JP4736590B2 - Candidate route creation device, method, program, traffic simulation device, method and program, route search device, method, and program - Google Patents

Description

  The present invention relates to a candidate route creation device, method, program, traffic simulation device, method and program, route search device, method, and program, and in particular, a candidate route creation device capable of processing at high speed with reduced calculation load, The present invention relates to a method, a program, a traffic simulation device, a method and a program, a route search device, a method, and a program.

  The current traffic network uses a traffic flow simulator that can quantitatively evaluate the effects of introducing policies and assist in planning when roads and traffic facilities are modified or newly established for the purpose of eliminating traffic congestion and smoothing traffic. Traffic flow simulation is required. The traffic flow simulation is a simulation that reproduces the influence on traffic as a whole by reflecting traffic regulations and vehicle routes in the traffic flow.

  Conventionally, there has been disclosed a route selection method in traffic flow simulation that can reduce the route determination processing time and total calculation time in each vehicle even when the network is enlarged (see, for example, Patent Document 1). .)

  The route selection method in the traffic flow simulation described in Patent Document 1 prepares and stores in advance the required cost of the shortest route from each starting point and each branch point to the destination, and from the current location to the destination. By sequentially selecting and determining the route, the route determination processing time in each vehicle is shortened.

  In such a traffic flow simulation, either a method for sequentially searching for a route (method A) or a method for selecting an optimum route from a candidate route created in advance (method B) is used.

In Method A, an optimal route is searched for by the Dijkstra method or the like when the vehicle is generated. In addition, depending on the traffic situation, a route search may be performed sequentially during travel. In method B, a road that travels stochastically is selected from among a plurality of candidate routes using a dial distribution method or the like in consideration of the road conditions at the time of vehicle occurrence. Further, when changing the route in the middle, as in Method A, the route may be searched at a point in the middle of traveling to change the route.
JP 2002-157672 A

  Method A has a problem that the amount of calculation is very large, and when many vehicles travel on a wide-area network to be evaluated, the calculation time of route search processing for each vehicle is very long. In addition, the searched route varies depending on traffic conditions, and it is difficult to confirm whether the route of each vehicle is realistic. Furthermore, when the traffic situation changes from moment to moment, an unrealistic route may be selected, such as going around a point or making a very long trip.

  In the method B, a plurality of candidate routes are selected in order from the smallest cost, and in this case, a part of the searched route may overlap with another route. Such a route is not suitable as a candidate route because most of the links overlap with links of other routes.

  There is also a method for searching for a route that minimizes the link including travel time as the link cost. In this method, the route is mainly a narrow street that avoids congestion of the main trunk line, and on the actual road, a route that is not selected by the driver is searched. Such a route is not realistic and has a problem that it is not suitable for a candidate route of a vehicle.

  The present invention has been proposed to solve the above-described problems, and an object thereof is to provide a candidate route creation apparatus, method, and program capable of creating a suitable route at high speed. Furthermore, it aims at providing the traffic simulation apparatus, method, and program which perform a traffic simulation using the said invention, the route search apparatus, method and program which search a route using the said invention.

Candidate route generation device according to the present invention, the road network data link multiple nodes, based on the OD traffic volume data for each node, for each combination of the origin and end point specified by the OD traffic volume data, An optimum route searching means for searching for an optimum route from a starting point to an end point; a branch node extracting means for extracting a node at which the optimum route searched by the optimum route searching means intersects or a node at which the optimum route branches as a branch node; A branching node of each of the optimum paths using a branching unit that connects the branch nodes extracted by the branching node extraction unit to create a simplified network; and a branch node of the simplified network created by the creating unit. A combination of a path extracting unit that extracts a path between the plurality of paths extracted by the path extracting unit, It includes a route candidate creating means for creating a candidate route suitable route, the.

  The optimum route searched by the optimum route searching means is one for the combination of the starting point and the ending point. Therefore, the optimum route search with a large calculation load is performed only once for the combination of the start point and the end point.

  The branch node where the optimum routes searched by the optimum route search means overlap is considered to be a place where a general driver passes relatively well. Therefore, the branch node extraction means extracts a branch node from each node of each optimum route based on the degree of overlap of the searched optimum routes.

  The creation means creates a simplified network by connecting the extracted branch nodes. This simplified network represents the optimum route when connecting places where general drivers often pass.

  The path extraction means extracts a path between the branch nodes of each optimum route using each branch node of the simplified network. This path is extracted so as not to overlap with the optimal path and other paths, and not to make a large detour from the optimal path.

  The candidate route creation means creates a candidate route for each optimum route by combining the plurality of extracted paths. As a result, candidate routes having the same start point and end point but different routes are created.

  Therefore, the candidate route creation device according to the present invention searches for the optimum route from the start point to the end point, extracts a branch node from each node of each optimum route based on the degree of overlap of each optimum route, and each branch node By extracting the paths between the branch nodes of each optimum route using and creating a candidate route for each optimum route by combining the plurality of extracted paths, an appropriate candidate route can be created at high speed.

  The present invention can also be realized by a candidate route creation method and program.

Traffic simulation apparatus according to the present invention, a cost representing the said candidate route generation device, and a vehicle generating means for generating a predetermined number of vehicles per unit time, the driving load of the searched optimum route by the candidate route generation device, Generated by the vehicle generation means, the route selection means for selecting a route stochastically from the optimum route and the candidate route using the cost representing the travel burden of the candidate route created by the candidate route creation device Vehicle moving means for moving each of the vehicles that have been selected from the starting point to the ending point of the route selected by the route selecting means.

  The vehicle generating means generates a predetermined number of vehicles per unit time. The vehicle is on the traffic simulation. The route selection means stochastically selects a route from the optimum route obtained by the candidate route creation device and the candidate route.

  In order to select a route stochastically, a cost that represents the travel burden of the optimum route searched by the candidate route creation device and a cost that represents the travel burden of the candidate route created by the candidate route creation device are used. Normally, as the cost increases, the driver does not select a route corresponding to the cost, and as the cost decreases, the driver selects a route corresponding to the cost. Therefore, each of the above-described costs is used, and one route is selected stochastically from the optimum route and the candidate route.

  The vehicle moving means moves each vehicle generated by the vehicle generating means from the start point to the end point of the route selected by the route selecting means.

  Therefore, the traffic simulation device according to the present invention uses the cost that represents the travel burden of the optimal route and the cost that represents the travel burden of the candidate route created by the candidate route creation device. By selecting a route stochastically from each other and moving a predetermined number of vehicles generated per unit time from the start point to the end point of the selected route, a traffic simulation can be performed at high speed. The present invention can also be realized by a traffic simulation method and program.

Route searching apparatus according to the present invention, the current position and the starting point, the candidate path to create a candidate route of the optimum route with the destination as the end point, searching an optimum route from the current position to the destination The route with the lowest cost based on the creation device, the cost representing the travel burden of the optimum route searched by the candidate route creation device, and the cost representing the travel burden of the candidate route created by the candidate route creation device Route search means for searching for.

  The candidate route creation device searches for one optimum route from the current position to the target and creates a candidate route different from the optimum route. The route search means searches for the route with the lowest cost based on the cost representing the travel burden of the searched optimum route and the cost representing the travel burden of the created candidate route.

  Therefore, the route search device according to the present invention can search a route from the current position to the destination at high speed. Note that the present invention can also be realized by a route search method and a program.

  The candidate route creation apparatus, method, and program according to the present invention can create an optimum route and its candidate route at high speed.

  The traffic simulation apparatus, method, and program according to the present invention can perform a traffic simulation of a vehicle traveling on a plurality of routes at high speed.

  The route search apparatus, method, and program according to the present invention can search for a route at high speed.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
FIG. 1 is a block diagram showing a configuration of a traffic simulation apparatus according to the first embodiment of the present invention.

(overall structure)
The traffic simulation apparatus includes an input unit 10 to which road network data and the like are input, a route search processing unit 20 that searches for an optimum route, a candidate route creation processing unit 30 that creates a plurality of candidate routes, a candidate route list and a branch A database 40 that stores a node list, a traffic flow calculation unit 50 that calculates a traffic flow, and a display unit 60 that displays a traffic flow simulation result by the traffic flow calculation unit 50 are provided.

  Road network data and OD traffic volume data are input to the input unit 10 from the outside. The input unit 10 supplies these data to the route search processing unit 20 and the traffic flow calculation unit 50. The input unit 10 is not particularly limited as long as the input data can be supplied to the route search processing unit 20 and the traffic flow calculation unit 50. For example, the input unit 10 is any of a keyboard, an input / output port that receives data from an external network, a sensor that receives traffic information, and a reading device that reads road network data from a recording medium such as an optical disk or a magnetic disk. Also good.

(Configuration of the route search processing unit 20)
Based on the road network data supplied from the input unit 10 and the start and end points of the OD traffic volume, the route search processing unit 20 searches for an optimal route that minimizes the total value of link costs. The OD traffic volume refers to the traffic volume between the starting point and the ending point from a certain starting point (starting point) to the destination point (ending point). is there. The starting point and the ending point of the route may be representative points in an area where there are facilities, factories, residential areas, commercial areas, and the like.

  FIG. 2 is a diagram illustrating a road network represented by road network data. The road network is obtained by simplifying roads by links that connect nodes representing intersections. In the figure, the nodes which are the starting point and the ending point of the vehicle inputted by the OD traffic are represented by black circles, and the other nodes are represented by white circles. In addition, the link between the nodes is represented by a single line, but in actuality, the links are separated in the upstream direction and the downstream direction. Thick links indicate major highways, and thin links indicate narrow streets.

  The route search processing unit 20 searches for an optimal route for a vehicle traveling from the starting point N1 to the ending point N2 as follows. The route search processing unit 20 uses, for example, the Dijkstra method to obtain a route that minimizes the total link cost on the route. The link cost represents a burden when the vehicle travels on the corresponding link, and is calculated by, for example, Expression (1).

In equation (1),
C: link cost T: (by direction) link travel time L: link length Rk: road type Rw: number of lanes D: presence / absence of right / left turn F: tolls w1-w6: weighting factor.

  The weighting coefficient w is adjusted so that a route close to the route selected by the driver on the actual road is calculated. The weight coefficient w may be set for a vehicle type such as a normal vehicle or a large vehicle. The link travel time T may be obtained for each direction such as going straight, turning right, turning left.

  Since the link travel time T changes according to traffic conditions, it is a dynamic (variable) parameter for the link cost C. Parameters other than the link travel time T are constant regardless of traffic conditions, and are static (fixed) parameters.

  As the link travel time T, a value obtained by dividing the link length by the speed limit is used at the time of creating a candidate route or before starting the simulation. During the simulation, the calculated value of the link travel time T based on the simulation result of the previous time is used. The value is calculated based on a travel time during which the vehicle travels on the corresponding link within a certain interval (for example, 5 minutes). Moreover, you may use the value measured on the actual road for speed and travel time.

FIG. 3 is a diagram illustrating the optimum route R1 _N1-N2 searched by the route search processing unit 20 before the simulation is started. This route R1 _N1-N2 is calculated so as to shorten the distance mainly through main trunk lines such as national roads. That is, the route R1 _N1-N2 is a route similar to the route traveled by the driver during the quiet time, the route determined by looking at the map, and the recommended route by the navigation system.

  In FIG. 3, the route from the starting point N1 to the ending point N2 is taken as an example. However, the route is not limited to this. The traffic demand input by the OD traffic volume represents various directions such as from outside to outside, from outside to inside, from inside to inside.

FIG. 4 is a diagram showing other traffic demands in the target area of FIG. Here, there are also traffic demands such as a starting point N3 to an ending point N4 and a starting point N5 to an ending point N6. The route search processing unit 20 searches the optimum routes R1 _N3-N4 and R1 _N5-N6 for these traffic demands. Then, the route search processing unit 20 searches for the optimum route as described above for all the origin / end point pairs of the OD traffic volume, and adds the searched route to the “candidate route list” of the database 40.

(Candidate route creation processing unit 30)
The candidate route creation processing unit 30 creates a plurality of candidate routes that are different from the optimum route for each set of starting points and ending points of the OD traffic volume. In order to create a candidate route, (1) extraction of branch nodes, (2) simplification of a traffic network, (3) extraction of paths between branch nodes, and (4) combination of paths are performed.

(Extraction of branch nodes)
When the optimal routes for every pair of starting points and ending points of the OD traffic are overlapped, there are nodes where the routes intersect. For example, as shown in FIG. 4, at the node N9, the route R1 _N1-N2 and the route R1 _N3-N4 intersect. Since many cars pass through a node where many routes intersect, the actual intersection corresponding to the node becomes a main intersection with a large amount of traffic. A node where the routes intersect in this way is hereinafter referred to as a “branch node”. The overlapping state of the branch nodes is not sufficient if the paths are simply overlapped, and it is necessary to satisfy certain conditions.

  FIG. 5 is a diagram for explaining conditions under which node e is extracted as a branch node when routes overlap at node e. In the case of FIG. 5A, when the vehicle moves from node a to node e, there is only node b as an option for the next destination node (excluded because returning from node e to node a is meaningless). Therefore, the node e is merely a passing point and is not a branch node.

  In the case of FIG. 5B, when the car moves from node a to node e, there are node b and node c as options for the next destination node. Therefore, node e is extracted as a branch node. In the case of FIG. 6C, since the two paths intersect at the node e, the node e is extracted as a branch node. In this embodiment, when two or more branches exist from a node, the node is set as a branch node. However, the present invention is not limited to this. For example, when it is necessary to reduce the number of branch nodes, only nodes having three or more branches from the nodes may be used as branch nodes.

  Then, as described above, the candidate route creation processing unit 30 overlaps all routes in the target area, sequentially extracts branch nodes under the above conditions, and adds the branch nodes to the branch node list.

(Simplification of traffic network)
Next, the candidate route creation processing unit 30 records a node string in which the branch nodes are connected on the optimum route as a portion connecting the branch nodes (hereinafter referred to as “path”).

FIG. 6 is a diagram illustrating paths when nodes N7, N8, and N9 are defined as branch nodes P2, P3, and P4, respectively. The branch nodes P2 and P3 pass through the route R1 _N1-N2 , and the branch nodes P2 and P3 are connected by a path connecting the nodes N7, N8, and N9. Therefore, the candidate route creation processing unit 30 connects the branch links P2-P3, the node strings N7, N8, N9 constituting the path between the branch nodes, and the connection links (hereinafter referred to as “branch”) between the branch nodes. It is called “node connection link.”) Record L _P2-P3 as follows.

FIG. 7 is a diagram showing a correspondence relationship table between branch nodes, node strings, and branch node connection links recorded by the candidate route creation processing unit 30. For example, for the inter-branch node P3-P4, the candidate route creation processing unit 30 records the inter-branch node P3-P4, the node strings N9, N10, and the branch node connection link L _P3-P4 in the correspondence table. Then, the candidate route creation processing unit 30 creates a simplified version of the traffic network shown in FIG. 2 (hereinafter referred to as “simplified network”) based on the branch node and the branch node connection link.

  FIG. 8 is a diagram illustrating a simplified network created by the candidate route creation processing unit 30. The simplified network is composed of a plurality of branch nodes and a plurality of branch node connection links connecting the branch nodes.

  In the above description, an example in which the candidate route creation processing unit 30 automatically extracts a branch point in accordance with an actual road condition has been described. However, the present invention is not limited to this. For example, when the user determines that a node extracted as a branch node is not suitable as a branch point compared with the actual road situation, the user may delete the node from the branch node candidates. Further, the user may newly register a node that has not been extracted as a branch node in the candidate route creation processing unit 30 as a branch node.

  When the branch node is edited by the user, the candidate route creation processing unit 30 creates a simplified network again. In addition, when branch nodes are added or deleted from the list and there are no paths between the branch nodes, the route search processing unit 20 can search for an optimum route between the branch nodes using the Dijkstra method or the like and find a path. That's fine. Information such as addition, deletion, and editing of branch nodes is input from the input unit 10.

  The user may manually register all branch nodes. In this case, the route search processing unit 20 may search for an optimum route between branch nodes, between a start node and a branch node, and between a branch node and an end node, and obtain a path. Then, the candidate route creation processing unit 30 may create a simplified network using branch nodes.

(Extracting paths between branch nodes)
Next, the candidate route creation processing unit 30 creates a plurality of candidate routes for each start point and end point based on the simplified network. Here, an example in which a plurality of candidate routes of the route R1 _N1-N2 from the starting point N1 to the ending point N2 is created will be described.

FIG. 9 is a diagram showing a route R1 _N1-N2 from the starting point N1 to the ending point N2 in the simplified network. The route R1 _N1-N2 is represented by a start node, a branch node, and an end node (N1 → P1 → P2 → P3 → P4 → P5 → P6 → P7 → N2). This route R1 _N1-N2 is divided and a path between branch nodes is examined. For example, the path from P2 to P6 is composed of a branch node sequence P2->P3->P4->P5-> P6. Therefore, the candidate route creation processing unit 30 extracts and registers this branch node sequence as the first path from P2 to P6.

  FIG. 10 is a diagram illustrating a branch node sequence registered by the candidate route creation processing unit 30 and its cost. The candidate route creation processing unit 30 calculates the cost of the branch node sequence P2-> P3-> P4-> P5-> P6, registers the branch node sequence as the first node (number 1), and also registers the cost. In addition, the calculation of cost may use Formula (1), and may use another formula.

FIG. 11 is a diagram showing another route _R1′N1-N2 from the start point N1 to the end point N2. The path from P2 to P6 in _R1′N1-N2 is composed of a branch node sequence P2 → P8 → P9 → P10 → P6. This is a different path from the first path. Therefore, the candidate route creation processing unit 30 calculates the cost of the branch node sequence P2->P8->P9->P10-> P6, registers the branch node sequence as the second node (number 2), and also registers the cost. To do.

  The candidate route creation processing unit 30 sequentially registers a plurality of paths between the branch nodes of all routes in this way. However, if the path does not satisfy a certain condition, the path is not registered.

  FIG. 12 is a diagram illustrating an example in which the second and subsequent paths are not registered. For example, when searching for the second path from P1 to P3, the following conditions are added, and the search for the optimum route is performed with the start point being P1 and the end point being P3.

Condition 1: Do not pass another branch node (other than P1 and P3) on the route R1 _N1-N2 on the way.

  Condition 2: The cost of the new path is within 1.5 times the cost of the first path.

  Condition 1 means that all or part of the new path must not pass through a route (or path) on the simplified network that has already been registered. As a result, if the path P1 → P2 → P3 is excluded between P1 and P3, the path P1 → P11 → P12 → P9 → P14 → P13 → P3 becomes candidates for the path as shown in FIG. If no path candidate is found even after the optimum route search is executed, it is assumed that there is no path.

  However, if the cost of another candidate path is very large compared to the cost of the first path, the path is considered a very detour path and is inappropriate as a path. Condition 2 is used to eliminate such a path.

  In the case of FIG. 12, the cost of the paths P1 → P11 → P12 → P9 → P14 → P13 → P3 is 1.5 times or more the cost of the paths P1 → P2 → P3. For this reason, the paths P1, P11, P12, P9, P14, P13, and P3 are not appropriate paths and are excluded.

The candidate route creation processing unit 30 searches for a plurality of paths between the branch nodes on the optimum route (route R1 _{N1-N2 in} this embodiment) as described above, and extracts the following plurality of paths. .

FIG. 13 is a diagram showing a plurality of paths of the route R1 _N1-N2 . FIG. 14 is a diagram illustrating a final branch node sequence registered by the candidate route creation processing unit 30 and its cost. According to FIG. 14, there is a path that can branch at P1, P2, P3, and P4 among the branch nodes of the optimum route R1 _N1-N2 .

(Combination of paths)
The candidate route creation processing unit 30 newly creates the following four candidate routes R2, R3, R4, and R5 in addition to the optimum route R1 _N1-N2 for the start point N1 to the end point N2 by combining paths between branch nodes. Can be created.

R2 _N1-N2 : N1 → P1 → P11 → P12 → P9 → P10 → P6 → P7 → N2
R3 _N1-N2 : N1 → P1 → P2 → P8 → P9 → P10 → P6 → P7 → N2
R4 _N1-N2 : N1 → P1 → P2 → P3 → P13 → P14 → P10 → P6 → P7 → N2
R5 _N1-N2 : N1 → P1 → P2 → P3 → P4 → P15 → P16 → P17 → P7 → N2
The underlined portion indicates a branch node branched from the route R1 _N1-N2 . Then, the candidate route creation processing unit 30 adds the routes R2, R3, R4, and R5 created in this way to the candidate route list in the database 40.

  Therefore, the route search processing unit 20 searches for the optimum route based on the road network data and the OD traffic volume data, and the candidate route creation processing unit 30 creates a plurality of candidate routes from the branch nodes of the optimum route. As a result, the search for the optimum route with a large amount of calculation is required only once, so that a plurality of routes from the start point to the end point can be created at high speed.

  In addition, the candidate route creation processing unit 30 extracts branch nodes based on the overlapping state of the optimum routes having different starting points and ending points, and creates a plurality of routes by linking the branch nodes. A likely route can be created. In other words, it is possible to avoid the creation of a path that is rarely used by a general driver.

(Traffic flow calculation unit 50)
The traffic flow calculation unit 50 includes a vehicle generation processing unit 51 that generates vehicles, a route selection processing unit 52 that probabilistically selects a route for each vehicle, and a vehicle movement processing unit 53 that moves the vehicle on the selected route. And.

  The vehicle generation processing unit 51 generates a predetermined number of vehicles on the traffic simulation per unit time. The generated vehicle is arranged at the starting point of the route selected by the route selection processing unit 52.

  For each vehicle generated by the vehicle generation processing unit 51, the route selection processing unit 52 stochastically selects a route from the optimal route searched by the route search processing unit 20 and the candidate route created by the candidate route creation processing unit 30. Select.

(Factors of route selection behavior)
In order to express the route selection behavior of the vehicle in the traffic simulation, it is necessary to classify the driver's route selection factors. The following are examples of drivers' route selection behavior.

(1) Do you know the normal traffic situation or do not know it at all? The driver who commute to school has already traveled many times in the area, knowing the approximate traffic jam area and traffic situation. ing. The driver like this selects “the route of the main trunk center in the early morning, and selects another route that avoids the congestion of the main trunk in the commuting hours” according to the traffic situation. decide. On the other hand, a driver who does not know the land for the first time in the area determines a route with reference to a map and a navigation system. In this case, since the driver does not know the traffic situation and determines the route by a static factor, the driver mainly selects the main trunk line as the route.

(2) Whether or not there is a means of receiving traffic information Vehicle drivers with a navigation system with VICS (Vehicle Information and Communication System) or a means of receiving traffic information services using mobile phones are far away from the current location. I can know the situation. Therefore, when such a driver receives traffic information of a point on a route farther from the current location, the driver considers a route change so as to bypass the far traffic point even if there is no traffic around the current location. On the other hand, a driver of a vehicle having no means for receiving traffic information can recognize only the surrounding traffic situation, and considers a route change only when encountering an immediate traffic jam.

(3) Do you like to change the route or not? There are drivers who change to another route immediately when there is a traffic jam or drivers who do not change the route very much. This is considered to be due to different conditions for becoming a trigger for examining route changes.

  Route selection is performed in consideration of driver / vehicle factors related to route selection behavior as described above.

(Processing of the route selection processing unit 52)
When the vehicle generated by the vehicle generation processing unit 51 selects a route, the route selection processing unit 52 determines the probability of selecting each route based on the logit model shown in Expression (2).

here,
n: Total number of paths for this starting point and end point P _k : Probability of selecting path k C _k : Total link cost of path k θ: Sensitivity parameter representing sensitivity of path selection probability to cost.

  θ will be described later. Moreover, the link cost of each link of the route k can be obtained by Expression (1). For the link travel time T for obtaining the link cost, one of the following two methods is used for each driver factor.

Method 1-A: Link travel time T obtained by simulation calculation of the previous time (for example, the time before one sampling when simulation calculation is performed every predetermined sampling time) or the actual link travel time T
Method 1-B: Link travel time T when the link travels at the maximum speed, or link travel time T = 0 (Also possible as coefficient w1 = 0)
The link travel time T changes in a situation where traffic congestion occurs. If the driver has knowledge about the traffic situation in the driving area, the link travel time T is determined according to method 1-A, and if there is no such knowledge, the total link cost of route k is calculated. . In method 1-B, the link travel time T is ignored by setting the link travel time T to a predetermined value (the travel time when the link travels at the maximum speed, or zero). Therefore, it is better to switch to method 1-A when there is information on travel time T and to method 1-B when there is no information about travel time T.

For example, in the case of a vehicle traveling from the starting point N1 to the ending point N2, the route selection processing unit 52 performs each route R1 _N1-N2 shown in FIG. 12: N1 → P1 → P2 → P3 → P4 → P5 → P6 → P7 → N2
R2 _N1-N2 : N1 → P1 → P11 → P12 → P9 → P10 → P6 → P7 → N2
R3 _N1-N2 : N1 → P1 → P2 → P8 → P9 → P10 → P6 → P7 → N2
R4 _N1-N2 : N1 → P1 → P2 → P3 → P13 → P14 → P10 → P6 → P7 → N2
R5 _N1-N2 : N1 → P1 → P2 → P3 → P4 → P15 → P16 → P17 → P7 → N2
Then, after calculating the sum of the link costs for each, the formula (2) is calculated, and one of the paths R1 to R5 is selected stochastically.

FIG. 15 is a diagram for explaining the stochastic selection of a route. P _{1 to} P ₅ are probabilities of R1 _N1 -N2 to R5 _N1 -N2 obtained by the equation (2). The route selection processing unit 52 calculates P _{1 to} P ₅ and accumulates them as shown in FIG.

For example, if P ₁ = 0.40, P ₂ = 0.26, P ₃ = 0.17, P ₄ = 0.10 and P ₅ = 0.07, the range of accumulated probabilities P ₁ to ₅ is It becomes as follows.

Range of accumulated probability P ₁ : 0.00 or more and less than 0.40 Range of accumulated probability P ₂ : 0.40 or more and less than 0.66 Range of accumulated probability P ₃ : 0.66 or more and 0.83 The range of the accumulated probability P ₄ : 0.83 or more and less than 0.93 The range of the accumulated probability P ₅ : 0.89 or more and less than 1.00 Next, the route selection processing unit 52 selects a random number (0 to 1). .0) and the route is selected based on which range of the cumulative probabilities P _{1 to} P ₅ the generated random number corresponds to. For example, if the random number is 0.5 is selected path R2 corresponding to the cumulative probability P _2, when the random number is 0.80 path R3 that corresponds to the cumulative probability P ₃ is selected.

  The vehicle movement processing unit 53 performs processing for moving the vehicle generated by the vehicle generation processing unit 51 from the starting point of the route selected by the route selection processing unit 52 toward the end point. Before the vehicle moves on the route and reaches the next branch node, the route selection processing unit 52 determines whether or not to change the route at the branch node.

(Route change condition)
FIG. 16 is a diagram for explaining determination as to whether or not to change the route at the branch node of the route R1 _N1-N2 . Here, it is assumed that the traffic jam A occurs from the branch nodes P2 to P3.

If the vehicle generated by the vehicle generator unit 51 selects a route R _N1-N2, the path is the path R _N1-N2 are respectively branched from the branch node P1, P2, P3, P4 are present. For this reason, the vehicle can select another path (change the route) before the branch node. However, the route is not necessarily changed at a point where the route can be changed. For example, when there is no traffic jam and the vehicle can travel smoothly, the driver does not need to change the route. Also, if there is a traffic jam on the currently selected route, the driver considers whether to change the route.

  Here, even if a traffic jam is encountered, the determination of whether or not to consider a route change differs depending on the driver. Therefore, the route selection processing unit 52 can set the preference of the route selection of the driver in advance.

  FIG. 17 is a diagram illustrating the preference of the driver for route selection. According to FIG. 17, a driver with a high preference for route change changes a route when the travel time becomes twice or more that of a quiet time. A driver with a low preference for changing the route changes the route when the travel time becomes four times or more that of a quiet time. A driver with a moderate preference for changing the route changes the route when the travel time becomes three times or more that of a quiet time.

  The degree of distribution of drivers having preference for route change may be set in accordance with driver route selection statistics.

  In addition, vehicle drivers who do not have the means to receive traffic information can only recognize the surrounding traffic situation, so consider changing the route only if the link just before the traffic is congested before the branch node where the route can be changed. To do. The driver of a vehicle with a means to receive traffic information can also know far-off traffic information, so if traffic congestion occurs on a link on the route, the route can be changed anywhere before the branch node that can change the route. Can be considered.

(Route selection method when changing routes)
In route selection when a route is changed during traveling, a partial route (hereinafter referred to as “partial route”) from a point in the middle (branch node) to the end point is selected. For example, when route selection is performed before branch node P2 in FIG. 15, the following partial route exists as a route from branch node P2 to end point N2 depending on the combination of paths between the branch nodes on the route. To do.
R1 _P1-N2 : P2->P3->P4->P5->P6->P7-> N2
R2 _P1-N2 : P2 → P8 → P9 → P10 → P6 → P7 → N2
R3 _P1-N2 : P2->P3->P13->P14->P10->P6->P7-> N2
R4 _P1-N2 : P2 → P3 → P4 → P15 → P16 → P17 → P7 → N2
The probability of selecting each partial path is determined by the logit model of Equation (3).

here,
n: Number of partial paths for this branch node a P _k : Probability of selecting the path k for the branch node a C _k : Total link cost of the path k for the branch node a θ: Sensitivity parameter representing the sensitivity of the path selection probability to the cost It is.

The link cost of each link of the target partial route is obtained by Expression (1). The target partial path is selected by one of the following two methods according to driver factors.
Method 2-A: Method of obtaining the total value of the link costs of partial routes for all partial routes from the branch node to the end point of the route change, and selecting the partial route stochastically by Equation (3) B: For the partial route from the branch node to which the route is changed to the foremost branch node on the route where the path exists, the total value of the link costs of the partial route is obtained, and the partial route is stochastically expressed by Equation (3). Method of Selection When the route is changed at the branch node P2 in FIG. 16, the following partial route is selected by the method 2-A or the method 2-B.

The partial path selected by method 2-A is R1 _P1-N2 : P2 → P3 → P4 → P5 → P6 → P7 → N2
R2 _P1-N2 : P2 → P8 → P9 → P10 → P6 → P7 → N2
R3 _P1-N2 : P2->P3->P13->P14->P10->P6->P7-> N2
R4 _P1-N2 : P2 → P3 → P4 → P15 → P16 → P17 → P7 → N2
There are four.

The partial path selected by method 2-B is R1 _P2-P6 : P2 → P3 → P4 → P5 → P6
R2 _P2-P6 : P2 → P8 → P9 → P10 → P6
R3 _P2-P6 : P2 → P3 → P13 → P14 → P10 → P6
It is three.

In method 2-B, the partial route is selected for a branch node where a path from the branch node P2 exists on the route, and the branch node nearest (closer) to the branch node P2 is P6. Because there is. The reason is that the branch node P3 is a branch node having a path (P2->P8->P9->P10->P14->P13-> P3) from P2, and is the foremost branch node from P2. However, since the cost of this path is 1.5 times or more the cost of the path P2 → P3, this path is regarded as inappropriate, and the branch node P3 is excluded from the target branch nodes in the method 2-B. The The branch nodes P4 and P5 do not have a path from the branch node P2. The branch node P6 has a path (P2->P8->P9->P10-> P6) from the branch node P2. Therefore, by the method 2-B, the paths R1 to R3 _P2-P6 from the branch nodes P2 to P6 become selectable partial routes.

For example, when R3 _P2-P6 is selected as the partial path, the partial path from the branch node P2 to the end point N2 is R3 _P2-N2 : P2 → P3 → P13 → P14 → P10 → P6 → P7 → N2
It becomes.

  As described above, the route selection processing unit 52 selects a plurality of partial routes according to the method 2-A or the method 2-B before the vehicle reaches the next branch node on the route, and determines the link cost of each partial route. After calculating the sum, Equation (3) is calculated, and any one partial path is selected probabilistically.

  In addition, since the vehicle which becomes the object of method 2-A is a vehicle having a means for receiving traffic information, the driver of the vehicle knows traffic conditions (traffic jams A and B) around and far from the vehicle. Therefore, the cost of each partial route is calculated in consideration of traffic jam A and traffic jam B shown in FIG. 16, and the partial route with the lowest cost is selected.

  On the other hand, since the vehicle that is the target of method 2-B is a vehicle that does not have a means for receiving traffic information, the driver of the vehicle knows only the traffic situation (traffic jam A) around the vehicle. In this case, the driver considers a route that avoids only the traffic A. Therefore, the cost of the partial route in the range close to the vehicle position is calculated, and the partial route with the lowest cost is selected.

  FIG. 18 is a diagram illustrating a relationship between a factor of route selection behavior of a driver and a route selection method. According to the figure, in the case of a driver of a vehicle having a means for obtaining (receiving) traffic information, the route selection processing unit 52 uses the method 1-A when the vehicle is generated regardless of whether or not there is knowledge about the traffic situation in the area. When the route is changed in the middle, the route is changed by the method 2-A. The point where the route change is determined is in front of all branch nodes that can change the route.

  In the case of a driver of a vehicle having no means for obtaining traffic information and having no knowledge about the traffic situation in the area, the route selection processing unit 52 selects a route by the method 1-B when the vehicle is generated, and starts the route from the middle. When changing the route, the route is changed by the method 2-B. In the case of a driver of a vehicle who has no means for obtaining traffic information and is a driver who has knowledge about the traffic situation in the area, the route selection processing unit 52 selects a route by the method 1-A when the vehicle is generated, and starts the route from the middle. When changing the route, the route is changed by the method 2-B. The point where the route change is determined is immediately before the branch node immediately before the occurrence of the traffic jam.

  The route selection processing unit 52 may randomly determine the driver's route selection behavior factor and route selection method for each vehicle generated by the vehicle generation processing unit 51, or statistical data (per unit number). The number of vehicles having the traffic information receiving means, the route selection pattern of the driver, etc.) may be taken into consideration.

  As described above, the traffic simulation apparatus according to the first embodiment searches for an optimum route based on road network data and OD traffic volume data, and creates a plurality of candidate routes from the branch nodes of the optimum route. As a result, a plurality of routes are created by searching for an optimum route with a large computation load only once, and traffic simulation can be calculated at high speed.

  In addition, the traffic simulation device extracts branch nodes based on the overlapping state of the optimum routes having different starting points and end points, and creates a plurality of routes by linking the branch nodes. Can be used to perform traffic simulation.

  Further, the traffic simulation apparatus can simulate what route is selected at the branch node in consideration of the characteristics of the route selection by the driver and the presence / absence of the vehicle traffic information receiving means. That is, since the traffic simulation can be executed in a situation close to the actual driver's route selection behavior, the accuracy of the simulation can be improved. In addition, since the traffic simulation apparatus can change a route in the middle of traveling without performing route search processing, the computation load can be reduced and a traffic simulation can be performed at high speed. Note that the traffic simulation apparatus can change the route again after the route is changed.

FIG. 19 is a diagram illustrating that the route change is performed again after the route change. Even when a route other than the route _RN1-N2 is selected when the vehicle is generated or when the route is changed midway, the route can be changed again. Here, the route R3 _N1-N2 : N1 → P1 → P2 → P8 → P9 → P10 → P6 → P7 → N2
Explain that changing to a different route from the middle of.

According to FIG. 14 described above, between the branch nodes P10 → P7, in addition to the path passing through the branch node P6, there is also a path passing through the branch node P18. Therefore, the route selection processing unit 52
R7 _N1-N2 : N1 → P1 → P2 → P8 → P9 → P10 → P18 → P19 → P7 → N2
It is also possible to change to this route.

  However, when the route change is repeated, there is a possibility that a route that goes around more than necessary may be selected. Therefore, the route selection processing unit 52 may add a condition that “the cost of the route to be selected is within twice the cost of the route selected when the vehicle is generated”.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the site | part same as 1st Embodiment, and the detailed description is abbreviate | omitted.

  FIG. 20 is a block diagram illustrating a configuration of a route search apparatus according to the second embodiment. The route search apparatus may be a vehicle-mounted type or a computer connected to a network.

  The route search apparatus includes an input unit 10 to which road network data and the like are input, a route search processing unit 20 that searches for an optimum route, a candidate route creation processing unit 30 that creates a plurality of candidate routes, candidate route data and a branch A database 40 for storing node data, a route selection unit 55 for selecting a route, and a display unit 60 for displaying the selected route are provided.

  The input unit 10 inputs road network data and OD traffic volume data from the outside, and supplies these data to the route search processing unit 20 and the route selection unit 55. The input unit 10 is not particularly limited as long as the input data can be supplied to the route search processing unit 20 and the route selection unit 55. For example, the input unit 10 includes an input / output port for receiving data from an external network, a sensor for receiving traffic information, a reading device for reading road network data from a recording medium such as an optical disk or a magnetic disk, and current vehicle position information. Any of the receiving GPS sensors or the like may be used.

  The route search processing unit 20 searches for an optimum route from the current position to the destination using the current position as the starting point and the destination as the end point. The current position may be a position represented by information received by the GPS sensor, or may be a position set by a driver or an operator. The candidate route creation processing unit 30 creates a candidate route for the optimum route searched by the route search processing unit 20.

  The route selection unit 55 calculates the total value of the link costs of the optimum route searched by the route search processing unit 20 and the total value of the link costs of the candidate routes created by the candidate route creation processing unit 30. The link cost is obtained by equation (1). Then, the route selection unit 55 selects a route having the largest total link cost and supplies the selection result to the display unit 60. Thereby, the route from the current position to the destination is displayed on the display unit 60.

  As described above, the route search device according to the second embodiment searches for the optimum route based on the road network data and the OD traffic data, and creates a plurality of candidate routes from the branch nodes of the optimum route. As a result, the search for the optimum route with a large amount of calculation is required only once, so that a candidate route from the current position to the destination can be created at high speed. Therefore, the route search from the current position to the destination can be performed at high speed.

  In addition, the route search device extracts branch nodes based on the overlapping state of optimum routes having different starting points and end points, and creates a plurality of routes by linking the branch nodes. It is possible to search for a route with a small travel burden from the routes.

  Note that the present invention is not limited to the above-described embodiment, and it is needless to say that the present invention can also be applied to a design modified within the scope of the claims.

  In the embodiment described above, “Condition 2: The cost of the new path is within 1.5 times the cost of the first path.”, “The cost of the route to be selected is the cost of the route selected when the vehicle is generated However, the numerical values of 1.5 times and 2 times are not particularly limited. The route search processing unit 20, the candidate route creation processing unit 30, and the traffic flow calculation unit 50 (or the route selection unit 55) may be configured by different computers or the same computer.

It is a block diagram which shows the structure of the traffic simulation apparatus which concerns on the 1st Embodiment of this invention. It is a figure which shows the road network represented by road network data. It is a figure which shows the optimal path | route R1 _N1-N2 searched by the path | route search process part before the simulation start. It is a figure which shows the other traffic demand in the object area shown in FIG. It is a figure explaining the conditions from which the node e is extracted as a branch node when the path | route has overlapped in the node e. It is a figure explaining a path when nodes N7, N8, and N9 are defined as branch nodes P2, P3, and P4, respectively. It is a figure which shows the correspondence table of branch nodes, a node row | line | column, and a branch node connection link recorded by the candidate route creation process part. It is a figure which shows the simple network produced by the candidate route creation process part. It is a figure which shows path | route R1 _N1-N2 from the starting point N1 to the end point N2 in a simple network. It is a figure which shows the branch node row | line | column registered by the candidate path | route creation process part, and its cost. It is a figure which shows another path | route _R1'N1-N2 from the starting point N1 to the end point N2. It is a figure which shows the example from which the path | pass after the 2nd is not registered. It is a figure which shows the some path _| route of path | route R1 _N1-N2 . It is a figure which shows the final branch node row | line | column registered by the candidate path | route creation process part, and its cost. It is a figure explaining selecting a route stochastically. It is a figure explaining the judgment whether a path _| route is changed in the branch node of path _| route R1 _N1-N2 . It is a figure which shows the preference of a route selection of a driver. It is a figure which shows the relationship between the factor of a route selection action of a driver, and a route selection system. It is a figure explaining performing a route change again after a route change. It is a block diagram which shows the structure of the route search apparatus which concerns on 2nd Embodiment.

Explanation of symbols

10 Input unit 20 Route search processing unit 30 Candidate route creation processing unit 40 Database 50 Traffic flow calculation unit 51 Vehicle generation processing unit 52 Route selection processing unit 53 Vehicle movement processing unit 55 Route selection unit 60 Display unit

Claims (19)

Based on the road network data in which a plurality of nodes are linked and the OD traffic volume data of each node, the optimum route from the start point to the end point is searched for each combination of the start point and the end point specified by the OD traffic volume data. An optimum route search means;
A branch node extracting means for extracting, as a branch node , a node at which the optimum route searched by the optimum route searching means intersects or a node at which the optimum route branches;
Creating means for creating a simplified network by connecting the branch nodes extracted by the branch node extracting means;
Using each branch node of the simplified network created by the creation means, path extraction means for extracting a path between the branch nodes of each optimum route;
Candidate route creation means for creating a candidate route for each optimum route by combining a plurality of paths extracted by the path extraction means;
A candidate route creation device comprising: The path extracting means extracts a path so as not to overlap an already extracted path in the simplified network .
The candidate route creation apparatus according to claim 1 . 3. The candidate route creation device according to claim 1, wherein the path extraction unit compares a cost representing a travel burden of each path between branch nodes with a threshold and extracts only a path having a cost lower than the threshold. . The optimum route searching means, by using the representative point of the road network data for at least one of the start point and the end point, any one of claims 1 to 3 for searching for the optimum route The candidate route creation device described in 1. A candidate route creation method executed in a candidate route creation device comprising an optimum route search means, a branch node extraction means, a creation means, a path extraction means, and a candidate route creation means,
The optimum route searching means is configured to start and end points for each combination of start and end points specified by the OD traffic volume data based on road network data in which a plurality of nodes are linked and OD traffic volume data of each node. Search for the best route to
The branch node extraction means, the node node or the optimal path optimal route searched by said optimum route searching means intersect branches, and extracted as a branch node,
The creation means creates a simplified network by connecting the branch nodes extracted by the branch node extraction means ,
The path extraction means uses each branch node of the simplified network created by the creation means to extract a path between the branch nodes of each optimum route,
The candidate route creation method, wherein the candidate route creation means creates a candidate route for each optimum route by combining a plurality of paths extracted by the path extraction means . The computer,
Based on the road network data in which a plurality of nodes are linked and the OD traffic volume data of each node, the optimum route from the start point to the end point is searched for each combination of the start point and the end point specified by the OD traffic volume data. Optimal route search means ,
A branch node extracting means for extracting, as a branch node , a node at which the optimum route searched by the optimum route searching means intersects or a node at which the optimum route branches ;
Creating means for creating a simplified network by connecting the branch nodes extracted by the branch node extracting means ;
Using each branch node of the simplified network created by the creating means , combining a plurality of paths extracted by the path extracting means for extracting paths between the branch nodes of the optimum route, and the path extracting means , Candidate route creation means for creating candidate routes for each optimum route
Candidate route creation program to function as . The candidate route creation device according to any one of claims 1 to 4 ,
Vehicle generating means for generating a predetermined number of vehicles per unit time;
Among the optimum route and the candidate route, the cost representing the travel burden of the optimum route searched by the optimum route search means and the cost representing the travel burden of the candidate route created by the candidate route creation means are used. Route selection means for selecting a route stochastically from,
Vehicle moving means for moving each vehicle generated by the vehicle generating means from the starting point to the ending point of the route selected by the route selecting means;
Traffic simulation device with The route selection means calculates the cost of each route using estimated or actually measured travel time between routes.
The traffic simulation device according to claim 7 . The route selecting means sets a travel time between routes to a predetermined value, and calculates each cost of the optimum route and the candidate route.
The traffic simulation device according to claim 7 . The route selection means stochastically selects a route from the partial routes using the cost of all the partial routes from the branch node to the end point until the vehicle reaches the branch node.
The traffic simulation device according to any one of claims 7 to 9 . The route selection means uses the cost of the partial route from the branch node to the foremost branch node on the optimum route until the vehicle reaches the branch node, and randomly selects a route from the partial routes. select
The traffic simulation device according to any one of claims 7 to 9 . The route selection means is a first mode for calculating the cost of each route using the estimated or actually measured travel time between routes according to the presence or absence of the travel time of the vehicle. Set to a value and switch to one of the second modes for calculating the costs of the optimum route and the candidate route
The traffic simulation device according to claim 7 . It said path selection means, depending on whether the driver of the traveling experience, probabilistic from all partial paths each partial path by using the cost of from the branch node until vehicles reaches the branch node to the end point A third mode in which a route is selected, and the cost of the partial route from the branch node to the foremost branch node on the optimum route until the vehicle reaches the branch node is stochastically selected from the partial routes Switch to any one of the fourth modes of route selection
The traffic simulation device according to claim 7 or 12 . The route selection means is probabilistic from among the partial routes using the cost of all the partial routes from the branch node to the end point until the vehicle reaches the branch node, depending on the presence or absence of the traffic information receiving means. A third mode for selecting a route, using the cost of the partial route from the branch node to the foremost branch node on the optimum route until the vehicle reaches the branch node Switch to one of the fourth modes to select
The traffic simulation device according to claim 7 or 12 . A traffic simulation method executed in a traffic simulation apparatus comprising the candidate route creation device according to any one of claims 1 to 4, a vehicle generation unit, a route selection unit, and a vehicle movement unit. There,
The candidate route creation device executes the candidate route creation method according to claim 5 ,
The vehicle generating means generates a predetermined number of vehicles per unit time;
The route selection means uses the cost that represents the travel burden of the optimum route searched by the optimum route search means and the cost that represents the travel burden of the candidate route created by the candidate route creation means. And a route from the candidate route stochastically,
The traffic simulation method , wherein the vehicle moving means moves each vehicle generated by the vehicle generating means from the starting point to the ending point of the route selected by the route selecting means . The computer, there by traffic simulation program for causing to function as each means, a vehicle generator, routing means, and a vehicle moving means constituting a candidate route generation device as claimed in any one of claims 4 And
The vehicle generating means generates a predetermined number of vehicles per unit time,
The route selection means uses the cost representing the travel burden of the optimum route searched by the optimum route search means and the cost representing the travel burden of the candidate route created by the candidate route creation means , And a route from the candidate route stochastically,
The said vehicle moving means is a traffic simulation program which moves each vehicle generated by the said vehicle generation means from the starting point of the path | route selected by the said route selection means to an end point. The current position and the starting point as the end point destination, any one of claims 1 to 4 to create a candidate route of the optimum path with searching the optimum route to the destination from the current position A candidate route creation device described in
A route that searches for the route with the lowest cost based on the cost that represents the travel burden of the optimum route searched by the optimum route search means and the cost that represents the travel burden of the candidate route created by the candidate route creation means Search means;
A route search device comprising: The current position and the starting point as the end point destination, any one of claims 1 to 4 to create a candidate route of the optimum path with searching the optimum route to the destination from the current position A route search method executed in a route search device comprising the candidate route creation device described in claim 1 and route search means,
The candidate route creation device executes the candidate route creation method according to claim 5 ,
The route search means is based on the cost representing the travel burden of the optimum route searched by the optimum route search means and the cost representing the travel burden of the candidate route created by the candidate route creation means . A route search method that searches for a small route. The computer, the current position and the starting point, either the destination as the end point, the claims 1 to create a candidate route of the optimum path thereby searches for an optimum route from the current position to the destination of claim 4 Each means constituting the candidate route creation device according to claim 1, and a route search program for functioning as route search means,
The route search means is based on the cost representing the travel burden of the optimum route searched by the optimum route search means and the cost representing the travel burden of the candidate route created by the candidate route creation means . A route search program that searches small routes.

Images