JP5348181B2 - Road estimation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a road estimation device capable of suitably selecting a core point to be a processed in a divided area of map data from distributed core point information. <P>SOLUTION: A core point (CP) just before crossing a border is acquired on the inside of a boundary of a target parcel (S500) and a distance from the acquired CP up to the boundary of the target parcel is measured (S510). Whether or not the distance from the boundary up to the CP just before crossing the border is equal to or more than a predetermined distance is determined (S520). When it is determining that the distance is less than the predetermined distance (S520: NO), a CP up to the CP just before crossing the border on the inside of the target parcel is set as a CP to be processed (S530). On the other hand, when it is determining that the distance is equal to or more than the predetermined distance (S520: YES), a CP up to a CP just after crossing the border in addition to the CP on the inside of the target parcel is set as a CP to be processed (S540). <P>COPYRIGHT: (C)2013,JPO&amp;INPIT

Description

  The present invention relates to a technique for extracting a link on a map corresponding to a road indicated by the core point based on a core point given along the road and having an attribute for specifying the road.

  Conventionally, VICS (Vehicle Information and Communication System) is known as a service for transferring traffic information by broadcasting. This service sends traffic information such as traffic congestion information of the VICS Center to the vehicle, searches the map data on the vehicle side to identify the corresponding road, and changes the display mode of the road to the received traffic information. It is changed accordingly and provided to the user. Thereby, the user can grasp traffic information such as traffic jam information in real time.

  By the way, in the map data format on the vehicle-mounted device side, roads are represented by links and nodes. A link indicates a road whose end point is a node. Here, the above-mentioned VICS center transmits information called a VICS link that identifies a road. Various traffic information and display mode change instruction information are given to the VICS link. Therefore, the in-vehicle device side has a position reference table for collating the VICS link and the link in the map data, and searches for a link corresponding to the VICS link based on this table. That is, in VICS, a position reference table is essential (for example, refer to Patent Documents 1 and 2).

JP 2006-275777 A (FIG. 3) JP 2009-270953 A (FIGS. 2 and 31)

  By the way, as a service other than VICS, use of traffic information data transmitted by TPEG (Transport Protocol Expert Group) on the terminal side such as a vehicle has been studied. However, in the data format or the like transmitted by TPEG, the position information is shown in the form of DLR data, for example. This position information is composed of core points having position coordinates and attributes for specifying a road. The core points are usually distributed as a plurality of arrays arranged along the road. If position information is indicated by such a core point, a position reference table that depends on the manufacturer of map data, the format of map data, and the version becomes unnecessary. That is, a road (link) can be specified on the map data without depending on the map data on the in-vehicle device side.

  On the other hand, various processes for specifying the road based on the core points are required. For example, since there are various types of map data as described above, the core point does not necessarily exist on the road of the map data. Therefore, the process (specific process) which specifies on the map the link applicable to the road which a core point shows becomes necessary.

  Incidentally, map data is managed in units of divided areas obtained by dividing map data into a plurality of areas. This divided area is called a “parcel”. For example, when displaying a map or the like, processing is performed in units of the divided areas.

  Here, the core points are distributed as a plurality of arrays arranged along the road, but generally exist across the boundaries of the divided areas of the map data as described above. Therefore, especially in the case of specific processing near the boundary, it is important to determine up to which core point the processing target is to be divided into the display target for the purpose of facilitating processing and ensuring specific accuracy. Come.

  The present invention has been made in order to solve the above-described problems, and its purpose is to appropriately select a core point to be processed for a divided area of map data from information on a distributed core point. Is to provide a simple road estimation device.

The road estimation device according to claim 1, which has been made to achieve the above-described object, includes a plurality of cores having attributes for identifying the road that are given along the road as position information transmitted together with the traffic information. Point data is received from the outside, and a road is estimated by extracting a link on a map corresponding to the road indicated by the core point.

  In this apparatus, the map data input means inputs map data. Map data is input in units of divided areas divided into a plurality of areas. Moreover, map data is comprised by the link which has an attribute according to the attribute which a core point has.

  Here, in particular, in the present apparatus, when the core point straddles the boundary of the display divided region that is the divided region to be displayed by the core point selecting unit, the display is performed on the map data input by the map data input unit. A core point existing inside the divided area is selected as a processing target core point.

At this time, the link extraction means extracts a link as a road candidate indicated by the core point from the map data based on the attribute of the link and the attribute of the processing target core point.
That is, in the present invention, considering that map data is displayed in units of divided areas, only the core points existing inside the display divided area that is the divided area to be displayed are processed. In this way, the time required for the processing of the core points is shortened compared to the case where the core points existing outside the display divided region are also processed. As a result, it contributes to an appropriate selection of core points to be processed.

  If the inner core point closest to the boundary of the display divided region is too far from the boundary, the accuracy of road estimation near the boundary may be reduced. Therefore, as shown in claim 2, the core point selection means, when the distance from the boundary of the display divided area to the closest inner core point is a predetermined distance or more, the core point selection means In addition, the core points immediately after crossing the border may be selected as processing target core points. The core point immediately after crossing the border is a core point immediately after crossing the boundary of the display divided region from the inside to the outside. In other words, the core point immediately after crossing the border is the display division area outside the target display division area when the road indicated by the core point jumps from the target display division area to the adjacent display division area outside. It is the first core point (first core point in the array of core points in the display divided area adjacent to the target display divided area) that is distributed to show the road. The same applies to the following. In this way, it is possible to improve the accuracy of road estimation in the vicinity of the boundary of the display divided area. As a result, it contributes to an appropriate selection of core points to be processed.

  By the way, when the distance from the boundary of the display divided region to the nearest inner core point is a predetermined distance or more, the distance from the boundary to the core point immediately before the crossing is also a predetermined distance or more. The core point immediately before crossing the border is a core point immediately before crossing the boundary of the display divided area from the inside to the outside. In other words, the core point immediately before crossing the border is distributed to indicate the road in the target display divided region when the road indicated by the core point jumps from the target display divided region to the adjacent display divided region outside. The last core point (the last core point in the array of core points in the target display divided region). The same applies to the following. Therefore, as shown in claim 3, the core point selection means may select the core point immediately after the crossing as the processing target core point when the distance to the core point immediately before the crossing is not less than a predetermined distance. In this way, it is possible to improve the accuracy of road estimation in the vicinity of the boundary of the display divided area. As a result, it contributes to an appropriate selection of core points to be processed.

  According to a fourth aspect of the present invention, when the distance from the boundary of the display divided region to the core point immediately before the crossing is less than a predetermined distance, the core point selection unit selects the core point immediately before the crossing as the processing target core point. It is possible to do. In this way, it is possible to prevent the time required for processing the core point from becoming extremely long, and as a result, it contributes to appropriate selection of the core point to be processed.

  By the way, from the viewpoint of improving the accuracy of the road estimation in the vicinity of the boundary of the display divided region, as shown in claim 7, the boundary of the display divided region which is the divided region to be displayed is determined by the core point selection means. In the case where there are core points straddling, in the map data input by the map data input means, in addition to the core points existing inside the display divided area, the core points immediately after the crossing may be selected as the processing target core points. Good. In this way, the time required for processing is longer than when only the inner core points are processed, but the accuracy of road estimation near the boundaries of the display divided regions can be improved. As a result, it contributes to an appropriate selection of core points to be processed.

  If the outer core point closest to the boundary of the display divided region is too far from the boundary, the time required for processing the core point may be extremely long. Therefore, as shown in claim 8, the core point selection means, when the distance from the boundary of the display divided region to the nearest outer core point is a predetermined distance or more, the core point existing inside the display divided region is It is good also as selecting as a processing target core point. In this way, it is possible to prevent the time required for processing the core point from becoming extremely long, and as a result, it contributes to appropriate selection of the core point to be processed.

  By the way, when the distance from the boundary of the display divided region to the nearest outer core point is equal to or greater than a predetermined distance, the distance from the boundary to the core point immediately after crossing the boundary is also equal to or greater than the predetermined distance. Therefore, as shown in claim 9, the core point selection means selects the core point existing inside the display divided region as the processing target core point when the distance to the core point immediately after the border crossing is a predetermined distance or more. It is good. In this way, it is possible to prevent the time required for processing the core point from becoming extremely long, and as a result, it contributes to appropriate selection of the core point to be processed.

  According to another aspect of the present invention, when the distance from the boundary of the display divided region to the core point immediately after the crossing is less than a predetermined distance, the core point selection unit selects the core point immediately after the crossing as the processing target core point. It is possible to do. In this way, it is possible to improve the accuracy of road estimation in the vicinity of the boundary of the display divided area. As a result, it contributes to an appropriate selection of core points to be processed.

  Further, as shown in claim 13, when the core point selection means exists across the boundary of the display divided area which is the divided area to be displayed across the core point, in the map data input by the map data input means, You may employ | adopt the structure which compares the distance from a boundary to the core point just before a crossing, and the distance from a boundary to a core point just after a crossing.

  At this time, if the distance to the core point immediately after the crossing is large, the core point immediately before the crossing is selected as the processing target core point. Therefore, it is possible to prevent the time required for processing the core point from becoming extremely long, and as a result, it contributes to appropriate selection of the core point to be processed.

  On the other hand, when the distance to the core point immediately before the crossing is large, the core point immediately after the crossing is selected as the processing target core point in addition to the core point existing inside the display divided region. Therefore, it is possible to improve the accuracy of road estimation in the vicinity of the boundary of the display divided region, and as a result, it contributes to an appropriate selection of a core point to be processed.

  By the way, as shown in claims 5, 11, and 14, the distance from the boundary of the display divided region to the core point is exemplified as a length of a perpendicular line extending from the core point to the boundary. In this way, the distance from the boundary of the display divided area to the core point can be measured relatively easily.

  However, in order to make the distance closer to the driving distance according to the road shape, as shown in claims 6, 12 and 15, the intersection of the straight line connecting the core points across the boundary and the boundary is obtained, and from the core point to the intersection It is exemplified that the length is measured as the length. In this way, there is a high possibility that a distance closer to the driving distance in accordance with the road shape is required.

It is a block diagram which shows the structure of a navigation apparatus. It is a functional block diagram which shows operation | movement of a control circuit. It is a flowchart which shows a matching process. It is explanatory drawing which shows the attribute of CP after conversion. It is explanatory drawing which shows the value and content of the attribute FC. It is explanatory drawing which shows the value and content of the attribute FW. It is explanatory drawing which shows the value and content of attribute IT. It is a flowchart which shows the conversion process in a matching process. It is explanatory drawing which shows that CP exists over a target parcel. It is explanatory drawing which shows the process in provision of virtual CP and a parcel boundary. It is explanatory drawing which shows the calculation rule of attribute PCI. It is a flowchart which shows the candidate link search process in a matching process. It is explanatory drawing which shows the search range with respect to object CP, and parcel acquisition. It is a flowchart which shows the link row selection process in a candidate link search process. It is explanatory drawing which shows extraction of the link row | line with respect to a search range. It is a flowchart which shows the candidate selection process in a matching process. It is a flowchart which shows the coincidence determination of the non-shape system attribute in candidate selection processing. It is a flowchart which shows the coincidence determination of the shape type attribute in a candidate selection process. It is explanatory drawing which shows the coincidence determination based on the attribute BR. It is a flowchart which shows the coincidence determination in attribute PCI in the coincidence determination of a shape type attribute. It is explanatory drawing which shows the coincidence determination based on the attribute CA and the attribute DCA. It is a flowchart which shows the road matching process in a matching process. It is explanatory drawing which shows the omission method of a road search. It is explanatory drawing which shows the omission method of a road search. It is explanatory drawing which shows the cancellation method of a road search. It is a flowchart which shows the road unit candidate selection in a road matching process. It is explanatory drawing which shows the road selection based on the attribute CA and the attribute DCA. It is explanatory drawing which shows the road selection based on the attribute BR and the attribute DMB. It is explanatory drawing which shows calculation of the road length based on attribute PD. It is explanatory drawing which shows the road selection based on attribute PD. It is explanatory drawing which shows the road selection based on attribute PDM. It is explanatory drawing which shows the determination method of a road. It is a flowchart which shows the conversion process of another embodiment. It is a flowchart which shows the conversion process of another embodiment. It is a flowchart which shows the conversion process of another embodiment.

Embodiments of the present invention will be described below with reference to the drawings.
<1. Configuration of navigation device>
First, the configuration of the navigation device will be described with reference to FIG. The navigation apparatus 10 shown in FIG. 1 functions as a “road estimation apparatus”. Specifically, the navigation apparatus 10 receives TPEG (Transport Protocol Expert Group) data, matches roads from position information in the data, and transmits the data together with position information. Displayed traffic information corresponding to the road.

  A navigation device 10 shown in FIG. 1 includes a reception device 11, a position detection device 12, a map data input device 13, an operation device 14, an audio output device 15, a display device 16, and a control circuit 17.

  The receiving device 11 is for receiving TPEG data from the center 20. In the navigation device 10, TPEG data is acquired via the receiving device 11 by selecting a channel in the control circuit 17.

  The position detection device 12 is for detecting the current position of a vehicle on which the navigation device 10 is mounted, and includes, for example, a known gyroscope, a distance sensor, a GPS receiver, and the like.

  The map data input device 13 has a recording medium (specifically, a hard disk or a DVD) in which map data is stored, and is configured so that the map data stored in this recording medium can be input to the control circuit 17. Yes. The map data input device 13 can be configured to include a DVD drive in addition to a hard disk device that stores and holds map data. If the map data input device 13 is configured in this way, the navigation device 10 can be configured such that additional data of map data sold as an option through DVD media can be installed in the hard disk device. Note that map data is managed in units called parcels and cached in units of parcels.

  The operation device 14 is used for inputting an instruction from the user to the control circuit 17, and includes a touch panel disposed on the display device 16, an operation switch group provided on the main body surface of the navigation device 10, and a remote controller. Composed. Through this operation device 14, the user can perform all operations of the navigation apparatus 10 such as map scale change, scrolling, and destination setting.

  The audio output device 15 is composed of a speaker or the like, and receives a signal from the control circuit 17 and outputs a guidance voice to the user. In addition, the display device 16 is capable of full color display, and traffic information based on TPEG data acquired via the receiving device 11 is input to the display device 16 from the map data input device 13. It is displayed superimposed on the map image based on the map data.

  The control circuit 17 has the same configuration as a known microcomputer, and includes a CPU 17a, a ROM 17b, a RAM 17c, an I / O, a bus line for connecting these configurations, and the like. The CPU 17a performs various processes according to the program stored in the ROM 17b. For example, TPEG data received via the receiving device 11 includes location information as DLR (Dynamic Location Referencing) data. Therefore, the control circuit 17 estimates a corresponding road on the map data side based on the position information.

<2. Function of control circuit>
Next, the function of the control circuit 17 relating to the processing of TPEG data will be described with reference to FIG. FIG. 2 is an explanatory diagram schematically showing the function of the control circuit 17.

The function of the control circuit 17 is divided into a channel selection block 171, an application block 172, a DLR library block 173, and a drawing block 174.
The channel selection block 171 is a block that receives TPEG data via the receiving apparatus 11 described above. That is, the channel selection block 171 has the above-described channel selection function. The received TPEG data is passed from the tuning block 171 to the application block 172.

  The application block 172 is realized as a function of the application program. The application program is stored in the ROM 17b and is executed by the CPU 17a.

  The application block 172 manages the TPEG data delivered from the channel selection block 171 and updates the TPEG data when new TPEG data is received. Further, the application block 172 has information for specifying a display target parcel (hereinafter referred to as “target parcel”) displayed on the screen. The display screen may be configured with one target parcel depending on the scale, or may be configured with a plurality of (for example, nine) target parcels. The application block 172 passes the target parcel information and the TPEG data position information to the DLR library block 173 as “target data”.

  The DLR library block 173 is realized as a function of the DLR library program. Like the application program, the DLR library program is stored in the ROM 17b and executed by the CPU 17a.

  The DLR library block 173 executes matching processing described later. The matching process estimates a corresponding road (link) in the map data input from the map data input device 13 based on the position information of the TPEG data. Prior to such matching processing, the DLR library 173 reads map data from the map data input device 13 into the RAM 17c in response to a request from the application block 172. The result of the matching process is output to the application block 172 as a matching result by the DLR library 173.

  The application block 172 manages the matching result. Then, the drawing block 174 updates the map based on the matching result. Thereby, as described above, the traffic information based on the TPEG data is displayed superimposed on the map image based on the map data input from the map data input device 13.

<3. Matching process>
The present embodiment is characterized in matching processing by the DLR library block 173. Therefore, the matching process will be described next. FIG. 3 is a flowchart showing the matching process.

  In the first S100, a conversion process is performed. This conversion process converts position information (binary data) as DLR of TPEG data into intermediate data. Although it has already been described that discrete points are mainly used for the position information of TPEG data, these discrete points are called core points (hereinafter referred to as “CP”). The CP has various attributes. Therefore, here, attributes necessary for road estimation are extracted and intermediate data is created.

<3.1 CP attributes after conversion>
The attributes of the CP after conversion into intermediate data are shown in FIG. Usually, a plurality of CPs are transmitted as position information of TPEG data. Therefore, CPs are managed in an array. Of course, it is also possible to estimate the position from only one CP, and this does not exclude the case where only one CP is transmitted.

  As shown in FIG. 4, the CP has a latitude, longitude, an IP flag, and a virtual CP flag as basic attributes. The latitude is the latitude coordinate of the CP, and the longitude is the longitude coordinate of the CP. The IP flag is a flag indicating whether the CP is an intersection. “1” is set when the CP indicates an intersection, and “0” is set otherwise. Furthermore, the virtual CP flag is a flag indicating whether or not the CP is a virtual CP. If the CP is a virtual CP, “1” is set, otherwise “0” is set. The virtual CP will be described later.

  Further, the CP has various attributes regarding the road connecting the CPs. These attributes include a shape-type attribute related to the road shape and a non-shape-type attribute not related to the road shape. First, non-shape attributes will be described.

<3.1.1 Non-shape attribute>
The attribute FC indicates a road segment. As shown in an example in FIG. 5, the road class is indicated by 10 levels from “0” to “9”. “0” is “Main road”, “1” is “First class road”, and “2” is “Second class road”. Similarly, “9” is ranked up to “Ninth class road”. For example, “Main road” is a road connected to the country and the capital, and “first class road” is ranked as a national road connecting major cities.

  The attribute FW indicates a physical road type. As shown in an example in FIG. 6, the physical road types are divided into 13 types “0” to “12”. Referring to FIG. 6, “0” for “unclear”, “1” for “highway”, “2” for “multiple roads (other than highways)”, “single road” "Is" 3 "," Rotary "is" 4 "," Traffic Square "is" 5 "," Enclosed Traffic Area "is" 6 ", and" Bypass "is" 7 " The “branch road” is “8”, the “parking lot entrance or exit” is “9”, the “service area entrance or exit” is “10”, and the “pedestrian zone” is “11”. "And" passage "is" 12 ".

  In the attribute RD, if a road number (for example, a national road number such as national road “1”) exists, the road number becomes an RD value. When no road number exists, the road name becomes an RD value. The road name is a maximum of five characters of the official name (see FIG. 4).

  The attribute IT indicates the classification of the intersection (see FIG. 4). As shown in an example in FIG. 7, the intersection classification is divided into seven types “0” to “6”. Referring to FIG. 7, “0” for “undefined”, “1” for “highway or restricted road interchange”, “2” for “rotary”, “1,2” “3”, “simple intersection” is “4”, “traffic square” is “5”, “bivalent intersection (intersection where road number and road name change) ) "Is" 6 ".

The attribute RDI indicates the name of the intersection (see FIG. 4).
The attribute DD indicates the permitted legal driving direction (see FIG. 4). For example, “undefined” is “0”, “forward” is “1”, and “reverse” is “2” and “both” is “3”. That is. The attribute AFR is a flag indicating whether or not the attribute DD can be referred to. If the attribute DD has a value of “0” to “3”, the attribute AFR is set to “1”, and if the attribute DD has no value, “0” is set as a reference impossible. .

<3.1.2 Shape system attributes>
Next, the attributes of the shape system will be described.
The attribute BR indicates a positional angle to the next CP (see FIG. 4). As described above, usually, a plurality of CPs are managed as an array. The BR value is a clockwise angle with reference to true north.

The attribute DMB indicates a linear distance to the next CP (see FIG. 4). As with the attribute BR, since the next CP usually exists, the straight line distance to the next CP is the DMB value.
The attribute CA indicates an angle with respect to the side road. A side road is a minor road to which no road number is added. When there is a side road, an angle in which clockwise is “positive” and counterclockwise is “negative” based on the angle of the attribute BR is set as the CA value. The attribute DCA indicates a connection distance to the side road.

  That is, when a vector from the CP is considered, the attribute CA indicates the direction of the vector, and the attribute DCA indicates the magnitude of the vector. Therefore, the position coordinates are determined by the attribute CA and the attribute DCA. This position coordinate indicates a point where a side road exists.

  The attribute PCI indicates which roadway is selected when there are a plurality of roadways in the same direction (see FIG. 4). This attribute PCI includes the number of roadways and a sequence number indicating the number of the target road among the roadways.

The attribute PDM indicates a separation distance of the target road from a straight line connecting to the next CP (see FIG. 4). The maximum distance to the target road is set as the separation distance.
The attribute PD indicates the driving distance along the road to the CP including the attribute PD next (see FIG. 4).

  The attribute of the CP after the conversion to the intermediate data has been described above, but the virtual CP is assigned and the attribute value of the CP is recalculated along with the conversion to the CP. The specific procedure is as shown in FIG.

<4. Details of the matching process (first half)>
In S110 in FIG. 8, a virtual CP is assigned to the boundary of the target parcel. FIG. 9 is an explanatory diagram showing the road shape represented by CP with a two-dot chain line for convenience. Of course, the ultimate goal is to estimate such a road shape and perform matching with the road indicated by the link string of the map data. Therefore, the road shape indicated by a two-dot chain line in FIG. 9 is for convenience and does not indicate an actual road shape. What is important here is that, as shown in FIG. 9, the arrangement of CPs may straddle not only the target parcel but also the surrounding parcels. In such a case, a virtual CP is assigned to the boundary of the target parcel.

  Specifically, as shown in FIG. 10A, when two CP (a) and CP (b) straddle the boundary of the target parcel, the apex of the road shape given by the value of the attribute PDM A line segment connecting V and CP (a) and CP (b) is calculated, and a virtual CP is assigned to the intersection of one of the line segments and the boundary of the target parcel.

In the following, symbols a, b, c,... Are used to distinguish a plurality of CPs, which are denoted as CP (a), CP (b), CP (c).
<4.1 Recalculation of attributes>
In the next S120, the attributes of the CP are recalculated. By assigning the virtual CP, the CP on the end side (CP (b) in FIG. 10A) is removed from the processing target CP, as will be described later. Therefore, the attributes of the CP on the start end side (CP (a) in FIG. 10A) and the virtual CP are recalculated. The attributes to be recalculated are attribute BR, attribute DMB, attribute PCI, attribute CA, attribute DCA, attribute PDM, and attribute PD.

  The attribute BR and the attribute DMB are an angle and a straight line distance to the next CP (see FIG. 4). Accordingly, the angle and straight line distance to the next CP are calculated and set including the newly assigned virtual CP.

  For the attribute CA and the attribute DCA, recalculation is not performed at the CP on the starting end side (CP (a) in FIG. 10A). An invalid value is set for the virtual CP. The attribute CA and the attribute DCA are related to the side road, and the side road information is not valid for the virtual CP.

  The attribute PCI relates to the number of parallel roadways (see FIG. 4). Therefore, recalculation is not performed on the CP at the start end (CP (a) in FIG. 10A). In the virtual CP, values according to the PCI set on the start side and the end side (CP (a) and CP (b) in FIG. 10A) are set.

  A specific calculation rule of the attribute PCI for the virtual CP is shown in FIG. When both the starting end CP and the terminating end CP hold the attribute PCI, and the attribute PCI in both CPs matches, the same value is set for both virtual CPs. On the other hand, if the attribute PCIs in both CPs do not match, an invalid value is set for the virtual CP.

  Also, when either the starting CP or the terminating CP holds the attribute PCI, and the virtual CP is in the vicinity of the CP holding the attribute PCI, the same value as that CP is set. . On the other hand, if the virtual CP is not in the vicinity of the CP holding the attribute PCI, an invalid value is set. Note that the neighborhood means, for example, a case where the straight line distance from the CP holding the attribute PCI to the virtual CP is 10% or less of the straight line distance between the CP on the start side and the CP on the end side.

Furthermore, if neither the start-side CP nor the end-side CP holds the attribute PCI, an invalid value is set for the virtual CP.
The attribute PDM is a distance of the road from a straight line connecting to the next CP (see FIG. 4). The attribute PDM is set by multiplying the value before assigning the virtual CP by the correction value. Specifically, the placement position ratio of the virtual CP is calculated, and the correction value is calculated by the following formula.

(I) When the arrangement position ratio is less than 50%

Correction value = 0.6 × arrangement position ratio / 100 Equation 1

(Ii) When the arrangement position ratio is 50% or more

Correction value = 1.4 × placement position ratio / 100−0.4 Equation 2

Here, the arrangement position ratio is calculated by the following equation.

The arrangement position ratio when recalculating the attribute PDM of the CP on the start side is

Straight line distance between start side CP and virtual CP / total straight line distance from start side CP to end side CP
... Formula 3

It becomes.

In addition, the arrangement position ratio when recalculating the attribute PDM of the virtual CP is

Linear distance between virtual CP and terminating CP / Total linear distance from starting CP to terminating CP
... Formula 4

It becomes.

  The total straight distance is the sum of the straight distances when passing through the virtual CP from the start end CP to the end CP, and the straight distance between the start CP and the virtual CP and the straight line between the virtual CP and the end CP. Total distance.

  The attribute PD is a driving distance to the next CP having the attribute PD (see FIG. 4), and is not necessarily an attribute given to the CP. Therefore, two CPs straddling the parcel boundary may not have the attribute PD. Therefore, recalculation is performed using the CP having the attribute PD closest to the virtual CP. If there is no CP having the attribute PD, the attribute PD is not recalculated.

  Specifically, the value of the attribute PD possessed by the CP is apportioned according to the straight line distance to the virtual CP. That is, the value of the CP attribute PD is obtained by multiplying the value of the original attribute value PD of the CP by the arrangement position ratio of Equation 3 above. Further, the value of the attribute PD of the virtual CP is obtained by multiplying the original attribute value PD of the start point CP by the arrangement position ratio of the above equation 4.

<4.2 Refinement of CP>
Here, returning to the description of FIG. 8, in S130, the search is narrowed down to the CPs in the target parcel. In this process, only the CP in the target parcel (including the virtual CP) is a processing target. It is conceivable that the CP array continues beyond the boundary of the target parcel, but it is sufficient to perform matching only for the target parcel that is the display target in the application.

<4.2.1 Method without Assigning Virtual CP>
In the above-described method, the virtual CP is assigned to process the virtual CP up to the virtual parcel, assuming that the CP exists at the boundary of the target parcel. However, as another embodiment, the original CP delivered as DLR data is changed. Processing may be performed as the end CP of the target parcel.

  For example, when the CP straddles the boundary of the target parcel as shown in FIG. 10B, the CP (a) in the target parcel may be the termination CP. Alternatively, CP (b) that appears first outside the target parcel may be set as the termination CP. However, since CP (a) and CP (b) may exist away from the boundary of the target parcel, for example, depending on the distance from the boundary of the target parcel of CP (a) and CP (b), It may be determined whether the last CP is the termination CP.

<4.2.2 Another Embodiment (1)>
FIG. 33 shows a conversion process as another embodiment (1).
In the first S500, a CP immediately before straddling the boundary inside the boundary of the target parcel (CP immediately before crossing the border) is acquired. The CP just before crossing is the last CP (in the array of CPs in the target parcel) that is delivered to indicate the road in the target parcel when the road indicated by the CP jumps from the target parcel to the adjacent parcel outside the target parcel. Last CP).

  In subsequent S510, the distance from the CP acquired in S500 to the boundary of the target parcel is measured. As an example, as shown in FIG. 10B, an intersection point K between a straight line connecting two CP (a) and CP (b) and the boundary of the target parcel is calculated, and a linear distance A to the intersection point K is measured. .

  In the next S520, it is determined whether or not the distance from the boundary of the target parcel to the CP just before the crossing is equal to or greater than a predetermined distance. This predetermined distance can be set in consideration of the accuracy of road estimation near the boundary of the target parcel. If it is determined that the distance is less than the predetermined distance (S520: NO), the CP inside the target parcel is set as the processing target CP in S530. On the other hand, if it is determined that the distance is equal to or greater than the predetermined distance (S520: YES), in S540, in addition to the CP inside the target parcel, the CP immediately after the crossing border is set as the processing target CP. The CP immediately after the crossing border is the first CP outside the target parcel to be distributed following the CP just before the crossing border.

  In the determination process of S520, the distance between the CP just before the crossing and the boundary of the target parcel is used, but the determination is made using the distance between the CP closest to the boundary of the target parcel inside the target parcel and the boundary of the target parcel. May be. This is because if this distance is more than a predetermined distance, the distance to the boundary of the CP just before the crossing is naturally more than the predetermined distance.

  In other words, if a negative determination is made in S520, the determination is made based on whether or not the distance to the CP just before the crossing is less than a predetermined distance. The determination may be made by determining whether the distance to the CP closest to the boundary is equal to or greater than a predetermined distance.

  Speaking of the termination CP shown in FIG. 10 (b), normally, CP (a) inside the target parcel is defined as the termination CP (S530 in FIG. 33), and when the linear distance A is equal to or greater than a predetermined distance, CP (b) outside the target parcel is defined as a termination CP. At this time, since processing up to CP (a) is normally processed, the time required for processing can be shortened as much as possible. In addition, for example, when CP (a) is extremely far from the boundary of the target parcel, if up to CP (a) is processed as usual, the road estimation accuracy decreases when there are many candidate roads. End up. Therefore, in this embodiment, up to CP (b) is the processing target in this case, so that the road estimation accuracy near the boundary of the target parcel is improved and the display of traffic information such as traffic jam information is suppressed from being insufficient. it can.

<4.2.3 Another embodiment (2)>
FIG. 34 shows a conversion process as another embodiment (2).
In the first S600, the CP immediately after straddling the boundary outside the boundary of the target parcel (CP immediately after crossing the border) is acquired. The CP immediately after crossing the border is as described above. When the road indicated by the CP jumps out from the target parcel to the adjacent parcel outside the target parcel, the first CP distributed in the parcel (in the arrangement of CPs in the parcel) First CP). In subsequent S610, the distance from the CP acquired in S600 to the boundary of the target parcel is measured. As an example, as shown in FIG. 10B, an intersection point K between a straight line connecting two CP (a) and CP (b) and the boundary of the target parcel is calculated, and a linear distance B to the intersection K is measured. .

  In the next S620, it is determined whether or not the distance from the boundary of the target parcel to the CP immediately after the crossing is equal to or greater than a predetermined distance. This predetermined distance can be set in consideration of the processing load on the core point. If it is determined that the distance is less than the predetermined distance (S620: NO), in S630, in addition to the CP inside the target parcel, the CP immediately after the crossing border is set as the processing target CP. On the other hand, when it is determined that the distance is greater than or equal to the predetermined distance (S620: YES), the CP inside the target parcel is set as the processing target CP.

  In the determination process of S620, the distance between the CP immediately after the crossing and the boundary of the target parcel is used, but the determination is made using the distance between the CP closest to the boundary of the target parcel outside the target parcel and the boundary of the target parcel. May be. This is because if this distance is more than a predetermined distance, the distance to the boundary of the CP immediately after crossing the border is naturally also more than a predetermined distance.

  That is, if a negative determination is made in S620, the determination is made based on whether or not the distance to the CP immediately after the crossing is less than a predetermined distance, while if an affirmative determination is made in S620, the CP immediately before the crossing or the outside The determination may be made by determining whether the distance to the CP closest to the boundary is equal to or greater than a predetermined distance.

  Speaking of the termination CP shown in FIG. 10 (b), CP (b) that is usually outside the target parcel is defined as the termination CP (S630 in FIG. 34), and when the linear distance B is equal to or greater than a predetermined distance, CP (a) inside the target parcel is defined as a termination CP. In this case, since the processing target is usually up to CP (b), the road estimation accuracy near the boundary of the target parcel is improved, and the display of traffic information such as traffic jam information can be suppressed. Further, when CP (b) is extremely far from the boundary of the target parcel, since CP (a) is a processing target, it is possible to suppress an increase in time required for processing.

<4.2.4 Another Embodiment (3)>
FIG. 35 shows a conversion process as another embodiment (3).
In the first S700, the CP immediately before crossing the boundary of the target parcel and the CP immediately after crossing are acquired. In subsequent S710, the distance from each CP acquired in S700 to the boundary of the target parcel is measured. As an example, as shown in FIG. 10B, an intersection K between a straight line connecting two CP (a) and CP (b) and the boundary of the target parcel is calculated, and linear distances A and B to the intersection K are calculated. taking measurement.

  In the next S720, it is determined whether or not the distance from the outer CP (CP immediately after crossing) to the boundary is greater than the distance from the inner CP (CP immediately before crossing) to the boundary. If it is determined that the distance from the outer CP is large (S720: YES), the CP inside the target parcel is set as the processing target CP in S730. On the other hand, when it is determined that the distance from the outer CP is not large (that is, the same or smaller) (S720: NO), the CP immediately after the boundary is set as the processing target CP in addition to the inner CP of the target parcel.

  10B, the straight line distances A and B are compared, and when the straight line distance B is larger than the straight line distance A, CP (a) inside the target parcel is set as the end CP. When the straight line distance B is less than the straight line distance A, CP (b) outside the target parcel is defined as the termination CP. At this time, since the end CP is determined according to the straight line distance, when CP (b) is a processing target, the road estimation accuracy near the boundary of the target parcel is improved, and traffic such as traffic jam information is transmitted. Information can be displayed sufficiently. On the other hand, when CP (a) is a processing target, the time required for processing can be shortened as much as possible.

  In any case, the time required for the process related to the virtual CP assignment (S110 in FIG. 8) is shortened compared to the configuration for assigning the virtual CP, and the attribute recalculation process associated with the virtual CP assignment This is advantageous in that the time required for (S120) is shortened.

<4.2.5 Inheritance of attributes>
By the way, in S130 in FIG. 8, an attribute takeover process is also performed as the CP to be processed is determined.

  In the CP as DLR data, only the CP indicating the intersection (the CP in which the IP flag is set) has various attributes. Therefore, the non-shape attribute of such a CP is inherited to the CP existing between the intersections. Accordingly, it is possible to save the trouble of acquiring the non-shape attribute every time processing described later. The attributes specifically inherited are the attribute FC, the attribute FW, the attribute RD, the attribute DD, and the attribute AFR.

  Through the conversion process described above, an array of CPs converted into intermediate data is created. That is, at this stage, the CP has various attributes shown in FIG. On the other hand, at least one of the nodes and links constituting the map data has an attribute corresponding to the attribute of the CP. Therefore, the road indicated by the CP is finally estimated by comparing various attributes.

<5. Details of the matching process (second half)>
Returning to the description of the matching process in FIG. 3, in S200, a candidate link search process is executed. In the candidate link search process, candidate link strings are searched, and then narrowed down in units of links. In the next S300, candidate selection processing is executed. The candidate selection process further narrows down the links narrowed down in S200. In subsequent S400, a road matching process is executed. The road matching process estimates a link connecting CPs. The processes in S200 to S400 are repeatedly executed by the number of CPs with each CP narrowed down in S130 in FIG. 8 as the target CP.

Therefore, in the following, these processes will be described in detail. An example of the candidate link search process of S200 is shown in FIG.
In the first S210, information on the target CP is acquired. This process acquires the type, latitude, and longitude of the CP. The type of CP is a distinction between a CP indicating an intersection, a virtual CP, and other CPs.

  In subsequent S220, a search range is set. In this process, a search range of links centering on the target CP is set. In particular, this search range is set as a polygon whose boundary is composed of vertical or horizontal line segments, as indicated by broken lines in FIG. In this embodiment, the range is set such that a square and a cross shape longer than one side of the square are combined.

  In the next S230, a parcel within the search range is acquired. In this process, since nodes and links are managed in units of parcels, all the parcels related to the search range are acquired when selecting a link string. For example, as shown in FIG. 13B, three parcels P1, P2, and P3 are acquired according to the search range.

  In subsequent S240, a link row selection process is executed. In this link string selection process, links are searched in link string units based on non-shape attributes. The “link train” is a series of links having the same road classification such as a highway and a general road. Here, a link string including a part of the parcel acquired in S230 is a search target. That is, at this stage, it is not determined whether it is in the search range, but the link string is first narrowed down by attribute.

<5.1 Link string selection processing>
Here, the link row selection process will be described in detail. An example of the link string selection process is shown in FIG. In this link string selection process, extraction is performed in units of link strings based on attributes FC, attribute FW, and attribute RD which are non-shape attributes.

  In the first step S241, the search is narrowed down to link strings having the same attribute FC. The attribute FC is an attribute indicating a road segment as described above (see FIG. 5). Here, the road segment assigned to the link constituting the link string is compared with the attribute FC of the target CP to narrow down the link string. However, the road segment may change in the middle of the link row. Therefore, when the road sections of all the links constituting the link string match the attribute FC, the link string is set as a candidate link string.

  In the next S242, the search is narrowed down to link strings having the same attribute FW. As described above, the attribute FW is an attribute indicating the physical road type (see FIG. 6). Here, the physical road type assigned to the links constituting the link string is compared with the attribute FW of the target CP to narrow down the link string. However, the physical road type may change in the middle of the link row. Therefore, when the physical road types of all the links constituting the link string match the attribute FW, the link string is set as a candidate link string.

  In subsequent S243, the search is narrowed down to link strings having the same attribute RD. As described above, the attribute RD is an attribute indicating a road number or a road name (see FIG. 4). Here, only when the attribute RD indicates a road number, a match by road number is determined. That is, the road number assigned to the link constituting the link string is compared with the attribute RD of the target CP to narrow down the link string. However, the road number may change in the middle of the link string. Therefore, when the road numbers of all the links constituting the link string match the attribute RD, the link string is set as a candidate link string.

By such a link string selection process, the link string narrowed down by the non-shape attribute remains as a candidate link string.
<5.2 Link extraction>
In S250 of FIG. 12, a link that partially includes the search range centered on the target CP is extracted. Since a link string is selected in the link string selection in S240, here, a link that includes a part in the search range is extracted from the links constituting the selected link string.

  Here, the “link including a part in the search range” will be described. In the following, the description will be continued on the assumption that nodes are normally given to intersections, nodes are given to parcel boundaries, and shape interpolation points are set between nodes as necessary.

<5.2.1 When the target CP is an intersection>
When the target CP indicates an intersection, processing that considers the node is performed. That is, when the target CP indicates an intersection, the target CP is associated with a node. Accordingly, in this case, a link in which one of the two nodes constituting the link is included in the search range is referred to as a “link in which a part is included in the search range”.

  For example, it is assumed that the target CP shown in FIG. In this case, the nodes A, B, and C included in the search range are associated with the target CP. For this reason, the links L1, L2, L3, L4, and L5 having the nodes A, B, and C as one of the end points are “links that are partially included in the search range”.

  However, the map data has a level corresponding to the scale (hereinafter referred to as “parcel level”). Since this parcel level is switched by switching the scale by the user, link extraction is performed at all parcel levels. The parcel level is set in order from the lower level as LV0 → LV2 → LV4 → LV6 → LV8 → LV10. And as the higher level is reached, intersections and roads are thinned out.

  CP is data corresponding to LV0 which is the lowest parcel level. Therefore, at the parcel level higher than LV2, a node corresponding to the intersection does not necessarily exist. Therefore, in the extraction process at a parcel level higher than LV2, a link in which a node does not exist in the search range is extracted when a certain condition is satisfied. “When a certain condition is satisfied” will be described later.

<5.2.2 When the target CP is a virtual CP>
Even in the case where the target CP is a virtual CP, processing is performed in consideration of the node. In the map data specification, nodes are always set at parcel boundaries. Therefore, when the target CP is a virtual CP, the virtual CP is assigned to the parcel boundary. Therefore, if there is a node existing on the same parcel boundary, the virtual CP is associated with the node.

  Therefore, if any of the nodes that make up a link is present at the parcel boundary, and the node existing at this parcel boundary is included in the search range, the link is defined as “a link that includes part of the search range”. .

  For example, assuming that the target CP shown in FIG. 15B indicates a virtual CP and the node D is a node existing on the parcel boundary, the links L6 and L8 with the node D as an end point are “partly included in the search range”. Link ".

<5.2.3 When the target CP is neither an intersection nor a virtual CP>
In this case, since it is not known whether the target CP is associated with the node, the link is extracted without being aware of the node. In other words, when the target CP is neither an intersection nor a virtual CP, a link whose node does not exist in the search range also satisfies a certain condition in addition to the case where at least one of the nodes constituting the link is included in the search range To extract. Link extraction in this case is performed in the same manner at all parcel levels.

<5.2.4 When certain conditions are met>
“When a certain condition is satisfied” refers to the following two cases. One is a case where a shape interpolation point is set between nodes and a part of a line segment connecting the shape interpolation points is included in the search range. For example, in the link L9 shown in FIG. 15C, since a part of a line segment connecting the shape interpolation points (small black circles) is included in the search range, the nodes E and F are not included in the search range. “Links partially included in the search range”. On the other hand, the link L10 does not become a “link in which a part is included in the search range” because a part of the line connecting the shape interpolation points is not included in the search range. The other is a case where a shape interpolation point is not set between the nodes and a part of a line segment connecting the two nodes is included in the search range. For example, the link L11 shown in FIG. 15D is a “link in which a part is included in the search range” because a part of the line segment connecting the two nodes G and H is included in the search range.

<5.2.5 Others>
Returning to the description of FIG. 12, in the subsequent S260, the links whose attributes RD match among the links partially included in the search range are narrowed down. If the attribute RD indicates a road number, the determination is made in S243 in FIG. 14. Therefore, here, when the attribute RD is a road name, a match by the road name is determined. Since the road name has a maximum of 5 characters, it is determined whether or not the character string as the value of the attribute RD is included in the road name given to the link.

  In the next S270, the number of links is narrowed down to be within “10”. Specifically, when the number of links exceeds “10” by narrowing down to S260, “10” links are selected in order from the closest to the target CP. Whether or not it is close to the target CP is determined by obtaining a distance, for example, by dropping a perpendicular line from the target CP to each link.

<5.3 Candidate selection processing>
Next, the candidate selection process of S300 in FIG. 3 will be described. In the candidate selection process, candidates for links and nodes are selected. An example of the candidate selection process is shown in FIG.

  In the first step S310, the non-shape attribute match determination is performed. Although details will be described later, here, a match determination using the attribute FC, the attribute FW, the attribute IT, the attribute RDI, the attribute DD, and the attribute AFR is performed on the link extracted in S200 or the nodes constituting the link. . Note that the attribute FC and the attribute FW are used when narrowing down the link string (S241 and S242 in FIG. 14), which are duplicated processes, but are performed again to ensure certainty. The match determination here is a fine match determination using the attribute IT, the attribute RDI, the attribute DD, the attribute AFR, and the like for the sake of certainty.

  In the next S320, candidate links are narrowed down based on the determination result of the non-shape attribute. Specifically, as a result of the matching determination by the non-shape attribute in S310, a link having a large number of matches with each attribute is set as a candidate link.

  In subsequent S330, the shape system attribute matching determination is performed. Although details will be described later, here, only the link that has been selected as a candidate link in S320 or a node that constitutes the link is subjected to matching determination using the attribute BR, attribute PCI, attribute CA, and attribute DCA.

  In subsequent S340, candidate links are narrowed down by the determination result of the shape system attribute. Specifically, a link having a large number of matches with each attribute as a result of the match determination by the shape system attribute in S330 is set as a candidate link.

<5.3.1 Non-shape attribute match determination>
Here, details of the non-shape attribute matching determination in S310 will be described. An example of non-shape attribute matching is shown in FIG.

  In the first step S311, a match determination with the attribute FC is performed. The attribute FC is an attribute indicating a road segment as described above (see FIG. 5). Here, it is determined whether the road segment assigned to the determination target link matches the attribute FC of the target CP.

  In the subsequent S312, a match determination with the attribute FW is performed. As described above, the attribute FW is an attribute indicating the physical road type (see FIG. 6). Here, it is determined whether the physical road type assigned to the determination target link matches the attribute FW of the target CP.

  In the next S313, a match determination with the attribute IT is performed. As described above, the attribute IT is an attribute indicating the classification of the intersection (see FIG. 7). Therefore, this match determination is performed only when the target CP indicates an intersection.

  Note that an attribute corresponding to the attribute IT has at least one of the node and the link. Therefore, when the node has intersection classification information indicating the classification of the intersection, it is determined whether the intersection classification information of the node corresponding to the target CP matches the attribute IT of the target CP. If the link has intersection classification information indicating the classification of the intersection, it is determined whether the intersection classification information of the link connected to the node corresponding to the target CP matches the attribute IT of the target CP. In the latter case, since the link may have intersection classification information of two nodes, it is determined whether any intersection classification information matches the attribute IT.

  In the subsequent S314, a match determination with the attribute RDI is performed. As described above, the attribute RDI indicates the name of the intersection (see FIG. 4). Therefore, this match determination is performed only when the target CP indicates an intersection. Here, it is determined whether the intersection name of the node corresponding to the target CP matches the attribute RDI of the target CP. This determination is made based on whether or not the character string that is the value of the attribute RDI is included in the node intersection name.

  In next step S315, a match determination is performed using the attribute DD and the attribute AFR. The attribute DD indicates the permitted legal driving direction, and the attribute AFR is a flag indicating whether or not the attribute DD can be referred to (see FIG. 4). Here, the attribute of the determination target link (one-way code) is compared with the attribute DD of the target CP for determination. The determination is performed except when the attribute AFR is set as “1” and the attribute DD is valid and the attribute DD is not “0” (undefined). By the way, the attribute DD indicates the direction in which the road can pass for certain traffic information, and the forward / reverse direction of both cannot be determined by comparison with the one-way code of the link. Therefore, here, it is not determined whether the traffic directions coincide with each other, but only to determine whether the road is a one-way road or a two-way road.

<5.3.2 Judgment of coincidence of shape system attributes>
Next, details of the coincidence determination of shape system attributes in S330 will be described. An example of shape system attribute matching determination is shown in FIG.

  In the first S331, a match determination with the attribute BR is performed. As described above, the attribute BR indicates a positional angle to the next CP (see FIG. 4). On the other hand, the link has a legal traveling direction as an attribute. Therefore, here, when the traveling direction of the link is greatly different from the direction based on the attribute BR of the target CP, the mismatch is determined. For example, when the angle formed by the vector indicating the traveling direction of the link and the vector based on the attribute BR toward the next CP is equal to or greater than a predetermined angle (for example, 90 degrees), it can be determined that there is a mismatch. When the link can travel in both directions, the determination is made with two vectors indicating the traveling direction.

  For example, in FIG. 19A, the determination is made at an angle formed by the vectors VL1, VL21, VL22, and VL3 indicating the traveling directions of the links L1, L2, and L3 and the vector VB based on the attribute BR. Here, when the angle formed by both vectors is 90 degrees or more, it is determined that there is a mismatch. The link L2 can travel in both directions, and is determined by two vectors VL21 and VL22 that indicate the traveling direction.

  Here, the angle formed by the vector VB and the vector VL1 is close to 180 degrees and is 90 degrees or more, so the link L1 is determined to be “mismatch”. Further, the angle formed by the vector VL22 and the vector VB is 90 degrees or more. However, since the angle formed by the vector VL21 and the vector VB is smaller than 90 degrees, the link L2 is determined as “match”. Furthermore, since the angle formed by the vector VL3 and the vector VB is smaller than 90 degrees, the link L3 is determined to be “match”.

  In other words, when the case of 90 degrees or more is regarded as a mismatch, as shown in FIG. 19B, when the travel direction of the link is indicated by the vectors V1 and V2 with respect to the vector VB based on the attribute BR, “ It is determined as “match”, and when indicated by the vectors V3 and V4, it is determined as “mismatch”.

  In S332 in FIG. 18, a match determination is performed using the attribute PCI. As described above, the attribute PCI indicates which roadway is selected when there are a plurality of roadways in the same direction (see FIG. 4).

  The attribute PCI includes a sequence number indicating the number of roadways and which one of the roadways. However, when trying to determine a match with this attribute, it is necessary to specify all the links that run in parallel.

  Therefore, as shown in FIG. 20, in the first S333, a search area centering on the target CP is set. This search area may be inside a circle centered on the target CP. For example, the inside of a circle whose radius is a predetermined distance (for example, 150 m) is used as a search area. Since a default value of the search area may be set as one of the PCI attributes, in this case, the inside of a circle whose radius is a value obtained by adding a predetermined distance (for example, 150 m) to the default value is set as the search area. It is possible to do.

  In continuing S334, the link in a search area is extracted. Here, since one of the attribute PCIs has an indicator type, a virtual straight line in the latitude direction or the longitude direction is drawn based on this indicator type, and a link that intersects this virtual straight line is extracted.

  In the next S335, it is determined whether or not the number of extracted links matches the number of roadways. If it is determined that the number of roadways is the same (YES in S335), the link is identified by the sequence number in S336, it is determined whether or not the link matches the determination target link, and the attribute PCI The match determination ends. On the other hand, if it is determined that the number of roadways does not match (NO in S335), the process of S336 is not executed and the match determination based on the attribute PCI is terminated. This is because if the number of roadways does not match, there is a risk of erroneous determination.

  Returning to FIG. 18, in S <b> 337, a match determination is performed using the attribute CA and the attribute DCA. The attribute CA indicates an angle with respect to the side road, and the attribute DCA indicates a connection distance with respect to the side road (see FIG. 4). Here, first, the direction of the side road is detected. Specifically, as shown in FIG. 21A, when the attribute CA is a negative value, it is determined that there is a side road on the left side with respect to the direction toward the next CP (the direction based on the attribute BR). On the other hand, as shown in FIG. 21B, when the attribute CA is a positive value, it is determined that there is a side road on the right side with respect to the direction toward the next CP (the direction based on the attribute BR). If the attribute CA is “0”, the direction of the side road cannot be specified, and therefore no match determination is performed. Next, a match determination with the attribute CA and the attribute DCA is performed from the positional relationship between the determination target link and the side road of the link.

  As a result of the candidate selection process in S300 described above, candidate links are selected around each CP. The process of estimating the link connecting these links, that is, the link connecting the CPs is the road matching process (S400) in FIG.

<5.4 Road matching process>
Next, the road matching process in S400 will be described. An example of the road matching process is shown in FIG. In the road matching process, two CPs are set as target CPs, a CP located on the start side is set as the start side CP, and a CP located on the end side is set as the end side CP.

<5.4.1 Road search>
In the first S410, a road connecting the start candidate link to the end candidate link is searched. The starting end candidate link is a link extracted with respect to the starting end side CP. The termination candidate link is a link extracted for the termination side CP. Here, all roads connecting the start candidate link and the end candidate link are searched. For example, as shown in FIG. 23 (a), the start end CP is CP (a), the end CP is CP (b), the link extracted for CP (a) is SL, and CP (b) Let GL be the link extracted for. At this time, when searching for a road between the start candidate link SL and the end candidate link GL, all combinations of the two nodes SN1, SN2 constituting the link SL and the two nodes GN1, GN2 constituting the link GL think of. Specifically, a road connecting SN1 → GN1, SN1 → GN2, SN2 → GN1, SN2 → GN2 is searched.

  In the next S420, road unit candidate selection is executed. This process is to determine candidate roads among the roads searched in S410 based on various shape system attributes. In subsequent S430, the road is narrowed down based on the determination result in S420.

  In the next S440, it is determined whether or not a road is uniquely determined. If the road is uniquely determined (YES in S440), the matching result is output in S450, and then the road matching process is terminated. On the other hand, if the road is not uniquely determined (NO in S440), the process of S450 is not executed and the road matching process is terminated.

<5.4.2 Omission of road search>
In addition, although such a road matching process (S410-S440) is repeated with respect to two object CP which becomes a pair, in S410 mentioned above, the road searched using the result of a previous road matching process is abbreviate | omitted. There is a case.

  For example, as shown in FIG. 23B, it is assumed that the road SL → CL → GL1 is uniquely determined as a result of the road matching process between CP (a) and CP (b). At this time, when a road matching process from CP (b) to the next CP is performed, a road from the link GL1 is searched. That is, even if the candidate link includes the link GL2, the road search from the link GL2 is omitted.

The road search is also omitted depending on the type of CP.
When CP (b) is an intersection, for example, as shown in FIG. 23C, in order to correspond to the node GN2 indicating the intersection, the link GL is searched for a road connected to one node GN1. That is, a road connecting SN1 → GN1 and SN2 → GN1 is searched.

  When CP (b) is a virtual CP, for example, if the node GN2 shown in FIG. 23 (d) is a node located at the boundary of the target parcel, this node becomes the start or end of the road. Therefore, in this case as well, for the link GL, a road connected to one node GN1 is searched. That is, a road connecting SN1 → GN1 and SN2 → GN1 is searched.

Furthermore, it is conceivable to reduce the time required for road search by not searching the road once searched.
For example, assume that a search of nodes N1 → N2 → N3 → N4 indicated by a solid line in FIG. At this time, when the road search from the node N1 is performed for the second time, the road search of the node N1 → N5 → N2 is performed, but since the search from the node N2 has already been performed, the search from the node N2 is performed. Absent.

  In addition, when a link branches from a node, a road search is performed with priority given to the link. For example, when a road search is performed from a certain node, an angle from a reference direction of a link connected to the node is obtained with reference to a direction connecting the target CP, and only a link within a predetermined angle is set as a road search target. . In FIG. 24B, the road search from the node N1 is performed, and the reference of the links L1, L2, and L3 connected to the node N1 is based on the direction from CP (a) to CP (b). The angles a1, a2, and a3 from the direction are obtained. Here, if the angles a1 and a2 are within a predetermined angle, the links L1 and L2 are targeted for road search. On the other hand, when the angle a3 exceeds the predetermined angle, the link L3 is excluded from the road search target.

  Furthermore, when a search from a node is performed, a link having the same road type as the road type of the link up to the node is set as a road search target. For example, in FIG. 24C, a road from CP (b) to CP (c) is searched from the node N1. Here, it is assumed that the road between CP (a) and CP (b) is determined as the link L1. At this time, since the road type of the link L1 is “national road”, the link L2 having the road type “national road” is set as a road search target when searching for a road from the node N1. The link L3 whose road type is “prefectural road” is excluded from road search targets.

<5.4.3 Canceling road search>
By the way, the road search in S410 does not necessarily end within a predetermined time. Therefore, the road search may be stopped halfway.

  For example, it is conceivable to use the attribute PD. In FIG. 25A, it is assumed that both CP (a) and CP (b) have an attribute PD. At this time, when searching for a road from the node N1 to the node N2, the cumulative distance at the shape interpolation points K1, K2, K3, K4,... Is calculated. When this cumulative distance becomes larger than the PD value by a certain value or more, the road search is stopped.

  For example, it is conceivable to use an attribute PDM. As shown in FIG. 25 (b), a U-shaped distance as indicated by a broken line is calculated using a linear distance between CP (a) and CP (b) and a PDM value. At this time, when searching for a road from the node N1 to the node N2, the cumulative distance at the shape interpolation points K1, K2, K3, K4,... Is calculated. Then, when the accumulated distance becomes larger than the distance indicated by the broken line, the road search is stopped.

  For example, it is conceivable to use the number of passing nodes. As shown in FIG. 25 (c), when searching for a road from the node N1 to the node N2, the number of passing nodes N3, N4, N5, N6 is counted. Then, when the number of nodes exceeds a certain value, the road search is stopped.

  For example, it can be considered that CP is an intersection. In FIG. 25D, it is assumed that both CP (a) and CP (b) are intersections. At this time, if the nodes N1 and N2 are intersections, the nodes N1 and N2 correspond to CP (a) and CP (b), respectively. Therefore, the linear distance between the nodes N1 and N2 is calculated in advance. . Then, when searching for a road from the node N1 to the node N2, the cumulative distance to the passing nodes N3, N4, N5, N6 is calculated, and this cumulative distance is the straight line distance between the nodes N1, N2 calculated in advance. If it becomes larger than a certain value, the road search is stopped.

Although four methods have been described here, the search for a road that seems to be wrong is interrupted in any way, so that the processing load is reduced.
<5.4.4 Selection of road unit candidates>
Here, the road unit candidate selection in FIG. 22 will be described. An example of road unit candidate selection is shown in FIG. In the road unit candidate selection, “OK” or “NG” determination is performed in units of roads extracted between CPs.

  In the first S421, the determination is made based on the attribute CA / DCA. As described above, the attribute CA indicates the angle with respect to the side road, and the attribute DCA indicates the connection distance with respect to the side road (see FIG. 4). Here, it is assumed that the link specified by the attribute CA and the attribute DCA in the searched road is a side road, and NG determination is performed on the road including the link. Specifically, as illustrated in FIG. 27A, if there is a link around the position coordinates indicated by the attribute CA and the attribute DCA (range indicated by a broken line), the road including the link is determined as “NG”. . For example, in FIG. 27B, road 1 and road 2 are shown as roads connecting CPs, but road 1 including a link determined as a side road from attribute CA and attribute DCA is determined as “NG”. Will be.

  In S422 in FIG. 26, the determination is based on the attribute BR / DMB. As described above, the attribute BR indicates an angle to the next CP, and the attribute DMB indicates a linear distance to the next CP (see FIG. 4). When there is a difference between the data indicated by the CP and the map data of the navigation device 10, a more probable road can be identified by using this attribute BR / DMB. Therefore, as shown in FIG. 28, the determination of the road between CP (b) and CP (c) is performed using the attribute BR and attribute DMB of CP (a) immediately before CP (b). Specifically, the road including the link L around the position coordinate specified from the attribute BR and the attribute DMB (the range indicated by the broken line of the symbol H) is determined as “OK”. In addition, since it is determination of the road between CP (b) -CP (c), it is on condition that the node or shape interpolation point which comprises the link L enters into a predetermined range. For example, it is conceivable that the predetermined range is inside a rectangle (indicated by a two-dot chain line in the figure) set from a line segment connecting CP (b) and CP (c) and the attribute PDM. Of course, the predetermined range is not limited to such a rectangle, and may be the inside of an ellipse passing through CP (b) and CP (c).

  In next S423, determination is made based on the attribute PD. As described above, the attribute PD is a driving distance to the next CP having the attribute PD (see FIG. 4). Therefore, it is determined whether or not the road can be a candidate based on the driving distance of the searched road.

  Specifically, as shown in FIG. 29, when calculating the driving distance between CP (a) and CP (b), each CP does not necessarily correspond to a node, and therefore, from CP (a) to the link L1. , And a perpendicular line from CP (b) to the link L3 is taken, and the intersections are set as M and N, respectively. Then, the driving distance from the intersection M to the intersection N is calculated.

  Note that the driving distance from the intersection M to the next node is obtained by calculating the ratio of the driving distance in the link L1. Here, when the distance from the intersection M to the next node is m (%) of the link L1, the operation distance from the intersection M to the next node is calculated by multiplying the operation distance of the link L1 by (m / 100). .

  Similarly, the driving distance from the intersection N to the immediately preceding node is obtained by calculating the ratio of this driving distance in the link L3. Here, when the distance from the intersection N to the immediately preceding node is n (%) of the distance of the link L1, the driving distance of the link L3 is multiplied by (n / 100) to drive from the intersection N to the immediately preceding node. Calculate the distance.

  By the way, not all CPs have the attribute PD. Therefore, if CP (a) has the attribute PD but CP (b) does not have the attribute PD, as shown in FIG. Hold. Next, assuming that CP (c) shown in FIG. 30 (b) has an attribute PD, road determination based on the attribute PD is performed between CP (a) and CP (c). That is, road L1-> L2-> L3-> L5-> L6-> L10; road L1-> L4-> L5-> L6-> L10; road L1-> L2-> L3-> L5-> L7-> L8-> L9-> L10; road L1-> L4 NG determination is performed using candidates of L5 → L7 → L8 → L9 → L10. Thereby, as shown in FIG. 30 (c), when the road L1 → L2 → L3 → L5 → L6 → L10 remains, the road L4 between CP (a) -CP (b) and CP (b ) -CP (c) road L7 → L8 → L9 is determined as “NG”.

  In subsequent S424, determination is made based on the attribute PDM. As described above, the attribute PDM indicates the separation distance of the road from the straight line connecting to the next CP (see FIG. 4). In this process, the driving distance along the road is obtained based on the attribute PDM, and it is determined whether or not the searched road is a candidate.

Specifically, the driving distance along the road based on the attribute PDM is calculated by the following formula.

PDM operating distance = half of the circumference with the straight line distance between CPs as the diameter x correction value Equation 5

Here, the correction value is determined by a table from the ratio between the value of the attribute PDM and half of the straight line distance between the CPs. For example, as shown in FIG. 31 (a), if half of the straight line distance between CP (a) and CP (b) is r, the value of the attribute PDM and the distance r are calculated according to the table shown in FIG. 31 (b). A correction value is obtained from the ratio (PDM / r). The table shown in FIG. 31 (b) is a partial extract.

Then, assuming that the driving distance of the candidate road is the road driving distance, when the road driving distance satisfies the following expression, the road is determined to be “OK”.

PDM driving distance x 0.5 <Road driving distance <PDM driving distance x 1.5 ... Equation 6

<5.4.5 Matching result output>
Then, as described above, the roads are narrowed down based on these determination results (S430 in FIG. 22). When the road is uniquely determined (YES in S440), the matching result is output in S450. The matching result is composed of link IDs of all links constituting the uniquely specified road, start point offset distance, end point offset distance, and the like. The starting point offset distance is a distance indicating from which position of the link that is the starting point of the road is the matching starting position. Similarly, the end point offset distance is a distance indicating which position of the link that is the end point of the road is the end position of the matching.

<5.4.6 Determination of road>
Normally, if the road is not uniquely determined (NO in S440), the road matching fails and the matching result is not output. However, it is assumed that the road is uniquely determined in the following cases.

  For example, in FIG. 32A, as a result of searching for a road from CP (a) to CP (b), a road of node N1 → N4 → N5 → N2 → N3 and a node N1 → N4 → N5 → N2 Assume that two of the roads remain. At this time, since the road (nodes N1 → N4 → N5 → N2) indicated by the bold line is included in both roads, this road is determined as a uniquely determined road.

  Further, for example, in FIG. 32B, as a result of searching for a road from CP (a) to CP (b), a road of nodes N1 → N4 → N5 → N2 → N3 and nodes N1 → N4 → N6 → N5 Suppose two roads of → N2 → N3 remain. At this time, since the roads (nodes N1 → N4, N5 → N2 → N3) indicated by bold lines are included in both roads, the two divided roads are determined uniquely.

  Alternatively, in the example shown in FIG. 32B, it is conceivable that the road that is the road is a uniquely determined road. As shown in FIG. 32 (c), the “road” refers to a case where an angle formed by links (for example, an angle indicated by symbol a) is equal to or less than a predetermined angle (for example, 15 degrees).

<6. Effect>
In the another embodiment (1), the CP just before the crossing is acquired inside the boundary of the target parcel (S500 in FIG. 33), and the distance from the acquired CP to the boundary of the target parcel is measured (S510). As an example, as shown in FIG. 10B, an intersection point K between a straight line connecting two CP (a) and CP (b) and the boundary of the target parcel is calculated, and a linear distance A to the intersection point K is measured. . Then, it is determined whether or not the distance to the CP just before the crossing is greater than or equal to a predetermined distance (S520). If it is determined that the distance is less than the predetermined distance (S520: NO), the process target CP is set up to the CP just before the crossing border inside the target parcel (S530). On the other hand, when it is determined that the distance is equal to or greater than the predetermined distance (S520: YES), the CP immediately after the crossing border is set as the processing target CP in addition to the CP inside the target parcel (S540).

  That is, in the case of the termination CP shown in FIG. 10B, the CP (a) that is usually inside the target parcel is the termination CP (S530 in FIG. 33), and the linear distance A is greater than or equal to the predetermined distance. Furthermore, CP (b) outside the target parcel is defined as the termination CP. Thereby, since CP (a) is normally a processing target, the time required for processing can be shortened as much as possible. Further, for example, when CP (a) is extremely far from the boundary of the target parcel, since CP (b) is a processing target, road estimation accuracy near the boundary of the target parcel is improved, and traffic congestion information, etc. It can suppress that the display of traffic information becomes insufficient.

  In the second embodiment (2), the CP immediately after crossing the boundary of the target parcel is acquired (S600 in FIG. 34), and the distance from the acquired CP to the boundary of the target parcel is measured (S610). As an example, as shown in FIG. 10B, an intersection point K between a straight line connecting two CP (a) and CP (b) and the boundary of the target parcel is calculated, and a linear distance B to the intersection K is measured. . Then, it is determined whether or not the distance to the CP immediately after the crossing is greater than or equal to a predetermined distance (S620). Here, when it is determined that the distance is less than the predetermined distance (S620: NO), in addition to the CP inside the target parcel, the CP immediately after the crossing border is set as the processing target CP (S630). On the other hand, if it is determined that the distance is equal to or greater than the predetermined distance (S620: YES), the process target CP is set up to the CP just before the crossing border inside the target parcel (S640).

  That is, in the case of the termination CP shown in FIG. 10B, the CP (b) that is usually outside the target parcel is the termination CP (S630 in FIG. 34), and the linear distance B is equal to or greater than the predetermined distance. Furthermore, CP (a) inside the target parcel is defined as the termination CP. As a result, since CP (b) is normally processed, the road estimation accuracy near the boundary of the target parcel is improved, and the display of traffic information such as traffic jam information can be suppressed. Further, for example, when CP (b) is extremely far from the boundary of the target parcel, since CP (a) is a processing target, the time required for processing can be shortened as much as possible.

  In the another embodiment (3), the CP immediately before crossing over the boundary of the target parcel and the CP immediately after crossing the boundary are acquired (S700 in FIG. 35), and the distance from each acquired CP to the boundary of the target parcel is measured ( S710). As an example, as shown in FIG. 10B, an intersection K between a straight line connecting two CP (a) and CP (b) and the boundary of the target parcel is calculated, and linear distances A and B to the intersection K are calculated. taking measurement. Then, it is determined whether or not the distance from the CP immediately before the crossing to the boundary is larger than the distance from the CP immediately before the crossing to the boundary (S720). Here, when it is determined that the distance from the CP immediately after the crossing is large (S720: YES), the CP immediately before the crossing inside the target parcel is set as the processing target CP (S730). On the other hand, when it is determined that the distance from the CP immediately after the crossing is not large (that is, the same or small) (S720: NO), the CP immediately after the crossing is set as the processing target CP in addition to the CP inside the target parcel (S740).

  That is, for the terminal CP shown in FIG. 10B, the straight line distances A and B are compared (S720 in FIG. 35), and the CP located inside the target parcel when the straight line distance B is larger than the straight line distance A. (A) is defined as the termination CP (S720: YES, S730). On the other hand, when the linear distance B is less than the linear distance A, CP (b) outside the target parcel is defined as the termination CP. At this time, since the end CP is determined according to the straight line distance, when CP (b) is a processing target, the road estimation accuracy near the boundary of the target parcel is improved, and traffic such as traffic jam information is transmitted. Information can be displayed sufficiently. On the other hand, when CP (a) is a processing target, the time required for processing can be shortened as much as possible.

  Further, in these other embodiments, the intersection K between the straight line connecting CP (a) and CP (b) and the boundary of the target parcel is calculated, and the linear distance to the intersection K is measured (S510, FIG. 33). S610 in FIG. 34, S710 in FIG. 35). Thereby, there is a high possibility that a distance closer to the road shape is required.

  Furthermore, in these other embodiments, the time required for the process related to the virtual CP assignment (S110 in FIG. 8) is shortened compared to the configuration of assigning the virtual CP, and the attributes associated with the virtual CP assignment This is advantageous in that the time required for the recalculation process (S120) is shortened.

  Note that the navigation device 10 of the present embodiment constitutes a “road estimation device”, the map data input device 13 constitutes a “map data input means”, the parcel corresponds to a “divided region”, and the target parcel is “display”. This corresponds to “divided area”.

  Further, the CPU 17a constitutes “core point selection means” and “link extraction means”, and the conversion processing shown in FIGS. 33 to 35 corresponds to processing as a function of “core point selection means”, and is shown in FIG. The candidate link search processing corresponds to processing as a function of “link extraction means”.

As mentioned above, this invention is not limited to embodiment mentioned above at all, In the range which does not deviate from the summary, it can be implemented with a various form.
In the above embodiment, when measuring the distance to the boundary of the target parcel (S510 in FIG. 33, S610 in FIG. 34, S710 in FIG. 35), the intersection of the straight line connecting the CP and the boundary across the boundary And the distance from each CP to the intersection was measured. On the other hand, you may make it measure the length of the perpendicular drawn from each CP to the boundary. In this way, the distance from the boundary of the display divided area to the CP can be measured relatively easily.

10: Navigation device 11: Reception device 12: Position detection device 13: Map data input device 14: Operation device 15: Audio output device 16: Display device 17: Control circuit 17a: CPU
17b: ROM
17c: RAM
20: Center 171: Tuning block 172: Application block 173: DLR library block 174: Drawing block

Claims (15)

As position information transmitted along with traffic information, data on a plurality of core points given along the road and having an attribute for specifying the road are received from the outside, and on the map corresponding to the road indicated by the core point A road estimation device that estimates the road by extracting a link,
Map data input means configured to input map data composed of links having attributes corresponding to the attributes of the core points, which are configured in units of divided areas divided into a plurality of areas;
When the core point exists across the boundary of the display divided area which is the display target divided area, the core point existing inside the display divided area is processed in the map data input by the map data input means A core point selection means for selecting as a target core point;
Link extraction means for extracting a link that is a candidate for the road indicated by the processing target core point from the map data based on the attribute of the link and the attribute of the processing target core point;
A road estimation device comprising: The road estimation device according to claim 1,
When the distance from the boundary of the display divided region to the nearest inner core point is equal to or greater than a predetermined distance, the core point selection unit adds the core point existing inside the display divided region to the boundary of the display divided region A road estimation device that selects a core point immediately after crossing, which is a core point immediately after crossing from the inside to the outside, as a processing target core point. The road estimation device according to claim 2,
The core point selection means, when the distance from the boundary of the display division area to the core point immediately before crossing the core point immediately before crossing the boundary from the inside to the outside is a predetermined distance or more, A road estimation device that is selected as a core point to be processed. In the road estimation device according to claim 2 or 3,
The core point selection means, when the distance from the boundary of the display division area to the core point immediately before crossing which is the core point immediately before crossing the boundary from the inside to the outside is less than a predetermined distance, A road estimation device that is selected as a core point to be processed. In the road estimation device according to any one of claims 2 to 4,
The said core point selection means measures the distance from the boundary of the said display division | segmentation area | region to the said core point as a length of the perpendicular drawn from the said core point to the said boundary. The road estimation apparatus characterized by these. In the road estimation device according to any one of claims 2 to 4,
The core point selection means obtains a distance from the boundary of the display divided region to the core point, and finds an intersection of the boundary between the straight line connecting the core point immediately after the crossing over the boundary and the core point immediately before the boundary, and the boundary. A road estimation device that measures the length from the core point to the intersection. As position information transmitted along with traffic information, data on a plurality of core points given along the road and having an attribute for specifying the road are received from the outside, and on the map corresponding to the road indicated by the core point A road estimation device that estimates the road by extracting a link,
Map data input means configured to input map data composed of links having attributes corresponding to the attributes of the core points, which are configured in units of divided areas divided into a plurality of areas;
When the core point exists across the boundary of the display divided region that is a display target divided region, in the map data input by the map data input means, in addition to the core point existing inside the display divided region, A core point selection means for selecting, as a processing target core point, a core point immediately after crossing the core point immediately after crossing the boundary of the display divided region from the inside to the outside;
Link extraction means for extracting a link that is a candidate for the road indicated by the processing target core point from the map data based on the attribute of the link and the attribute of the processing target core point;
A road estimation device comprising: The road estimation device according to claim 7,
When the distance from the boundary of the display divided region to the closest outer core point is a predetermined distance or more, the core point selecting unit selects a core point existing inside the display divided region as a processing target core point. A road estimation device characterized by this. The road estimation device according to claim 8,
The core point selection unit selects a core point existing inside the display divided region as a processing target core point when a distance to the core point immediately after the crossing is a predetermined distance or more. . The road estimation device according to claim 8 or 9,
When the distance to the core point immediately after the crossing is less than a predetermined distance, the core point selecting unit selects the core point immediately after the crossing as a processing target core point. In the road estimation device according to any one of claims 8 to 10,
The said core point selection means measures the distance from the boundary of the said display division | segmentation area | region to the said core point as a length of the perpendicular drawn from the said core point to the said boundary. The road estimation apparatus characterized by these. In the road estimation device according to any one of claims 8 to 10,
The core point selection means calculates the distance from the boundary of the display divided region to the core point, and obtains an intersection of the boundary between the straight line connecting the core point immediately before the crossing and the core point immediately preceding the boundary, and the boundary. A road estimation device that measures the length from the core point to the intersection. As position information transmitted along with traffic information, data on a plurality of core points given along the road and having an attribute for specifying the road are received from the outside, and on the map corresponding to the road indicated by the core point A road estimation device that estimates the road by extracting a link,
Map data input means configured to input map data composed of links having attributes corresponding to the attributes of the core points, which are configured in units of divided areas divided into a plurality of areas;
Immediately after crossing the boundary of the display divided region from the inside to the outside in the map data input by the map data input means when the core point exists across the boundary of the display divided region which is the divided region to be displayed Compare the distance from the core point immediately after crossing that is the core point to the core point immediately before crossing that is the core point immediately before crossing the boundary from the inside to the outside. Up to the core point immediately before the cross-border that is present inside the divided area is selected as the processing target core point. On the other hand, if the distance to the core point immediately before the cross-border is large, in addition to the core point that exists inside the display divided area Core point selection means for selecting up to the core point immediately after the crossing as a processing target core point;
Link extraction means for extracting a link that is a candidate for the road indicated by the processing target core point from the map data based on the attribute of the link and the attribute of the processing target core point;
A road estimation device comprising: The road estimation device according to claim 13,
The said core point selection means measures the distance from the boundary of the said display division | segmentation area | region to the said core point as a length of the perpendicular drawn from the said core point to the said boundary. The road estimation apparatus characterized by these. The road estimation device according to claim 13 or 14,
The core point selection means obtains a distance from the boundary of the display divided region to the core point, and obtains an intersection of the straight line connecting the core point immediately before the crossing and the core point immediately after the crossing across the boundary, and the boundary, A road estimation device that measures the length from the core point to the intersection.

Images