JP2005004668A - Traffic information providing device and traffic information providing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To precisely forecast traffic information which includes day kinds, such as the day of the week, the days of the month which can numerically be divided by five without a remainder and so on. <P>SOLUTION: A forecast computing means is provided, which reflects the day kinds (for example, the day of the week, the days of the month which can numerically be divided by five without a remainder, a weekday/holiday, large consecutive holidays, a period of time when a school is closed, and weather) in a numerical traffic information forecast, the forecast being made for a travelling time with interval units of several minutes which in not of day units, and forecasts the traffic information by adequately combining accumulated past traffic information under the condition where targeted day kinds are combined, even if a day having the same combination of day kinds as a forecast object day (day of the same event) does not exist in a period of time, when the traffic information is accumulated. The forecast computing means approximately restores and forecasts the phenomenon, by combining several day-unit feature quantities of the traffic information. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to the provision of traffic information, and more particularly to the presentation of traffic information that takes into account the day of the week, fifty days, and the like.
[0002]
[Prior art]
The method using quantification I, which is one of the mainstream traffic prediction methods, can predict numerical traffic information such as travel time, traffic congestion, traffic volume, etc., reflecting the day type. It is. The application target of this traffic information prediction method is, for example, prediction of traffic information on a daily basis such as a point traffic volume per day and a degree of traffic jam in a specific time zone. However, when the day type is the explanatory variable of the quantification class I, the explanatory variable is information on a daily basis, and is not suitable for prediction of traffic information that is not on a daily basis, such as travel time of several minutes.
[0003]
In this regard, as a technology that can predict traffic information with an arbitrary period, such as travel time of several minutes, traffic information is classified by day type, for example, as required time guidance of the Metropolitan Expressway Public Corporation. It is known to use an accumulation method.
[0004]
[Patent Document 1]
JP 2001-118188 A
[Non-Patent Document 1]
http: // mex. survey. ne. jp / mexntime /
[0005]
[Problems to be solved by the invention]
When the prediction is performed using the traffic information classified by the day type, the day type may be combined such that the day to be predicted is Monday and the fifty days.
[0006]
However, in the above technique, the event of the prediction target day, that is, the past traffic information under the combination of day types is searched and the prediction calculation is performed, so the event of the prediction target day does not exist in the period in which the traffic information is accumulated. In some cases, accurate predictions cannot be made.
[0007]
In view of the above problems, an object of the present invention is to predict traffic information under a desired event.
[0008]
[Means for Solving the Problems]
An object of the present invention is to display a map of a region including a prediction target route between two points, accept an input of a day type of a prediction target date or a date of a prediction target date, and a past traffic information time series of the prediction target route. This is achieved by obtaining traffic information time-series data corresponding to a plurality of input day types based on the data and displaying the obtained traffic information time-series data as a graph.
[0009]
Another object of the present invention is to obtain a base component from past traffic information time-series data and a feature amount as a coefficient for each base component when approximating traffic information time-series data from the base component. The coefficient of the feature quantity estimation model that approximates the feature quantity as a function of a plurality of day types is obtained from the combination information and the feature quantity of the day type for each day, and the feature quantity is estimated for the combination information of the day type in the future. This is achieved by obtaining a feature amount estimated value by a model and providing traffic information obtained by combining the base component using the feature amount estimated value as a coefficient.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
The provision of traffic information according to the present invention will be described below with reference to the drawings.
[0011]
FIG. 1 is an operation screen of a traffic information prediction system according to an embodiment of the present invention. The date input field 101 is an interface for inputting the date of the prediction target date. The recommended route input field 102 is an interface for designating a route to be predicted. A departure point 104 and a destination 105 are set in the map display field 103, and a route search button 106 is selected. The traffic information prediction target route 107 is displayed. When a date is specified in the date input field 101 and a prediction target route is determined in the recommended route input field 102 and then the prediction execution button 108 is selected, a travel time prediction graph 110 is displayed in the travel time display field 109 on the recommended route 107. Is displayed. The error range 111 represents, in a time series, a range of errors that can be taken with an actual travel time with a probability of 80%, for example, when traveling on a prediction target route at a certain time.
[0012]
The map displayed in the map display field 103 can be scrolled by selecting the scroll button 112 arranged around it or by dragging the map itself, and selecting an appropriate position on the scale display field 113. You can change the scale. The route information display field 114 in the recommended route input field 102 displays prediction information such as the travel distance on the recommended route, travel time at a specific time, and fuel consumption. The time for displaying the travel time and fuel consumption in the route information display field 114 can be specified by the time cursor 116 by selecting an appropriate position on the time axis 115 of the travel time display field 109. . In addition, when a dynamic route search algorithm that reflects traffic information is used in determining the prediction target route 107, the prediction target route changes with time, and the time can also be designated by the time cursor 116. . Search conditions such as time priority and distance priority when the prediction target route is determined by the route search algorithm can be specified in the search condition specification field 117.
[0013]
In FIG. 1, the date of the prediction target date is input from the date input field 101. Instead, from the day type input field 201 of FIG. 2, the day of the week, fifty days, weekdays / holidays, large consecutive holidays, school holidays, weather It is also possible to predict traffic information by inputting a combination of day types such as. In the operation screen of FIG. 1, the date input field and the date type input field are exclusively switched using a tab type interface. In this embodiment, the day type input field 201 includes a “weekday / holiday” input field 202, a “season” input field 203, and a “weather” input field 204, and items in each field are selected exclusively. . For example, in the category of “weekday / holiday” input field, “weekdays”, “holidays / holidays”, and “continuous holidays” are exclusive, and only one of them can be selected as the day type of the prediction target day. In the example of FIG. 2, this exclusive selection is realized by a radio button type interface. In the figure, ◯ represents selectable and ◎ represents actually selected.
[0014]
On the other hand, the fields are not exclusive, and the day type can be expressed by combining the fields selected in each field. In the example of FIG. 2, “weekdays” is selected from the “weekdays / holidays” input field, “spring” is selected from the “season” input field, and “rain” is selected from the “weather” input field, and the combination “Weekday Hunchun” is selected. The traffic information is predicted for the day type called “Rainy Rain”.
[0015]
In addition, the day types “50 days” and “school leave period” in the “others” input field 205 are day types that are not mutually exclusive, and are configured by a check box type interface in the example of FIG. The day type in the “other” input field 205 can be selected simultaneously with the day type of any other field. For example, in the example of FIG. 2, both “50 days” and “school holiday period” are selected. In this case, the traffic information is predicted under the combination of the day type “weekday, spring, rain, rainy day, school holiday period” together with the day type selected in the field 201-203. Note that, as shown in the date input field 101, a date type independent of the date such as weather can be input simultaneously with the date.
[0016]
In the example of FIG. 1, the starting point and the destination are specified in the recommended route input field 102, and the prediction target route is automatically set by the route search algorithm. Instead, the free route input field 301 in FIG. It is also possible for the user to freely specify the route to be predicted using. At that time, the user selects a link on the map display field 103 to input a route, and selects the route confirmation button 302 to confirm the prediction target route. When the prediction target route is input again, the route clear button 303 is selected.
[0017]
Further, if the section input field 401 in FIG. 4 is used, it is possible to predict traffic information using a section prepared in advance on the traffic information prediction system side. The section list 402 is a list represented by section names. By selecting a section from the section list 402 and selecting a section selection button 403, a section to be predicted is fixed, and a prediction target route is displayed in the map display field 103. Is done. In the operation screen of FIG. 1, the recommended route input field, the free route input field, and the section input field are exclusively switched using a tab-type interface.
[0018]
FIG. 5 is a conceptual diagram of a prediction method necessary for the traffic information prediction system. As shown in FIG. 1 and FIG. 2, in order to predict traffic information that is not a day unit (24 hour unit) such as 1 hour unit, 10 minute unit, 1 minute unit, etc. under any combination of day types, Conventional methods such as quantification type I using day type as explanatory variables or day type classification cannot be used. Therefore, in S501 of FIG. 5, from the past traffic information time series data obtained by receiving VICS information and collecting information of the probe car, the traffic information time series data is approximated by one day by linear summation. A plurality of base components that can be combined, and a daily feature amount as a coefficient for each base component when traffic information is combined from the base component, and a prediction that represents the feature amount as a function of a plurality of day types in S502 Determine the model's prediction coefficients.
[0019]
In this system, the idea that traffic information time-series data is represented by basis components and feature quantities is based on the concept that arbitrary time-series data is approximately expressed as a composite value of a plurality of orthogonal functions. Fourier transform, which is a signal processing technique, represents an arbitrary signal as a composite waveform of a plurality of trigonometric functions having different periods. The basis component and the feature amount referred to in the present application correspond to a coefficient related to each trigonometric function when linearly synthesizing a trigonometric function in Fourier transform and a plurality of trigonometric functions.
[0020]
In addition, the concept of predicting the feature quantity in association with the day type in this system is based on an empirical rule that the same result can be obtained from the same cause. A conceptual response is like a weather forecast or a technical analysis of stock prices. If the precipitation probability of tomorrow's weather forecast is 10%, the current weather map is compared with a similar weather map in the past, and the relative frequency of raining the next day from a similar weather map in the past is 0. It was about .10. Therefore, the system called weather forecast is the same concept as this system. Further, a system in which a buy sign is issued if the relative frequency at which the stock price rises for a predetermined period from the next day onward is 0.90 or more from the past stock price transition is the same concept as this system. It is easy to understand when the feature amount and the predicted value of the feature amount referred to in the present application are considered as PER, RSI, and Bollinger band.
[0021]
In order to predict traffic information using such a prediction model when the prediction execution button 108 is selected in FIG. 1, the day information is input to the prediction model in S503, and the feature amount is calculated using the prediction model in S504. Time series data of traffic information predicted values can be obtained by calculating a predicted value and combining the base component calculated in S501 as traffic information in S505 using the feature value predicted value as a coefficient. In this way, the daily feature quantity is predicted and calculated as a function of day type, and the calculated feature quantity predicted value is converted to traffic information, so that traffic information of an arbitrary period can be obtained under any combination of day types. Predictive calculations can be made. Therefore, using the interface such as the day type input field 201 shown in FIG. 2, the predicted value of the time series data such as the travel time and the congestion degree is provided in the form of the travel time prediction graph 110 of FIG. Can do.
[0022]
FIG. 6 is a block diagram of an embodiment of a traffic information prediction system provided with the prediction method whose concept is shown in FIG. The feature quantity extraction device 601 calculates the base component and the daily feature quantity as follows. That is, data obtained by receiving VICS information or collecting information by a probe car, and accumulated in a traffic information database (hereinafter referred to as DB) 602, such as travel time, congestion degree, traffic volume, etc. Based on the time series data, the base component and the feature amount are calculated by a method such as principal component analysis. A plurality of basis components are necessary for synthesizing the original traffic information using the feature quantity as a coefficient, and a plurality of basis components are obtained. A plurality of basis components can be synthesized approximately for each day by linear addition of the traffic information. A plurality of feature quantities are also obtained, and each base component is related as a coefficient when traffic information is synthesized from the base component. At this time, the feature quantity extraction device 601 outputs the feature quantity to the prediction coefficient determination device 603 and outputs the base component to the traffic information synthesis device 604. The above processing that calculates the base component and the feature quantity from the traffic information and outputs them to the traffic information synthesis device and the prediction coefficient determination device respectively corresponds to the processing of S501 in FIG.
[0023]
When the principal component analysis is used for calculating the feature amount, the base component is uniquely calculated simultaneously with the feature amount. On the other hand, it is also possible to calculate a feature amount by designating a plurality of functions orthogonal to each other in advance as basis components. The Fourier transform is an example, and when the Fourier transform is used for calculating the feature quantity, a plurality of trigonometric functions having different periods are used as the base component.
[0024]
The prediction coefficient determination device 603 uses a method such as quantification I from the feature amount input from the feature amount extraction device 601 and the day type information of the period for which the feature amount extraction device 601 is the target of feature amount calculation. Calculate the prediction coefficient. This prediction coefficient is used to calculate the prediction value of the daily feature amount by the prediction model using the day type as a parameter in the feature amount prediction apparatus 605. The calculated prediction coefficient is recorded in the feature amount prediction apparatus 605. The day type information includes days of the week, fifty days, weekdays / holidays, large consecutive holidays, school holidays, weather, etc., and is recorded in the day type DB 606. The above processing for calculating the prediction coefficient in the prediction coefficient determination device corresponds to the processing of S502 in FIG.
[0025]
When quantification type I is used for the prediction calculation of the feature amount, the function form of the prediction model is a linear sum of day types, and binary description that indicates whether it corresponds to N types of day types by 1 and 0, respectively. Using the variables d1, d2,..., DN and the prediction coefficients a1, a2,.
T = a1 * d1 + a2 * d2 + ... + aN * dN
It is expressed. On the other hand, for example, when reflecting numerical data such as temperature and precipitation in the prediction model, the terms of multi-value explanatory variables x1, x2,.
T = a1 * d1 + a2 * d2 + ... + aN * dN + b1 * x1 + b2 * x2 + ... + bM * xM
Is used. In the above equation, the term of the multi-value explanatory variable is a first-order term, but a prediction model having terms such as second-order and third-order is also conceivable.
[0026]
When the traffic information is predicted, a day type parameter required for the prediction model is input to the feature amount prediction apparatus 605 according to the event on the prediction target day. The feature amount prediction device 605 calculates a feature amount prediction value using the day type parameter input by a person and the prediction coefficient input and recorded from the prediction coefficient determination device 603, and the feature amount prediction value is calculated as traffic information. The data is output to the synthesizer 604. This corresponds to the processing of S504 in FIG.
[0027]
The traffic information synthesizing device 604 performs a calculation for synthesizing the base components input and recorded from the feature value extracting device 601 using the predicted feature value as a coefficient. This composite value is a predicted value of traffic information corresponding to the day type parameter input to the feature quantity predicting device 605, and is output to the traffic information display device 607. In FIG. 6, the predicted value of the traffic information is output to the traffic information display device 607. However, the predicted traffic information may be used as an input of route search means in the car navigation device or dispatch plan means of the dispatch management system. Is possible.
[0028]
FIG. 12 is a processing flow for predicting traffic information. Hereinafter, the processing flow of FIG. 12 will be described according to the embodiment of FIG.
[0029]
S1201: In the feature amount extraction apparatus 601, the above-described base component that can be combined approximately from past traffic information accumulated in the traffic information DB 602, and a coefficient when combining the original traffic information from the base component The feature amount is calculated.
[0030]
S1202: The base component calculated in S1201 is recorded in the traffic information synthesizer 604.
[0031]
S1203: In the prediction coefficient determination device 603, the prediction coefficient for calculating the predicted value of the feature amount using the prediction model with the day type as a parameter is read from the feature amount calculated in S1201 and the day type DB 606. Calculate from day type information.
[0032]
S1204: The prediction coefficient calculated in S1203 is recorded in the feature amount prediction apparatus 605.
[0033]
S1205: When predicting traffic information on a specific day, the process proceeds to S1206. When predicting traffic information by specifying an event, the process proceeds to S1207.
[0034]
S1206: The date input device 801 receives the date of the prediction target date, reads the day type schedule DB 802 and, if necessary, weather information, reads the day type information from the weather forecast DB 804 and inputs it to the feature amount prediction unit 605. To do.
[0035]
S1207: The day type information representing the event to be predicted is received by the day type input device 805, and the day type information is input to the feature amount prediction device 605.
[0036]
S1208: In the feature amount prediction apparatus 605, the predicted value of the feature amount corresponding to the event of the prediction target day is calculated from the day type information input in S1206 or S1207 using the prediction coefficient recorded in S1204. .
[0037]
S1209: The traffic information synthesizing device 604 calculates the predicted value of traffic information from the feature amount predicted value calculated in S1208 using the basis component recorded in S1202.
[0038]
In the traffic information prediction system according to the embodiment of FIG. 6, the process from feature amount extraction to traffic information prediction is treated as a series of processes, but a process of creating a DB necessary for prediction in advance as shown in FIG. In addition, it is possible to perform the processing asynchronously by dividing the processing for each prediction. The feature quantity extraction device 601 and the prediction coefficient determination device 603 shown in FIG. 6 constitute a traffic information prediction DB creation device 701 in the embodiment of FIG. The base component and the prediction coefficient calculated by the feature amount extraction device 702 and the prediction coefficient determination device 703 are recorded in the traffic information prediction DB 704. In the traffic information prediction DB creation device 701, the processing beyond calculating the base component and the prediction coefficient from the traffic information time series data and recording it in the traffic information prediction DB 704 is executed in advance separately from the calculation processing required for each prediction. It is possible to update the basis component and the prediction coefficient recorded in the traffic information prediction DB 704 by this processing when the traffic information in the traffic information DB 705 used for prediction is updated or This is when the model is changed.
[0039]
In FIG. 6, the feature amount predicting device 605 and the traffic information synthesizing device 604 used in the process of predicting the traffic information constitute the traffic information predicting device 706 in the embodiment of FIG. When the day information parameter is input, the traffic information prediction device 706 passes the parameter to the feature amount prediction device 707, and the feature amount prediction device 707 and the traffic information synthesis device 708 receive the prediction coefficient and the base component from the traffic information prediction DB 704, respectively. Read out and output the predicted value of traffic information according to the day type parameter. The processing beyond the calculation of the traffic information prediction value from the day type parameter in the traffic information prediction device 706 is necessary every time the traffic information is predicted for different day type parameters.
[0040]
The processing flow for predicting traffic information will be supplemented with reference to FIG. In contrast to the processing flow in accordance with the embodiment of FIG. 6, in the embodiment of FIG. 7, the basis component and the prediction coefficient are recorded in the traffic information prediction DB 704 in S1202 and S1204, respectively. In S1208 and S1209, the basis component and The prediction coefficient is read out from the traffic information prediction DB 704.
[0041]
FIG. 8 is an example of inputting the day type parameter, and is a block diagram showing a means for inputting the day type parameter to the traffic information prediction system in the embodiment of FIGS. 6 and 7. The date input device 801 reads the day type information of the corresponding date from the day type schedule DB 802 storing the future day type information according to the date input from the user through the operation screen illustrated in FIG. The data is output to the feature amount prediction apparatus 803 shown in FIG. At this time, if the feature amount prediction apparatus 803 performs prediction calculation using weather information as a day type parameter and uses weather forecast information instead of the weather specified by the user as the weather information, the weather forecast The weather forecast information of the date corresponding to the date type DB 802 is read from the DB 804 and output to the feature amount prediction apparatus 803 together with other day type information.
[0042]
In the day type input device 805, the user can freely combine day types such as day of the week, fifty days, weekdays / holidays, large holidays, school holidays, weather, etc. through the operation screen illustrated in FIG. It is possible to input to 803 and predict the traffic information. The day type displayed on the day type selection operation screen is configured according to the day type parameter required in the prediction model in the provision of traffic information, and is limited to the embodiment of FIG. is not.
[0043]
In FIG. 8, it is assumed that the user inputs the date or date type. However, the program constituting the car navigation device, the vehicle allocation management system, etc. automatically inputs the date or date type to the date input device 801 or the date type input device 805. You can enter it.
[0044]
FIG. 9 is a block diagram showing an embodiment applied to a communication navigation device and a traffic information center for providing traffic information. According to the traffic information prediction system of the present application, the traffic information can be provided by calculating traffic information from the day parameters shown in FIGS. 6 and 7, the prediction coefficient of the feature quantity, and the base component of the traffic information. The prediction amount of the feature amount and the information amount of the base component of the traffic information are smaller than the information amount of the traffic information prediction value calculated based on these pieces of information. That is, it can be said that this is a preferred embodiment when the traffic can be kept small and traffic information is provided using a communication line. In the embodiment of FIG. 9, the traffic information prediction database 704 of FIG. 7 is placed in the traffic information center 901, and the traffic information prediction device 706 is installed in the communication type navigation device 902, so that transmission of predicted traffic information is provided. The necessary information is not the traffic information prediction value itself, but only the day parameter, the prediction coefficient, and the base component.
[0045]
The traffic information prediction DB creation device 701 and the traffic information prediction DB 704 in FIG. 7 are installed in the traffic information center 901, and the traffic information prediction DB creation device 903 calculates the prediction coefficient and the base component, and the traffic information prediction DB 904 stores them. The recording is performed asynchronously as a process independent of provision of predicted traffic information from the traffic information center 901 to the communication navigation device 902. When predicting traffic information in the communication type navigation device 902, the date of the prediction target date is transmitted to the traffic information center 901, and the date corresponding to the prediction coefficient, the base component, and the transmitted date is transmitted from the traffic information center 901. The seed parameter is received, and the traffic information prediction device 905 calculates the predicted value of the traffic information.
[0046]
The communication data when providing the predicted traffic information from the traffic information center 901 to the communication type navigation device 902 follows the format of FIG. 10 and the day type parameter, prediction coefficient, basis for each link or section which is the prediction calculation unit of the traffic information. Consists of ingredients. In FIG. 10, the day type parameter, the prediction coefficient, and the base component are collected for each prediction link / section. However, as shown in FIG. 11, each prediction link / section is set for each information type such as the day type parameter, the prediction coefficient, and the base component. It is also possible to take a format that summarizes the above information, and if it is a communication format that can transmit day parameters, prediction coefficients, and base components necessary for prediction calculation by the traffic information prediction device 905 for each prediction section to be predicted, The embodiment is not limited to the formats shown in FIGS. In addition, when predicting a plurality of links / sections, the day type parameter may be the same for each prediction link / section, and in such a case, it is not necessary to transmit the day type parameter in duplicate. .
[0047]
In the embodiment of FIG. 9, the date of the prediction target date is transmitted from the communication type navigation device to the traffic information center 901, and the day type parameter generated by the traffic information center 901 via the date input device 801 shown in FIG. It is sent back to the communication type navigation device 902. On the other hand, the date input device 906 and the day type schedule DB 907 placed in the traffic information center 901 are mounted on the communication type navigation device 902, and the process of generating the day type parameter from the date is performed on the communication type navigation device 902 side. It is also possible to carry out offline, or to install the day type input device 805 in the communication type navigation device 902 and directly input the day type parameter offline.
[0048]
At this time, the day type parameter is excluded from the communication formats of FIGS. The feature amount prediction device 707 and the traffic information composition device 708 constituting the traffic information prediction device 905 can be separated, the former can be placed in the traffic information center 901, and the latter can be mounted on the communication type navigation device 902. In this case, the prediction of the feature amount is performed in the traffic information center 901, and the day type parameter and the prediction coefficient are excluded from the communication formats of FIGS. 10 and 11, and the feature amount prediction value is added instead. In the embodiment in which the date input device 906 and the day type schedule DB 907 are installed in the communication type navigation device 902, when the traffic information prediction device 905 needs weather as a day type parameter, the weather forecast information is downloaded in advance. When the forecast DB is installed in the communication type navigation device 902 or when the traffic information prediction calculation is performed, the weather forecast DB connected to the network is accessed to obtain the weather forecast information on the prediction target day.
[0049]
Here, the provision of information is applied to a communication type navigation device, but it can be applied to a mobile phone, a PDA connected to a network, a personal computer or the like instead. In addition, when the information is not provided through communication but is provided by a recording medium such as a DVD-ROM, for example, the traffic information predicting apparatus shown in FIG. 706 functions can be provided. Of course, also in this case, the information to be recorded on the recording medium includes only the prediction coefficient, the base component, and the day type parameter for each link / section, that is, only the information stored in the traffic information prediction DB 904 and the day type schedule DB 907. can do. By doing so, the amount of information can be reduced as described above, so that the load on the storage device can be reduced more than the traffic information predicted value itself.
[0050]
In this traffic information prediction system, the day type is reflected in the prediction of numerical traffic information that is not a daily unit, such as travel time every few minutes, and there is no day of the same event as the prediction target day in the period when traffic information is accumulated Even in this case, it is possible to accurately predict traffic information under a desired event by combining past traffic information.
[0051]
【The invention's effect】
According to the present invention, traffic information under a desired event can be predicted.
[Brief description of the drawings]
FIG. 1 shows an example of an operation screen of a traffic information prediction system.
FIG. 2 shows an example of a day type input field on the operation screen of the traffic information prediction system.
FIG. 3 shows an example of a free route input field on the operation screen of the traffic information prediction system.
FIG. 4 shows an example of a section input field on the operation screen of the traffic information prediction system.
FIG. 5 is a conceptual diagram of a prediction method necessary for a traffic information prediction system.
FIG. 6 is a block diagram showing a traffic information prediction system.
FIG. 7 is a block diagram showing a traffic information prediction system.
FIG. 8 is a block diagram of a means for inputting date and date type.
FIG. 9 is a block diagram showing an embodiment applied to a communication navigation device and a traffic information center.
FIG. 10 shows an example of a communication format in an embodiment applied to a communication navigation device and a traffic information center.
FIG. 11 shows an example of a communication format in an embodiment applied to a communication navigation device and a traffic information center.
FIG. 12 shows an example of a processing flow for predicting traffic information.

Claims (9)

Traffic information database that records past traffic information time series data,
A map display device for displaying a map of a region including a predicted route between two points;
An input device for inputting the day type or the date of the prediction target date;
A traffic information calculation device for obtaining traffic information time-series data corresponding to a plurality of day types input from the day-type input device based on the past traffic information time-series data of the prediction target route;
A traffic information providing device comprising: a traffic information display device that displays the obtained traffic information time-series data as a graph. In claim 1,
The input device is a date input device for inputting the date of the prediction target date,
A weather database that records weather information for dates after the forecast date;
A day type database that records combination information of day types other than the weather such as day of the week, weekday / holiday, season, applicable to the fifty-day, applicable to school holidays, etc.
Based on the combination information of the day type other than the weather read from the day type database and the weather information read from the weather database based on the date of the prediction target date input to the date input device A traffic information providing apparatus, wherein the traffic information calculation apparatus obtains traffic information time-series data. In claim 1,
A prediction target route input device that directly inputs the prediction target route without searching,
A traffic information providing apparatus, wherein the traffic information calculation apparatus obtains traffic information time-series data for an inputted prediction target route. In claim 1,
The prediction target path is obtained by input to a prediction target section input device that inputs a section name representing the prediction target path,
A traffic information providing apparatus, wherein the traffic information calculation apparatus obtains traffic information time-series data for the obtained prediction target route. From the past traffic information time series data recorded in the traffic information database, when calculating the basis component and the feature amount as a coefficient for each base component when approximating the traffic information time series data from the basis component,
A coefficient of a feature amount estimation model that approximates the feature amount as a function of a plurality of day types is calculated from the combination information of the day type for each day in the calculation target period and the feature amount,
For the combination information of the day type of a certain day in the future, the feature amount estimation value is calculated by the feature amount estimation model,
A traffic information providing method for providing traffic information obtained by synthesizing the base components using the estimated feature value as a coefficient. In claim 5,
When providing traffic information, the feature quantity estimation model coefficient and the base component are read from the traffic information prediction database, and the feature quantity estimation value is calculated and combined with the base component.
A process of calculating the base component and the feature quantity estimation model coefficient from the past traffic information time-series data and recording them in the traffic information prediction database; and the feature quantity estimation model coefficient and the base component in the traffic information prediction database A method for providing traffic information, comprising: asynchronously executing a process for calculating time series data of future traffic information read out from the network. In claim 5,
When providing future traffic information from the traffic information center to a communication type navigation device, mobile phone, network-connected personal computer, PDA or other communication terminal at the current time, the basis component and the feature amount estimation model coefficient are Transmitted from the traffic information center to the communication terminal,
In the communication terminal, the feature amount estimation value is calculated by inputting the combination information of the day type in a future day to the feature amount estimation model,
Traffic information time-series data corresponding to the combination information of the day type input to the communication terminal is transmitted from the traffic information center by synthesizing the base component using the estimated feature value as a coefficient in the communication terminal. A traffic information providing method comprising: calculating at the communication terminal using only the base component and the feature quantity estimation model coefficient. In claim 5,
When providing future traffic information from a traffic information center to a communication terminal such as a communication type navigation device, a mobile phone, a networked personal computer or a PDA at the current time, the date of a certain future day To the traffic information center,
Transmitting the basis component, the feature amount estimation model coefficient, and the combination information of the day type corresponding to the date from the traffic information center to the communication terminal,
In the communication terminal, the combination information of the day type is input to the feature amount estimation model to calculate a feature amount estimation value,
The base information transmitted from the traffic information center is transmitted to the traffic information time-series data according to the future date input to the communication terminal by synthesizing the base component using the estimated feature value as a coefficient in the communication terminal. A traffic information providing method comprising: calculating in a communication terminal using only the combination information of the component, the feature quantity estimation model coefficient, and the day type. In claim 5,
When providing future traffic information from a traffic information center to a communication type navigation device, mobile phone, network-connected personal computer, PDA or other communication terminal at the current time, the combination information of the day of the day in the future Or the date of a certain future day is transmitted from the communication terminal to the traffic information center,
In the traffic information center, the feature amount estimation value is calculated by inputting the combination information of the day type or the combination information of the day type corresponding to the date into the feature amount estimation model,
Transmitting the basis component and the feature amount estimated value from the traffic information center to the communication terminal;
By combining the base component with the feature value estimated value as a coefficient in the communication terminal, the traffic information time-series data according to the combination information of the day type or the future date input to the communication terminal is obtained. A traffic information providing method characterized in that a calculation is performed in a communication terminal using only the base component and the feature amount estimated value transmitted from the communication terminal.

Images