DEVICE FOR GENERATING DATA OF A ROAD WITHIN AN INTERSECTION, PROGRAM FOR GENERATING DATA OF A ROAD WITHIN AN INTERSECTION, AND STORAGE MEDIUM

Abstract

Route data generation device for generating data of a route within an intersection for automated driving, comprising: - a route data generation unit (2) that generates the route data such that the route data generation unit (2), using an absolute motion path of actual driving of a vehicle within the intersection, fits an estimated motion path of the actual driving of the vehicle within the intersection into lane network data associated with the intersection, wherein - the route data generation unit (2) has: - a lane network data specification unit (7) that specifies inbound lane network data describing a lane network from which the intersection is entered and outbound lane network data describing a lane network onto which the intersection is exited; - a center point coordinate specification unit (8) that specifies an intersection point between an extension line of the inbound lane network data and an extension line of the outbound lane network data as a central coordinate point of the intersection; - a nearest coordinate specification unit (9) which, as a nearest coordinate point, specifies a coordinate point that is closest to the central coordinate point of the intersection from several coordinate points that form the absolute trajectory; - a segment setting unit (10) which defines a motion path detection segment and a fitting segment based on the nearest coordinate point for both an input-side absolute motion path and an output-side absolute motion path, wherein the input-side absolute motion path is an input-side section of the absolute motion path and the output-side absolute motion path is an output section of the absolute motion path; - a mean calculation unit (11) that calculates a mean of distance differences between the fitting segment of the input-side absolute motion path and an input-side estimated motion path, wherein the input-side estimated motion path is a section of the estimated motion path corresponding to the fitting segment of the input-side absolute motion path; - a first translation unit (12) that performs a translation of the estimated trajectory by the mean of the distance differences; - a straight-line distance calculation unit (13) that calculates a straight-line distance from a rotation reference point to the input lane network data, wherein the rotation reference point is defined as an intermediate point of the fitting segment of the input estimated trajectory; - a second translation unit (14) that performs a translation of the estimated trajectory by the straight-line distance; - an estimated motion path rotation unit (15) that rotates the estimated motion path such that a distance difference between the fitting segment of the input estimated motion path and the input lane network data is minimized; - an input path segment vector setting unit (16) that specifies an input path segment vector extending from a starting point to an endpoint of the fitting segment of the input-side estimated trajectory; and - a third translation unit (17) that performs a translation of the estimated motion path in the direction of the input track segment vector such that a distance difference between the output estimated motion path and the output lane network data is minimized.

Description

TECHNICAL AREA

[0001] The present disclosure relates to a device for generating data about routes within intersections, a program for generating data about routes within intersections and a storage medium. STATE OF THE ART

[0002] A known method exists for measuring the shape, course, and position of a road with high accuracy using specialized vehicles and for generating route data for automated driving. This method inherently requires enormous effort due to expensive sensors and labor, and can only generate route data in limited areas, such as highways or main roads. Therefore, it is not possible to generate route data for local roads or similar infrastructure, nor is it possible to create route data for intersections. Consequently, there is a need for a technology to generate route data within intersections.

[0003] JP 2017-97088 A, for example, discloses a map data update method in which a new road is estimated using a GPS trajectory that represents the GPS (Global Positioning System) positions of a vehicle as an absolute path of movement, and a connection between the new road and an existing road is estimated. Furthermore, JP 2010-26875 A, for example, discloses a method for generating data of a driving path within an intersection, in which an entry lane, from which the intersection is "entered," and an exit lane, onto which the intersection is exited, are connected by an arc (a quadratic Bézier curve).

[0004] The method disclosed in JP 2017-97088A has a problem in that the GPS positions are widely dispersed, and the route data generated by the above method has lower accuracy. The method disclosed in JP 2010-26875A is unrealistic because actual vehicle routes within intersections vary depending on the shape of the intersection, and it is very likely that the route data generated by the above method will deviate from actual trajectories.

[0005] From JP 2009 - 276 224 A, a device and a program for calculating the current position on a road segment are also known, wherein the device for calculating the current position calculates a candidate for the current position based on absolute position information, calculates a drivable area in which a vehicle travels based on road width data or the like, and further calculates a center point of the drivable area, which indicates the center of the drivable area, calculates an estimated driving position at which the vehicle is expected to travel, and evaluates a current position candidate that is present on the road segment corresponding to the center point of the drivable area based on the distance between the center point of the drivable area and the estimated driving position and selects a candidate for the current position.which is displayed on a map.

[0006] JP 2008 - 256 620 A relates to a device for correcting map data, a method for correcting map data and a program for correcting map data, and in particular a device for correcting map data, a method for correcting map data and a program for correcting map data for correcting map data relating to a lane. BRIEF SUMMARY OF THE INVENTION

[0007] The purpose of the present disclosure is to provide a route data generation device, a route data generation program and a storage medium capable of generating route data within intersections for automated driving in a suitable manner.

[0008] The problem is solved by the subject matter of the independent claims. Advantageous further developments are specified in the dependent claims.

[0009] According to the invention, the route data is generated by fitting an estimated route of the actual vehicle movement within the intersection into lane network data connected to the intersection, using an absolute path of the vehicle's actual movement within the intersection. This contrasts with a conventional method for generating route data within the intersection using only the absolute path of the vehicle's actual movement within the intersection. By fitting the estimated path into the lane network data using the absolute path, it is possible to optimize the connection between the route data within the intersection and the lane network data connected to the intersection. This makes it possible to generate route data within intersections suitable for automated driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The functions, features, and advantages of this disclosure are more clearly evident from the following detailed description with reference to the accompanying drawings. The drawings show: Fig. 1. A functional block diagram to illustrate an overall configuration of an embodiment; Fig. 2 a functional block diagram of a route data generation unit; Fig. 3 a functional block diagram of a correct route data selector unit; Fig. 4. A flowchart illustrating a route data generation process; Fig. 5. An illustration to demonstrate one way of generating route data; Fig. 6. A figure illustrating one way of specifying lane network data; Fig. 7. An illustration to demonstrate a way of specifying a central coordinate point of an intersection; Fig. 8. An illustration to demonstrate a method for specifying the nearest coordinate point; Fig. 9. An illustration to demonstrate one way of setting a motion trajectory detection segment and a fitting segment; Fig. 10. An illustration to demonstrate one way of calculating an average of distance differences; Fig. 11 a figure to illustrate a way of translating an estimated trajectory around the mean of distance differences; Fig. 12. An illustration to demonstrate a method for calculating a straight-line distance; Fig. 13. An illustration to demonstrate one way of translating the estimated trajectory by the straight-line distance; Fig. 14 an illustration to demonstrate one way of rotating the estimated trajectory; Fig. 15 a figure to illustrate one way of setting an input segment vector; Fig. 16 a figure to illustrate one way of translating the estimated trajectory in one direction of the input segment vector; Fig. 17 a figure to illustrate the estimated trajectory of motion that has undergone a translation in the direction of the input segment vector; Fig. 18 a flowchart to illustrate a correct route data selection process; Fig. 19 An illustration to demonstrate one way of selecting correct route data; Fig. 20 an illustration to demonstrate one way of limiting route data in accordance with vehicle speed; Fig. 21 a figure to illustrate one way of specifying a boundary point of input lane network data; Fig. 22 a figure to illustrate a way of calculating distance differences up to a calculation segment; Fig. 23 an illustration to demonstrate one way of calculating a total distance difference value; Fig. 24 an illustration to demonstrate a way of calculating a total value of distance differences; Fig. 25 a figure to illustrate an extraction method in which route data for which a total value of distance differences satisfies a first predetermined condition are extracted as a candidate for selection; Fig. 26 a diagram illustrating a way of limiting route data in accordance with a curvature; Fig. 27 a figure to illustrate a way of calculating a total value of an input-side distance difference and an output-side distance difference; Fig. 28 a figure to illustrate a selection method in which route data for which a total value satisfies a predetermined condition are selected as correct route data; Fig. 29 an illustration to demonstrate one way of removing an overlapping section; Fig. 30 an illustration to demonstrate a way of judging close passing and / or crossing; Fig. 31 an illustration to demonstrate one way of adding information about a stopping position for a vehicle; and Fig. 32 An illustration to demonstrate one way of adding information about a vehicle's drivable width. MODES FOR EXECUTING THE INVENTION

[0011] One embodiment is described below with reference to the drawings. As shown in Fig. As shown in Figure 1, a route data generation device 1 is a device that generates data about routes within intersections for automated driving and comprises a route data generation unit 2 and a correct route data selection unit 3. The route data generation unit 2 receives inputs of absolute and estimated trajectories recorded by vehicles and inputs of lane network data stored in a lane network data storage unit 4, and generates data about routes within intersections. The absolute trajectory is, for example, a GPS trajectory that describes GPS positions. The estimated trajectory is, for example, a trajectory described by sensor values ​​from a gyroscope. The lane network data is data about routes outside of intersections.The route data generation unit 2 outputs the generated route data and stores this intersection-internal route data in a route data storage unit 5. The correct route data selection unit 3 receives input data on multiple routes within an intersection, which is stored in the route data storage unit 5, and selects the data of one correct route from the data of multiple routes. The correct route data selection unit 3 then outputs the selected correct route data and stores this correct route data in a correct route data storage unit 6.

[0012] As in Fig. As shown in Figure 2, the route data generation unit 2 includes a lane network data specification unit 7, a center point coordinate specification unit 8, a nearest coordinate specification unit 9, a segment setting unit 10, a mean calculation unit 11, a first translation unit 12, a straight-line distance calculation unit 13, a second translation unit 14, an estimated trajectory rotation unit 15, an input track segment vector setting unit 16, and a third translation unit 17. As shown in Fig. As shown in Figure 3, the correct route data selection unit 3 includes a first boundary unit 18, a boundary point specification unit 19, a distance difference calculation unit 20, a first total value calculation unit 21, a candidate selection unit 22, a second boundary unit 23, a second total value calculation unit 24, a selection unit 25, an overlap section distance unit 26, and a density pass / cross evaluation unit 27. These functional blocks are implemented by a microcomputer with a CPU (central processing unit), a ROM (read-only memory), a RAM (random access memory), and an I / O (input / output).The microcomputer executes a computer program stored in a non-volatile tangible storage medium to perform processing according to the computer program and controls the overall operation of the route data generation device 1. The computer program executed by the microcomputer includes a route data generation program.

[0013] First, the individual functional blocks of the route data generation unit 2 are described. The lane network data specification unit 7 reads the lane network data stored in the lane network data storage unit 4. Based on the read lane network data, the lane network data specification unit 7 specifies the input lane network data, which is data from a lane network from which an intersection is entered, and specifies output lane network data, which is data from a lane network onto which the intersection is exited.

[0014] The center point coordinate specification unit 8 specifies an intersection point between a line provided by extending the inbound lane network data and a line provided by extending the outbound lane network data, and specifies the intersection point as the center point of the intersection. The nearest coordinate specification unit 9 captures the absolute trajectory supplied by the vehicle and specifies, from among several coordinate points that comprise the captured absolute trajectory, a coordinate point that is closest to the center point of the intersection specified by the center point coordinate specification unit 8.

[0015] For an input-side absolute motion path, which is an input segment of the absolute motion path, the segment setting unit 10 defines a motion path detection segment and a fitting segment based on the nearest coordinate point specified by the nearest coordinate specification unit 9. For an output-side absolute motion path, which is an output segment of the absolute motion path, the segment setting unit 10 defines a motion path detection segment and a fitting segment based on the nearest coordinate point specified by the nearest coordinate specification unit 9.The mean calculation unit 11 records the estimated trajectory supplied by the vehicle and calculates an average of distance differences between: the fitting segment for the input-side absolute trajectory determined by the segment setting unit 10, and an input-side estimated trajectory which is a section of the estimated trajectory that corresponds to the fitting segment for the input-side absolute trajectory.

[0016] The first translation unit 12 translates (shifts) the estimated path of motion by the mean of the distance differences calculated by the mean calculation unit 11. The straight-line distance calculation unit 13 defines the midpoint of the fitting segment of the input estimated path of motion as a rotation reference point and calculates the straight-line distance from the rotation reference point to the input lane network data. The second translation unit 14 translates (shifts) the estimated path of motion by the straight-line distance calculated by the straight-line distance calculation unit 13.

[0017] The estimated trajectory rotation unit 15 rotates the estimated trajectory to minimize the distance difference between the input estimated trajectory's fitting segment and the input lane network data. The input track segment vector setting unit 16 sets an input track segment vector extending from the start point to the end point of the input estimated trajectory's fitting segment. The third translation unit 17 translates (shifts) the estimated trajectory in the direction of the input track segment vector set by the input track segment vector setting unit 16, minimizing the distance difference between the output estimated trajectory and the output lane network data.

[0018] The following describes the respective functional blocks of the Correct Route Data Selection Unit 3. The first delimitation unit 18 delimits the data from multiple routes according to the vehicle speed. The boundary point specification unit 19 specifies a boundary point of the input lane network data. The distance difference calculation unit 20 selects data from one route from the data of multiple routes as an evaluation target and selects the data of all other routes as comparison target(s). The distance difference calculation unit 20 calculates a distance difference between the evaluation target route data and the comparison target route data at predetermined intervals from the boundary point specified by the boundary point specification unit 19 to a calculation segment.

[0019] The first total value calculation unit 21 calculates a total value of the distance differences calculated for the predetermined intervals. From the multiple route data, the selector extraction unit 22, acting as a selector, extracts the route data for which the total value of the distance differences calculated by the first total value calculation unit 21 satisfies the first predetermined condition. The second delimitation unit 23 delimits the multiple route data according to the route curvature.The second total value calculation unit 24 calculates a total value of an inbound distance difference and an outbound distance difference, where the inbound distance difference is a distance difference between the inbound lane network data and the inbound route data, which is a section of the route data extracted as the candidate from whose section the intersection is entered, and the outbound distance difference is a distance difference between the outbound lane network data and the outbound route data, which is a section of the route data extracted as the candidate to whose section the intersection is exited.

[0020] The selection unit 25 selects correct route data from the data of several routes extracted as the candidate route. This correct route data is that of a route for which the total value of the inbound distance difference and the outbound distance difference, as calculated by the second total value calculation unit 24, satisfies a second predetermined condition. The overlap section removal unit 26 removes sections of the correct route data selected by the selection unit 25. The removed sections are those that overlap the inbound lane network data and the outbound lane network data. The density passing / crossing evaluation unit 27 evaluates the correct route that passes close to and / or crosses another correct route.

[0021] Below is an example of operating the above configuration with reference to the Fig. Sections 4 to 30 are described. The route data generation device 1 executes a route data generation program to perform a route data generation process and a correct route data selection process. The processes are described below. (1) ROAD DATA GENERATION PROCESS

[0022] In the route data generation device 1, the route data generation unit 2 carries out the in Fig. The route data generation process shown in section 4 responds to the fulfillment of a start event of the route data generation process. As shown in Fig. As shown in Figure 5, the route data generation unit 2 receives the following inputs: the absolute and estimated trajectories recorded by the vehicle; and the lane network data stored in the lane network data storage unit; and generates the data for a route within an intersection and outputs the generated route data. The processing performed by each function block of the route data generation unit 2 is described below.

[0023] First, the lane network data specification unit 7 reads the lane network data stored in the lane network data storage unit 4. Based on the read lane network data, the lane network data specification unit 7 specifies how in Fig. Figure 6 shows that from the lane network data describing approach lanes to intersections and exit lanes of the intersections, input lane network data describing an approach lane(s) to a specific intersection, and output lane network data describing an exit lane(s) from the specific intersection that is the destination for generating the driving data (step S1, which corresponds to a lane network data specification process).

[0024] The center point coordinate specification unit 8 is then specified, as in Fig. 7 shown as the central coordinate point of the intersection, an intersection point between an extension line of the inbound lane network data and an extension line of the outbound lane network data, identified by the lane network data specification unit 7 (step S2, which corresponds to a center point coordinate specification process).

[0025] Next, the nearest coordinate specification unit 9 is recorded, as in Fig. Figure 8 shows the absolute trajectory supplied by the vehicle and specifies, from several coordinate points forming the absolute trajectory, one coordinate point closest to the central coordinate point of the intersection as the nearest coordinate point (step S3, corresponding to a nearest coordinate specification process). Specifically, the nearest coordinate specification unit 9, from the GPS coordinate points "N1" to "N7" forming the GPS trajectory, gives GPS coordinate point "N4" as the nearest coordinate point to the central coordinate point of the intersection.

[0026] The segment setting unit 10 then places, as described in the Fig. Figure 9 shows a motion path detection segment and a fitting segment for both an inbound absolute motion path and an outbound absolute motion path, wherein the inbound absolute motion path is a portion of the absolute motion path from which the intersection is entered, and the outbound absolute motion path is a portion of the absolute motion path onto which the intersection is exited (step S4, corresponding to a segment setting process). Specifically, the segment setting unit 10 sets the motion path detection segment "A1" and the fitting segment "B1" of the inbound absolute motion path using the nearest coordinate point as a reference, and sets the motion path detection segment "A2" and the fitting segment "B2" of the outbound absolute motion path using the nearest coordinate point as a reference.The input motion trajectory detection segment "A1" and the output motion trajectory detection segment "A2" are, for example, "50 meters" long, and the input fitting segment "B1" and the output fitting segment "B2" are, for example, "30 meters" long. The input motion trajectory detection segment "A1" and the output motion trajectory detection segment "A2" can be the same length or different lengths. Furthermore, the input fitting segment "B1" and the output fitting segment "B2" can be the same length or different lengths.

[0027] Next, the mean calculation unit 11 records, as in Fig. The segment adjustment unit 10 shows the estimated trajectory supplied by the vehicle and calculates a mean value of the distance difference between the fitting segment of the input-side absolute trajectory, specified by the segment adjustment unit 10, and an input-side estimated trajectory, where the input-side estimated trajectory is a segment of the estimated trajectory corresponding to the fitting segment of the input-side absolute trajectory (step S5, which corresponds to an averaging procedure). Specifically, the averaging unit 11 specifies the input-side estimated trajectory corresponding to the fitting segment "B1" of the input-side absolute trajectory and a mean value of the distances "D1" to "Dn" (n is a natural number) between the fitting segment "B1" of the input-side absolute trajectory and the input-side estimated trajectory.

[0028] The first translation unit then takes 12, as in Fig. Figure 11 shows a translation of the estimated motion path I by the mean of the distance differences calculated by the mean calculation unit 11 (step S6, which corresponds to a first translation process). In particular, the first translation unit 12 performs a translation of the estimated motion path by the mean “Da” of the distance differences “D1” to “Dn”.

[0029] Following this, the straight-line distance calculation unit 13 calculates, as in Fig. Figure 12 shows a straight-line distance from a rotation reference point to the input lane network data, where the midpoint of the fitting segment of the input estimated trajectory is placed on the rotation reference point (step S7, which corresponds to a straight-line distance calculation process). In particular, the Straight-line distance calculation unit 13 calculates the straight-line distance “E” from the rotation reference point to the input lane network data.

[0030] The second translation unit then takes place, 14, as in Fig. Figure 13 shows a translation of the estimated trajectory by the straight-line distance calculated by the straight-line distance calculation unit 13 (step S8, which corresponds to a second translation process). In particular, the second translation unit 14 performs a translation of the estimated trajectory by the straight-line distance “E”.

[0031] Next, the estimated trajectory rotation unit 15 rotates, as shown in Fig. Figure 14 shows the estimated trajectory such that the distance between the fitting segment of the input estimated trajectory and the input lane network data is minimized (step S9, which corresponds to an estimated trajectory rotation procedure). In particular, the estimated trajectory rotation unit 15 rotates the estimated trajectory such that the distance difference “F1” to “Fn” between the fitting segment “B1” of the input estimated trajectory and the input lane network data is minimized.

[0032] The input segment vector setting unit 16 then sets, as in Fig. Figure 15 shows that an input path segment vector is set, extending from the start point to the end point of the fitting segment of the input-side estimated trajectory (step S10, corresponding to an input path segment vector setting process). In particular, the input path segment vector setting unit 16 sets the input path segment vector “G”, extending from the start point to the end point of the fitting segment “B1” of the input-side estimated trajectory.

[0033] Following this, the third translation unit 17 takes place, as in Fig. Figure 16 shows a translation of the estimated trajectory in the direction of the input track segment vector, such that the distance difference between the output estimated trajectory and the output lane network data is minimized (step S11, which corresponds to a third translation process). In particular, the third translation unit 17, as shown in Fig. Figure 17 shows a translation of the estimated trajectory in the direction of the input track segment vector, so that the distance difference “H1” to “Hn” between the output estimated trajectory and the output lane network data is minimized.

[0034] The route data generation unit 2 performs the series of processes described above, wherein the route data generation unit 2 generates the data about the route within the intersection using the absolute path of motion and the estimated path of motion detected by the vehicle and the lane network data stored in the lane network data storage unit 4, and stores the generated route data in the route data storage unit 5. (2) Correct route data dial processing

[0035] In the route data generation device 1, the correct route data selection unit 3 performs the in Fig. Figure 18 shows the correct route data selection process responding to the fulfillment of a start event of the correct route data selection process. As in Fig. As shown in Figure 19, the correct route data selector 3 receives inputs of data from multiple routes stored in the route data storage unit 5, selects data from one of the multiple correct routes, and outputs the selected correct route data. The following describes the processing performed by a respective function block of the correct route data selector 3.

[0036] First, the first restriction unit 18 limits the data of several routes according to the vehicle speed (step S21, which corresponds to a first restriction process). In particular, the first restriction unit 18 excludes, as an exception, the route data generated based on a vehicle with an extremely high vehicle speed or the route data generated based on a vehicle with an extremely low vehicle speed from the data of several routes.

[0037] The boundary point specification unit 19 is then specified, as in Fig. Figure 21 shows a boundary point of the input lane network data (step S22, which corresponds to a boundary point specification process).

[0038] Next, the distance difference calculation unit selects 20, as in the Fig. Figure 22 shows how, from the data of travel paths, the data of one travel path data set is selected as the evaluation target, and the data of all other travel path data sets are selected as the comparison target(s). The distance difference calculation unit 20 calculates, at predetermined distances (intervals) from the boundary point specified by the boundary point specification unit 19 to a calculation segment, the distance difference between the travel path data of the evaluation target and the travel path data of the comparison target (step S23, which corresponds to one distance difference calculation process). The predetermined interval is, for example, "1 meter".

[0039] Subsequently, the first total value calculation unit 21 calculates the total value of the distance differences calculated for the predetermined distances from the boundary point (step S24, which corresponds to a first total value calculation process). The first total value calculation unit 21 changes the evaluation goal and calculates the total value of the distance differences for all of the travel path data, as in the Fig. 23 and Fig. 24 shown.

[0040] Following this, the candidate extraction unit 22 extracts, as in Fig. Figure 25 shows how to extract the data of specific routes as candidate routes from the route data. The extracted candidate route describes the data of a route whose total distance difference value, calculated by the first total value calculation unit 21, fulfills a first predetermined condition (step S25, which corresponds to a candidate route extraction process). For example, the first predetermined condition is such that the candidate routes belong to 10% of the route data and each has a total distance difference value smaller than that of the remaining 90% of the routes.

[0041] The second boundary unit 23 then borders, as in Fig. Figure 26 shows the route data extracted by the candidate extraction unit 22 as the candidates, with the narrowing down being performed in accordance with a route curvature (step S26, which corresponds to a second narrowing process). In particular, the second narrowing unit 23, as an exception, excludes from the route data the route data whose curvature partially exceeds a threshold, thus narrowing down the route data extracted as the candidates.

[0042] Next, the second boundary unit calculates 23, as in Fig. 27 shows, using the route data provided by the delimitation by the second delimitation unit 23, a total value of an inbound distance difference and an outbound distance difference, where the inbound distance difference is a distance difference between the data of inbound route data (of an inbound route data set) and the inbound lane network data on the inbound side of the intersection, and the outbound distance difference is a distance difference between the data of outbound route data (of an outbound route data set) and the outbound lane network data on the outbound side of the intersection (step S27, which corresponds to a second total value calculation process).

[0043] Next, the dialing unit dials 25, as in Fig. Figure 28 shows how, from the route data selected as the candidates, the data of a correct route is selected. This correct route is the data of a route whose total value of the input-side distance difference and the output-side distance difference fulfills a second predetermined condition (step S28, which corresponds to a selection process). The second predetermined condition is, for example, such that the total value of the input-side distance difference and the output-side distance difference is the smallest.

[0044] The overlap section removal unit 26 then removes, as shown in Fig. Figure 29 shows sections of the correct route data selected by the selector unit 25, wherein the removed sections are sections that overlap the input lane network data and the output lane network data (step S29, which corresponds to an overlap section removal process). In particular, the overlap section removal unit 26 smooths a junction point between the correct route data and the input lane network data and a junction point between the correct route data and the output lane network data.

[0045] Subsequently, the Density Passing / Crossing Evaluation Unit 27 assesses a correct path that passes close to and / or crosses another correct path (step S29, which corresponds to an overlap section removal process). Specifically, the Density Passing / Crossing Evaluation Unit 27 determines a vehicle collision possibility by assessing a correct path that passes close to and / or crosses another correct path.

[0046] The correct route data selector 3 performs the series of processes described above, wherein the correct route data selector 3 selects the correct route data from the data of several routes stored in the route data storage unit 5 and stores the selected correct route data in the correct route data storage unit 6.

[0047] The embodiment described above achieves the following effects. To generate the route data, the route data generation device 1 fits the estimated trajectory of the vehicle, as it actually travels through the intersection, into the lane network data associated with the intersection using the absolute trajectory. By fitting the estimated trajectory of the vehicle's actual movement within the intersection into the lane network data using the absolute trajectory, it is possible to optimize the connection between the route data within the intersection and the lane network data associated with the intersection. This makes it possible to generate route data within intersections for automated driving in a suitable manner.

[0048] Furthermore, the route data generation device 1 selects the data from one of several routes as the correct route data. In particular, when data from multiple routes is generated, not all routes may be suitable, and there is a possibility that an incorrect route will be generated. However, by selecting the correct route data, it is possible to choose suitable route data.

[0049] Although the present disclosure is described above with reference to its embodiments, it should be noted that it is not limited to embodiments and structures. The present disclosure encompasses various examples and modifications within a range of equivalence. Furthermore, various combinations and forms, as well as other combinations and forms containing only a single element, more elements, or fewer elements, also fall within the scope and spirit of the present disclosure.

[0050] The process for narrowing down the data of multiple routes according to vehicle speed by the first narrowing unit 18 and / or the process for narrowing down the data of multiple routes according to route curvature by the second narrowing unit 23 can be executed at desired timings. In particular, the process for narrowing down the routes according to route curvature can be performed immediately after the process for narrowing down the data according to vehicle speed. These processes can be omitted.

[0051] When the correct route data is selected, the Correct Route Data Selector Unit 3 can add additional information to the correct route data. Specifically, the Correct Route Data Selector Unit 3, as shown in Fig. Figure 31 shows how, using vehicle speeds or vehicle speed data as criteria for narrowing down the route data in accordance with the vehicle speed, a stopping position of the vehicle within the intersection can be specified. The correct route data selector unit 3 can add information about the specified stopping position of the vehicle to the correct route data.

[0052] As in Fig.As shown in Figure 32, the correct route data selector 3 can specify, as a vehicle-travelable width, a range between the data of a route that runs furthest inward at the intersection and the data of a route that runs furthest outward at the intersection, among the data of multiple routes, and add the information about the specified vehicle-travelable width to the correct route data. By providing a margin, the correct route data selector 3 can specify a vehicle-travelable width smaller than the range between the innermost route data (of the innermost path) and the outermost route data (of the outermost path). Furthermore, the correct route data selector 3 can specify only the inner traversable range based on the innermost route data and add information about the specified inner traversable range to the correct route data.Alternatively, only the outer drivable area can be specified by the outermost route data, and information regarding the specified outer drivable area can be added to the correct route data.

Claims

[1] Route data generation device for generating route data within an intersection for automated driving, comprising: - a route data generation unit (2) that generates the route data such that the route data generation unit (2), using an absolute motion path of actual driving of a vehicle within the intersection, fits an estimated motion path of the actual driving of the vehicle within the intersection into lane network data associated with the intersection, wherein - the route data generation unit (2) has: - a lane network data specification unit (7) that specifies inbound lane network data describing a lane network from which the intersection is entered and outbound lane network data describing a lane network onto which the intersection is exited; - a center point coordinate specification unit (8) that specifies an intersection point between an extension line of the inbound lane network data and an extension line of the outbound lane network data as a central coordinate point of the intersection; - a nearest coordinate specification unit (9) which, as a nearest coordinate point, specifies a coordinate point that is closest to the central coordinate point of the intersection from several coordinate points that form the absolute trajectory; - a segment setting unit (10) which defines a motion path detection segment and a fitting segment based on the nearest coordinate point for both an input-side absolute motion path and an output-side absolute motion path, wherein the input-side absolute motion path is an input-side section of the absolute motion path and the output-side absolute motion path is an output section of the absolute motion path; - a mean calculation unit (11) that calculates a mean of distance differences between the fitting segment of the input-side absolute motion path and an input-side estimated motion path, wherein the input-side estimated motion path is a section of the estimated motion path corresponding to the fitting segment of the input-side absolute motion path; - a first translation unit (12) that performs a translation of the estimated trajectory by the mean of the distance differences; - a straight-line distance calculation unit (13) that calculates a straight-line distance from a rotation reference point to the input lane network data, wherein the rotation reference point is defined as an intermediate point of the fitting segment of the input estimated trajectory; - a second translation unit (14) that performs a translation of the estimated trajectory by the straight-line distance; - an estimated motion path rotation unit (15) that rotates the estimated motion path such that a distance difference between the fitting segment of the input estimated motion path and the input lane network data is minimized; - an input path segment vector setting unit (16) that specifies an input path segment vector extending from a starting point to an endpoint of the fitting segment of the input-side estimated trajectory; and - a third translation unit (17) that performs a translation of the estimated motion path in the direction of the input track segment vector such that a distance difference between the output estimated motion path and the output lane network data is minimized. [2] Route data generation device according to claim 1, further comprising a correct route data selection unit (3) which selects the data of a correct route as correct route data from the data of several routes generated by the route data generation unit (2). [3] Route data generation device according to claim 2, wherein the correct route data selection unit (3) comprises: - an edge point specification unit (19) that specifies an edge point of the input lane network data; - a distance difference calculation unit (20) which selects data from one route as an evaluation target and data from all other routes as a comparison target from the data of the multiple routes and calculates distance differences between the evaluation target and the comparison target at predetermined intervals from the boundary point to a calculation segment; - a first total value calculation unit (21) that calculates a total value of the distance differences calculated from the boundary point in the predetermined intervals; - a candidate selection unit (22) which extracts from the route data, as a candidate selection unit, the data of a route whose total value of the distance differences satisfies a first predetermined condition; - a second total value calculation unit (24) that calculates a total value of: an input-side distance difference between the input-side lane network data and input-side route data representing an intersection-inside section of the route data selected as the candidate; and an output-side distance difference between the output-side lane network data and output-side route data representing an intersection-outside section of the route data selected as the candidate; - a selection unit (25) which selects from the route data selected as the selectors the data of a correct route which represents the data of a route whose total value of the inbound distance difference and the outbound distance difference satisfies a second predetermined condition; - an overlap section removal unit (26) that removes the data of a section of the correct lane data, wherein the section overlaps the inbound lane network data and the outbound lane network data; and - a density passing / crossing assessment unit (27) that assesses the correct route data that pass close to and / or cross other correct route data. [4] Route data generation device according to claim 2 or 3, wherein the correct route data selection unit (3) has a first limitation unit (18) that limits the data from multiple routes in accordance with the vehicle speed. [5] Route data generation device according to one of claims 2 to 4, wherein the correct route data selection unit (3) has a second limiting unit (23) that limits the data from several routes in accordance with a curvature. [6] Route data generation device according to one of claims 2 to 4, wherein when the correct route data is selected, the correct route data selection unit (3) adds additional information to the correct route data. [7] Route data generation program for a route data generation device to generate route data within an intersection for automated driving, wherein the program causes the device to execute: - a lane network data specification process for specifying inbound lane network data that describes a lane network from which the intersection is entered, and outbound lane network data that describes a lane network onto which the intersection is exited; - a center point coordinate specification process to specify an intersection point between an extension line of the input lane network data and an extension line of the output lane network data as a central coordinate point of the intersection; - a nearest coordinate specification process to specify as a nearest coordinate point, a coordinate point that is closest to the central coordinate point of the intersection, from several coordinate points that form the absolute trajectory; - a segment setting process to define a motion path detection segment and a fitting segment based on the nearest coordinate point for both an input-side absolute motion path and an output-side absolute motion path, wherein the input-side absolute motion path is an input-side section of the absolute motion path and the output-side absolute motion path is an output section of the absolute motion path; - a mean calculation process for calculating a mean value of distance differences between the fitting segment of the input-side absolute motion path and an input-side estimated motion path, wherein the input-side estimated motion path is a section of the estimated motion path corresponding to the fitting segment of the input-side absolute motion path; - a first translation process to perform a translation of the estimated trajectory around the mean of the distance difference; - a straight-line distance calculation process for calculating a straight-line distance from a rotation reference point to the input lane network data, wherein the rotation reference point is defined as an intermediate point of the fitting segment of the input estimated trajectory; - a second translation process to perform a translation of the estimated trajectory by the straight-line distance; - an estimated trajectory rotation process to rotate the estimated trajectory such that a distance difference between the fitting segment of the input estimated trajectory and the input lane network data is minimized; - an input path segment vector setting process for determining an input path segment vector that extends from a starting point to an endpoint of the fitting segment of the input-side estimated trajectory; and - a third translation process to translate the estimated trajectory in the direction of the input track segment vector in such a way that the distance difference between the output estimated trajectory and the output lane network data is minimized. [8] Route data generation program according to claim 7, wherein the program causes the device to execute: - a boundary point specification process for specifying a boundary point of the input lane network data; - a distance difference calculation process for selecting, from the data of several routes, data of one route as an evaluation target and for selecting data of the remaining routes as a comparison target and for calculating, in predetermined intervals from the boundary point to a calculation segment, distance differences between the evaluation target and the comparison target; - an initial total value calculation process to calculate a total value of the distance differences calculated from the boundary point in the predetermined intervals; - a candidate selection process for extracting from the data of routes, as a candidate that has data of a route whose total value of distance differences satisfies a first predetermined condition; - a second total value calculation process to calculate a total value of: an input-side distance difference between the input-side lane network data and input-side route data representing an intersection-inside section of the route data selected as the candidate; and an output-side distance difference between the output-side lane network data and output-side route data representing an intersection-outside section of the route data selected as the candidate; - a selection process for selecting from the data of routes selected as the candidates, the data of a correct route, which represents the data of a route whose total value of the inbound distance difference and the outbound distance difference satisfies a second predetermined condition; - an overlap section removal process for removing the data of a section of the correct lane data, wherein the section to be removed is a section that overlaps the inbound lane network data and the outbound lane network data; and - a proximity / crossing assessment process to assess the correct route data that closely passes and / or crosses other correct route data. [9] Computer-readable, non-volatile storage medium that stores the route data generation program according to claim 7 or 8.

Images