Vehicle position estimation device and driving position estimation method

Abstract

The present invention provides a vehicle position estimation device and a driving position estimation method. The vehicle position estimation device comprises: a road end information acquisition unit (F31) for acquiring position information of a road end relative to the vehicle using at least one of a camera (11) and a distance measuring sensor (19A, 19B), wherein the camera captures a predetermined range around the vehicle, and the distance measuring sensor detects an object existing in a predetermined direction of the vehicle by sending a detection wave or a laser; a map acquisition unit (F2) for acquiring map information including the number of lanes of a road on which the vehicle is traveling from a map storage unit arranged inside or outside the vehicle; and a driving lane determination unit (F4) for determining the driving lane of the vehicle based on the position information of the road end acquired by the road end information acquisition unit and the number of lanes included in the map information acquired by the map acquisition unit.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is based on patent application No. 2020-107959 filed in Japan on June 23, 2020, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The present disclosure relates to a technique for determining the driving position of a host vehicle on a road. Background Art

[0004] Patent Document 1 discloses a method for determining the lane in which a vehicle is traveling (hereinafter referred to as the driving lane) from the left or right end of the road based on the type of lane boundary lines located to the left or right of the vehicle. Furthermore, Patent Document 2 discloses a mechanism for determining the driving lane using trajectory information of other vehicles traveling around the vehicle.

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 2013-242670

[0006] Patent Document 2: Japanese Patent Application Laid-Open No. 2019-91412

[0007] The technology disclosed in Patent Document 1 cannot identify the driving lane in situations where there are multiple lanes with the same left and right lane boundary lines, such as on multi-lane roads with four or more lanes. Furthermore, it is difficult to determine the driving lane in areas where the lane boundary lines are blurred, making it difficult to identify the lane boundary lines. For example, if a lane boundary line that was originally a solid line is blurred or contaminated and is recognized as a dotted line, the vehicle's driving lane may be misidentified.

[0008] The technology disclosed in Patent Document 2 can be expected to determine the driving lane even in environments that cannot be handled by the structure disclosed in Patent Document 1. However, the structure disclosed in Patent Document 2 has the problem that the driving lane cannot be determined if there are no other vehicles around. Summary of the Invention

[0009] An object of the present disclosure is to provide a vehicle position estimation device and a vehicle position estimation method that can determine a vehicle lane even when there are no surrounding vehicles.

[0010] A vehicle position estimation device for achieving this purpose, as an example, comprises: a road end information acquisition unit, which uses at least any one of a camera and a ranging sensor to acquire position information of the road end relative to the vehicle, wherein the camera captures a specified range around the vehicle, and the ranging sensor detects objects existing in a specified direction of the vehicle by sending detection waves or lasers; a map acquisition unit, which acquires map information including the number of lanes of the road on which the vehicle is traveling from a map storage unit arranged inside or outside the vehicle; and a driving lane determination unit, which determines the driving lane of the vehicle based on the position information of the road end acquired by the road end information acquisition unit and the number of lanes included in the map information acquired by the map acquisition unit.

[0011] In the above structure, the number of lanes shown on the map and the position information of the road end relative to the vehicle are obtained. In other words, the position information of the road end relative to the vehicle is equivalent to the distance information from the vehicle to the road end. Generally speaking, since the width of each lane is stipulated by laws, etc., as long as the distance from the road end is known, it is possible to estimate which lane the vehicle is located in. In addition, if the number of lanes on the road is known, the width of the curb, which is an area outside the lanes, can also be estimated. In other words, according to the above structure, the width of the curb can be taken into consideration to determine the driving lane of the vehicle. Moreover, in the above structure, when determining the driving lane, the type of lane boundary line and the trajectory information of surrounding vehicles are not required. In other words, even in a multi-lane road with four or more lanes, the driving lane can be determined even if there are no surrounding vehicles.

[0012] In addition, a driving position inference method for achieving the above-mentioned purpose is a driving position inference method for determining the lane in which the vehicle is traveling, which is executed by at least one processor and comprises: a position acquisition step, using at least any one of a camera and a distance measuring sensor to acquire position information of a road end relative to the vehicle, wherein the camera captures a specified range around the vehicle, and the distance measuring sensor detects an object existing in a specified direction of the vehicle by sending a detection wave or a laser; a map acquisition step, acquiring map information including the number of lanes of the road in which the vehicle is traveling from a map storage unit arranged inside or outside the vehicle; and a driving lane determination step, determining the driving lane of the vehicle based on the position information of the road end acquired in the position acquisition step and the number of lanes included in the map information acquired in the map acquisition step.

[0013] According to the above-mentioned method, the driving lane can be determined even if there are no surrounding vehicles, by the same function as the vehicle position estimation device.

[0014] In addition, the reference numerals in parentheses described in the claims indicate a correspondence relationship with specific elements described in an embodiment described later as one aspect, and do not limit the technical scope of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a block diagram showing the configuration of the driving support system 1 .

[0016] Figure 2 It is a block diagram showing the structure of the front camera 11.

[0017] Figure 3 It is a block diagram showing the structure of the position estimator 20.

[0018] Figure 4 This is a flowchart regarding the lane determination process performed by the position estimator 20 .

[0019] Figure 5 This is a diagram for explaining position correction of a detected feature using the yaw rate.

[0020] Figure 6 This is a diagram showing the in-image slope of the regression line when the regression line at the road edge can be accurately calculated.

[0021] Figure 7 This is a diagram showing the in-image slope of the regression line when the regression line at the left road end is erroneously calculated.

[0022] Figure 8 This is a flowchart regarding the road end slope determination process.

[0023] Figure 9 This is a flowchart regarding lane validity determination processing.

[0024] Figure 10 This is a diagram for explaining the operation of the lane validity determination process.

[0025] Figure 11 This is a diagram for explaining the operation of the lane validity determination process.

[0026] Figure 12 This is a flowchart of the process for calculating curb width.

[0027] Figure 13 This is a flowchart regarding the road side validity determination process.

[0028] Figure 14 This is a flowchart regarding the map compatibility determination process.

[0029] Figure 15 This is a flowchart regarding individual lane position determination processing.

[0030] Figure 16 This is a conceptual diagram of a system that generates a map based on curb width data calculated in a vehicle.

[0031] Figure 17 This is a diagram for explaining a road structure portion detected as a road end.

[0032] Figure 18 It is a diagram showing a modified example of the system configuration.

[0033] Figure 19 It is a diagram showing a modified example of the system configuration.

[0034] Figure 20 This is a diagram showing an example of equipment that can be used for detecting road edges. DETAILED DESCRIPTION

[0035] Hereinafter, embodiments of the present disclosure will be described using the drawings. Figure 1 1 is a diagram showing an example of a schematic configuration of a driving assistance system 1 to which the position estimator of the present disclosure is applied.

[0036] <Overall Structure Overview>

[0037] like Figure 1 As shown, the driving assistance system 1 includes a front camera 11, an inertial sensor 12, a GNSS receiver 13, a V2X onboard device 14, a map storage unit 15, an HMI system 16, a position inference unit 20, a driving assistance ECU 30, and a driving recorder 50. The term "ECU" in the component names stands for Electronic Control Unit, meaning an electronic control unit. Furthermore, "HMI" stands for Human Machine Interface. V2X, short for Vehicle to Everything, refers to the communication technology that connects vehicles to various objects.

[0038] The various devices or sensors constituting the driving assistance system 1 are connected as nodes to the in-vehicle network Nw, which is a communication network constructed in the vehicle. Nodes connected to the in-vehicle network Nw can communicate with each other. In addition, specific devices can also be configured to communicate directly with each other without going through the in-vehicle network Nw. For example, the position inference device 20 and the driving assistance ECU 30 can also be directly electrically connected via a dedicated line. In addition, Figure 1In the present invention, the in-vehicle network Nw is configured as a bus, but this is not limited to this. The network topology may also be a mesh, star, or ring structure. The network configuration can be modified as appropriate. Various standards, such as the Controller Area Network (hereinafter, CAN: registered trademark), Ethernet (Ethernet: registered trademark), and FlexRay (registered trademark), can be adopted as standards for the in-vehicle network Nw.

[0039] Hereinafter, a vehicle equipped with the driving assistance system 1 will also be referred to as the host vehicle, and the occupant sitting in the driver's seat of the host vehicle (in other words, the driver's seat occupant) will also be referred to as the user. Furthermore, the front-back, left-right, and up-down directions in the following description are defined with respect to the host vehicle. The front-back direction corresponds to the longitudinal direction of the host vehicle. The left-right direction corresponds to the width direction of the host vehicle. The up-down direction corresponds to the height direction of the vehicle. From another perspective, the up-down direction corresponds to a direction perpendicular to a plane parallel to both the front-back and left-right directions.

[0040] <Overview of Each Component>

[0041] The front camera 11 is a camera that captures the front of the vehicle at a predetermined angle of view. The front camera 11 is, for example, disposed at the upper end of the windshield on the inner side of the vehicle, the front grille, the roof, etc. Figure 2 As shown, the front camera 11 includes a camera body 40 that generates image frames and an ECU (hereinafter referred to as camera ECU 41) that detects a specified detection object by performing recognition processing on the image frames. The camera body 40 is a structure that includes at least an image sensor and a lens, and generates and outputs captured image data at a specified frame rate (for example, 60 fps). The camera ECU 41 can be implemented using an image processing chip including a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), etc. The camera ECU 41 includes a recognizer G1 as a functional module. The recognizer G1 is a structure that recognizes the type of an object based on a feature vector of an image generated by the camera body 40. The recognizer G1 can, for example, utilize a CNN (Convolutional Neural Network) or a DNN (Deep Neural Network) that applies deep learning.

[0042] The detection objects of the front camera 11 include, for example, pedestrians, other vehicles, and other moving objects. Other vehicles also include bicycles, bicycles with engines, and motorcycles. In addition, the front camera 11 is configured to detect specified ground objects. Ground objects that are detection objects of the front camera 11 include road edges, road surface markings, and structures located along the road. The so-called road surface markings refer to spray patterns drawn on the road surface for traffic control and traffic restrictions. For example, road markings include lane boundary lines (so-called lane markings) that indicate lane boundaries, crosswalks, stop lines, guide strips, safety zones, restriction arrows, etc. Lane boundary lines also include simple dividing signs, Botts' Dots, and other boundaries implemented by road spikes. The so-called structures located along the road include, for example, guardrails, curbs, trees, utility poles, road signs, traffic lights, etc. The camera ECU 41 separates and extracts the background and detection objects from the captured image based on image information including color, brightness, and contrast related to color and brightness.

[0043] The camera ECU 41 uses SfM (Structure from Motion) processing and other techniques to calculate the relative distance and direction (in other words, relative position) and movement speed of features such as lane boundaries and road edges from the vehicle. The relative position of features relative to the vehicle can also be determined based on the size and slope of the features within the image. Furthermore, the camera ECU 41 generates driving path data representing the track's shape, such as its curvature and width, based on the position and shape of lane boundaries and road edges.

[0044] The camera ECU 41 also calculates regression equations for the right and left road edges, using the set of points corresponding to the road edges in the image coordinate system as the parent cluster. The regression equations are functions that approximate the distribution of multiple detection points, such as a straight line or curve, and are calculated using, for example, the least squares method. For example, the image coordinate system uses the center of the pixel in the upper left corner of the image as its origin, with the right side of the image defined as the positive X-axis direction and the bottom of the image defined as the positive Y-axis direction. In the image coordinate system, pixel centers can be defined as integer values.

[0045] The regression equation for each road edge is represented, for example, by a quadratic function with the Y coordinate as a variable. The coefficient parameters of the regression equation can be adjusted sequentially based on the image recognition results for the road edge. The dimension of the regression equation can be modified as appropriate, and the regression equation can be a linear or cubic function. Furthermore, the camera ECU 41 can also calculate a regression equation for lane boundaries. Hereinafter, a straight line or curve in the image coordinate system represented by a regression equation will also be referred to as a regression line. The regression equation and regression line mentioned below primarily refer to the regression equation and regression line for the road edge.

[0046] The camera ECU 41 also calculates the yaw rate (rotational angular velocity) acting on the vehicle based on the SfM. The camera ECU 41 provides detection result data, including the relative position and type of the detected object, including the parameters of the regression equation for the road edge, to the position estimator 20, the driving assistance ECU 30, and the operation recording device 50 via the in-vehicle network Nw.

[0047] The front camera 11 can also be configured to provide image frames serving as observation data for object recognition to the driving assistance ECU 30 or the like via the in-vehicle network Nw. The observation data corresponds to the raw data observed by the sensor, or the data before the recognition process is performed. Object recognition processing based on the observation data can also be performed by an ECU outside the sensor, such as the driving assistance ECU 30. In addition, the calculation of the relative position of lane boundary lines, etc. can also be performed by the position inference device 20. Part of the functions of the camera ECU 41 (mainly the object recognition function) can also be provided in the position inference device 20 and the driving assistance ECU 30. In this case, the front camera 11 can provide the image data serving as observation data to the position inference device 20 and the driving assistance ECU 30.

[0048] The inertial sensor 12 is a sensor that detects a predetermined physical state quantity, such as acceleration. The inertial sensor 12 includes, for example, a 3-axis gyroscope sensor and a 3-axis acceleration sensor. The driving assistance system 1 may also include a magnetic sensor as the inertial sensor 12. Furthermore, the driving assistance system 1 may also include an atmospheric pressure sensor or a temperature sensor as the inertial sensor 12. The atmospheric pressure sensor and the temperature sensor may also be used to correct the output values ​​of other sensors. Various inertial sensors 12 may also be packaged as an inertial measurement unit (IMU).

[0049] The inertial sensors 12 output data representing the current value of the physical state quantity being detected (in other words, the detection result) to the in-vehicle network Nw. The output data of each inertial sensor 12 is acquired by the position estimator 20 and the like via the in-vehicle network Nw. The types of sensors used as inertial sensors 12 in the driving assistance system 1 can be appropriately designed, and it is not necessary to include all of the sensors described above.

[0050] The GNSS receiver 13 detects its current position sequentially (e.g., in 100-millisecond increments) by receiving navigation signals transmitted from positioning satellites constituting the GNSS (Global Navigation Satellite System). Examples of GNSS systems include GPS (Global Positioning System), GLONASS, Galileo, IRNSS, QZSS, and Beidou.

[0051] The V2X vehicle-mounted device 14 is used to conduct wireless communications between the vehicle and other devices. The "V" in V2X refers to the vehicle itself, while the "X" can refer to various entities outside the vehicle, such as pedestrians, other vehicles, road equipment, networks, and servers. The V2X vehicle-mounted device 14 includes a wide-area communication unit and a narrow-area communication unit as communication modules. The wide-area communication unit is a communication module for conducting wireless communications based on a prescribed wide-area wireless communication standard. Examples of such wide-area wireless communication standards include LTE (Long Term Evolution), 4G, and 5G. In addition to communicating via wireless base stations, the wide-area communication unit can also be configured to directly communicate with other devices using a method based on the wide-area wireless communication standard, in other words, without going through a base station. In other words, the wide-area communication unit can also be configured to implement cellular V2X. Equipped with the V2X vehicle-mounted device 14, the vehicle becomes a connected car capable of connecting to the internet. For example, the position estimator 20 can update the map data stored in the map storage unit 15 by downloading the latest high-precision map data from a predetermined server in cooperation with the V2X vehicle-mounted device 14 .

[0052] The narrow-area communication unit of the V2X vehicle-mounted device 14 is a communication module for directly conducting wireless communication with other mobile objects and roadside equipment in the vicinity of the vehicle in accordance with a narrow-area communication standard. The narrow-area communication standard is a communication standard that limits the communication distance to within several hundred meters. Other mobile objects are not limited to vehicles but can also include pedestrians, bicycles, etc. As a narrow-area communication standard, any standard such as the WAVE (Wireless Access in Vehicular Environment) standard disclosed in IEEE1609 or the DSRC (Dedicated Short Range Communications) standard can be adopted.

[0053] The map storage unit 15 is a non-volatile memory that stores high-precision map data. The high-precision map data here is equivalent to map data that represents the position coordinates of the road structure and the features arranged along the road with an accuracy that can be used for autonomous driving. The so-called accuracy that can be used for autonomous driving is, for example, equivalent to suppressing the error between the actual position involved in each map element and the position registered on the map to a level below 10cm to 20cm. The high-precision map data, for example, has three-dimensional shape data of the road, lane data, feature data, etc. The above-mentioned three-dimensional shape data of the road includes node data related to the places (hereinafter referred to as nodes) where multiple roads intersect, merge, and branch, and link data related to the roads (hereinafter referred to as links) connecting the places.

[0054] Road segment data includes information such as road end points and road width. The road end point information represents the coordinates of the road end points. Road segment data may also include data indicating the road type, such as whether it is a dedicated road for vehicles or a general road. Dedicated roads here refer to roads where pedestrians and bicycles are prohibited, such as toll roads like expressways. Road segment data may also include attribute information indicating whether autonomous driving is permitted on the road.

[0055] The lane data indicates the number of lanes, the location information of the lane boundary lines of each lane, the direction of travel of each lane, and the branch / merge locations at the lane level. For example, the lane data may also include information indicating whether the lane boundary lines are realized by a pattern of solid lines, dashed lines, or Bosch points. The location information of lane boundary lines and road ends (hereinafter referred to as lane boundary lines, etc.) is represented as a coordinate group (in other words, a point group) of the locations where lane boundary lines are formed. In addition, as another method, the location information of lane boundary lines, etc. can also be represented as a polynomial. The location information of lane boundary lines, etc. can also be a collection of line segments represented as polynomials (in other words, a line group).

[0056] Feature data includes the location and type of road markings such as temporary stop signs, as well as the location, shape, and type of landmarks. Landmarks include three-dimensional structures along roads, such as traffic signs, traffic lights, poles, and commercial billboards. Furthermore, the map storage unit 15 may be configured to temporarily store high-precision map data within a specified distance from the vehicle. Furthermore, the map data stored in the map storage unit 15 may be navigation map data, also known as navigation map data. Navigation map data is map data with a relatively lower accuracy than high-precision map data.

[0057] The HMI system 16 provides an input interface function for accepting user operations and an output interface function for presenting information to the user. The HMI system 16 includes a display 161 and an HCU (HMI Control Unit) 162. In addition to the display 161, other means for presenting information to the user may include a speaker, a vibrator, or a lighting device (e.g., an LED).

[0058] Display 161 is a device that displays images. For example, display 161 is a so-called center display, located at the uppermost portion of the center section of the instrument panel in the vehicle's width direction. Display 161 is capable of full-color display and can be implemented using a liquid crystal display, an OLED (Organic Light Emitting Diode) display, a plasma display, or the like. Alternatively, the HMI system 16 may include a head-up display (HUD) that projects a virtual image onto a portion of the windshield in front of the driver's seat. Alternatively, display 161 may be a so-called instrument display, located in the area of ​​the instrument panel directly in front of the driver's seat.

[0059] The HCU 162 is a component that comprehensively controls information presented to the user. The HCU 162 is implemented using, for example, a processor such as a CPU or GPU, RAM, and flash memory. The HCU 162 controls the display screen of the display 161 based on information provided by the driving assistance ECU 30 and signals from an input device (not shown). For example, the HCU 162 displays a route guidance image on the display 161 in response to a request from the navigation device. The route guidance image includes, for example, so-called turn-by-turn images that show the direction of travel and recommended lanes at intersections, junctions, and lane addition locations.

[0060] The position estimator 20 is a device that determines the current position of the vehicle. The position estimator 20 is equivalent to the vehicle position estimation device. Details of the functions of the position estimator 20 will be described later. The position estimator 20 is primarily composed of a computer comprising a processing unit 21, RAM 22, memory 23, a communication interface 24, and a bus connecting these components. The processing unit 21 is hardware used for computational processing in conjunction with the RAM 22. The processing unit 21 is equivalent to a processor. The processing unit 21 includes at least one computing core, such as a CPU. The processing unit 21 executes various processes by accessing the RAM 22. The memory 23 includes a non-volatile storage medium, such as flash memory. The memory 23 stores a position estimation program, which is a program executed by the processing unit 21. Execution of the position estimation program by the processing unit 21 is equivalent to executing the method corresponding to the position estimation program (in other words, the driving position estimation method). The communication interface 24 is a circuit used to communicate with other devices via the in-vehicle network Nw. The communication interface 24 can be implemented using analog circuit elements, integrated circuits, etc. The position estimator 20 may be configured to sequentially output the lane ID of the vehicle's driving lane (in other words, the driving lane number), the calculated result of the curb width, the type of features used to determine the vehicle's position, etc. to the in-vehicle network Nw.

[0061] The driving assistance ECU 30 controls the driving actuator 18 based on the detection results of the front camera 11 and the inference results of the position estimator 20, thereby performing some or all driving operations on behalf of the driver. The driving assistance ECU 30 can also function as an automatic driving device that enables the vehicle to drive autonomously. The driving actuator 18 refers to actuators used for driving. The driving actuator 18 includes mechanical elements for accelerating, decelerating, and steering the vehicle. The driving actuator 18 includes, for example, a brake device, an electronic throttle valve, and a steering actuator. The so-called braking device is, for example, a brake actuator.

[0062] The driving assistance ECU 30 includes a lane tracking control (LTC) unit H1, which provides a lane tracking control (LTC) function as one of the vehicle control functions. LTC is a function that causes the vehicle to travel along and within its lane. LTC unit H1 generates a predetermined travel line along the vehicle's lane and controls the steering angle via a steering actuator to maintain this predetermined travel line. For example, LTC unit H1 generates a steering force toward the center of the vehicle's lane to cause the vehicle to travel along the center of the lane.

[0063] Similar to the position estimator 20, the driving assistance ECU 30 is primarily composed of a computer equipped with a processing unit, RAM, memory, a communication interface, and a bus connecting these components. Illustration of these components is omitted. The memory of the driving assistance ECU 30 stores a driving assistance program, which is a program executed by the processing unit. The execution of the driving assistance program by the processing unit is equivalent to executing the method corresponding to the driving assistance program.

[0064] The operation recording device 50 records data representing the conditions inside the vehicle and outside the vehicle cabin while the vehicle is traveling. The conditions inside the vehicle while the vehicle is traveling can include the operating status of the position estimator 20, the operating status of the driving assistance ECU 30, and the status of the driver's seat occupants. Data representing the operating status of the position estimator 20 includes information such as the road edge recognition status, the regression equation representing the road edge, the calculated value of the curb width, and the types of features used to determine the vehicle's position. Data representing the operating status of the driving assistance ECU 30 also includes the results of the driving assistance ECU 30's recognition of the surrounding environment, the driving plan, and calculations such as the target control variables of each driving actuator. Data to be recorded is acquired from ECUs and sensors installed in the vehicle, such as the position estimator 20, the driving assistance ECU 30, and surrounding monitoring sensors including the front camera, via the in-vehicle network Nw. For example, when a specified recording event occurs, the operation recording device 50 stores data on the items specified for recording in a non-volatile storage medium. Data stored by the operation recording device 50 can also be stored on an external server.

[0065] <Functions of Position Estimator 20>

[0066] Here, use Figure 3 The function and operation of the position estimation device 20 will be described. The position estimation device 20 provides the position estimation program stored in the memory 23. Figure 3 The various functional blocks shown correspond to the functions. Specifically, the position estimator 20 includes a provisional position estimating unit F1, a map acquiring unit F2, a course information acquiring unit F3, a lane identifying unit F4, and a detailed position calculating unit F5 as functional blocks.

[0067] The provisional position estimation unit F1 sequentially determines the vehicle's position by combining the positioning results from the GNSS receiver 13 with the detection results from the inertial sensors. For example, in tunnels, where GNSS positioning is unavailable, the provisional position estimation unit F1 uses the yaw rate and vehicle speed to perform dead reckoning (autonomous navigation). The yaw rate used for dead reckoning can be the yaw rate detected by the camera ECU 41 using SfM technology or the yaw rate detected by a yaw rate sensor.

[0068] The map acquisition unit F2 reads out map data (map information) of a specified range based on the current position from the map storage unit 15. The current position used for map reference can be the position determined by the provisional position inference unit F1 or the position determined by the detailed position calculation unit F5. For example, when the detailed position calculation unit F5 is able to calculate the current position, the position information is used to acquire map data. On the other hand, when the detailed position calculation unit F5 is unable to calculate the current position, the position coordinates calculated by the provisional position inference unit F1 are used to acquire map data. On the other hand, just after the ignition power is turned on, for example, the map reference range is determined based on the previous position calculation result stored in the memory. This is because the previous position calculation result stored in the memory is equivalent to the end point of the previous trip, that is, the parking position. In addition, the map acquisition unit F2 can also be configured to sequentially download high-precision map data of the surrounding area of ​​the vehicle from an external server, etc. via the V2X vehicle-mounted device 14. The map storage unit 15 can also be set outside the vehicle.

[0069] The lane information acquisition unit F3 acquires travel path data from the camera ECU 41 included in the front camera 11. Specifically, it acquires the relative positions of lane boundaries and road edges (hereinafter referred to as lane boundaries, etc.) recognized by the front camera 11, as well as regression line parameters for the road edges. The structure that acquires the relative position information of the road edges corresponds to the road edge information acquisition unit F31. The structure that acquires the relative position information of the lane boundaries, etc., corresponds to the boundary line information acquisition unit F32.

[0070] The position of a road edge, for example, is represented by an XY coordinate system, or vehicle coordinate system, with the vehicle's reference point as the origin. The X-axis constituting this vehicle coordinate system is set parallel to the left-right direction of the vehicle, with the right direction being the positive direction, for example. The Y-axis is set parallel to the front-back direction of the vehicle, with the direction toward the front of the vehicle being the positive direction.

[0071] Furthermore, the coordinate system representing the position of a road edge, etc., can employ a variety of coordinate systems. For example, if the image recognition software of the camera ECU 41 uses the world coordinate system (WCS) or program coordinate system (PCS) used in CAD, etc. to represent the position of the detected object, the relative position of the road edge, etc., can also be represented using the WCS or PCS. The vehicle coordinate system can also be configured with the front of the vehicle as the positive X-axis direction and the left side of the vehicle as the positive Y-axis direction. The runway information acquisition unit F3 can also acquire data representing the position of the road edge, etc., using an image coordinate system.

[0072] In addition, the runway information acquisition unit F3 can also convert the relative position coordinates of the lane boundary line, etc. obtained from the camera ECU 41 into position coordinates in the global coordinate system (hereinafter also recorded as observation coordinates). The observation coordinates of the lane boundary line, etc. can be calculated by combining the current position coordinates of the vehicle and the relative position information of the lane boundary line, etc. relative to the vehicle. When the detailed position calculation unit F5 is able to calculate the current position, the current position coordinates of the vehicle used for calculating the observation coordinates of the lane boundary line, etc. can use this position information. On the other hand, when the detailed position calculation unit F5 is unable to calculate the current position, the position coordinates calculated by the provisional position estimation unit F1 can be used. In addition, the calculation of the observation coordinates of the lane boundary line, etc. using the current position coordinates of the vehicle can also be implemented by the camera ECU 41. Hereinafter, the lane boundary line detected by the front camera 11 will also be recorded as a detected boundary line. In addition, the road end detected by the front camera 11 will also be recorded as a detected road end.

[0073] The lane determination unit F4 is configured to determine the lane in which the vehicle is traveling, or the driving lane, based on the relative position information of the road ends and lane boundaries acquired by the track information acquisition unit F3. The lane determination unit F4 includes a curb width calculation unit F41, which calculates the width of the curb provided on the road in which the vehicle is traveling, or the driving path. Details of the lane determination unit F4 and the curb width calculation unit F41 will be described later. Furthermore, the lane determination unit F4 may also be configured to determine the driving lane and driving position using relative position information of landmarks such as direction signs. The lane determination unit F4 is equivalent to the driving lane determination unit.

[0074] The detailed position calculation unit F5 determines the detailed position of the vehicle within the travel lane based on the determination results of the lane determination unit F4 and the data acquired by the lane information acquisition unit F3. Specifically, it calculates the left-right offset from the center of the travel lane based on the distance from the left and right lane boundaries. Furthermore, the lateral position of the vehicle within the travel path is determined by combining the lane information determined by the lane determination unit F4 with the offset from the lane center. The offset from the lane center determined by the detailed position calculation unit F5 is used, for example, by the LTC unit H1.

[0075] Alternatively, the detailed position calculation unit F5 may determine the detailed position of the vehicle on the map based on the determination result of the lane determination unit F4 and the landmark information detected by the front camera 11. For example, when the distance from the left road end to the center of the vehicle is determined to be 1.75 m as a result of image analysis, the vehicle is determined to be located at a position offset by 1.75 m to the right from the coordinates of the left road end shown on the map. Furthermore, for example, when the distance to a direction sign existing in front of the vehicle is determined to be 100 m as a result of image analysis, the vehicle is determined to be located at a position offset by 100 m to the near front side from the position coordinates of the direction sign registered in the map data. The near front side here refers to the direction opposite to the direction of travel of the vehicle. When driving forward, the so-called near front side corresponds to the rear of the vehicle.

[0076] Furthermore, the correspondence between the landmark detected by the front camera 11 and the landmark registered on the map can be implemented, for example, by comparing the observed coordinates of the landmark with the coordinate information registered on the map. For example, the landmark registered on the map that is closest to the observed coordinates of the landmark is inferred to be the same landmark. When comparing landmarks, it is preferred to use feature quantities such as shape, size, and color, and adopt the landmark with the highest degree of feature consistency. Once the correspondence between the observed landmark and the landmark on the map is completed, the detailed position calculation unit F5 sets the longitudinal position of the vehicle on the map to a position that is longitudinally offset from the position of the landmark on the map by the distance between the observed landmark and the vehicle.

[0077] As described above, according to the structure of the detailed position calculation unit F5 that calculates not only the lateral position of the vehicle but also the longitudinal position of the vehicle, the remaining distance to feature points on the road such as intersections, curve entrances / exits, the end of traffic jams, and road branch points (in other words, POIs) can be calculated with high precision. In addition, by inferring the lateral position of the vehicle, it is possible to determine whether a lane change should be made in order to turn left or right, and when a lane change must be made if necessary. In addition, the process of determining the current position of the vehicle on the map using the detection position information of landmarks and road ends as described above is also called positioning processing. The position of the vehicle as a result of the positioning processing can be represented by the same coordinate system as the map data (for example, latitude, longitude, altitude). The vehicle position information can be represented by an arbitrary absolute coordinate system such as WGS84 (World Geodetic System 1984), for example.

[0078] Lane determination processing

[0079] Next, use Figure 4 The flowchart shown explains the lane determination process executed by the position estimator 20 (mainly the lane determination unit F4 ). Figure 4The flowchart shown is executed at a predetermined cycle (for example, every 100 milliseconds) while the vehicle's driving power supply is on. The driving power supply is, for example, an ignition power supply in an engine vehicle. In an electric vehicle, the system main relay is equivalent to the driving power supply. In this embodiment, as an example, the lane determination process includes steps S0 to S10. Steps S1 to S10 are performed by the lane determination unit F4. This series of processing flows is repeatedly executed, and when the processing is successful, the vehicle position information calculated in this process is saved to the memory. In addition, step S0 is a process in which the map acquisition unit F2 acquires the map data of the driving route. The process in which the map acquisition unit F2 acquires the map data of the driving route can also be executed independently of this process as a preliminary preparation process. Step S0 is equivalent to the map acquisition step.

[0080] In step S1, the lateral position (X coordinate) of various ground features such as the road end and lane boundary line at the judgment point located at a specified distance in front of the vehicle is calculated. Step S1 is equivalent to a position acquisition step. The judgment point can be, for example, a location 10.0m in front of the vehicle. In addition, the judgment point can also be a location 5.0m in front or a location 8.5m in front. The judgment point can be a specified location included in the shooting range of the front camera 11. The judgment point is a linear concept that includes a point located at a specified distance in front of the vehicle and locations to the left and right of the location. The judgment point can be replaced by a judgment line. The lateral position of the road end at the judgment point is calculated based on, for example, the regression line parameters of the road end.

[0081] The positions of various objects on the ground are expressed, for example, in a vehicle coordinate system based on the vehicle itself. In addition, the lane boundary line and the lateral position of the road end at the determination point (here, the X coordinate) use the vehicle speed and yaw rate of the vehicle itself, as shown in the following example. Figure 5 As shown, corrections are made for the lateral position along the curve. Specifically, the time t required for the vehicle to actually reach the decision point is calculated using the vehicle's speed and yaw rate. Furthermore, the lateral displacement ΔX of the vehicle at the time of reaching the decision point (in other words, in the X-axis direction) is calculated based on the time required to reach the decision point, the yaw rate, and the vehicle speed.

[0082] Specifically, if the time required to reach the decision point is t, the distance to the decision point is D, the yaw rate is ω, and the vehicle speed is v, then the relationship D = (v / ω)sin(ωt) holds. The required time t is determined by solving this equation. Furthermore, the lateral displacement ΔX can be calculated using the equation ΔX = (v / ω){1-cos(ωt)}.

[0083] The lane identification unit F4 corrects the lateral position coordinates to a curve shape by subtracting ΔX from the lateral position of the road end currently detected by the front camera 11 . Figure 5P1 and P2 represent the pre-correction position coordinates of the lane boundary lines on the left and right sides of the vehicle at the determination point, while P1a and P2a represent the post-correction positions. By using the yaw rate to correct the positions of the lane boundary lines and road edges at the determination point, the influence of road curvature can be suppressed. After step S1 is completed, step S2 is executed.

[0084] In step S2, the road end slope determination process is executed. The road end slope determination process is a process for not using information about road ends with a possibility of misidentification in subsequent processing based on the slope of the regression line of the road end. The road end slope determination process is described separately. If step S2 is completed, road end information with a certain degree of appropriateness is obtained after removing road ends with a high possibility of misidentification. If step S2 is completed, step S3 is executed. Step S2 is not an essential element and can be omitted as appropriate. However, by executing step S2, the possibility of miscalculating the width of the curb or mistaking the driving lane can be reduced.

[0085] In step S3, a lane appropriateness determination process is performed. The lane appropriateness determination process is a process for checking whether the detected lane boundary line has a certain degree of appropriateness as a lane boundary line constituting the lane. In other words, it is equivalent to a process for removing information about lane boundary lines that have the possibility of being misdetected. As a viewpoint for checking the appropriateness of the detection result of the lane boundary line, for example, whether the lateral position of the detected lane boundary line is closer to the outside than the road end position, or whether the interval between the lane boundary lines is too small to be used as the lane width, etc. The details of the lane appropriateness determination process will be described later. By executing step S3, boundary line information with a certain degree of appropriateness can be obtained after removing lane boundary lines with a high possibility of being misdetected. If step S3 is completed, step S4 is executed.

[0086] In step S4, the curb width calculation unit F41 performs a curb width calculation process. This process is a structure for calculating the width of the curb at the determination point. The so-called curb here refers to the area on the road other than the lane. The curb here also includes the shoulder. In addition, the curb can include an area sandwiched between the lane outer line and the road end. The lane outer line refers to a dividing line drawn near the end of the lane. In addition, the curb can include a zebra crossing (in other words, a guide strip) adjacent to the road end. The zebra crossing refers to an area of ​​the road surface with stripes. The details of the curb width calculation process as step S4 will be described later. By executing step S4, the calculated value of the curb width is obtained. If step S4 is completed, step S5 is executed.

[0087] In step S5, the road end validity determination process is executed. The road end validity determination process is a process for verifying whether the lateral position coordinates of the road end calculated in step S1 are valid. For example, when the lateral position of the road end calculated in step S1 enters the area where the existence of the lane is confirmed by the front camera 11, that is, the lane detection range, it is determined that the lateral position of the road end is wrong, and the erroneously calculated lateral position information of the road end is discarded. The details of the road end validity determination process will be described separately later. In addition, the determination content of step S5 can be included in the lane validity determination process of step S3. Step S5 is not a necessary element. If step S5 is completed, step S6 is executed.

[0088] In step S6, a map matching determination process is executed. The map matching determination process is a process for determining whether the road information obtained in the above process matches the road information registered on the map. The details of the map matching determination process will be described separately later. By executing this map matching determination process, it is possible to verify, for example, whether the width of the curb calculated in step S4 is an appropriate value or an error. Step S6 is not a required structure and can be omitted. However, by executing step S6, the possibility of misidentifying the driving lane can be reduced. After step S6 is completed, step S7 is executed.

[0089] In step S7, the width of the driving lane is calculated. For example, the distance from the nearest lane boundary line on the right side of the vehicle to the nearest lane boundary line on the left side of the vehicle is calculated as the driving lane width. If more than three lane boundary lines can be detected, the lane width can also be calculated as the average of the distances between the lane boundary lines. In step S7, the width of other lanes, that is, lanes in which the vehicle is not traveling, can also be calculated. The width of the lane in which the vehicle is traveling can also be regarded as the width of other lanes. By completing step S7, the lane width in the driving path is determined. If step S7 is completed, step S8 is executed.

[0090] In step S8, individual lane position determination processing is performed. This process calculates the range of each lane. It also includes discarding lanes that are located in incorrect positions. Details of step S8 will be described later. Once step S8 is complete, step S9 is executed. Through the above processing, the position of each lane is determined based on the width of the curb. Specifically, the width of the curb is taken into account to determine how far from the road edge the first and second lanes are located.

[0091] In step S9, the driving lane of the vehicle is determined based on the distance from the road end to the vehicle and the calculated curb width. Since the position of each lane is represented by a vehicle coordinate system based on the vehicle, the lane containing the location of X=0.0 is equivalent to the driving lane. For example, when the left end X coordinate of the second lane is -2.0 and the right end X coordinate is +1.0, it is determined that the vehicle is in the second lane. The driving lane can be represented by a lane ID, where the lane ID indicates the number of the lane from the left road end. Since the position of each lane is determined by taking into account the curb width, the above processing is equivalent to determining the structure of the driving lane based on the lane adjacent to the curb (for example, as the first lane). In addition, the lane ID can also be assigned based on the right road end. Step S9 is equivalent to the driving lane determination step. If step S9 is completed, step S10 is executed.

[0092] In step S10, the determined lane and curb width information is output externally. Examples of output destinations include the driving assistance ECU 30, the operation recorder 50, the navigation system, and a map server installed externally to the vehicle. The driving assistance ECU 30, the operation recorder 50, the navigation system, and the map server are considered external devices.

[0093] <Step S2: Road End Slope Determination Process>

[0094] The regression line at the road end is usually as follows Figure 6 As shown in , the left side of the road in the image has a negative slope, and the right side of the road has a positive slope. Figure 7 As shown in the example, when another vehicle is captured near the edge of the image or near the road edge, the edge of the other vehicle is mistakenly recognized as part of the road edge, and the regression line of the road edge deviates significantly from the actual road edge. Of course, if the regression line of the road edge deviates significantly, the deviation between the recognized position of the road edge at the judgment point and the actual position increases, which may cause the incorrect judgment of the driving lane. In addition, Figure 6 The dashed line shown conceptually represents a lane boundary line, the dashed-dotted line represents a regression line at the left road end, and the double-dashed-dotted line represents a regression line at the right road end. Figure 7 The components of Figure 6 Likewise, the relatively thin one-dot chain line represents the true left road end, and the relatively thick one-dot chain line represents the regression line of the left road end calculated by the camera ECU 41 .

[0095] This process is introduced with an eye on the above-mentioned issues and characteristics, and is configured to discard the recognition results of road ends whose slopes at the decision points are out of the normal range. Figure 8 The road end slope determination process will be described. Figure 8The flow of the road end slope determination process is shown as Figure 4 Step S2 is performed.

[0096] As an example, the road edge slope determination process of this embodiment includes steps S201 to S205. The positive or negative slope may vary depending on how the image coordinate system is defined. Consequently, the normal slope range for each road edge also changes depending on how the image coordinate system is defined. Here, as an example, the upper left corner of the image frame is used as the origin, the right side is used as the positive X-axis direction, and the bottom side is used as the positive Y-axis direction.

[0097] In step S201, the regression line parameters for the left and right road edges are obtained, and the process proceeds to step S202. In step S202, the slopes of the left and right road edges at the decision point within the image are calculated. The slopes of the road edges are calculated, for example, by substituting the Y coordinate value at the decision point into the function obtained by first-order differentiation of the regression line. Once the slopes within the image at the decision point, in other words, the slopes in the image coordinate system, are calculated for the left and right road edges, step S203 is executed.

[0098] In step S203, a determination is made as to whether the slope of the road end at the determination point in the image coordinate system falls within a predetermined normal range. This process is performed separately for the left and right road ends. Specifically, a determination is made as to whether the slope of the left road end at the determination point in the image coordinate system falls within a predetermined normal range for the left road end. Separately, a determination is made as to whether the in-image slope of the right road end falls within a predetermined normal range for the right road end. The normal range for the left road end can be set to a value greater than zero, for example. Alternatively, the normal range for the right road end can be set to a value less than zero. Furthermore, the normal range for the slope of each road end within the image can be dynamically adjusted, taking into account the curvature of the road registered in the map data.

[0099] If the slope of a road edge within the image (in other words, the slope of the regression line) falls within the normal range, the recognized position of that road edge is used as the actual road edge position. On the other hand, if the slope of a road edge within the image falls outside the normal range, the recognized position data for that road edge is discarded. In other words, information about road edges where the slope of the regression line at the decision point falls outside the normal range is not used in the following processing. This reduces the risk of misidentification of driving lanes due to misidentification of road edges.

[0100] <Step S3: Lane Validity Determination Process>

[0101] Here, use Figure 9 The lane validity determination process is described below. Figure 4As an example, the lane appropriateness determination process includes steps S301 to S311. In step S301, the number of lane boundary lines detected by the front camera 11, that is, the number of detected boundary lines, is obtained, and the process moves to step S302. In step S302, the number of lane candidates Cn is calculated based on the number of detected boundary lines obtained in step S301. The number of lane candidates Cn is equivalent to the area sandwiched by mutually adjacent detection boundary lines, that is, the number of lane candidates. The number of lane candidates Cn is the value obtained by subtracting 1 from the number of detected boundary lines. For example, Figure 10 As shown, when six lane boundary lines are detected, the number of lane candidates Cn is five.

[0102] also, Figure 10 "B1" to "B6" shown in the figure represent the detected lane boundary lines. "EgL" represents the left road edge, and "EgR" represents the right road edge. "Hv" represents the vehicle. Hereinafter, the i-th lane candidate from the left will be referred to as the i-th lane candidate. Figure 10 In the example shown, the vehicle is located in the third lane candidate. Here, as an example, lane candidates and lane boundary lines are numbered starting from the left. Of course, as another example, numbers may be assigned starting from the right to identify multiple lane candidates and lane boundary lines.

[0103] Once the processing in step S302 is complete, step S303 is executed. In step S303, the variable k used in the process is initialized (specifically, set to 1), and the process proceeds to step S304. In step S304, the kth lane candidate is set as the target of the following processing, and the process proceeds to step S305.

[0104] In step S305, a determination is made as to whether the width WLC of the k-th lane candidate is less than the minimum lane width LWmin. If the width WLC of the k-th lane candidate is less than the minimum lane width LWmin, the process proceeds to step S306. On the other hand, if the width WLC of the k-th lane candidate is greater than or equal to the minimum lane width LWmin, the process proceeds to step S307. A k-th lane candidate width WLC less than the minimum lane width LWmin indicates that the k-th lane candidate is not actually a lane but a curb and / or that the lane boundary forming the outer edge of the k-th lane candidate has been erroneously detected. Furthermore, a k-th lane candidate width WLC greater than or equal to the minimum lane width LWmin indicates that there is a possibility that the k-th lane candidate is a valid lane.

[0105] The minimum lane width LWmin is the minimum value of the range of lane widths that can be used. The minimum lane width LWmin is set based on the laws and regulations of the area where the vehicle is used. For example, the minimum lane width LWmin is set to 2.5m. Of course, the minimum lane width LWmin can also be 2.2m, 2.75m, 3.0m, 3.5m, etc. In order to prevent misjudgments, the minimum lane width LWmin used in this process is preferably set to a value that is smaller than the minimum value of the actual lane width specified by laws and regulations by a specified amount (for example, 0.25m). The set value of the minimum lane width LWmin can also be changed according to the type of road on which the vehicle is traveling. For example, the minimum lane width LWmin when the vehicle is traveling on a highway can be set to be larger than the minimum lane width LWmin when the vehicle is traveling on a general road. According to this structure, since the set value corresponding to the road type is applied as the minimum lane width LWmin, it is possible to prevent misjudgments of the validity of the lane boundary detection results.

[0106] In step S306, the lane boundary line that is located outside the k-th lane candidate as viewed from the vehicle is considered an erroneously detected lane boundary line. Furthermore, the inner lane boundary line of the k-th lane candidate is set as the end of the lane detection range. This is equivalent to determining that the lane detection range extends to the inner lane boundary line of the k-th lane candidate.

[0107] For example, when the width WLC of the fifth lane candidate is less than the minimum lane width LWmin, the lane boundary line B6 that exists on the outside as viewed from the vehicle among the lane boundary lines B5 and B6 that constitute the fifth lane candidate is discarded, and the lane boundary line B5 is set as the end of the lane detection range. By executing steps S305 to S306, the concern about using lane boundary line information with a high probability of false detection in the following processing can be reduced. The right end of the lane detection range corresponds to the rightmost outermost detection line of the lane boundary lines detected by the front camera 11, which is located on the rightmost side and is a valid lane boundary line. The so-called valid lane boundary line refers to a lane boundary line that is judged to have a low probability of false detection based on the distance from the adjacent lane boundary line and the positional relationship with the road end, and is also used in the following processing. In addition, the left end of the lane detection range corresponds to the leftmost outermost detection line of the lane boundary lines detected by the front camera 11, which is located on the leftmost side and is a valid lane boundary line.

[0108] In step S307, it is determined whether the left lane boundary line, which is the lane boundary line on the left side of the lane boundary lines constituting the k-th lane candidate, is located outside the left road end (here, on the left side) as viewed from the vehicle. If the left lane boundary line is located outside the left road end, step S308 is executed. On the other hand, if the left lane boundary line is not located outside the left road end, step S309 is executed.

[0109] In step S308, the lane boundary line located on the right side of the lane boundary lines constituting the k-th lane candidate, that is, the right boundary line, is set as the left end of the lane detection range. Figure 10 As shown, if lane boundary line B1, which corresponds to the left boundary line of the first lane candidate, is located outside the left road end EgL, the detection result of lane boundary line B1 is discarded, and lane boundary line B2 is set as the left end of the lane detection range. This process is equivalent to treating the detection of the left boundary line of the k-th lane candidate as an error.

[0110] In step S309, a determination is made as to whether the right boundary line of the k-th lane candidate is located outside (here, on the right side) of the right road end as viewed from the vehicle. If the right boundary line is located outside the right road end, step S310 is executed. On the other hand, if the right boundary line is not located outside the right road end, step S311 is executed. In step S310, the left boundary line of the k-th lane candidate is set as the right end of the lane detection range. This process is equivalent to treating the detection of the right boundary line of the k-th lane candidate as an error.

[0111] By executing steps S307-S310, lane boundary data outside the road edge is not used in the following processing. Lane boundary data detected outside the road edge is more likely to be misdetected. In other words, by executing steps S307-S310, the concern that lane boundary data with a high probability of misdetection may be used in the following processing (in other words, lane determination) is reduced.

[0112] In step S311, the value of variable k is incremented by 1, in other words, it is incremented by 1, and the process proceeds to step S312. In step S312, it is determined whether variable k is less than the number of lane candidates Cn. If the relationship k≤Cn is satisfied, step S304 is executed. On the other hand, if the relationship k>Cn is satisfied, this process ends and the process proceeds to step S304. Figure 4Step S4 is shown. Step S312 corresponds to the process of determining whether any lane candidates remain unprocessed. In other words, it corresponds to the process of determining whether the processes of steps S304 to S310 have been executed for all lane candidates. Furthermore, while this embodiment discloses a method of incrementing the variable k from 1 to the number of lane candidates Cn, this is not limiting. The variable k may also be decremented from the number of lane candidates Cn to 1.

[0113] By executing the above processing, for example, Figure 10 The test results shown are corrected to Figure 11 That is, information on the lane boundary line B1 located to the left of the left road end and the lane boundary line B6 whose distance to the adjacent lane boundary line is less than the minimum lane width LWmin is discarded / not used.

[0114] <Step S4: Curb Width Calculation Process>

[0115] Here, use Figure 12 The curb width calculation process will be described. Figure 12 The curb width calculation process shown is, for example, as Figure 4 Here, as an example, the curb width calculation process includes steps S401 to S404. Steps S401 to S404 are independently executed for both the right and left ends of the lane detection range. For convenience, the steps are described using the right end of the lane detection range as an example, but the same process is also performed for the left end of the lane detection range.

[0116] In step S401, a determination is made as to whether the distance between the right end of the lane detection range and the right road end is less than the minimum lane width LWmin. This process is equivalent to determining whether the road end is located outside the lane detection range by more than the minimum lane width. If the distance between the right end of the lane detection range and the right road end is less than the minimum lane width LWmin, step S403 is executed. Furthermore, if the distance between the lane detection range and the road end is less than the minimum lane width LWmin, this means that the area from the end of the lane detection range to the road end is not a lane, but rather a curb. On the other hand, if the distance between the right end of the lane detection range and the right road end is greater than the minimum lane width LWmin, step S402 is executed.

[0117] In step S402, a determination is made as to whether there are clearly no undetected lane boundaries on the right side of the lane detection range. For example, if the road edge is visible within the lane detection range of the front camera 11, or if the road surface from the lane detection range to the road edge is captured, a positive determination is made in step S402, and the process proceeds to step S403. The lane detection range of the front camera 11 is defined by the specifications of the front camera 11, such as the horizontal angle range of three lanes. Capturing the road surface from the lane detection range to the road edge, for example, includes the absence of other vehicles on the right side of the lane detection range.

[0118] On the other hand, if it is unclear that there is no undetected lane boundary line on the right side of the lane detection range, a negative determination is made in step S402, and the process proceeds to step S404. The situation where it is unclear that there is no undetected lane boundary line on the right side of the lane detection range includes, for example, a situation where the road end is located one lane or more outside the lane detectable range of the front camera 11. Furthermore, situations where it is unclear that there is no undetected lane boundary line on the right side of the lane detection range include situations where part or all of the road surface from the lane detection range to the road end cannot be captured due to other vehicles on the right side of the lane detection range. In other words, if there is a possibility that there is an undetected lane boundary line on the right side of the lane detection range, a negative determination is made in step S402, and step S404 is executed. The possibility that there is an undetected lane boundary line on the right side of the lane detection range also includes situations where the detection reliability of the lane boundary line at the right end of the specified lane detection range is low, or where the detection reliability of the road end itself is low.

[0119] In step S403, the distance from the right end of the lane detection range to the right end of the road is calculated as the width of the curb strip set on the right side of the road. This process is equivalent to determining that the area from the end of the lane detection range to the road end is not a lane, that is, a curb strip. On the other hand, in step S404, the width of the curb strip set on the right side of the road is set to a specified value (for example, 2.5m) and the process ends. In addition, the content of step S404 can also be regarded as the curb strip width being unknown, and the lane determination process can be interrupted. In the case where the lane determination process is interrupted, it can also be determined that the driving lane is unknown.

[0120] By performing the above processing on the road component located to the left of the vehicle, in other words, the left end of the lane detection range and the left road end, the width of the curb located on the left side of the road can also be calculated. The curb width calculated by this process is verified in step S6 described later. Therefore, the curb width calculated by this process corresponds to the width of the curb-like area, in other words, the candidate curb width.

[0121] According to the above configuration, since the curb width is calculated based on the distance between the lane detection area and the road end, an actual value can be set as the curb width, compared to a configuration in which the curb size is set to a fixed value. In a configuration in which the curb size is set to a fixed value, in the presence of a wide curb, such as in a setback zone, the fixed curb width and the actual curb width may deviate significantly, potentially leading to the driver mistaking the lane. In contrast, the above configuration can reduce the risk of mistaking the lane due to a deviation between the system-used curb width value and the actual value.

[0122] Furthermore, the above processing corresponds to a configuration in which information about the detected road end is used for lane determination when the distance between the outermost identifiable lane boundary, i.e., the outermost detection line, and the road end is less than the specified minimum lane width. Furthermore, the above processing corresponds to a configuration in which information about the detected road end is used for lane determination when there is clearly no lane boundary between the outermost detection line and the road end. Furthermore, the outermost detection line is a concept that applies independently to the left and right sides of the vehicle. Furthermore, the above processing corresponds to a configuration in which, based on the distance between the outermost detection line and the road end being less than the specified minimum lane width, the area from the outermost detection line to the road end is considered a curb.

[0123] <Step S5: Road End Validity Determination Process>

[0124] Here, use Figure 13 The road end validity determination process is explained. This process verifies whether the lateral position of the road end determined based on the regression line of the road end is correct. The road end does not exist within the lane, that is, within the lane detection range. This process is created with this positional relationship in mind and, as an example, includes steps S501 to S506. Figure 13 The process shown as Figure 4 The lane determination process is performed in step S5 shown.

[0125] In step S501, a determination is made as to whether the position of the left road end at the determination point determined by the regression equation is to the left of the lane detection range calculated in step S3. For example, a determination is made as to whether the lateral position of the left road end determined by the regression equation is to the left of the lateral position of the left end of the lane detection range. Furthermore, if the lateral position of the left road end is to the right of the lateral position of the left end of the lane detection range, this means that the calculated position of the left road end, determined by the regression equation, is within the lane detection range.

[0126] If the estimated position of the left road end is on the left side of the lane detection range, a positive determination is made in step S501, and the process proceeds to step S502. On the other hand, if the estimated position of the left road end is to the right of the left end of the lane detection range, a negative determination is made in step S501, and the process proceeds to step S503.

[0127] In step S502, the estimated position of the left road end determined by the regression equation is adopted as the lateral position of the left road end, and the process proceeds to step S504. In step S503, the estimated position of the left road end determined by the regression equation is discarded, and the process proceeds to step S504. This process corresponds to a determination that the calculated position of the left road end is incorrect. If step S503 is executed, the position of the left road end is considered unknown.

[0128] In step S504, a determination is made as to whether the estimated position of the right road end, determined by the regression equation, lies to the right of the lane detection range calculated in step S3. For example, a determination is made as to whether the lateral position of the right road end, determined by the regression equation, lies to the right of the right end of the lane detection range. If the lateral position of the right road end lies to the left of the right end of the lane detection range, this means that the estimated position of the right road end, determined by the regression equation, lies within the lane detection range.

[0129] If the estimated position of the right road end is on the right side of the lane detection range, a positive determination is made in step S504, and the process proceeds to step S505. On the other hand, if the estimated position of the right road end is on the left side of the right end of the lane detection range, a negative determination is made in step S504, and the process proceeds to step S506.

[0130] In step S505, the estimated position of the right road end determined by the regression equation is adopted as the lateral position of the right road end, and the process ends. In step S506, the estimated position of the right road end determined by the regression equation is discarded, and the process ends. This process is equivalent to determining that the calculated position of the right road end is incorrect. In the case of step S506, the position of the right road end is treated as unknown.

[0131] The above process can detect incorrect values ​​for the lateral position of the road edge at the decision point determined by the regression equation, such as due to misidentification of another vehicle's edge as part of the road edge, and discard the calculated result. This reduces the risk of incorrect lane determination due to misidentification of the road edge position.

[0132] <Step S6: Map Compatibility Determination Process>

[0133] Here, use Figure 14 The map compatibility determination process is described. The map compatibility determination process is a process for determining whether the curb width, lane detection range, road end position, etc. calculated in the process so far are properly compared with the data registered on the map. In this embodiment, as an example, the map compatibility determination process includes steps S601 to S606. Figure 14 The process shown as Figure 4 The lane determination process is performed in step S6 shown.

[0134] First, in step S601, a determination is made as to whether the lateral positions of the left and right road ends at the determination point can be acquired. For example, if the position information for at least one of the left and right road ends was discarded in steps S205 of the road end slope determination process, S503 of the road end validity determination process, or S506, the determination in step S601 is negative. If the lateral positions of the left and right road ends at the determination point can be acquired, the determination in step S601 is positive, and step S602 is executed. On the other hand, if the lateral position of either of the left and right road ends at the determination point cannot be acquired, the determination in step S601 is negative, and step S606 is executed.

[0135] In step S602, the difference between the left and right road end positions is calculated as the road width RW, and the process proceeds to step S603. In step S603, the estimated number of lanes, Nf, is calculated by subtracting the calculated curb width RsW from the road width RW calculated in step S602 and dividing it by the standard lane width LWstd. In other words, the estimated number of lanes Nf is calculated as (RW - RsW) / LWstd. The estimated number of lanes Nf is a real number that includes decimal points. The estimated number of lanes Nf can be expressed, for example, using a floating-point notation format.

[0136] The standard lane width LWstd is set to a standard value for the width of the lane based on the laws and regulations of the area where this vehicle is used. For example, the standard lane width LWstd is set to 3.0m. Of course, the minimum lane width LWmin can also be 2.5m, 2.75m, 3.5m, etc. The set value of the standard lane width LWstd can also be changed according to the type of road on which this vehicle is traveling. For example, the standard lane width LWstd when this vehicle is traveling on a highway can be set to be larger than the standard lane width LWstd when this vehicle is traveling on a general road. According to such a structure, since the set value corresponding to the road type is applied as the standard lane width LWstd, the concern about misjudgment in step S604 can be suppressed. If the calculation processing in step S603 is completed, step S604 is executed.

[0137] In step S604, a determination is made as to whether the absolute value of the difference between the number of lanes registered in the map data, i.e., the mapped lane number Nm, and the estimated lane number Nf, is below a predetermined threshold. The threshold used here serves as the upper limit of the permissible error range. If all recognition results are completely correct, the difference between the mapped lane number Nm and the estimated lane number Nf, i.e., the lane number difference, is 0. However, since the estimated lane number Nf has a decimal component, even if the front camera 11 can accurately detect the road edge and various lane boundaries, the lane number difference can take values ​​such as 0.1 or 0.3. The threshold for the absolute value of the lane number difference is set based on the permissible error level, for example, 0.5. Of course, the threshold for the lane number difference can also be 0.3, 1, etc. Furthermore, if the lane number difference exceeds 1, this means that there is a recognition error of one or more lanes within the road width or curb width. Therefore, it is preferable to set the threshold for the lane number difference to less than 1.

[0138] If the absolute value of the lane number difference is less than the specified threshold, a positive determination is made in step S604, and the process proceeds to step S605. On the other hand, if the absolute value of the lane number difference is greater than the specified threshold, a negative determination is made in step S604, and the process proceeds to step S606. Furthermore, a lane number difference greater than the specified threshold indicates a significant discrepancy between the content registered on the map and the content identified by the position estimator 20 (in other words, a mismatch). The content identified by the position estimator 20 here can be, for example, at least one of the road width and the curb width.

[0139] In step S605, it is determined that the content registered on the map matches the content identified by the position inference unit 20, and the process ends. In step S606, the lane determination process is interrupted because some part of the acquired data related to the road structure (for example, the calculated value of the curb width) is incorrect. If the lane determination process is interrupted, it can be determined that the driving lane is unknown. Alternatively, in step S606, the curb width can be reset to a specified value and subsequent processing can be continued.

[0140] The above configuration makes it possible to determine whether the calculated values ​​for the curb width and the calculated values ​​for the road end position are correct. For example, if the lane boundary adjacent to the road end cannot be detected, treating the area corresponding to the lane as the curb may result in a significant difference in the curb width. Regarding this issue, if the calculated value for the assumed curb width is too large relative to the actual curb width, the lane number difference calculated using the above processing may be a value greater than 1, such as 1.5, and thus be incorrectly processed. In other words, the above processing can detect anomalies in the calculated curb width and perform error processing.

[0141] In addition, in step S606, the value obtained by multiplying the number of map lanes Nm by the standard lane width LWstd can also be calculated as the map basic driving area width, and the value obtained by subtracting the map basic driving area width from the road width RW is used as the curb width RsW.

[0142] Furthermore, while the above disclosure uses the difference between the number of mapped lanes and the estimated number of lanes as a metric for determining the compatibility between the map and the recognition / calculation content of road structure-related information, such as curb width, in the position inference unit 20, this is not limiting. For example, the difference between the recognition-based driving area width (defined as the road width RW minus the curb width RsW) and the map-based driving area width determined based on data registered on the map could also be used as a metric. The driving area here refers to the entire area of ​​the road where lanes are formed. In other words, the driving area corresponds to the portion of the road excluding the curb and sidewalk.

[0143] The lane determination unit F4 also calculates the map-based curb width by subtracting the map-based driving area width from the road width registered on the map or the road width RW calculated using the regression line. Alternatively, the difference between the curb width RsW calculated in the above process and the map-based curb width may be used as a determination indicator.

[0144] Alternatively, if the road width is registered on a map, the difference between the road width calculated in step S602 and the registered road width can be used as the determination indicator. In this case, while the validity of the calculated curb width cannot be evaluated, the validity of the calculated road end position can be determined.

[0145] <Step S8: Individual Lane Position Identification Process>

[0146] Here, use Figure 15 The individual lane position determination process is described. The individual lane position determination process is a process for calculating the lateral position range of each detected lane. In this embodiment, as an example, the individual lane position determination process includes steps S801 to S803. Figure 15 The process shown as Figure 4 The lane determination process is performed in step S8 shown.

[0147] First, in step S801, the road width is calculated by multiplying the lane width by the number of lanes and adding the calculated curb width RsW. Furthermore, the road width is used to determine the upper and lower limits of the possible lateral position range of the lane boundary line (in other words, the range of X coordinates), with the left or right road end as a reference. For example, if the road width is calculated to be 15 meters, and the X coordinate of the left road end, which serves as the reference, is 5 meters to the left of the vehicle, the lower limit of the lateral position of the lane boundary line is set at a point 5.0 meters to the left of the vehicle (X: -5.0). Furthermore, the upper limit of the lateral position of the lane boundary line is set at an X coordinate 10 meters to the right of the vehicle (X: +10.0).

[0148] The road end that serves as the reference can be any road end whose lateral position can be calculated. For example, if the lateral position of the right road end has been successfully calculated, but the lateral position of the left road end is unknown, the upper and lower limits of the lateral position of the lane boundary line are set using the right road end as the reference. In addition to cases where the lateral position of the road end is unknown, this also includes cases where the road end information is discarded during the road end validity determination process in step S4, for example. Furthermore, assuming that the lateral positions of both the left and right road ends can be obtained, the road end closest to the vehicle can be used as the reference. This is because the position estimation accuracy of the road end closest to the vehicle can be expected to be higher than that of the road end farther away. The lane width used in this step S801 can be, for example, the width of the driving lane calculated in step S7. Alternatively, the number of lanes registered on the map can be used. Once the processing in step S801 is completed, the process proceeds to step S802.

[0149] In step S802, pairs of adjacent detected boundary lines are set as lane candidates, and the process proceeds to step S803. In step S803, lane candidates whose lateral positions (in other words, X coordinates) of both left and right detected boundary lines fall within the range of possible lane boundary line positions set in step S801 are identified as lanes. Furthermore, the position coordinates of the right and left boundary lines that constitute the identified lane are determined.

[0150] Through the above processing, the position of the lane boundary line detected by the front camera 11 can be determined, and the lane number from the right / left that the vehicle is traveling in can be determined. In other words, the lane ID of the driving lane can be determined.

[0151] <Effects of the above structure>

[0152] The proposed structure described above determines the driving lane based on the distance from the road end, taking into account the actual curb width. This eliminates the need to use the trajectories of surrounding vehicles. Therefore, the driving lane can be determined even when there are no surrounding vehicles.

[0153] Furthermore, in the above configuration, the curb width is calculated based on the results of identifying the road ends and lane boundaries, and the calculated curb width is used to determine the driving lane. As a comparative configuration, a configuration using a suitably designed fixed curb width to determine the driving lane is considered. However, the curb width varies depending on the location. For example, in a configuration where the curb width is set to a relatively small value, such as 0.5m, the deviation between the designed curb width and the actual curb width increases in road sections with wide curbs, such as setback sections. This can lead to areas that are actually curbs being mistakenly identified as the first lane, potentially misidentifying the driving lane. Alternatively, in a configuration where the curb width is set to a relatively large value, such as 2.2m, areas that are actually the first lane can be mistakenly identified as the curb in road sections with little curb, potentially misidentifying the driving lane. In such a comparative configuration where the curb width is set to a fixed value, the deviation between the actual curb width and the designed curb width can lead to misidentification of the driving lane.

[0154] In contrast to this comparative configuration, the proposed configuration calculates the distance between the lane's lateral position (lane detection area) and the road edge and uses it as the curb width. This configuration dynamically applies a value close to the actual curb width to determine the lane. Consequently, the vehicle's lane can be determined with greater accuracy than with the comparative configuration.

[0155] Furthermore, road edges don't skim past like white lines. Furthermore, road edges are often three-dimensional. Therefore, even in snowy or flooded conditions, road edges are easily detected from lane boundaries. Therefore, the proposed structure for determining lanes based on road edges offers the advantage of improved robustness compared to structures that determine lanes based on the type of lane boundary.

[0156] The embodiments of the present disclosure have been described above, but the present disclosure is not limited to the above-mentioned embodiments, and the various modifications described below are also included in the technical scope of the present disclosure, and in addition to the following, various changes can be made and implemented without departing from the main purpose. For example, the various modifications described below can be appropriately combined and implemented within the scope that does not produce technical contradictions. In addition, for components having the same functions as the components described in the aforementioned embodiments, the same figure marks are marked and their descriptions are omitted. In addition, when only a part of the structure is mentioned, the structure of the embodiment described above can be applied to the other parts.

[0157] <About the operation of the calculated value of the curb width>

[0158] It can also be configured so that when the calculated value RsW of the curb width exceeds a predetermined curb threshold, it is determined that the calculated value RsW of the curb width is erroneous, and the lane determination process is interrupted. The curb threshold is the maximum value of the possible width of the curb. For example, in Japan, the curb threshold can be set to 2.5m. The curb width threshold can also be set to 0.75m, 1.5m, etc. The curb threshold is preferably set based on the laws and regulations of the region where the position inference device 20 is used. In addition, the curb threshold can also be changed according to the type of road on which the vehicle is traveling. On highways, a relatively large value such as 2.0m can be applied, while on general roads, it can be set to 0.75m or 0.8m.

[0159] In addition, if Figure 16 As shown, the curb width calculated by the position inference unit 20 can also be uploaded to the map generation server 6 that generates and updates the road map together with the detector data from multiple vehicles. The V2X vehicle-mounted device 14 can also be configured to wirelessly transmit the curb width data determined by the position inference unit 20 together with the position coordinates of the determination point to the map generation server 6. According to such a structure, the map generation server 6 can create a high-precision map containing curb width information based on the curb width information uploaded from multiple vehicles. In addition, Figure 16 , some components of the driving assistance system 1, such as the front camera 11, are omitted. Similarly, the position estimator 20 may be configured to cooperate with the V2X vehicle-mounted device 14 to upload the calculated position coordinates of the road end and the lane boundary to the map generation server 6.

[0160] Furthermore, the driving assistance ECU 30 can utilize the curb width calculated by the position estimator 20 as described below. Specifically, when implementing an MRM (Minimum Risk Maneuver), the driving assistance ECU 30 can use the curb width information obtained from the position estimator 20 to determine whether a curb exists in front of the vehicle where the vehicle can safely park. For example, if the position estimator 20 notifies the driver that a curb of sufficient width exists, the driving assistance ECU 30 can create a driving plan toward the curb as the MRM. Alternatively, if the position estimator 20 notifies the driver that a curb does not exist, the driving assistance ECU 30 can create a driving plan to park within the current lane. In other words, by obtaining curb width information from the position estimator 20, the driving assistance ECU 30 can choose between slowly parking within the current lane or parking on the curb as the MRM action. By being able to select parking on the curb as the MRM, the risk of rear-ending another vehicle after parking can be reduced. Furthermore, there is an advantage in that the risk of a following vehicle coming into contact with a vehicle stopped by MRM can be reduced.

[0161] <About the processing of detection results at the roadside>

[0162] In addition, when there is a road end in one or more lanes to the right of the rightmost lane boundary line (in other words, the outermost detection line on the right) detected by the front camera 11, it is unclear whether the entire area from the rightmost detection line to the right road end is a curb or there is an undetected lane. Therefore, it can also be configured so that when the distance from the rightmost detection line to the right road end is greater than the minimum lane width LWmin, the right road end is not used as the road end. In this case, the lane position can be determined by using the road end on the opposite side (in other words, the left side). In addition, when the distance from the rightmost detection line to the right road is less than the minimum lane width LWmin, the right road end is used as the road end. The right curb width can be the distance from the rightmost detection line to the right road end. The same is true when there is a road end in one or more lanes to the left of the leftmost lane boundary line (in other words, the outermost detection line on the left) detected by the front camera 11.

[0163] <Supplementary information on roadside detection methods>

[0164] The front camera 11 of this embodiment detects the terminal portion of the road surface as the road end. For example, if the outer side of the road is unpaved, the boundary between the paved road and the unpaved portion is detected as the road end by analyzing the brightness distribution of the image frame (e.g., through edge detection). Furthermore, if a three-dimensional structure, such as a sidewalk step, is formed at the end of the road, the junction between the road end structure and the road surface—in other words, the lowest portion of the road end structure—is detected as the road end. The portion of the structure that rises from the road surface can also be detected as the road end by analyzing the brightness distribution of the image frame.

[0165] The following effects are achieved by detecting the boundary of a structure erected from the road surface as a road edge. Figure 17 As shown, there is a case where a step portion that rises several tens of centimeters from the road surface is erected on the outer side of the side wall of the highway. In a structure that assumes that the position of the side wall is used as the road end, there is a concern that the portion where the step portion is formed will be mistakenly identified as a drivable portion. As a result, there is a concern that the tire or the vehicle body will come into contact with the step portion. For such an assumed structure, the concern about the above-mentioned problem can be reduced by using a structure that uses the terminal portion of the road surface plane as the road end as in the present embodiment. As another embodiment, the position inference device 20 can also be configured to use the position of the side wall, guardrail, etc. as the position of the road end.

[0166] <About the use of driving lane information>

[0167] The lane ID of the driving lane determined by the above-mentioned position inference device 20 is, for example, information necessary when implementing autonomous driving accompanied by lane changes. For example, the driving assistance ECU 30 as an autonomous driving device can also be configured to create a driving plan accompanied by lane changes based on the lane ID of the driving lane determined by the position inference device 20. In addition, as an operational design domain (ODD), there may be a case where autonomous driving roads are specified in units of lanes. Under such constraints, the autonomous driving system needs to determine with high precision whether the vehicle is in a lane that can be driven autonomously. For example, when it is determined that the vehicle is not in a lane that can be driven autonomously, or has to exit a lane that does not allow autonomous driving, it is necessary to implement system responses such as transferring driving authority to the driver's seat occupant, the operator, or executing MRM. For such needs, the structure disclosed in the present invention can be useful because it has high robustness to the driving environment.

[0168] Furthermore, in HMI systems that include navigation devices, it is necessary to display turn-by-turn information, facility guidance information, and other features on a head-up display (HUD) in a manner that matches the real world. If the vehicle's lane identification deviates from a lane, the image displayed on the HUD deviates significantly from the real world. In other words, there is a need for a structure that can accurately determine the vehicle's lane not only in the field of autonomous driving but also in the field of navigation technology. The proposed structure can also be useful in meeting such needs.

[0169] <Supplementary information on system structure>

[0170] In the above embodiment, the position estimator 20 is arranged outside the front camera 11, but the arrangement of the position estimator 20 is not limited to this. Figure 18 As shown in FIG. 4 , the function of the position inference device 20 may also be built into the camera ECU 41. Figure 19 As shown, the function of the position inference device 20 can also be built into the driving assistance ECU 30. The driving assistance ECU 30 including the function of the position inference device 20 is equivalent to the driving control device. In addition, the driving assistance ECU 30 can also have the function of the camera ECU 41 (mainly the identifier G1). That is, the front camera 11 can also be configured to output image data to the driving assistance ECU 30, and the driving assistance ECU 30 performs image recognition and other processing. Figure 18 、 Figure 19 In FIG. 1 , a part of the configuration included in the driving support system 1 is omitted from illustration.

[0171] In addition, the above discloses a structure in which the position estimator 20 uses the front camera 11 to detect the position of the road side relative to the vehicle, but the present invention is not limited to this. Figure 20 As shown, the device for detecting the position of the road end relative to the vehicle may also be a millimeter wave radar 19A, LiDAR (Light Detection and Ranging / Laser Imaging Detection and Ranging: Light Detection and Ranging / Laser Imaging Detection and Ranging) 19B, sonar, etc. Millimeter wave radar 19A, LiDAR19B, sonar, etc. are equivalent to ranging sensors. In addition, it is also possible to use a side camera that takes pictures of the side of the vehicle and a rear camera that takes pictures of the rear to detect the road end and lane boundary line. The front camera 11, the side camera, and the rear camera are equivalent to the shooting device. The position inference device 20 may also be configured to use a plurality of devices to detect the road end, etc. In other words, the position inference device 20 may also determine the position of the road end and the lane boundary line by sensor fusion. The position of the road end may also use the position recognized by the recognition software as it is without using the regression line.

[0172] <Postscript (1)>

[0173] The control unit and methods described in this disclosure may also be implemented by a dedicated computer, wherein the dedicated computer comprises a processor programmed to execute one or more functions embodied by a computer program. Furthermore, the apparatus and methods described in this disclosure may also be implemented by dedicated hardware logic circuits. Furthermore, the apparatus and methods described in this disclosure may also be implemented by one or more dedicated computers, wherein the dedicated computers comprise a combination of a processor that executes a computer program and one or more hardware logic circuits. Furthermore, the computer program may be stored as computer-readable, non-transitory, tangible recording media as instructions executed by the computer. For example, the units and / or functions provided by the position estimator 20 may be provided by software stored in a physical memory device and a computer that executes the software, software alone, hardware alone, or a combination thereof. Some or all of the functions provided by the position estimator 20 may be implemented as hardware. Implementing a function in hardware includes using one or more integrated circuits, etc. The processing unit 21 may also be implemented using an MPU, GPU, or DFP (Data Flow Processor) instead of a CPU. Furthermore, the processing unit 21 may be implemented by combining multiple types of processing devices, such as CPUs, MPUs, and GPUs. ECUs can also be implemented using FPGAs (field-programmable gate arrays) or ASICs (application-specific integrated circuits). Programs can be stored on non-transitory tangible storage media. Program storage media include HDDs (hard disk drives), SSDs (solid state drives), EPROMs (erasable programmable read-only memories), flash memory, USB flash drives, and SD (secure digital) memory cards.

[0174] <Postscript (2)>

[0175] The present disclosure also includes the following technical ideas.

[0176] [Structure (1)]

[0177] A curb identification device comprising:

[0178] A boundary line information acquisition unit (F32) acquires position information of a detected lane boundary line by analyzing an image generated by a camera (11) that captures the surrounding environment of the vehicle;

[0179] a road edge information acquisition unit (F31) that acquires position information of a road edge relative to the vehicle using at least one of a camera and a distance measuring sensor (19A, 19B), wherein the distance measuring sensor detects an object in a predetermined direction of the vehicle by transmitting a detection wave or a laser; and

[0180] A curb width calculation unit (F41) calculates a lateral distance from an outermost detection line to a road end as a candidate curb width, wherein the outermost detection line is the outermost lane boundary line among the detected lane boundary lines.

[0181] Under the condition that the curb candidate width is less than a specified value, the area from the outermost detection line to the road end is determined to be the curb.

[0182] Generally speaking, since the width of the curb (in other words, the shoulder) varies depending on the road section, it is difficult to distinguish whether it is a curb or a lane. According to the above structure, the area equivalent to the curb can be identified with high precision. As an assumed structure related to the identification of the curb, a structure is also considered to determine whether the outer side of the lane boundary line is a curb based on the type of lane boundary line (solid line, dashed line). However, since the lane boundary line is a solid line / dashed line, its outer side may not necessarily correspond to the curb. In other words, in the assumed structure, there is a concern that whether it is a curb or not may be incorrectly determined.

[0183] According to the above-mentioned configuration (1), whether the area between the road edge and the detected lane boundary line corresponds to the curb is determined based on the distance therebetween, thereby reducing the possibility of misjudging whether the area corresponds to the curb.

[0184] [Structure (2)]

[0185] The curb recognition device according to the above configuration (1) is configured to transmit the calculated curb width to the automatic driving device.

[0186] [Structure (3)]

[0187] According to the roadside strip identification device described in the above structure (1),

[0188] A vehicle position acquisition unit (F1, F5) is provided for acquiring the position of the vehicle.

[0189] The curb recognition device is configured to transmit information indicating the calculated curb width together with the position information of the vehicle to a map generation server disposed outside the vehicle.

[0190] [Structure (4)]

[0191] A method for identifying a curb, comprising:

[0192] Step (S1) is to obtain the position information of the detected lane boundary line by analyzing the image generated by the camera (11) for shooting the surrounding environment of the vehicle, and

[0193] Using at least one of a camera and a distance measuring sensor (19A, 19B) to obtain position information of a road end relative to the vehicle, wherein the distance measuring sensor detects an object existing in a predetermined direction of the vehicle by transmitting a detection wave or a laser; and

[0194] The curb width calculation step (S4) calculates the lateral distance from the outermost detection line to the road end as the candidate curb width, wherein the outermost detection line is the outermost lane boundary line among the detected lane boundary lines.

[0195] Under the condition that the curb candidate width is less than a specified value, the area from the outermost detection line to the road end is determined to be the curb.

[0196] [Structure (5)]

[0197] A curb identification device comprising:

[0198] A boundary line information acquisition unit (F32) acquires position information of a detected lane boundary line by analyzing an image generated by a camera (11) that captures the exterior of the vehicle;

[0199] A road edge information acquisition unit (F31) acquires position information of a road edge relative to the vehicle using at least one of a distance measuring sensor (19A, 19B) and a camera, wherein the distance measuring sensor detects an object in a predetermined direction of the vehicle by transmitting a detection wave or a laser;

[0200] a curb width calculation unit (F41) for calculating a lateral distance from an outermost detection line to a road end as the curb width, wherein the outermost detection line is the outermost lane boundary line among the detected lane boundary lines;

[0201] A map acquisition unit (F2) acquires map information including the number of lanes of a road on which the vehicle is traveling from a map storage unit disposed inside or outside the vehicle; and

[0202] A driving lane determination unit (F4) determines a driving lane of the vehicle based on the position information of the road end acquired by the road end information acquisition unit, the number of lanes included in the map information acquired by the map acquisition unit, and the curb width calculated by the curb width calculation unit.

[0203] The driving lane identification unit is configured to output information on the driving lane identified by the driving lane identification unit to a predetermined operation recording device.

Claims

1. A vehicle position estimation device comprising: The road end information acquisition unit acquires the distance from the vehicle to the road end using at least one of a camera and a distance measuring sensor, wherein: The camera captures a predetermined area around the vehicle, and the distance sensor detects objects in a predetermined direction of the vehicle by transmitting a detection wave or a laser. a map acquisition unit that acquires map information including the number of lanes of a road on which the vehicle is traveling from a map storage unit disposed inside or outside the vehicle; a boundary line information acquisition unit, which acquires position information of the detected lane boundary line by analyzing the image generated by the above-mentioned camera device; a curb width calculation unit that calculates a lateral distance from an outermost detection line to the road end as the curb width, wherein the outermost detection line is the outermost lane boundary line among the detected lane boundary lines; and a driving lane determination unit that determines a driving lane of the host vehicle based on the distance from the host vehicle to the road end, the curb width, and the number of lanes included in the map information acquired by the map acquisition unit. The curb width calculation unit performs a map matching determination process in which the lateral distance from the outermost detection line, which is the outermost lane boundary line, to the road end, calculated as the curb width, is compared with data registered on the map to determine whether the calculated curb width is appropriate. If the result of the map matching determination process is that the calculated curb width matches the data registered on the map, the vehicle position estimation device determines the driving lane of the vehicle using the calculated curb width.

2. The vehicle position estimation device according to claim 1, wherein: If the width of the curb is less than a specified threshold, the area from the outermost detection line to the road end is considered the curb. The driving lane determination unit determines the driving lane based on a lane adjacent to the curb.

3. The vehicle position estimation device according to claim 1 or 2, wherein: The driving lane identification unit is configured not to use information on the lane boundary line detected outside the road end for identification of the driving lane.

4. The vehicle position estimation device according to any one of claims 1 to 3, wherein: The driving lane determination unit is configured to not use information on the outer lane boundary line of the two adjacent lane boundary lines for determining the driving lane when the distance between the two adjacent lane boundary lines is less than a predetermined minimum lane width.

5. The vehicle position estimation device according to any one of claims 1 to 4, wherein: When the driving lane determination unit can obtain position information of both a left road end existing on the left side of the vehicle and a right road end existing on the right side of the vehicle as the road end, An estimated value of the number of lanes, namely, the estimated number of lanes, is calculated using the distance between the left road end and the right road end and the curb width, and If the difference between the number of lanes registered on the map and the estimated number of lanes is equal to or greater than a predetermined threshold, it is determined that the calculated curb width is incorrect, or the determination of the driving lane is interrupted.

6. The vehicle position estimation device according to any one of claims 1 to 5, wherein: The driving lane determination unit determines a lane detection range, which is an area where at least a lane exists, based on the position information of the detected lane boundary line. When the position of the road end acquired by the road end information acquisition unit is inside the lane detection range, the position information of the road end is erroneous, or the information of the road end is not used for determining the driving lane, or the process of determining the driving lane is interrupted.

7. The vehicle position estimation device according to any one of claims 1 to 6, wherein: The data of the curb width calculated by the curb width calculation unit is uploaded to a map generation server.

8. The vehicle position estimation device according to any one of claims 1 to 7, wherein: The road edge information acquisition unit determines the slope of the road edge at a predetermined determination point in the image generated by the imaging device based on a regression line of the road edge generated using the detection points of the road edge as a parent group. When the slope of the road end is outside a predetermined normal range, the driving lane determination unit determines that the recognition result of the road end is erroneous, does not use the information of the road end for determining the driving lane, or interrupts the process of determining the driving lane.

9. The vehicle position estimation device according to any one of claims 1 to 8, wherein: When the driving lane identification unit successfully identifies the driving lane, the driving lane number, which is the number of the driving lane, is output to the outside.

10. A driving position estimation method for determining a lane in which a vehicle is traveling, executed by at least one processor, comprising: The position acquisition step is to obtain the position information of the detected lane boundary line by analyzing the image generated by the camera, and to obtain the distance from the vehicle to the road end using at least one of the camera and the distance measuring sensor, wherein: The camera captures a predetermined area around the vehicle, and the distance sensor detects objects in a predetermined direction of the vehicle by transmitting a detection wave or a laser. A map acquisition step of acquiring map information including the number of lanes of a road on which the vehicle is traveling from a map storage unit disposed inside or outside the vehicle; a curb width calculation step of calculating a lateral distance from an outermost detection line to the road end as the curb width, wherein the outermost detection line is the outermost lane boundary line among the detected lane boundary lines; and a driving lane determining step of determining the driving lane of the vehicle based on the distance from the vehicle to the road end obtained in the position obtaining step, the curb width calculated in the curb width calculating step, and the number of lanes included in the map information obtained in the map obtaining step; In the curb width calculation step, a map matching determination process is performed. In this map matching determination process, the lateral distance from the outermost detection line, which is the outermost lane boundary line, to the road end, which is calculated as the curb width, is compared with data registered on the map to determine whether the calculated curb width is appropriate. If the result of the map matching determination process is that the calculated curb width matches the data registered on the map, the travel lane of the host vehicle is determined using the calculated curb width.

Images