JP4137672B2 - Traffic jam prediction system and traffic jam prediction method - Google Patents

Description

[0001]
[Technical field to which the invention belongs]
  The present invention relates to a technology for predicting a traffic jam in consideration of the occurrence of a traffic jam and the duration of the traffic jam.
[0002]
[Prior art]
  As a related prior art, JP 2002-296057 was extracted. According to the technique disclosed therein, an “expected congestion area” is set and reflected in route search. The “expected traffic jam area” is an area where a constant traffic jam is expected depending on the time zone because, for example, a predetermined number or more department stores exist within a predetermined range. It is said that it is a technology that enables route search that avoids a traffic jam section without using road traffic information.
[0003]
[Patent Document 1]
      JP 2002-296057 A
[0004]
[Problems to be solved by the invention]
  In the technique disclosed in Japanese Patent Application Laid-Open No. 2002-296057, factors such as weather and events to be held that are considered to be one of the causes of traffic congestion are not taken into consideration.
[0005]
  In addition, there is no prediction that takes into account data related to the elimination of traffic jams. Therefore, it is difficult to predict with high accuracy in short-term prediction such as 10 minutes or 1 hour later.
[0006]
  Therefore, the problem to be solved by the present invention is to provide a technology capable of predicting traffic jams taking into consideration the occurrence probability of traffic jams and experience values of the time to eliminate traffic jams.
  From claim 1Claim 5An object of the described invention is to provide a traffic jam prediction system capable of traffic jam prediction taking into account the occurrence probability of traffic jams and experience values of the time to eliminate jams.
  Claim 6An object of the described invention is to provide a traffic jam prediction process capable of traffic jam prediction taking into account the occurrence probability of traffic jams and experience values of the time to eliminate traffic jams.
[0007]
[Means for Solving the Problems]
  In order to solve the above-described problems, the present invention acquires past road traffic information and accumulates correlation data, and predicts occurrence of traffic jams together with prediction of traffic jam elimination. It is more suitable for short-term predictions such as 10 minutes or 1 hour later than medium- and long-term predictions such as several hours or days later.
[0008]
  (Claim 1)
  The invention described in claim 1 relates to a system for calculating traffic jam prediction data using traffic jam correlation data for traffic jam occurrence prediction and traffic jam pattern data for traffic jam duration time.
  That is,Based on past road traffic information, when other links are congested, this is the traffic congestion correlation, which is the probability that the link will be congested, and the time interval until the congested traffic that occurred on other links reaches the link. Per link including time gapBased on correlation data creation means to create traffic jam correlation data and past road traffic informationFor each day of the week and time of day, including average time and standard deviation during which traffic jams persistCongestion pattern data, which is actual dataFor each linkCongestion pattern data creation means to create,With the correlation data creation meansA correlation database that accumulates the created traffic jam correlation data;With the above traffic pattern data creation meansCongestion pattern database that accumulates the created congestion pattern data,Road traffic information receiving means for acquiring road traffic information for each link of the current status;  Acquired by the road traffic information receiving meansCurrent road traffic information for each linkAboveCongestion probability calculation means for calculating the probability of occurrence of traffic congestion for each link based on the traffic congestion correlation data, current road traffic information for each link, andAboveA congestion duration probability calculating means for calculating a congestion probability for each link based on the congestion pattern data;For each link, use the traffic jam persistence probability and the traffic jam occurrence probability multiplied by the calculation coefficient.A traffic jam prediction system comprising traffic jam forecast data calculating means for calculating traffic jam forecast data.
[0009]
  (Glossary)
  The “road traffic information for each link” is obtained from, for example, a road traffic information providing system provided separately from the system according to the present application. Here, the “road traffic information providing system” is, for example, a Japanese road traffic information center that provides road traffic information.
  The “link” is generally a section between main intersections, but the direction of travel is also specified. That is, it is data having elements of position information and direction like a vector.
[0010]
  When creating “congestion correlation data” and “congestion pattern data”, the past traffic data for each link is simply averaged, and the most recent data takes precedence over the previous data. In some cases, the weighted average is used. For example, when traffic data for the past three years is used, calculation is performed by weighting so that traffic data of one year before is more important than traffic data of three years ago.
[0011]
  (Function)
  Based on the past road traffic information for each link, the correlation data creation means creates traffic jam correlation data with other links for a predetermined link, and the created traffic jam correlation data is stored in the correlation database.
  On the other hand, based on the past road traffic information for each link, the congestion pattern data creation means creates the congestion pattern data that is the actual data of the congestion duration of the predetermined link, and the created congestion pattern data is stored in the congestion pattern database. To accumulate.
[0012]
  Next, on the basis of the current road traffic information for each link and the traffic jam correlation data, the traffic jam occurrence probability calculating means calculates the traffic jam occurrence probability for each link. On the other hand, based on the current road traffic information for each link and the traffic pattern data, the traffic congestion duration calculation means calculates the traffic jam persistence probability for each link.
  Then, the traffic jam prediction data calculation means calculates the traffic jam prediction data for each link using the traffic jam persistence probability and the traffic jam occurrence probability. Here, not only the occurrence probability of the traffic jam but also the traffic jam prediction taking into account the traffic jam duration, that is, the time for eliminating the traffic jam is performed. For this reason, improvement in prediction accuracy can be expected, and it is suitable for short-term prediction such as 10 minutes or 1 hour later.
[0013]
[0014]
  (Glossary)
  The “congestion degree of correlation” refers to the probability that a link (A) for which traffic jam prediction is to be made when a certain link (D) is congested.
  "Time gap" means a certain link (D)I want to predict traffic jamsTime to reach link (A).
[0015]
  (Function)
  The correlation database accumulates traffic jam correlation data with other links for a given link. At the time of accumulation, the degree of correlation and the time gap relating to the congestion of other links with respect to a predetermined link are calculated and accumulated based on the road traffic information in units of links. Since the traffic jam correlation degree of other links related to the link related to the link specific data is accumulated together with the time gap, the traffic jam occurrence probability can be calculated easily and rationally.
[0016]
[0017]
  Congestion pattern data is accumulated for each day of the week and for each time period, and the average time and standard deviation for which the congestion continues are accumulated. What is necessary is just to select applicable traffic pattern data. For this reason, the calculation is simple, and even with road traffic information that is added (changed) from moment to moment, traffic jam predictions can be calculated one after another.
[0018]
  (Claim 2)
  Claim 2The invention described in is limited to the traffic jam prediction system described in claim 1,
  Traffic jamThe pattern data creation means creates the traffic jam pattern data for each weather based on the past weather data, and the traffic jam persistence calculation means calculates the traffic jam persistence probability for each link based on the traffic jam pattern data corresponding to the weather data. The present invention relates to a traffic jam prediction system to be prepared.
[0019]
  (Glossary)
  The “weather data” is generally actual weather recorded every region and every predetermined time. In addition to the weather such as weather = sunny, rain, and snow, it may include accompanying data such as how many millimeters of rain per hour and how many centimeters of snow. The traffic jam pattern database has traffic jam pattern data when the weather is sunny and traffic jam pattern data when the weather is raining separately, and the traffic jam persistence probability calculation means uses traffic pattern data corresponding to the weather data. Is extracted, and the congestion probability is calculated.
  For “weather data”, it is also possible to use a weather forecast forecasted for each region and time zone instead of the current weather data. Need to take into account the correct probability (for example, about 80%).
[0020]
  (Function)
  The traffic pattern database stores traffic pattern data for each weather. Then, the traffic jam duration probability calculating means extracts the traffic jam pattern data corresponding to the weather data, and calculates the traffic jam duration probability for each link using it. For this reason, more accurate congestion probability can be calculated, and as a result, the accuracy of the congestion prediction data is increased.
[0021]
  (Claim 3)
  Claim 3The invention described is from claim 1Claim 2Limited to the traffic jam prediction system described in either
  Traffic jamThe pattern data creation means creates congestion pattern data for each event based on past event data, and the congestion duration probability calculation means determines the congestion for each link based on the congestion pattern data corresponding to the current event data. The present invention relates to a traffic jam prediction system that calculates the sustainability probability.
[0022]
  (Glossary)
  “Event data” is data related to an event held near a road related to traffic jam prediction. In addition to formal data such as the date and time of the event, it may include information related to the event, such as an expected turnout. For example, professional sports games, concerts in event halls, traffic restrictions associated with festivals and demonstrations.
[0023]
  (Function)
  In the traffic pattern database, traffic pattern data is created for each event. Then, the traffic jam duration probability calculating means extracts the traffic jam pattern data corresponding to the presence / absence of an event, and calculates the traffic jam duration probability for each link using the data. For this reason, more accurate congestion probability can be calculated, and as a result, the accuracy of the congestion prediction data is increased.
[0024]
  (Claim 4)
Claim 4The invention described is from claim 1Claim 3Limited to the traffic jam prediction system described in either
Using the error data storage means for comparing the traffic jam prediction data calculated by the prediction data calculation means with the actual traffic data and storing the error data, and using the error data stored in the error data storage means, the traffic jam prediction The present invention relates to a traffic jam prediction system including a calculation coefficient correction unit that corrects a calculation coefficient used in a data calculation unit.
[0025]
  (Glossary)
  The “error data storage means” may be a random access memory, but is preferably a nonvolatile storage device such as a hard disk. This is because it is possible to cope with a case where the calculation coefficient corrected for accuracy improvement is changed later.
  The “calculation coefficient etc.” refers to a coefficient such as (m, n) when calculated as “congestion prediction data = congestion occurrence probability × m + congestion persistence probability × n”, for example. The “calculation coefficient correction means” is a means for calculating a new calculation coefficient (m ′, n ′) in order to reduce the error based on the error data.
[0026]
  (Function)
  The traffic jam prediction data calculated by the prediction data calculation unit is compared with the actual traffic jam data, and the error data is stored in the error data storage unit. Then, using the error data stored in the error data storage means, the calculation coefficient correction means for use in the traffic jam prediction data calculation means is corrected by the calculation coefficient correction means. By using the corrected calculation coefficient or the like, the accuracy of the traffic jam prediction data calculated after the next time is further increased.
[0027]
  (Claim 5)
Claim 5The invention described is from claim 1Claim 4Limited to the traffic jam prediction system described in either
Position data etc. receiving means for receiving position data etc. transmitted from the transmission / reception means such as position data capable of transmitting the current position data etc., and the position data of the traffic jam prediction data calculated by the traffic jam prediction data calculation means The present invention relates to a traffic jam prediction system including a traffic jam prediction data transmission unit that transmits traffic jam prediction data related to the location data to the position data etc. transmission / reception unit.
[0028]
  (Glossary)
  “Position data and the like” means that at least position data and a traveling direction are required. In addition to location data, there are links that want to know traffic jam prediction data. If the position data is received a plurality of times at predetermined intervals, the position data receiving means can acquire data such as direction data and whether or not the traffic is in traffic.
  Examples of the “position data transmission / reception means” include a car navigation device, a mobile phone having a GPS function, and a mobile information terminal.
[0029]
  (Function)
  The position data etc. receiving means receives the position data etc. transmitted from the position data etc. transmitting / receiving means capable of transmitting the current position data etc.
Of the traffic jam prediction data calculated by the traffic jam prediction data calculation means, the traffic jam prediction data related to the position data is transmitted by the traffic jam prediction data transmission means to the position data etc. transmission / reception means.
[0030]
  (Claim 6)
Claim 6The described invention relates to a traffic jam prediction method for calculating traffic jam prediction data using traffic jam correlation data for traffic jam occurrence prediction and traffic jam pattern data for traffic jam duration time.
  That is,Based on past road traffic information, when other links are congested, this is the traffic congestion correlation, which is the probability that the link will be congested, and the time interval until the congested traffic that occurred on other links reaches the link. Per link including time gapBased on the correlation data creation procedure to create traffic jam correlation data and past road traffic informationFor each day of the week and time of day, including average time and standard deviation during which traffic jams persistCongestion pattern data, which is actual dataFor each linkCongestion pattern data creation procedure to create,With the correlation data creation meansCorrelation data accumulation procedure for accumulating the created traffic jam correlation data in the correlation database;With the above traffic pattern data creation meansCongestion pattern data accumulation procedure for accumulating the created congestion pattern data in the congestion pattern database,Road traffic information reception procedure to obtain road traffic information for each current link,  Obtained by the road traffic information reception procedureCurrent road traffic information for each linkAboveCongestion probability calculation procedure for calculating the probability of occurrence of traffic jam for each link based on the traffic jam correlation data, and current road traffic information for each linkAboveCongestion probability calculation procedure for calculating congestion probability for each link based on traffic pattern data,For each link, use the traffic jam persistence probability and the traffic jam occurrence probability multiplied by the calculation coefficient.This is a traffic jam prediction method in which a traffic jam forecast data calculation procedure for calculating traffic jam forecast data is executed.
[0031]
  (Variation 1)
It is also possible to provide a system for calculating traffic jam prediction data for each link as follows.
  That is, based on past road traffic information for each link, a correlation data creating means for creating traffic jam correlation data with other links for a given link, a correlation database for accumulating the created traffic jam correlation data, and for each link A traffic jam occurrence probability calculating means for calculating the traffic jam occurrence probability for each link based on the current road traffic information and the traffic jam correlation data, the traffic jam correlation data for the link relating to the link specific data; This is a traffic jam prediction system that accumulates a traffic jam correlation degree and a time gap for each other link related to the traffic jam.
[0032]
  The traffic jam prediction probability may be calculated by combining the traffic jam occurrence probability for each link calculated by the traffic jam probability calculation means with other data calculated for traffic jam prediction. The probability of occurrence may be used alone.
[0033]
  (Variation 2)
It is also possible to provide a system for calculating traffic jam prediction data for each link as follows.
  That is, a traffic jam pattern data creation means for creating traffic jam pattern data that is actual data of traffic jam duration time in a predetermined link based on past road traffic information for each link, and a traffic jam pattern database that accumulates the created traffic jam pattern data And a traffic jam persistence probability calculating means for calculating the traffic jam persistence probability for each link based on the current road traffic information and the traffic jam pattern data for each link. It is a traffic jam prediction system that accumulates the average time and standard deviation that lasts.
[0034]
  The traffic jam prediction data may be finally calculated by combining the traffic jam persistence probability calculated by the traffic jam persistence calculation means with other data calculated for traffic jam prediction. The persistence probability may be used alone.
[0035]
DETAILED DESCRIPTION OF THE INVENTION
  Hereinafter, the present invention will be described in more detail based on embodiments and drawings. The drawings used here are FIGS. 1 to 8.
[0036]
  (Figure 1)
  FIG. 1 shows an overall outline of the present system that calculates traffic jam prediction data for each link using traffic jam correlation data for traffic jam occurrence prediction and traffic jam pattern data for traffic jam duration time.
This system includes a correlation database (abbreviated as “DB” in the figure) that accumulates traffic jam correlation data with other links for a given link based on road traffic information in link units acquired from the road traffic information providing system, Similarly, a traffic jam pattern database is provided for creating and storing traffic jam pattern data, which is actual data of traffic jam continuation time on a predetermined link, based on road traffic information acquired from the road traffic information providing system.
[0037]
  (Congestion correlation data)
  The traffic jam correlation data accumulated in the correlation database is the traffic jam correlation degree for other links (for example, B, C, etc.) that are deeply related to traffic jams occurring in a certain link (for example, A), This data includes the time gap to the link. The correlation data creating means creates based on the past traffic data for each link. Details will be described later with reference to FIG.
[0038]
  (Congestion pattern data)
  The traffic pattern data accumulated in the traffic pattern database is the time from the occurrence of traffic jams to the resolution of a given link by obtaining past traffic data for each link from the road traffic information provision system. The mean value and standard deviation are calculated. Based on the past traffic data for each link, the traffic pattern data creating means creates. For ease of use and accuracy when used for traffic jam prediction, it is created for each day of the week and each time zone and stored in the traffic jam pattern database. Details will be described later with reference to FIG.
In addition, the traffic pattern data is created by acquiring past weather data from the weather data providing system, acquiring past event data from the event data providing system, and determining whether or not there is an event for each weather. Details will be described later with reference to FIG.
[0039]
  (Congestion prediction data)
  Using the traffic jam correlation data extracted by the correlation data extraction unit and the current traffic jam data for each link, the traffic jam occurrence probability calculation unit calculates the traffic jam occurrence probability for each link. On the other hand, using the current traffic data and the current event data, the traffic congestion pattern data extracted by the traffic pattern data extraction means and the current traffic data for each link, the traffic jam persistence probability calculation means uses the current traffic data for each link. Probability is calculated.
[0040]
  In the calculation, a calculation coefficient such as (m, n) in the case of calculating as “congestion prediction data = congestion occurrence probability × m + congestion persistence probability × n” is used. The calculation coefficient is sequentially corrected in order to improve accuracy. Details will be described later with reference to FIG.
  The traffic jam prediction data calculation means calculates the traffic jam prediction data for each link using the traffic jam occurrence probability and the traffic jam persistence probability calculated respectively. The calculated result is output by the traffic jam prediction data output means in the required data format.
[0041]
  (Figure 2)
  FIG. 2 is a conceptual diagram showing a procedure for calculating a traffic jam occurrence probability using the traffic jam correlation data and the current status data.
The road traffic information provided from the road traffic information providing system is received by the road traffic information receiving means, and the correlation data creating means creates correlation data for past data. The current data is used when calculating the occurrence probability of traffic jams. The following will be described by taking “link A” as an example.
[0042]
  (Congestion correlation data)
  In the correlation database, traffic jam correlation data for a plurality of links (B to F) deeply related to the traffic jam of each link (A in this case) is accumulated. The traffic jam correlation data includes a combination of the traffic jam correlation degree and the time gap. The traffic jam correlation degree is a probability that a link (link A) to be calculated when the link is jammed will be jammed. The time gap is the linkI want to predict traffic jamsThis is the time required to reach the link (link A). In links BF, the traffic jam correlation degree is 70, 60, 50, 40, 30%, and the time gap is 10, 15, 20, 25, 30 minutes.
[0043]
  (Present data and calculation of traffic jam probability)
  It is assumed that links B and C are not congested and links D, E and F are congested in the current state data received by the road traffic information receiving means.
The congestion occurrence probability of the link A to be calculated is calculated using the congestion correlation data and the current state data. If the current time is 7:00 am on Monday, the probability that link A is congested after 10 minutes, 15 minutes, 20 minutes, 25 minutes, and 30 minutes is 0% and 0%, respectively. , 50%, 40% and 30%.
For example, the probability of being involved in a traffic jam on the link A that will reach after 20 minutes is 50% displayed as 20 minutes later.
[0044]
  (Figure 3)
  Next, calculation of the congestion congestion probability will be described based on FIG.
The road traffic information provided from the road traffic information providing system is received by the road traffic information receiving means, and the congestion pattern data creating means creates the congestion pattern data for past data. The current data is used to calculate the congestion probability. The following will be described by taking “link A” as an example.
[0045]
  (Congestion pattern data)
  The traffic congestion pattern data is accumulated by calculating the average duration μ and standard deviation σ of traffic congestion calculated from past data for every day of the week and every three hours. When the assumption that the past traffic jam data has a substantially normal distribution is largely broken, it may be corrected or another calculation method may be adopted. The congestion pattern data is created separately for each weather, whether there is an event, etc., but is omitted here for the sake of simplicity. Moreover, although the time zone is divided every three hours, it is not particularly defined and may be changed in operation.
[0046]
  (Present data and congestion probability calculation)
  From the road traffic information system, the link A receives that 15 minutes have passed since the occurrence of the traffic jam. On the other hand, since the current time is 7:00 am on Monday, the average time of traffic jam μ and standard deviation σ, which is the time zone of Monday 6: 00 to 8: 59, is extracted, and traffic jam duration probability calculating means However, the congestion probability is calculated. The probability that traffic jams will continue on Link A at 7:00 am on Monday is 30%, 20%, 10%, 10 minutes, 15 minutes, 20 minutes, 25 minutes, and 30 minutes, respectively. 5% and 3%.
[0047]
  An example of calculation after 20 minutes is shown. Since 15 minutes have passed since the occurrence of the traffic jam, and further 20 minutes later from the present, 35 minutes have passed since the traffic jam occurred. Therefore, the probability that the traffic jam is eliminated is calculated to be 10% from the statistical probability that the average duration of the traffic jam is 20 minutes and the standard deviation is 5 minutes. That is, the probability that the traffic jam continues even after 20 minutes is 10%.
[0048]
  (Fig. 4)
  FIG. 4 illustrates the traffic jam prediction data after the traffic jam occurrence probability and the traffic jam persistence probability are calculated, and correction (tuning) for the traffic jam prediction data calculation means.
  When the traffic jam occurrence probability and the traffic jam persistence probability are calculated, it is calculated as “traffic jam prediction data = traffic jam occurrence probability × m + traffic jam persistence probability × n”. If (m, n) = (1.0,1.0), according to the example shown in FIGS. 2 and 3, 50% + 10% = 60%.
  In order to improve the accuracy of the traffic jam prediction data calculated as 60%, the following means are provided.
[0049]
  (Correction means)
  As shown in FIG. 4, the traffic jam prediction data calculated by the traffic jam prediction data calculation unit and the actual traffic jam data are compared, and the difference data is stored, and the error data storage unit stores the difference data. And calculating coefficient correction means for correcting the calculation coefficient used in the traffic jam prediction data calculating means.
A new calculation coefficient (m ′, n ′), for example, (1.12, 0.95) calculated by various calculation methods such as a weighted average is used by the correction coefficient calculation unit, and the calculation is performed next time. Increase the accuracy of traffic jam prediction data.
[0050]
  (Fig. 5)
  FIG. 5 conceptually illustrates a traffic jam pattern database in which several traffic jam pattern tables are prepared.
This system uses a road traffic information providing system, an event data providing system, and a weather information providing system as external information systems.
[0051]
  The road traffic information provided from the road traffic information providing system includes traffic jam information, accident information, and construction information. The weather data provided from the weather information providing system is data on the weather for each region and time zone. Event data provided from the event data providing system, such as a schedule for holding professional baseball, soccer, concerts, festivals, etc., is used.
The traffic congestion pattern data constitutes a plurality of pattern data by a combination of event data and weather data. For example, it is stored in the traffic pattern database as classified traffic pattern data such as good weather and no event, rain and no event, good weather and event.
[0052]
  (Fig. 6)
  FIG. 6 shows a case where the weather data provided from the weather data providing system shown in FIG. 5 and the event data provided from the event data providing system are used as means for correcting congestion pattern data. That is, in FIG. 5, a plurality of traffic pattern data is provided. However, FIG. 6 shows the basis of traffic pattern data with good weather and no events. Used by multiplying the correction coefficient and event correction coefficient.
[0053]
  For example, a weather correction coefficient is prepared for the traffic pattern data when the weather is sunny, such that the probability of traffic congestion is 1.3 times that of a certain day of the week when it is raining. The congestion probability is calculated. The same applies to the event correction coefficient. In other words, for the congestion pattern data when there is no event, prepare an event correction coefficient such that the congestion probability is 1.6 times that when there is no event in a certain time zone on the event date. Is calculated.
[0054]
  (Fig. 7)
  FIG. 7 illustrates an example of using the traffic jam prediction data individually. The user transmits position data to the traffic jam prediction system from an information communication terminal that can detect and transmit the position data of the user. In some cases, the location data is used to specify a link for which traffic jam prediction is desired.
The transmitted position data or the like is received by a position data etc. receiving means provided in the present system. Then, individual traffic jam prediction data related to the position data and the like is extracted from the traffic jam prediction data output unit, and transmitted from the traffic jam prediction data transmission unit to the information communication terminal related to the user.
By providing the above configuration, it is possible to realize a traffic jam prediction data providing service to the user. Details of the billing system for the user are omitted.
[0055]
  (Fig. 8)
  FIG. 8 shows an embodiment in which each function in the traffic jam prediction system shown in FIG. 1 is distributed. In other words, the congestion occurrence probability calculation system for calculating the occurrence probability of traffic for each link, the congestion probability calculation system for calculating the congestion probability for each link, and the occurrence probability and congestion probability for each link calculated by both systems. The traffic jam prediction system that calculates the traffic jam prediction data by using them is made independent of each other.
[0056]
  By making the three systems independent, there are various merits such that one or both of the traffic jam occurrence probability calculation system and the traffic jam persistence probability calculation system can be independently improved or operated.
  Further, for example, the traffic jam prediction data may be finally calculated by combining the traffic jam occurrence probability for each link calculated by the traffic jam occurrence probability calculating means and other data calculated for traffic jam prediction. Further, the traffic jam prediction data may be finally calculated by combining the traffic jam persistence probability for each link calculated by the traffic jam duration probability calculation means and other data calculated for traffic jam prediction.
[0057]
【The invention's effect】
  From claim 1Claim 5According to the described invention, it is possible to provide a traffic jam prediction system that can perform traffic jam prediction in consideration of the occurrence probability of traffic jams and the experience values of the time to eliminate traffic jams.
  Claim 6According to the described invention, it is possible to provide a traffic jam prediction process capable of traffic jam prediction taking into consideration the occurrence probability of traffic jams and the experience values of the time for eliminating jams.
[Brief description of the drawings]
FIG. 1 is a conceptual diagram showing an overall outline of the present system for calculating traffic jam prediction data.
FIG. 2 is a conceptual diagram showing a procedure for calculating a traffic jam occurrence probability using traffic jam correlation data and current status data.
FIG. 3 is a conceptual diagram showing a procedure for calculating a congestion congestion probability.
FIG. 4 is a conceptual diagram for explaining processing after a traffic jam occurrence probability and a traffic jam persistence probability are calculated.
FIG. 5 is a conceptual diagram of a traffic jam pattern database in which several traffic jam pattern tables are prepared.
FIG. 6 is a conceptual diagram showing a case where event data provided from an event data providing system is used as a means for correcting congestion pattern data.
FIG. 7 is a conceptual diagram illustrating a case where traffic jam prediction data is individually used.
FIG. 8 is a conceptual diagram showing an overall outline of a system that divides and integrates each function for a system that calculates traffic jam prediction data.

Claims (6)

A system that calculates traffic jam forecast data using traffic jam correlation data for traffic jam forecasting and traffic jam pattern data for traffic jam duration,
Based on past road traffic information, when other links are congested, this is the traffic jam correlation, which is the probability that the link will be congested, and the time interval until the traffic generated on the other link reaches the link. time gap, the correlation data creation means for creating a traffic jam correlation data for each link including,
Based on past road traffic information, for each day of the week and each time zone , a traffic jam pattern data creating means for creating traffic jam pattern data, which is actual data including average time and standard deviation for which traffic jams continue, for each link ;
A correlation database for storing traffic jam correlation data created by the correlation data creation means ;
A traffic jam pattern database for accumulating the traffic jam pattern data created by the traffic jam pattern data creating means ;
Road traffic information receiving means for acquiring road traffic information for each link of the current status;
A traffic occurrence probability calculating means for calculating a traffic jam occurrence probability for each link based on the current road traffic information for each link acquired by the road traffic information receiving means and the traffic jam correlation data;
A traffic jam persistence probability calculating means for calculating a traffic jam persistence probability for each link based on the current road traffic information for each link and the traffic jam pattern data;
With multiplied by the respective calculation coefficient in traffic sustained probability and the congestion probability of that, traffic jam prediction system and a traffic jam prediction data calculation means for calculating a traffic jam prediction data for each link. The traffic pattern data creation means creates traffic pattern data for each weather based on past weather data,
The traffic jam prediction system according to claim 1, wherein the traffic jam duration probability calculating means calculates a jam duration probability for each link based on the traffic jam pattern data corresponding to the weather data. The traffic pattern data creation means creates traffic pattern data based on past event data for each event,
3. The traffic jam prediction system according to claim 1, wherein the traffic jam duration probability calculating means calculates a jam duration probability for each link based on the traffic jam pattern data corresponding to the current event data. Error data storage means for comparing the traffic jam prediction data calculated by the prediction data calculation means and the actual traffic jam data and storing the error data;
Using the error data stored in the error data storing means, any one of claims 1 to 3 and a calculation coefficient such correcting means for correcting the calculation coefficient and the like used in the traffic jam prediction data calculation means The traffic jam prediction system described in. It has position data etc. receiving means for receiving the position data etc. transmitted from the position data etc. transmitting / receiving means capable of transmitting the current position data etc.,
The traffic jam prediction data related to the position data, etc. of the traffic jam prediction data congestion prediction data calculating means is calculated, of claims 1 to 4 having a traffic-congestion prediction data transmission means for transmitting to said position data, such as transmitting and receiving means The traffic jam prediction system described in either. A traffic jam prediction method for calculating traffic jam forecast data using traffic jam correlation data for traffic jam forecasting and traffic jam pattern data for jam duration,
Based on past road traffic information, when other links are congested, this is the traffic jam correlation, which is the probability that the link will be congested, and the time interval until the traffic generated on the other link reaches the link. Correlation data creation procedure for creating traffic jam correlation data for each link including a time gap ,
Based on the past road traffic information, for each day of the week and every time zone , a traffic jam pattern data creation procedure for creating traffic jam pattern data, which is actual data including average time and standard deviation that traffic jams last, for each link ,
A correlation data accumulation procedure for accumulating traffic jam correlation data created by the correlation data creation means in a correlation database;
A congestion pattern data storing procedure for storing the congestion pattern data prepared in traffic pattern database in the traffic jam pattern data creating means,
Road traffic information reception procedure to obtain road traffic information for each current link,
A traffic jam occurrence probability calculating procedure for calculating a traffic jam occurrence probability for each link based on the current road traffic information for each link acquired in the road traffic information receiving procedure and the traffic jam correlation data;
A traffic jam duration calculation procedure for calculating a jam duration probability for each link based on the current road traffic information for each link and the traffic jam pattern data,
With multiplied by the respective calculation coefficient in traffic sustained probability and the congestion probability of that, a traffic jam prediction method to perform a congestion prediction data calculation procedure for calculating the traffic jam prediction data for each link.

Images