability that the section is a congested section output by the learning model 611.

[0097] Next, when a positive judgment is made in step S205 and the process proceeds to the next step S306, the CPU 11, as the cause prediction unit 332, predicts the cause of the traffic jam from the probability of each label included in the output of the learning model 611 obtained in step S304.

[0098] As described above, the traffic congestion prediction device according to the third embodiment uses a learning model constructed by machine learning when detecting the occurrence of traffic congestion and predicting the cause of the congestion, thereby achieving the same effects as those of the first and second embodiments.

[0099] In the third embodiment, as in the second embodiment, a case has been described in which a learning model is used to detect the occurrence of congestion and predict the cause of congestion when operation information is used in addition to speed information and location information, but this is not limiting. A learning model may also be used in the configuration of the first embodiment. In this case, a learning model in which machine learning is performed using average speed and road network data as training data may be used.

[0100] In addition, in the third embodiment, a case has been described in which the machine learning device that performs machine learning of the learning model is configured as a device separate from the traffic congestion prediction device, but the two may also be configured as a single device.

[0101] <Fourth embodiment> Next, a fourth embodiment will be described. In the traffic congestion prediction device according to the fourth embodiment, components similar to those of the traffic congestion prediction device 100 according to the first embodiment and the traffic congestion prediction device 200 according to the second embodiment will be assigned the same reference numerals, and descriptions thereof will be omitted. Furthermore, components having some functions in common with the functional units in each embodiment will be assigned reference numerals having the same last two digits, and detailed descriptions thereof will be omitted.

[0102] 19 is a block diagram showing an example of the functional configuration of the traffic congestion prediction device 400. As shown in FIG. 19, the traffic congestion prediction device 400 includes, as its functional configuration, an acquisition unit 410, a calculation unit 420, a prediction unit 430, and an output unit 440. The prediction unit 430 further includes an occurrence detection unit 131, a cause prediction unit 232, and a resolution detection unit 433. In addition, an acquisition information DB 401, a road NW 102, and a section information DB 203 are stored in a predetermined storage area of ​​the traffic congestion prediction device 400. Each functional configuration is realized when the CPU 11 reads out a traffic congestion prediction program stored in the ROM 12 or the storage 14, expands it in the RAM 13, and executes it.

[0103] The acquisition unit 410 acquires speed information, position information, and operation information from each of the multiple vehicles, and also acquires the number of passing vehicles, speed information, and measurement position information from fixed sensors such as vehicle detectors and cameras installed on the road. The number of passing vehicles is the number of vehicles that pass in front of the fixed sensor per predetermined time unit. The speed information is the speed of the passing vehicles. The measurement position information is the latitude and longitude at which the fixed sensor is installed. The measurement position information may be coordinate values ​​on the road NW102.

[0104] The acquired information DB 401 differs from the acquired information DB 201 of the second embodiment in that it stores information acquired from the fixed sensors described above in addition to the speed information, position information, and operation information of each of a plurality of vehicles.

[0105] The calculation unit 420 calculates the average speed and the average operation amount for each section and each time period, similar to the calculation unit 220 according to the second embodiment. The calculation unit 420 also calculates the reliability of each of the average speed and the average operation amount, and corrects the average speed and the average operation amount with low reliability. During these processes, the calculation unit 420 may correct the position information and the speed information using information acquired from fixed sensors. For example, if there is a discrepancy between the speed information acquired from the vehicle and the speed information acquired from the fixed sensors, the calculation unit 420 may use the average, maximum value, or minimum value of both the speed information from the vehicle and the speed information from the fixed sensors.

[0106] The resolution detection unit 433 detects that congestion in a congested section has been resolved when the average speed in that section becomes equal to or greater than the first threshold, starting from the first section of the congested section. For example, as shown in Figure 20, at 16:20, the average vehicle speed near the beginning of the congested section is 40 km / h or less, but at 16:25, the average speed begins to return to 40 km / h or more, indicating that the congestion has begun to be resolved. The resolution detection unit 433 passes resolution information for the section where congestion has been resolved to the output unit 440.

[0107] The resolution detection unit 433 may detect the resolution of traffic congestion using the learning model 611, as in the third embodiment.

[0108] The output unit 440 outputs congestion information including information on congested sections and predicted causes, and also outputs congestion information including resolution information passed from the resolution detection unit 433 .

[0109] The hardware configuration of the traffic congestion prediction device 400 according to the fourth embodiment is similar to the hardware configuration of the traffic congestion prediction device 100 according to the first embodiment shown in FIG. 1, and therefore a description thereof will be omitted.

[0110] Next, the operation of the traffic congestion prediction device 400 will be described. Fig. 21 is a flowchart showing the flow of traffic congestion prediction processing by the traffic congestion prediction device 400. The CPU 11 reads a traffic congestion prediction program from the ROM 12 or storage 14, deploys it in the RAM 13, and executes it, thereby performing the traffic congestion prediction processing. Note that in the traffic congestion prediction processing according to the fourth embodiment, processing that is the same as the traffic congestion prediction processing according to the second embodiment (Fig. 16) is assigned the same step numbers, and detailed description thereof will be omitted.

[0111] After steps S201 to S203, in the next step S401, the CPU 11, functioning as the occurrence detection unit 131, determines whether or not the target section has already been identified as a congested section. If it has been identified as a congested section, the process proceeds to step S402, and if it has not been identified, the process proceeds to step S204.

[0112] In step S402, the CPU 11, functioning as the resolution detection unit 433, detects that congestion in a section has been resolved when the average speed in that section becomes equal to or greater than the first threshold, starting from the first section of the congested section.

[0113] Next, in step S403, the CPU 11, functioning as the resolution detection unit 433, determines whether or not resolution of the congestion has been detected in step S403. If resolution has been detected, the process proceeds to step S404, and if not, the process proceeds to step S206. In step S404, the CPU 11, functioning as the resolution detection unit 433, outputs congestion information including resolution information, and returns to step S201.

[0114] As described above, the traffic congestion prediction device according to the fourth embodiment detects the elimination of traffic congestion based on whether the average speed of the first section of the traffic congestion section is equal to or greater than the first threshold value. This makes it possible to predict the elimination of traffic congestion at an early stage.

[0115] In the fourth embodiment, similarly to the third embodiment, a learning model may be used to detect the occurrence and resolution of traffic congestion and to predict the cause.

[0116] Furthermore, in each of the above embodiments, if there is a "road branch" near the beginning of a traffic jam, a prediction may be made that the cause of the traffic jam will affect roads via the branch. For example, as shown in FIG. 22 , if traffic congestion on road B, such as an interchange on a highway, affects road A, such as the main highway, the average vehicle speed of vehicles on road A will recover immediately after passing the branch. In this case, the change over time in the average speed near the beginning of the traffic jam section on road A will show a trend similar to that of traffic jams caused by accidents or broken-down vehicles. Therefore, the cause of the traffic jam section on road B may be predicted, and it may be predicted that this is the cause of the traffic jam on road A.

[0117] Furthermore, in each of the above embodiments, speed information may lack reliability if it is data obtained from a small number of vehicles, including a single vehicle. For example, consider a situation as shown in Figure 23 where only the left lane is congested due to traffic entering a drive-through, while the other lanes are not congested. In this case, if speed information of the shaded vehicles were collected, the average speed would be less than the first threshold, and a traffic jam would be detected.

[0118] Therefore, if there are any vehicles with high speeds among the vehicles for which the average is calculated, a correction may be made to assume that factors other than congestion exist. For example, the average may be calculated by further dividing at least one of the time period and the section, or variance, maximum value, minimum value, etc. may be taken into consideration. Furthermore, past route trends and driving lane information of vehicles for which speed information can be collected may be taken into consideration.

[0119] In the above embodiments, the time period for calculating the average speed is set to 5 minutes and the section is set to 125 meters, but these values ​​may be variable or may be fixed to other values. Furthermore, considering that the number of vehicles for which the average speed can be obtained varies depending on the region, these values ​​may be changed depending on the region.

[0120] Furthermore, the speed information is not limited to that detected by a vehicle speed sensor, but may be calculated from changes in position information measured by GPS, or may be calculated using sensors installed in various locations that measure cross-sectional traffic volume or vehicle speed. Furthermore, a value similar to the speed measured by these sensors or that can be converted to speed, such as the number of vehicles passing in one minute, may be used instead.

[0121] Furthermore, when predicting the cause of congestion, trends by time period (morning, evening, etc.) or by day of the week may be taken into consideration. For example, when comparing only time periods (between 7:00 a.m. and 5:00 p.m.) when traffic volume is high and congestion due to traffic concentration occurs, changes in the initial congestion length and average speed may be taken into consideration.

[0122] Furthermore, if an evaluation limited to a wider or more limited section reveals a different trend from the difference in the trend of time-dependent changes in average speed depending on the cause of congestion described in the above embodiment, a threshold appropriate for the section may be set and compared. For example, on a road with only one lane in each direction, in a location where a reversal phenomenon of a trend is observed, such as the impact of congestion caused by an accident or a broken-down vehicle extending faster than congestion caused by traffic concentration at the same time, a threshold appropriate for this trend may be set.

[0123] There are also cases where a traffic jam caused by traffic concentration develops into a traffic jam caused by multiple factors along the way, such as a traffic jam caused by an accident on the way, or a traffic jam caused by traffic concentration that occurs when the traffic volume increases due to vehicles detouring due to an accident congestion on another route.Even in such cases, it is possible to predict multiple factors by combining the trends described in the above embodiment and setting an appropriate threshold value.

[0124] The technology disclosed herein can also be applied to predicting the cause of congestion for some lanes, such as only one lane or only two lanes out of three lanes, that is, predicting the cause of congestion on a lane-by-lane basis.

[0125] For example, the lane in which each vehicle is traveling, whose speed information has been collected, may be identified based on the position information collected from each vehicle and its transition. Alternatively, video from a drive recorder, sensor data such as LiDAR, etc. may be acquired, and the lane in which each vehicle is traveling may be identified based on the acquired information. If low-speed vehicles are concentrated in a particular lane and can travel at a sufficient speed in other lanes, it may be predicted that the cause of the congestion is near the beginning of the section in which low-speed vehicles are concentrated in the lane in which the low-speed vehicles are concentrated. Whether low-speed vehicles are concentrated in a particular lane may be determined, for example, by dividing the "section" into lanes rather than roads, calculating the average speed for each lane section, and comparing the average speeds for each lane.

[0126] 23, when only the first lane from the left is congested, the system checks the road network and map information to see if there are any commercial facilities with drive-throughs or parking lots, or intersections or traffic lights at the beginning of the congested section or the section immediately following it. If these exist, the system predicts that they are the cause, and if not, the system predicts that an accident or a broken-down vehicle is the cause.

[0127] Furthermore, in the above embodiments, the moving body is a vehicle, but the present invention is not limited to this. The moving body may be a train, a ship, a person (pedestrian), a bicycle, or the like. Specifically, in the case of pedestrians, a huge number of people pass through major stations in the metropolitan area during rush hour. In this case, the occurrence of congestion or jamming may be detected based on the number of pedestrians per unit time and per section and their average walking speed. Furthermore, based on the change in average walking speed over time, it may be possible to predict trends that differ from normal, such as the elimination of congestion or the malfunction of a moving walkway or escalator ahead.

[0128] Similarly, for example, if the average speed of a ship traveling on a river decreases, it may be predicted that the cause, such as a large ship running aground, is near the front section, and it may be assumed that the cause has been resolved within a unit time when the speed begins to improve due to the change in average speed over time.

[0129] In addition, the traffic congestion prediction process executed by the CPU by reading the software (program) in each of the above embodiments may be executed by various processors other than the CPU. Examples of processors in this case include dedicated electrical circuits, such as programmable logic devices (PLDs) (such as field-programmable gate arrays (FPGAs)) whose circuit configuration can be changed after manufacture, and application-specific integrated circuits (ASICs) that are processors having a circuit configuration designed specifically to execute specific processes. The traffic congestion prediction process may be executed by one of these various processors, or by a combination of two or more processors of the same or different types (e.g., multiple FPGAs, a combination of a CPU and an FPGA, etc.). The hardware structure of these various processors is, more specifically, an electrical circuit that combines circuit elements such as semiconductor elements.

[0130] In addition, in each of the above embodiments, the congestion prediction program is described as being pre-stored (installed) in the storage 14, but the present invention is not limited to this. The program may be provided in a form stored on a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), or a USB (Universal Serial Bus) memory. The program may also be downloaded from an external device via a network.

[0131] The following additional notes are provided regarding the above-described embodiments.

[0132] (Additional Item 1) A traffic congestion prediction device including: an acquisition unit that acquires speed information and position information of a moving body; and a prediction unit that predicts the cause of traffic congestion in a section specified by the position information based on a change over time in the speed information for each section.

[0133] (Supplementary Item 2) The traffic congestion prediction device described in Supplementary Item 1, wherein the prediction unit detects the occurrence of traffic congestion in the section when the average speed obtained by averaging the speed information of each of the multiple moving bodies for each section is less than a predetermined first threshold value.

[0134] (Supplementary Item 3) The traffic congestion prediction device described in Supplementary Item 2, wherein the prediction unit predicts the cause of the traffic congestion based on at least one of the following as the time change of the speed information: traffic congestion length, which is the length of the traffic congestion section, which is a plurality of consecutive sections where the occurrence of traffic congestion is detected at the time of the traffic congestion occurrence; change in traffic congestion length after the traffic congestion occurrence; average speed after the traffic congestion occurrence; recovery tendency of the average speed in sections ahead of the beginning of the traffic congestion section; and coordinates of the beginning of the traffic congestion section.

[0135] (Supplementary Item 4) The traffic congestion prediction device described in Supplementary Item 2 or Supplementary Item 3 further includes a calculation unit that calculates an average speed by averaging the speed information of each of the multiple moving bodies for each section, calculates the reliability of the average speed, and if the reliability is lower than a predetermined second threshold, sets the average of the average speeds of the target section and the sections before and after it as the average speed of the target section.

[0136] (Supplementary Item 5) The traffic congestion prediction device according to any one of Supplementary Items 2 to 4, wherein the prediction unit detects that the traffic congestion in a section has been resolved when the average speed of the section becomes equal to or greater than the first threshold, starting from the first section of the traffic congestion section.

[0137] (Supplementary Item 6) The traffic congestion prediction device according to any one of Supplementary Items 1 to 5, further including an output unit that outputs a prediction result by the prediction unit to an information processing device used by a user associated with the mobile object.

[0138] (Supplementary Item 7) A traffic congestion prediction program for causing a computer to function as each part of the traffic congestion prediction device according to any one of Supplementary Items 1 to 6.

[0139] (Supplementary Item 8) A traffic congestion prediction device including: a memory; and at least one processor connected to the memory, wherein the processor is configured to acquire speed information and position information of a moving object, and predict the cause of traffic congestion in each section identified by the position information based on changes over time in the speed information.

[0140] (Supplementary Item 9) A non-transitory storage medium storing a program executable by a computer to execute a traffic congestion prediction process, wherein the traffic congestion prediction process acquires speed information and position information of a moving object, and predicts the cause of traffic congestion in each section identified by the position information based on changes over time in the speed information.

[0141] DESCRIPTION OF SYMBOLS 11 CPU 12 ROM 13 RAM 14 Storage 15 Input unit 16 Display unit 17 Communication I / F 19 Bus 100, 200, 300, 400 Traffic congestion prediction device 101, 201, 401 Acquired information DB 103, 203 Section information DB 110, 210, 410 Acquisition unit 120, 220, 420 Calculation unit 130, 230, 330, 430 Prediction unit 131, 331 Occurrence detection unit 132, 232, 332 Cause prediction unit 433 Resolution detection unit 140, 440 Output unit 600 Machine learning device 601 Learning unit 611 Learning model

Claims

1. A congestion prediction device including: an acquisition unit that acquires speed information and position information of a moving object; and a prediction unit that predicts a cause of congestion in a section based on a temporal change of the speed information for each section specified by the position information.

2. The congestion prediction device according to claim 1, wherein the prediction unit detects occurrence of congestion in the section when an average speed obtained by averaging the speed information of each of a plurality of moving objects for each section is less than a predetermined first threshold value.

3. The congestion prediction device according to claim 2, wherein, as the temporal change of the speed information, the prediction unit predicts the cause of the congestion based on at least one of: a congestion length which is a length of a congestion section that is a plurality of consecutive sections in which occurrence of congestion is detected at the time of congestion occurrence; a change in the congestion length after congestion occurrence; an average speed after congestion occurrence; a recovery tendency of an average speed of a section ahead of the head of the congestion section; and a coordinate of the head of the congestion section.

4. The congestion prediction device according to claim 2 or claim 3, further including a calculation unit that calculates an average speed obtained by averaging the speed information of each of the plurality of moving objects for each section, calculates a reliability of the average speed, and when the reliability is lower than a predetermined second threshold value, sets an average of the average speeds of a target section and sections before and after the target section as the average speed of the target section.

5. The congestion prediction device according to claim 2 or claim 3, wherein the prediction unit detects that congestion in a section has been resolved when the average speed of the section becomes greater than or equal to the first threshold value in order from a section at the head of a congestion section.

6. The congestion prediction device according to any one of claims 1 to 3, further including an output unit that outputs a prediction result by the prediction unit to an information processing device used by a user associated with the moving object.

7. A congestion prediction method, wherein an acquisition unit acquires speed information and position information of a moving object, and a prediction unit predicts a cause of congestion in a section based on a temporal change of the speed information for each section specified by the position information.

8. A congestion prediction program for causing a computer to function as each unit of the congestion prediction device according to any one of claims 1 to 3.

Images