KR20230063199A - Turn-based optimal path search method and system - Google Patents

Abstract

An optimal path finding system, which reads node data consisting of node IDs and coordinates, link data consisting of link IDs, departure node IDs, and arrival node IDs, and rotation constraint data containing information about rotatability from a server to find the optimal path, comprises: a rotation-based rotation data construction unit which generates the rotation movable from a link included in the link data to an adjacent link, and constructs a rotation-based rotation data structure including rotation IDs, departure link IDs, arrival link IDs, or travel times of the generated rotation; a travel time extraction unit which extracts travel times corresponding to the rotation IDs from rotation data stored on the server and stores the travel times in the rotation-based rotation data constructed by the rotation-based rotation data construction unit; and a rotation-based path finding unit which extracts the optimal path based on the rotation-based rotation data constructed by the rotation-based rotation data construction unit. Therefore, the optimal path finding system can extract the optimal path considering travel times.

Description

Turn-based optimal path search method and system {TURN-BASED OPTIMAL PATH SEARCH METHOD AND SYSTEM}

The present invention relates to an optimal route search method and system, and more particularly, to a rotation-based optimal route search method and system that searches for an optimal route in consideration of rotation time so that it can be used in the field of vehicle navigation and the field of providing traffic information.

With the development of LTE and 5G mobile communication technologies, the number of users of mobile-based navigation and in-vehicle navigation continues to increase. However, current navigation systems often provide incorrect routes in complex networks as link-based travel time is applied due to a technical standard established 10 years ago.

At intersections, directions are divided into straight ahead, left turn, and right turn, and traffic characteristics may appear differently due to differences in signal time and traffic volume for each direction. Traffic information obtained by dividing and aggregating these different characteristics by direction is referred to as rotational traffic information. Existing optimal route search algorithms do not reflect turn traffic information of urban intersections as they are made based on links, and thus do not consider left turn and straight forward asymmetric travel times. can

1 is a conceptual diagram of an optimal path search method, FIG. 1 (a) shows a conceptual diagram showing network information provided from a server, and FIG. 1 (b) shows a conceptual diagram of a link-based path search method. 1(c) shows a conceptual diagram of the rotation-based route search method.

Referring to (a) of FIG. 1, traffic information on general urban roads may be provided as link travel time, which is the time required to move from node to node, and turn travel time, which is the time required to move from link to link. there is. Here, the turn travel time means the travel time for each direction of a turn divided into a left turn, a right turn, or going straight at an intersection.

Referring to (b) of FIG. 1, as a concept of the existing link-based path search, when searching for an optimal path from a starting point to a destination, it can be calculated using a link travel time. Using the link travel time, a route that takes 396.7 seconds can be extracted as an optimal route.

Referring to (c) of FIG. 1, as a concept for a turn-based route search, it can be calculated using a turn travel time when searching for an optimal route. Using the turnaround time, a route that takes 380 seconds can be extracted as an optimal route. In this way, if the rotation travel time is considered, an optimal route that can move faster than in the case of using the link travel time can be extracted. A path extracted by considering only the link travel time without considering rotation may be a false path. Due to this misroute, left-turn traffic at a specific intersection may increase compared to the existing pattern, resulting in side effects such as increased congestion.

If travel time per turn is generated in the future, it is necessary to change the technical standard of the navigation system. The optimal path search technology, which is the base technology of the navigation system, also requires a change from a link-based to a rotation-based one, and accordingly, the establishment of the concept of a rotation-based network structure is required.

Functional improvement of search technology is also required to solve the increase in search time due to the increase in network size. In the case of existing algorithms, one data structure that is most preferred by developers, such as heap or bucket, is used as the data structure. When data is built on a rotation basis, the data size increases compared to link-based data structures, so search time increases when a single data structure is used. will do

Establish rotation constraints or rotation permitting data to implement rotation prohibition, U-Turn, or P-Turn in the existing path search algorithm, and calculate rotation constraints or rotation allowances when the model is configured to achieve realistic path search by reflecting it to the algorithm. The search time may increase as a separate operation is performed to determine whether the node should do this.

As described above, since the existing link-based optimal route search does not consider the round trip time, a wrong route can be extracted, but such a critical mind has not been presented. In addition, a data structure and algorithm for reducing the search time are required due to the increase in data size when the round trip time is considered.

An object of the present invention is to provide an optimal path search method and system capable of extracting an optimal path considering rotation time by constructing rotation-based data.

Another object of the present invention is to improve algorithm performance using a hybrid data structure.

In the present invention, node data composed of a node ID and coordinates, link data composed of a link ID, a starting node ID, and an arrival node ID, or rotation constraint data including information on rotation possibility are read from a server and an optimal path is searched. In an optimal path search system for generating a rotation that can move from a link included in the link data to an adjacent link, and generating a rotation ID, a departure link ID, an arrival link ID, or rotation-based rotation data including a travel time of the generated rotation a rotation-based rotation data construction unit for constructing a structure; a travel time extraction unit extracting travel time corresponding to a rotation ID from rotation data stored in the server and storing the travel time in the rotation-based rotation data constructed by the rotation-based rotation data construction unit; It is characterized in that it includes a rotation-based path search unit for extracting an optimal path based on the rotation-based rotation data built by the rotation-based rotation data construction unit.

Preferably, when the rotation data stored in the server is link-based rotation data including a departure link ID, an arrival link ID, and a travel time, the travel time extractor converts the converted rotation-based rotation data into rotation-based rotation data. The travel time corresponding to the rotation ID can be extracted from the data.

Preferably, when the rotation data stored in the server is node-based rotation data composed of a departure node ID, an intermediate node ID, an arrival node ID, and a travel time, the travel time extraction unit corresponds to the departure node ID and the intermediate node ID. The link ID is stored as the departure link ID, and the link ID corresponding to the intermediate node ID and arrival node ID is stored as the destination link ID, and converted into link-based rotation data including the departure link ID, arrival link ID, and travel time, , the converted link-based rotation data can be converted into rotation-based rotation data, and the travel time corresponding to the rotation ID can be extracted from the converted rotation-based rotation data.

Preferably, the rotation-based rotation data construction unit deletes the rotation prohibited from the generated rotation if the rotation constraint data is rotation prohibition data in the link with rotation constraint data, and generates a rotation that makes a U-turn if the rotation constraint data is U-turn permission data. In a link without rotation constraint data, it is possible to create a rotation that can move to an adjacent link, except for a rotation that makes a U-turn.

Preferably, the rotation-based path search unit includes: a first labeling module for searching for a rotatable adjacent link in a link directly connected to the starting node and assigning a label including a previous link number and travel time; a marker module for searching for a link having the shortest travel time among labeled links when there is no label with a marker, and displaying a marker on the label of the searched link; a second labeling module for searching for a rotatable adjacent link among labeled links having a mark and assigning a label thereto; a mark change module for searching for a link with the shortest travel time among labeled links excluding labels with a mark, displaying a mark on the label of the searched link, and deleting the label with the previous mark; An optimal path extraction module for extracting an optimal path by connecting previous link numbers included in the label of a link directly connected to the destination node in reverse order may be included.

Preferably, the rotation-based route search unit may further include a label removal module that removes labels other than a label having the smallest travel time when a plurality of labels are assigned to one link.

Preferably, the rotation-based path search unit may search for an optimal path using a double bucket data structure constituting a lower bucket inside an upper bucket.

In addition, the present invention reads node data composed of a node ID and coordinates, link data composed of a link ID, a departure node ID, and an arrival node ID, or rotation constraint data including information on rotation possibility from a server to determine an optimal path. In an optimal path search method for searching, a rotation-based rotation including a rotation ID, a departure link ID, an arrival link ID, or travel time of the generated rotation is generated by generating a rotation capable of moving from a link included in the link data to an adjacent link. Rotation-based rotation data construction step of building a data structure; a travel time extraction step of extracting a travel time corresponding to a rotation ID from rotation data stored in the server and storing the travel time in the rotation-based rotation data constructed in the rotation-based rotation data construction step; Another feature is a rotation-based path search step of extracting an optimal path based on the rotation-based rotation data constructed in the rotation-based rotation data construction step.

Preferably, the rotation-based path search step includes: searching for an adjacent rotatable link in a link directly connected to the starting node, and assigning a label including a previous link number and travel time; If there is no label with a mark, searching for a link having the shortest travel time among the labeled links, and displaying a mark on the label of the searched link; searching for adjacent rotatable links in the labeled links having a marker and assigning labels thereto; searching for a link with the shortest travel time among labeled links excluding labels with a mark, displaying a mark on a label of the searched link, and deleting a label with a previous mark; and extracting an optimal path by connecting previous link numbers included in labels of links directly connected to the destination node in reverse order.

According to the present invention, there is an advantage in that the rotation-based rotation data building unit can extract an optimal path considering rotation time by constructing rotation data based on rotation in order to search for an optimal path based on rotation.

In addition, the present invention has an advantage that the rotation-based route search unit can reduce route search time by using a bucket & heap hybrid data structure.

1 is a conceptual diagram of an optimal path search method, FIG. 1 (a) shows a conceptual diagram showing network information provided from a server, and FIG. 1 (b) shows a conceptual diagram of a link-based path search method. 1(c) shows a conceptual diagram of the rotation-based route search method.
2 shows a configuration diagram of an optimal path search system according to an embodiment of the present invention.
3 shows an example network of nodes, links and rotations according to an embodiment of the present invention.
4 shows the structure of node data and link data according to an embodiment of the present invention.
5 is a structure of rotation constraint data, FIG. 5(a) shows node-based rotation constraint data, and FIG. 5(b) shows link-based rotation constraint data.
6 shows the structure of rotation-based rotation data according to an embodiment of the present invention.
7 shows a data conversion process of a rotation-based rotation data conversion unit according to an embodiment of the present invention.
8 shows a configuration diagram of a rotation-based path search unit according to an embodiment of the present invention.
9A to 9D show an algorithm development process performed by a rotation-based path search unit according to an embodiment of the present invention.
10 shows a bucket & heap hybrid data structure used by a rotation-based path search unit according to an embodiment of the present invention.
11 shows a flow chart of a path search algorithm of an optimal path search system according to an embodiment of the present invention.

Hereinafter, the present invention will be described in detail with reference to the contents described in the accompanying drawings. However, the present invention is not limited or limited by exemplary embodiments. The same reference numerals in each figure indicate members performing substantially the same function.

The objects and effects of the present invention can be naturally understood or more clearly understood by the following description, and the objects and effects of the present invention are not limited only by the following description. In addition, in describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted.

2 shows a configuration diagram of an optimal path search system 1 according to an embodiment of the present invention. Referring to FIG. 2 , the optimal path search system 1 may include a rotation-based rotation data construction unit 100 , a travel time extraction unit 300 , and a rotation-based path search unit 500 . The optimal route search system 1 can be used in the field of vehicle navigation and the field of providing traffic information provided by traffic information centers and portals (ATIS: Advanced Transportation Information System). In addition, the optimal route search system 1 can be utilized in the field of traffic S/W for traffic planning and traffic operation.

Optimal route search system 1 according to this embodiment is node data composed of node ID and coordinates, link data composed of link ID, departure node ID, and destination node ID, or rotation constraint data including information on rotation possibility. can be read from the server 3 and an optimal path can be searched. The server 3 may be understood as an Internet environment for receiving information such as turnaround time collected through a smartphone, navigation, CCTV, or probe vehicle.

The optimal route search system 1 may form a network structure capable of performing turn-based route search so that actual traffic conditions may be reflected, and may have a data structure as a hybrid structure for quick route search. The optimal route search system 1 can perform turn-based route search that can consider time for left turn, right turn, or straight ahead.

Rotation information can be organized into data structures in three ways. First, there is an enumeration of node IDs on the path that proceeds in the way used in the existing rotation constraint and rotation permission. The second has a link ID enumeration formula, and the third has a rotation ID construction method.

3 shows an example network of nodes, links and rotations according to an embodiment of the present invention. Referring to FIG. 3, the node enumerated expression of starting from node 2 and arriving at node 3 by going straight is (2, 1, 3). Departing from node 2 and arriving at node 3 in a straight line is expressed as a link list formula (1111, 1113). Departing from node 2 and arriving at node 3 in a straight line is expressed by the rotation ID construction method as (x0011). Departing from node 2 and arriving at node 4 with a left turn is expressed as a node sequence expression (2, 1, 4). Departing from node 2 and arriving at node 4 by taking a left turn is expressed as a link list formula (1111, 1116). Departing from node 2 and arriving at node 4 with a left turn is expressed as a turn ID construction method (x0012). As such, rotation can organize data in three ways. The most excellent method is the rotation ID construction method, but since it is not standardized in ITS (Intelligent Transport System) traffic information exchange, the link listing method can be the most suitable method. Therefore, the optimal path search system 1 is a technology of constructing a rotation data structure using node or link data, a technology of converting link-listed rotation travel time into rotation-based rotation data, or a technology of searching for a path based on rotation. can be configured.

The rotation-based rotation data construction unit 100 generates a rotation that can move from a link included in the link data to an adjacent link, and includes a rotation ID, a departure link ID, an arrival link ID, or a travel time of the generated rotation. You can build data structures. The rotation-based rotation data construction unit 100 may perform a rotation-based function of constructing a rotation data structure for the optimal path search system 1 to search for an optimal path based on rotation. In order to explain the data structure built by the rotation-based rotation data construction unit 100, the structures of nodes, links, and rotation constraint data will be described in FIGS. 4 and 5.

4 shows the structure of node data and link data according to an embodiment of the present invention. Referring to (a) of FIG. 4 , node data is composed of a node ID and coordinates. Each node can be identified by a node ID. Each node ID may be specified by an x-coordinate value and a y-coordinate value. Referring to (b) of FIG. 4, link data may include a link ID, a source node ID (FNode ID), and a destination node ID (TNode ID). Each link may be identified by a link ID. Each link ID may be specified by a source node ID and a destination node ID.

Node data can take the node ID as a key. Link data can take the starting node ID and destination node ID as keys within the scope of the node ID in the node data. Link data can be configured to use link ID as a key.

5 is a structure of rotation constraint data, FIG. 5(a) shows node-based rotation constraint data, and FIG. 5(b) shows link-based rotation constraint data.

Referring to (a) of FIG. 5 , node-based rotation constraint data may include a starting node ID, an intermediate node ID, an arrival node ID, or a rotation constraint. In the node-based rotation constraint data, a starting node ID, an intermediate node ID, and an arrival node ID may specify a rotation. In node-based rotation constraint data, rotation constraints may indicate whether a specified rotation is rotatable or non-rotatable.

Referring to (b) of FIG. 5, the link-based rotation constraint data may include a starting link ID, an arrival link ID, or a rotation constraint. In the link-based rotation constraint data, the originating link ID and the destination link ID may specify rotation. In link-based rotation constraint data, rotation constraints may indicate whether a specified rotation is rotatable or non-rotatable.

6 shows the structure of rotation-based rotation data according to an embodiment of the present invention. Referring to FIG. 6 , the rotation-based rotation data constructed by the rotation-based rotation data construction unit 100 may include a rotation ID, a departure link ID, an arrival link ID, or travel time. In rotation-based rotation data, each rotation can be distinguished by a rotation ID. Rotation-based rotation data may specify each rotation with an ID of a starting link where rotation starts and an ID of an arrival link where rotation ends. In rotation-based rotation data, travel time may represent the time required to rotate.

An embodiment in which the rotation-based rotation data construction unit 100 constructs rotation-based rotation data will be described. The rotation-based rotation data construction unit 100 may read node data from the server 3 and form a key using a node ID. The rotation-based rotation data construction unit 100 may read link data from the server 3 and form a key using the departure node ID and destination node ID. The rotation-based rotation data construction unit 100 may read link data and form a key using a link ID. The rotation-based rotation data construction unit 100 may read rotation constraint data and input an indication that rotation constraints exist in a link where rotation starts. The rotation restriction data read by the rotation-based rotation data construction unit 100 may include data such as prohibiting left turn, prohibiting entry, or permitting a U-turn.

The rotation-based rotation data construction unit 100 may generate a rotation movable from a link to an adjacent link in order to build a rotation-based rotation data structure. Steps for the rotation-based rotation data construction unit 100 to generate rotation are as follows. The rotation-based rotation data construction unit 100 may extract all links starting with the arrival node ID by extracting the arrival node ID of the departure link ID. The rotation-based rotation data construction unit 100 may generate a rotation using the extracted link as the destination link ID.

The rotation-based rotation data construction unit 100 deletes the prohibited rotation from the generated rotation if the rotation constraint data is rotation prohibition data in the link with rotation constraint data, and generates a rotation that makes a U-turn if the rotation constraint data is U-turn permission data. can The rotation-based rotation data construction unit 100 may generate a rotation that is movable to an adjacent link except for a rotation that makes a U-turn in a link without rotation constraint data. The rotation-based rotation data construction unit 100 may generate a key using the starting link ID of the generated rotation. In this way, the rotation-based rotation data construction unit 100 can enable a realistic optimal path search by reflecting rotation constraints.

The travel time extraction unit 300 may extract the travel time corresponding to the rotation ID from the rotation data stored in the server and store it in the rotation-based rotation data constructed by the rotation-based rotation data construction unit 100 . The travel time extractor 300 may extract rotational travel time, which is the time required to move from link to link, from traffic information collected in real time by the server 3 . The travel time extraction unit 300 may extract the rotation travel time and store it in the travel time of the rotation ID where the departure link and the arrival link coincide. When the traffic information stored in the server 3 is node-based or link-based, the travel time extractor 300 may convert it into rotation-based data so that it can be stored in the rotation data constructed by the rotation-based rotation data construction unit 100.

The travel time extractor 300 may play a role of converting the rotation travel time generated through a system such as a navigation into rotation-based rotation data generated by the rotation-based rotation data construction unit 100 . Since the existing traffic information format is configured based on links or nodes, the travel time extractor 300 can convert the traffic information format into rotation-based rotation data so as to be matched with a rotation ID.

When the rotation data stored in the server 3 is link-based rotation data including a departure link ID, an arrival link ID, and a travel time, the travel time extractor 300 may convert it into rotation-based rotation data. The travel time extractor 300 may convert link-based rotation data into rotation-based rotation data so that the traffic information stored in the server 3 can be stored in the rotation data constructed by the rotation-based rotation data construction unit 100 . The travel time extractor 300 may extract a travel time corresponding to a rotation ID from rotation-based rotation data. Since the link-based rotation data is composed of a departure link ID and an arrival link ID, the travel time extractor 300 extracts the rotation ID using the departure link ID and the arrival link ID as keys, and stores the travel time in a corresponding field.

When the rotation data stored in the server 3 is node-based rotation data composed of a departure node ID, an intermediate node ID, an arrival node ID, and a travel time, the travel time extractor 300 responds to the departure node ID and the intermediate node ID. Stores the corresponding link ID as the departure link ID and stores the link ID corresponding to the intermediate node ID and arrival node ID as the destination link ID, and converts it into link-based rotation data including the departure link ID, arrival link ID, and travel time And, the converted link-based rotation data can be converted into rotation-based rotation data. The travel time extractor 300 may convert node-based rotation data into link-based rotation data so that traffic information stored in the server 3 can be stored in rotation data constructed by the rotation-based rotation data construction unit 100 . The travel time extractor 300 may convert the converted link-based rotation data into rotation-based rotation data once again.

Since the node-based rotation data is composed of a starting node ID, an intermediate node ID, and an arrival node ID, the travel time extractor 300 extracts the corresponding link ID using the starting node ID and the intermediate node ID as keys and stores the corresponding link ID as the starting link ID. can The travel time extraction unit 300 may extract a corresponding link ID using the intermediate node ID and the arrival node ID as keys and store the corresponding link ID as the arrival link ID. Since the link-based rotation data is composed of a departure link ID and an arrival link ID, the travel time extractor 300 extracts the rotation ID using the departure link ID and the arrival link ID as keys, and stores the travel time in a corresponding field.

7 shows a data conversion process of the travel time extractor 300 according to an embodiment of the present invention. Referring to FIG. 7 , the travel time extractor 300 converts node-based rotation data stored in the server 3 into rotation-based rotation data to extract the travel time of rotation 1 moving from link 1 to link 3. The travel time extraction unit 300 may extract link 1 using the start node 1 and the intermediate node 2 as keys from the node-based rotation data and store it as the start link 1. The travel time extractor 300 may extract link 3 using intermediate node 2 and arrival node 3 as keys from node-based rotation data and store it as arrival link 3. Through this process, the travel time extractor 300 transfers node-based rotation data composed of departure node 1, intermediate node 2, arrival node 3, and travel time 0.15 to a link composed of departure link 1, arrival link 3, and travel time 0.15. It can be converted into base rotation data.

The travel time extractor 300 may extract rotation 1 using the departure link 1 and arrival link 3 as keys from the converted link-based rotation data, and store 0.15 in the travel time of the corresponding field. Through this process, the travel time extractor 300 converts link-based rotation data composed of departure link 1, arrival link 3, and travel time 0.15 into rotation-based rotation data composed of rotation 1, departure link 1, arrival link 3, and travel time 0.15. It can be converted into rotation data.

8 shows a configuration diagram of a rotation-based path search unit 500 according to an embodiment of the present invention. Referring to FIG. 8 , the rotation-based path search unit 500 includes a first labeling module 510, a mark module 520, a second labeling module 530, a mark change module 540, and an optimal path extraction module 550. ), and a label removal module 560.

The rotation-based path search unit 500 may extract an optimal path based on the rotation-based rotation data constructed by the rotation-based rotation data construction unit 100 . The rotation-based path search unit 500 extracts an optimal path based on rotation-based rotation data in which rotation travel times such as a left turn, a right turn, a U-turn, or a pit-turn are taken into account, so that the search for a wrong path can be minimized. A path search method of the rotation-based path search unit 500 will be described through the embodiments shown in FIGS. 9A to 9D.

9A to 9D show an algorithm development process performed by the rotation-based path search unit 500 according to an embodiment of the present invention. 9A to 9D, circles represent nodes, solid lines represent links, square boxes represent link IDs, and dotted lines represent rotations. The number next to the dotted line means the rotation time, and the (x,y) next to the box means the label. Here, x of the label is the previous link number, and y is the time until reaching the corresponding link.

Referring to FIG. 9A , the rotation-based path search unit 500 may use a label marking technique based on the Dijstra algorithm, which is a path search technique. The rotation-based route search unit 500 may perform route search by displaying a label on a link, unlike conventional route search in which a label is displayed on a node. The biggest difference from the existing link-based algorithm is that the rotation-based path search unit 500 displays labels on links. Hereinafter, rotation-based label display will be described with starting from node 1 as an example.

First, since links directly connected to node 1 are link 1 and link 2, the rotation-based path search unit 500 may search for rotation based on link 1 and link 2. The rotation-based path search unit 500 may connect links 9 and 11 when a rotation is searched for based on link 1, and links 4 and 5 may be connected when a rotation is searched for based on link 2. At this time, the rotation-based path search unit 500 labels (1,1.5) for link 9, (1,1.2) for link 11, (2,1.3) for link 4, and (2,0.7) for link 5. can be attached This function of the rotation-based path search unit 500 may be performed by the first labeling module 510 .

Since the travel time from the first search to link 5 is the shortest, the turn-based route search unit 500 may assign * to the label of link 5. This function of the rotation-based path search unit 500 may be performed by the mark module 520 .

Referring to FIG. 9B , the rotation-based path search unit 500 may search for a link connected to link 5 after link 5 is determined. The rotation-based path search unit 500 may search for links 13, 14, and 15 connected to link 5 and label them as (5, 1.2), (2, 2.9), and (5, 2.0). This function of the rotation-based path search unit 500 may be performed by the second labeling module 530 .

The rotation-based route search unit 500 may assign * to link 11 recording the shortest time among the remaining labels except for link 5 and may remove the label attached to link 5. This function of the rotation-based path search unit 500 may be performed by the mark change module 540 .

Referring to FIG. 9C , the rotation-based path search unit 500 may search for a link connected to link 11 after link 11 is determined. The rotation-based path search unit 500 may search for the link 20 connected to the link 11 and attach a label of (11, 2.7) to the link 20. The rotation-based path search unit 500 may assign * to the link 13 recording the shortest time among the remaining labels except for the link 11 and may remove the label attached to the link 11.

After the link 13 is determined, the rotation-based path search unit 500 may search for a link connected to the link 13 . The rotation-based path search unit 500 may search for links 8 and 11 connected to link 13. Since the destination node ID of link 8 is the starting point (node 1), the rotation-based path search unit 500 can exclude it from labeling. The rotation-based path search unit 500 may exclude link 11 from labeling because labeling has already been completed. The rotation-based route search unit 500 may assign * to link 9 recording the shortest time among the remaining labels excluding link 13 and may remove the label attached to link 13.

The rotation-based path search unit 500 may search for a link connected to link 9 after link 9 is determined. The rotation-based path search unit 500 may search for links 12, 14, and 15 connected to link 9 and label them as (9,3.4), (9,2.8), and (9,2.1). In the case of the link 14, the turn-based path search unit 500 can replace the existing label because the travel time is 0.1 faster than the values of the existing labels (2, 2.9). In the case of link 15, the turn-based route search unit 500 has a slower travel time than the existing labels (5, 2.0), so it cannot replace the existing label and may drop it. As such, the function of removing labels other than the label having the smallest traveling time of the rotation-based route search unit 500 may be performed by the label removal module 560 .

Referring to FIG. 9D , the rotation-based path search unit 500 may derive a final labeling result by repeating the above process. The rotation-based path search unit 500 may extract an optimal path by tracking backward labels in the final labeled result.

In this embodiment, when the rotation-based path search unit 500 extracts the path to node 9, since the label of node 9 is (23,3.8), it extracts the back link ID of 23 and extracts the path by referring to the label of link 23. can do. The rotation-based path search unit 500 can extract the back link ID of link 23 as link 15, the back link ID of link 15 as link 5, and the back link ID of link 5 as link 2. The rotation-based path search unit 500 may extract an optimal path as 9-8-5-2-1 as a node ID. The rotation-based path search unit 500 may extract the optimal path as 23-15-5-2 as a link ID. This function of the rotation-based path search unit 500 may be performed by the optimal path extraction module 550 .

10 shows a bucket & heap hybrid data structure used by the rotation-based path search unit 500 according to an embodiment of the present invention. Referring to FIG. 10 , the rotation-based path search unit 500 may search for an optimal path using a double bucket data structure constituting a lower bucket inside an upper bucket. The rotation-based path search unit 500 may use a structure capable of aligning the inside of a double bucket by configuring it. A heap structure may be applied to the rotation-based path search unit 500 as a data structure for internal alignment of double buckets. Preferably, a binary heap structure may be applied to the rotation-based path search unit 500 as a data structure for internal alignment of double buckets.

The rotation-based path search unit 500 may use a Bucket data structure, which is a structure for storing data having similar values in one bucket. In this embodiment, the rotation-based route search unit 500 may store accumulated travel time in a node based on one starting point. Assuming that the rotation-based path search unit 500 configures buckets in units of 1 minute, the 11th bucket can manage node IDs with a travel time greater than 10 minutes and less than 11 minutes as a list.

The rotation-based path search unit 500 may use a double bucket structure composed of an upper bucket and a lower bucket. The rotation-based path search unit 500 may be configured in such a way that 10 lower buckets exist inside the upper bucket when lower buckets are configured in units of 0.1 minutes. In this embodiment, the rotation-based path search unit 500 manages only the nodes whose cumulative travel time is greater than 10.5 minutes and less than 10.6 minutes to a node inside the bucket where the upper bucket is the 11th and the lower bucket is the 5th as a list. can do. The rotation-based path search unit 500 may keep the number of data finally stored in lower buckets small by using double buckets.

The rotation-based path search unit 500 can be operated by using a bucket data structure to store data having a value within a specific range, and rearranging the data inside when the order of the corresponding bucket comes. Since the rotation-based path search unit 500 performs sorting while the process of searching for a path using the labeling described above is in progress, when using the bucket&heap hybrid data structure, when the travel time is large, it can be managed in a small range because it is delayed. there is. Accordingly, the rotation-based path search unit 500 can reduce the size of data to be sorted and the number of times of sorting, thereby shortening the optimal path search time.

The first labeling module 510 may search for rotatable adjacent links among links directly connected to the starting node, and assign a label including a previous link number and travel time.

When there is no label with the mark, the mark module 520 may search for a link having the shortest travel time among the labeled links, and display a mark on the label of the searched link.

The second labeling module 530 may search for a rotatable adjacent link among labeled links having a mark and assign a label thereto.

The mark change module 540 may search for a link having the shortest travel time among labeled links excluding labels with a mark, display a mark on the label of the searched link, and delete a label with a previous mark.

The optimal path extraction module 550 may extract an optimal path by connecting previous link numbers included in labels of links directly connected to the destination node in reverse order.

When a plurality of labels are assigned to one link, the label removal module 560 may remove labels other than a label having the smallest travel time.

11 shows a flow chart of a path search algorithm of the optimal path search system 1 according to an embodiment of the present invention. Referring to FIG. 11 , the optimal path search system 1 may read node data, link data, and rotation constraint data from the server 3 . The optimal path search system 1 may construct a rotation-based rotation data structure using the read data. The optimal path search system 1 may change node or link-based rotation data stored in the server 3 into rotation-based rotation data. The optimal path search system 1 may extract the travel time of the changed rotation-based rotation data and store it in a corresponding field of the constructed rotation-based rotation data structure. The optimal path search system 1 may create a bucket structure and create a heap in the created bucket. The optimal path search system 1 may proceed with labeling of links using a labeling technique based on the Dijstra algorithm, which is a path search technology. The optimal route search system 1 can extract the optimal route to the destination by extracting the backward route.

As another embodiment of the present invention, the optimal route search method may include a turn-based rotation data construction step, a travel time extraction step, a turn-based rotation data conversion step, and a turn-based route search step.

The rotation-based rotation data construction step creates a rotation that can move from a link included in the link data to an adjacent link, and builds a rotation data structure including the rotation ID, departure link ID, arrival link ID, or travel time of the generated rotation. can The rotation-based rotation data building step refers to an operation performed by the above-described rotation-based rotation data building unit.

The travel time extraction step may extract the travel time corresponding to the rotation ID from the rotation data stored in the server and store it in the rotation data constructed in the rotation-based rotation data construction step. The travel time extraction step refers to an operation performed by the aforementioned travel time extraction unit.

In the rotation-based path search step, an optimal path may be extracted based on rotation data constructed in the rotation-based rotation data construction step. The rotation-based path search step refers to an operation performed in the aforementioned rotation-based path search step.

The turn-based route search step searches for adjacent rotatable links among the links directly connected to the starting node and assigns a label including the previous link number and travel time. If there is no label with a marker, travel time among the labeled links Retrieving this shortest link, labeling the retrieved link, searching for adjacent rotatable links from the links labeled with the marker and labeling them, assigning labels excluding the labeled with the marker Retrieving the link with the shortest travel time among the found links, displaying a mark on the label of the searched link and deleting the label with the previous mark, and the previous link number included in the label of the link directly connected to the destination node in reverse order. It may include a step of extracting an optimal path by connecting.

Although the present invention has been described in detail through representative embodiments, those skilled in the art will understand that various modifications are possible to the above-described embodiments without departing from the scope of the present invention. will be. Therefore, the scope of the present invention should not be limited to the described embodiments and should not be defined, and should be defined by all changes or modifications derived from the claims and equivalent concepts as well as the claims to be described later.

1: Optimal path search system
3: Server
100: rotation-based rotation data construction unit
300: travel time extraction unit
500: rotation-based path search unit
510: first labeling module
520: mark module
530: second labeling module
540: mark change module
550: optimal path extraction module
560: label removal module

Claims (9)

Node data consisting of node ID and coordinates, link data consisting of link ID, starting node ID, and arrival node ID, or rotation constraint data including rotation possibility information are read from the server to search for an optimal path. in the system,
Rotation-based rotation data for constructing a rotation-based rotation data structure that generates a rotation that can move from a link included in the link data to an adjacent link and includes the rotation ID, departure link ID, arrival link ID, or travel time of the generated rotation. construction department;
a travel time extraction unit extracting travel time corresponding to a rotation ID from rotation data stored in the server and storing the travel time in the rotation-based rotation data constructed by the rotation-based rotation data construction unit;
An optimal path search system comprising a rotation-based path search unit for extracting an optimal path based on the rotation-based rotation data built by the rotation-based rotation data construction unit.
According to claim 1,
The travel time extraction unit,
If the rotation data stored in the server is link-based rotation data including a departure link ID, an arrival link ID, and travel time, it is converted into rotation-based rotation data, and travel time corresponding to the rotation ID in the converted rotation-based rotation data. Optimal path search system, characterized in that for extracting.
According to claim 1,
The travel time extraction unit,
If the rotation data stored in the server is node-based rotation data composed of a start node ID, an intermediate node ID, an arrival node ID, and a travel time, a link ID corresponding to the start node ID and the intermediate node ID is stored as a start link ID, Link IDs corresponding to the intermediate node ID and destination node ID are stored as arrival link IDs, and converted into link-based rotation data including departure link ID, arrival link ID, and travel time;
The converted link-based rotation data is converted into rotation-based rotation data, and a travel time corresponding to a rotation ID is extracted from the converted rotation-based rotation data.
According to claim 1,
The rotation-based rotation data construction unit,
In a link with rotation constraint data, if the rotation constraint data is rotation prohibited data, the rotation prohibited from the generated rotation is deleted, and if the rotation constraint data is U-turn permitted data, a rotation that makes a U-turn is created,
An optimal path search system characterized in that in a link without rotation constraint data, a rotation that can be moved to an adjacent link is generated except for a rotation that makes a U-turn.
According to claim 1,
The rotation-based path search unit,
A first labeling module for searching for adjacent rotatable links from links directly connected to the starting node and assigning a label including a previous link number and travel time;
a marker module for searching for a link having the shortest travel time among labeled links when there is no label with a marker, and displaying a marker on the label of the searched link;
a second labeling module for searching for a rotatable adjacent link among labeled links having a mark and assigning a label thereto;
a label change module that searches for a link with the shortest travel time among labeled links excluding labels with a label, displays a new label on the label of the searched link, and deletes the label with the previous label;
An optimal route search system comprising an optimal route extraction module for extracting an optimal route by connecting previous link numbers included in the label of a link directly connected to the destination node in reverse order.
According to claim 5,
The rotation-based path search unit,
When a plurality of labels are assigned to one link, the optimal path search system further comprises a label removal module for removing labels other than the label with the smallest travel time.
According to claim 1,
The rotation-based path search unit,
An optimal path search system characterized by searching for an optimal path using a double bucket data structure constituting a lower bucket inside an upper bucket.
Node data consisting of node ID and coordinates, link data consisting of link ID, starting node ID, and arrival node ID, or rotation constraint data including rotation possibility information are read from the server to search for an optimal path. in the method,
Rotation-based rotation data for constructing a rotation-based rotation data structure that generates a rotation that can move from a link included in the link data to an adjacent link and includes the rotation ID, departure link ID, arrival link ID, or travel time of the generated rotation. construction phase;
a travel time extraction step of extracting a travel time corresponding to a rotation ID from rotation data stored in the server and storing the travel time in the rotation-based rotation data constructed in the rotation-based rotation data construction step;
and a rotation-based path search step of extracting an optimal path based on the rotation-based rotation data constructed in the rotation-based rotation data construction step.
According to claim 8,
In the rotation-based path search step,
Searching for an adjacent rotatable link from links directly connected to the starting node and assigning a label including a previous link number and travel time;
If there is no label with a mark, searching for a link having the shortest travel time among the labeled links, and displaying a mark on the label of the searched link;
searching for adjacent rotatable links in the labeled links having a marker and assigning labels thereto;
searching for a link with the shortest travel time among labeled links excluding labels with a mark, displaying a mark on a label of the searched link, and deleting a label with a previous mark;
An optimal route search method comprising the step of extracting an optimal route by connecting previous link numbers included in labels of links directly connected to the destination node in reverse order.

Images