JP2012118026A - Road node position management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To compensate a difference between a position of merging/outgoing node read from road information and a point of actually starting a merging/outgoing travel.SOLUTION: A road node position management system includes: a vehicle position information acquiring unit for acquiring vehicle position information expressing a present position of a vehicle; a road information acquiring unit for acquiring road information expressing a road by a connection relationship between a plurality of links and nodes; and a merging/outgoing node position learning unit that calculates a correcting position on the road of the merging/outgoing node as an intersection between a merging link or an outgoing link and a main road link based on the vehicle position information when detecting that the vehicle crosses a zone identification line through an image recognition of the zone identification line that is provided between the main lane and a merging/outgoing lane, and that stores the calculation result as a leaning node position, with its associated with the road information.

Description

  The present invention relates to a road node position management system that manages road nodes included in road information. This road node position management system uses route transition, map matching, vehicle control, and other processing as a reference point for the transition area between the main road and merging / exiting roads that exist on the travel route to reach the destination. The present invention can be applied to a vehicle navigation apparatus that performs the above.

  For example, Patent Document 1 describes a vehicle navigation system that provides travel guidance for a branch road in a special section having a branch point with an introduction lane that extends from a main lane such as an expressway or a main road to the branch road via the introduction lane. Has been. In this system, vehicle position information acquisition means for detecting a vehicle position, guide route calculation means for calculating a guide route from the vehicle position and a set destination, and an introduction lane that branches from the main line to a branch road through the introduction lane Special section detecting means for detecting a special section having an attached branch point from the guide route ahead of the vehicle position, introduction lane transition detecting means for detecting that the vehicle has shifted from the main line to the introduction lane of the special section, and A branch guidance unit that performs branch guidance for a branch point with an introduction lane, and that performs travel guidance for the branch road in response to the introduction lane shift detection unit detecting a shift of the vehicle to the introduction lane. ing. Therefore, in response to detecting the transition of the vehicle to the actual introduction lane, it is possible to provide travel guidance for the branch road before actually passing through the subsequent branch point with the introduction lane. The correct driving guidance for the vehicle can be given to the driver with a sufficient margin. However, in this technique, the next processing to be performed is performed appropriately based on the fact that the introduction lane transition detection means has detected the transition of the vehicle to the introduction lane, so that the next processing to be performed is appropriately performed. However, it does not serve as a starting criterion for processing to be performed before the vehicle moves to the introduction lane.

  Further, in an intersection area or the like, a marking line peculiar to an intersection exists as feature information, and a vehicle position recognition device that appropriately corrects the vehicle position using this is described in Patent Document 2. In this apparatus, vehicle position information acquisition means for acquiring vehicle position information representing the current position of the vehicle, road information acquisition means for acquiring road information, and target features included in the acquired image information around the vehicle. Image recognition means for performing image recognition processing, course change learning information storage means for storing course change learning information, and vehicle position correction means for correcting the vehicle position information based on the course change learning information are provided. . In the course change learning information, course change information that is generated by learning a past course change operation of the vehicle and that can identify a course change intersection that is an intersection where the course change operation of the vehicle has been performed, and the course change operation Information on distances from the recognition position of the target feature to the position where the vehicle's course change operation has been performed is included for the target feature that has been image-recognized earlier. Specifically, when the target feature is recognized by the image recognition unit, the vehicle position correcting unit is configured to detect the target based on the vehicle position information based on the course change learning information about the target feature. The vehicle position information is corrected so that a position that has traveled a distance indicated by the distance information from a recognized position of the feature coincides with a position in the road information of the course change intersection specified by the course change information. That is, it is possible to appropriately correct the vehicle position information using information obtained by learning the detection result of the past course changing operation of the vehicle and the image recognition result of the target feature before the detection. However, with this device, correction of the vehicle position information at the route change intersection is possible, but the node position included in the road information does not change. For this reason, for example, even if the start timing of the course changing operation at the intersection differs depending on the driver, guidance and control associated with the course changing operation are performed at the same timing. That is, the node position in the road information is used as being universally correct.

  In addition, when there is a nearby intersection, it was extracted from the road information in order to solve the problem that proper route guidance cannot be performed due to the error between the distance between the intersections obtained from the road information and the distance between the actual intersections. When the distance between the proximity guidance points is calculated and the calculated distance is less than the predetermined distance, the distance between the proximity guidance points in the road information is corrected to the distance between the actual proximity guidance points, and the corrected distance is used for guidance. A navigation device that creates information and performs guidance guidance is described in Patent Document 3. The correction in this device is a correction that eliminates the error between the distance between the intersections obtained from the road information and the actual distance between the intersections. Therefore, the correction is appropriate for each vehicle that travels with various traveling tracks depending on circumstances. It is difficult to make correct corrections.

JP 2009-109340 A (paragraph number 0005-0014, FIG. 1 and FIG. 2) JP 2009-8591 A (paragraph number 0008-0029, FIG. 1) JP 2008-286750 A (paragraph number 0004-0015, FIG. 3)

  For example, if the junction / exit node provided in the area between the main lane and the junction / exit lane is used as a reference point for the guidance control and subsequent control for this junction / exit driving, The point at which merging / leaving travel starts is more important than the position of the merging / leaving node in the road information. The point where such merging / exiting starts is changed depending on the road shape, the driving tendency of the driver, and the like. In addition, the travel locus calculated from the direction sensor, the distance sensor, etc. is compared with the road shape (specified by the link and the node) indicated by the road information, and the best matching road link and the position on the link are obtained. Also in map matching, the relationship between the actual travel locus and the road shape is important.

  Therefore, a technique that can compensate for the difference in position between the position of the joining / leaving node read from the road information and the point where the joining / leaving travel is actually started is desired.

  The characteristic configuration of the road node position management system according to the present invention is a vehicle position information acquisition unit that acquires vehicle position information that represents the current position of a vehicle, and road information that represents a road by a connection relationship of a plurality of links and nodes. A road information acquisition unit, an identification line straddling detection unit for detecting the vehicle straddling the area identification line through image recognition of the area identification line provided between the main lane and the merge / exit lane, and the straddle Based on the vehicle position information at the time of traveling detection, the corrected position on the road of the joining / exiting node as the intersection of the joining link or the leaving link and the main line link is calculated, and the road information is obtained using the calculation result as the learning node position. And a merging / leaving node position learning unit that stores the information in association with each other.

  According to this configuration, when it is detected that the vehicle has traveled across the area identification line provided between the main lane and the merge / exit lane, the vehicle position at the time of detection, that is, the vehicle travel locus Find the intersection point of the area identification line. Based on this intersection point position, a more appropriate position of the junction / exit node as the intersection of the corresponding junction link or exit link and main line link is calculated and derived. This derived position is the position of the joining / leaving node that matches the actual straddling travel (joining traveling or leaving traveling between the main lane and the joining / leaving lane). Therefore, this position is stored in association with road information as a learning node position. As a result, when the joining / leaving node is extracted during the next run, this learning node position is also read out and can be used. Therefore, the learning node position can be used in place of the position of the merge / exit node included in the road information, and the route guidance or map is based on the position that matches the travel locus for the past merge / exit. Matching can be performed.

  In one preferred embodiment, using the learning node position, the node position in the road information corresponding to a connection portion between the main lane and the merge / exit lane where the learning node position is calculated is corrected. And a node position correction unit. With this configuration, the node position in the road information corresponding to the learning node position is corrected using the learning node position as the position of the merging / leaving node suitable for actual merging traveling or leaving traveling. That is, the position of the junction / exit node, which is the point connecting the main link and the junction / exit link in the road information, is the actual vehicle junction / exit between the main lane and the junction / exit lane. Correction is made in accordance with the vehicle transition point between the main lane and the merge / exit lane obtained based on traveling, that is, the vehicle position at the time of detection of straddling travel. Such correction enables more appropriate processing such as route guidance and map matching.

  As one of the more effective use of the learning node position, the learning node position obtained by learning is used for the main lane and the merge / exit lane similar to the road shape of the main lane and merge / exit lane at the time of learning. However, it is possible to apply the same learning node position to a non-running merging / leaving node by changing the relationship between the position of the merging / leaving node and the learning node position. The joining / leaving node position learning unit of such a road node position management system stores the learning node position in association with the road shape of the main lane and the joining / leaving lane corresponding to the learning node position, The node position correcting unit is configured to correct the position of a joining / leaving node different from the joining / leaving node corresponding to the learning node position based on the learning node position and the road shape. As a result, even if a merging / leaving node that has not yet traveled appears, appropriate route guidance, map matching, and the like can be performed.

  The travel trajectory in actual merging traveling or leaving traveling is greatly affected by the degree of road congestion during the traveling. In other words, the traffic congestion on the road is a factor that determines the travel locus. Therefore, in one of the preferred embodiments of the present invention, the joining / leaving node position learning unit is configured to connect the learning node position to the main lane and / or joining / leaving lane corresponding to the learning node position. It is configured to store the information in association with the traffic condition of the lane. As a result, it is possible to set a joining / exit node position suitable for a traffic jam situation.

  In the present invention, the learning node position as the position of the joining / exiting node suitable for the crossing traveling is derived based on the vehicle position at the time of the crossing traveling detection. One of the specific examples is the detection. The position that is the starting point of the crossing travel that has been made is the learning node position. From the viewpoint that a steering operation is performed at this starting point, in one of the preferred embodiments of the present invention, the learning node position is a cause of the corresponding straddling travel detected by the identification line straddling detection unit. It is calculated as a position corresponding to the steering operation start point.

  Detection of crossing over the area identification line can be performed by examining whether or not the area identification line is included in the captured image of the road surface using a known image recognition technique. However, in order to effectively perform the image recognition, it is preferable to refer to the steering operation history. For example, since it is possible to estimate the possibility that the area identification line appears in the captured image from the steering operation history, it is possible to narrow down the captured image to be subjected to image recognition processing. For this purpose, the identification line straddling detection unit may be configured to detect the straddling travel with reference to a steering operation history in addition to the image recognition.

Furthermore, the present invention provides a road node position management method and road node position management for compensating for the difference in position between the position of the merge / leave node read from the road information and the point where the merge / leave travel is actually started. The program is also subject to rights. For example, such a road node position management method includes a step of obtaining vehicle position information representing a current position of a vehicle, a step of obtaining road information representing a road by a connection relation of a plurality of links and nodes, a main lane, The step of detecting the crossing of the vehicle with respect to the area identification line through image recognition of the area identification line provided between the merging / exiting lanes, and the merging based on the vehicle position information at the time of detecting the crossing driving A merging / leaving node position learning step of calculating a correction position on the road of the merging / leaving node as an intersection between the link or the leaving link and the main line link, and storing the calculation result in association with the road information as a learning node position; It has.
In addition, such a road node position management program includes a function of acquiring vehicle position information indicating the current position of a vehicle, a function of acquiring road information indicating a road by a connection relationship of a plurality of links and nodes, a main lane, Based on the image recognition of the area identification line provided between the merging / exiting lanes, the function of detecting the crossing of the vehicle with respect to the area identification line, and the merging based on the vehicle position information at the time of detection of the crossing driving A merging / leaving node position learning function for calculating a corrected position on the road of the merging / leaving node as an intersection of the link or the leaving link and the main line link, and storing the calculation result as a learning node position in association with the road information; Is executed on the computer. Naturally, such a road node position management method and a road node position management program can also obtain the effects described in the above-described road node position management system, and some additional effects described as preferred embodiments thereof. It is also possible to incorporate technology.

It is explanatory drawing which has shown typically the basic principle of the road node position management system by this invention. It is explanatory drawing which has shown typically the learning node position calculation algorithm in the basic principle of the road node position management system by this invention. It is the schematic diagram which modeled an example of the road information handled with the road node position management system by this invention. It is a functional block diagram of the navigation device for vehicles concerning one embodiment of the present invention. It is a flowchart which shows an example of a joining / leaving node learning routine. -A chart. It is a flowchart which shows an example of a straddle detection routine. -A chart. It is a flowchart which shows an example of a joining / exit node position learning routine. -A chart. It is a flowchart which shows an example of a similar node correction routine. -A chart.

Before describing a specific embodiment of the road node position management system according to the present invention, its basic principle will be described with reference to FIGS. FIG. 1 is an explanatory view schematically showing the basic principle of a road node position management system according to the present invention. FIG. 2 is a schematic diagram illustrating three examples of learning node position calculation algorithms that can be used in the road node position management system according to the present invention.
Fig. 1 shows an example of exit / running that branches from a main line such as an expressway or a main road to a destination road through the exit lane as an example of merge / exit travel performed between the main lane and the junction / exit lane. The flow of various information (data) executed by the road node position management system according to the present invention in the leaving traveling is illustrated.

FIG. 3 schematically shows the road information of the road region in FIG. 1, and usually such road information is made into a database and stored in the map database in a form according to the specification. Here, as an example, the road network layer m1, the road shape layer m2, and the road attribute layer m3 are divided into three layers and stored.
The road network layer m1 is a layer indicating connection information between roads. Specifically, information on a large number of nodes N having position information on a map expressed by latitude and longitude, and information on a number of links K constituting a route corresponding to a road by connecting the two nodes N, Is recorded. Each link K records information such as road type (type of highway, toll road, national road, prefectural road, etc.) and link length as link information. Note that the node N connected to the branching link K is called a branch node. However, since this branch node is a reference point for joining and leaving the vehicle, it is also called a joining / leaving node here. In addition, a link connected by a branch node to a main link that is a link representing a main lane (main road) is called a merge link or an exit link.

The road shape layer m2 is a layer that is stored in association with the road network layer m1 and indicates the shape of the road. Specifically, information on the shape interpolation point s that is arranged between two nodes N (on the link K) and represents the detailed shape of the link K, and at each shape interpolation point s that represents the detailed shape of the road. Information on road width is recorded.
The road attribute layer m3 is configured in association with the road network layer m1 and the road shape layer m2, and is a layer in which detailed information regarding the road is recorded. Information recorded in the road attribute layer m3 includes, for example, information related to branch nodes, lane information of each road, and the like. Here, the lane information includes information such as the number of lanes of each road, the lane width of each lane, and the lane length. The lane information also includes information such as the positional relationship and connection relationship (link angle, etc.) of each lane between a plurality of roads that are branched and connected at a branch node. Here, the road information is a general term for information stored in each of the layers m1 to m3, and means any information included in the information.

  FIG. 1 shows a merging / exit node (simply referred to as “exit node” in the following description of the basic principle using FIG. 1 and FIG. 2) as an exit travel route that shifts from a main lane to an exit lane, A link connected to the link and a vehicle trajectory of the vehicle traveling on the exit travel route are illustrated. Further, in order to explain the leaving travel with time, a time point indicating the passage of time, t1 to t5, is set, and the situation at that point will be described below. 1 and 2, the leaving node N2 is indicated by a black dot, and each link connected to the leaving node is indicated by a thick solid line.

(1) Time point: t1
At this time: the vehicle is still traveling on the main lane in the time zone indicated by t1. Time point: t1 is a time (time stamp) when the vehicle position: P1 is generated.
(2) Time point: t2
At this time: t2 is a time (time stamp) when the steering operation is performed to move to the exit lane from the main lane to the exit lane. At this time: vehicle position P2 at t2 is acquired. The steering operation is indicated by a change in the steering angle. At this time point: t2, if a steering angle: θ indicating a predetermined steering operation occurs, steering operation information including the steering angle: θ is generated. Further, as an event of the steering operation history, the vehicle position: P2 at the time point: t2 and the steering operation information: θ are associated. With the detection of the steering operation represented by this steering operation information: θ as a trigger, the image recognition processing of the white thick broken line as the area identification line drawn on the road surface between the main lane and the exit lane starts. This image recognition process is performed based on a road surface image captured by the in-vehicle camera.

(3) Time point: t3
At this time point: t3, the vehicle travels across the area identification line drawn in white on the road surface by a white thick broken line in this embodiment. In the image recognition process for the road surface image taken by the in-vehicle camera, the area identification line is recognized, and the straddle recognition information X indicating this is issued. At this time: t3 is the start timing, and the following learning node position calculation process starts.
(3-1)
Crossing recognition information: When X is issued: Vehicle position: P3 acquired at t3 is regarded as a zone identification line crossing travel position: Px. A learning node position, which is a new node position, is calculated using this straddling travel position as an input parameter. Various algorithms can be considered as this calculation algorithm: G, but the present invention is not limited to a specific algorithm. Here, the three algorithms illustrated in FIG. 2 are taken up and described below.
(A) A straddling travel position at which straddling travel is recognized: A perpendicular line is drawn from the link P1 of the main lane from Px, and the intersection is set as a learning node: Nx.
(B) Crossing travel position: The angle formed by the straight line from Px between the main lane and the exit lane, that is, the branch link angle formed by the links K2 and K3 (also simply referred to as the link angle): the point at which α intersects the link K1 Learning node: Nx.
(C) A point on the link of the main lane closest to the vehicle position: P2 steered to shift from the main lane to the exit lane is defined as a learning node: Nx.
(3-2)
The calculated position of the learning node: Nx is stored in association with the leaving node: N2 as the corrected position of the leaving node: N2. Even if the calculation algorithm is different, the calculated position of the learning node: Nx is likely to be closer to the position where the exit travel is actually started than the position of the exit node: N2. Therefore, at the time of the next travel, the learning node: Nx once stored is used in place of the exit node: N2, thereby improving the timing of guidance before exit travel and the accuracy of map matching after exit travel.

(4) Time point: t4
At this time point: at t4, the transition of the vehicle to the exit lane is completed, and the vehicle is traveling in the exit lane.
(5) Time point: t5
At this time point: t5, the vehicle moves to the exit lane, and the vehicle position is set to a predetermined position in the exit lane by map matching for the link set in the exit lane.

[Use of learning node position]
When the leaving node N2 is included in the planned travel route determined by the map matching, it is checked whether or not the learning node Nx is associated with the leaving node N2. If the learning node Nx is associated, the route guidance process and the subsequent map matching process are performed using the position of the learning node Nx instead of the leaving node N2.

  The basic feature of this road node position management system is that the vehicle position information indicating the current position of the vehicle and the road information on the road are acquired, and the area identification provided between the main lane and the merge / exit lane is identified. Based on the vehicle position information when the vehicle straddling to the area identification line is detected through line image recognition, the position of the merge / exit node on the road as the intersection of the merge link or the exit link and the main line link is calculated The calculation result is stored in association with road information as a learning node position.

  The timing at which the steering operation is performed to start leaving the vehicle is roughly determined by the driving tendency of the driver, but varies depending on the situation. Examples of the types of situations that cause such changes include the degree of congestion on the main lane or the exit lane, the length and width of the exit lane, the travel time (night or noon), and the weather. Therefore, it is preferable to store the position of the learning node for each kind of situation, and use it appropriately according to the situation when the learning node is used.

  In the above description of the principle, the calculation of the learning node for the leaving node using exit traveling is taken as an example, but the learning node for the joining node can also be calculated using the same procedure using joining traveling. Therefore, here, the exit node and the join mode are regarded as substantially the same, and are treated as the exit / join mode.

  FIG. 4 is a functional block diagram showing an example of a vehicle navigation apparatus that employs a road node position management system that operates based on the basic principle described above. This navigation apparatus includes a navigation electronic control unit (hereinafter simply referred to as a navigation ECU) 3 as a core element, a dead reckoning navigation unit 5, a GPS unit 6, an image processing unit 7, a vehicle state detection management unit 8, and an in-vehicle camera 11. , A monitor 12 with a touch panel 13, a speaker 14, and the like.

Signals from the distance sensor 8a and the azimuth sensor 8b are input to the dead reckoning unit 5 via the vehicle state detection management unit 8. The distance sensor 8a and the direction sensor 8b send detection signals to the dead reckoning unit 5 via the vehicle state detection management unit 8. The dead reckoning navigation unit 5 calculates dead reckoning position coordinates based on the signals related to the movement distance and the signals relating to the azimuth, which are sent every moment, and sends them to the navigation ECU 3 as vehicle position information by dead reckoning.
The GPS unit 6 includes a GPS measurement unit that receives GPS signals from GPS satellites. The GPS unit 6 analyzes the signal from the GPS satellite received by the GPS measurement unit, calculates the current position (latitude and longitude) of the vehicle, and sends it to the navigation ECU 3 as vehicle position information based on the GPS position coordinates.

The image processing unit 7 includes a captured image acquisition unit 7a and an identification line straddling detection unit 70.
The captured image acquisition unit 7a acquires a captured image output by the camera 11 and performs necessary preprocessing such as distortion correction and density correction. The camera 11 includes a lens optical system and an image pickup device, and is at least a section line of a road around the vehicle, for example, a roadway center line, a lane boundary line, or an area provided between a main lane and a merge / exit lane An identification line (in this example, a white thick broken line) is provided at a position where photographing is possible. As such a camera 11, for example, a back camera or the like that is standard or optional on the vehicle is preferably used. The captured image acquisition unit 7a captures captured images from the camera 11 at a predetermined time interval, for example, about 10 to 50 ms.

  The identification line straddling detection unit 70 starts to operate in response to a command from the navigation ECU 3, and detects that the vehicle has entered the exit lane across the area identification line from the main lane. In this embodiment, the identification line straddling detection unit 70 detects the vehicle identification line stride based on the recognition of the area identification line obtained through the image recognition processing of the captured image from the camera 11. For this purpose, the identification line crossing detection unit 70 includes an identification line extraction unit 71, an identification line image buffer 72, and an identification line crossing determination unit 73. The identification line extraction unit 71 performs binarization processing, edge detection processing, and the like on the captured image captured from the captured image acquisition unit 7a, and indicates a partition line indicating the outline of the area identification line included in the captured image. Extract images. The identification line image buffer 72 temporarily stores the generated lane marking images in order over time. The identification line straddling determination unit 73 compares the lane line images stored in the identification line image buffer 72 over time, obtains a motion vector from the lane line image, for example, and finally the vehicle straddles the area identification line. Judge that. When traveling across the area identification line of the vehicle is detected, the identification line straddling detection unit 70 sends identification line straddling detection information to the navigation ECU 3.

  The navigation ECU 3 includes a map database 30, a vehicle position information acquisition unit 31, a road information acquisition unit 32, and a navigation processing unit 4 as basic navigation function units. The vehicle position information acquisition unit 31 reads the current position of the vehicle from the vehicle position information sent from the dead reckoning navigation unit 5 or the GPS unit 6. The road information acquisition unit 32 acquires road information including necessary map information from the map database 30 based on the current position of the vehicle read by the vehicle position information acquisition unit 31. The road information acquisition part 32 gives the acquired road information to the function part of other navigation ECU3. Further, the road information acquisition unit 32 acquires the map information of the area requested for execution of the navigation processing in the navigation processing unit 4 from the map database 30 and transfers it to the navigation processing unit 4. As schematically shown in FIG. 3, the map database 30 includes data in a format divided into layers such as a road network layer m1, a road shape layer m2, a road attribute layer m3, and data in a list format. Various road information is stored.

  The navigation processing unit 4 operates in accordance with various installed application programs for executing navigation functions such as vehicle position display, route search from a departure point to a destination, route guidance to a destination, and destination search. An arithmetic processing function unit is included. For example, the navigation processing unit 4 displays a map image included in the road information around the current vehicle position acquired by the road information acquisition unit 32 on the monitor 12, and on the map image, based on the vehicle position information. A process of displaying the vehicle position mark (symbol) in an overlapping manner is performed. Furthermore, the navigation processing unit 4 uses the monitor 12 and / or the speaker 14 based on the guidance route from the departure point to the destination calculated by a known method and the vehicle position information, and the route including the branch guidance. Give guidance. Furthermore, when the vehicle position sent from the vehicle position information acquisition unit 31 includes a measurement error or the like and deviates from the road, the navigation processing unit 4 sets the vehicle position on the road indicated on the road map. It also has a map matching function unit that performs correction. The vehicle position corrected through the map matching process is given to the vehicle position information acquisition unit 31.

  The navigation ECU 3 includes a communication interface 20, a display control unit 21, and a sound processing unit 22. The communication interface 20 connects other modules or ECUs connected via the in-vehicle LAN and the navigation ECU 3 so that data can be transmitted. The display control unit 21 controls the display of route guidance images and captured images on the monitor 14. The voice processing unit 22 generates a voice signal obtained by converting the route guidance information and the like into voice, and outputs the voice signal to the speaker 14.

  Further, the navigation ECU 3 implements the travel path calculation unit 33, the identification line straddling detection confirmation unit 34, the steering operation history management unit 35, and the merge / exit node position learning in order to realize the road node position management system described above. Part 36 and a node position correcting part 37. The travel locus calculation unit 33 calculates the vehicle travel locus from the requested predetermined start timing based on the vehicle position information from the vehicle position information acquisition unit 31 and temporarily stores it. The identification line straddling detection confirming unit 34, with respect to the identification line straddling detection unit 70, straddles the vehicle with respect to the area identification line through image recognition of the area identification line provided between the main lane and the merge / exit lane. An identification line straddling travel detection process for detecting travel is requested, and based on the result, traveling over the identification line of the vehicle is confirmed. Further, the identification line straddling detection confirmation unit 34 holds the vehicle position information at the time when the identification line straddling has occurred. The steering operation history management unit 35 records the steering operation history at a predetermined timing, for example, when a joining / exiting node as an intersection between the joining link or the leaving link and the main line link is present in the traveling direction of the vehicle. , And it becomes possible to grasp the time point or point at which the steering operation is performed to join or leave the vehicle.

  The joining / leaving node position learning unit 36 joins / leaves as the intersection of the joining link or leaving link and the main line link based on the vehicle position information at the time of detecting the crossing travel confirmed by the identification line straddling detection confirmation unit 34. The corrected position of the node on the road is calculated, and the calculation result is stored as a learning node position in association with road information, for example, the original joining / leaving node. As the algorithm for calculating the correction position, (a), (b), (c), etc. in the description of the basic principle described above with reference to FIGS. 1 and 2 can be adopted. The joining / leaving node position learning unit 36 also has a function of storing the learning node position in association with the traffic congestion status of the main lane and / or the joining / leaving lane corresponding to the learning node position. Is possible.

Using the learning node position obtained by the node position correcting unit 37 and the joining / leaving node position learning unit 36, the above-mentioned corresponding to the connecting part between the main lane and the joining / leaving lane where the learning node position is calculated. A process for correcting the node position in the road information is performed. As a result, correction is performed so that the joining / exiting node position in the road information originally stored in the map database 30 matches this learning node position.

  In addition, if the road shape of the main lane and the merge / exit lane are substantially the same, it is conceivable that the driver behaves in the same manner during the merge / exit travel. Therefore, in this embodiment, the joining / leaving node position learning unit 36 stores the learning node position in association with the main lane and the road shape of the joining / leaving lane corresponding to the learning node position, Based on the learning node position and the road shape, the node position correcting unit 37 can also correct the position of a joining / leaving node different from the joining / leaving node corresponding to the learning node position. More specifically, at a junction / exit node at a place where the road shape approximates, it is assumed that the traveling locus of the vehicle also approximates, and the learning node position already stored in such an approximate junction / exit node is used. The position of the merge / leave node is corrected. The approximate judgment condition parameters of the road shape used at that time are the length of the area identification line, the road width of the main lane, the number of lanes of the main lane, the road width of the merge / exit lane, the main lane and the merge / exit lane The link angle made by A connection section between the main lane and the merge / exit lane corresponds to the connection portion of the present invention. The connection section here is a section where the lane of the vehicle can be transferred between the main lane and the merge / exit lane, and more specifically, the section where the main lane and the merge / exit lane are in contact. A section that is adjacent to each other and extends in parallel is a connection section. Specifically, as shown in FIGS. 1 and 2, a section in which an area identification line is provided between the main lane and the merge / exit lane is a connection section.

A flow of control of learning management of the node position in the road node position management system configured as described above will be described with reference to the flowchart of FIG.
When a guide route to the destination input through the touch panel 13 is calculated and travel is started, a joining / exit node learning routine is performed for the latest planned travel route sequentially taken from the guide route along with the travel. Done. Therefore, first, the road information acquisition unit 32 takes out the latest planned travel route (# 01). Whether the identification line crossing detection confirmation unit 34 includes the merge / exit node scheduled to travel from the confluence lane to the main lane or from the main lane to the exit lane in the extracted planned travel route. Check (# 02). If such a joining / leaving node is not included (# 02 No branch), it waits for the next latest travel planned route to be taken out. If such a merge / leave node is included (Yes in # 02), the identification line crossing detection confirmation unit 34 further checks whether the merge / leave node exists within a predetermined distance (#). 03). The vehicle waits until the vehicle enters within a predetermined distance from the joining / leaving node. When the vehicle enters within the predetermined distance from the joining / leaving node (# 03 Yes branch), the identification line straddling detection confirmation unit 34 identifies the image processing unit 7. The line crossing detection is requested (# 04). The image processing unit 7 receives the request for detection of the identification line crossing and executes a crossing detection routine (# 05).

  When the straddle detection routine shown in FIG. 6 is executed, the captured image acquisition unit 7a sequentially captures captured images of the in-vehicle camera 11, performs necessary preprocessing, and develops them in the memory (# 51). The identification line straddling detection unit 70 performs the following straddle detection processing on the captured image developed in the memory. First, the identification line extraction unit 71 extracts an area identification line that is a thick broken line between the main lane and the merge / exit lane using an edge detection filter or the like (# 52). When the area identification line is not extracted (# 53 No branch), the process proceeds to the next photographed image. When the area identification line is extracted (Yes in # 53), the extracted area identification line is marked and the identification line image is temporarily stored in the identification line image buffer 72 (# 54). The identification line straddling determination unit 73 determines the timing at which the vehicle straddles the area identification line from the identification line images stored in the identification line image buffer 72 in the order of time (# 55). If it is not determined that the vehicle crosses the area identification line (# 56 No branch), the elapsed time of the crossing detection process is checked (# 57). If the elapsed time is within the predetermined time (# 57 No branch), the crossing detection process is further continued. When the elapsed time exceeds the predetermined time (# 57 Yes branch), the identification line straddling detection unit 70 sends straddle detection information whose content is “non-detection” to the navigation ECU 3 (# 58), and ends this routine. When it is determined in step # 55 that the vehicle has crossed the area identification line (Yes in # 56), the identification line crossing detection unit 70 sends the crossing detection information including “detection” to the navigation ECU 3 (# 59). End the routine.

  When the straddle detection routine ends, the identification line straddle detection confirmation unit 34 confirms straddle detection from the straddle detection information sent from the image processing unit 7 (# 6). If it is confirmed that the identification line crossing is not detected (# 6 No branch), the process of step # 7 is jumped and the process proceeds to step # 8. When the identification line crossing is confirmed (# 6 Yes branch), the time (time stamp) when the identification line crossing is detected, the vehicle position information at that time, and the merge / exit node learning process are the targets. The process proceeds to the join / leave node position learning routine by the join / leave node position learning unit 36 using the information of the join / leave node as an argument (# 7).

This joining / leaving node position learning routine is shown in FIG. 7, and an explanatory schematic diagram thereof is shown in FIG. In this embodiment, (c) in the learning node position calculation algorithm exemplified in the description of the basic principle using FIG. 2, that is, the vehicle position steered to shift from the main lane to the exit lane. A method is adopted in which a point on the link of the nearest main lane is used as a learning node.
First, the joining / leaving node position learning unit 36 reads vehicle position information at the time when the identification line straddling is detected (# 71). Further, the joining / leaving node position learning unit 36 goes back from the steering operation history recorded by the steering operation history management unit 36 starting from the time when the identification line straddling is detected, and is detected by the identification line straddling detection unit 70. Further, the steering operation start time point that causes the corresponding straddling travel is extracted (# 72). Then, the joining / leaving node position learning unit 36 obtains the vehicle position at the start point of the extracted steering operation (# 73). Next, the joining / leaving node position learning unit 36 calculates the position on the link in the road information corresponding to the steering operation start point that is the obtained vehicle position as the learning node position (# 74). Here, the position on the link of the main lane closest to the steering start point is set as the learning mode position. Finally, the obtained learning node position is recorded in a predetermined learning node position storage location associated with the corresponding joining / leaving node (# 75), and this routine is terminated.

  When the navigation ECU 3 returns from the joining / leaving node position learning routine, it is checked whether or not the termination of the navigation processing is instructed (# 8), and unless the termination is instructed (# 8 No branch), this joining / leaving node When the learning routine is repeated and the termination is instructed (# 8 Yes branch), the navigation process is terminated.

  The learning node position generated and recorded through the joining / leaving node position learning routine described above can be used by the node position correcting unit 37 in place of the position of the joining / leaving node associated with the learning node position. Further, when the accuracy of the recorded learning node position reaches a predetermined level, the position of the joining / leaving node in the map database 30 can be rewritten with the corresponding learning node position.

  Furthermore, even if the learning node position is not recorded through the joining / leaving node position learning routine, the joining / leaving node in which the surrounding situation of the joining / leaving node records the learning node position. Can be used as the learning node position of the merging / leaving node that is approximated. A routine of such processing will be described with reference to the flowchart of FIG. Here, it is assumed that the joining / leaving node is provided with the surrounding situation / node table recorded in association with the surrounding situation information regarding the joining / leaving node.

  First, ambient condition information is acquired (# 81). Here, the ambient state information is first a road shape, and additionally includes weather, time (brightness), road congestion degree, and the like. Approximate judgment condition parameters related to road shape include the length of the area identification line, the road width of the main lane, the number of lanes of the main lane, the road width of the merge / exit lane, the link angle between the main lane and the merge / exit lane, etc. It is. The ambient situation / node table is searched from the obtained ambient situation information to determine whether or not an ambient situation similar to the actual surrounding situation that is the surrounding situation of the merging / leaving node scheduled to travel is recorded (# 82). If the search result is checked (# 83) and a similar situation exists (Yes branch at # 83), it is further checked whether the learning node position is associated with the joining / exiting node associated with the similar situation (# 84). . If there is no similar situation (# 83 No branch), and if the learning node position is not associated with the joining / leaving node (# 84 No branch), this routine is terminated as it is. If the learning node position is associated with the joining / leaving node (# 84 Yes branch), a more appropriate learning node position is selected (# 85). Then, the position of the joining / leaving node scheduled to travel is corrected at the selected learning node position (# 86). In the learning node selection at step # 85, when there is only one learning node position, an appropriate one is selected from the learning node position and the original joining / leaving node position.

[Other Embodiments]
(1) In the above-described embodiment, the image recognition function for detecting that the vehicle has traveled across the area identification line is provided in the image processing unit 7, but the image processing unit 7 is a target of image recognition. A configuration may be adopted in which only pre-processing of the captured image is performed and actual image recognition is performed by another unit, for example, the navigation ECU 3.

(2) In the above-described embodiment, the configuration in which the learning node position calculation algorithm selected in advance is employed in the joining / leaving node position learning unit 36, but a plurality of learning node position calculation algorithms are prepared. However, it may be arbitrarily selectable.

(3) In the above-described embodiment, the correction node based on the calculated learning node position is stored in association with the corresponding joining / leaving node. However, even in the same junction / exit node, the timing of steering operation for junction or exit may be different depending on road conditions or driving conditions. Therefore, even for the same joining / exiting node, the learning node position may be calculated every time such a road situation or driving situation is different, and a correction node may be stored for each road situation or driving situation. In this configuration, when the learning node position is used, a more appropriate correction node can be used according to the road condition or the driving condition. Such use of the learning node position can be performed not only for the merging / leaving node corresponding to the learning node position but also for other merging / leaving nodes that approximate road conditions and driving conditions.
Road conditions include: (1) Congestion level of main lanes, (2) Congestion level of merge / exit lanes, (3) Length of area identification line as a kind of road shape, that is, length of merge / exit lanes (4) Road widths of main lanes and merging / exiting lanes as a kind of road shape, and (5) time zone during which merging / exiting occurs. In addition, driving conditions include (11) driver fatigue, (12) familiarity (running frequency) with respect to joining / exiting nodes, and the like.

(4) In the embodiment described above, an image recognition process for a road surface image taken by a vehicle-mounted camera is employed to detect that the vehicle has traveled across the area identification line. However, in consideration of bad weather and night driving, in order to improve the accuracy of straddle detection, detection of straddle over the area identification line is performed from the travel locus calculated by referring to the steering operation history, and image recognition is performed. A technique that collates with the result of the above may be adopted.

(5) In the embodiment described above, associating the learning node position with the road information means that the learning node position is associated with the corresponding joining / leaving node included in the road information. For example, the link may be associated with a link. It may be associated with a link included in other road information, or a specific feature or symbol.

  The road node position management system according to the present invention can be applied not only to car navigation but also to a technical field in which various controls are executed based on nodes included in road information.

3: Navigation ECU
31: Vehicle position information acquisition unit 32: Road information acquisition unit 33: Travel locus calculation unit 34: Identification line crossing detection confirmation unit 35: Steering operation history management unit 36: Join / exit node position learning unit 37: Node position correction unit 4 : Navigation processing unit 5: Dead reckoning unit 6: GPS unit 7: Image processing unit 7a: Captured image acquisition unit 70: Identification line crossing detection unit 71: Identification line extraction unit 72: Identification line image buffer 73: Identification line crossing determination unit 8: Vehicle state detection management unit

Claims (8)

A vehicle position information acquisition unit for acquiring vehicle position information representing the current position of the vehicle;
A road information acquisition unit that acquires road information representing a road by a connection relationship of a plurality of links and nodes;
An identification line straddling detector that detects the travel of the vehicle with respect to the area identification line through image recognition of the area identification line provided between the main lane and the merge / exit lane;
Based on the vehicle position information at the time of detection of the straddling travel, the correction position on the road of the junction / exit node as the intersection of the junction link or the exit link and the main line link is calculated, and the calculation result is used as the learning node position. A merging / leaving node position learning unit that stores information associated with road information;
Road node position management system with   A node position correcting unit that corrects a node position in the road information corresponding to a connecting unit between the main lane and the joining / exiting lane in which the learning node position is calculated using the learning node position. The road node position management system according to Item 1.   The joining / leaving node position learning unit stores the learning node position in association with road shapes of the main lane and the joining / leaving lane corresponding to the learning node position, and the node position correcting unit stores the learning node position. The road node position management system according to claim 2, wherein the position of a joining / leaving node different from the joining / leaving node corresponding to the learning node position is corrected based on the node position and the road shape.   4. The joining / leaving node position learning unit stores the learning node position in association with a congestion situation of a main lane and / or a joining / leaving lane corresponding to the learning node position. The road node position management system according to any one of the above.   5. The learning node position according to claim 1, wherein the learning node position is calculated as a position corresponding to a steering operation start point detected by the identification line straddling detection unit and causing a corresponding straddling travel. Road node location management system.   The road node position management system according to any one of claims 1 to 5, wherein the identification line straddling detection unit detects the straddling travel with reference to a steering operation history in addition to the image recognition. Obtaining vehicle position information representing the current position of the vehicle;
Acquiring road information representing a road by a connection relationship of a plurality of links and nodes;
Detecting the crossing of the vehicle with respect to the area identification line through image recognition of the area identification line provided between the main line lane and the merge / exit lane;
Based on the vehicle position information at the time of detection of the straddling travel, the correction position on the road of the junction / exit node as the intersection of the junction link or the exit link and the main line link is calculated, and the calculation result is used as the learning node position. A joining / leaving node position learning step for storing in association with road information;
Road node position management method comprising: A function of acquiring vehicle position information representing the current position of the vehicle;
A function of acquiring road information representing a road by a connection relation of a plurality of links and nodes;
A function of detecting the crossing of the vehicle with respect to the area identification line through image recognition of the area identification line provided between the main line lane and the merge / exit lane;
Based on the vehicle position information at the time of detection of the straddling travel, the correction position on the road of the junction / exit node as the intersection of the junction link or the exit link and the main line link is calculated, and the calculation result is used as the learning node position. A join / leave node location learning function that stores information associated with road information,
Road node position management program for causing a computer to execute.

Images