Course generation method and course generation device

Abstract

To appropriately generate a course.SOLUTION: A controller 100 extracts a division line of a road from detection information acquired by an image capturing section 220 on an area ahead of a vehicle 1 in a travel direction, and generates a course along which the vehicle 1 travels, on the basis of the extracted division line. The course can be updated according to traveling of the vehicle 1. Before the vehicle 1 enters a crossing, the controller 100 updates the course on the basis of a division line up to an entrance of the crossing and a division line following an exit of the crossing. The controller does not update the course in a crossing section from the entrance of the crossing up to the exit of the crossing, and updates the course in a single road section up to the entrance of the crossing and a single road section following the exit of the crossing, on the basis of respective division lines of the single road sections.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a route generation method and a route generation device. [Background technology]

[0002] A driving path generation system has been proposed that, when a vehicle turns, acquires information about the lane centerline of the driving path along which the vehicle will turn, generates multiple clothoid curves whose start and end points are located on the acquired lane centerline, and generates a target driving path based on the clothoid curve that minimizes the integral value of the error indicating the difference between the clothoid curve and the lane centerline (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-160625 Summary of the Invention [Problem to be solved by the invention]

[0004] Patent Document 1 describes that information on lane centerlines is included in map information stored in a navigation system, and that an ECU (Electronic Control Unit) acquires the information on lane centerlines from the navigation system. Therefore, with the technology described in Patent Document 1, if the map information stored in the navigation system does not include information on the vehicle centerline, there is a risk that the technology will not be able to generate a suitable route.

[0005] The present invention has been made in view of the above circumstances, and aims to provide a lane generation method and the like that can generate lane appropriately. [Means for solving the problem]

[0006] To achieve the above object, the present invention provides a lane generation method that acquires the current position of a vehicle, extracts road markings from detection information ahead of the vehicle in the direction of travel acquired by a sensor equipped on the vehicle, and generates a lane for the vehicle to travel on based on the extracted markings. The lane can be updated as the vehicle travels, and before the vehicle enters an intersection, the lane is updated based on the markings up to the entrance of the intersection and the markings after the exit of the intersection. The lane is not updated in the intersection section from the entrance to the intersection to the exit of the intersection, but is updated in the single lane section up to the entrance of the intersection and the single lane section after the exit of the intersection based on the markings of the single lane section. [Effects of the Invention]

[0007] According to the present invention, a running path can be generated suitably. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a block diagram illustrating an example of a functional configuration of a route generation device according to an embodiment of the present invention. [Figure 2] FIG. 2 illustrates an example of a hardware configuration of a controller according to an embodiment. [Figure 3] FIG. 2 is a diagram showing an example of map information and a driving route. [Figure 4] FIG. 10 is a diagram illustrating an example of reference routes and classification of sections. [Figure 5] 4 is a flowchart illustrating an example of a vehicle control process. [Figure 6A] FIG. 10 is a diagram illustrating an example of acquiring a lane marking; [Figure 6B] FIG. 1 is a diagram showing track boundaries and tracks. [Figure 7A] FIG. 10 is a diagram illustrating an example of acquiring a lane marking; [Figure 7B] FIG. 1 is a diagram showing track boundaries and tracks. [Figure 8] FIG. 10 is a diagram showing input information in each section. [Figure 9] 10A and 10B are diagrams showing input information for each section type in a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0009] A route generation device and a route generation method executed by the route generation device according to an embodiment of the present invention will be described with reference to the drawings. In the drawings, the same or equivalent parts are designated by the same reference numerals.

[0010] The path generation device 10 according to the embodiment of the present invention is a device that generates a path along which a vehicle 1 (host vehicle) travels on a road based on input information. The path generation device 10 also functions as a vehicle control device that inputs control data based on the information about the generated path to actuators 240 that drive each part of the vehicle 1. The path generation device 10 thereby controls the automatic driving and driving assistance of the vehicle 1.

[0011] The vehicle 1 is a passenger vehicle such as an electric vehicle, hybrid vehicle, or gasoline vehicle, and has an autonomous driving function that automatically drives along a route generated by the route generation device 10. In this embodiment, the route generation device 10 controls, for example, level 4 autonomous driving, but other levels may also be used. Note that the route generation device 10 is not limited to controlling autonomous driving, and may also control driving assistance such as accelerator assistance, brake assistance, and steering assistance.

[0012] Also, a vehicle control device may be provided separately from the path generation device 10, and the vehicle control device may control the automatic driving of the vehicle 1 based on the path generated by the path generation device 10. In other words, the path generation device 10 may be a device that generates a path for automatic driving based on input information.

[0013] 1 is a diagram showing an example of the functional configuration of a path generation device 10 provided in a vehicle 1 according to this embodiment. The vehicle 1 according to this embodiment includes the path generation device 10 and an actuator 240. The vehicle 1 also includes an engine, a body, a chassis, a drive train, electrical parts, etc., but a description of the general configuration of these components will be omitted.

[0014] 1, the path generation device 10 includes a controller 100 that generates a path for the vehicle 1, a position information acquisition unit 210 that acquires position information of the vehicle 1, an imaging unit 220 that captures images of at least the area ahead of the vehicle 1, and a map database 230 that stores map information (map data) including information about the roads on which the vehicle 1 is traveling. Note that the position information acquisition unit 210, the imaging unit 220, and the map database 230 may be configured to be shared with other systems or devices such as a navigation system.

[0015] The position information acquisition unit 210 is any device, such as a GNSS (Global Navigation Satellite System) receiver, that can acquire the position of the vehicle 1. The GNSS receiver receives orbit information and time information from a plurality of positioning satellites, and outputs position information indicating the position of the vehicle 1 calculated based on the received signals to the controller 100.

[0016] The imaging unit 220 is, for example, one or more cameras that capture images of at least the area ahead of the vehicle 1, and each camera is installed in a position where it can capture images of the lane markings (white lines) of the lane in which the vehicle 1 is traveling and the marking lines (white lines) near the intersection ahead of the vehicle 1. The imaging unit 220 inputs image data of the captured images to the controller 100. Note that the imaging unit 220 may be provided so as to be able to capture images of the surroundings (rear, etc.) other than the area ahead of the vehicle 1. Note that the imaging unit 220 is an example of a sensor that acquires information ahead of the vehicle 1. The image captured by the imaging unit 220 is an example of information detected by a sensor.

[0017] The map database 230 stores map information (map data) including road information and intersection information, and is a database used, for example, in a navigation system. The map information stored in the map database 230 includes at least information on roads, such as roads on a map, such as roads and intersections. A road is a continuous section of road without intersections or railroad crossings, and in this embodiment, it also refers to roads other than intersections. In this embodiment, the map information is assumed to be data with a graph structure that represents roads on a map using graph points indicating characteristic points on the road, such as intersections and T-junctions, and graph lines indicating roads. For roads with opposing roads or medians, the graph line indicating a road is represented as two roads in each direction. Furthermore, the graph line indicating a road is a single graph line indicating the center of the road, regardless of the number of lanes. Therefore, the map information handled in this embodiment does not include information on the center lines of each lane or the number of lanes. As such, the map information handled in this embodiment is composed of the minimum information required for the navigation system to navigate a route. Therefore, when the travel route of the vehicle 1 is shown on the map information of this embodiment, the travel route is described as a series of graph lines.

[0018] The controller 100 is a control device that controls the operation of the path generation device 10. The controller 100 generates a path on a road for the vehicle 1 based on the position information acquired by the position information acquisition unit 210, the map information in the map database 230, a driving route searched by a navigation system or the like, and an image of the area ahead of the vehicle captured by the imaging unit 220. Then, the controller 100 controls the automatic driving of the vehicle 1 based on the generated path.

[0019] Fig. 2 is a diagram showing an example of the hardware configuration of the controller 100. In the example of Fig. 2, the controller 100 includes a processor 1011, a memory 1012, a storage 1013, and a communication interface (referred to as communication I / F in the figure) 1014, which are connected to each other via a bus 1010.

[0020] The processor 1011 includes, for example, one or more CPUs (Central Processing Units) and their peripheral circuits, and executes various types of arithmetic processing. The processor 1011 loads a control program stored in a storage 1013 into a memory 1012 and executes the program. The processor 1011 may further include arithmetic circuits such as a logical operation unit and a numerical operation unit.

[0021] The memory 1012 includes, for example, a volatile semiconductor memory such as a RAM (Random Access Memory), and functions as a work memory for the processor 1011. The memory 1012 also temporarily stores the control program that the processor 1011 reads from the storage 1013 and various data used in the processor 1011's arithmetic processing.

[0022] The storage 1013 includes, for example, a nonvolatile semiconductor memory such as an EEPROM (Electrically Erasable and Programmable Read Only Memory), a flash memory, etc. The storage 1013 stores the control program executed by the processor 1011 and various data used in the arithmetic processing of the processor 1011.

[0023] The communication interface 1014 includes an interface circuit for connecting the controller 100 to an in-vehicle network that complies with standards such as CAN (Controller Area Network). The communication interface 1014 receives signals from the position information acquisition unit 210, the imaging unit 220, the map database 230, and other in-vehicle components, and passes the signals to the processor 1011.

[0024] Furthermore, the communication interface 1014 transmits the vehicle control signal generated by the processor 1011 to the actuator 240 that operates the vehicle 1. The vehicle control signal is generated based on information necessary for vehicle control, such as information about surrounding moving objects, traffic light information, and traveling speed information, in addition to information about the route generated by the route generation device 10 of this embodiment.

[0025] Returning to FIG. 1 , the actuator 240 is communicatively connected to the trajectory generation device 10 via an in-vehicle network, and the trajectory generation device 10 controls the actuator 240. The actuator 240 includes, for example, a drive device (at least one of an engine and a motor) for accelerating the vehicle 1, a brake actuator for braking the vehicle 1, a steering motor for steering the vehicle 1, etc. In this way, the trajectory generation device 10 controls the actuator 240 to realize automatic driving or driving assistance of the vehicle 1.

[0026] In the controller 100, for example, a processor 1011 and a memory 1012 cooperate to realize the functions shown in Fig. 1. That is, functionally, the controller 100 includes a driving route acquisition unit 101, a map information acquisition unit 102, a reference route generation unit 103, a route section classification unit 104, a vehicle section classification determination unit 105, a lane marking extraction unit 106, a lane boundary generation unit 107, a lane generation unit 108, and a vehicle control unit 109. Note that the functional configuration of the controller 100 in this embodiment is an example, and the functional units may be integrated or subdivided as desired.

[0027] The driving route acquisition unit 101 acquires a driving route from a navigation system (not shown), a user's smartphone, or the like connected to the controller 100. Therefore, the driving route acquisition unit 101 is connected to the navigation system, smartphone, or the like by wire or wirelessly. The driving route may be, for example, one that is searched for by a map application or the like installed in the navigation system or smartphone based on the current position and a destination specified by the user.

[0028] The map information acquisition unit 102 acquires map information for the area around the current location. The map information acquisition unit 102 also acquires, from the map database 230, map information for an area that includes at least the travel route acquired by the travel route acquisition unit 101.

[0029] The reference route generation unit 103 generates a reference route indicating the route length and direction of travel of the road along which the vehicle 1 will travel, based on the driving route acquired by the driving route acquisition unit 101 and the position information of the vehicle 1 acquired by the position information acquisition unit 210.

[0030] As shown in Fig. 3, the map information of this embodiment is made up of graph lines indicating a plurality of roads (main roads) and graph points indicating intersections and T-junctions. In the example shown in Fig. 3, main roads r1 to r5, etc. are shown. Also, intersections c1 to c3, etc. are shown. For example, main roads r1 to r4 are roads with two-way traffic or a median strip, and are shown with two lines. Also, main road r5 is a one-way road, and is shown with one line. Intersections c1 to c3 are shown with points indicating the intersections of the main roads and lines connecting each point (lines crossing the intersections).

[0031] Moreover, Fig. 3 shows an example of a driving route on map information. The driving route indicated by the thick line in Fig. 3 is a route that passes through single roads r1 to r4 from the current position, goes straight at intersection c1, turns left at intersection c2, and turns right at intersection c3. As shown in Fig. 3, the driving route on map information shows right and left turn points as linear. Therefore, it differs from the actual behavior, traveling direction, and trajectory of the vehicle 1 when turning right or left. In this embodiment, in order to generate a driving path for autonomous driving of the vehicle 1, first a reference route indicating the traveling direction (azimuth) of the vehicle 1 is generated based on the driving route on map information.

[0032] For straight sections of the travel route other than the right or left turn points, the direction indicated by the travel route can be estimated as the traveling direction (azimuth) of the vehicle 1, so for example, the reference route generation unit 103 applies the travel route as is to the reference route. Next, the reference route generation unit 103 identifies the start point (entrance) of the intersection at the right or left turn point on the travel route and the end point (exit) of the intersection. Then, by drawing a curve or arc with a predetermined curvature connecting the start point and end point of the intersection, a reference route is generated that indicates the traveling direction (azimuth) of the vehicle 1 at the intersection (right or left turn point). The curvature in this case may be a value based on statistical data, for example.

[0033] Fig. 4 shows a reference route generated at intersection c2. The reference route generation unit 103 identifies a start point S3 and an end point S4 of intersection c2 based on the distance from the intersection c2, etc., and generates a reference route by drawing a curve with a predetermined curvature that connects the start point S3 and the end point S4, as shown in Fig. 4.

[0034] Furthermore, for example, the reference route generating unit 103 may generate a reference route for an intersection by determining a control point and drawing a Bezier curve based on the azimuth angle of the main road on the approaching side of the intersection (intersection approach attitude), the azimuth angle of the main road on the exiting side of the intersection (intersection exit attitude), the intersection point of the two straight lines formed by the intersection approach attitude and the intersection exit attitude, etc.

[0035] The reference route generating unit 103 may also generate a reference route by correcting the travel route for a straight intersection on the travel route by estimating the trajectory of the vehicle 1 within the intersection based on the attitude of entering the intersection, the attitude of leaving the intersection, etc. Furthermore, when two main roads are adjacent to each other on the travel route, the reference route may be generated by correcting the connection between the main roads. Furthermore, a well-known method may be used to generate the reference route.

[0036] Returning to FIG. 1 , the route section classification unit 104 classifies the reference route into a plurality of sections. The lane generation device 10 of this embodiment updates the lane boundary model and the lane depending on the classification of the section to which the current position of the vehicle 1 belongs. Therefore, in this embodiment, the route section classification unit 104 classifies the reference route generated by the reference route generation unit 103 into a plurality of sections. The route section classification unit 104 classifies the reference route into either a single road section or an intersection section. Alternatively, the route section classification unit 104 classifies the area around each intersection on the reference route into either a first section, a second section, a third section, or a fourth section. In this way, by classifying the reference route into a plurality of sections, processing can be realized to generate a lane according to the section.

[0037] Specifically, as shown in FIG. 4, the route section classification unit 104 determines points S1 to S5 for the intersection to be entered (intersection c2 in this example). The determined positions S1 to S5 are stored, for example, as the value of the route length S from the travel start point. Here, S3 is the start point of the intersection, and S4 is the end point of the intersection. S2 and S5 are points for determining sections (sections S2 to S3 and S4 to S5) where it is assumed that there are no lane markings due to the presence of a crosswalk or the like near the intersection. S1 is a point that becomes a single road section just before entering the intersection, and is a point for determining a section (section S1 to S2) where it is assumed that lane markings can be acquired from the captured image after leaving the intersection. The distance from the travel start point to point S1 is also referred to as the first distance. The distance from the travel start point to point S2 is also referred to as the second distance. The distance from the travel start point to point S5 is also referred to as the third distance. In this embodiment, the input information when generating a route and whether or not to update the route are switched based on the first distance, the second distance, and the third distance.

[0038] For example, the route section classification unit 104 determines as S2 a point that is a first predetermined distance (e.g., 5 meters) from S3 in the opposite direction to the traveling direction, and determines as S2 a point that is a second predetermined distance (e.g., 5 meters) from S4 in the traveling direction of the vehicle 1. Next, the route section classification unit 104 determines as S1 a point that is a third predetermined distance (e.g., 3 meters) from S2 in the opposite direction to the traveling direction.

[0039] The first, second, and third predetermined distances may be the same or different. At least one of the first, second, and third predetermined distances may be determined according to the route length of the main road (main roads r2 and r3 in FIG. 4) connecting to the intersection. For example, at least one of the first, second, and third predetermined distances may be a predetermined percentage of the route length of the main road.

[0040] In this way, the route section classification unit 104 determines points S1 to S5 for each intersection on the reference route. Then, the route section classification unit 104 may store the route lengths S1 to S5 at each intersection, for example, at a given intersection, S1 is a point with a route length of 1000 meters, S2 is a point with a route length of 1003 meters, S3 is a point with a route length of 1008 meters, S4 is a point with a route length of 1020 meters, and S5 is a point with a route length of 1025 meters. In this way, by managing points S1 to S5 by the route length of the reference route, each section can be easily determined.

[0041] At a given intersection, the section from just before the intersection to point S1 is the first section in this embodiment, which is the single road section. The section from point S1 to point S2 is the second section in this embodiment, which is the section immediately before entering the intersection. The section from point S2 to point S5 is the third section in this embodiment, which is the intersection section. Furthermore, the section after point S5 is the fourth section in this embodiment, which is the single road section. In this way, the route section classification unit 104 classifies the reference route around the intersection into a plurality of sections by determining points S1 to S5 around the intersection.

[0042] Returning to FIG. 1, the vehicle section classification determination unit 105 determines, while the vehicle 1 is traveling, which section of the sections classified by the route section classification unit 104 the current position of the vehicle 1 belongs to.

[0043] The lane marking extraction unit 106 detects lane markings (white lines, yellow lines, etc.) on the road from the captured image of a predetermined range ahead of the vehicle 1 captured by the imaging unit 220 using known methods such as edge extraction and pattern matching, and further extracts characteristic points of the lane markings. The lane marking extraction unit 106 then identifies the position of the lane markings by calculating the relative position of the extracted characteristic points of the lane markings from the vehicle 1, for example. The position of the lane markings is identified as coordinates in a Cartesian coordinate system with latitude and longitude or the start point of travel as the origin. The lane marking extraction unit 106 inputs information on the extracted characteristic points of the lane markings to the lane boundary generation unit 107.

[0044] In this way, the lane marking extraction unit 106 can detect the lane markings of the lane in which the vehicle is traveling. The lane marking extraction unit 106 also detects the lane markings of the intersection ahead after the vehicle has left the intersection just before entering the intersection. The lane marking extraction unit 106 then adds information about the characteristic points of the lane markings after the vehicle has left the intersection to the input information to the lane boundary generation unit 107.

[0045] The lane boundary generation unit 107 converts the Cartesian coordinates of the characteristic points of the lane lines extracted by the lane line extraction unit 106 into a curvilinear coordinate system based on the reference route. Then, the lane boundary generation unit 107 generates or updates a lane boundary model that indicates the boundaries of the lane based on the coordinates of the converted characteristic points of the lane lines in the curvilinear coordinate system. In this embodiment, the lane boundary model refers to information including left and right boundary lines that the vehicle 1 must not cross (boundaries that are dangerous if crossed) while the vehicle 1 is traveling in autonomous driving mode. The lane boundary model may also include information about the median lines between the left and right boundary lines. A specific method for generating the lane boundary model will be described later.

[0046] Each characteristic point of the lane marking has coordinates (X, Y) in a Cartesian coordinate system when extracted by the lane marking extraction unit 106. The reference route also has coordinate information in the Cartesian coordinate system according to the route length from the starting point (for example, the starting point of travel). The Cartesian coordinates (X, Y) in the Cartesian coordinate system can be expressed as curvilinear coordinates (S = route length from the starting point to the coordinate (X, Y), D = distance in the normal direction from the reference route to the coordinate (X, Y)) in a curvilinear coordinate system based on the reference route. Alternatively, the position vector X(s) in the Cartesian coordinate system is expressed by the following equation (1): r(s), n r (s) and d(s) respectively indicate the reference path in the path length, the unit normal vector on the reference path, and the signed distance from the reference path in the normal direction to the position vector.

[0047]

number

[0048] Based on this relationship, the lane boundary generating unit 107 converts the Cartesian coordinates of the characteristic points of the lane markings into a curvilinear coordinate system based on the reference route.

[0049] The lane boundary generation unit 107 generates the lane boundary model when generating the lane boundary model for the first time for a given driving route, and updates the lane boundary model based on new input information in other cases.

[0050] Furthermore, the lane boundary generation unit 107 can generate a lane boundary model using a reference route as input information. For example, in a section where lane markings have not been detected, the lane boundary generation unit 107 estimates that the azimuth angles of the reference route and the lane boundary model match, and generates a lane boundary model based on the reference route and known lane width information. In this way, the lane boundary generation unit 107 can generate a lane boundary model based on multiple pieces of input information.

[0051] The lane generation unit 108 generates a lane for the vehicle by converting the lane boundary model in the curvilinear coordinate system generated by the lane boundary generation unit 107 into a Cartesian coordinate system (e.g., information indicating latitude and longitude), and updates the lane as needed while the vehicle is traveling. For example, the lane generation unit 108 generates or updates information obtained by converting the positions of the center lines of the left and right boundary lines indicated by the lane boundary model into latitude and longitude as the lane. The conversion of the lane boundary model from the curvilinear coordinate system to the Cartesian coordinate system can be performed, for example, using an algorithm that is the reverse of the conversion from the Cartesian coordinate system to the curvilinear coordinate system performed by the lane line extraction unit 106.

[0052] The lane generation unit 108 generates the lane when generating the lane for the first time for a predetermined travel route, and updates the lane based on the updated lane boundary model in other cases.

[0053] The vehicle control unit 109 generates and outputs control signals for controlling each actuator of the vehicle 1 based on the route generated by the route generation unit 108. This allows the vehicle 1 to drive automatically along the travel route set for the vehicle 1. In this embodiment, the vehicle control unit 109 is included in the route generation device 10, but it may also be provided outside the route generation device 10. In this case, the vehicle control unit receives data on the route generated by the route generation device 10 and controls the actuators based on the data on the route.

[0054] In this embodiment, the controller 100 determines whether to update the lane boundary model using the lane boundary generation unit 107 depending on the section to which the vehicle 1 belongs, as determined by the host vehicle section classification determination unit 105, and varies the input information to the lane boundary generation unit 107. That is, the lane generation device 10 determines whether to update the lane depending on the section to which the vehicle 1 belongs, and updates the lane based on the different input information. Before the vehicle 1 enters an intersection, the lane generation device 10 updates the lane based on the lane markings up to the entrance of the intersection and the lane markings after the exit of the intersection. The lane generation device 10 does not update the lane at the intersection from the entrance to the exit of the intersection, but updates the lane based on the lane markings of the single lane section up to the entrance of the intersection and the single lane section after the exit of the intersection. In this way, lane updates can be achieved depending on the section to which the vehicle 1 belongs.

[0055] In this manner, the controller 100 of this embodiment acquires a driving route, generates a reference route based on the acquired driving route, extracts road lane markings from the captured image, generates a lane boundary model indicating the boundaries of the lane based on the coordinates of feature points of the extracted lane markings in a curvilinear coordinate system that is based on the reference route, and generates a lane by converting the generated lane boundary model into a Cartesian coordinate system. This method of generating a lane by the controller 100 is one example. For example, the controller 100 may at least extract lane markings from detection information ahead of the vehicle acquired by a sensor, and generate a lane along which the vehicle 1 will travel based on the extracted lane markings.

[0056] The operation of the path generation device 10 configured as above will be described in detail with reference to the flowchart in Fig. 5. Fig. 5 is a flowchart showing an example of vehicle control processing executed when the vehicle 1 is traveling. The vehicle control processing shown in Fig. 5 mainly shows processing for generating a path, but other controls for automatic driving, such as control for following a vehicle ahead, automatic braking control, and signal judgment control, may be executed in parallel with this processing.

[0057] 5 is started when the vehicle 1 starts up or begins to travel. First, the controller 100 acquires the current position of the vehicle 1 (host vehicle position) based on the output signal of the position information acquisition unit 210 (step S101).

[0058] Next, the map information acquisition unit 102 of the controller 100 acquires map information from the map database 230 based on the current position of the vehicle 1 acquired in step S101 (step S102).

[0059] Next, the travel route acquisition unit 101 acquires the travel route of the vehicle 1 from a navigation system or the like (step S103). Note that in steps S101 to S103, it is sufficient to acquire the current position, map information, and travel route of the vehicle 1, and the order of the processes may be changed as appropriate.

[0060] Thereafter, the reference route generating unit 103 generates a reference route based on the current position, map information, and driving route of the vehicle 1 (step S104). The reference route generating unit 103 generates a reference route indicating the traveling direction (azimuth) of the vehicle 1 by drawing curves for all right and left turning points included in the driving route, as shown in Fig. 4.

[0061] Next, the route section classification unit 104 classifies the reference route into a plurality of sections by setting points S1 to S5 shown in FIG. 4 at each intersection on the reference route (step S105).

[0062] Next, the lane marking extraction unit 106 extracts feature points of lane markings on the road from the captured image of a predetermined range ahead of the vehicle 1 captured by the imaging unit 220 (step S106). For example, if the road on which the vehicle 1 is traveling is a single lane, the lane markings on the left and right ahead of the lane on which the vehicle 1 is traveling are detected, as shown by the white circles in FIG. 6A, and multiple feature points are extracted from the lane markings. Any point on the lane markings detected from the captured image may be used as the feature point. For example, feature points may be set at predetermined intervals on the detected lane markings.

[0063] Then, the lane marking extraction unit 106 converts the extracted characteristic points of the lane markings into coordinates of a curvilinear coordinate system based on the reference route (step S107).

[0064] The curvilinear coordinate system based on the reference route is expressed as (S, D) using route length S from the start point of the reference route (the starting position of vehicle 1) and distance D in the normal direction from the reference route.

[0065] For example, when vehicle 1 is on a single-lane section, as shown in Fig. 6A, the characteristic points of the lane markings on the left and right sides in front of vehicle 1 are converted to coordinates (S, D) in a curvilinear coordinate system. For example, the coordinates of characteristic point fp shown in Fig. 6A are (Sn, Dn). Note that the reference route shown by the dotted line in Fig. 6A is shown outside the road for ease of viewing, but in reality it is written in the center of the road.

[0066] Returning to FIG. 5, the host vehicle section classification determination unit 105 then determines which of the sections classified by the route section classification unit 104 the current position of the vehicle 1 belongs to (step S108). Specifically, the host vehicle section classification determination unit 105 determines whether the route length S from the travel start position of the current position of the vehicle 1 corresponds to the first section from just before the intersection to point S1, the second section from point S1 to point S2, the third section from point S2 to point S5, or the fourth section from point S5 onward. Note that which section the current position of the vehicle 1 belongs to is determined by comparing the route length S from the travel start position of the vehicle 1 with points S1 to S5 of the intersection to be entered. The intersection to be entered is the nearest intersection in the direction of travel. For example, when the host vehicle section classification determination unit 105 detects an intersection within a predetermined range in the direction of travel, it acquires information on points S1 to S5 of the intersection.

[0067] If the route length S from the starting position of vehicle 1 is less than S1 or greater than S5 of the intersection to be entered (step S109; Yes), that is, if the current position of vehicle 1 is on the single-lane section of the first or fourth section, the controller 100 sets the lane line coordinates, which are the coordinates of the characteristic points of the lane line converted in step S107, and the reference route azimuth, which is the azimuth indicated by the direction of travel of the reference route, as input information for generating the lane boundary model, as shown in Figure 8 (step S110).

[0068] The first section is the fourth section at the nearest intersection that has already been passed. That is, this section is the section between point S5 of the nearest intersection that has already been passed and point S1 of the nearest intersection in the direction of travel. That is, when looking at the entire reference route, the first section and the fourth section are similar sections. Therefore, in step S109, it is determined whether the section is the first section or the fourth section, and if it is the first section or the fourth section, the same control is executed.

[0069] For example, in the case of the reference route shown in FIG. 4, the reference route azimuth angle set in step S110 is upward (e.g., north) for the section up to S3, tangent to the curve for the section from S3 to S4, and leftward (e.g., west) for the section after S4.

[0070] If the route length S from the starting position of the current position of vehicle 1 is greater than or equal to S1 and less than S2 of the intersection to be entered (step S109; No, step S111; Yes), that is, if the current position of vehicle 1 is in the second section before entering the intersection, the controller 100 sets the lane boundary coordinates, reference route azimuth angle, intersection entry extension line, and intersection exit extension line as input information for generating the lane boundary model, as shown in Figure 8 (step S112).

[0071] When the current location of vehicle 1 is in the second section between S1 and S2, in addition to the left and right lane markings ahead of the currently traveling lane, it is also possible to extract characteristic points of lane markings near the entrance and exit of the intersection, as shown by the white circles in FIG. 7A. The intersection approach extension line is the characteristic point of the lane marking extracted at the entrance of the intersection (at or near point S3) extended in the opposite direction of the reference route (black circles in FIG. 7A). The intersection exit extension line is the characteristic point of the lane marking extracted at the exit of the intersection (at or near point S4) extended in the direction of the reference route (black circles in FIG. 7A). The intersection approach extension line is the intersection approach route of vehicle 1 estimated based on the lane markings near the entrance of the intersection. Furthermore, the intersection exit extension line is the intersection exit route of vehicle 1 estimated based on the lane markings near the exit of the intersection. In the second section, these intersection approach extension lines and intersection exit extension lines are added to the input information for generating a lane boundary model.

[0072] In the second section near the intersection, the intersection approach extension line and intersection exit extension line are added to the input information for generating the lane boundary model. This allows the lane boundary when entering the intersection to be generated based on the intersection approach extension line. Also, the lane boundary when exiting the intersection can be generated based on the intersection exit extension line. Even if there are few characteristic points of the lane markings that can be extracted near the intersection exit, these characteristic points can be used to generate input information based on the reference route, making it possible to generate an optimal lane when exiting the intersection.

[0073] After the input information for generating the lane boundary model is set in step S110 or S112, the lane boundary generating unit 107 generates or updates the lane boundary model based on the input information (step S113).

[0074] The first or fourth section, where the route length S is less than S1 or greater than S5, is a single-lane section before entering an intersection or after exiting an intersection, as shown in FIG. 4. In this case, as shown in FIG. 8, lane line coordinates and a reference route azimuth angle are set as input information. The lane boundary generation unit 107 generates or updates a lane boundary model indicating the left and right lane boundaries shown by the dashed lines in FIG. 6B, for example, by connecting the lane line coordinates along the reference route azimuth angle. In other words, the lane boundary generation unit 107 updates the lane boundary model in the first or fourth section based on the distance of the characteristic point of the lane line from the reference route in the normal direction. In this way, a lane boundary model based on lane lines is generated in the first or fourth section, which is a single-lane section before entering an intersection or after exiting an intersection, thereby enabling lane following based on the lane boundary model. Furthermore, in the fourth section, which is a single-lane section after exiting an intersection, lane following based on lane lines after exiting an intersection can be restored. For example, the lane boundary generation unit 107 may generate lane boundaries by connecting lane line coordinates, and may generate a lane boundary model by complementing the lane boundaries using a reference route azimuth angle for areas where the lane lines are blurred and cannot be detected.

[0075] The second section, whose route length S is equal to or greater than S1 but less than S2, is a single-lane road near the intersection entrance, as shown in FIG. 4. In this case, as shown in FIG. 8, the input information includes lane line coordinates, a reference route azimuth, an intersection entrance extension, and an intersection exit extension. The lane boundary generation unit 107 generates a lane boundary model showing the left and right lane boundaries shown by the dashed lines in FIG. 7B, for example, by connecting the lane line coordinates before the intersection entrance, the lane line coordinates after the intersection exit, the intersection entrance extension, and the intersection exit extension along the reference route azimuth. In other words, the lane boundary generation unit 107 updates the lane boundary model for the second section based on the normal distance from the reference route to the characteristic points of the lane lines up to the intersection entrance and after the intersection exit. Furthermore, for the second section, the lane boundary generation unit 107 adds to the input information to the lane boundary generation unit 107 an intersection exit route, which extends the characteristic points of the lane lines after the intersection exit in the direction indicated by the reference route, and an intersection entrance route, which extends the characteristic points of the lane lines up to the intersection entrance in the direction indicated by the reference route. The lane boundary of the curved section within the intersection can be generated by drawing a curve that follows the curves S3 to S4 of the reference route. In this way, for the second section, which is a single road near the intersection entrance, a lane boundary model for the area around the intersection, including the intersection itself, can be generated based on the lane line coordinates from the intersection exit onward, the intersection approach extension, the intersection exit extension, and the reference route azimuth. Furthermore, since it is possible to generate lane boundaries for sections without lane markings, such as within an intersection, automated driving in sections without lane markings can be realized. For example, the lane boundary generation unit 107 may generate lane boundaries by connecting lane line coordinates, and then generate a lane boundary model by interpolating the lane boundary for sections where the lane lines are blurred and cannot be detected using the reference route azimuth. Furthermore, if lane lines at the entrance or exit of the intersection cannot be detected and the intersection approach extension and intersection exit extension cannot be generated, the lane boundary model can be generated using the lane line coordinates of the detected lane lines.

[0076] When moving from the first section to the second section, or from the third section to the fourth section, the input information to the lane boundary generation unit 107 changes, which may result in a deviation between the lane boundary generated in the new section and the lane boundary generated in the previous section. In this case, a relaxation time or relaxation distance may be set to eliminate the deviation between the different sections, and the lane boundary may be corrected to gradually transition from the lane boundary generated in the previous section to the lane boundary generated in the new section, thereby generating a lane boundary model. This can mitigate sudden changes in the lane boundary model and lane, preventing sudden steering during autonomous driving or driver assistance of the vehicle 1.

[0077] Next, the lane generation unit 108 generates or updates a lane for the vehicle by converting the lane boundary model generated by the lane boundary generation unit 107 into an orthogonal coordinate system (e.g., information indicating latitude and longitude) (step S114). For example, as shown by the thick solid arrows in Figures 6B and 7B, a line passing through the center of the left and right lane boundaries indicated by the lane boundary model is generated or updated as the lane.

[0078] If the route length S from the travel start position of the current position of vehicle 1 is equal to or greater than S2 but less than S5 of the intersection to be entered (step S109; No, step S111; No), that is, if the current position of vehicle 1 is in the third section, the process proceeds to step S115. The third section, where the route length S is equal to or greater than S2 but less than S5, is an intersection within or near an intersection, as shown in FIG. 4. In this case, since it is within or near an intersection and it is not expected that lane lines will be extracted from the captured image, the controller 100 does not set input information for lane boundary model generation and does not update the lane, as shown in FIG. 8. This prevents the generation of an inappropriate lane due to erroneous or inaccurate detection of lane lines.

[0079] If the lane markings cannot be extracted at the start of travel, such as when the vehicle 1 starts travelling within an intersection, the controller 100 may not be able to generate a suitable route. In this case, the controller 100 may output an error to the vehicle 1 or the user.

[0080] After the processing of step S114 or when the current position of the vehicle 1 is in the third section, the vehicle control unit 109 outputs a control signal to the actuator 240 so that the vehicle 1 travels along the generated or updated route (step S115). This allows the vehicle 1 to travel along the route generated by the route generation device 10. In other words, automatic driving of the vehicle 1 based on the route generated by the route generation device 10 can be realized.

[0081] Next, the controller 100 determines whether or not the vehicle has arrived at the destination set on the travel route (step S116). If the vehicle has not arrived at the destination (step S116; No), the process returns to step S106 and repeats the process of updating the travel route according to the travel of the vehicle 1. If the vehicle has arrived at the destination (step S116; Yes), the vehicle control process ends.

[0082] In this way, the lane generation device 10 of this embodiment can update the lane as the vehicle 1 travels. In the first section up to point S1, which is a single-lane section up to the entrance of the intersection, and in the fourth section from point S5 onward, which is a single-lane section after the exit of the intersection, the lane generation device 10 sets the input information to the lane boundary generation unit 107 as lane line coordinates and a reference route azimuth, and updates the lane based on the lane lines of the single-lane section. In the second section from point S1 to point S2 near the intersection before entering the intersection, an intersection entrance extension line and an intersection exit extension line are added to the input information to the lane boundary generation unit 107, and the lane near the intersection is updated based on the lane lines up to the entrance of the intersection and the lane lines after the exit of the intersection. In the third section from point S2 to point S4, which is the intersection, the lane is not updated, and the vehicle 1's travel is controlled based on the lane generated up to point S2, for example. In this way, the path generation device 10 determines whether to generate or update a path and varies the input information used for path generation and updating depending on the section to which the current position of the vehicle 1 belongs, i.e., the distance to the intersection, etc. That is, the path generation device 10 of this embodiment classifies the reference route into multiple sections and varies the input information used for path generation and updating, taking into account whether or not lane markings on the road can be detected and the detection accuracy. This prevents the generation of inappropriate paths due to erroneous or inaccurate detection of lane markings, and because information based on detectable lane markings is set as input information, paths can be generated and updated appropriately according to the travel of the vehicle 1.

[0083] (Variation) The present invention is not limited to the above-described embodiment, and various modifications and applications are possible. For example, parts of the above-described embodiment may be omitted or replaced, or any configuration may be added. Furthermore, the hardware configurations, functional configurations, flowcharts, sequences, etc. shown in the above-described embodiment are merely examples and may be modified as appropriate.

[0084] In the above embodiment, when the vehicle 1 starts traveling, points S1 to S5 are set for all intersections on the reference route at once, thereby classifying the reference route into a plurality of sections. However, this is not limiting, and the process of classifying the reference route into a plurality of sections may be performed in separate steps. For example, while the vehicle 1 is traveling, points S1 to S5 may be set for the intersection each time the distance to the intersection falls within a predetermined distance.

[0085] In the above embodiment, as shown in Figures 4, 7A, and 7B, an example was described in which a lane boundary model and a lane are generated and updated at a left-turn intersection, but lane boundary models and lane can be generated and updated in a similar manner at a right-turn intersection or a straight-through intersection. That is, the controller 100 of the lane generation device 10 sets the lane line coordinates, reference route azimuth, intersection approach extension line, and intersection exit extension line as input information for the second section with route length S of S1 to S2 near a right-turn intersection or a straight-through intersection, and generates and updates the lane boundary model and lane based on the lane lines up to the entrance of the intersection and the lane lines after the exit of the intersection.

[0086] The method of generating a lane boundary model by the lane boundary generation unit 107 is not limited to the method described in the above embodiment. The lane boundary generation unit 107 may generate a lane boundary model by arbitrarily utilizing input information such as lane line coordinates, a reference route azimuth, an intersection approach extension line, and an intersection exit extension line. For example, the lane boundary generation unit 107 may first generate a lane boundary model based on a reference route azimuth that is initially obtained based on the driving route, and then sequentially update the lane boundary model based on lane line coordinates and extension lines that are sequentially obtained as the vehicle travels.

[0087] In addition, the intersection approach extension line and the intersection exit extension line do not need to be included in the input information to the lane boundary generation unit 107. In this case, the process of generating the intersection approach extension line and the intersection exit extension line may be omitted.

[0088] 5, lane markings are extracted from the captured image in steps S106 and S107, and then the section of the current location of vehicle 1 is determined in step S108. However, it is also possible to first determine the section of the current location of vehicle 1, and then extract lane markings from the captured image based on the determination result. In this case, if the section of the current location of vehicle 1 is the third section (S2 to S5) which is an intersection, the process of extracting lane markings from the captured image can be omitted. Also, if the section of the current location of vehicle 1 is the second section (S1 to S2) near the intersection, it is easier to detect lane markings after the intersection exit by, for example, focusing on a position in the captured image that can be estimated to be near the intersection exit.

[0089] The running path generation device 10 in the above embodiment generates the running path using a curvilinear coordinate system based on the reference route, but the running path may be generated by processing everything in an orthogonal coordinate system (latitude and longitude information, etc.).

[0090] Furthermore, in the above embodiment, the route section classification unit 104 classifies sections based on the route length of the reference route from the travel start point, but it may also classify sections based on the distance from an intersection.

[0091] (Variation 1) As shown in FIG. 3 , the map information in the above embodiment includes information on roads and intersections, and the travel route is described as a series of graph lines. The route section classification unit 104 may classify each road and intersection on a reference route generated based on the travel route as one section. The vehicle section classification determination unit 105 may acquire section information indicating whether each section is a road or an intersection, and determine the section type by determining whether the section to which the current vehicle 1 belongs and the sections before and after it are roads or intersections based on the division information. The controller 100 may generate and update a road by determining whether to generate and update a road boundary model using the lane boundary generation unit 107 and by changing the input information to the lane boundary generation unit 107, depending on the section type determined by the vehicle section classification determination unit 105.

[0092] 9, for example, if the current section to which the current vehicle 1 belongs is a single road, the previous section is an intersection, and the next section is a single road, the controller 100 determines that the type of section to which the current vehicle 1 belongs is a single road after leaving the intersection (for example, the fourth section in the above embodiment). At this time, the controller 100 sets the lane boundary coordinates and the reference route azimuth angle as input information to the lane boundary generation unit 107.

[0093] When the current section is a single road, the previous section is a single road, and the next section is an intersection, the controller 100 determines that the section type to which the current vehicle 1 belongs is the single road before entering the intersection (for example, the first or second section in the above embodiment). At this time, the controller 100 sets the lane line coordinates, the reference route azimuth angle, the intersection entrance extension line, and the intersection exit extension line as input information to the lane boundary generation unit 107.

[0094] Furthermore, when the current section is a single road and the preceding and following sections are intersections, the controller 100 determines that the section type to which the current vehicle 1 belongs is a single road after leaving an intersection or before entering an intersection. Whether the current section is after leaving an intersection or before entering an intersection may be further determined based on the distance between the preceding section and the following section, etc. Then, the controller 100 may set input information according to the determination result.

[0095] If the section to which the current vehicle 1 belongs is an intersection, the controller 100 determines that the section type to which the current vehicle 1 belongs is passing through an intersection (for example, the third section in the above embodiment). At this time, the controller 100 does not set input information to the lane boundary generation unit 107 and does not update the lane.

[0096] According to this modification 1, the section type to which the current vehicle 1 belongs can be determined based on road information. Then, the controller 100 can predict the presence or absence of lane markings on the road and switch input information according to the determined section type, and can switch the lane update method. In this case, it is no longer necessary to set points S1 to S5 at each intersection as in the above embodiment, and the processing load on the controller 100 and the lane generation device 10 can be reduced.

[0097] In the above embodiment, the location information of the vehicle 1 is acquired based on the GNSS signal, but the method of acquiring the location information of the vehicle 1 is not limited to this, and any other method may be used. For example, the location information may be acquired using a wireless signal from a wireless communication network, or may be acquired based on the output of a motion sensor mounted on the vehicle 1.

[0098] In addition, in the first embodiment, the lane markings are detected from the image captured by the imaging unit 220, but this is not limiting. For example, instead of the imaging unit 220, the lane markings may be detected by any sensor such as a LiDAR (Light Detection and Ranging), a RADAR (Radio Detection and Ranging), an LRF (Laser Range-Finder), or a SONAR (Sound Navigation and Ranging).

[0099] Furthermore, in the above embodiment, the configuration is described assuming that the vehicle is traveling on the left side, but the present invention can also be applied to the case of traveling on the right side.

[0100] The route generation device 10 and route generation method of the present invention may also be applied to automatic driving of vehicles traveling on roads other than public roads, for example, roads within facilities.

[0101] Furthermore, in the above embodiment, an example has been described in which the processor 1011 executes a control program to realize each function, but the controller 100 may also be configured with dedicated hardware that realizes each function.

[0102] Furthermore, a control program for executing the operations of the above-described embodiments may be stored and distributed on a computer-readable recording medium such as a CD-ROM (Compact Disc Read-Only Memory), a DVD (Digital Versatile Disc), an MO (Magneto Optical Disc), or a memory card, and the program may be installed on a computer to configure the controller 100 that can realize each function. When each function is realized by sharing the work between an OS (Operating System) and an application, or by cooperation between the OS and an application, only the parts other than the OS may be stored on the recording medium.

[0103] The present invention allows various embodiments and modifications without departing from the broad spirit and scope of the present invention. Furthermore, the above-described embodiments are intended to explain the present invention and do not limit the scope of the present invention. In other words, the scope of the present invention is defined by the claims, not by the embodiments. Various modifications made within the scope of the claims and the meaning of the disclosure equivalent thereto are considered to be within the scope of the present invention. [Explanation of symbols]

[0104] 1 vehicle, 10 roadway generation device, 100 controller, 101 driving route acquisition unit, 102 map information acquisition unit, 103 reference route generation unit, 104 route section classification unit, 105 host vehicle section classification determination unit, 106 lane marking extraction unit, 107 roadway boundary generation unit, 108 roadway generation unit, 109 vehicle control unit, 210 position information acquisition unit, 220 imaging unit, 230 map database, 240 actuator, 1010 bus, 1011 processor, 1012 memory, 1013 storage, 1014 communication interface.

Claims

1. Obtain the current position of the vehicle, A lane generation method for generating a lane along which the vehicle will travel, the lane generation method comprising: extracting road markings from detection information of a road ahead in a traveling direction of the vehicle, the detection information being acquired by a sensor provided in the vehicle; and generating a lane along which the vehicle will travel based on the extracted markings, the lane generation method comprising: The route can be updated according to the traveling of the host vehicle, Before the vehicle enters an intersection, the route is updated based on the dividing line up to the entrance of the intersection and the dividing line after the exit of the intersection; The route is not updated at the intersection from the entrance to the exit of the intersection, In the single-lane section up to the entrance of the intersection and the single-lane section after the exit of the intersection, the lane is updated based on the lane markings of the single-lane section. Track generation method.

2. acquiring a driving route of the vehicle on map data including at least information on roads and intersections; generating a reference route indicating a traveling direction of the road on which the host vehicle is traveling based on the acquired traveling route; the detection information is a captured image, and road dividing lines are extracted from the captured image; generating a lane boundary model indicating a boundary of the lane based on the coordinates of the extracted characteristic points of the lane markings in a curvilinear coordinate system based on the reference route; The road is generated by converting the generated road boundary model into an orthogonal coordinate system. The route generation method according to claim 1 .

3. classifying the reference route around an intersection to be entered into a plurality of sections based on a route length from the current position of the vehicle to the intersection to be entered; The route can be updated according to the classified sections. The route generation method according to claim 2 .

4. In a single section that is a first section in which the route length is less than a first distance among the classified sections, the lane boundary model is updated based on the distance in the normal direction from the reference route of a feature point of a lane marking extracted from the captured image, thereby updating the lane. The route generation method according to claim 3.

5. In a second section among the classified sections, which is a section before entering an intersection and in which the route length is equal to or greater than the first distance and less than the second distance, the lane boundary model is updated based on the normal distance from the reference route to a characteristic point of a lane line up to the entrance of the intersection extracted from the captured image, and the normal distance from the reference route to a characteristic point of a lane line onward from the exit of the intersection extracted from the captured image, thereby updating the lane. The route generation method according to claim 4.

6. In the second section among the classified sections, an intersection escape route obtained by extending characteristic points of the lane markings after the exit of the intersection extracted from the captured image in the direction indicated by the reference route is added to the input information, and the lane boundary model is updated based on the intersection escape route, thereby updating the lane. The route generation method according to claim 5.

7. In the second section among the classified sections, an intersection approach route obtained by extending the characteristic points of the lane markings to the entrance of the intersection extracted from the captured image in the direction indicated by the reference route is added to the input information, and the lane boundary model is updated based on the intersection approach route, thereby updating the lane. The route generation method according to claim 6.

8. Among the classified sections, at an intersection portion that is a third section where the route length is equal to or greater than the second distance and less than the third distance, the travel path is not updated. The route generation method according to claim 3.

9. In the classified sections, in a single section that is a fourth section in which the route length is equal to or greater than the third distance, the lane boundary model is updated based on the distance in the normal direction from the reference route of the characteristic point of the lane marking extracted from the captured image, thereby updating the lane. The route generation method according to claim 8.

10. a sensor for acquiring information about the area ahead of the vehicle; a location information acquisition unit that acquires a current location of the vehicle; A path generation device including: a controller that generates a path on which the host vehicle will travel; The controller extracting road dividing lines from the detection information acquired by the sensor, and generating the running path based on the extracted dividing lines; The route can be updated according to the traveling of the host vehicle, Before the vehicle enters an intersection, the route is updated based on the dividing line up to the entrance of the intersection and the dividing line after the exit of the intersection; The route is not updated at the intersection from the entrance to the exit of the intersection, In the single-lane section up to the entrance of the intersection and the single-lane section after the exit of the intersection, the lane is updated based on the lane markings of the single-lane section. Track generation device.

Images