JP2007219633A - Travel time prediction device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To precisely predict the travel time of a road after the occurrence of such a sudden event as an accident on a road or disaster. <P>SOLUTION: The prediction object section of a travel time including a section where any sudden event has occurred is set, and the prediction object section is divided into a plurality of blocks, and the traffic density of each block at a certain time t is calculated, and movable traffic volume is calculated between adjacent blocks based on the traffic density of each block, and the traffic density of each block at the precious time t+ΔT is predicted based on the traffic density of each block and the traffic volume movable between the adjacent blocks, and a travel time is predicted. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a travel time predicting apparatus and method for predicting a travel time on a road after occurrence of a sudden event such as an accident or disaster on the road.

A vehicle detector is installed on a road to acquire current travel time data on the road, and a future travel time is predicted based on the travel time data.
As this prediction method, past travel time data is accumulated as statistical information, and the travel time is predicted on the assumption that the current travel time increases / decreases in the same way as the past (changes in the same pattern). .

On the other hand, in order to calculate the travel time of vehicles traveling on the road, a block model of unit distance is connected to form a road model, and the traffic volume between blocks according to the traffic density of each block and the set traffic volume-density curve There is a method for expressing the traffic flow by calculating the number and moving the number corresponding to this traffic volume.
JP 2002-260142 A JP-A-9-147285

However, when a sudden event such as an accident or disaster occurs on the road, it is impossible to predict with high accuracy even if travel time data accumulated in the past is used.
Therefore, an object of the present invention is to provide a travel time prediction apparatus and method that can accurately predict a travel time on a road after occurrence of a sudden event such as an accident or a disaster on the road.

  The travel time prediction apparatus of the present invention includes a setting means for setting a travel time prediction target section including a section in which a sudden event occurs, and the prediction target section is divided into a plurality of blocks, and traffic in each block at a certain time A traffic density calculating means for calculating a density, a movable traffic volume calculating means for calculating a traffic volume movable between adjacent blocks at the certain time based on the traffic density in each block, and in each block Traffic density predicting means for predicting the traffic density of each block at the previous time based on the traffic density of each block and the traffic volume movable between the adjacent blocks, and the traffic density predicting means predicted by the traffic density predicting means Travel time prediction means for predicting travel time based on the traffic density of each block at the time of day.

This travel time prediction apparatus sets a prediction target section including a section in which a restriction has occurred due to a sudden event, and divides the block into unit length blocks. For the divided blocks, the travel time is allocated to each block from the travel time of each acquired link, the travel time is converted into a speed, and the traffic density of each block is calculated.
To convert speed to traffic density, a speed-traffic density curve can be used.

  When the speed-traffic density curve is used, if the acquired traffic speed is a free flow speed, the traffic density cannot be uniquely determined because the traffic flow is not saturated. Therefore, the traffic density may be determined by referring to the traffic density information related to at least one or more combinations of the route, area, day type (day of week, holiday, etc.) and time zone to which the prediction target section belongs stored in advance.

Further, it is preferable to predict the travel time by changing the traffic density, obtain an average error from the actual travel time acquired under the same conditions, and store the traffic density that minimizes the average error as data.
And based on the traffic density in each said block, the traffic volume which can move between the adjacent blocks in the said certain time is calculated.
Specifically, based on the traffic density of each block, the traffic volume that can exit the block and the traffic volume that can enter the block are calculated for each block. Between adjacent blocks, the traffic volume that can exit from the upstream block and the traffic volume that can enter the downstream block are compared, and the smaller traffic volume is defined as the movable traffic volume between the blocks.

A traffic volume-traffic density curve can be used to calculate the traffic volume that can move between the blocks.
In this case, it is considered that the movable traffic volume is restricted by the lane restriction in the block where the sudden event has occurred. That is, different traffic volume-traffic density curves are used according to the lane restriction ratio of the section where the sudden event has occurred.

The “different traffic volume-traffic density curve” is, for example, a curve in which different critical traffic density and jam traffic density are set according to the throttle rate of the section where the sudden event occurs.
Based on the movable traffic volume and the traffic density in each block, the traffic density of each block at the previous time is predicted. If the traffic density can be predicted, the travel density can be converted into the vehicle speed and the travel time can be obtained.

Then, it is possible to predict the travel time at any future time point by repeating the above-described processing at regular time intervals.
The traffic volume flowing in from the upstream end of the prediction target section or the traffic volume flowing out from the downstream end is obtained separately from the “traffic movable traffic volume”.
For example, the speed is obtained based on the travel time of the link adjacent to the upstream end or the downstream end of the prediction target section, the traffic density of the link is obtained from the speed, and the traffic flowing in from the upstream end using the traffic density Alternatively, the traffic volume flowing out from the downstream end can be calculated.

  Since the “traffic volume flowing in from the upstream end of the prediction target section or the traffic volume flowing out from the downstream end” is the traffic volume in the section deviating from the prediction target section, the influence of the sudden event is small. Therefore, for example, a value obtained by using a conventional method (a prediction method using pattern matching of travel time transition and statistical information) may be employed. Further, the travel time of the link at the current time point may be used as a constant value without fluctuation.

The timing for starting the travel time prediction process is obtained when information on the sudden event is acquired, or information on the travel time of the link is acquired over a plurality of time zones, and the increase rate of the travel time in the time zone is out of a predetermined range. When it becomes, it is good to do.
In the latter case, since the travel time increases rapidly, it can be determined that a sudden event has occurred.

If it is determined that a sudden event has occurred, the travel time prediction corresponding to the sudden start is started for the target link.
The end of the travel time prediction may be a time when it is considered that the regulation based on the sudden event is canceled.
The restriction cancellation time may be set in accordance with the type of sudden event.

  The travel time prediction method of the present invention is a method according to the substantially same invention as the travel time prediction apparatus of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
In the embodiment of the present invention, the travel time prediction device always monitors road regulation information from the outside and lists the road regulation information in advance. Usually, the travel time is predicted by the method using pattern matching and statistical information described in [Background Art]. Once a sudden event occurs, road regulation information is obtained from the regulation list, and a prediction target section including the point where the sudden occurrence is targeted is used for prediction using this method.

  The switching timing from the normal prediction to the travel time prediction corresponding to the sudden event is performed when information on the sudden event, for example, road regulation information is acquired from a predetermined external organization. In addition, when the increase in travel time acquired from a predetermined external organization is abrupt, for example, when the transition rate of the acquired travel time is large compared with the result of analyzing the increase in travel time at the time of a past accident, Assume that a sudden event occurs.

Hereinafter, the travel time prediction process corresponding to the sudden event will be described.
First, terms are defined.
Link: A unit of road section of a route.
Prediction target section: road section for which travel time is to be predicted. A series of links including at least a link where an unexpected event has occurred. As the prediction target section, a range in which congestion is expected to occur due to a sudden event is selected.

Free flow speed Vf: Speed at which the vehicle can travel freely. Set to a constant value.
Scan time ΔT: processing time for one cycle for travel time prediction processing, a constant value. All blocks to be predicted are processed in this one cycle.
Block B: A section obtained by further dividing the link and passing during the scan time ΔT when the vehicle travels at the free flow velocity Vf. The distance L of one block is called “block length”. The block length is constant.

Traffic Q: Number of vehicles passing through the point in unit time (units / hour).
Traffic density K: Number of vehicles (unit / distance) per unit distance existing in the block. Between the traffic volume Q, the traffic density K, and the speed V, there is a relationship of “traffic volume Q = traffic density K × speed V”.
Critical traffic density Kc: Traffic density when traffic density K increases and traffic congestion begins.
Jam traffic density Kjam: Traffic density when the degree of traffic congestion increases and vehicle movement stops.

FIG. 1 is a road diagram showing a prediction target section. The prediction target section includes a series of linked links 1 to 3. Each link 1 to 3 is divided into blocks. A series of numbers is assigned to the blocks to form blocks B0 to B8.
In the present embodiment, the block numbers are assigned so that the numbers increase from the downstream in the traveling direction of the vehicle to the upstream. That is, the traveling direction of the vehicle is the direction of B8 → B0. The most upstream block B8 is referred to as a start block, and the most downstream block B0 is referred to as a termination block.

In FIG. 1, the number of links constituting the prediction target section is 3 and the number of blocks is 9, but the implementation of the present invention is not limited to these numbers.
When a sudden event occurs, the link travel time predicting device of the present invention specifies a prediction target section including a link where the sudden event has occurred, acquires the traffic volume transmitted between the blocks and the traffic density of each block, and the previous time The traffic density of each block is calculated, the speed of each block is obtained based on the traffic density of each block, and the travel time is predicted.

In the following, the subscript representing a block is assumed to be lead.
The traffic volume from the block Bi + 1 to the block Bi at time t is expressed as Qi + 1, i (t), and the traffic density of the block Bi is expressed as Ki (t).
FIG. 2 is a schematic diagram illustrating the relationship between the traffic volume between blocks and the traffic density within the block. FIG. 2 shows how the traffic volume and traffic density at time t transition to the traffic volume and traffic density at time t + ΔT. For example, when attention is paid to block Bi, traffic density Ki (t at time t is determined by traffic volume Qi + 1, i (t) flowing into block Bi and traffic volume Qi, i-1 (t) flowing out from block Bi. ) Shifts to traffic density Ki (t + ΔT) at time t + ΔT.

  FIG. 3 schematically shows the state of the road when the sudden event occurs. FIG. 3 shows a state where an accident or a disaster has occurred in the block Bi on a two-lane road, and one lane cannot travel. However, in the adjacent blocks Bi-1 and Bi + 1, both lanes can travel. Such a lane decrease state of the block Bi is referred to as “aperture” (iris), and a ratio of all lanes to a passable lane after the lane decrease is referred to as “aperture ratio”. In FIG. 3, the aperture ratio in the block Bi is 0.5. However, in the case of a road with only one lane, if the number of lanes decreases, there will be no traffic lanes and the throttle rate will be 0. In this case, however, the throttle rate is greater than 0, assuming that it can run on the shoulder. Set to a value (0.1 to 0.3).

Hereinafter, a travel time prediction algorithm after the occurrence of the sudden event will be described.
If a sudden event occurs, the prediction target section for which the traffic jam is to be predicted is specified. The prediction target section may be only one link or a plurality of links may be connected.
The link is broken down into blocks, and for each block, the traffic volume flowing into and out of the block and the traffic density within the block are determined.

FIG. 4 is a block diagram showing the overall flow.
The whole process includes a sudden event detection process 12 for detecting a sudden event and a travel time prediction process 13. In addition to this, although not shown, there is a normal travel time prediction process that proceeds even if a sudden event does not occur.
The overall flow will be described. First, link travel time / regulation data received from an external organization is always stored in the memory 14. The link travel time / restriction data includes information on the presence / absence of a sudden event, occurrence location, presence / absence of lane restrictions, and throttle rate, along with link travel time data.

In the sudden event detection process 12, the memory 14 is read every certain time, for example, every minute.
Further, the sudden event detection process 12 goes to read the road map data when necessary in order to see the link connection.
The sudden event detection process 12 creates a sudden event occurrence table as shown in Table 1 and stores it in the memory 16 when a sudden event occurs based on the link travel time / regulation data and the road map data.

  In Table 1, “Time” refers to the time (for example, every minute) when the memory 14 is read. The “regulation occurrence mesh number” is a map number including a place where the sudden event has occurred. The “regulation occurrence link number” is a number for identifying the link where the sudden event has occurred. The “restriction start position” is the upstream end position of the lane-restricted road, and the “restriction end position” is the downstream end position of the lane-restricted road. “Regulation type information” indicates the type of sudden events such as accidents and disasters, the number of traffic lanes, and the throttle rate. The “link travel time / regulation data” is data acquired from the memory 14.

In the travel time prediction process, the unexpected event occurrence table in the memory 16 is read every prediction cycle, for example, every 5 minutes.
The travel time prediction process performs a travel time prediction calculation based on the traffic density in the block read from the sudden event occurrence table and the traffic volume flowing into and out of the block.
FIG. 5 is a graph showing the relationship between the traffic density K and the traffic volume Q in the block when the traffic density is taken on the horizontal axis and the traffic volume is taken on the vertical axis. This graph is called “traffic volume-traffic density curve”. When the traffic density K is less than or equal to the critical traffic density Kc, it is referred to as a free flow region, and the vehicle can travel at a free flow speed Vf without congestion. If it is below the critical traffic density Kc, the traffic volume Q increases in proportion to the traffic density. The traffic volume Qc at the critical traffic density Kc indicates the maximum traffic volume that the block can hold. If the critical traffic density Kc is exceeded, the traffic volume Q decreases as the traffic density increases. When the traffic density K becomes the jam traffic density Kjam, the traffic volume becomes zero. This means that the vehicle has stopped moving due to traffic jams.

The above graph can be expressed as follows. The subscript i is a block subscript. (T) represents a function of time.
Qi (t) = Ki (t) x Vf (when K (t) ≤ Kc) (1)
Qi (t) = Qci (Kjam-Ki (t)) / (Kjam-Kci)
(Kc <K (t) ≤ Kjam) (2)
FIG. 6 is a graph showing the relationship between the traffic density K in the block and the traveling speed V when the horizontal axis represents the traffic density and the vertical axis represents the traveling speed. This graph is called “speed-traffic density curve”. When the traffic density K is less than or equal to the critical traffic density Kc, it is assumed that the vehicle is traveling at a constant speed (free flow speed) Vf. If the critical traffic density Kc is exceeded, the speed V decreases. When the traffic density K becomes the jam traffic density Kjam, the speed becomes zero.

If the above graph is expressed by an equation,
Ki (t) = αKci (when V = Vf) (3)
Ki (t) = KjamQci / [V (Kjam-Kci) + Qci]
(When V <Vf) (4)
It becomes. α is a constant from 0 to 1, but when V = Vf, Ki (t) is not uniquely determined even if the traveling speed Vf is given. The handling in this case will be described later.

The flowable traffic volume Aout from the block Bi is expressed by the following equation.
Aout, i (t) = min {Ki (t), Kci} Vf (5)
That is, the flowable traffic volume Aout from the block Bi is represented by the product of the smaller value of the traffic density K and the critical traffic density Kc at that time and the free flow velocity Vf.
The inflowable traffic volume Ain to the block Bi is expressed by the following equation.

Ain, i (t) = (Kjam−Ki (t)) Vf (when K (t) ≦ Kc) (6)
Ain, i (t) = Kci (Kjam−Ki (t)) Vf / (Kjam−Kci) (when Kc <K (t) <Kjam) (7)
Ain, i (t) = 0 (when K (t) = Kjam) (8)
This equation is the graph of FIG.

The movable traffic volume Qi + 1, i (t) between blocks (exiting block Bi + 1 and entering block Bi) is the flowable traffic volume Ain, i (t) of block Bi and can flow out of block Bi + 1 Expressed as the smaller of the traffic volumes Aout, i + 1 (t).
Qi + 1, i (t) = min {Ain, i (t), Aout, i + 1 (t)} (9)
The traffic density of the block Bi at time t + ΔT is obtained by subtracting the traffic density (value obtained by dividing the traffic volume by the speed) from the block Bi from the traffic density Ki (t) of the block Bi at time t. It can be determined by adding the incoming traffic density (the traffic volume divided by the speed).

Ki (t + .DELTA.T) = Ki (t) -Qi, i-1 (t) / Vf + Qi + 1, i (t) / Vf (10)
Thus, based on the traffic density of the block Bi at time t, the traffic density of the block Bi at time t + ΔT can be predicted.
Further, if the equations (5) to (9) are used, the traffic density at the time t + 2ΔT can be calculated using the traffic density of the block Bi at the time t + ΔT. In this way, the traffic density at the time point of t + nΔT (n ≧ 1) can be calculated.

If the traffic density for each block is calculated in this way, the traffic density can be converted into a speed based on equations (3) and (4) (FIG. 6). The travel time of the block is obtained using this speed, and the travel time of the prediction target section is obtained by adding the travel time for the blocks constituting the prediction target section.
Next, a calculation processing procedure when the link travel time prediction apparatus of the present invention executes the prediction method described above will be described based on a flowchart.

An overall flowchart is shown in FIG. The overall flowchart is divided into an initialization process, a prediction accuracy determination process, a traffic density calculation process, and a travel time prediction process.
The initialization process (step S1) is a process for setting various constants after the occurrence of the sudden event.
In the prediction accuracy determination process (step S2), the predicted value of the travel time obtained one scan time ago is compared with the value of the current travel time stored in the memory 14, and the absolute value of the difference is equal to or greater than the threshold value. If the ratio of the travel time obtained one scan time ago and the current travel time divided by the smaller one is equal to or greater than a threshold value, the travel time prediction is canceled. This is for the purpose of saving computing resources because it is recognized that the travel time is not predicted for some reason.

The traffic density calculation process (step S3) is a process for calculating the traffic density using the above-described equations (5) to (10).
As described above, the travel time prediction process (step S4) is a process for converting the traffic density into a speed based on the equations (3) and (4) and obtaining the travel time of the prediction target section using this speed. is there.
Hereinafter, the contents of the initialization process (step S1) will be described in more detail with reference to FIG.

  When the trip time prediction process 13 detects the occurrence of the sudden event with reference to the trip time / regulation data in the memory 14, the trip time prediction process 13 specifies the prediction target section (step S11). In this identification method, a predetermined number (a) of links is selected upstream from the block including the location of the occurrence of the sudden event, and a predetermined number (b) of links is selected downstream. Both the numbers a and b are set to the maximum number of sections in which traffic congestion is predicted due to a sudden event. In general, a> b because the traffic jam affects the upstream side of the occurrence of the sudden event. For example, a is the number of links corresponding to 5 km to 10 km, and b is the number of links corresponding to 0 km to 2 km. Then, the link is divided into blocks. The length of one block is about 270 m when the free flow velocity Vf is 100 km / h and the scan time ΔT is 10 seconds.

  Next, prediction accuracy determination is performed (step S12). As described above, this process compares the travel time obtained one prediction cycle before the current travel time value stored in the memory 14, and if the absolute value of the difference is equal to or greater than the threshold, This is a process for canceling the time prediction. When canceling the travel time prediction, “2: cancel prediction” is set in the memory 16 as in step S16.

If the prediction is not stopped, the variables in each block are acquired and set (step S14). Variables include “free flow velocity”, “critical traffic density Kci”, “jam traffic density Kjam”, and “squeezing rate”.
The free flow velocity, the critical traffic density, and the jam traffic density are values specific to the route or the link, and are stored in the road map table of the memory 15.

The aperture ratio is a numerical value that is set based on restriction information given from an external organization and is stored in the table of the memory 16 as sudden event occurrence information.
The travel time prediction process 13 sets the throttle rate in the block where the sudden event has occurred. The critical traffic density Kc and jam traffic density Kjam of the block are reduced in accordance with the aperture ratio. This is indicated by arrows in FIG. For example, if the aperture ratio is 0.5, the critical traffic density Kc and the jam traffic density Kjam are also halved.

Next, the current traffic density Ki (t) in the block is acquired, and inflow / outflow possible traffic volumes Ain, i (t), Aout, i (t) are calculated (step S15).
This traffic density Ki (t) is calculated based on travel time data given from an external organization and stored in the memory 16. That is, the speed is obtained from the travel time, and the traffic density is obtained from the speed.

  In the method of obtaining the speed from the travel time, the travel time of the link is divided into blocks according to the block length, and the divided travel time is divided by the block length. The calculation formulas for obtaining the traffic density are the above-mentioned formulas (3) and (4) (FIG. 6). When V <Vf, the traffic density is uniquely determined based on equation (4). However, when V = Vf, the traffic density Ki (t) is not uniquely determined even if the travel speed Vf is given. .

  Therefore, the value of α is set in a predetermined table such as a road map table according to the route (national road, local road, etc.), day type, time zone, and weather of the prediction target section. For example, set α to a lower value for free time, set α to a higher value for busy hours, set α to a lower value for fine weather, and set α to a higher value for rainy weather. is there. If the route is crowded on a special day, consider the day type. For determining α, α is set to be small because the traffic density is low in experience in situations where it is difficult to get crowded (for example, when there is a vacant time, fine weather). In the case of a crowded situation (for example, a busy time zone, rainy weather), α is set to be large because the traffic density is high according to experience.

The traffic density information managed from these routes, regions, day types, and time zones may be updated based on link travel time data received from an external organization.
In addition, using past performance information at the time of sudden events, predictions are made by changing the pattern of traffic density at free flow speed, and it is judged using the evaluation index whether it matches the transition of actual travel time. , The pattern when the travel time of the predicted result and the actual travel time best match is the parameter of the corresponding time zone and the relevant route, and this is repeated, and the traffic at the free flow speed by route, region, day type, and time zone is repeated. The density may be determined.

Thus, if the value of α is obtained, the traffic density K (t) can be known.
If the traffic density K (t) is known, the inflow / outflow possible traffic volume Ain, i (t), Aout, i (t) can be calculated from the equations (5) and (6).
Details of the traffic density calculation process (step S3) are shown in the flowchart of FIG.
First, the number J of calculation processes is set. This J depends on how many hours ahead you want to predict your travel time. For example, if one scan time is predicted to be 1 second and 3 hours ahead, J is determined to be 10800.

First, the number j of calculation processes is set to 1 (step S31).
Next, the current traffic flowing in from the upstream end of the start block of the prediction target section, that is, the current traffic flowing out of the link adjacent to the upstream end of the start block of the prediction target section is set as a boundary condition (step) S32). The traffic volume is obtained by dividing the travel time of the link adjacent to the upstream of the starting block by the link length to obtain the speed, and the obtained speed is applied to the graph of FIG. 6 to obtain the traffic density of the adjacent link. At this time, if V <Vf, the traffic density is uniquely determined, but if V = Vf, the traffic density is not uniquely determined even if the travel speed is given. It is determined uniquely by setting α. The traffic volume is calculated using the traffic density and the traffic volume-traffic density curve.

In addition to setting the start block of the prediction target section as a boundary condition, the traffic volume at the downstream end of the terminal block can also be set as a boundary condition.
Next, it is determined whether or not a time when the restriction is considered to be canceled (referred to as a restriction cancellation time) has passed (step S33). This “regulation elimination time” may be determined empirically and statistically in relation to the type of sudden event such as an accident or disaster. If there is an accident, if there is past statistical data indicating that the lane restriction will be resolved in about U time (for example, U = 2) by post-accident processing, a numerical value called U time is set. If there is a disaster, if there is past statistical data indicating that the lane restriction will be resolved in about normal W time (for example, W = 12), a numerical value called W time is set.

It is determined whether or not the time has elapsed (step S34), and if it has elapsed, the aperture ratio is returned to normal (step S35). By setting the regulation elimination time in this way, it is possible to perform prediction after the regulation is eliminated at the stage where the regulation is not eliminated.
If the time has not elapsed, the traffic density of each block in the prediction target section at the time t + ΔT is calculated based on the equations (9) and (10), and the traffic density Ki (t) of the block Bi at the time t. Then, prediction is made based on the inflow / outflow possible traffic volume Ain, i (t), Aout, i (t) (step S36).

Next, it is determined whether or not the number of calculation processes has reached J (step S37). If the number has not reached J, 1 is added to the number j of the calculation process (step S38), and the boundary condition at the upstream end at that time is set (step S39).
As the upstream boundary condition in this case, since traffic data at a future time corresponding to the number j of the calculation processes is necessary, the travel time data stored at the current time stored in the memory 16 as in step S32 is used. In principle, it cannot be used. Therefore, the following method (1) or (2) is used.

(1) The travel time data at the present time is used as travel time data at the future time.
(2) Select the pattern with the most recent travel time transition pattern from the past statistical travel time time transition patterns under the same conditions such as day of the week and weather, and the current travel time data along the pattern is selected. Obtained by extension (see Patent Document 1).
FIG. 10 is a graph for explaining a method of extending the travel time data at the present time point, and a similar pattern selected from the time transition patterns of the past statistical travel time under the same conditions such as day of the week and weather. P represents the actual value of the travel time data up to the present time point. The future travel time data R ′ is obtained by extending the travel time data at the current time point along the time transition pattern P in parallel.

Thereafter, the process proceeds to step S34, and the block traffic density prediction process (step S36) is repeated until the number of calculation processes reaches J.
If the number of calculation processes reaches J, the travel time estimation process starts (step S40). This travel time estimation process can be obtained by using the equation (3) and equation (4) based on the block traffic density K of each block.

Since the travel time of each block is obtained in this manner, the travel time of the prediction target section can be obtained by summing up the prediction target sections.
After obtaining the travel time of the prediction target section, the prediction result of this method is overwritten on the predicted value of travel time by the method using the normal pattern matching / statistical information.
Although the embodiments of the present invention have been described above, the embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made within the scope of the present invention.

It is a road map which shows the prediction object area which is going to predict travel time. It is a schematic diagram which shows the relationship between the traffic volume between blocks, and the traffic density in a block. It is a schematic diagram which shows the state of the road when a sudden event occurs. It is a block diagram which shows the whole flow of a travel time prediction process. It is a graph which shows the relationship between the traffic density in a block, and traffic volume. It is a graph which shows the relationship between the traffic density in a block, and the driving speed V. It is a whole flowchart which shows the calculation processing procedure in the case of performing the link travel time prediction method of this invention. It is a flowchart for demonstrating the content of the initialization process. It is a flowchart for demonstrating the content of a traffic density calculation process. It is a flowchart for demonstrating the method to extend the travel time data at the present time.

Explanation of symbols

12 Sudden event detection process 13 Travel time prediction process 14-18 Memory

Claims (14)

An apparatus for predicting travel time at a previous time of a vehicle traveling on a road,
A setting means for setting a travel time prediction target section including a section where the sudden event has occurred;
A traffic density calculating means for dividing the prediction target section into a plurality of blocks and calculating a traffic density in each block at a certain time;
Based on the traffic density in each block, a movable traffic volume calculating means for calculating a traffic volume that can move between adjacent blocks at the certain time;
Traffic density predicting means for predicting the traffic density of each block at the previous time based on the traffic density in each block and the traffic volume movable between the adjacent blocks;
A travel time prediction device comprising travel time prediction means for predicting travel time based on the traffic density of each block at the previous time predicted by the traffic density prediction means.   The travel time prediction device according to claim 1, wherein the traffic density calculation means acquires traffic speed information of a prediction target section, and calculates a traffic density using a speed-traffic density curve based on the traffic speed information.   When the acquired traffic speed is a free flow speed, the traffic density calculating means is a traffic related to a combination of at least one or more of a route, a region, a day type, and a time zone to which the prediction target section stored in advance belongs. The travel time prediction apparatus according to claim 2, wherein the traffic density is determined with reference to the density information.   The travel time prediction apparatus according to claim 1, wherein the movable traffic volume calculating means calculates the movable traffic volume using a traffic volume-traffic density curve based on the traffic density in each block.   5. The travel time prediction device according to claim 4, wherein the movable traffic volume calculation means uses different traffic volume-traffic density curves according to the lane restriction ratio of the section where the sudden event occurs.   The travel time prediction device according to claim 5, wherein a critical traffic density and a jam traffic density are set according to a lane restriction ratio of a section where the sudden event occurs. Boundary traffic volume calculating means for calculating the traffic volume flowing in from the upstream end of the prediction target section or the traffic volume flowing out from the downstream end,
The traffic density predicting means includes a traffic density in each block, a traffic volume flowing in from an upstream end of the prediction target section or a traffic volume flowing out from a downstream end, and a traffic volume movable between the adjacent blocks. The travel time prediction device according to claim 1, wherein the traffic density of each block at the previous time is predicted based on the above.   The boundary traffic volume calculating means obtains a speed based on a travel time of a link adjacent to an upstream end or a downstream end of the prediction target section, obtains a traffic density of the link from the speed, and uses this traffic density The travel time predicting device according to claim 7, wherein the travel time predicting apparatus calculates the traffic volume flowing in from the end or the traffic volume flowing out from the downstream end.   The said boundary traffic calculation means uses the travel time of the said link predicted by the prediction method using the pattern matching and statistical information of travel time transition as the travel time of the said adjacent link. Travel time prediction device.   The travel time prediction device according to claim 8, wherein the boundary traffic volume calculation means uses a travel time of the link at a current time point as a travel time of the adjacent link. The travel time prediction device according to claim 1, wherein the travel time prediction means starts a travel time prediction process triggered by either of the following (1) and (2).
(1) When information about sudden events is acquired,
(2) When the travel time information of the link is acquired over a plurality of time zones, and the increase rate of the travel time in the time zones is out of the predetermined range.   The travel time prediction apparatus according to claim 1, wherein the travel time to the time before the regulation is canceled is predicted.   The travel time prediction device according to claim 12, wherein the restriction cancellation time is set according to the type of sudden event. A method for predicting travel time at a previous time of a vehicle traveling on a road,
Set the travel time prediction target section including the section where the sudden event occurred,
Dividing the prediction target section into a plurality of blocks, calculating the traffic density in each block at a certain time,
Based on the traffic density in each block, calculate the traffic volume that can move between adjacent blocks,
Based on the traffic density in each block and the traffic volume movable between the adjacent blocks, predict the traffic density of each block at the previous time,
A travel time prediction method for predicting travel time based on the traffic density of each block predicted by the traffic density prediction means.

Images