JP2906943B2 - Route calculation method and device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention reads out route network data in a range including a departure point (may be the current position of a vehicle) and a destination from a road map memory in accordance with the setting of a destination or the like by a driver. The present invention also relates to a route calculation method and apparatus which can calculate an optimum route to a destination based on the route network data and indicate the route to a driver.

[0002]

2. Description of the Related Art Conventionally, there has been known a navigation device developed to display the position and orientation of a vehicle on a screen and to facilitate traveling on an unknown land or at night.
The navigation device has a display, a direction sensor, a distance sensor, a road map memory, and a computer mounted on a vehicle, and is stored in a direction data input from the direction sensor, a traveling distance data input from the distance sensor, and a road map memory. The vehicle position is detected based on the coincidence with the road pattern, and the vehicle position is displayed on a display together with the road map.

In this case, in order to select a traveling route from the departure point to the destination, the current route from the departure point to the destination is automatically calculated by the computer according to the input of the destination setting by the driver. A method has been proposed (see JP-A-5-53504). In this method, the road to be calculated is divided into several parts, and the divided points are used as nodes,
A path connecting nodes is set as a link, a node or link closest to a departure point (or a destination) is set as a start point, a node or link closest to a destination (or a departure point) is set as an end point, and a start point to an end point. , The link costs of the routes constituting the tree are sequentially added, and only the route having the lowest link cost to reach the destination or the departure point is selected.

[0004] If a route is calculated in this way and the vehicle is driven along the route, the vehicle will surely reach the destination, which is convenient for a driver who does not know the road. The following is a summary of the conventionally proposed methods for the route calculation. Normal Dijkstra method: In the process of searching the tree,
One route with the lowest cost is sequentially searched from the search range,
A method that ends the loop processing when any route reaches the destination. Potential method: A method in which all routes within the search range are searched, and the route with the lowest cost is finally selected.

In the above method, the Dijkstra method seems to be superior in terms of calculation time. However, in actuality, a process of "sequentially searching one route with the minimum cost" is required. The calculation time is much longer than the potential method (Kobayashi et al., "Navigation System with Recommended Route Display Function", Sumitomo Electric No. 141, PP.1
55-160, September 1992). When the calculation time is long, the utility value for the driver is significantly reduced. Therefore, a method capable of shortening the calculation time is desired.

Therefore, at present, the following potential method is employed.

[0007]

However, recently, the route network database has been enriched, and the calculation time tends to be longer even by the potential method. More specifically, conventionally, the number of links in the path network is relatively small, and the number of loops in the path calculation is small. However, when the number of links is increased, for example, by including the roads to be calculated to thin roads, the tree is expanded, and the number of link searches increases.

As a result, the calculation time greatly increases.
Therefore, the time ratio of the route calculation process by the potential method in the entire process (reading of route network data, rewriting to a link table, route calculation, map display, etc.) has begun to increase dramatically. Therefore, in the end, it has become impossible to reduce the overall time unless the path calculation processing time is reduced. For that purpose, the number of loops must be reduced.

The present invention has been made in view of the above-mentioned problem, and even if the number of route networks increases, the number of loops can be reduced by prioritizing processing to links, thereby shortening the calculation time. It is an object to provide a route calculation method and device.

[0010]

Before describing the configuration of the invention for achieving the above object, a so-called potential method as a premise of path calculation will be described. In this potential method, as described in the section "in..." Of claim 1, starting from a calculation start link, a link connected to the link is searched, and the total cost of the searched connection link is set in advance. If the total cost of the link becomes smaller than the total cost of the same link, the procedure of rewriting the value to the total cost of the connection link is repeated. It is a method of determining.

According to the route calculation method of the present invention, when this potential method is executed, the links to be searched after the second search have a small total cost during the previous search, and the first link has a small total cost.
If the link has been rewritten in the table, the link is searched for prior to the link searched for the first time. According to the method of claim 2, the second table is divided into two parts, the connection link searched for the first time is stored in a non-priority part, and the link searched after the second time is set in a priority part. To store.

According to a third aspect of the present invention, there is provided a route calculation device for executing the route calculation method according to the first aspect.

[0013]

According to this route calculation method and apparatus, the connection link to be searched is a link that has already been searched once, and the total cost is small during the previous search and the link rewritten in the first table is used. If so, the link
It is relatively more likely to form an optimal route than other links that are searched for the first time. Therefore, the connection link of the link is searched with priority.

As a result, one or more connecting links are searched. On the other hand, if there are other links, the connection links are searched for, but if they match any of the searched connection links, the total costs are compared, and if the total costs are equal or greater, they are excluded. Therefore, thereafter, the search is continued only for the route including the link searched preferentially, and the search for the other route is terminated.

This will be described more specifically with reference to the drawings. In Figure 1, a departure point O, and the destination D, and departure link as L ₁ and L _2, the destination link and L _12. Route computation is started from the starting point links L ₁ and _{L 2, L 3, L 4} , L 5, L 6, will be searched and L _8, L _7. Here, the route of L ₁ → L ₃ → L ₅ →.
The route of L ₂ → L ₄ → L ₆ →...

[0016] Directing attention to when it reaches, for example, in the link L _8, at which time, the total cost of the path A is written in the first table. The next link L ₇ according to route B is searched, and thereafter, if the conventional method link according to the route A L _9, links L ₈ according to the path B, the link L ₁₀ according to the path A, according to the path B Link L ₉ and Route B are Route A
It is searched as if chasing after.

If route A is less expensive, route B
Is terminated somewhere, but if the reverse is true, the path B will continue to be searched after the path A forever. For this reason, the same route is searched twice, which wastes time. Therefore, in the present invention, when the link L ₈ according to the path B is searched, since this is determined to be the second search, the lower the total cost of the link L _8, preferentially a link L ₈ Treat it. And rewrites the total cost of the link L ₈ to the lower value, rather than the link L _9, start the search from the link L _8. Then, the search for the route related to the high-cost route B is eventually terminated, and only the search for the route related to the low-cost route A is continued. Therefore, there is no waste when the same route is searched twice.

According to the second aspect of the present invention, since the second table is divided into two sections, "the link has already been searched once, and the total cost is small at the time of the previous search.
The process can proceed without attaching an identifier to the effect that "the table has been rewritten".

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings. As shown in FIG. 2, a navigation device that implements the route calculation method of the present invention includes a direction sensor 1
1, a distance sensor 12, a shift sensor 13, a memory drive 15 for reading stored data from a disk 14 storing road map data and route network data,
The travel distance detected by the distance sensor 12 and the travel direction change detected by the direction sensor 11 are integrated, and the vehicle position is detected based on a comparison between the integrated data and the road map data read from the memory drive 15. Locator 16 and readout of a road map in a predetermined range, calculation of an optimum route using route network data, creation of display data for guiding a vehicle, voice output device 1
8 and a controller 17 that performs various controls such as control of the locator 16, a display 19, and touch keys 20 for setting initial data and the like. Note that a remote control key may be used instead of the touch key 20.

More specifically, the direction sensor 11
Is for detecting a change in azimuth accompanying the traveling of the vehicle, and a gyro or the like can be used. The distance sensor 12 detects the traveling distance based on the speed of the vehicle, the number of rotations of the wheels, and the like, and a wheel speed sensor, a vehicle speed sensor, or the like can be used. The locator 16 calculates travel trajectory data by integrating the distance data detected by the distance sensor 12 and the azimuth change data detected by the azimuth sensor 11 respectively, and calculates the travel trajectory data and the road data stored in the disk 14. The vehicle position is detected based on a comparison with a pattern (so-called map matching method, see JP-A-64-53112). Note that a beacon receiver or a GPS receiver may be added to increase the accuracy of position detection.

The display 19 displays a menu screen on a screen such as a CRT or a liquid crystal panel, displays a road map, displays the current position and direction of the vehicle, displays an optimal route, displays destinations and points serving as landmarks, It displays the azimuth and distance to the destination, displays the window of the enlarged shape of the complex intersection, and the like. Further, a transparent touch key 20 is attached to the display 19. Touch keys 20
On the initial setting menu screen, the user inputs the criteria for selecting the optimum route (the shortest time route, the shortest distance route, etc.), the scale of the map, the destination, and the like. That is, it mediates the dialog between the route guidance device and the driver.

The disk 14 is a road map (including highway national roads, urban highways, general national roads, major local roads, general prefectural roads, general city roads in designated cities, and other living roads).
Is divided into meshes, and data obtained by combining nodes and links in each mesh unit is stored. In addition, background data for display such as railways, rivers, place name columns, famous facilities, points registered by the driver in advance, contour lines, and the like may be included.

[0023] The mesh is a map of Japan with a longitude difference of one.
A primary mesh that is divided into degrees and a latitude difference of 40 minutes and has a vertical and horizontal distance of about 80 km x 80 km, and a secondary mesh that divides this primary mesh into 8 vertical and horizontal parts and has a vertical and horizontal distance of about 10 km x 10 km And has a double structure. Here, a node is generally a coordinate position for specifying a road intersection or a bending point. A node representing an intersection is an intersection node, and a node representing a road bending point (excluding an intersection) is interpolated. Sometimes called a point node.

What connects the nodes is a link. The link data is based on the link number, the address of the start node and the end node of the link, the distance of the link, the direction of passing through the link, the required time in that direction, the road type, the road width, the traffic regulation data such as one-way and toll roads Become. As described above, since the addresses of the start node and the end node of the link are included in the link data, the point can be specified only by the link. When the start point of a link is specified by a link, the link is called an “outgoing link”, and when the end point of the link is specified, the link is called an “incoming link”. Further, since the link data includes the direction in which the vehicle passes through the link, by specifying one link, the traveling direction of the vehicle can also be specified. FIG. 3 illustrates four outgoing links specifying a crossroad, and FIG. 4 illustrates four incoming links specifying a crossroad. In the following embodiment, an outgoing link is mainly used to specify a point.

FIG. 5 shows the details of the controller 17. The controller 17 includes a memory control unit 21 that obtains necessary data from the disk 14 through the memory drive 15, a voice control unit 25 that utters voice that is electronically recorded in the voice output device, and a display that displays a required image on the display 19. A control unit 22, an input processing unit 23 that processes input information set by the touch keys 20, and a vehicle position recognition processing unit 2 that captures the vehicle position calculated by the locator 16 as data.
4. A route calculation processing unit 26 that calculates an optimum route from the destination to the departure point by the potential method.

The route calculation processing section 26 includes a link table and an arc table for searching for a link connected to each link, based on link information in a certain range including a departure place and a destination taken out of the disk 14. And calculate the optimal route based on these tables.
A PU 26a, a memory 26b for storing a state of an initial search flag (described later) indicating whether the link has been searched for the first time, and a main memory 26c for storing a link table, an arc table, and a pivot table. I have. Note that the link table and the arc table are
The pivot table corresponds to the "second table", and the pivot table corresponds to the "second table".

The link table contains a sequence number, an exit link mesh number, and a link number for a certain range of links including the departure point and the destination extracted from the disk 14.
The mesh number of the incoming link, the pointer to the arc table, the road type, the pointer to the previous link on the route of the link (hereinafter, referred to as the "previous link" or the "entering link"), the link cost, the link length, and the total cost I remember.

The sequence numbers are integers set in ascending order from 0. The link cost is a time expressed in seconds when traveling on the link. Actually, the link cost changes due to traffic congestion or the like, but here, the time required for traveling at the legal speed is used. The total cost is the cost of reaching the link in consideration of the course of the route calculation. More precisely, the total cost of a certain link is the sum of the arc costs of each link on the route from the departure value to the link. Here, as shown in FIG. 6 (a), the arc is from immediately after the start node of a link to immediately after the start node of the next link. The arc cost is, as shown in FIG. 6B, the sum of the link cost of the link and the cost of turning right or left or going straight ahead to leave the link. For example, in the case of no entry, even if the link cost is finite, the straight traveling cost is infinite, so the arc cost is also infinite.
If there is a signal, the arc cost is a cost obtained by adding the link cost taking into account the average signal waiting time when turning right or left or going straight.

Although not shown in the table, the link table has a bit indicating whether each link is registered in the pivot table. The arc table includes, for each sequence number, an arc cost, another mesh flag, an arc termination identifier indicating whether or not the arc is the last one of the arcs coming out of the same link, a pointer to a link table of a connection link connected to the arc (link table link). (Indicated by a sequence number).

Although the arc cost is as described above, for example, if traffic congestion information on the road comes in through a beacon receiver, the arc cost can be changed in consideration of the information. Also, the driver can change the cost according to his / her preference. For example, costs can be increased or decreased only for specific types of roads (such as highways and ferries).

For example, assuming that three arcs are emitted from one link, the first and second arc end identifiers are
For the third arc, the arc end identifier is assigned 0, and for the third arc, 1 is assigned. Therefore, when searching for a link connected to one link using the arc table, if there is another connection link while the arc end identifier is 0, and if the arc end identifier becomes 1, no further connection is established. It is clear that there is no link.

The pivot table is a first-in first-out table storing a sequence number of a link currently required for calculation among a large number of links stored in the link table, and includes a priority pivot table and a normal pivot table. There is. The link stored in the priority pivot table is subjected to the link search processing before the link stored in the normal pivot table.

The description will be made assuming a more specific route network. FIG. 7 shows the nodes within the range for which the route calculation including the departure point and the destination is to be performed, A, B, C, D, E,
A link route diagram is shown in which F and G are set, A is a departure point node, and F is a destination node. The total number of links is A → B, A
→ C, B → C, B → D, C → A, C → B, D → E, E →
F, F → G.

Tables 1 and 2 show the structures of the stored tables, respectively. Table 1 is a link table. For each link, a sequence number, an exit link mesh number, an entry link mesh number, a pointer to an arc table, a pointer to a road type or a previous link, a link cost, a link length, and a total cost are shown. It has a column.

[0035]

[Table 1]

Table 2 is an arc table, which includes columns for a sequence number, an arc cost, another mesh flag, an arc end identifier, and a pointer to a link table.

[0037]

[Table 2]

The numerical values entered in the link table and the arc table do not indicate the actual storage states of the link table and the arc table, but are merely entered for the sake of explaining the path calculation procedure later. I refuse that. Next, a route search method using a link table, an arc table, and a pivot table will be described. In the following calculation, a method of creating a tree from the departure point to the destination and calculating the route will be described. On the contrary, a method of creating a tree from the destination to the departure point and calculating the route may be used.

Referring to FIG. 7, it is assumed that the link A → B or the link A → C starts. It is assumed that the destination is near node F. Looking at the column of sequence number 0 corresponding to the link A → B in the link table, the pointer to the arc table indicates 0, so the column of sequence number 0 of the arc table is viewed. Pointer to link table is 2
It has become. Therefore, looking at the column of sequence number 2 in the link table, the link with sequence number 2 is a link B → C specified by the start node B and the end node C.
It is. Therefore, it is understood that the link B → C is connected to the link A → B. Since the arc end identifier of the arc table is 0, another arc following the link A → B exists. Then, the user goes down the arc table and looks at the column of sequence number 1. Since the pointer to the link table is 3, the link B → D is seen in the column of the sequence number 3 in the link table. Therefore, it can be seen that the link B → D is also connected to the link A → B. Here, since the arc end identifier is 1, the search for the link connecting the link A → B is completed.

Next, a search is made for a link connected to the link B → C. Looking at the column of sequence number 2 in the link table, the pointer to the arc table is 3.
Therefore, looking at the column of sequence number 3 in the arc table, the link is link C → A. Since the arc end identifier in the column of the sequence number 3 in the arc table is 1, the arc connected to the link B → C is only C → A.

Next, a search is made for a link connected to the link B → D. Looking at the column of sequence number 3 in the link table, the pointer to the arc table is 4.
Then, looking at the column of sequence number 4 in the arc table, the link is link D → E. Since the arc end identifier in the sequence number 4 column of the arc table is 1, there is no other connection link.

As described above, connection links are obtained one after another like a tree. The obtained tree is as shown in FIG. FIG. 8A shows a tree extending from link A → B, and FIG. 8B shows a tree extending from link A → C. Although the tree extends indefinitely, the calculation is terminated in the middle of the tree by modifying the processing as described later.

The method of successively searching for links as described above has been described. In the route calculation processing, in addition to searching for a route, arc costs are integrated, and an optimal route having the least cost is obtained. Needs to be calculated. Therefore, the method of calculating the optimum route will be described in detail with reference to a flowchart (FIGS. 9-11).

First, the CPU 26a retrieves a certain range of link data including the departure place and the destination from the disk 14, and creates a link table based on a predetermined program. At this time, if a certain range including the departure place and the destination includes a plurality of meshes, the mesh numbers are renumbered and stored. This facilitates map management and enables high-speed access.

When the link table is created, the total cost value of the link table is all infinite (actually, it is sufficient to set the total cost to about two orders of magnitude larger than the expected total cost), and there is no previous link. Initialize.
Hereinafter, assuming the route network of FIG. 7, the optimal route is calculated by setting the departure point links (referred to as calculation start links) to A → B and A → C, and the destination links (referred to as calculation end links) to F → G.

The CPU 26a takes out the calculation start links A → B and A → C, and registers these calculation start links A → B and A → C in the ordinary pivot table (step S1). The link registered in the pivot table is hereinafter referred to as “pivot link”. The total cost corresponding to the pivot link is set to 0 on the link table, and the initial search flag in the memory 26b is turned off (step S2).

The CPU 26a refers to the priority pivot table and determines whether there is an unprocessed pivot link (step S3). If not, it is determined whether there is an unprocessed pivot link by referring to the ordinary pivot table (step S4). Since no processing is performed at first, an unprocessed pivot link, for example, a calculation start link A
→ B usually exists in the pivot table. This is taken out as an unprocessed pivot link (step S6).

The CPU 26a refers to the link table corresponding to the link A → B in order to search for a link connecting the link A → B, and acquires a pointer to the arc table and an arc cost. As a result, as described above, the link B → C is found in the arc table. Then, it is determined whether the search for all arcs has been completed (step S7). If the search has not been completed, the arc cost of link B → C is extracted. Arc cost is link A → B
Is the link cost 40 plus the right turn cost 5 at the intersection B. Further, the arc cost 45 is added to the total cost of the link A → B (it is 0), and this is set as the total cost 45 of the link B → C (step S8). It is determined whether or not this total cost is smaller than the total cost of the connection links B → C written in the link table (its initial value is infinite) (step S9).

At first, since the size is always small, FIG.
It is determined whether the connection link B → C has been searched for the first time (step S11). This determination can be easily understood from the fact that the total cost of the link B → C is infinite. If it has been searched for the first time, the CPU 26a turns on the above-described first search flag (step S12). Then, the total cost of the connection link B → C is rewritten on the link table (step S13).

Next, it is checked whether the link B → C exists in the pivot table (step S14). In this check, it takes time to check each link of the pivot table one by one, so the check is performed based on the predetermined bits on the link table described above. If it does not exist in the pivot table, link B → C is registered. The detailed procedure of this registration is as shown in FIG. First, it is determined whether or not the above-described initial search flag is turned on (step S15a).
That is, if it is the first search, it is registered in the low-priority ordinary pivot table (step S15b). If the first search flag is off, that is, if the search is the second or subsequent search, it is registered in the priority pivot table with a high priority (step S15c).

Returning to FIG. 10, the first search flag is turned off (step S16), and a pointer to the previous link, link A → B, is recorded on the link table (step S17). Then, returning to step S7, in order to search for another link connecting the link A → B,
Refer to the column of sequence number 1 in the arc table. As a result, the link B → D is found. The arc cost is the sum of the link cost 40 of the link A → B and the straight-ahead cost 2 at the intersection B (the straight-ahead cost is not 0 because the signal waiting at the intersection is considered).
Then, the arc cost 42 is added to the total cost of the links A → B (which is 0) to obtain the total cost of the links B → D. And this total cost 42
Is determined to be smaller than the total cost of the link B → D written in the link table (its initial value is infinite) (step S9).

If it is smaller, the process proceeds to step S11 in FIG. 10, and it is determined whether the connection link B → D has been searched for the first time (step S11). This judgment is made for link B →
It is easy to see if the total cost of D is infinite. If it is the first search, CPU
26a turns on the above-described initial search flag (step S12). Then, the total cost of the connection link B → D is rewritten on the link table to be 42 (step S13). Next, it is determined whether the link B → D exists in the pivot table (step S14). If the link B → D does not exist in the pivot table, the link B → D is registered (step S15).

The detailed procedure of this registration is the same as that described above with reference to the flowchart of FIG. Then, the link A → B is set as the previous link of the connection link B → D (step S17). Thereafter, the process returns to step S7, and when the arc end identifier in the second row of the arc table corresponding to the link A → B is referred to as 1, the link A →
Since there is no other link leading to B, it can be seen that the search for all arcs has been completed. Proceeding to step S3, reference is made to the priority and normal pivot tables to determine whether there is an unprocessed pivot link (step S3).
4).

At this stage, the calculation start links A → C, B → C and B → D are registered as unprocessed in the ordinary pivot table. Then, the unprocessed pivot link A → C is taken out and the same processing as above is performed. That is, the link C → B which is a connection link of the link A → C is searched, and the total cost (0) of the pivot link A → C is added to the arc cost (the link cost 15 of the link A → C to the node C). Left turn cost at 2 is added)
Add. Then, this is set as the total cost of the link C → B, the total cost of the link C → B on the link table is rewritten to 17, and the link C → B is registered in the ordinary pivot table.

Next, processing is performed for the link B → C. That is, the link C → A which is a connection link of the link B → C is searched, and the total cost of the link B → C (4
5), the arc cost of the link C → A (the link cost 10 of the link B → C plus the right turn cost 5 at the node C) is added. 45 + 10 + 5 = 60 This is the new total cost of the link C → A on the link table. After searching for the link C → A, if a link connected to the link C → A is picked up from the arc table, the link returns to the calculation start link A → B, and there is a possibility that the search becomes infinite and the search continues indefinitely. However, in this case, in step S9, the new total cost of the link A → B is compared with a value registered as the total cost of the link A → B (initial value = 0).
The new total cost is (new total cost)> (previous total cost =
0) is established, and the process returns to step S7. Therefore, there is no possibility of falling into an infinite loop.

Next, regarding the link B → D, the link B →
A search is made for the link D → E leading to D, and the total cost is rewritten to 62 and registered in the pivot table. At this time, since the link D → E is the first search, the link D → E is registered in the low-priority ordinary pivot table (step S15b). Next, with respect to the link C → B, a search is made for a link B → D that is connected to the link C → B, and an arc cost 20 thereof is obtained. (Step S9). Since the total cost is small, the total cost is updated to the smaller value 37 (step S13). The link B → D is registered in the pivot table (step S15). At this time, the link B → D is not the link searched for the first time because the previous total cost was the finite value 42. Therefore, the first search flag is turned on in step S12. Then, the link B → D is registered in the priority pivot table having the higher priority (step S15c).

Therefore, in the next search, step S
In 3, the link B → D is extracted from the priority pivot table having the higher priority, and the search processing of the connection link is continued. That is, the link D → E following the link B → D is searched, and the total cost (37 + 20) is compared with the value 62 of the link table. Since the total cost of the link D → E is small, the total cost of the connection link D → E is rewritten on the link table. Then, a link B → D is set as a link preceding the connection link D → E (step S11). Furthermore, this connection link D
→ Since E is the second search, it is registered in the priority pivot table (step S12).

Next, in step S3, a link D → E is extracted from the normal pivot table having a high priority. Then, the link E → F following the link D → E is searched, and the total cost 67 is written in the link table. Next, the link D → E is extracted from the ordinary pivot table having the lower priority. Then, the link E → F following the link D → E is searched, and the total cost is compared with the value of the link table. In this case, link E →
Since the total cost of F is the same as the previous value, the process returns from step S9 to step S7, and the search is terminated.

Therefore, the links A → C, C → B, B →
The total cost of the link E → F continuing from D, D → E is determined, and only the connection link F → G exiting from the link E → F is searched. As described above, the reason why the search can be aborted is that the link B → D and the link D → E to be connected therefrom are registered in the priority pivot table.

Because, if there is only one pivot table, link B → link connected from link C → B
In the link B → D, which is connected from D to the link A → B, the latter is searched first and the cost of the latter is high. Therefore, if a link with a low total cost is searched later, all the results searched earlier are useless. Become. Therefore, if the total cost is low as in the embodiment, if the result of the second search is registered in the priority pivot table, the next search for the same link is immediately terminated because the total cost is high or equal. be able to. Thereafter, the search can be continued for only the route with the low total cost.

Finally, it is determined whether there is any unprocessed pivot link via steps S3 and S4. When all the pivot links have been processed, the route search has been completed. As described above, finally, the route A → C, C → B, B → D, D → E, E → F, F by tracing the link preceding the calculation end link F → G to the destination →
G is determined as the optimal route.

Referring to the total cost column of the calculation end link F → G, the total cost 77 is the optimum route cost from the departure point to the destination. By the method of the present invention as described above, Osaka-Tokyo (the number of links is 2.3 × 10 ⁴ )
The optimal route was calculated. In the conventional potential method without a priority table, the number of loops is about 3 × 10 ⁵ because the link once searched is searched again. On the other hand, in the potential method of the present invention described in the embodiment, The number of loops was reduced to about 4.8 × 10 ⁴ . As a result, the net calculation time was reduced from about 6 minutes 30 seconds to about 1 minute.

[0063]

As described above, according to the route calculation method and apparatus of the present invention, the connection link to be searched is a link that has already been searched once, and the total cost during the previous search is small and the first connection is small. If the link has been rewritten in the table, the connection link of the link is searched with priority, so that the possibility of double search for the same route is reduced. Therefore, the calculation processing time can be shortened.

[Brief description of the drawings]

FIG. 1 is a link configuration diagram for explaining a method of calculating an optimum route according to the present invention.

FIG. 2 is a block diagram showing a navigation device that executes the route calculation method of the present invention.

FIG. 3 is a diagram showing an example of an outgoing link at a crossroads.

FIG. 4 is a diagram showing an example of an incoming link at a crossroad.

FIG. 5 is a detailed configuration diagram of a controller.

FIG. 6 is a simple link configuration diagram for explaining total cost.

FIG. 7 is a link configuration diagram for which route calculation including a departure point and a destination is performed.

FIG. 8 is a diagram showing a tree of links obtained by successively obtaining connection links.

FIG. 9 is a flowchart illustrating a method of calculating an optimum route.

FIG. 10 is a flowchart (continued from FIG. 9) for explaining a method of calculating an optimum route.

FIG. 11 is a flowchart showing a part of the flowchart (FIG. 10) in detail.

[Explanation of symbols]

 14 disk 17 controller 26 route calculation processing unit 26a CPU 26b first search flag 26c main memory

────────────────────────────────────────────────── ─── Continuing on the front page (72) Kazuo Hirano, 1-3-1 Shimaya, Konohana-ku, Osaka-shi Sumitomo Electric Industries, Ltd. Osaka Works (56) References JP-A-4-204881 (JP, A) JP-A-2-201600 (JP, A) JP-A-2-184998 (JP, A) JP-A-4-372985 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G01C 21 / 00 G08G 1/0969 G09B 29/10

Claims (3)

(57) [Claims] 1. A path network memory storing costs of links constituting a path network and connection relations between the links, one or more connection links connected next to the link for each link, and A first table having a column for entering a link on a calculated route connected before the link, a total cost of the calculated route including the link, and a link for searching for a connection link is entered. Using a second table having columns, (a) reading a predetermined range of route network data from the route network memory according to the setting of the departure point and the destination, and (b) configuring the read route network The initial value of the total cost of each link to be set is set to a sufficiently large value and written in the corresponding column of the first table. Writing the table, the total cost column of the first table corresponding to the link written in the second table is set to an initial value, the links that are filled in (d) of the second table, the first A connection link is searched for by referring to the table, the link cost of the searched connection link is added to the cost of the link entered in the second table, and the added value is added to the first cost.
(E) As a result of the comparison, if the added value is large, the processing for the connected link is terminated. If the added value is small, the value is added to the first table. Rewrite the total cost of the connection link and write the link entered in the second table as a link on the calculated route to this connection link, and (f) store the connection link in the second table. A new entry is made, and the above processes (d) and (e) are repeated. (G) When the search for the connection link is completed for all the links entered in the second table, the calculation end link is changed to the first link. In the route calculation method for determining an optimal route by searching from the table of the above and sequentially tracing the links on the calculated route to the link, In step (d), a link to be searched for a connection link is already searched, and as a result of the comparison in step (e), the added value is small and rewriting is performed in the first table. Link
In preference to the link being searched for the first time,
(D) A route calculation method characterized by performing the following processing. 2. The method according to claim 1, wherein said second table is divided into two sections.
The link searched for the first time in (d) is stored in a non-priority part, and the link searched in the procedure (d) for the second time and thereafter and the link rewritten in the first table last time is stored in the priority part. The route calculation method according to claim 1, wherein 3. A route network memory storing costs of links constituting the route network and connection relationships between the links, one or more connection links connected next to the link for each link, and A first table having a column for entering a link on a calculated route connected before the link, a total cost of the calculated route including the link, and a link for searching for a connection link is entered. A route comprising: a second table having columns; and route calculation processing means for calculating an optimum route between two points on a road map by executing the procedures (a) to (g) according to claim 1. In the computing device, the route calculation processing means may be configured such that, in the step (d), a link to be searched for a connection link has already been searched in the step (d), and a result of the comparison in the step (e) at that time If the added value is small and the link has been rewritten in the first table, the link performs the following processing (d) in preference to the link searched for the first time. Path calculation device.