Driving assistance apparatus and driving assistance program

Abstract

To provide a technique for improving the accuracy of lane change determination.SOLUTION: A driving assistance apparatus (1) for assisting the operation of a vehicle (C) is provided with a lane change determining unit (44). The lane change determining unit determines that the vehicle has changed lanes from its own lane (Hj), which is the lane in which the vehicle is currently traveling, to an adjacent lane (Hr) based on recognition results based on a captured image of a road demarcation line (G) provided on a road surface (DF) of a road (D) in front of the vehicle. The lane change determination unit has a determination reference line selection unit (45) which selects a determination reference line (Lc) as a basis for lane change determination. The determination reference line selection unit selects a virtual line offset by a predetermined amount from its own lane among the virtual lines included in the image recognition results, which are line information to be candidates for the boundary of the lane (H).SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a driving assistance device and a driving assistance program for assisting driving of a vehicle. [Background technology]

[0002] In recent years, vehicles equipped with driving assistance devices or automatic driving devices capable of implementing functions such as lane keeping control, lane departure prevention, and lane change assistance have become widespread. The lane keeping control function is a function that automatically controls the motion of the vehicle so that it travels within a certain lane. The lane departure prevention function is a function that warns the driver and / or automatically controls the steering so that the vehicle does not deviate from the lane in which it is currently traveling. The lane change assistance function is a control that automatically moves the vehicle from the lane in which it is currently traveling to an adjacent lane. In order to properly execute these driving assistance functions or automatic driving functions, it is necessary for the vehicle to accurately recognize the lane in which it is currently traveling.

[0003] In this regard, Patent Document 1 discloses a technology for recognizing the lane in which a vehicle is traveling by comparing the pattern of road dividing lines obtained from map information with the pattern of road dividing lines around the vehicle recognized from an image captured by a camera. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2018-203018 A Summary of the Invention [Problem to be solved by the invention]

[0005] In this type of device, when the vehicle intentionally deviates from the lane in which the vehicle is currently traveling in order to change lanes, overtake, or the like, a determination is made as to whether the vehicle has deviated from the lane and moved to an adjacent lane or the like. In this case, in this type of conventional device, the accuracy of such a determination may be reduced depending on the driving scene. The present invention has been made in consideration of the circumstances exemplified above. That is, the present invention provides a technology for improving the accuracy of lane change determination, for example. [Means for solving the problem]

[0006] A driving assistance device (1) for assisting driving of a vehicle (C) according to claim 1, a lane change determination unit (44) that determines that the vehicle has changed lanes from a current lane (Hj) in which the vehicle is currently traveling to an adjacent lane (Hr) based on a recognition result based on a photographed image of a road dividing line (G) provided on a road surface (DF) of a road (D) ahead of the vehicle, The lane change determination unit includes a determination reference line selection unit (45) that selects a determination reference line (Lc) that is a reference for lane change determination, The determination reference line selection unit selects a virtual line that is offset by a predetermined amount from the own lane from among virtual lines that are line information that are included in the recognition result and serve as boundary candidates for a lane (H). A driving assistance program according to claim 15 is a computer program executed by a driving assistance device (1) that assists driving of a vehicle (C), The process executed by the driving assistance device is as follows: A lane change determination process is included, which determines that the vehicle has changed lanes from a current lane (Hj) in which the vehicle is currently traveling to an adjacent lane (Hr) based on a recognition result based on a photographed image of a road dividing line (G) provided on a road surface (DF) of a road (D) in front of the vehicle, The lane change determination process includes a determination reference line selection process for selecting a determination reference line (Lc) that is a reference for lane change determination, The determination reference line selection process selects a virtual line that is offset by a predetermined amount from the own lane from among virtual lines that are line information that are included in the recognition result and are candidates for boundaries of a lane (H).

[0007] In addition, in each section of the application documents, each element may be given a reference number in parentheses. However, such a reference number merely indicates an example of the correspondence between the element and the specific means described in the embodiment described later. Therefore, the present invention is not limited in any way by the description of the reference numbers above. [Brief description of the drawings]

[0008] [Figure 1] 1 is a side view showing a state in which a host vehicle equipped with a driving assistance device according to an embodiment of the present invention is traveling on a road. [Diagram 2] FIG. 2 is a plan view of the host vehicle shown in FIG. [Diagram 3] 2 is a block diagram showing a schematic configuration of a first embodiment of the driving assistance device shown in FIG. 1. [Figure 4] 4 is a block diagram showing a schematic functional configuration realized by an image recognition unit shown in FIG. 3. [Diagram 5] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 6] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 7] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 8] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 9] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 10]1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 11] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 12] 10 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a comparative example; FIG. [Figure 13] 10 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a comparative example; FIG. [Figure 14] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 15] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 16] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 17] 10 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a comparative example; FIG. [Figure 18] 10 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a comparative example; FIG. [Figure 19] 5 is a flowchart showing a specific example of a lane change determination process in the first embodiment of the present invention. [Figure 20] 5 is a flowchart showing a specific example of a lane change determination process in the first embodiment of the present invention. [Figure 21] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 22] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a left lane to a right lane in a first embodiment of the present invention; [Figure 23]1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a straight lane to a branching lane in a first embodiment of the present invention; FIG. [Figure 24] 1 is a schematic diagram showing an overview of lane recognition when a host vehicle changes lanes from a straight lane to a branching lane in a first embodiment of the present invention; FIG. [Diagram 25] 2 is a block diagram showing a schematic configuration of a driving assistance device shown in FIG. 1 according to a second embodiment and a third embodiment. FIG. [Figure 26] 26 is a block diagram showing an example of a schematic functional configuration realized by an image recognition unit and a lane recognition unit shown in FIG. 25. FIG. [Figure 27] 27 is a schematic diagram showing an overview of road shoulder recognition performed by the lane recognition unit shown in FIG. 26. [Figure 28] 27 is a flowchart showing a specific example of a process related to a road shoulder recognition function performed by a lane recognition unit shown in FIG. 26. [Figure 29] 26 is a block diagram showing another example of the schematic functional configuration realized by the image recognition unit and the lane recognition unit shown in FIG. 25. FIG. [Diagram 30] 30 is a flowchart showing a specific example of processing related to a road shoulder recognition function performed by a lane recognition unit shown in FIG. 29. [Diagram 31] 26 is a block diagram showing yet another example of the schematic functional configuration realized by the image recognition unit and the lane recognition unit shown in FIG. 25. FIG. [Diagram 32] 32 is a flowchart showing a specific example of processing related to a road shoulder recognition function performed by a lane recognition unit shown in FIG. 31. [Diagram 33] 26 is a block diagram showing yet another example of the schematic functional configuration realized by the image recognition unit and the lane recognition unit shown in FIG. 25. FIG. [Diagram 34] 34 is a flowchart showing a specific example of processing related to a road shoulder recognition function performed by a lane recognition unit shown in FIG. 33. [Diagram 35] 26 is a block diagram showing yet another example of the schematic functional configuration realized by the image recognition unit and the lane recognition unit shown in FIG. 25. FIG. [Diagram 36] 36 is a flowchart showing a specific example of processing related to a road shoulder recognition function performed by a lane recognition unit shown in FIG. 35. [Figure 37] 26 is a block diagram showing yet another example of the schematic functional configuration realized by the image recognition unit and the lane recognition unit shown in FIG. 25. FIG. [Figure 38] 38 is a flowchart showing a specific example of a process related to a road shoulder recognition function performed by a lane recognition unit shown in FIG. 37. [Figure 39] FIG. 38 is a schematic diagram showing an overview of road shoulder recognition performed by the lane recognition unit shown in FIG. 37. [Diagram 40] FIG. 38 is a schematic diagram showing an overview of road shoulder recognition performed by the lane recognition unit shown in FIG. 37. [Diagram 41] FIG. 38 is a schematic diagram showing an overview of road shoulder recognition performed by the lane recognition unit shown in FIG. 37. [Diagram 42] 38 is a flowchart showing another specific example of the process related to the road shoulder recognition function performed by the lane recognition unit shown in FIG. 37. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] (Embodiment) Hereinafter, an embodiment of the present invention will be described based on the drawings. Note that the configuration, functions, and implementation examples of the embodiment can be modified as appropriate. Therefore, multiple embodiments and variations may be described in the specification. In this case, in the following description, the same reference symbols will be given to parts that are identical or equivalent between one embodiment and another embodiment, between one embodiment and its variation, or between one variation and another variation. And, the preceding description of such parts may be appropriately applied to other embodiments or variations that follow, unless there is a technical contradiction or special additional explanation.

[0010] (Driving assistance device) 1 and 2, the vehicle C is an automobile traveling on a road D and has a box-shaped body. The vehicle C is equipped with a driving assistance device 1. The driving assistance device 1 is configured to assist the driving of the vehicle C by using an image recognition result based on an image captured by an image capturing unit 2, which is a camera. Specifically, the driving assistance device 1 is configured to be capable of executing various operations such as driving control operations of the vehicle C and warning operations for occupants. In addition, when the driving assistance device 1 is configured to realize the automatic driving of the vehicle C, the driving assistance device 1 is called an "automatic driving device". "Automatic driving" is a driving automation level corresponding to SAE levels 3 to 5, in which the driving automation system is responsible for, i.e., executes, all dynamic driving tasks. In contrast, "driving assistance" is a driving automation level corresponding to SAE levels 1 to 2, in which the driving automation system is responsible for, i.e., executes, some dynamic driving tasks. In other words, "driving assistance" also includes "advanced driving assistance" of SAE level 2. "DDT" refers to all operational and tactical functions, excluding strategic functions, that need to be performed in real time when operating a vehicle in road traffic. Driving behavior in general can be categorized into three types of functions: strategic, tactical, and operational. "Strategic" functions include planning a route, selecting stops, etc., and specifically, deciding or choosing "to go or not, when, where and how to go". "Tactical" functions involve vehicle operation in traffic situations, such as deciding whether and when to overtake or change lanes, and choosing an appropriate speed. "Operational" functions include steering, braking, accelerating, etc. SAE stands for Society Of Automotive Engineers.

[0011] The image capturing unit 2 is a so-called forward camera, and is fixed on the body or inside the vehicle cabin of the vehicle C so as to capture an image of the area ahead of the vehicle C. That is, the image capturing unit 2 is provided so as to be able to capture various targets within a capture range E ahead of the vehicle C. Targets to be captured include, for example, three-dimensional objects such as traffic lights, road signs, and obstacles around the vehicle C, and road markings such as road dividing lines G provided on the road surface F of the road D ahead. The image capturing unit 2 is provided with an imaging element such as a CCD or a CMOS, and is configured to be able to capture color images. CCD stands for Charge Coupled Device. CMOS stands for Complementary Metal Oxide Semiconductor.

[0012] (First embodiment: configuration) 3 and 4 show a schematic configuration of a driving support device 1 according to a first embodiment of the present invention. Referring to FIG. 3, the driving support device 1 includes an image capturing unit 2, a vehicle state quantity sensor unit 3, an image recognition unit 4, and an alarm / vehicle control unit 5. The image capturing unit 2 repeatedly captures an image of the front of the vehicle C at a preset time interval. The image capturing unit 2 is connected to the image recognition unit 4 via an in-vehicle communication line conforming to a communication standard such as CAN (international registered trademark) so as to output captured image data to the image recognition unit 4. CAN (international registered trademark) is an abbreviation for Controller Area Network. Note that the in-vehicle communication line may have a sub-network conforming to LIN, FlexRay, or the like in addition to a main network conforming to CAN (international registered trademark). LIN is an abbreviation for Local Interconnect Network.

[0013] The host vehicle state quantity sensor unit 3 is connected to the image recognition unit 4 via an in-vehicle communication line so as to measure the state quantity of the host vehicle C and output the measurement result to the image recognition unit 4. That is, the host vehicle state quantity sensor unit 3 is a collective name for various sensors, such as a vehicle speed sensor and a yaw rate sensor, that measure state quantities related to the driving state and behavior of the host vehicle C for the sake of simplicity of illustration and explanation. Each sensor included in the host vehicle state quantity sensor unit 3 repeats measurement at a preset time interval and outputs the measurement result to the image recognition unit 4. The in-vehicle sensor may be equipped with a well-known navigation device, and the navigation device may store map information including information on road shapes and the like.

[0014] The image recognition unit 4 has a configuration centered on a well-known microcomputer equipped with a CPU, a ROM, a RAM, a non-volatile rewritable memory, and the like. The CPU stands for Central Processing Unit. The ROM stands for Read Only Memory. The RAM stands for Random Access Memory. The non-volatile rewritable memory is a storage device that allows information to be rewritten while the power is on, but holds information in an unrewritable manner while the power is off, such as a flash memory or a hard disk. The image recognition unit 4 is configured to execute various functions as described below by the CPU reading and executing a program stored in the ROM and / or the non-volatile rewritable memory. The ROM and the non-volatile rewritable memory correspond to non-transient physical recording media. The image recognition unit 4 outputs the image recognition result to the alarm / vehicle control unit 5.

[0015] The warning / vehicle control unit 5 has a configuration centered on a well-known microcomputer including a CPU, a ROM, a RAM, a non-volatile rewritable memory, and the like. The warning / vehicle control unit 5 is connected to the image recognition unit 4 via an in-vehicle communication line so as to execute a warning operation for the occupants of the vehicle C and / or a driving behavior control operation of the vehicle C based on the image recognition result by the image recognition unit 4. The driving behavior control operation includes steering control, braking control, acceleration control, and the like. The driving support device 1 is configured to execute a lane keeping control function, a lane departure suppression function, a lane change support function, an automatic lane change function, an automatic overtaking control function, and the like based on the image recognition result of the image capturing unit 2 of the road dividing line G, and the like. These functions may be collectively referred to as "lane-related control functions" below.

[0016] (Image Recognition Device) Fig. 4 shows part of the functional configuration realized on a microcomputer by executing a program by the CPU in the image recognition unit 4. As shown in Fig. 4, the image recognition unit 4 includes a lane marking detection unit 41, a line type determination unit 42, a road shape estimation unit 43, and a lane change determination unit 44 as functional configuration realized on the in-vehicle microcomputer.

[0017] The lane marking detection unit 41 is adapted to detect road markings G in the captured image. Specifically, the lane marking detection unit 41 is adapted to detect, i.e., extract, white or yellow painted portions on the road surface F that constitute the road markings G by using a well-known image recognition technique. The line type determination unit 42 is adapted to determine the line type of the detected road markings G. Examples of "line types" include straight lines, dashed lines, double lines consisting of straight lines and dashed lines, double lines consisting of straight lines and straight lines, etc.

[0018] 5, the lane detection unit 41 performs lane recognition, that is, recognition of the lane H on the road D on which the vehicle C is currently traveling, based on the detection result of the road lane G. In the specific example shown in FIG. 5, the left lane H1 is the leftmost lane H between the outer roadway line G1 and the lane boundary line G2 on the right side thereof based on the traveling direction of the vehicle C, and is the "traffic lane" (hereinafter simply referred to as the "traffic lane") under the Road Traffic Act of Japan. The right lane H2 is the lane H adjacent to the right side of the left lane H1, and is a traffic lane or an overtaking lane. In other words, the road lane G that defines the right end of the right lane H2 is the lane boundary line G2 or the center line.

[0019] Lane recognition includes recognition of lane structure, lane position, and host vehicle position. Lane structure includes the number and type of lanes H in the road width direction. When there are multiple lanes H in the road width direction, the lane position indicates which of the multiple lanes H the host vehicle's lane Hj, which is the lane H in which the vehicle is currently traveling, is in. The host vehicle position is the lateral position of the host vehicle center line CL within the host vehicle lane Hj, i.e., the position in the road width direction. The host vehicle center line CL is a virtual line segment that corresponds to the vehicle overall length of the host vehicle C, is parallel to the vehicle overall length direction of the host vehicle C, passes through the center in the vehicle width direction, and corresponds to the vehicle overall length of the host vehicle C. In other words, lane recognition includes recognition of the host vehicle lane Hj and recognition of the adjacent lane Hr, which is the lane H adjacent to the host vehicle lane Hj.

[0020] The recognition of the own lane Hj includes the recognition of a first own lane virtual line Kj1 and a second own lane virtual line Kj2. The first own lane virtual line Kj1 and the second own lane virtual line Kj2 are virtual lines along the extension direction of the road D (hereinafter referred to as the "road extension direction") that serve as a reference for determining whether to activate an alarm operation or a driving behavior control operation for keeping the vehicle in the lane or preventing it from leaving the lane. The first own lane virtual line Kj1 is set on the left side of the own lane Hj. Specifically, the first own lane virtual line Kj1 is set at the edge position on the inside of the own lane Hj, i.e., on the left side, of the road dividing line G located on the left side of the own lane Hj, or at a position offset by a predetermined amount further inside the own lane Hj, i.e., to the left side. The second own lane virtual line Kj2 is set on the right side of the own lane Hj. Specifically, the second own lane virtual line Kj2 is set at the inside, i.e., right-side edge position of the road dividing line G located to the right of the own lane Hj, or at a position offset a predetermined amount further inside, i.e., to the right, of the own lane Hj.

[0021] The recognition of the adjacent lane Hr includes the recognition of a first adjacent virtual line Kr1 and the recognition of a second adjacent virtual line Kr2. The first adjacent virtual line Kr1 is set on the left side of the adjacent lane Hr. Specifically, the first adjacent virtual line Kr1 is set at the inside edge position of the adjacent lane Hr, i.e., the left edge position, of the road dividing line G located on the left side of the adjacent lane Hr, or at a position offset by a predetermined amount further inside the adjacent lane Hr, i.e., to the left side. The second adjacent virtual line Kr2 is set on the right side of the adjacent lane Hr. Specifically, the second adjacent virtual line Kr2 is set at the inside edge position of the adjacent lane Hr, i.e., the right edge position, of the road dividing line G located on the right side of the adjacent lane Hr, or at a position offset by a predetermined amount further inside the adjacent lane Hr, i.e., to the right side. Machine learning by deep learning is used for the recognition of the own lane Hj and the adjacent lane Hr. Since lane recognition using deep learning was already publicly known or well-known at the time of filing this application, further details will not be explained in this specification.

[0022] The road shape estimation unit 43 is configured to estimate the shape of the road D on which the host vehicle C is currently traveling, using a model such as a Kalman filter. The road shape estimated by the road shape estimation unit 43 includes, for example, lane width, lane inclination, lane curvature, and the like. Referring to FIG. 5, the lane width is the distance between a pair of road dividing lines G provided on the left and right of the lane H. The lane inclination is the inclination of the tangent of the lane center line with respect to the vehicle traveling direction at the traveling direction position of the host vehicle C. The lane center line is a virtual curve that passes through a middle position equidistant from each of the pair of road dividing lines G provided on the left and right of the lane H and extends along the road extension direction. The lane curvature is the curvature of the lane center line. Road shape estimation using a model such as a Kalman filter is already publicly known or well known at the time of filing of the present application, so further details will be omitted in this specification.

[0023] The lane change determination unit 44 is adapted to determine that the vehicle C has changed lanes from the own lane Hj to the adjacent lane Hr based on the recognition result based on the captured image of the road dividing line G. As shown in FIG. 4, the lane change determination unit 44 includes a determination reference line selection unit 45. The determination reference line selection unit 45 is adapted to set the lane change determination reference line Lc shown in FIG. 5 and the like. The lane change determination reference line Lc is a virtual line that serves as a reference for lane change determination. FIGS. 5 to 8 show a schematic view of the vehicle C changing lanes from the left lane H1 to the right lane H2. In the example of FIG. 5, the lane change determination reference line Lc is a virtual line that serves as a reference for lane change determination of the vehicle C from the left lane H1, which is the own lane Hj, to the right lane H2, which is the adjacent lane Hr. Note that, as will be described later, the "lane change determination" is different from a determination of whether or not a lane change has been completed. That is, the determination that the lane change is completed refers to the determination that the entire body of the vehicle C is within the lane H of the destination lane. In contrast, the "lane change determination" is the determination of a state in which the vehicle C is evaluated as having substantially moved from the original vehicle lane Hj to the destination lane H, and is considered to be satisfied even when the vehicle C is straddling the road dividing line G between the original vehicle lane Hj and the destination lane H.

[0024] In this embodiment, the judgment reference line selection unit 45 is adapted to select the lane change judgment reference line Lc from among the recognized virtual lines. Specifically, the judgment reference line selection unit 45 selects a virtual line offset by a predetermined amount from the own lane Hj among the virtual lines that are line information that are boundary candidates of the lane H included in the lane recognition result. More specifically, the judgment reference line selection unit 45 selects a virtual line generated based on a position difference between the own lane Hj and the adjacent lane Hr. Such a position difference can be acquired from the lane recognition result using deep learning or map information. In the example of FIG. 5, the judgment reference line selection unit 45 selects the first adjacent virtual line Kr1 on the side close to the own lane Hj as the lane change judgment reference line Lc among the first adjacent virtual line Kr1 and the second adjacent virtual line Kr2 recognized in the adjacent lane Hr. In other words, the judgment reference line selection unit 45 sets the reference position for judging a lane change from the own lane Hj to the adjacent lane Hr on the adjacent lane Hr side, thereby judging a lane change from the adjacent lane Hr rather than from the own lane Hj.

[0025] (Effects of the First Embodiment) Below, an overview of the operation of the driving assistance device 1 according to this embodiment, i.e., the image recognition unit 4, or the driving assistance program, i.e., the image recognition program, which is a computer program executed thereby, will be described together with the effects achieved by these. Hereinafter, the driving assistance device 1, the image recognition unit 4, the driving assistance program, and the image recognition program according to this embodiment may be collectively referred to as "this embodiment."

[0026] FIG. 5 shows a state when the vehicle C starts to change lanes from the left lane H1, which is the vehicle's own lane Hj, to the right lane H2, which is the adjacent lane Hr. FIG. 6 and FIG. 7 show a state during the lane change. FIG. 8 shows a state after the vehicle's own lane Hj is switched from the left lane H1 to the right lane H2 by lane change. First, referring to FIG. 5, this embodiment recognizes a first vehicle's own lane virtual line Kj1 along the outer roadway line G1 located at the left end of the left lane H1, which is the vehicle's own lane Hj at the current time. In addition, this embodiment recognizes a second vehicle's own lane virtual line Kj2 along the lane boundary line G2 located at the right end of the left lane H1, which is the vehicle's own lane Hj at the current time. In addition, this embodiment recognizes a first adjacent virtual line Kr1 along the lane boundary line G2 located at the left end of the right lane H2, which is the adjacent lane Hr at the current time. In addition, in this embodiment, a second adjacent virtual line Kr2 is recognized along the road dividing line G, ie, the lane boundary line G2 or center line, located at the right end of the right lane H2, which is the current adjacent lane Hr.

[0027] In this embodiment, the first adjacent virtual line Kr1, which is a virtual line on the side of the own lane Hj in the right lane H2, which is the adjacent lane Hr, is selected as the lane change judgment reference line Lc. In addition, in this embodiment, the second own lane virtual line Kj2, which is a virtual line on the side of the adjacent lane Hr in the left lane H1, which is the own lane Hj, is selected as the crossing judgment reference line Lf, which is a reference line for the crossing judgment. The crossing judgment is a judgment of whether or not the own vehicle C is crossing the road division line G that constitutes the own lane Hj. As shown in FIG. 6, when the start crossing judgment amount d1, which is the shortest distance between the own vehicle center line CL and the crossing judgment reference line Lf, becomes smaller than half the vehicle width of the own vehicle C, the crossing judgment is turned ON. The start crossing judgment amount d1 can be calculated based on the image recognition result and the vehicle body dimensions of the own vehicle C.

[0028] As shown in FIG. 6, when the vehicle C crosses the lane boundary line G2 and the crossing determination is turned ON, the vehicle C moves further to the right, i.e., toward the adjacent lane Hr, as shown in FIG. 7, the vehicle C protrudes into the adjacent lane Hr. Therefore, in this embodiment, a protrusion amount d2 is calculated. The protrusion amount d2 corresponds to the protrusion amount of the vehicle C from the lane change determination reference line Lc toward the adjacent lane Hr. The protrusion amount d2 can be calculated based on the image recognition result and the vehicle body dimensions of the vehicle C. In this embodiment, when the protrusion amount d2 becomes larger than a reference value, a lane change from the left lane H1 to the right lane H2 is determined. The reference value is, for example, 3 / 4 of the vehicle width of the vehicle C.

[0029] As a result of the lane change determination, as shown in FIG. 8, the recognition of the own lane Hj is switched from the left lane H1 to the right lane H2. Then, in this embodiment, the first own lane virtual line Kj1 is recognized along the lane boundary line G2 located at the left end of the right lane H2, which is the own lane Hj at the current time. In addition, in this embodiment, the second own lane virtual line Kj2 is recognized along the road division line G located at the right end of the right lane H2, which is the own lane Hj at the current time. In addition, in this embodiment, the first adjacent virtual line Kr1 is recognized along the outer roadway line G1 located at the left end of the left lane H1, which is the adjacent lane Hr at the current time. In addition, in this embodiment, the second adjacent virtual line Kr2 is recognized along the lane boundary line G2 located at the right end of the left lane H1, which is the adjacent lane Hr at the current time.

[0030] However, immediately after the recognition of the own lane Hj is switched, the own vehicle C is in a state of straddling the lane boundary line G2 between the left lane H1 and the right lane H2. That is, the lane change judgment does not judge the "completion" of the lane change, but judges the time when the recognition of the own lane Hj is switched. Therefore, in this embodiment, as shown in FIG. 8, the first own lane virtual line Kj1 is selected as the crossing judgment reference line Lf. Then, when the crossing judgment amount d3 at the time of completion, which is the shortest distance between the own vehicle center line CL and the crossing judgment reference line Lf, becomes larger than half the vehicle width of the own vehicle C, the crossing judgment is turned OFF, and the lane change is completed. The crossing judgment amount d3 at the time of completion can be calculated based on the image recognition result and the vehicle body dimensions of the own vehicle C. In addition, in a scene such as overtaking, the vehicle C may change lanes from the left lane H1 to the right lane H2 once, and then change lanes from the right lane H2 back to the left lane H1 again. In this case, the present embodiment selects the second adjacent virtual line Kr2 as the lane change judgment reference line Lc as shown in FIG.

[0031] 5 to 8 use a standard example in which the lane boundary line G2 between the left lane H1 and the right lane H2 is a single line in order to briefly explain the outline of the lane change judgment according to this embodiment. In this example, by machine learning the lane width and the position of the adjacent lane Hr by deep learning, even if the lane change judgment reference line Lc is the same virtual line as the crossing judgment reference line Lf, that is, the virtual line on the own lane Hj side, good lane recognition is possible before and after lane change regardless of disturbances such as lane width fluctuation. That is, the recognition of the own lane Hj is continued or maintained well before and after lane change. In contrast, FIG. 9 to FIG. 11 show an example in which a double line G3 exists between the left lane H1 and the right lane H2 as a specific example in which the effect of this embodiment is more clearly manifested. The double line G3 may be a solid line or at least one of them may be a broken line depending on the national laws and regulations of each country. Of the double lines G3, one on the left lane H1 side is called the left lane G4, and the other on the right lane H2 side is called the right lane G5. In this case, the left lane H1 is formed between the outer roadway line G1 and the left lane G4. On the other hand, the right lane H2 is formed between the right lane G5 and the road marking G to the right of it.

[0032] FIG. 9 shows a state at the time when the vehicle C starts changing lanes from the left lane H1, which is the own lane Hj, to the right lane H2, which is the adjacent lane Hr. In the example of FIG. 9, the present embodiment recognizes a first own lane virtual line Kj1 along the outer roadway line G1 located at the left end of the left lane H1, which is the own lane Hj at the current time. In addition, the present embodiment recognizes a second own lane virtual line Kj2 along the left lane dividing line G4 located at the right end of the left lane H1, which is the own lane Hj at the current time. In addition, the present embodiment recognizes a first adjacent virtual line Kr1 along the right lane dividing line G5 located at the left end of the right lane H2, which is the adjacent lane Hr at the current time. In addition, the present embodiment recognizes a second adjacent virtual line Kr2 along the road dividing line G located at the right end of the right lane H2, which is the adjacent lane Hr at the current time. In this embodiment, the first adjacent virtual line Kr1 is selected as the lane change judgment reference line Lc, and the second own lane virtual line Kj2 is selected as the crossing judgment reference line Lf.

[0033] FIG. 10 shows a state at a time point during lane change. In this embodiment, when the protruding amount d2 shown in FIG. 10 becomes larger than a reference value, a lane change from the left lane H1 to the right lane H2 is determined. Due to the lane change determination, the recognition of the own lane Hj is switched from the left lane H1 to the right lane H2 as shown in FIG. 11. Then, in this embodiment, a first own lane virtual line Kj1 is recognized along the right lane marking G5 located at the left end of the right lane H2, which is the own lane Hj at the current time. In addition, in this embodiment, a second own lane virtual line Kj2 is recognized along the road marking G located at the right end of the right lane H2, which is the own lane Hj at the current time. In addition, in this embodiment, a first adjacent virtual line Kr1 is recognized along the outer roadway marking G1 located at the left end of the left lane H1, which is the adjacent lane Hr at the current time. In addition, this embodiment recognizes a second adjacent virtual line Kr2 along the left-side dividing line G4 located at the right end of the left-side lane H1, which is the current adjacent lane Hr. Then, this embodiment selects the second adjacent virtual line Kr2 as the lane change judgment reference line Lc and selects the first own lane virtual line Kj1 as the crossing judgment reference line Lf. When the crossing judgment amount d3 at the completion time shown in Fig. 11 becomes larger than half the vehicle width of the own vehicle C, the crossing judgment is turned OFF, and the lane change is completed.

[0034] FIG. 12 shows a comparative example in which the lane change judgment reference line Lc is set on the own lane Hj side in the scene of FIG. 9. In this comparative example, the second own lane virtual line Kj2 on the lane change destination side of the first own lane virtual line Kj1 and the second own lane virtual line Kj2, which are a pair of virtual lines in the own lane Hj, is selected as the lane change judgment reference line Lc. In this case, when lane change judgment is made based on the lane change judgment reference line Lc and the recognition of the own lane Hj is switched from the left lane H1 to the right lane H2, as shown in FIG. 13, lane recognition may not be performed correctly, such that the first own lane virtual line Kj1 after switching is recognized between the left lane G4 and the right lane G5. In contrast, in this embodiment, the lane change judgment reference line Lc is set on the adjacent lane Hr side instead of the own lane Hj side, and the lane change is judged from the adjacent lane Hr instead of from the own lane Hj. This makes it possible to make the lane change judgment at the correct timing.

[0035] 14 to 16 show an example in which a guidance strip G6 exists between the left lane H1 and the right lane H2 as another specific example in which the effect of this embodiment is more clearly manifested. In the guidance strip G6, the white line extending along the left lane H1 is referred to as the left zebra forming line G7, and the white line extending along the right lane H2 is referred to as the right zebra forming line G8. In other words, the guidance strip G6 is a zebra zone formed between the left zebra forming line G7 and the right zebra forming line G8, with the left end being the left zebra forming line G7 and the right end being the right zebra forming line G8, based on the traveling direction of the vehicle C.

[0036] FIG. 14 shows a state at the time when the vehicle C starts changing lanes from the left lane H1, which is the own lane Hj, to the right lane H2, which is the adjacent lane Hr across the guidance strip G6. In the example of FIG. 14, the present embodiment recognizes a first own lane virtual line Kj1 along the outer roadway line G1 located at the left end of the left lane H1, which is the own lane Hj at the current time. In addition, the present embodiment recognizes a second own lane virtual line Kj2 along the left zebra marking line G7 located at the right end of the left lane H1, which is the own lane Hj at the current time. In addition, the present embodiment recognizes a first adjacent virtual line Kr1 along the right zebra marking line G8 located at the left end of the right lane H2, which is the adjacent lane Hr at the current time. In addition, the present embodiment recognizes a second adjacent virtual line Kr2 along the road division line G located at the right end of the right lane H2, which is the adjacent lane Hr at the current time. In this embodiment, the first adjacent virtual line Kr1 is selected as the lane change judgment reference line Lc, and the second own lane virtual line Kj2 is selected as the crossing judgment reference line Lf. In this embodiment, too, when the start crossing judgment amount d1, which is the shortest distance between the own vehicle center line CL and the crossing judgment reference line Lf, becomes smaller than half the vehicle width of the own vehicle C, the crossing judgment is turned ON.

[0037] FIG. 15 shows a state at a time point during lane change. In this embodiment, when the protruding amount d2 shown in FIG. 15 becomes larger than a reference value, a lane change from the left lane H1 to the right lane H2 is determined. Due to the lane change determination, the recognition of the own lane Hj is switched from the left lane H1 to the right lane H2 as shown in FIG. 16. Then, in this embodiment, a first own lane virtual line Kj1 is recognized along the right zebra marking line G8 located at the left end of the right lane H2, which is the own lane Hj at the current time. In addition, in this embodiment, a second own lane virtual line Kj2 is recognized along the road division line G located at the right end of the right lane H2, which is the own lane Hj at the current time. In addition, in this embodiment, a first adjacent virtual line Kr1 is recognized along the outer roadway line G1 located at the left end of the left lane H1, which is the adjacent lane Hr at the current time. In addition, this embodiment recognizes a second adjacent virtual line Kr2 along the left zebra marking G7 located at the right end of the left lane H1, which is the current adjacent lane Hr. Then, this embodiment selects the second adjacent virtual line Kr2 as the lane change judgment reference line Lc and selects the first own lane virtual line Kj1 as the crossing judgment reference line Lf. When the crossing judgment amount d3 at the completion time shown in Fig. 16 becomes larger than half the vehicle width of the own vehicle C, the crossing judgment is turned OFF, and the lane change is completed.

[0038] FIG. 17 shows a comparative example in which the lane change judgment reference line Lc is set on the own lane Hj side in the scene of FIG. 14. In this comparative example, of the pair of virtual lines in the own lane Hj, the first own lane virtual line Kj1 and the second own lane virtual line Kj2, the second own lane virtual line Kj2 on the lane change destination side is selected as the lane change judgment reference line Lc. In this case, when lane change judgment is made based on the lane change judgment reference line Lc and the recognition of the own lane Hj is switched from the left lane H1 to the right lane H2, as shown in FIG. 18, recognition loss of the virtual lines corresponding to both ends of the own lane Hj occurs. That is, even if the first own lane virtual line Kj1 is recognized along the left zebra forming line G7, the second own lane virtual line Kj2 is not recognized because the road dividing line G is not detected within a predetermined distance range on the right side of the first own lane virtual line Kj1.

[0039] Fig. 19 and Fig. 20 are flowcharts showing an example of an operation of the image recognition unit 4 according to this embodiment. In Fig. 10 and the flowcharts shown in Fig. 10, "S" is an abbreviation of "step." The same applies to other flowcharts from Fig. 19 onwards.

[0040] Referring to FIG. 19, first, in step 1101, the image recognition unit 4 acquires information on the own lane Hj and the adjacent lane Hr. That is, the image recognition unit 4 recognizes the own lane Hj and the adjacent lane Hr based on the image recognition result of the road dividing line G formed on the road surface F of the road D ahead of the own vehicle C. As a result, the own lane Hj and the adjacent lane Hr are specified, and the lateral position of the own vehicle C in the own lane Hj and virtual lines such as the first own lane virtual line Kj1 corresponding to both ends of the own lane Hj and the adjacent lane Hr are recognized. Next, in step 1102, the image recognition unit 4 determines whether the own vehicle C is straddling the road dividing line G that constitutes the own lane Hj. The processing of step 1102 corresponds to a straddling determination. If the straddling determination is ON (i.e., step 1102=YES), the image recognition unit 4 advances the processing to step 1103. On the other hand, if the straddling determination is OFF (that is, step 1102==NO), the processing from step 1103 onwards is skipped.

[0041] In step 1103, the image recognition unit 4 acquires a reference position for lane change judgment. That is, the image recognition unit 4 sets a lane change judgment reference line Lc. At this time, the image recognition unit 4 sets the lane change judgment reference line Lc at a position offset by an arbitrary width from the own lane Hj. Specifically, referring to FIG. 20, in step 1201, the image recognition unit 4 judges whether or not there is adjacent lane information. If there is adjacent lane information (i.e., step 1201=YES), the image recognition unit 4 executes the process of step 1202. In step 1202, the image recognition unit 4 sets a reference position, i.e., the lane change judgment reference line Lc, at a position offset by the position difference between the own lane Hj and the adjacent lane Hr. In other words, the image recognition unit 4 selects a virtual line generated based on the position difference between the own lane Hj and the adjacent lane Hr as the lane change judgment reference line Lc. If there is no adjacent lane information (ie, step 1201=NO), the process of step 1202 is skipped.

[0042] Referring again to FIG. 19, following the process of step 1103, the process of step 1104 is executed. In step 1104, the image recognition unit 4 executes lane change determination based on the positional relationship between the reference position, i.e., the lane change determination reference line Lc, and the vehicle C. When the lane change determination is not established (i.e., step 1104=NO), the process of step 1105 is skipped. On the other hand, when the lane change determination is established (i.e., step 1104=YES), the image recognition unit 4 executes the process of step 1105. In step 1105, the image recognition unit 4 switches the lane information between the own lane Hj and the adjacent lane Hr. That is, the image recognition unit 4 switches the recognition of the first own lane virtual line Kj1, the second own lane virtual line Kj2, the first adjacent virtual line Kr1, and the second adjacent virtual line Kr2 from information before the lane change (see, for example, FIG. 14) to information after the lane change (see, for example, FIG. 16).

[0043] In this manner, in this embodiment, lane changes are determined based on the adjacent lane Hr, not the own lane Hj, which improves the accuracy of lane change determination, thereby enabling lane-related control functions to be executed with high accuracy.

[0044] (First embodiment: Modification 1) As shown in Fig. 14, when a guidance strip G6 exists between the left lane H1, which is the own lane Hj, and the right lane H2, which is the adjacent lane Hr, the position difference between the own lane Hj and the adjacent lane Hr becomes large. In such a case, that is, when the position difference is larger than a predetermined value, the judgment reference line selection unit 45 may select a virtual line generated midway between the own lane Hj and the adjacent lane Hr as the lane change judgment reference line Lc, as shown in Fig. 21. This makes it possible to improve the accuracy of lane change judgment, and thus to execute lane-related control functions with high accuracy.

[0045] (First embodiment: Modification 2) As shown in FIG. 22, there are cases where a road dividing line G is drawn along the road extension direction inside the guidance strip G6. In such a case, that is, when a road dividing line G exists between the own lane Hj and the adjacent lane Hr in a form different from the normal lane boundary line G2, the judgment reference line selection unit 45 may select a virtual line corresponding to the road dividing line G as the lane change judgment reference line Lc. This makes it possible to improve the accuracy of the lane change judgment, and thus makes it possible to execute the lane-related control function with high accuracy. In this case, the lane change judgment reference line Lc may be provided at the edge position of the road dividing line G or at a position offset by a predetermined amount therefrom, or may be provided inside the road dividing line G.

[0046] (First embodiment: Modification 3) FIG. 23 shows a situation in which the vehicle C is moving from the straight lane H3 to the branch lane H4, i.e., changing lanes. In the example of FIG. 23, the branch lane H4 is provided so as to extend from the straight lane H3 toward the left front. The branch lane H4 is a lane H between the branch lane left dividing line G11 and the branch lane right dividing line G12. The branch lane left dividing line G11 is a road dividing line G provided so as to define the left end of the branch lane H4. The branch lane right dividing line G12 is a road dividing line G provided so as to define the right end of the branch lane H4. In this case, currently, i.e., before the lane change determination, the vehicle C's lane Hj is the straight lane H3, and the adjacent lane Hr to which the lane is to be changed is the branch lane H4. Therefore, the first adjacent virtual line Kr1 is set along the branch lane left dividing line G11. In addition, the second adjacent virtual line Kr2 is set along the branch lane right dividing line G12.

[0047] In the example of FIG. 23, a first own lane virtual line Kj1, which is a virtual line corresponding to the boundary position of the own lane Hj on the adjacent lane Hr side, and a second adjacent virtual line Kr2, which is a virtual line corresponding to the boundary position of the own lane Hj on the adjacent lane Hr side, intersect at an intersection point X. The intersection point X is located in an intersection area where the own lane Hj, i.e., the straight lane H3, and the adjacent lane Hr, i.e., the branch lane H4, intersect. The second adjacent virtual line Kr2, which is located on the front side of the intersection point X, i.e., on the lower side in the figure, invades the own lane Hj and is located inside the own lane Hj more than the first own lane virtual line Kj1. For this reason, in such a case, the entire second adjacent virtual line Kr2 on the own lane Hj side of the pair of virtual lines in the adjacent lane Hr cannot be selected, i.e., set, as the lane change judgment reference line Lc. Therefore, when the own lane Hj intersects with the adjacent lane Hr, the judgment reference line selection unit 45 selects, from the first own lane virtual line Kj1 or the second adjacent virtual line Kr2, the one that is farther laterally from the own lane Hj as the lane change judgment reference line Lc. Specifically, the lane change judgment reference line Lc is formed in a broken line shape by a portion of the first own lane virtual line Kj1 on the front side of the intersection point X and a portion of the second adjacent virtual line Kr2 on the rear side of the intersection point X, i.e., on the traveling destination side of the own vehicle C. This makes it possible to improve the accuracy of lane change judgment, and thus makes it possible to execute lane-related control functions with high accuracy.

[0048] (First embodiment: Modification 4) When the length of the intersection area between the straight lane H3 and the branch lane H4 in the road extension direction is long, the entire intersection area may not fit within the shooting range E of the image shooting unit 2. On the other hand, the branch road structure can be easily acquired from map information, especially HD map information. HD stands for High Definition. Therefore, as shown in FIG. 24, a virtual line generated from the boundary between the straight lane H3 and the branch lane H4 acquired from the map information can be selected, that is, set as the lane change judgment reference line Lc. This makes it possible to improve the accuracy of lane change judgment even when the length of the intersection area between the straight lane H3 and the branch lane H4 in the road extension direction is long, and thus makes it possible to execute lane-related control functions with high accuracy.

[0049] Second Embodiment This embodiment is a modification of a part of the configuration in the first embodiment. Specifically, as shown in FIG. 25, the driving support device 1 includes an image capturing unit 2, a vehicle state quantity sensor unit 3, an image recognition unit 4, a lane recognition unit 60, and an alarm / vehicle control unit 5. The image recognition unit 4 outputs the image recognition result to the alarm / vehicle control unit 5 and the lane recognition unit 60. The lane recognition unit 60 recognizes the lane H on the road D based on the white line recognition result by the image recognition unit 4, i.e., the recognition result of the road dividing line G. That is, in this embodiment, part of the functions of the image recognition unit 4 in the first embodiment, such as the lane recognition function, is provided to the lane recognition unit 60. In other words, in this embodiment, the lane recognition unit 60 recognizes the lane H based on the white line recognition result by the image recognition unit 4. The lane recognition unit 60 outputs the recognition result to the alarm / vehicle control unit 5. With such a configuration, the same effects as those of the first embodiment can be achieved.

[0050] In this way, the white line recognition function and the lane recognition function can be configured as the same or separate functional blocks or functional modules. The white line recognition function is a function that recognizes the position, shape, and line type of white lines, i.e., road dividing lines G. The lane recognition function is a function that recognizes the number, position, arrangement, and type of lanes H on road D in the road width direction. Types of lanes H include driving lanes, overtaking lanes, etc. The lane recognition function also includes a function to identify the current lane Hj, a lane change determination function, and a function to switch between the current lane Hj and adjacent lane Hr due to lane changes.

[0051] Third embodiment FIG. 25 shows a schematic configuration of the driving support device 1 according to this embodiment. FIG. 26 shows a schematic functional configuration realized by the image recognition unit 4 and the lane recognition unit 60 shown in FIG. 25. This embodiment relates to driving behavior control of the host vehicle C in so-called EDSS or MRM. EDSS stands for Emergency Driving Stopping System. MRM stands for Minimal Risk Maneuver. That is, this embodiment relates to functions and configurations for the host vehicle C, which is traveling in the driving lane H5 as the host lane Hj, to shift from the driving lane H5 to the road shoulder H6, i.e., to make an emergency stop on the road shoulder H6, as shown in FIG. 27, and to "change lanes."

[0052] Conventional lane-related control functions are designed on the assumption that a lane H exists in the lane change destination from the own lane Hj, and the road shoulder H6 has not been recognized in such control functions. On the other hand, when the own vehicle C shifts from the driving lane H5 to the road shoulder H6 for an emergency stop on the road shoulder H6 in an emergency, it is required to execute a smooth "lane change" to the road shoulder H6 by recognizing the road shoulder H6 as the lane H. Therefore, in this embodiment, the lane recognition unit 60 recognizes the own lane Hj based on the recognition result of the road division line G by the image recognition unit 4, and recognizes the road shoulder H6 as an adjacent lane Hr as a lane change destination from the own lane Hj when making an emergency evacuation to the road shoulder H6. Specifically, referring to FIG. 26, the image recognition unit 4 includes a division line detection unit 41 and a line type determination unit 42. The lane recognition unit 60 includes a road edge detection unit 61, a road shoulder virtual line generation unit 62, a road shape estimation unit 43, and a lane change determination unit 44. Each unit constituting this embodiment will be described below with reference to Fig. 25 to Fig. 27. Note that "emergency evacuation" can also be rephrased as "emergency evacuation".

[0053] The road edge detection unit 61 is configured to detect the road edge boundary N1 shown in Fig. 27. The road edge boundary N1 is a boundary in the road width direction of an area on the road D where vehicles can travel. Specifically, the road edge detection unit 61 detects the road edge boundary N1 based on the recognition results of the captured images of the road D and three-dimensional objects such as guardrails around the road D. The recognition of the road edge boundary N1 and its types will be described later.

[0054] The shoulder virtual line generating unit 62 generates a right shoulder virtual line P1 and a left shoulder virtual line P2, which are virtual lines corresponding to the boundary position of the shoulder H6 in the road width direction. The right shoulder virtual line P1 is a virtual line corresponding to the boundary position of the right side of the shoulder H6, i.e., the driving lane H5 side in the road width direction. The left shoulder virtual line P2 is a virtual line corresponding to the boundary position on the opposite side of the right shoulder virtual line P1 in the road width direction, in other words, the boundary position on the left side of the shoulder H6, i.e., the road edge boundary N1 side. The right shoulder virtual line P1 is used as a reference line for determining whether the host vehicle C will "change lanes" from the driving lane H5, which is the host lane Hj, to the shoulder H6, which is the adjacent lane Hr. The right shoulder virtual line P1 and the left shoulder virtual line P2 are used as a reference for control of "lane keeping" or "lane departure suppression" to safely stop the host vehicle C within the shoulder H6. However, the shoulder virtual line generating unit 62 is configured not to set the shoulder right virtual line P1 and the shoulder left virtual line P2 except in an emergency.

[0055] In this embodiment, the shoulder virtual line generating unit 62 generates a shoulder right virtual line P1 and a shoulder left virtual line P2 based on the recognition result of the own lane Hj. Specifically, for example, the shoulder virtual line generating unit 62 sets the shoulder right virtual line P1 at a position away from a first own lane virtual line Kj1 corresponding to the boundary position on the shoulder H6 side of the own lane Hj by a road side margin W1 having a predetermined width corresponding to the width of the road dividing line G. In addition, the shoulder virtual line generating unit 62 sets the shoulder left virtual line P2 at a position away from the road edge boundary N1 by a road edge side margin W2 having a predetermined width. The road side margin W1 and the road edge side margin W2 may be the same or different. Alternatively, for example, the shoulder virtual line generating unit 62 may set the shoulder right virtual line P1 at the outer edge, i.e., the left edge, of the outer roadway line G1 located on the shoulder H6 side of the pair of road dividing lines G that define the own lane Hj, or at a position offset a predetermined amount further outside, i.e., to the left, of the own lane Hj.

[0056] FIG. 28 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this embodiment. In the figure and the following description, LC stands for Lane Change. Referring to FIG. 28, first, in step 2101, the lane recognition unit 60 turns off the LC determination flag. The LC determination flag is turned on when a lane change is determined. Next, in step 2102, the lane recognition unit 60 determines whether or not there is an emergency evacuation signal. The emergency evacuation signal is a signal that is generated and output when there is a need to urgently evacuate the host vehicle C to the road shoulder H6, and is generated, for example, by a driver status monitor function used for automatic driving or advanced driving assistance. That is, when the lane recognition unit 60 receives an emergency evacuation signal from a driving control ECU in automatic driving or advanced driving assistance, or an ECU provided in a driver status monitor device, the lane recognition unit 60 determines "YES" in step 2102. ECU stands for Electronic Control Unit.

[0057] If an emergency evacuation signal has not been received (i.e., step 2102=NO), the lane recognition unit 60 advances the process to step 2103. In step 2103, the lane recognition unit 60 executes the normal LC mode. In this case, the shoulder H6 is not recognized as the lane H to which the lane is to be changed. In contrast, if an emergency evacuation signal has been received (i.e., step 2102=YES), the lane recognition unit 60 advances the process to steps 2104 and 2105. In step 2104, the lane recognition unit 60 executes the emergency LC mode. As a result, the shoulder H6 is recognized as the lane H to which the lane is to be changed for emergency evacuation. That is, in step 2105, the lane recognition unit 60 generates a shoulder right virtual line P1 and a shoulder left virtual line P2.

[0058] In this way, according to this embodiment, by properly recognizing the road shoulder H6 as a lane change destination during an emergency evacuation, a smooth "lane change" to the road shoulder H6 and lateral position control for maintaining the host vehicle C within the road shoulder H6 can be properly realized. Therefore, according to this embodiment, an emergency stop of the host vehicle C on the road shoulder H6 can be properly performed.

[0059] (Third embodiment: modified example) Various modifications of this embodiment will be described below.

[0060] (Variation 1) As described above, the outer roadway line G1, which is a road dividing line G, is usually provided between the travel lane H5 and the shoulder H6. However, for example, the outer roadway line G1 may be blurred or disappear due to aging, and the road dividing line G corresponding to the outer roadway line G1 may not be detected. Therefore, in this modified example, when a road dividing line G is not detected near the shoulder right virtual line P1, the shoulder right virtual line P1 is determined to correspond to the boundary position of the shoulder H6. This allows the vehicle C to make an emergency stop on the shoulder H6 in a good manner.

[0061] Fig. 29 shows a configuration according to this modified example. As shown in Fig. 29, the lane recognition unit 60 includes a virtual line determination unit 63. When a road dividing line G is not detected near a shoulder right virtual line P1, the virtual line determination unit 63 registers the shoulder right virtual line P1 as a lane H. In other words, the virtual line determination unit 63 determines the shoulder right virtual line P1 as one that corresponds to the boundary position of the shoulder H6.

[0062] FIG. 30 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this modified example. In the flowchart, the road dividing line G is written as a "boundary line" for the sake of simplicity. This is based on the fact that the lane dividing line G2 as the road dividing line G is the boundary line between two adjacent lanes H, and the outer roadway line G1 as the road dividing line G is the boundary line between the driving lane H5 and the shoulder H6, and therefore these are often conventionally referred to as "boundary lines" by those skilled in the art. In addition, such conventional expressions as "boundary lines" may be used in the following explanations in the specification. However, instead of the road dividing line G, a virtual line corresponding to the outer edge of a certain area or the boundary between adjacent areas may also be referred to as a "boundary line".

[0063] Referring to FIG. 30, first, in step 2201, the lane recognition unit 60 acquires own lane information. That is, the lane recognition unit 60 recognizes the own lane Hj based on the white line recognition result by the image recognition unit 4. Next, in step 2202, the lane recognition unit 60 acquires white line width information. That is, the lane recognition unit 60 estimates the width dimension of the road dividing line G on the road D on which the own vehicle C is currently traveling based on the white line recognition result by the image recognition unit 4. Next, in step 2203, the lane recognition unit 60 calculates a shoulder right virtual line P1, which is a virtual line on the side closer to the own lane Hj on the shoulder H6 adjacent to the own lane Hj, at the position of the recognition position of the own lane Hj + the white line width. Then, in step 2204, the lane recognition unit 60 determines whether or not a road dividing line G is detected near the calculated shoulder right virtual line P1.

[0064] If there is a road dividing line G near the shoulder right virtual line P1 (i.e., step 2204=YES), the lane recognition unit 60 advances the process to step 2205. In step 2205, the lane recognition unit 60 executes the process as the virtual line determination unit 63. That is, the lane recognition unit 60 determines the shoulder right virtual line P1 as corresponding to the boundary position of the shoulder H6. In contrast, if there is no road dividing line G near the shoulder right virtual line P1 (i.e., step 2204=NO), the lane recognition unit 60 advances the process to step 2206. In step 2206, the lane recognition unit 60 does not determine the shoulder right virtual line P1 as corresponding to the boundary position of the shoulder H6. That is, the lane recognition unit 60 does not execute the process as the virtual line determination unit 63.

[0065] (Variation 2) For example, due to reasons such as the outer lane G1 being blurred or disappearing due to aging, recognition of the own lane Hj, which is the basis for generating the shoulder right virtual line P1, may become unstable or the own lane Hj may not be recognized. In this way, when fluttering occurs in the recognition of the own lane Hj, it is preferable not to generate the shoulder right virtual line P1 until the fluttering is eliminated and the recognition of the own lane Hj is stabilized. Therefore, in this modification, as shown in FIG. 31, the lane recognition unit 60 includes a lane fluttering determination unit 64. The lane fluttering determination unit 64 determines whether fluttering occurs in the recognition of the own lane Hj. The lane recognition unit 60 according to this modification may further include a virtual line determination unit 63 shown in FIG. 29.

[0066] FIG. 32 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this modified example. Referring to FIG. 32, first, in step 2301, the lane recognition unit 60 acquires the own lane information. Next, in step 2302, the lane recognition unit 60 judges whether or not there is fluttering in the recognition of the own lane Hj. If there is fluttering in the recognition of the own lane Hj (i.e., step 2302=YES), the lane recognition unit 60 advances the process to step 2303. In step 2303, the lane recognition unit 60 does not generate the right virtual shoulder line P1. In contrast, if there is no fluttering in the recognition of the own lane Hj (i.e., step 2302=NO), the lane recognition unit 60 advances the process to step 2304 and step 2305. In step 2304, the lane recognition unit 60 generates the right virtual shoulder line P1. That is, the lane recognition unit 60 executes the process as the shoulder virtual line generation unit 62. In step 2305, the lane recognition unit 60 registers the generated shoulder right virtual line P1 and shoulder left virtual line P2 as adjacent lane information.

[0067] (Variation 3) When detection information of a road dividing line G is output near a virtual line corresponding to the shoulder H6 determined by the virtual line determination unit 63, there is a possibility that the road dividing line G may be erroneously recognized or may have a flutter in recognition. Therefore, it is preferable to determine the reliability of such detection information and determine the continuation of use of the virtual line for each frame. Therefore, in this modification, as shown in FIG. 33, the lane recognition unit 60 includes a virtual line determination unit 63 and a reliability determination unit 65. When a road dividing line G is detected near a virtual line corresponding to the shoulder H6, the reliability determination unit 65 determines the reliability of the detection of the road dividing line G. Then, the lane recognition unit 60 registers either the virtual line or the road dividing line G as corresponding to the boundary position of the shoulder H6 based on the reliability. Note that the lane recognition unit 60 according to this modification may further include a lane flutter determination unit 64 shown in FIG. 31.

[0068] Fig. 34 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this modified example. Referring to Fig. 34, first, in step 2401, the lane recognition unit 60 acquires determined virtual line information. That is, the lane recognition unit 60 acquires information on a virtual line corresponding to the shoulder H6, determined by the virtual line determination unit 63. Next, in step 2402, the lane recognition unit 60 determines whether or not there is a boundary line near the determined virtual line.

[0069] If there is no boundary line near the confirmed virtual line (i.e., step 2402=NO), the lane recognition unit 60 advances the process to step 2403. In step 2403, the lane recognition unit 60 registers the confirmed virtual line as one corresponding to the boundary position of the road shoulder H6. In contrast, if there is a boundary line near the confirmed virtual line (i.e., step 2402=YES), the lane recognition unit 60 advances the process to steps 2404 and 2405. In step 2404, the lane recognition unit 60 determines the reliability of the boundary line near the confirmed virtual line. That is, the lane recognition unit 60 executes the process as the reliability determination unit 65. Then, in step 2405, the lane recognition unit 60 determines whether the reliability is high or not, that is, whether the reliability is higher than a threshold value or not. If the reliability is high (i.e., step 2405=YES), the lane recognition unit 60 advances the process to step 2406. In step 2406, the lane recognition unit 60 registers a boundary line near the determined virtual line as a boundary position of the road shoulder H6. On the other hand, if the reliability is low (i.e., step 2405=NO), the lane recognition unit 60 advances the process to step 2403.

[0070] (Variation 4) When the virtual line is determined by the virtual line determination unit 63, the recognition of the original own lane Hj may become unstable or unrecognized. In this case, it is possible to generate the reference own lane Hj from the virtual line. In this modification, the lane recognition unit 60 includes a virtual line determination unit 63 and a virtual own lane generation unit 66 as shown in FIG. 35. When the virtual line is determined to correspond to the boundary position of the shoulder H6 and the recognition of the own lane Hj becomes unstable or unrecognized, the virtual own lane generation unit 66 recognizes the own lane Hj based on the virtual line. Note that the lane recognition unit 60 according to this modification may further include an own lane flutter determination unit 64 shown in FIG. 31 and / or a reliability determination unit 65 shown in FIG. 33.

[0071] FIG. 36 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this modified example. Referring to FIG. 36, first, in step 2501, the lane recognition unit 60 acquires determined virtual line information. That is, the lane recognition unit 60 acquires information on the virtual line corresponding to the shoulder H6 determined by the virtual line determination unit 63. Next, in step 2502, the lane recognition unit 60 acquires own lane information. Then, in step 2503, the lane recognition unit 60 determines whether the recognition of the own lane Hj is unstable or unrecognized. If the recognition of the own lane Hj is unstable or unrecognized (i.e., step 2503=YES), the lane recognition unit 60 advances the process to step 2504. In step 2504, the lane recognition unit 60 executes the process as the virtual own lane generation unit 66. On the other hand, if the own lane Hj is stably recognized (that is, step 2503=NO), the lane recognition unit 60 skips the process of step 2504.

[0072] (Variation 5) The above description of this embodiment is mainly about the generation or recognition of the right virtual shoulder line P1, which is one of a pair of virtual lines corresponding to the boundary position of the shoulder H6 in the road width direction. In contrast, the following describes the selection or position setting of the left virtual shoulder line P2, which is the other virtual line.

[0073] Fig. 37 shows a functional configuration of the lane recognition unit 60 according to this modification. As shown in Fig. 37, the shoulder virtual line generation unit 62 includes a shoulder edge virtual line generation unit 621 and a road edge virtual line generation unit 622. The shoulder edge virtual line generation unit 621 generates a shoulder right virtual line P1 as a first virtual line. The shoulder right virtual line P1 is generated as in the above modifications 1 to 4. The road edge virtual line generation unit 622 generates a shoulder left virtual line P2 as a second virtual line. Specifically, the road edge virtual line generation unit 622 sets the position of the shoulder left virtual line P2 in the road width direction based on the detection result of the road edge boundary N1 corresponding to the end of the road D in the road width direction. In addition, the lane recognition unit 60 in this modified example may further include at least one of the virtual line determination unit 63 shown in Figure 29, the own lane flutter determination unit 64 shown in Figure 31, the reliability determination unit 65 shown in Figure 33, and the virtual own lane generation unit 66 shown in Figure 35.

[0074] FIG. 38 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this modified example. Referring to FIG. 38, first, in step 2601, the lane recognition unit 60 acquires own lane information that is a reference for generating the left virtual shoulder line P2. Next, in step 2602, the lane recognition unit 60 judges whether or not there is a boundary line near the left end of the road D in the road width direction. If there is a boundary line (i.e., step 2602=NO), the lane recognition unit 60 advances the process to step 2603 and step 2604. In step 2603, the lane recognition unit 60 acquires boundary line information, i.e., recognition information of the outer road line G1. In step 2604, the lane recognition unit 60 judges whether or not the fluttering of the boundary line recognition is large. If the fluttering is not large (i.e., step 2604=NO), the lane recognition unit 60 advances the process to step 2605. In step 2605, the lane recognition unit 60 adopts the boundary line as the road edge, i.e., the position of the edge of the shoulder H6 opposite to the right virtual shoulder line P1. If the fluttering is large (i.e., step 2604=YES), the lane recognition unit 60 advances the process to step 2606. In step 2606, the lane recognition unit 60 generates the left virtual shoulder line P2, which is a road edge virtual line. If there is no boundary line (i.e., step 2602=YES), the lane recognition unit 60 advances the process to step 2306.

[0075] (Variation 6) As shown in Figures 39 to 41, this modification secures as much vehicle entry space as possible on the road shoulder H6 by changing the road shoulder width W3, i.e., the width recognized as the road shoulder H6, depending on the type of road edge boundary N1. Specifically, for example, as shown in Figure 39, the road edge boundary N1 may be formed by a roadside structure N2 such as a guardrail. Alternatively, as shown in Figure 40, the road edge boundary N1 may be formed as a boundary between a paved surface N3 and an unpaved surface N4. The former is referred to as a solid object road edge boundary N11, and the latter is referred to as a planar road edge boundary N12.

[0076] As shown in Fig. 39, when the road edge boundary N1 is the solid object road edge boundary N11, it is preferable to set the road edge side margin W2 as large as possible in consideration of avoiding contact between the vehicle body of the vehicle C and the roadside structure N2 and convenience for passengers to get on and off. On the other hand, when the road edge boundary N1 is the planar road edge boundary N12, the road edge side margin W2 is set smaller as shown in Fig. 41 than the same as the solid object road edge boundary N11 as shown in Fig. 40, thereby making it possible to maximize the vehicle entry space on the shoulder H6. Therefore, when the distance between the shoulder right virtual line P1 and the shoulder left virtual line P2 is equal to or less than the threshold value and the road edge boundary N1 is the planar road edge boundary N12, the lane recognition unit 60 sets the road edge side margin W2 smaller than the solid object road edge boundary N11. Alternatively, in this case, the lane recognition unit 60 recognizes the area between the shoulder right virtual line P1 and the planar road edge boundary N12 as the shoulder H6. That is, the lane recognition unit 60 makes the shoulder left virtual line P2 approximately coincide with the planar road edge boundary N12.

[0077] FIG. 42 is a flowchart showing an example of an operation of the lane recognition unit 60 according to this modified example. Referring to FIG. 42, first, in step 2701, the lane recognition unit 60 acquires recognition information of the shoulder left virtual line P2, which is a road edge virtual line. Next, in step 2702, the lane recognition unit 60 acquires recognition information of the shoulder right virtual line P1, which is the other virtual line. Then, in step 2703, the lane recognition unit 60 calculates the shoulder width W3. That is, the lane recognition unit 60 calculates the interval between the shoulder right virtual line P1 and the shoulder left virtual line P2. Then, in step 2704, the lane recognition unit 60 determines whether the calculated shoulder width W3 is equal to or greater than a threshold value.

[0078] If the calculated shoulder width W3 is equal to or greater than the threshold (i.e., step 2704=YES), the lane recognition unit 60 advances the process to step 2705. In step 2705, the lane recognition unit 60 recognizes the shoulder H6 using the left virtual shoulder line P2, which is a road edge virtual line. On the other hand, if the calculated shoulder width W3 is less than the threshold (i.e., step 2704=NO), the lane recognition unit 60 advances the process to step 2706. In step 2706, the lane recognition unit 60 determines whether the road edge boundary N1 is the planar road edge boundary N12. If the road edge boundary N1 is the planar road edge boundary N12 (i.e., step 2706=YES), the lane recognition unit 60 advances the process to step 2707. In step 2707, the lane recognition unit 60 recognizes the shoulder H6 using the planar road edge boundary N12. On the other hand, if the road edge boundary N1 is not the planar road edge boundary N12 (i.e., step 2706=NO), the lane recognition unit 60 advances the process to step 2708. In step 2708, the lane recognition unit 60 does not recognize the boundary position on the opposite side of the shoulder right virtual line P1 in the shoulder H6.

[0079] (Other variations) The present invention is not limited to the above-mentioned embodiment and modified examples. Therefore, the above-mentioned embodiment and the like can be modified as appropriate. Representative other modified examples will be described below. In the following description of the modified examples, differences from the above-mentioned embodiment and the like will be mainly described. In addition, the same reference numerals are used for parts that are the same or equivalent in the above-mentioned embodiment and the following modified examples. Therefore, in the following description of the modified examples, the description in the above-mentioned embodiment and the like may be used as appropriate for components having the same reference numerals as the above-mentioned embodiment and the like, unless there is a technical contradiction or special additional explanation.

[0080] The present invention is not limited to the specific device configuration shown in the above embodiment. For example, the image capturing unit 2 is not limited to a so-called front camera, and may be a full-surround camera. There is also no particular limitation on the number and mounting positions of the image capturing units 2.

[0081] The image recognition unit 4 may be provided in the warning / vehicle control unit 5. That is, the image recognition unit 4 may be an image recognition function provided in the autonomous driving ECU or the driving assistance ECU. The same applies to the lane recognition unit 60. As described above, the image recognition unit 4 and the lane recognition unit 60 do not have to be provided separately. That is, the road edge detection unit 61, the road shoulder virtual line generation unit 62, the virtual line determination unit 63, the own lane fluttering determination unit 64, the reliability determination unit 65, and the virtual own lane generation unit 66 may be provided as functions of the image recognition unit 4. Alternatively, for example, the division line detection unit 41, the line type determination unit 42, the road shape estimation unit 43, and the lane change determination unit 44 may be provided as functions of the lane recognition unit 60.

[0082] All or part of the image recognition unit 4 may be configured to include a digital circuit configured to enable the above-mentioned operation, such as an ASIC or FPGA. ASIC is an abbreviation for Application Specific Integrated Circuit. FPGA is an abbreviation for Field Programmable Gate Array. That is, in the image recognition unit 4, an on-board microcomputer part and a digital circuit part may coexist. The same applies to the lane recognition unit 60.

[0083] The program according to the present invention, which enables the execution of various operations, procedures, or processes described in the above embodiment, can be downloaded or upgraded via V2X communication. V2X is an abbreviation for Vehicle to X. Alternatively, such a program can be downloaded or upgraded via a terminal device provided in a manufacturing plant, a maintenance plant, a dealer, or the like of the vehicle C. Such a program may be stored on a memory card, an optical disk, a magnetic disk, or the like.

[0084] In this way, each of the above functional configurations and processes may be realized by a dedicated computer provided by configuring a processor and a memory programmed to execute one or more functions embodied in a computer program. Alternatively, each of the above functional configurations and processes may be realized by a dedicated computer provided by configuring a processor with one or more dedicated hardware logic circuits. Alternatively, each of the above functional configurations and processes may be realized by one or more dedicated computers configured by combining a processor and a memory programmed to execute one or more functions with a processor configured with one or more hardware logic circuits. Furthermore, the computer program may be stored in a computer-readable non-transient tangible storage medium as instructions executed by the computer. In other words, each of the above functional configurations and processes may be expressed as a computer program including a procedure for realizing it, or as a non-transient tangible storage medium storing the program.

[0085] Therefore, the functional blocks shown in each drawing are merely set for the sake of convenience in order to facilitate understanding of the contents of the present invention. Even if these functional blocks are not actually implemented as subroutines or hardware within the image recognition unit 4 or the lane recognition unit 60, the requirements of the present invention can be met as long as the functions or processes prescribed in the present invention are implemented.

[0086] The present invention is not limited to the specific operation modes shown in the above embodiment. For example, in the above embodiment, the control on the so-called road D with two lanes on each side is illustrated, but the present invention is not limited to such a mode. That is, for example, the present invention can be well applied to lane-related control functions on the road D with three or more lanes on each side. In addition, the above embodiment is based on the premise that vehicles drive on the left side in accordance with the Road Traffic Act of Japan, but the present invention is not limited to such a mode. That is, the present invention can be well applied to the so-called case of vehicles driving on the right side. In this case, the illustration of FIG. 5 and the like may be mirror-inverted left and right.

[0087] Needless to say, the elements constituting the above embodiment or its modified examples are not necessarily essential, except when expressly stated as essential or when clearly considered essential in principle. Furthermore, when the numbers, amounts, ranges, etc. of the components are mentioned, the present disclosure is not limited to the specific values, except when expressly stated as essential or when clearly limited to a specific value in principle. Similarly, when the shapes, directions, positional relationships, etc. of the components are mentioned, the present disclosure is not limited to the shapes, directions, positional relationships, etc., except when expressly stated as essential or when clearly limited to a specific shape, direction, positional relationship, etc. in principle.

[0088] The modified examples are not limited to the above examples. For example, multiple embodiments may be combined with each other as long as there is no technical contradiction. Furthermore, multiple specific examples or modified examples may be combined with each other as long as there is no technical contradiction. Furthermore, all or part of the embodiment and all or part of the modified examples may be combined with each other as long as there is no technical contradiction. [Explanation of symbols]

[0089] 1 Driving assistance devices 2. Image capture unit 4. Image Recognition Section 44 Lane change judgment unit 45 Judgment reference line selection section 60 Lane Recognition Unit 62 Road shoulder virtual line generation section C Vehicle Hj own lane Hr adjacent lane

Claims

1. A driving assistance device (1) that assists driving of a vehicle (C), a lane change determination unit (44) that determines that the vehicle has changed lanes from a current lane (Hj) in which the vehicle is currently traveling to an adjacent lane (Hr) based on a recognition result based on a photographed image of a road dividing line (G) provided on a road surface (DF) of a road (D) ahead of the vehicle, The lane change determination unit includes a determination reference line selection unit (45) that selects a determination reference line (Lc) that is a reference for lane change determination, the determination reference line selection unit selects a virtual line that is offset by a predetermined amount from the own lane from among virtual lines that are line information that are candidates for boundaries of a lane (H) included in the recognition result; Driving assistance device.

2. the determination reference line selection unit selects the virtual line generated based on a position difference between the own lane and the adjacent lane. The driving assistance device according to claim 1 .

3. the determination reference line selection unit acquires the position difference from map information; The driving assistance device according to claim 2 .

4. the determination reference line selection unit selects the virtual line generated midway between the own lane and the adjacent lane when the position difference is greater than a predetermined value. The driving assistance device according to claim 2 .

5. the determination reference line selection unit, when another road dividing line exists between the road dividing line constituting the own lane and the road dividing line constituting the adjacent lane, selects the virtual line corresponding to the other road dividing line. The driving assistance device according to claim 1 .

6. When the own lane and the adjacent lane intersect, the judgment reference line selection unit selects, as the judgment reference line, one of a first virtual line (Kj1) that is the virtual line corresponding to a boundary position of the own lane on the adjacent lane side and a second virtual line (Kr2) that is the virtual line corresponding to a boundary position of the adjacent lane on the own lane side, which is farther lateral from the own lane. The driving assistance device according to claim 1 .

7. the determination reference line selection unit, when the adjacent lane is a branching road intersecting the own lane, selects the virtual line generated from a boundary between the own lane and the adjacent lane acquired from map information. The driving assistance device according to claim 1 .

8. A lane recognition unit (60) is further provided which recognizes the own lane based on the captured image and recognizes the road shoulder (H6) as the adjacent lane to which the own lane is to be changed when an emergency evacuation to the road shoulder is performed, The lane recognition unit includes a virtual line generation unit (62) that generates the virtual line corresponding to a boundary position of the road shoulder, The virtual line generating unit generates the virtual line based on a result of the recognition of the own lane. The driving assistance device according to claim 1 .

9. the virtual line generating unit sets the virtual line at a position spaced a predetermined width away from an own lane virtual line (Kj1) corresponding to a boundary position on the shoulder side of the own lane, the virtual line corresponding to a width of the road dividing line; When the road dividing line is not detected near the virtual line, the lane recognition unit determines that the virtual line corresponds to a boundary position of the road shoulder. A driving assistance device according to claim 8.

10. The lane recognition unit does not generate the virtual line when the recognition of the own lane is unstable or unrecognized. A driving assistance device according to claim 8.

11. the lane recognition unit includes a reliability determination unit (63) that, when the road dividing line is detected near the virtual line, determines a reliability of the detection of the road dividing line, registering either the virtual line or the road dividing line as a line corresponding to a boundary position of the road shoulder based on the reliability; A driving assistance device according to claim 9.

12. the lane recognition unit recognizes the own lane based on the virtual line when the virtual line is determined to correspond to a boundary position of the road shoulder and recognition of the own lane becomes unstable or unrecognized. A driving assistance device according to claim 9.

13. The virtual line generating unit A first virtual line (P1) is generated as the virtual line corresponding to a boundary position on the driving lane (H1) side of the road shoulder in a road width direction, and a second virtual line (P2) is generated as the virtual line corresponding to a boundary position on the opposite side thereof; A position of the second virtual line in the road width direction is set based on a detection result of a road edge boundary (N1) corresponding to an end of the road in the road width direction. A driving assistance device according to claim 8.

14. When the distance between the first virtual line and the second virtual line is equal to or less than a threshold value and the road edge boundary is a planar road edge boundary (N12) that is a boundary of a paved surface (N2), the lane recognition unit recognizes the area between the first virtual line and the planar road edge boundary as the road shoulder. A driving assistance device according to claim 13.

15. A driving assistance program executed by a driving assistance device (1) that assists in driving a vehicle (C), The process executed by the driving assistance device is as follows: a lane change determination process for determining that the vehicle has changed lanes from a current lane (H2) in which the vehicle is currently traveling to an adjacent lane (H3) based on a recognition result based on a photographed image of a road dividing line (G) provided on a road surface (DF) of a road (D) ahead of the vehicle, The lane change determination process includes a determination reference line selection process for selecting a determination reference line (Lc) that is a reference for lane change determination, The determination reference line selection process selects a virtual line that is offset by a predetermined amount from the own lane from among virtual lines that are line information that are candidates for boundaries of a lane (H) included in the recognition result. Driver assistance programs.

Images