JP2013057544A - Traffic situation route guide device, route guide method, and route guide program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a traffic situation route guide device, route guide method and route guide program capable of mitigating a traffic jam.SOLUTION: A route guide device calculates a first passage prediction time to be required for a vehicle to pass a link being a predetermined section on the basis of the movement speed of the vehicle, calculates a second passage prediction time to be required for the vehicle to pass a route from a present travel position to a target position on the basis of the calculated first passage prediction time, and selects a route on the basis of the calculated second passage prediction time.

Description

  The present invention relates to a route guidance device, a route guidance method, and a route guidance program for a traffic situation for predicting a traffic situation on a road from information related to the traffic situation of a vehicle.

The traffic situation prediction calculation device collects road information data related to the traffic situation of a vehicle, predicts a traffic jam situation that occurs based on the road information data, and provides information on the predicted traffic jam situation. Here, the road information data is periodically collected by a traffic information system. A system for predicting traffic conditions such as traffic jam information whose situation changes every moment based on road information data collected periodically as described above is known (see, for example, Non-Patent Documents 1 to 3).
Information on the predicted traffic situation is provided via a road information display board provided on the roadside, a road traffic information communication system (VICS (Vix)), broadcast media, and the like.

  For example, in a road traffic information communication system, an aggregation center collects vehicle traffic data via a fixed sensor installed on a road. The tally center processes the collected data to generate road traffic information such as traffic jams, and multiplexes the generated road traffic information with FM broadcasts and transmits them. For this reason, in the conventional road traffic information communication system, road traffic information can be provided only on the road where the fixed sensor is installed. Therefore, the VICS traffic information can cover only about 70,000 [km] of main roads where fixed sensors are installed out of the total length of about 830,000 [km] of roads with a width of 5.5 [m] or more in Japan. It wasn't.

  For this reason, the following system for providing road traffic information (hereinafter referred to as a road traffic information system) has started to be provided. In the road traffic information system, the vehicle transmits probe traffic information in which time information is added to travel information indicating the position and travel state of the host vehicle. The road traffic information system provides road traffic information using the received probe traffic information. Such services are currently performed based on information from vehicles using car navigation systems of each automobile manufacturer or car navigation manufacturer.

Shiraishi, Kuwahara, Horiguchi, “Development of Real-time Predictive Traffic Flow Simulation”, 30th Civil Engineering Research Conference, November 21-23, 2004. Shiraishi, Horiguchi, "Development of traffic flow prediction system using real-time simulation", 54th Theoretical and Applied Mechanics Lecture Meeting (NCTAM2005), Science Council of Japan, January 27, 2005. Shiraishi, Kuwahara, Horiguchi, “A Development of Traffic Prediction System Based on Real-time Simulation”, 12th ITS World Congress San Francisco 2005, November 6-10, 2005.

  However, when the number of vehicles that provide probe traffic information or vehicles that use probe traffic information increases, if the optimal route is provided to the vehicle as in the conventional car navigation system, the vehicle concentrates on the optimal route. As a result, there was a problem that traffic congestion could not be alleviated.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a route guidance device, a route guidance method, and a route guidance program in a traffic situation that can alleviate traffic congestion.

  In order to achieve the above object, the route guidance apparatus according to the present invention calculates a first predicted passing time required for the vehicle to pass a link, which is a predetermined section, based on the moving speed of the vehicle, Based on the calculated first predicted passage time, a second predicted passage time required for the vehicle to pass the route from the current travel position to the target position is calculated, and the calculated second predicted passage time is calculated. Based on this, the route is selected.

  In the present invention, the first estimated passage time is a time estimated to be required for the vehicle to pass the link. The second predicted passage time is a time predicted to be required for the vehicle to pass the route from the current travel position to the target position. That is, the present invention is characterized in that the route guidance device searches for a route candidate using the first predicted passage time of the link, not the route.

  In the route guidance device of the present invention, a traffic information collection unit that acquires traffic information including the moving speed of the vehicle, a link on the route is extracted, and the vehicle is set to a target position based on the moving speed of the vehicle. Predicted time information for extracting a link where a traffic jam has occurred by the time of arrival and generating a first predicted pass time required for the vehicle to pass through the link with the traffic jam based on the moving speed of the vehicle And a generation unit.

  In the route guidance device of the present invention, a route candidate information generating unit that generates route candidate information indicating a route candidate from the current travel position of the vehicle to the target position, and the vehicle arrives at the target position by the route candidate. A dynamic marginal cost information generating unit that calculates a first dynamic marginal cost based on the first estimated passage time for a link that has traffic jams up to the time of the link, and a link in which the traffic jam has not occurred in the route candidate The second dynamic marginal cost is calculated based on the estimated transit time required for the vehicle to pass through, and the second dynamic marginal cost and the dynamic marginal cost information are generated for all the links of the route candidate The first dynamic marginal cost calculated by the section is summed for each of the route candidates, the total dynamic marginal costs are compared, and as a result of the comparison, the route candidate that minimizes the total dynamic marginal cost A Pareto improvement route guidance unit that searches, may be provided with a.

  In the present invention, the dynamic limit cost of a link means an increase (decrease) in total cost due to (or did not) one vehicle selecting a certain link. The first dynamic marginal cost is a dynamic marginal cost for a link that is congested before the time when the vehicle arrives at the target position in the route candidate. The second dynamic marginal cost is a dynamic marginal cost for a link where there is no congestion. The marginal cost is generally a cost per unit of production, and in the present invention, it is a cost per unit for movement between routes. That is, the present invention is characterized in that the Pareto improvement route guide unit searches for a route candidate using a dynamic limit cost of a link instead of a route.

  In the route guidance device of the present invention, the predicted time information generation unit extracts one link from all links on the route candidate, and until the time at which the vehicle on the extracted link arrives at the destination position. If it is determined that the moving speed of the vehicle is less than a predetermined first speed, the extracted link is predicted to be a link in which traffic congestion has occurred by the time when the vehicle arrives at the destination position. It may be.

  Further, in the route guidance device of the present invention, the predicted time information generation unit extracts a link where the traffic jam ends as a bottleneck link based on information indicating the link where the traffic jam occurs, and extracts the bottleneck In the link, the time at which the bottleneck / link congestion on the route candidate ends may be estimated using the moving speed of the vehicle for each link and the route candidate information.

  In the route guidance device of the present invention, the predicted time information generation unit may move the vehicle at the time of reaching the subordinate link in all the subordinates downstream of the link where the traffic jam occurs. Is determined to be equal to or higher than a predetermined second speed, the link where the traffic jam occurs is weighted, and a value indicating the weight of the link where the traffic jam occurs is a predetermined value. If it is larger, the link where the traffic jam has occurred may be extracted as a bottleneck link.

  In the route guidance device according to the present invention, the predicted time information generation unit may continue the link at a speed less than a predetermined speed based on information indicating the link where the traffic jam occurs and a moving speed of the vehicle. In such a case, it may be determined that the link has one continuous traffic jam.

  In the route guidance device of the present invention, the Pareto improvement route guidance unit may determine the degree of influence of the bottleneck link based on the number of terminals in the graph structure of the link subordinate to the upstream side of the bottleneck link. The dynamic marginal cost may be calculated using the calculated degree of influence.

  Further, in the route guidance device of the present invention, the Pareto improvement route guidance unit calculates the total of the link cost in the current route that is the route from the current position that the vehicle is currently selected to the target position, If the total of the link cost of the selected route candidate is smaller than the total of the link cost in the current route, compared with the total of the link cost of the selected route candidate, the selected route candidate is determined as the route. You may do it.

  In order to achieve the above object, the present invention provides a route guidance method in a route guidance device, wherein a first estimated passage time required for a vehicle to pass a link, which is a predetermined section, is determined by moving the vehicle. Means for calculating on the basis of speed; means for calculating a second predicted passage time required for the vehicle to pass the route from the current travel position to the target position based on the calculated first predicted passage time; And a means for selecting the route based on the calculated second estimated passage time.

  In order to achieve the above object, the present invention provides a route guidance program for causing a computer to execute a route guidance method process in a route guidance device, in order for a vehicle to pass a link that is a predetermined section. It is necessary for the vehicle to pass the route from the current traveling position to the target position based on the procedure for calculating the required first passing predicted time based on the moving speed of the vehicle and the calculated first predicted passing time. It is characterized by causing a computer to execute means for calculating a second predicted passage time and means for selecting the route based on the calculated second predicted passage time.

The route guidance device according to the present invention calculates a time (first predicted pass time) that is predicted to be required for the vehicle to pass the link. Next, the route guidance device according to the present invention is based on the calculated first estimated passage time, and is estimated to be required for the vehicle to pass the route from the current travel position to the target position (second Predicted passage time), and a route is selected based on the calculated second predicted passage time.
As a result, the route guidance device according to the present invention reduces the travel time from the travel time in the user optimum state with the shortest route for the user, and further the total travel time of the user who uses the route guidance device It is possible to alleviate traffic congestion in the Pareto optimal state, which can reduce the increase in the travel time of the user that occurs in a small system optimal state.

It is a figure explaining the framework of the route guidance system concerning this embodiment. It is a figure explaining a user equilibrium state, a Pareto optimal state, and a system optimal state in a static allocation problem with a paradox. It is a schematic structure figure of route guidance system 1 concerning this embodiment. It is a figure explaining the number of inflow vehicles to a link, and the number of outflow vehicles from a link. It is a figure explaining the limit cost of the path | route which concerns on this embodiment. It is a figure explaining the dynamic marginal cost which concerns on this embodiment. It is a figure explaining the traffic congestion area which concerns on this embodiment. It is a figure explaining an example of the identification method of the bottleneck link which concerns on this embodiment. It is a figure explaining the influence degree of the bottleneck link which concerns on this embodiment. It is a flowchart of the process sequence of the route guidance system 1 which concerns on this embodiment. In the route guidance system 1 which concerns on this embodiment, it is the map used for the simulation. It is the result of having simulated the route guidance system 1 which concerns on this embodiment in the area shown in FIG. It is a figure showing the number of vehicles and travel time change at the time of a route change in a system optimal state. It is a figure showing the number of vehicles and travel time change at the time of the route change in the pareto optimal state concerning this embodiment.

First, the outline of the present invention will be described.
The route guidance system according to the present invention calculates a marginal cost in a Pareto optimal state to be described later. Further, the route guidance system according to the present invention is characterized in that it calculates the marginal cost of a link, which will be described later, instead of calculating the marginal cost of the route. The marginal cost is generally a cost per unit of production, and in the present invention, it is a cost per unit for movement between routes. And the route guidance system which concerns on this invention produces | generates the path | route of each vehicle used as a Pareto optimal state based on the marginal cost of this link, and provides it to each vehicle.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, an outline of the present invention will be described. FIG. 1 is a diagram for explaining a framework of a route guidance system according to this embodiment.
In FIG. 1, the symbol ter represents a group of vehicles using the route guidance system 1. Vehicles M1-1 to M1-4 represent probe vehicles. The probe vehicle is a vehicle equipped with a device that acquires and transmits vehicle travel data (speed information, position information, time information, etc.). The data acquired by this device is referred to as probe traffic information, and the probe vehicle transmits the acquired probe traffic information to the probe traffic information system Sys1 (also referred to as a probe transmission center).
Vehicles M10-1 to M10-16 represent vehicles using the route guidance system 1 other than the probe vehicle.
Vehicles M20-1 to M20-4 represent vehicles that do not use the route guidance system 1.

  Information i1 to i4 represents information transmitted from the probe vehicles M1-1 to M1-4 to the probe traffic information system Sys1. Information i5 schematically represents information i1 to i4. Information i6 represents information transmitted from the probe traffic information system Sys1 to the real-time traffic jam prediction system Sys2. The information i7 represents information transmitted from the real-time traffic jam prediction system Sys2 to the environment improvement Pareto route guidance system Sys3. Information i8 represents information transmitted from the environmental improvement Pareto route guidance system Sys3 to the vehicle using the route guidance system 1.

The probe vehicles M1-1 to M1-4 and the probe traffic information system Sys1 serve as sensors that capture changes in traffic conditions immediately and in detail.
The real-time traffic jam prediction system Sys2 generates future situation information (referred to as predicted travel time information) that can be predicted from the current situation. The real-time congestion prediction system Sys2 obtains dynamic marginal cost (DMC) information necessary for the calculation performed by the environment improvement Pareto route guidance system Sys3 based on the generated predicted travel time information.
The environmental improvement Pareto route guidance system Sys3 generates route guidance information in the Pareto optimal state based on the predicted travel time information and the dynamic marginal cost information generated by the real-time traffic jam prediction system Sys2, and the generated route guidance information is transmitted to the vehicle (M1). -1 to M1-4, M10-1 to M10-16).
The probe traffic information system Sys1, the real-time traffic jam prediction system Sys2, the environmental improvement Pareto route guidance system Sys3, and the vehicles (M1-1 to M1-4, M10-1 to M10-16) constitute an open loop system.
Note that the dynamic limit cost of a route in the present embodiment means an increase (decrease) in total cost due to (or did not) one vehicle selecting a certain route.

Here, the concept of the Pareto optimum state will be described with reference to FIG.
FIG. 2 is a diagram for explaining a user equilibrium state, a Pareto optimum state, and a system optimum state in a static allocation problem with a paradox. The Pareto optimal state is a state in which the travel time of the own vehicle cannot be further shortened without increasing the travel time of other vehicles. Under such conditions, it is known that, under certain conditions, a socially lower cost can be achieved than a user equilibrium state where the principle of equal cost between routes is established. The system optimum state is an optimum state when viewed from the system, in which the total cost on the route is minimized. In addition, the user equilibrium state is a state where costs are equalized for all network users.

  In FIG. 2, the route is represented as a network. Symbols p1 to p4 each represent a point. Reference numerals l12, l13, l23, l24, and l34 each represent a link. Reference numeral l12 represents a link (12) from point 1 to point 2, and reference numeral l24 represents a link (24) from point 2 to point 4. Reference numeral l13 represents a link (13) from the point 1 to the point 3, and reference numeral l34 represents a link (34) from the point 3 to the point 4. The symbol l23 represents the link (23) from the point 2 to the point 3.

  In the example shown in FIG. 2A to FIG. 2D, when moving from the point p1 to p4, there are the following three routes. The first route is the route (124) that first passes through the link (12) 112 and then passes through the link (24) 112. The second route is a route (134) that first passes through the link (13) l13 and then passes through the link (34) l34. The third route is a route (1234) that first passes through the link (12) 112, then passes through the link (23) 112, and finally passes through the link (34) 114.

FIG. 2A is a diagram for explaining service levels and marginal costs.
In FIG. 2A, a link (12) 112 and a link (34) 134 are links having a high service level and a large marginal cost. The link (13) l13 and the link (24) l24 are links having a low service level and a small marginal cost. The link (23) l23 is a link with a small marginal cost that connects the link (12) l12 with a large marginal cost and the link (34) l34. The marginal cost is a cost per unit for movement between routes in the network as shown in FIG.
In FIG. 2A, the symbol x12 represents the cost of the link (12) 112, and the symbol x24 represents the cost of the link (24) 112. The symbol x13 represents the cost of the link (13) l13, and the symbol x34 represents the cost of the link (34) l34. The symbol x23 represents the cost of the link (23) I23.
Of the cost “10x ₁₂ ” of the link (12) 112, “10” multiplied by “x ₁₂ ” represents the magnitude of the marginal cost. “50” added to “x ₁₃ ” out of the cost “x ₁₃ +50” of the link (23) l23 represents the size of the service level.
Further, in FIG. 2 (a), the cost of each link, the cost of the link 12 is _{10x 12,} the cost of the link 13 is _x 13 +50, the cost of the link 23 is _x 23 +10, the link 24 The cost is x ₂₄ +50, and the cost of the link 34 is 10x ₃₄ .

FIG. 2B is a diagram illustrating an example of a user equilibrium state.
The state of FIG. 2B is the route cost of the route 124 (hereinafter referred to as route cost) (10x ₁₂ (= 4) + (x ₂₄ (= 2) +50) = 92) and the route of the route (134). Cost ((x ₁₃ (= 2) +50) + 10x ₃₄ (= 4) = 92) and route cost (10x ₁₂ (= 4) + (x ₂₃ (= 2) +10) + 10x ₃₄ (=) 4) = 92) is in equilibrium, that is, coincides. Therefore, the total cost in FIG. 2B is 92 × 3 = 276.

FIG. 2C illustrates an example of the system optimum state.
The costs in the state of FIG. 2C are the route cost (10x ₁₂ (= 3) + (x ₂₄ (= 3) +50) = 83) of the route (124) and the route cost ((x ₁₃ (= 3) +50) + 10x ₃₄ (= 3) = 83), route cost of route (1234) (10x ₁₂ (= 3) + (x ₂₃ (= 0) +10) + 10x ₃₄ (= 3) = 70) It is. Therefore, the total cost 236 (= 83 + 83 + 70) in the system optimum state in FIG. 2C is lower than the total cost 276 in the user equilibrium state in FIG.

FIG. 2D is a diagram for explaining an example of the Pareto optimal state.
Each cost in the state shown in FIG. 2 (d), total cost of the route of the path _{(124) (10x 12 (=} 3.5) + (x 24 (= 2.5) +50) = 87.5), route (134) total cost of route ((x ₁₃ (= 2.5) +50) + 10x ₃₄ (= 3.5) = 83), total cost of route (1234) (10x ₁₂ (= 3. 5) + (x ₂₃ (= 1.0) +10) + 10x ₃₄ (= 3.5) Therefore, the total cost 256 of the Pareto optimal state in FIG. It is higher than the total cost 236 of the state and lower than the total cost 276 of the user equilibrium state of FIG.

In a general route guidance system, an optimal route that is in a user balanced state as shown in FIG. 2B is provided to all vehicles that use the route guidance system. In the case of such a user equilibrium state, there is a problem that vehicles are concentrated on the optimum route and traffic congestion cannot be alleviated. In the user equilibrium state, as described with reference to FIG. 2B, the path cost and the total cost are larger than the Pareto optimal state and the system optimal state.
A case will be described in which a route that is the optimum system optimum state for the route guidance system (hereinafter referred to as a system optimum route) is introduced to alleviate traffic congestion. In the system optimum state, the route guidance system does not guide all vehicles to the optimal route, but guides to a route that forces some vehicles to detour to prevent the vehicle from concentrating on the optimal route To do. In this case, since the driver of the vehicle forced to detour suffers a disadvantage, it is difficult to accept the route guidance of such a route. However, as described in FIG. 2D, the route cost and the total cost are smaller than those in the user equilibrium state.
On the other hand, as described in FIG. 2C, in the Pareto optimum state, the route cost and the total cost (= 81) are lower than those in the user equilibrium state, and the route of some users (route 1234). The cost is lower than the route cost (= 83) of the route (124) and the route (134) in the system optimum state.
For this reason, the route guidance system according to this embodiment does not deteriorate the efficiency of traffic congestion mitigation by preventing the vehicle from concentrating on the optimum route, and provides route guidance in the Pareto optimal state that is not disadvantageous to the driver. Do.

Next, the configuration of the route guidance system 1 will be described with reference to FIG.
FIG. 3 is a schematic configuration diagram of the route guidance system 1 according to the present embodiment. As shown in FIG. 3, the route guidance system 1 includes a probe traffic information collection device 101, a traffic information storage device 102, an external information storage unit 103, a statistical information processing device 104, a predicted travel time information generation device 105, a dynamic limit. An expense information generation device 106 and a Pareto improvement route guidance device 107 are provided.

  The probe traffic information collection device 101 (traffic information collection unit) acquires probe traffic information i10 transmitted by the probe vehicle M30. The probe traffic information collection device 101 uses the received probe traffic information i10, traffic volume information from sensors installed on the road, past traffic volume information, and the like to detect current traffic volume and vehicle travel by a known method. Real-time traffic information i11 indicating speed or the like is generated. The traffic information is information such as the position of the vehicle, the moving speed, and the current time. The probe traffic information collection device 101 outputs the generated real-time traffic information i11 to the predicted travel time information generation device 105, and further stores the generated real-time traffic information i11 in the traffic information storage device 102.

The traffic information storage device 102 stores real-time traffic information i11 at each time.
The external information storage unit 103 stores traffic information acquired from sensors, historical statistical traffic information, and traffic information generated based on information transmitted from sensors installed on the road. Yes.

The statistical information processing apparatus 104 reads the real-time traffic information i12 stored in the traffic information storage apparatus 102. The statistical information processing apparatus 104 reads the traffic information i13 stored in the external information storage unit 103. The statistical information processing apparatus 104 generates a prediction model based on the read real-time traffic information i12 and traffic information i13.
Note that the prediction model is a prediction model that receives the fluctuation data of the previous travel time based on the current time as an input and outputs the fluctuation information of the future travel time based on the current time.
The prediction model shows the fluctuation characteristics of travel time information learned from the statistical data of the read real-time traffic information i12 and traffic information i13. Note that the fluctuation characteristics are characteristics indicating the relationship of output to input by, for example, regression analysis of data input immediately before and subsequent output.
The predicted travel time information generation apparatus 105 receives the real-time traffic information as input (past information based on the current time) and outputs the predicted travel time information (the future based on the current time). Information).
The traffic information storage device 102 and the external information storage unit 103 each store information for one week, for example. Then, the statistical information processing device 104 reads each information stored in the traffic information storage device 102 and the external information storage unit 103, for example, once a week, and generates the above-described prediction model, Information i14 indicating the generated prediction model is output to the predicted travel time information generation device 105.

The predicted travel time information generation device 105 (route candidate information generation unit, predicted time information generation unit) uses the prediction model generated by the statistical information processing device 104 for the real-time traffic information i11 generated by the probe traffic information collection device 101. The predicted travel time information (first predicted pass time) for each link is generated. The predicted travel time information generation device 105 is configured such that a vehicle transmitted by a vehicle M40 equipped with a traffic information terminal (not shown) is currently traveling route information (hereinafter referred to as current route information) and the vehicle is traveling. Information i15 including the existing destination information is acquired.
Based on the acquired current route information and destination information, the predicted travel time information generation device 105 generates route candidate information from the current position of the vehicle M40 to the destination position using a known method. The predicted travel time information generation device 105 generates the generated predicted travel time information, the generated route candidate information to the destination, the real-time traffic information i11 output from the probe traffic information collection device 101, and the statistical information processing device 104. Using the predicted model, the bottleneck link is extracted from the route of the route candidate information as will be described later. The bottleneck link will be described later.
The predicted travel time information generation device 105 determines the time when the bottleneck link on the route ends in the extracted bottleneck link (hereinafter referred to as the predicted traffic end time), the predicted travel time information, the route candidate information, and Estimate using. The predicted travel time information generation device 105 outputs the generated predicted travel time information and the estimated route candidate information to the dynamic marginal cost information generation device 106.
Note that the predicted travel time information generation device 105 uses the prediction model generated by the statistical information processing device 104 to calculate the real-time traffic information i11 output from the probe traffic information collection device 101, for example, for every five minutes. Generate travel time information. The period in which the predicted travel time information generation device 105 generates the predicted travel time information may be selected according to the time zone to be predicted, season, place, and the like, for example.

  The dynamic marginal cost information generating device 106 (dynamic marginal cost information generating unit) uses the predicted travel time information for each link, the route candidate information, the estimated congestion end time, and the bottleneck / link as described later. Calculate the dynamic marginal cost of the route. Based on the input route candidate information, information indicating the dynamic limit cost and predicted travel time information i16, and the calculated dynamic limit cost of each route, the dynamic margin cost information generation device 106 predicts the estimated travel time of each route candidate. Calculate (second estimated passage time) and dynamic marginal cost. The dynamic marginal cost information generation device 106 outputs information i17 including the calculated estimated travel time and dynamic marginal cost of each route candidate to the Pareto improvement route guidance device 107.

The Pareto improvement route guide device 107 (Pareto improvement route guide unit) is based on the predicted travel time and the dynamic limit cost of each route candidate output from the dynamic marginal cost information generation unit 106, as will be described later. The degree of influence on the route candidate is calculated. The Pareto improvement route guide device 107 is based on the predicted travel time and the dynamic limit cost of each route candidate output from the dynamic marginal cost information generation unit 106, and the route that the vehicle M40 is currently traveling and the generated routes Each candidate travel time and each dynamic marginal cost will be compared sequentially. The Pareto improvement route guidance device 107 searches for a route candidate that minimizes the dynamic marginal cost as will be described later.
The Pareto improvement route guide device 107 searches the calculated route candidates for a route candidate whose predicted travel time is less than or equal to the current route and whose dynamic margin cost is less than or equal to the current route and has the smallest dynamic margin cost. . The Pareto improvement route guidance apparatus 107 determines whether there is a recommended route candidate that has a predicted travel time that is less than or equal to the current route and that has a dynamic limit cost that is less than or equal to the current route and that minimizes the dynamic limit cost. When it is determined that there is a recommended route candidate, the Pareto improvement route guide device 107 transmits recommended route information i18 as a search result to the vehicle M40 equipped with the traffic information terminal. When it is determined that there is no recommended route candidate, the Pareto improvement route guide device 107 does not transmit the search result to the vehicle M40 equipped with the traffic information terminal.

Next, the dynamic limit cost in the route search algorithm will be described.
Route search algorithms used in general car navigation systems and traffic information centers are calculated based on the cost of links. However, the dynamic marginal cost is not defined in units of links but is a concept defined in units of travel routes. For this reason, the dynamic marginal cost cannot be calculated unless it is calculated based on the cost of the link and the route search result is used. Furthermore, since a vehicle on a general road is affected by the control by a traffic light, the travel time (also referred to as travel time) is delayed to some extent even when there is no traffic. Therefore, most routes will always have a travel time with a delay from the simulation start time to the end time. As a result, it is difficult to calculate the marginal cost strictly with a general route search algorithm.
For this reason, in this embodiment, the dynamic limit cost is approximately calculated for each link described in FIG. The reason for the approximate calculation is: “Dynamic marginal cost is the sum of the time required for a vehicle to exit the traffic jam section and the time it takes for the traffic to clear from the time the vehicle arrives at the end of the traffic jam. It is shown by research by the inventors of the present invention (Masao Kuwahara, Ikuo Yoshii, Kotaro Kumagai: Study on optimal allocation of dynamic system and ramp inflow control-Basic analysis in simplified network-, Civil engineering (See Journal of Society, No.667 / IV-50, pp.59-71, Japan Society of Civil Engineers, 2001.01).
Note that the dynamic limit cost of the route in this embodiment is required to pass the difference between the time when the vehicle arrives at the traffic jam section on the route and the time until the traffic is resolved thereafter, that is, the traffic jam. Equal to time.

FIG. 4 is a diagram illustrating the number of vehicles flowing into the link and the number of vehicles flowing out from the link. In FIG. 4, one link will be described for the sake of simplicity.
In FIG. 4, the horizontal axis represents time, and the vertical axis represents cumulative traffic. A curve g101 represents a change in the inflow amount (number of vehicles) to the link with respect to time. A symbol g102 represents a change in the outflow amount (number of vehicles) from the link with respect to time. The curve g111 period _{t b} from the time _{t a} denotes the dynamic marginal cost at time _{t a.} A broken line g131 represents the Nth unit.
As shown in FIG. 4, when the inflow amount to the link (curve g101) is different from the outflow amount from the link (curve g102), the outflow is delayed at this link, that is, the traffic congestion occurs. I mean.
In this case, as indicated by a broken line g131, the Nth vehicle that has flowed in at time ta flows out of this link at the Nth vehicle. Therefore, the time _tab in FIG. 4 represents the travel time from when the Nth vehicle enters the traffic jam until it gets out of the traffic jam.
The curve g111, the traffic jam that is happening in this link, in the previous time from the time t _a, which indicates whether congestion to which the vehicle is continuing.

Next, the limit cost of the route will be described with reference to FIG.
FIG. 5 is a diagram for explaining the marginal cost of the route according to the present embodiment. FIG. 5A illustrates a route. FIG. 5B is a diagram for explaining the marginal cost in the route traffic volume (route flow). FIG.5 (c) is a figure explaining the marginal cost in the traffic volume (link flow) of a link unit. In FIG. 5B, the horizontal axis represents time, and the vertical axis represents the accumulated traffic volume of the route k. In FIG. 5C, the horizontal axis represents time, and the vertical axis represents link traffic.
In FIG. 5A, reference numerals p201 to p205 represent the positions of the path k and the path j. Reference numerals l201 to l206 denote links of the path k and the path j. In FIG. 5A, the moving direction of the vehicle is a direction from a position p201 (upstream) to p204 (downstream). Also, link 1201 is the i-th link of route k, links 1202 and 1205 are the i + 1-th link of route k, and links 1203 and l206 are the i + 2-th link of route k.
As shown in FIG. 5A, on the route k, the vehicle flows in from the position p201, passes through the link 1201, and flows out of the position 204 through the links 1202 and l203. On the path j, the vehicle flows in from the position p205, passes through the link l205, and flows out of the position 204 through the link l206. In this case, the link 202 of the route k and the link 1205 of the route j merge at the position p203. That is, the links l203 and l206 are the same road.

Here, in the case of a simple network, as shown in FIG. 5B, the dynamic marginal cost of the route is calculated based on the accumulated traffic curve and the passage at the bottleneck link for the traffic volume (route flow) of a route. It is obtained by drawing a cumulative curve and calculating it graphically. In FIG. 5B, a curve A ^k _i represents a change with time of the traffic volume (the number of vehicles) flowing in (also referred to as arrival traffic volume) from the upstream position p201 in the link 1201. A curve A ^k _{i + 1} represents a change with time of the traffic flowing in from the upstream position p202 in the link 1202. A curve A ^k _{i + 2} represents a change with time of the traffic flowing in from the upstream position p203 in the link l203. A curve D ^k _{i + 2} represents a change with time of the traffic flowing out from the downstream position p204 (bottleneck link) in the link l203.
Also, broken lines g201 and g202 schematically represent an inflow amount equivalent to the link 1201 to the link 1203. That is, the number of vehicles indicated by a broken line g201 flows into the link 1201. However, in the link 1203, only the curve A ^k _{i + 2} indicated by the solid line flows in, indicating that the difference vehicle is congested. Also, in the curve D ^k _{i + 2} indicating the outflow amount, only the number of vehicles lower than the inflow amount shown by the broken line g202 can pass, indicating that the difference vehicles are congested.
Curve d201, similar to that described in FIG. 4, represents the marginal cost of a path k at time t _1.

In this embodiment, as shown in FIG. 5 (c), a cumulative figure of traffic volume (link flow) is drawn not for each route but for each link, and the bottleneck link located at the end of the congestion section on a certain route is drawn. The dynamic marginal cost is simply obtained from the cumulative chart.
A curve A _{i in} FIG. 5C represents a change with time of the traffic flowing in from the upstream position p201 in the link 1201.
A curve A _{i + 1} represents a change with respect to time of the traffic volume obtained by adding the traffic volume flowing in from the upstream position p202 in the link 1202 and the traffic volume flowing in from the upstream position p205 in the link l205. A curve A _{i + 2} represents a change with time of the traffic flowing in from the upstream position p203 in the link l203 and the link 206. A curve D _{i + 2} represents a change in the traffic volume flowing out from the downstream position p204 in the link 1203 with respect to time.
A broken line g211 schematically represents an inflow amount equivalent to the link 1201 to the link 1203.
Curve d202 represents the marginal cost of path k in links l203 and l206 with delay.
Further, as shown in FIG. 5A and FIG. 5C, since the traffic jam has ended at the link 1203, this link 1203 is defined as a bottleneck link.

In FIG. 5 (c), the better the slope of the curve _{A i ~A i + 2, D} i + 2 is large, indicating that heavy traffic. In addition, the slope of the curve D _{i + 2} is smaller than the slope of the curve A _i , even if the amount of traffic that has arrived is large, the amount of traffic that can be passed through the route is less than the amount of traffic that has arrived. . Here, the slope of the curve _{Di + 2} is defined as the bottleneck capacity. The bottleneck capacity is a traffic volume that can pass per unit time.
Therefore, the deviation amount between the slope and the slope of the curve A _{i + 1} of the curve A _i represents the deviation of the inflow traffic actually flows traffic volume, i.e., the traffic volume that remains from the section i in the interval i + 1 Represents.
Similarly, the divergence between the slope of the curve A _{i + 1 and} the slope of the curve A _{i + 2} represents the divergence between the inflow traffic volume and the actual traffic volume, that is, the traffic volume accumulated from the section i + 1 to the section i + 2. Represents.

In the case of the route described with reference to FIG. 5, the traffic jam is simply resolved on a link basis from the above-described time required for a vehicle to exit the traffic jam section and the time when the vehicle arrives at the end of the traffic jam. For example, if a bottleneck link is congested to the upstream link, the marginal cost will be It is calculated redundantly.
For this reason, in this embodiment, for a section where links with delays continue, the marginal cost, that is, the time from the inflow time of the link to the prediction cancellation time of the traffic congestion is calculated only at the most downstream link of the section. For other traffic jam links, the required travel time is calculated. As a result, duplication is eliminated and an approximate value of the dynamic marginal cost can be obtained by calculation in link units.

Next, the traffic jam prediction method will be described with reference to FIGS.
FIG. 6 is a diagram for explaining the dynamic marginal cost according to the present embodiment. In FIG. 6, the horizontal axis represents time, and the vertical axis represents the link travel speed, which is the moving speed of the vehicle in the link. A broken line g301 represents a threshold for determining a traffic jam. The solid line g311 represents the change in the measured value of the link travel speed with respect to time during the period of t ₃₀₁ from time 0, the dashed line g312 represents the change in the prediction value of the link travel speed with respect to time in the after time t _301. Time t ₃₀₁ is the current time.
Predicted travel time information generating unit 105 is input with real-time traffic information is a probe traffic information, using a predictive model to predict the link travel speed of the current time t ₃₀₁ after.

As shown in FIG. 6, the predicted travel time information generating apparatus 105 considers a traffic jam when the link travel speed g311 is smaller than a predetermined traffic jam determination threshold g301. In FIG. 6, the predicted travel time information generating apparatus 105 determines that traffic congestion has occurred during the period from time t ₃₀₁ to time t ₃₀₂ after the current time. As described above, in this embodiment, the period (t ₃₀₁ to t ₃₀₂ ) when the predicted value g 312 of the link travel speed is less than the traffic congestion determination threshold g ₃₀₁ is defined as the dynamic marginal cost.

Next, determination of a link having a traffic jam will be described with reference to FIG.
FIG. 7 is a diagram for explaining a traffic jam section according to the present embodiment. In FIG. 7, the traveling direction of the vehicle is the x direction, and the direction perpendicular to the x direction is the y direction. In FIG. 7, reference numeral r401 represents a road extending in the x direction, and reference numerals c401 to c404 represent roads extending in the y direction perpendicular to the road r401. The intersections of roads r401 and c401, the intersections of roads r401 and c402, the intersections of road r401 and road c403, and the intersections of roads r401 and c404 represent intersections.
In FIG. 7, a section l401 from 0 to x402 in the x direction, a section l402 from x402 to x404, and a section l403 from x404 to x405 are each a section of a traffic light and a traffic light and are links.

As shown in FIG. 7, a symbol g401 is a section from x401 to x402 in the x direction, and a section where a speed reduction occurs. The predicted travel time information generation apparatus 105 does not consider traffic congestion because the speed decrease section (hereinafter referred to as speed decrease section) g401 is shorter than the link 1401.
Further, as illustrated in FIG. 7, a symbol g402 is a section from x403 to x405 in the x direction, and a section in which a speed decrease occurs. The predicted travel time information generating apparatus 105 regards the traffic as a traffic jam because the speed decrease section g402 in which the speed decreases in this way is longer than the link l403. In this case, the section regarded as a traffic jam includes not only the link l403 but also the link l402. For this reason, the predicted travel time information generating apparatus 105 connects continuous speed reduction sections and regards it as one speed reduction section when the speed reduction section straddles a plurality of links.
Note that a unit section for predicting travel speed (hereinafter referred to as a travel speed prediction unit section) has a distance of about a signalized intersection. Since the distance at the signalized intersection is about several hundred meters, the unit section of the travel speed prediction should be longer than the staying length (about 200m at the longest) that makes the amount of vehicles staying in one cycle of the signal. desirable.
For example, when a section of 300 m or more is predicted to be less than 20 km / h, the predicted travel time information generation device 105 predicts that a traffic jam will occur. This is because when the length of the section is 300 m or more, the vehicle is in a state where two or more cycles of the signal can be passed and it is not a decrease in speed due to signal waiting.

Next, a method for identifying a bottleneck link will be described with reference to FIG. It is not simple to specify where the bottleneck link that is the end of the traffic jam is. For this reason, in this embodiment, the past traffic volume and traffic jam data are accumulated, and the location where the traffic jam occurs is extracted in advance.
FIG. 8 is a diagram illustrating an example of a bottleneck / link specifying method according to the present embodiment. FIG. 8A is a diagram showing a section from the x-axis direction x403 to x405 in FIG. FIG. 8B is a diagram showing the section of FIG. In FIG. 8A, the traveling direction of the vehicle is the x direction, and the direction perpendicular to the x direction is the y direction.
In FIG. 8B, a symbol p402 is an intersection between the roads r401 and c403 in FIG. 8A, and a symbol p403 is an intersection between the roads r401 and c404 in FIG.
As shown in FIG. 8B, each input of links 1402, 1402-2, and 1402-2 is connected to the point p402. The output of the point p402 is connected to the link l403. Similarly, the inputs of links l403, l403-1, and l403-2 are connected to the point p403. Further, the output of the point p403 is connected to the link l404.

For example, the predicted travel time information generation device 105 of the route guidance system 1 uses external information via the statistical information processing device 104 for data relating to the link travel speed for 5 minutes among traffic data for the past month to several months. Read from the storage unit 103. The 5-minute link travel speed data is 5-minute vehicle travel speed data measured for each link.
As shown to Fig.8 (a), the link travel speed of the area g401 is less than 10 km / h (1st speed). For this reason, the estimated travel time information generation apparatus 105 determines that traffic congestion has occurred in the section of links l402 and l403 on the road r401. On the other hand, in the section g402 (link l404) after c404 in the x direction of the road r401, the link travel speed is 30 km / h (first speed) or more. For this reason, the estimated travel time information generation apparatus 105 determines that no traffic jam occurs on the link l404 on the road r401. Since traffic congestion has occurred at links l402 and l403 but not at link l404, the predicted travel time information generation apparatus 105 sets the link where the traffic congestion has occurred as a bottleneck link, and adds 1 point to this link. .
In this way, the predicted travel time information generation apparatus 105 compares travel speed data between adjacent links for all links to be analyzed.
The predicted travel time information generation apparatus 105 adds the given points in this way, and extracts a link having an average value of 6 to 12 points per day as a bottleneck link. That is, the predicted travel time information generating apparatus 105 determines the downstream point that is the bottleneck in this way.

Next, the influence degree of the bottleneck link on the route will be described with reference to FIG.
In general roads, a plurality of roads are subordinately connected as shown in FIG. For this reason, in this embodiment, weighting is also performed on how much route one bottleneck affects.
In this embodiment, based on past statistical data, the average traffic congestion information is used to count how many roads are connected to the bottleneck and the number of connected roads, thereby determining the degree of influence. Calculate as follows.

A state where several roads as shown in FIG. 9 intersect each other will be described.
FIG. 9 is a diagram for explaining the degree of influence of the bottleneck link on the route according to the present embodiment.
In FIG. 9, the left-right direction is the x direction, and the direction perpendicular to the x direction is the y direction. In FIG. 9, reference numerals r501, r502, and r503 are roads parallel to the x direction, and reference numeral c501 is a road perpendicular to the roads r501, r502, and r503. Reference symbol c502 is a road that intersects the road r503 and merges with the road r502.
Reference numerals l501, l502, l503, and l504 each represent a bottleneck link. Reference numerals p501, p502, p503, p504, p505, p506, p507, and p508 represent points on the road or intersections.
A point p501 is one end of the road r502 in the negative direction of the x axis, and is an inflow point of the bottleneck link l501. Point p502 is one end of the negative direction of the x-axis of road r503, and is an inflow point of bottleneck link l502. Point p503 is one end of the negative direction of the y-axis of road c502, and is an inflow point of bottleneck link l503. Point p504 is one end of the negative direction of the y-axis of road c501, and is an inflow point of bottleneck link l502. Point p505 is the intersection of road r503 and road c502, and is a passing point on bottleneck link l503. A point p506 is an intersection of the road r503 and the road c501, and is a passing point on the bottleneck link l504. A point p507 is an intersection of the road r502 and the road c502, a junction of the road r502 and the road c502, and a passing point on the bottleneck links l501, l502, l503, and l504.
Further, in the bottleneck link l504, the negative direction of the y axis is referred to as downstream, and the positive direction of the x axis is referred to as upstream.

As shown in FIG. 9, the bottleneck link l501 is composed of a section between points p501 and p507 and a section between points p507 and p508. Further, the bottleneck link l501 starts to be congested from the point p501, on the road r502 from the point p501 to the point p507 in the positive x-axis direction, and on the road c501 from the position p507 to the position p508 in the positive y-axis direction. Congestion continues at the link.
The bottleneck link l502 includes a section between the points p502 and p505, a section between the points p505 and p507, and a section between the points p507 and p508. Further, the bottleneck link l502 starts to be congested from the point p502, is a link from the point p502 to the point p505 in the positive direction of the x-axis on the road r503, and the road c502 in the direction of the angle α with respect to the x-axis from the position p505 Congestion continues on the link to the position p507 and the link from the position p507 to the position p508 in the positive direction of the y-axis on the road c501.
The bottleneck link l503 is composed of a section between points p503 and p505, a section between points p505 and p507, and a section between points p507 and p508. Further, the bottleneck link 1503 starts to be congested from the point p503, and two links from the point p503 to the position p507 through the point p505 in the direction of the angle α with respect to the x axis on the road c502, from the position p507 on the road c501. Congestion continues at the link to the position p508 in the positive direction of the y-axis.
The bottleneck link l504 includes a section between the points p504 and p506, a section between the points p506 and p507, and a section between the points p507 and p508. The bottleneck link l504 is congested from the point p504, and two links from the point p504 to the position p507 via the point p506 in the y-axis positive direction on the road c501, and from the position p507 to the positive y-axis on the road c501. Congestion continues on the link up to position p508 in the direction.

As described with reference to FIG. 8, the predicted travel time information generation device 105 determines whether or not there is traffic on each link when the vehicle passes the link. Specifically, the predicted travel time information generation device 105 reads, for example, data related to the link travel speed for 5 minutes from the external information storage unit 103 from the traffic data for the past month to several months. The predicted travel time information generating apparatus 105 determines whether or not the link travel speed is less than 10 km / h (first speed).
The predicted travel time information generation apparatus 105 extracts one link, the link travel speed of the extracted link is less than 10 km / h, and any link travel speed connected to the link downstream is also less than 10 km / h. In the case of, 1 point is given to the extracted link. Note that the points used here are different from the points described in FIG.
The estimated travel time information generation device 105 regards links calculated in this way having an average of about 6 to 12 points per day as links (terminal roads) that are subordinate to traffic congestion from the bottleneck. .

Specifically, in FIG. 9, points p507 to p508 are merging points and are upstream. As shown in FIG. 9, this link includes a link from point p501 to p507, a link connected from point p502 to point p507 through point p505, a link connected from point p503 through point p505 to point p507, The four links connected from the point p504 to the point p507 via the point p506 are connected to the upstream point p507 to p508.
Then, the Pareto improvement route guide device 107 calculates the number of graph structures of the subordinate links upstream of the bottleneck link. In FIG. 9, since there are four subordinate links in the section from points p507 to p508 upstream of the bottleneck link, the Pareto improvement route guide device 107 adds the influence degree 1 for each link, and the influence degree 4 Is defined as The implication of high impact means that the traffic jam rate is high.
The influence degree shown in FIG. 9 is generated based on data acquired from the probe. When weighting is performed for all the links in this way, the predicted travel time information generating apparatus 105 extracts, for example, links weighted on the average of about 6 to 12 points per day as bottleneck links. To do. When a sensor capable of measuring the moving speed of the vehicle is installed, the Pareto improvement route guide device 107 may similarly extract the bottleneck link using the speed data acquired by the sensor. .
Note that a link weighted by 6 points may be a road where traffic congestion has continued for 30 minutes (5 minutes × 6). Similarly, a link weighted by 12 points indicates that there is a possibility that the road is congested for 60 minutes (5 minutes × 12).

Next, the processing procedure of the route guidance system 1 will be described with reference to FIG. FIG. 10 is a flowchart of the processing procedure of the route guidance system 1 according to the present embodiment.
First, the outline of the process performed below will be described.
The route guidance system 1 searches for a route that is optimal for all the vehicles. Then, for each vehicle route, each link of the route has a merit in terms of cost (travel time) for the vehicle, and there is also a merit in terms of system cost. Is determined as recommended route information for the vehicle.

(Step S101) The probe traffic information collection device 101 acquires the probe traffic information i10 transmitted by the probe vehicle M30.
Next, the probe traffic information collecting apparatus 101 uses the received probe traffic information, the traffic information from sensors installed on the road, the past traffic information, and the like to detect the current traffic volume and vehicle by a known method. Real-time traffic information i11 indicating the travel speed and the like is generated. The probe traffic information collection device 101 outputs the generated real-time traffic information i11 to the predicted travel time information generation device 105, and further stores the generated real-time traffic information i11 in the traffic information storage device 102. After step S101 ends, the process proceeds to step S102.

(Step S102) The statistical information processing apparatus 104 reads real-time traffic information i12 stored in the past including the real-time traffic information i11 stored in step S101 in the traffic information storage apparatus 102.
Next, the statistical information processing apparatus 104 reads the traffic information i13 stored in the external information storage unit 103.
Next, the statistical information processing apparatus 104 generates a prediction model based on the read real-time traffic information i12 and traffic information i13. After step S102, the process proceeds to step S103.

(Step S103) The predicted travel time information generating apparatus 105 uses the prediction model generated by the statistical information processing apparatus 104 to calculate the predicted travel time information for each link from the real-time traffic information i11 output by the probe traffic information collecting apparatus 101. Generate. After step S103 ends, the process proceeds to step S104.

(Step S104) The predicted travel time information generating apparatus 105 obtains information i15 including the current route information of the vehicle transmitted by the vehicle M40 equipped with a traffic information terminal (not shown) and the destination information where the vehicle is traveling. get. After step S104 ends, the process proceeds to step S105.

(Step S105) The predicted travel time information generating apparatus 105 generates route candidate information from the current position of the vehicle M40 to the destination position using a known method based on the acquired current route information and destination information. After step S105 ends, the process proceeds to step S106.

(Step S106) The predicted travel time information generation apparatus 105 uses the route candidate information generated in step S105 and the predicted time information generated in step S103, from the route candidate information route as described above. Extract bottleneck links.
Specifically, the predicted travel time information generation device 105 extracts a traffic jam section using the predicted time information generated as described with reference to FIG. 7 at the time when the vehicle reaches each link, and extracts the extracted section. Is determined to be congested in a section predicted to be less than a predetermined travel speed (for example, 20 km / h). In addition, the estimated travel time information generation apparatus 105 performs this determination, for example for every time. Next, as described with reference to FIG. 8, the predicted travel time information generation device 105 extracts links (or sections) in which the average points given per day are greater than a predetermined value as bottleneck links. To do. After step S106 ends, the process proceeds to step S107.

(Step S107) The predicted travel time information generating apparatus 105 generates the estimated travel time information of the bottleneck link on the route in the extracted bottleneck link, in step S105, and the predicted travel time information generated in step S103. It estimates using the route candidate information made. For example, as described with reference to FIG. 5, the predicted travel time information generation apparatus 105 estimates the expected end time of the congestion of the bottleneck link based on the graph of the time change with respect to the accumulated traffic volume of the route of each link. The predicted travel time information generation device 105 outputs the estimated bottleneck link congestion end predicted time, predicted travel time information, and the generated route candidate information to the dynamic marginal cost information generation device 106. After step S107 ends, the process proceeds to step S108.

(Step S108) The dynamic marginal cost information generating device 106 has a traffic jam in each route candidate based on the input estimated traffic jam end time of the bottleneck link, the predicted travel time information, and the generated route candidate information. Calculate the dynamic marginal cost of the link. The dynamic marginal cost information generation device 106 outputs the calculated dynamic marginal cost of the link where the traffic congestion has occurred in each route candidate to the Pareto improvement route guidance device 107. After step S108 ends, the process proceeds to step S109.

(Step S109) The Pareto improvement route guide device 107 uses FIG. 9 based on the dynamic limit cost and the estimated travel time of the link where the congestion occurs in each route candidate output from the dynamic limit cost information generation device 106. As described above, the degree of influence of each bottleneck link on the route candidate is calculated.
Next, the Pareto improvement route guide device 107 is based on the dynamic limit cost and the estimated travel time of the link where the traffic jam occurs in each route candidate output from the dynamic limit cost information generation device 106. The travel route, the predicted travel time of each generated route candidate, and the dynamic marginal cost are sequentially calculated. After step S109 ends, the process proceeds to step S110.

(Step S110) Next, the Pareto improvement route guidance apparatus 107 sequentially compares the values calculated in step S109 to search for a route candidate that minimizes the dynamic marginal cost.
The Pareto improvement route guidance device 107 compares the predicted travel time information of the current route with the predicted travel time information of each route candidate, and searches for a route candidate that is less than or equal to the predicted travel time of the current route (first search). Further, the Pareto improvement route guide device 107 compares the dynamic limit cost of the current route with the dynamic limit cost of each route candidate, and searches for a route candidate equal to or less than the dynamic limit cost of the current route (second search). . As a result of performing the first search and the second search, the Pareto improvement route guide device 107 calculates the route having the smallest dynamic marginal cost using the following equation (1).

c _k (t) = a ₁ T _k (t) + a ₂ M _k (t) (1)

In Equation (1), c _k (t) is the cost at each time t in the route candidate k. T _k (t) is, for example, the time required for passing a link (link travel time) generated from probe traffic information or information obtained from a sensor. M _k (t) is the travel time in the bottleneck link calculated by the following equation (2).

M _k (t) = (time until congestion is resolved) × degree of influence (2)

Note that M _k (t) in equation (2) is calculated only for the bottleneck link.
In the formula (1), a ₁ and a ₂ are predetermined constants. In Equation (1), the state where the constant a ₁ = 0 and the constant a ₂ = 1 is the user optimum state. In Equation (1), the state where the constant a ₁ = 1 and the constant a ₂ = 0 is the system optimum state. Therefore, if the constants a ₁ and a ₂ are balanced, the Pareto optimal state can be obtained.
For the balance between the constants a ₁ and a ₂ , for example, depending on how much of the vehicle using the route guidance information generated in this way is calculated numerically by simulation to calculate a ratio according to the actual situation. After step S110 ends, the process proceeds to step S111.

(Step S111) Next, the Pareto improvement route guide device 107 selects a dynamic limit whose predicted travel time is less than or equal to the current route and whose dynamic limit cost is less than or equal to the current route from among the route candidates calculated in Step S110. It is determined whether there is a recommended route that is a route candidate that minimizes the cost. If it is determined that there is a recommended route (step S111; Yes), the process proceeds to step S112. If it is determined that there is no recommended route (step S111; No), the process proceeds to step S113.

(Step S112) When it is determined that there is a recommended route (Step S111; Yes), the Pareto improvement route guide device 107 transmits the recommended route information as a search result to the vehicle M40 equipped with the traffic information terminal. In this case, for example, the user of the vehicle M40 changes the travel route from the route in the user optimum state searched by the conventional car navigation system to the recommended route transmitted from the route guidance system 1.

(Step S113) When it is determined that there is no recommended route (Step S111; No), the Pareto improvement route guide device 107 does not transmit the search result to the vehicle M40 equipped with the traffic information terminal. In this case, for example, the user of the vehicle M40 continues to travel on the route of the user optimum state searched by the conventional car navigation system.

In the cost calculation by the conventional car navigation system, only the term of T _k (t) is calculated in the equation (1). However, in the present embodiment, there is a difference in calculating the dynamic limit cost for the bottleneck link of the two items M _k (t) in the equation (1). Moreover, in this embodiment, the dynamic marginal cost by a bottle link is also considered based on travel time like the prior art.
That is, in the present embodiment, when the own vehicle selects this route, it is taken into account how much influence is exerted on other vehicles.

Next, an example of the effect of the route guidance system 1 according to the present embodiment will be described with reference to FIGS.
FIG. 11 is a map used for simulation in the route guidance system 1 according to the present embodiment. In FIG. 11, the left-right direction is the x direction, and the direction perpendicular to the x direction is the y direction. The area shown in FIG. 11 has a length of 10 km in the x direction and a length of 20 km in the y direction, and the number of road links in this area is 7600. The traffic volume of vehicles is 270,000 in 3 hours.

FIG. 12 shows the result of simulating the route guidance system 1 according to this embodiment in the area shown in FIG.
In FIG. 12, the line indicated by reference numeral g601 indicates the number of vehicles changed from the route guidance in the user optimum state (DUO) 75% to the recommended route (DPO) presented by the route guidance system 1, and the system optimum state (DSO). ) Represents the number of vehicles. The row indicated by reference sign g602 represents the ratio of the number of vehicles changed from the route guidance of DUO 75% to the recommended route presented by the route guidance system 1, and the ratio of the number of vehicles in the DSO to the total number.
The row indicated by reference numeral g603 represents the total travel time (= number of vehicles × travel time) in each of the DUO 75%, DPO, and DSO states.

The line indicated by g604 indicates the difference between the total travel time in DUO 75% and the total travel time in DPO, and the difference between the total travel time in DUO 75% and the total travel time in DSO is the percentage of the total travel time in DUO 75%. It represents whether or not.
The row indicated by reference numeral g605 is changed from the route guidance of DUO 75% to the DSO from the total travel time reduced per vehicle in the vehicle changed to DPO presented by the route guidance system 1 and the route guidance of DUO 75%. It represents the total travel time reduced per vehicle in the vehicle.
The line indicated by reference numeral g606 represents the total emission amount of CO ₂ per three hours in each state of DUO 75%, DPO, and DSO.
The line indicated by reference numeral g607 represents DUO75% of CO ₂ DPO CO ₂ emissions reduction rate with respect to emissions, the CO ₂ emissions reduction rate of DSO for DUO75% of CO ₂ emissions.

  The reason for comparing DUO 75% is that when the vehicle equipped with car navigation is set to 100% (DUO 100%), for example, in the user optimum state, all vehicles are presented with the optimal It goes to the route. As a result of performing a simulation for a case where the vehicle equipped with the car navigation is set to 100% (DUO 100%), for example, the efficiency deteriorated. As a result of the simulation, the state of DUO 75% was the most efficient when the route guidance system 1 of this embodiment and the DSO were not used, so the state of DUO 75% was set as a comparison value.

As shown in FIG. 12, from the state of 75% DUO, the vehicle whose route was changed was 1.7% for DPO (this embodiment) and 2.4% for DSO (lines g601, g602). ). In terms of the number of units, the number of vehicles that have changed their routes from 75% of DUO is more in DSO.
However, when compared with the total travel time, the reduced travel time per changed vehicle is more in the DPO (lines g603 to g605).
Furthermore, the CO ₂ reduction amount of DPO with respect to DUO 75% is 0.98%.
As described above, in the route guidance system 1 according to the present embodiment, it is suggested that when the route is changed from the user optimum state, the travel time can be reduced, thereby reducing the CO ₂ exhaust amount.

FIG. 13 is a diagram illustrating changes in the number of vehicles and travel time when there is a route change in the system optimum state. In the case of a vehicle without a route change, the route in the user optimum state set by the car navigation system is selected.
In FIG. 13, the horizontal axis represents the travel time change time, and the middle 0 seconds on the horizontal axis means that the travel time did not change. In FIG. 13, the vertical axis represents the number of vehicles.
As shown in FIG. 13, in the case where there is a route change, the change in travel time is distributed almost symmetrically and in a normal distribution with respect to the travel time of 0 seconds. In this case, as shown in FIG. 13, there is a user whose travel time is shortened by 300 seconds as compared with the case of the user optimum state, while there is a user whose travel time is increased by 300 seconds. Furthermore, as indicated by the broken-line long circle g701, it indicates that there are 1000 or more users whose travel time has increased.

FIG. 14 is a diagram showing the number of vehicles and the travel time change when there is a route change in the Pareto optimal state according to the present embodiment. In the case of a vehicle without a route change, the route in the user optimum state set by the car navigation system is selected.
In FIG. 14, the horizontal axis represents the travel time change time, and the middle 0 seconds on the horizontal axis means that the travel time did not change. In FIG. 14, the vertical axis represents the number of vehicles.
As shown by a broken line ellipse g702 in FIG. 14, it is greatly different from FIG. 13 and represents that the number of users whose travel time is increasing is about several hundreds. On the other hand, the user whose travel time is shortened is equivalent to the system optimum state. That is, according to the route guidance system 1 of the present embodiment, for the user who has changed the route, the travel time is almost always the same or shortened, and the case where the travel time increases is small. ing.

As described above, in the route guidance system 1 of the present embodiment, the dynamic marginal cost information generation device 106 uses the current route and the route based on the information indicating the link section where the traffic jam occurs and the information indicating the time when the traffic jam ends. Calculate each candidate dynamic marginal cost. Then, the Pareto improvement route guidance device 107 compares the predicted travel time and each dynamic margin cost of the current route and each route candidate based on each predicted travel time and each dynamic margin cost of each route candidate, As a result, a route candidate that minimizes the dynamic marginal cost is searched.
As a result, the traffic congestion mitigation efficiency can be reduced by reducing the cost for the user from the user optimum state with the shortest route and further reducing the increase in the travel time of the user from the system optimum state with the lowest cost for the route guidance system 1. Can be improved.

In the present embodiment, the following processing is performed in the calculation of the dynamic marginal cost.
The predicted travel time information generation device 105 extracts one link from all links on the current route and on the route candidate, and the moving speed in the future up to the arrival time of the vehicle on the extracted link to the destination of the vehicle is determined in advance. When it is determined that the speed is lower than the first speed, the extracted link is predicted to be a link where traffic congestion occurs.
Further, the predicted travel time information generation apparatus 105 extracts a bottleneck link at which the traffic jam ends based on information indicating a link where the traffic jam occurs, and the bottles on the current route and the route candidate in the extracted bottleneck link. The time at which the congestion of the neck link ends is estimated using predicted travel time information, current route information, and route candidate information for each link.
In addition, the predicted travel time information generation device 105 is a second speed at which a future moving speed up to the arrival time of the vehicle at all links connected downstream of the link where the traffic jam occurs is determined in advance. If it is determined that the traffic congestion occurs, the link where the traffic jam occurs is weighted. If the value indicating the weight of the link where the traffic jam occurs is larger than a predetermined value, the link where the traffic jam occurs is extracted as a bottleneck link. I did it.
Further, the Pareto improvement route guidance device 107 calculates the influence degree of the bottleneck link based on the number of terminals in the graph structure of the link subordinate to the upstream side of the bottleneck link, and uses the calculated influence degree. To calculate the dynamic marginal cost.
As a result, in this embodiment, the bottleneck link that is the link where the traffic jam ends can be identified, and the influence on the route of the bottleneck link is calculated to calculate the dynamic marginal cost. Therefore, it is possible to improve the traffic congestion mitigation efficiency that is more suitable for the user than the user optimum state and the system optimum state.

  In this embodiment, the example using the traffic information transmitted from the probe vehicle has been described. For example, a sensor (not shown) collects the moving speed information of the vehicle and transmits it to the probe traffic information collecting apparatus 101. May be.

DESCRIPTION OF SYMBOLS 1 ... Route guidance system, 101 ... Probe traffic information collection apparatus, 102 ... Traffic information storage apparatus, 103 ... External information storage part, 104 ... Statistical information processing apparatus, 105 ... Prediction Travel time information generating device, 106 ... Dynamic marginal cost information generating device, 107 ... Pareto improvement route guide device, l201 to l206, l401 to l404, l501 to l504 ... link, M30 ... probe vehicle

Claims (11)

Calculating a first estimated passing time required for the vehicle to pass through a link that is a predetermined section based on the moving speed of the vehicle;
Calculating a second predicted passage time required for the vehicle to pass the route from the current travel position to the target position based on the calculated first predicted travel time;
The route guidance device, wherein the route is selected based on the calculated second estimated passage time. A traffic information collecting unit for acquiring traffic information including a moving speed of the vehicle;
A link on the route is extracted, a link where traffic congestion has occurred by the time when the vehicle arrives at a target position is extracted based on the moving speed of the vehicle, and the traffic congestion occurs based on the moving speed of the vehicle A predicted time information generating unit that generates a first predicted predicted time required for the vehicle to pass through a link;
The route guidance apparatus according to claim 1, further comprising: A route candidate information generating unit that generates route candidate information indicating a route candidate from the current travel position of the vehicle to the target position;
A dynamic marginal cost information generating unit that calculates a first dynamic marginal cost based on the first predicted predicted passage time for a link in which traffic congestion occurs before the time when the vehicle arrives at the target position in the route candidate;
A second dynamic marginal cost is calculated based on a predicted passing time required for the vehicle to pass through the link in which no congestion has occurred in the route candidate, and the second dynamic limit cost is calculated for all links of the route candidate. The total marginal cost and the first dynamic margin cost calculated by the dynamic margin cost information generation unit are summed for each of the route candidates, and the summed dynamic margin costs are compared. A Pareto improved route guide for searching for the route candidate with the smallest dynamic marginal cost;
The route guidance device according to claim 1, further comprising: The predicted time information generation unit
One link is extracted from all links on the route candidate, and it is determined that the moving speed of the extracted link up to the time when the vehicle arrives at the target position is less than a predetermined first speed. 4. The method according to claim 1, wherein the extracted link is predicted to be a link in which traffic congestion has occurred before the time when the vehicle arrives at the destination position. 5. Route guidance device. The predicted time information generation unit
Based on the information indicating the link where the traffic jam occurs, a link where the traffic jam ends is extracted as a bottleneck link, and the traffic jam of the bottleneck link on the route candidate ends at the extracted bottleneck link. 5. The route guidance device according to claim 1, wherein the time is estimated using a moving speed of the vehicle for each link and the route candidate information. The predicted time information generation unit
When it is determined that the moving speed of the vehicle at the time of reaching the connected link is equal to or higher than a predetermined second speed in all the links connected downstream of the link where the traffic jam occurs , When the link in which the traffic jam occurs is weighted, and the value indicating the weight of the link in which the traffic jam occurs is larger than a predetermined value, the link in which the traffic jam occurs is set as a bottleneck link The route guidance device according to any one of claims 1 to 5, wherein the route guidance device is extracted. The predicted time information generation unit
Based on information indicating the link where the traffic jam occurs and the moving speed of the vehicle, if the link is lower than a predetermined speed per hour, it is determined that the link is one continuous traffic jam. The route guidance device according to any one of claims 1 to 6, wherein The Pareto improvement route guide unit is
Based on the number of terminals in the graph structure of the link subordinate to the upstream side of the bottleneck link, the impact level of the bottleneck link is calculated, and the dynamic marginal cost is calculated using the calculated impact level. The route guidance device according to any one of claims 3 to 7, wherein: The Pareto improvement route guide unit is
The total of the link cost in the current route that is the route from the current position that the vehicle is currently selected to the target position is calculated, compared with the total of the link cost of the selected route candidate, and the selected 9. The selected route candidate is determined as the route when the sum of the link costs of the route candidates is smaller than the sum of the link costs in the current route. 9. Route guidance device. A route guidance method in a route guidance device,
Means for calculating a first estimated passage time required for the vehicle to pass through a link that is a predetermined section based on the moving speed of the vehicle;
Means for calculating a second predicted passage time required for the vehicle to pass the route from the current travel position to the target position based on the calculated first predicted travel time;
Means for selecting the route based on the calculated second predicted passage time;
A route guidance method comprising: A route guidance program for causing a computer to execute the route guidance method processing in the route guidance device,
A procedure for calculating a first estimated passage time required for the vehicle to pass through a link that is a predetermined section based on the moving speed of the vehicle;
Means for calculating a second predicted passage time required for the vehicle to pass the route from the current travel position to the target position based on the calculated first predicted travel time;
Means for selecting the route based on the calculated second predicted passage time;
A route guidance program to make a computer execute.

Images