JP2008152566A - Predicted traffic information providing method, predicted traffic information providing device, on-vehicle terminal and predicted traffic information providing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce the data quantity necessary for prediction in predicted traffic information considering a non-day type factor without reducing the accuracy and also without increasing the calculation quantity. <P>SOLUTION: A prediction coefficient calculation part 202 calculates a non-day type fluctuation quantity for each day type from traffic information, and then calculates, based on the non-day type fluctuation quantity and non-day type information as a base component, a prediction coefficient that is a coefficient related to the non-day type information. A traffic information prediction part 203 calculates a predicted non-day type fluctuation quantity, based on non-day type information of a prediction object date and the prediction coefficient corresponding to the day type of the prediction object date, and performs prediction of traffic information from the predicted non-day type fluctuation quantity and day type-based statistic traffic information that is statistic traffic information for each day type. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a traffic information prediction technique.

In recent years, a traffic information system aimed at preventing traffic accidents and alleviating traffic jams using a navigation system has attracted attention, and various technologies have been provided.
As a part of such a traffic information system, a traffic information prediction technique for predicting a traffic jam at a future date and time is provided.
For traffic information prediction technology, traffic information such as travel time is acquired and accumulated from, for example, VICS (Vehicle Information and Communication System: registered trademark) and probe cars, and statistical information such as probability processing is stored in the accumulated past traffic information. A technique for calculating current or future traffic jam information (traffic information) using the processed statistical traffic information is disclosed (see, for example, Patent Document 1).
Also, a characteristic traffic information pattern (feature pattern) is calculated from past traffic information using principal component analysis, etc., and the calculated feature pattern is multiplied by a predetermined intensity to predict traffic information A technique for spatial prediction is disclosed (for example, see Patent Document 2).

Furthermore, when predicting future traffic information from past traffic information, there is a need for prediction that takes into account non-daily factors such as temperature and precipitation.
Japanese Patent Laying-Open No. 2005-122461 (paragraph 0029 to paragraph 0033, FIG. 7) Japanese Patent Laying-Open No. 2005-4668 (paragraph 0028 to paragraph 0037, FIG. 12)

However, if you want to take into account the effects of non-daily factors in statistical traffic information, statistical traffic information corresponding to each non-daily factor is required, so the amount of data required for prediction will be enormous .
In other words, if statistical traffic information covers non-daily factors, statistical traffic information will be created with a combination of daily factors and non-daily factors (weekdays / rain, weekdays / clear, weekdays / snow, etc.) The amount of data increases. Moreover, if the conditions for this combination are narrowed down, the number of parameters for creating statistical traffic information will be reduced, and the prediction accuracy will be reduced.
Also, in the feature space prediction, in principle, a pattern with a low contribution as a feature amount is discarded, so that accuracy is limited. In addition, when reflecting non-daily factors, it is necessary to take into account the characteristics of daily factors and non-daily factors in each feature pattern. However, as the number of feature patterns increases, the amount of calculation increases. The load on the computer will increase.
That is, in the feature space prediction, dozens of kinds of feature patterns in which day / non-day factors are mixed are required as patterns of day factors and non-day factors, and the amount of data and calculation amount increase.

  In view of such a background, the present invention has been made, and the present invention reduces the amount of data and the amount of calculation required for prediction in the predicted traffic information that takes into account non-daily factors such as weather and season, It aims at suppressing the fall of the prediction accuracy at the time of reducing.

  In order to solve the above problems, the present invention has been completed. That is, the traffic information prediction providing method of the present invention calculates the non-day variation for each day type from the traffic information, and based on the non-day variation information and the non-day information as a base component, Calculating a prediction coefficient that is a coefficient related to the species information, calculating a predicted non-day variation amount based on the non-daily information of the prediction target date and time and a prediction coefficient corresponding to the day type of the prediction target date and time, The traffic information is predicted from the predicted non-daily fluctuation amount and the daily statistical traffic information which is the statistical traffic information for each day type.

  According to the present invention, it is possible to reduce the amount of data and the amount of calculation necessary for the prediction in the predicted traffic information taking into account non-daily factors such as weather and season, and to suppress the decrease in the prediction accuracy when the data is reduced. Is possible.

  Next, the best mode for carrying out the present invention (referred to as “embodiment”) will be described in detail with reference to the drawings as appropriate.

FIG. 1 is a schematic diagram of a predicted traffic information providing system according to the present embodiment.
The predicted traffic information providing system 1 obtains traffic information such as travel time from a weather information center 4 that obtains weather information from a weather center, a traffic information provision center 9 such as VICS (registered trademark), or a probe car 10. The traffic information center 3 calculates the predicted traffic information based on the traffic information and the non-daily information sent from the weather information center 4 and sends the calculated predicted traffic information to the in-vehicle terminal 7 mounted on each vehicle 8. The predicted traffic information transmitted from the predicted traffic information providing device 2 is installed in the base station 6 that transmits the predicted traffic information transmitted to the in-vehicle terminal 7 and the vehicle 8, and the predicted traffic information wirelessly transmitted from the predicted traffic information providing device 2 is received. The vehicle-mounted terminal 7 for displaying is included. The traffic information center 3, the weather information center 4, and the base station 6 are connected via a communication network 5.
The traffic information center 3 includes a predicted traffic information providing device 2 that calculates predicted traffic information of a date and time at a specified link and provides the calculated predicted traffic information to the in-vehicle terminal 7 via the communication network 5 and the base station 6. is set up.

  The in-vehicle terminal 7 receives the predicted traffic information and the prediction coefficient sent from the predicted traffic information providing device 2 and predicts the link, date, time, etc., to be predicted input to the main body 71 via the input unit 73. A transmitter / receiver 74 to be sent to the traffic information providing apparatus 2, and a main body 71 that displays the predicted traffic information and prediction coefficient received by the transmitter / receiver 74 on the display unit 72 as visually recognizable information are provided.

Next, the predicted traffic information providing apparatus according to this embodiment will be described with reference to FIG. 2 while referring to FIG.
FIG. 2 is a schematic diagram of a predicted traffic information providing apparatus installed in an information center.
The predicted traffic information providing apparatus 2 represents a statistical traffic information creation unit 201 that creates day-class statistical traffic information that is statistical traffic information for each day type, and represents the degree of influence of non-day-type factors on non-day-type fluctuations. A prediction coefficient calculation unit 202 that calculates a prediction coefficient is provided.
The predicted traffic information providing apparatus 2 includes a traffic information prediction unit 203 that calculates predicted traffic information based on the daily statistical traffic information, the prediction coefficient, and the non-daily information.
Further, the predicted traffic information providing apparatus 2 includes a traffic information database (DB) 204 in which traffic information sent from the traffic information providing center 9 and the probe car 10 is stored, and temperature and precipitation sent from the weather information center 4. Non-day type information DB 205 storing non-day type factor information (non-day type information), day-type statistical traffic information DB 206 storing day-type statistical traffic information created by the statistical traffic information creating unit 201, and A prediction coefficient DB 207 in which the prediction coefficient calculated by the prediction coefficient calculation unit 202 is stored is provided.
Possible non-daily factors include seasonal components, bargain periods at nearby stores, and vacation events in addition to the weather.

The predicted traffic information providing apparatus 2 also includes a calendar DB 208 that stores a daily calendar that associates a day type with a date, a link that stores a list of link numbers uniquely assigned to links, and the like. The database includes a DB 209 and a day type DB 210 in which a list of types of day types is stored.
The predicted traffic information providing apparatus 2 includes a traffic information acquisition unit 212 that acquires the traffic information sent from the traffic information providing center 9 or the probe car 10 and stores the acquired traffic information in the traffic information DB 204.
Further, the predicted traffic information providing device 2 acquires the non-day type information sent from the weather information center 4, and stores the acquired non-day type information in the non-day type information DB 205, an in-vehicle information A predicted traffic information request receiving unit 214 that receives a predicted traffic information request including the link number, date, and time of a link to be predicted from the terminal 7 and sends the request to the traffic information prediction unit 203 is provided.
The predicted traffic information providing apparatus 2 includes a predicted traffic information transmission unit 215 that transmits the predicted traffic information calculated by the traffic information prediction unit 203 to the in-vehicle terminal 7.
Further, the predicted traffic information providing device 2 is connected to the communication network 5 and receives information sent from the weather information center 4 and the like, and sends it to the traffic information acquisition unit 212 and the non-daily information acquisition unit 213. A communication interface unit 211 that sends the predicted traffic information calculated by the information prediction unit 203 to the in-vehicle terminal 7 is provided.
Each unit (201 to 203, 211 to 215, 71, 72, 73) in the predicted traffic information providing apparatus 2 and the in-vehicle terminal 7 shown in FIGS. 1 and 2 is stored in a storage device such as a hard disk mounted on the computer. The developed program is developed in a storage means such as a RAM (Random Access Memory) or a flash memory, and the developed program is implemented by a CPU (Central Processing Unit) (not shown).

FIG. 3 is a diagram illustrating a configuration example of data in the traffic information DB.
In the traffic information DB 204, the link number, which is a number uniquely assigned to each link, the date when the traffic information was acquired, and the traffic information (travel time) on the link corresponding to the link number and the date change over time. It is saved as data. In FIG. 3, the time transition of the traffic information is shown in the form of a graph for easy viewing, but actually, for example, the travel time value at every predetermined time (for example, every 5 minutes) is stored. . In the non-daily information of FIG. 4 and the daily statistical traffic information (daily statistical travel time) of FIG. 5 described below, the data is stored in the same format.

FIG. 4 is a diagram illustrating a configuration example of data in the non-daily information DB.
The non-day type information DB 205 stores a date, elements of non-day type factors such as temperature and precipitation, and non-day type information. Non-day type information is data of temporal transition of elements on a corresponding date. For example, for the temperature on July 15, 2006, the daily temperature transition on the corresponding date is stored as data.

FIG. 5 is a diagram illustrating a configuration example of data in the daily statistical traffic information DB.
The day type statistical traffic information DB 206 stores the day type and day type statistical traffic information (day type statistical travel time) corresponding to the day type. The calculation method of the daily statistical traffic information will be described later with reference to FIG.

FIG. 6 is a diagram showing a configuration example of a daily calendar in the calendar DB.
As shown in Fig. 6, the daily calendar includes the second day of weekdays, the fifty days (every day, the fifth and tenth day of the month), three consecutive holidays, etc. It is attached. A plurality of day types usually correspond to one date, but more specific day types are preferentially associated.
For example, in FIG. 6, there is a possibility that July 15th is associated with day types such as Saturday, holidays, three consecutive holidays, three consecutive holidays, and fifty days, among which the most important for traffic relations. And, special 50th day will be associated.

Next, the outline of the calculation method of the non-day change amount which is the change amount due to the non-day type factor in the traffic information is shown.
In the following embodiment, the traffic information so far will be described as travel time in order to provide a more specific description. That is, the traffic information is converted into travel time, the statistical traffic information is converted into statistical travel time, and the daily statistical traffic information is converted into daily statistical travel time. However, since the traffic information is not limited to the travel time, for example, the average speed for each link, the traffic congestion level (leveled for each road type speed), and the like may be used as the traffic information.

With reference to FIG. 7 and FIG. 8, the outline of the preparation method of the prediction coefficient in this embodiment and the preparation method of prediction travel time is shown.
FIG. 7 is a diagram showing an outline of a procedure for calculating the non-daily variation amount.
As an example, a case will be described where data on July 15, 2006 (hereinafter referred to as July 15) is used.
Referring to the day type calendar in FIG. 6, it can be seen that the day type on July 15 is fifty days. Furthermore, it is assumed that the weather on this day was raining all day.
T1 (t) in FIG. 7 represents a temporal transition of travel time on a certain link on July 15th.
Further, T2 (t) represents a temporal transition of the day-type statistical traffic information data, which is a characteristic travel time regarding the day type. A method for calculating the daily statistical travel time will be described with reference to FIG.
Then, T3 (t) obtained by subtracting the day-type statistical travel time T2 (t) from the travel time T1 (t) is the non-day type fluctuation amount on July 15.
T3 (t) = T1 (t) -T2 (t) Formula (1)

Next, the outline of the calculation process of the prediction coefficient will be described with reference to FIG.
FIG. 8 is a diagram illustrating an example of non-daily information.
FIG. 8 also assumes July 15 data following FIG.
As shown in FIG. 8, if the curve representing the temperature fluctuation on July 15 is X1 (t) and the curve representing the fluctuation in precipitation is X2 (t), the non-day variation T3 (t ) Is based on X1 (t), X2 (t).
T3 (t) = a (t) X1 (t) + b (t) X2 (t) + Expression (2)
It can be expressed as.
Here, a (t), b (t),... Are referred to as prediction coefficients. The prediction coefficient can be obtained by performing principal component analysis or regression analysis on T3 (t).

If this forecast coefficient can be used to obtain the predicted temperature transition X1 (t) ', forecast precipitation transition X2 (t)',... Non-daily fluctuation amount (predicted non-daily fluctuation amount) T3 (t) ′ predicted on the corresponding date
T3 (t) ′ = a (t) X1 (t) ′ + b (t) X2 (t) ′ + (3)
It can be expressed as.
If the date type of the corresponding date is known, the corresponding day-type statistical travel time T2 (t) ′ is obtained from the day-type statistical traffic information DB 206, so that the predicted travel time T1 (t ) ',
T1 (t) ′ = T2 (t) ′ + T3 (t) ′ Equation (4)
Can be obtained as
Hereinafter, a set of a (t), b (t),... Is described as a set of prediction coefficients.

(Creation processing of daily statistical travel time)
Prior to the processing, data acquisition processing of the traffic information DB 204 and the non-day type information DB 205 will be described with reference to FIGS. 1 and 2.
The traffic information acquisition unit 212 of the predicted traffic information providing apparatus 2 acquires the travel time for each link from the traffic information provision center 9 or the probe car 10 via the communication interface unit 211, and acquires the traffic information obtained by acquiring the travel time and the like. The unit 212 stores this in the traffic information DB 204.
The weather information center 4 sends non-daily information such as temperature and precipitation for each date and time to the predicted traffic information providing device 2 of the traffic information center 3 via the communication network 5.
The non-day type information acquisition unit 213 of the predicted traffic information providing apparatus 2 acquires non-day type information via the communication interface unit 211 and stores it in the non-day type information DB 205.

With reference to FIG. 1 and FIG. 2, the day-type statistical travel time creation process stored in the day type DB 210 will be described along FIG. 9.
FIG. 9 is a flow showing the flow of the day-type statistical travel time creation process.
This process is performed when the user inputs an update of the daily statistical traffic information DB 206 or at a constant cycle (for example, every year).
First, the day type of the day-type statistical travel time to be created is set (S101). As for the setting method, the name of the day type or the like is stored in the day type DB 210 in advance, the statistical traffic information creation unit 201 may read the day type data from the head of the day type DB 210, or the input unit 73 Then, the day type may be set by inputting the name of the day type or the like to the statistical traffic information creating unit 201.

Next, the statistical traffic information creation unit 201 acquires a day type calendar from the calendar DB 208, and acquires all dates associated with the set day type from the acquired day type calendar (S102).
Next, the link number of the target link is set (S103). The link number setting method is performed, for example, by the statistical traffic information creation unit 201 reading the link number from the top of the link DB 209.
Then, the statistical traffic information creation unit 201 acquires the corresponding travel time from the traffic information DB 204 based on the date acquired in step S102 and the link number set in step S103 (S104).

Next, the statistical traffic information creation unit 201 determines whether or not all travel times related to the date acquired in step S102 have been acquired (S105). For example, the statistical traffic information creating unit 201 sets a flag for the date acquired in step S102 and the travel time of the traffic information DB 204 corresponding to the link number set in step S103. The statistical traffic information creation unit 201 makes a determination by referring to this flag.
As a result of the determination, if it is determined that all travel times related to the acquired date have not been acquired (S105 → No), the process returns to step S104.
As a result of the determination, if it is determined that all travel times related to the acquired date have been acquired (S105 → Yes), the process proceeds to step S106.
Next, the statistical traffic information creation unit 201 performs a process such as averaging the travel time for every predetermined time set in advance for all the acquired links, so that the statistical travel of the corresponding day type is performed. Time is calculated (S106). The statistical travel time calculated here is the daily statistical travel time.

Then, the statistical traffic information creation unit 201 determines whether or not the daily statistical travel time has been calculated for all links (S107). The determination method is performed by, for example, determining whether the link number of the currently read link is the last data of the link DB 209 by the statistical traffic information creating unit 201 with reference to the link DB 209.
As a result of the determination, if it is determined that the daily statistical travel time has not been calculated for all the acquired links (S107 → No), the process returns to step S103.
As a result of the determination, if it is determined that the daily statistical travel time has been calculated for all the acquired links (S107 → Yes), the process proceeds to step S108.
Next, the statistical traffic information creation unit 201 determines whether or not the day-type statistical travel time has been calculated for all day types stored in the day type DB 210 (S108). The determination method is performed, for example, when the statistical traffic information creation unit 201 determines whether or not the name of the currently read day type is the last data in the day type DB 210.
As a result of the determination, when it is determined that the day-type statistical travel time has not been calculated for all day types (S108 → No), the processing returns to step S101.
As a result of the determination, when it is determined that the day-type statistical travel time is calculated for all day types (S108 → Yes), the process proceeds to step S109.
Then, the statistical traffic information creation unit 201 stores the created daily type statistical travel time in the daily type statistical traffic information DB 206 in association with the corresponding day type (S109).

(Prediction coefficient calculation process)
With reference to FIG. 1 and FIG. 2, the calculation process of the prediction coefficient will be described along FIG.
FIG. 10 is a flowchart showing the flow of the prediction coefficient calculation process.
First, the day type of the day-type statistical travel time to be created is set (S201). As for the setting method, the name of the day type or the like is stored in the day type DB 210 in advance, the statistical traffic information creation unit 201 may read the day type data from the head of the day type DB 210, or the input unit 73 Then, the day type may be set by inputting the name of the day type or the like to the statistical traffic information creating unit 201.

Next, the statistical traffic information creation unit 201 acquires a day type calendar from the calendar DB 208, and acquires all dates associated with the set day type from the acquired day type calendar (S202).
Next, the link number of the target link is set (S203). The link number setting method is performed, for example, by the statistical traffic information creation unit 201 reading the link number from the top of the link DB 209.
Then, the statistical traffic information creation unit 201 acquires the corresponding travel time from the traffic information DB 204 based on the date set in step S202 and the link number set in step S203 (S204).

Next, the prediction coefficient calculation unit 202 acquires the daily statistical travel time from the daily statistical traffic information DB 206 based on the day type set in step S201 and the link number set in step S203 of the target link. (S205).
Then, the prediction coefficient calculation unit 202 calculates the non-daily variation amount of the target link by subtracting the day-type statistical travel time acquired in step S205 from the travel time acquired in step S204 (S206). .
In this embodiment, in the day type calendar, the date and the day type are associated with each other on a one-to-one basis, so that one day type statistical travel time is subtracted from the travel time. A date usually has multiple date types. Therefore, for example, a pattern of temporal transition of travel time for each day type in the corresponding link is artificially created, and data such as taking the average of the day type pattern corresponding to the corresponding date is used as the daily statistical travel time. Also good.

Next, the prediction coefficient calculation unit 202 determines whether or not the non-daily fluctuation amount of the target link has been calculated for all dates of the relevant day type (S207). The determination method includes, for example, the same method as the determination performed in step S105 of FIG.
As a result of the determination, when it is determined that the non-daily variation amount of the target link has not been calculated for all dates (S207 → No), the process returns to step S206.
As a result of the determination, if it is determined that the non-daily fluctuation amount of the target link has been calculated for all dates (S207 → Yes), the process proceeds to step S208.
Then, the prediction coefficient calculation unit 202 acquires non-day type information corresponding to the target date from the non-day type information DB 205 (S208).

  Next, the prediction coefficient calculation unit 202 uses principal component analysis, regression analysis, or the like based on the non-day type information acquired in step S208 and the non-day type fluctuation amount calculated in step S206. A set is calculated (S209).

The prediction coefficient calculation unit 202 stores the calculated set of prediction coefficients in the prediction coefficient DB 207 (S210).
Then, the prediction coefficient calculation unit 202 determines whether or not prediction coefficients have been calculated for all target links (S211). As a determination method, for example, a flag is set to the link number of the link DB 209 that has finished calculating a set of prediction coefficients, and the prediction coefficient calculation unit 202 refers to the link DB 209 to determine whether all the links are flagged. Determine by searching.
As a result of the determination, when it is determined that the prediction coefficient has not been calculated for all links (S211 → No), the processing returns to step S203.
As a result of the determination, if it is determined that the prediction coefficient has been calculated for all links (S211 → Yes), the process proceeds to step S212.
And the prediction coefficient calculation part 202 determines whether the prediction coefficient was calculated about all the day types (S212). The determination method is the same as the method in step S211.
As a result of the determination, when it is determined that the prediction coefficient has not been calculated for all day types (S212 → No), the process returns to step S201.
As a result of the determination, when it is determined that the prediction coefficient has been calculated for all the day types (S212 → Yes), the process ends.

(Traffic information prediction process)
The traffic information prediction process will be described along FIG. 11 with reference to FIGS. 1 and 2.
FIG. 11 is a flow showing the flow of traffic information prediction processing.
First, the link number, date, time, and the like of a link to be predicted are input to the in-vehicle terminal 7 via the input unit 73 (S301). The link number of the link to be predicted may be directly input. For example, when the map is displayed on the display unit 72 and the user specifies the road to be predicted, the in-vehicle terminal 7 corresponds to the specified road. A method such as conversion to a link number may also be used.
A plurality of link numbers may be input. Moreover, you may identify the link number of a link from an area, a route, a point (coordinates), etc.

Next, the in-vehicle terminal 7 transmits a predicted traffic information request to the predicted traffic information providing apparatus 2 via the base station 6 and the communication network 5 (S302). The predicted traffic information request includes the link number, date, and time of the link to be predicted.
Next, the predicted traffic information request receiving unit 214 receives the predicted traffic information request via the communication interface unit 211. Then, the predicted traffic information request receiving unit 214 sends the received predicted traffic information request to the traffic information prediction unit 203, and the traffic information prediction unit 203 acquires the link number, date, time, etc. included in the predicted traffic information request. (S303).
Then, the traffic information prediction unit 203 acquires the day type calendar from the calendar DB 208, refers to the acquired day type calendar, searches for which day type the date input in step S303 corresponds to, and The day type to be acquired is acquired (S304).

Then, the traffic information prediction unit 203 refers to the daily statistical traffic information DB 206 based on the link number acquired in step S303 and the day type acquired in step S304, and acquires the corresponding daily statistical travel time. (S305).
Next, the traffic information prediction unit 203 refers to the non-day type information DB 205 based on the date input in step S303, and acquires all the non-day type information on the date (S306). The non-daily species information at this time is data such as expected temperature and expected precipitation on the date.
Then, the traffic information prediction unit 203 refers to the prediction coefficient DB 207 and acquires a set of prediction coefficients for the corresponding day type (S307). There are only a set of prediction coefficients for the date when the prediction coefficient is calculated. For example, the set of prediction coefficients is narrowed down to one as follows. First, the traffic information prediction unit 203 narrows down the set of prediction coefficients based on the link location, time, day type, and the like. Here, if it is finally possible to narrow down to one set of prediction coefficients, the set of prediction coefficients is used. If it is not possible to narrow down to one set, the set of prediction coefficients to be used is narrowed down to one set by, for example, taking the average for each prediction coefficient in the remaining set of prediction coefficients.

Next, the traffic information prediction unit 203 calculates a predicted non-day variation amount for the corresponding date based on the acquired set of prediction coefficients and non-day information (S308).
Then, the traffic information prediction unit 203 adds the predicted non-daily variation amount multiplied by an appropriate correction coefficient α to the daily statistical travel time acquired in step S305 to calculate the predicted travel time for the corresponding link. (S309).

Here, the correction coefficient α will be described.
The following calculation is processing performed by the traffic information prediction unit 203.
The predicted travel time T1 (t) ′, which is the predicted travel time in the corresponding link, is obtained by inserting a correction coefficient α, which is a parameter representing the influence of the predicted non-daily fluctuation amount, into the following formula (4). It can be rewritten as an expression.
T1 (t) ′ = T2 (t) ′ + α · T3 (t) ′ Equation (5)
Here, when both sides of the equation (5) are divided by T2 (t) ′,
T1 (t) ′ / T2 (t) ′ = 1 + α · T3 (t) ′ / T2 (t) ′ Equation (6)
When the denominator on the left side is divided by the link length L of the corresponding link, Equation (7) is obtained.
(T1 (t) ′ / L) / (T2 (t) ′ / L) = V2 (t) ′ / V1 (t) ′ = 1 + α · T3 (t) ′ / T2 (t) ′ Equation (7)

Here, V1 (t) ′ is a predicted speed based on the calculated predicted travel time, and V2 (t) ′ is a predicted speed based on the daily statistical travel time.
When the equation (7) is further modified as follows, the equation (8) is obtained.
V1 (t) ′ = V2 (t) ′ / (1 + α · T3 (t) ′ / T2 (t) ′) = β (t) · V2 (t) ′ Equation (8)
β (t) will be described later.

Then, the traffic information prediction unit 203 first calculates V1 (t) ′ by setting the correction coefficient α to 1, and determines the calculated value of V1 (t) ′. The traffic information prediction unit 203 determines whether or not this value is an abnormal value (for example, 200 km / h), that is, whether or not it is equal to or lower than a predetermined speed threshold, and is equal to or lower than the speed threshold. If so, T1 (t) ′ is used as the estimated travel time. If V1 (t) ′ is larger than the speed threshold value, the correction coefficient α is subtracted by a predetermined value between 0 and 1, V1 (t) ′ is calculated again, and the calculated V1 It is determined whether (t) ′ is an abnormal value.
This process is repeated until the correction coefficient α is updated until V1 (t) ′ becomes equal to or lower than the speed threshold value.
Here, although the correction coefficient α is a value that does not depend on time, the correction coefficient α may be set to a time-dependent value α (t) by performing the calculation procedure of the correction coefficient α every time.

Next, the traffic information prediction unit 203 determines whether or not the predicted travel time has been calculated for the links (prediction target links) corresponding to all the link numbers acquired in step S303 (S310).
As a result of the determination, when it is determined that the predicted travel time has not been calculated for all the prediction target links (S310 → No), the process returns to step S305.
As a result of the determination, when it is determined that the predicted travel time has been calculated for all the prediction target links (S310 → Yes), the process ends.

  The calculated predicted travel time and the used prediction coefficient are sent to the predicted traffic information transmitting unit 215 together with the link number, date, and time, and the predicted traffic information transmitting unit 215 includes the communication interface unit 211, the communication network 5, and the base station. 6, the calculated estimated travel time is sent to the in-vehicle terminal 7.

Here, β (t) in Expression (8) will be described.
β (t) is an intensity parameter that indicates the degree of influence of non-day type factors on the predicted speed V2 (t) ′ (day type statistical traffic information) based on the day type statistical travel time.

The traffic information prediction unit 203 may store the predicted speed V2 (t) ′ based on the daily statistical travel time and the intensity parameter β (t) in an intensity parameter storage unit (not shown).
FIG. 12 is an example of a graph showing the relationship between the predicted speed based on the daily statistical travel time and the intensity parameter in a certain link.
In FIG. 12, the horizontal axis indicates the predicted speed (V2 (t) ′ in equation (8)) based on the daily statistical travel time, and the vertical axis indicates the intensity parameter β (t).
The traffic information predicting unit 203 stores the predicted speed V2 (t) ′ based on the daily statistical travel time and the intensity parameter β (t) in association with each other in advance in the intensity parameter storage unit (not shown) as shown in FIG. By doing so, it is easy to calculate the predicted speed V1 (t) ′ (predicted traffic information) based on the predicted travel time of Expression (8).
For example, the traffic information predicting unit 203 may be a traffic information predicting unit when the predicted speed based on the daily statistical travel time input via the input unit (not shown) or the communication interface unit 211 (see FIG. 1) is V2c (t). In step 203, an intensity parameter βc (t) (see FIG. 12) is obtained from an intensity parameter storage unit (not shown). The traffic information prediction unit 203 uses this strength parameter βc (t) to calculate the predicted speed V1w (t ′) based on the predicted travel time at time t ′ as V1w (t ′) = βc (t ′) · V2c (t ′ ), The processing time can be shortened.

  In this way, the strength parameter β (t) for the predicted speed V2 (t) ′ based on the non-daily statistical travel time may be provided for each link, but the predicted speed V2 (t based on the road type or daily statistical travel time ) ′, The representative strength parameter β (t) is determined and used for roads having the same tendency with respect to the strength parameter β (t), that is, roads having similar shapes in the graph of FIG. By doing so, it is possible to reduce the amount of data stored in the intensity parameter storage unit (not shown). Furthermore, the result of the route search using the predicted travel time described later with reference to FIGS. 13 and 14 and the response time until the traffic congestion point is displayed and output can be shortened.

In addition, the traffic information prediction unit 203 selects a link that is greatly affected by a change in the predicted speed due to a non-daily factor, for example, as a link in which the difference between the maximum value and the minimum value of the predicted speed is a predetermined threshold value or more. Only for this link, the predicted speed V2 (t) ′ based on the non-daily statistical travel time and the intensity parameter β (t) may be associated with each other and stored in an intensity parameter storage unit (not shown).
Even in this way, it is possible to reduce the amount of data stored in the intensity parameter storage unit (not shown), and further, as a result of the route search using the predicted travel time described later with reference to FIGS. It is possible to shorten the response time until the traffic congestion point is displayed and output.

  Here, the predicted speed V2 (t) ′ based on the day-by-day statistical travel time and the intensity parameter β (t) are associated with each other and stored in an intensity parameter storage unit (not shown), and an input unit (not shown) or the communication interface unit 211 ( The traffic information prediction unit 203 acquires the intensity parameter β (t) based on the predicted speed V2 (t) ′ based on the daily statistical travel time input via FIG. The intensity parameter β (t) may be calculated using this parameter.

Next, referring to FIG. 1, a screen example in the in-vehicle terminal of the present embodiment will be described along FIGS. 13 and 14.
FIG. 13 is a diagram illustrating a screen example on the display unit of the in-vehicle terminal according to the present embodiment.
The main body 71 of the in-vehicle terminal 7 acquires the predicted travel time, the prediction coefficient, the link number, the date, and the time sent from the predicted traffic information providing apparatus 2 via the transceiver 74. And the main-body part 71 displays the estimated travel time exceeding the predetermined value on the link corresponding to the acquired link number on the display part 72 as traffic jam information.
For example, when the predicted travel time calculated by the traffic information prediction process described in FIG. 11 is displayed, the largest prediction coefficient in the set of sent prediction coefficients is used as a factor of the predicted travel time such as traffic congestion. To that effect.
For example, assuming that traffic congestion information is displayed based on the predicted travel time, the traffic congestion in the section A has the largest b (t) among the prediction coefficients calculated in the prediction coefficient calculation process of FIG. Suppose it was a value. If each prediction coefficient follows the description with reference to FIG. 8, b (t) is a coefficient related to precipitation. Therefore, the main body 71 determines that the cause of the traffic jam (predicted traffic jam) predicted in the section A is the influence of precipitation, that is, the rain. Accordingly, the main body 71 displays that the predicted traffic congestion corresponding to the display unit 72 is due to rain.
In the predicted traffic jam in the section B, it is assumed that a (t) and b (t) are the largest values. As in the section A, if each prediction coefficient follows the description with reference to FIG. 8, a (t) is a coefficient related to temperature, and b (t) is a coefficient related to precipitation. Therefore, in the section B, it is considered that temperature and precipitation are the cause of traffic congestion. Accordingly, for example, if the season is winter, the main body 71 determines that the cause of the predicted traffic congestion is snowfall. Therefore, the main body 71 displays that the predicted traffic congestion corresponding to the display unit 72 is an influence of snowfall.
In FIG. 13, “S” represents a departure point, and “G” represents a destination point.

FIG. 14 is a diagram illustrating a screen example on the display unit of the in-vehicle terminal according to the present embodiment.
FIG. 14 is an example in which the in-vehicle terminal 7 or the traffic information center 3 has a function of route guidance, and the display unit 72 has a function of displaying a plurality of routes obtained as a result of route guidance and the predicted congestion level.
Since a technique for calculating a plurality of routes and the predicted congestion level by route guidance is known, the description thereof is omitted here.
As shown in FIG. 14, routes C, D, and E are calculated by the multiple route calculation function, and the degree of traffic congestion is calculated based on the predicted traffic information sent from the predicted traffic information providing device. This is an example of a screen displayed in the small windows c, d, e.
And the factor of the traffic congestion in each route is displayed from the prediction coefficient of each prediction traffic information.
For example, in the predicted traffic jam on the route C, if the prediction coefficient b (t) is the largest, the fact that a traffic jam due to rain is expected is displayed as shown in FIG. Then, in the predicted traffic congestion in the routes D and E in the illustrated city, if the prediction coefficients a (t) and b (t) are large and the season is winter, as in FIG. Displays that traffic congestion due to snow is expected.

Further, the cause of the traffic jam may be displayed when the mouse pointer is over the route, and the factor in the corresponding route is displayed.
At this time, for example, the main body 71 acquires a link number associated with the link of the corresponding route by overlaying the mouse pointer on the route. And the main body part 71 corresponds by the main body part 71 collating the predicted travel time, the prediction coefficient, the link number, the date, the time, and the acquired link number sent from the predicted traffic information providing device 2. By specifying the estimated travel time and prediction coefficient of the link, the corresponding factors are identified and displayed.

In the present embodiment, the predicted traffic information providing apparatus 2 is installed in the traffic information center 3, but the present invention is not limited to this and may be installed in the in-vehicle terminal 7.
In this case, the predicted traffic information request reception unit 214 and the predicted traffic information transmission unit 215 are deleted, and the prediction coefficient calculated by the prediction coefficient calculation unit 202 and the information related to the predicted traffic information calculated by the traffic information prediction unit 203 are displayed. A display processing unit to be displayed on the unit 72 is added.
9 is the same as the process shown in FIGS. 9 to 11 except that the process shown in the traffic information prediction process in FIG. 11 from the day-type statistical travel time creation process in FIG.

Moreover, among each component shown in FIG. 2, the traffic information prediction unit 203 and the display processing unit described above may be mounted on the in-vehicle terminal 7 and other components may be installed in the information center.
In this case, the daily statistical travel time creation process in FIG. 9 and the prediction coefficient calculation process in FIG. 10 are performed in the traffic information center 3, and the traffic information prediction process in FIG. 11 is performed in the in-vehicle terminal 7.
In this case, the in-vehicle terminal 7 requests the information center for the day type, non-day type statistical traffic information, prediction coefficient set, non-day type information, etc. necessary for the traffic information prediction process shown in FIG. Each data described above may be downloaded from the information center to the in-vehicle terminal 7, and the in-vehicle terminal 7 may perform the traffic information prediction process based on the downloaded data.

According to the present embodiment, at least statistical traffic information for each day type, daily non-day type information, and prediction coefficient are used to perform traffic information prediction. There is no need to prepare statistical traffic information that reflects non-day type factors for each type of statistical traffic, unlike traffic information prediction that reflects the influence of factors, and many features such as feature space prediction. Since there is no need to prepare a pattern, the amount of data necessary for prediction can be reduced.
Furthermore, since the amount of calculation at the time of traffic information prediction in this embodiment is based on the number of types of non-day type information, compared with feature space prediction based on the number of feature patterns in which day type / non-day type factors are mixed. Traffic information can be predicted with a small amount of calculation.
Moreover, the responsiveness of an apparatus can be improved by reducing the calculation amount at the time of predicting traffic information with day type statistical traffic information and a prediction coefficient.

It is the schematic of the prediction traffic information provision system which concerns on this embodiment. It is the schematic of the prediction traffic information provision apparatus installed in an information center. It is a figure which shows the structural example of the data preserve | saved in traffic information DB. It is a figure which shows the structural example of the data in non-daily species information DB. It is a figure which shows the structural example of the data in daily classification statistical traffic information DB. It is a figure which shows the structural example of the day type | mold calendar in calendar DB. It is a figure which shows the outline of the procedure which calculates a non-day variation amount. It is a figure which shows an example of non-day type information. It is a flow which shows the flow of creation processing of a day type statistical travel time. It is a flow which shows the flow of the calculation process of a prediction coefficient. It is a flow which shows the flow of traffic information prediction processing. It is an example of the graph which shows the relationship between the prediction speed by a day type statistical travel time in a certain link, and an intensity parameter. It is a figure which shows the example of a screen in the display part of the vehicle-mounted terminal concerning this embodiment. It is a figure which shows the example of a screen in the display part of the vehicle-mounted terminal concerning this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Predictive traffic information provision system 2 Predictive traffic information provision apparatus 3 Traffic information center 4 Weather information center 5 Communication network 6 Base station 7 In-vehicle terminal 8 Vehicle 9 Traffic information provision center 10 Probe car 71 Main part 72 Display part 73 Input part 74 Transmission / reception Machine 201 Statistical traffic information creation unit 202 Prediction coefficient calculation unit 203 Traffic information prediction unit 204 Traffic information DB
205 Non-Japanese information DB
206 Daily statistical traffic information DB
207 Prediction coefficient DB
208 Calendar DB
209 Link DB
210 day DB
211 Communication Interface Unit 212 Traffic Information Acquisition Unit 213 Non-Japanese Information Acquisition Unit 214 Predictive Traffic Information Request Receiving Unit 215 Predictive Traffic Information Transmitting Unit

Claims (16)

A predicted traffic information providing method in a predicted traffic information providing apparatus for predicting traffic information,
Traffic information that can be acquired for each day type sent from the traffic information providing center or probe car, non-day type information that is information on factors other than day type including weather data sent from the non-day type information center, and It has a storage unit that holds information that associates an arbitrary date and day type,
The predicted traffic information providing method executes a first process and a second process,
The first process includes
From the storage unit, obtain traffic information for each day type,
Based on the acquired traffic information for each day type, calculate the day type statistical traffic information which is the statistical traffic information for each day type,
The calculated day type statistical traffic information and the corresponding day type are stored in association with each other,
Based on the traffic information and the day-type statistical traffic information, a non-day variation amount that is a variation amount of traffic information not depending on the day type is calculated,
Non-Japanese type information corresponding to the target date is acquired from the storage unit,
Based on the calculated non-daily variation amount and the acquired non-daily information, a process of calculating a prediction coefficient that is a weight of the non-daily information,
The second process includes
When the prediction target link and the prediction target date and time are specified via the input unit,
Obtaining the day type corresponding to the prediction target date and time from the storage unit;
The day-type statistical traffic information corresponding to the acquired day type is acquired from the storage unit,
Obtain non-daily information of the prediction target date and time from the storage unit,
Based on the calculated prediction coefficient and the acquired non-day type information, calculate a predicted non-day variation amount at the prediction target date and time,
A predicted traffic information providing method, wherein predicted traffic information is calculated based on the predicted non-daily fluctuation amount and the acquired daily statistical traffic information.   The method for providing predicted traffic information according to claim 1, wherein the calculation of the prediction coefficient is a principal component analysis or regression analysis of the non-daily variation amount based on non-daily information.   The method for providing predicted traffic information according to claim 1, wherein, in calculating the predicted traffic information, the predicted non-daily variation amount is multiplied by an intensity that is a weight.   Based on the predicted traffic information, the predicted speed at the corresponding link is calculated, and based on the predicted speed, it is determined whether the predicted traffic information is an abnormal value. The predicted traffic information providing method according to claim 3, wherein the value is reduced. Providing predicted traffic information in an in-vehicle terminal that receives information including predicted traffic information calculated by the predicted traffic information providing device from a predicted traffic information providing device that executes the predicted traffic information providing method according to claim 1. A method,
The in-vehicle terminal has a main body for processing information and a display for displaying information,
The main body of the in-vehicle terminal is
A method for providing predicted traffic information, comprising: displaying the received predicted traffic information on a display unit. The information sent to the in-vehicle terminal includes a prediction coefficient used to calculate the prediction coefficient,
The main body of the in-vehicle terminal is
Of the sent prediction coefficients, determine the largest prediction coefficient,
The predicted traffic information providing method according to claim 5, wherein non-daily information corresponding to the largest prediction coefficient is displayed on the display unit. A predicted traffic information providing method in a predicted traffic information providing apparatus for predicting traffic information,
It has a storage unit that holds daily statistical traffic information that is statistical traffic information for each day type, and non-day type fluctuation amount that is the fluctuation amount of traffic information that does not depend on the day type,
From the storage unit, obtain the daily statistical traffic information and the non-daily fluctuation amount,
Calculating an intensity parameter that is an intensity value of the non-daily fluctuation amount with respect to the daily statistical traffic information;
The day-type statistical traffic information and the calculated intensity parameter are stored in a storage unit in association with each other,
When the daily statistical traffic information is input via the input unit or the communication unit, the strength parameter corresponding to the input daily statistical traffic information is acquired from the storage unit,
A predictive traffic information providing method, wherein predictive traffic information is calculated from the input daily statistical traffic information and the acquired intensity parameter. A predictive traffic information providing device that predicts traffic information,
Traffic information that can be acquired for each day type sent from the traffic information providing center or probe car, non-day type information that is information on factors other than day type including weather data sent from the non-day type information center, and A storage unit that stores information in which an arbitrary date and day type are associated;
From the storage unit, traffic information is acquired for each day type, and based on the acquired traffic information for each day type, day type statistical traffic information that is statistical traffic information for each day type is calculated, and the calculated date Statistical traffic information creation unit having a function of storing the type statistical traffic information and the corresponding day type in association with each other,
Based on the traffic information and the day-type statistical traffic information, the non-day-type fluctuation amount that is the fluctuation amount of the traffic information not depending on the day-type is calculated, and the non-day-type information corresponding to the target date, A prediction coefficient calculation unit that has a function of calculating a prediction coefficient that is a weight of the non-daily information, based on the calculated non-daily variation amount and the acquired non-daily information, obtained from the storage unit When,
When the prediction target link and the prediction target date and time are specified via the input unit, the day type corresponding to the prediction target date and time is acquired from the storage unit, and the day type statistical traffic corresponding to the acquired day type Information is acquired from the storage unit, non-daily information on the prediction target date and time is acquired from the storage unit, and the prediction target date and time is based on the calculated prediction coefficient and the acquired non-daily information. A traffic information prediction unit having a function of calculating predicted traffic information based on the predicted non-daily fluctuation amount and the acquired day-type statistical traffic information;
A predictive traffic information providing apparatus characterized by comprising:   The predicted traffic information providing apparatus according to claim 8, wherein the prediction coefficient calculation unit further has a function of performing principal component analysis or regression analysis on the non-daily variation amount based on non-daily information.   The predicted traffic information providing apparatus according to claim 8, wherein the traffic information prediction unit further has a function of multiplying the predicted non-daily variation amount by an intensity that is a weight.   The traffic information prediction unit calculates a predicted speed at a corresponding link based on the predicted traffic information, determines whether the predicted traffic information is an abnormal value based on the predicted speed, and outputs an abnormal value. If it is, it has further the function to make the said intensity | strength value small, The prediction traffic information provision apparatus of Claim 10 characterized by the above-mentioned. A prediction traffic information providing device that is mounted on a vehicle and predicts traffic information,
A storage unit that holds the daily statistical traffic information that is the statistical traffic information for each day type, and the non-daily fluctuation amount that is the fluctuation amount of the traffic information that does not depend on the day type,
From the storage unit, obtain the daily statistical traffic information and the non-daily fluctuation amount,
Calculating an intensity parameter that is an intensity value of the non-daily fluctuation amount with respect to the daily statistical traffic information;
The day-type statistical traffic information and the calculated intensity parameter are stored in a storage unit in association with each other,
When the daily statistical traffic information is input via the input unit or the communication unit, the strength parameter corresponding to the input daily statistical traffic information is acquired from the storage unit,
A predicted traffic information providing apparatus, comprising: a traffic information prediction unit that calculates predicted traffic information from the input daily type statistical traffic information and the acquired intensity parameter. An in-vehicle terminal that receives information including predicted traffic information calculated by the predicted traffic information providing device from the predicted traffic information providing device according to any one of claims 8 to 11.
The in-vehicle terminal includes a main body for processing information and a display for displaying information,
The main body of the in-vehicle terminal is
An in-vehicle terminal characterized in that the received predicted traffic information is displayed on a display unit. The information sent to the in-vehicle terminal includes a prediction coefficient used to calculate the prediction coefficient,
The main body of the in-vehicle terminal is
Of the sent prediction coefficients, determine the largest prediction coefficient,
The in-vehicle terminal according to claim 13, wherein non-daily information corresponding to the largest prediction coefficient is displayed on the display unit.   The vehicle-mounted terminal carrying the prediction information provision apparatus as described in any one of Claims 8-11. A predicted traffic information providing system comprising: a predicted traffic information providing device that calculates data necessary for traffic information prediction; and an in-vehicle terminal that obtains data calculated by the predicted traffic information and predicts traffic information. And
The predicted traffic information providing device includes:
Traffic information that can be acquired for each day type sent from the traffic information providing center or probe car, non-day type information that is information on factors other than day type including weather data sent from the non-day type information center, and A storage unit that stores information in which an arbitrary date and day type are associated;
From the storage unit, traffic information is acquired for each day type, and based on the acquired traffic information for each day type, day type statistical traffic information that is statistical traffic information for each day type is calculated, and the calculated date Statistical traffic information creation unit having a function of storing the type statistical traffic information and the corresponding day type in association with each other,
Based on the traffic information and the day-type statistical traffic information, the non-day-type fluctuation amount that is the fluctuation amount of the traffic information not depending on the day-type is calculated, and the non-day-type information corresponding to the target date, A prediction coefficient calculation unit that has a function of calculating a prediction coefficient that is a weight of the non-daily information, based on the calculated non-daily variation amount and the acquired non-daily information, obtained from the storage unit And comprising
The in-vehicle terminal is
When a prediction target link and a prediction target date and time are specified via the input unit, a day type corresponding to the prediction target date and time is acquired from the predicted traffic information providing device, and a day corresponding to the acquired day type is acquired. Type statistical traffic information is acquired from the predicted traffic information providing device, non-daily information of the prediction target date and time is acquired from the predicted traffic information providing device, and the prediction coefficient is acquired from the predicted traffic information providing device, Based on the acquired prediction coefficient and the acquired non-day type information, a predicted non-day type fluctuation amount at the prediction target date and time is calculated, and the predicted non-day type fluctuation amount and the acquired daily type statistical traffic are calculated. A predicted traffic information providing system comprising a traffic information prediction unit that calculates predicted traffic information based on the information.

Images