External identification device

Abstract

An outside world recognition device (50) is provided with: a vehicle-mounted detector (5); a recognition unit (111) that recognizes, as road surface information, the road surface of a road on which the host vehicle travels and a three-dimensional object on the road on the basis of the point cloud data for each frame acquired by the vehicle-mounted detector (5); a determination unit (113) that determines the interval of the detection points required for the point cloud data of the next frame on the basis of the size of a predetermined three-dimensional object that is predetermined as a recognition target and the measurement distance from the host vehicle to the three-dimensional object on the basis of the point cloud data; and a gradient prediction unit (114) that predicts the gradient of the road surface that is not recognized by the recognition unit (111) on the basis of gradient information associated with map information in which the road is recorded, and the determination unit (113) determines the gradient of the road surface that is not recognized by the recognition unit (111) when the gradient is predicted by the gradient prediction unit (114). On the basis of the size of a predetermined three-dimensional object and an estimated distance from the host vehicle to the three-dimensional object, which is estimated on the basis of map information and a gradient, the interval of detection points required for point cloud data of the next frame is determined for the range from the farthest distance to the necessary distance.

Description

Technical Field

[0001] This invention relates to an external identification device for recognizing the external conditions of a vehicle. Background Technology

[0002] As such a device, there is a known device that scans by varying the irradiation angle of the laser emitted from the lidar around a first axis parallel to the height direction and a second axis parallel to the horizontal direction, and detects the external environment of the vehicle based on the position information of each detection point (see, for example, Patent Document 1).

[0003] In the above-mentioned device, a large number of detection points are obtained through scanning, resulting in a heavy processing burden for obtaining the location information based on each detection point.

[0004] Monitoring a vehicle's external conditions allows it to move smoothly, improving traffic convenience and safety. This, in turn, contributes to the development of sustainable transportation systems.

[0005] Existing technical documents Patent documents Patent document 1: Japanese Patent Application Publication No. 2020-149079 (JP2020-149079A). Summary of the Invention

[0006] An external identification device according to a technical solution of the present invention comprises: a vehicle-mounted detector that scans and irradiates electromagnetic waves in a first direction and a second direction intersecting the first direction within its field of view, and acquires point cloud data frame by frame, including three-dimensional position information of detection points on the surface of objects based on reflected waves from objects around the vehicle; an identification unit that identifies the road surface and three-dimensional objects on the road as road surface information based on the point cloud data of each frame; and a decision unit that determines the next frame based on the size of a predetermined three-dimensional object as the object to be identified and the measured distance from the vehicle to the three-dimensional object based on the point cloud data. The interval of detection points required for the point cloud data; and the slope prediction unit, which, when the farthest distance of the road surface in the direction of travel of the vehicle identified by the recognition unit is shorter than the necessary distance based on the vehicle speed, predicts the slope of the road surface not identified by the recognition unit based on slope information associated with map information containing the road, and the determination unit, when the slope prediction unit predicts the slope, determines the interval of detection points required for the next frame of point cloud data for the range from the farthest distance to the necessary distance, based on the size of the specified three-dimensional object and the estimated distance from the vehicle to the three-dimensional object estimated based on the map information and the slope. Attached Figure Description

[0007] Figure 1A It is a diagram showing vehicles driving on the road; Figure 1BThis is a schematic diagram illustrating an example of lidar detection data; Figure 2 It is a block diagram that roughly shows the main structural components of the vehicle control unit; Figure 3A It is a graph that uses a three-dimensional coordinate system to represent the position of point cloud data in three-dimensional space; Figure 3B It is a diagram illustrating the mapping of point cloud data from three-dimensional space to two-dimensional XZ space; Figure 3C This is a schematic diagram illustrating point cloud data divided into grids; Figure 3D This is a schematic diagram showing the road surface slope in the direction of depth; Figure 4A This is a schematic diagram showing the projection angle and depth distance; Figure 4B This is a schematic diagram showing the distance measured by lidar; Figure 5A This is a schematic diagram illustrating an example of the relationship between depth distance and vertical projection angle; Figure 5B This is a schematic diagram illustrating an example of the relationship between depth distance and angular resolution in the vertical direction; Figure 6A This is a schematic diagram showing an example of the illumination point when the illumination light of a lidar is illuminated in a grating scanning manner; Figure 6B This is a schematic diagram illustrating an example of a lidar illumination point where the illumination light only illuminates a predetermined grid of points arranged in a grid pattern within the detection area; Figure 7 It shows the direction Figure 6B A diagram showing an example of the irradiation sequence when the irradiation point is illuminated by the irradiation light; Figure 8A This is a diagram illustrating one example of a prediction; Figure 8B This is a diagram illustrating one example of a prediction; Figure 9 It is shown by Figure 2 A flowchart illustrating an example of the processing performed by the CPU of the controller; Figure 10 This is an explanation Figure 9 The flowchart for the S20 process. Detailed Implementation

[0008] Hereinafter, embodiments of the invention will be described with reference to the accompanying drawings.

[0009] The external identification device of this invention can be applied to vehicles with autonomous driving capabilities, i.e., autonomous vehicles. It should be noted that sometimes the vehicle using the external identification device of this embodiment is distinguished from other vehicles and referred to as "this vehicle." This vehicle can be any of the following: an engine vehicle with an internal combustion engine as the driving source, an electric vehicle with a drive motor as the driving source, or a hybrid vehicle with both an engine and a drive motor as driving sources. This vehicle can operate not only in an autonomous driving mode without driver intervention but also in a manual driving mode based on driver intervention.

[0010] When an autonomous vehicle is driving in autonomous driving mode (hereinafter referred to as automatic driving or autonomous driving), it identifies the external conditions around the vehicle based on detection data from onboard detectors such as cameras and LiDAR (Light Detection and Ranging). Based on the identification results, the autonomous vehicle generates a driving trajectory (target trajectory) that has elapsed for a specified period of time from the current moment, and controls the driving actuators to make the vehicle travel along the target trajectory.

[0011] Figure 1A This is a diagram showing the vehicle 101, which is an autonomous driving vehicle, driving on road RD. Figure 1B This is a schematic diagram illustrating an example of detection data obtained by a lidar mounted on the vehicle 101 and moving in the direction of travel of the vehicle 101. The measurement point (also called the detection point) of the lidar is the point information reflected back from a point on the surface of an object by the irradiated laser. The point information includes the distance from the laser source to that point, the intensity of the reflected laser, and the relative velocity between the laser source and that point. Furthermore, [the following will be explained by...] Figure 1B The data consisting of multiple detection points shown is called point cloud data. Figure 1B It shows the basis Figure 1A The point cloud data of the detected points on the surface of the object contained within the field of view (FOV) of the lidar. The FOV can be, for example, 120 degrees in the horizontal direction (also known as the road width direction) and 40 degrees in the vertical direction (also known as the up-down direction) of vehicle 101. The FOV value can be appropriately changed according to the specifications of the external identification device. Vehicle 101, according to... Figure 1B The point cloud data shown is used to identify the external conditions around the vehicle, and more specifically, to identify the road structure and objects around the vehicle, and to generate a target track based on the identification results.

[0012] Furthermore, as a method to fully identify the external conditions around the vehicle, increasing the number of electromagnetic wave illumination points from vehicle-mounted detectors such as lidar is considered (in other words, increasing the density of electromagnetic wave illumination points, thus increasing the number of detection points constituting the point cloud data). On the other hand, increasing the number of electromagnetic wave illumination points (increasing the number of detection points) may increase the processing load on the vehicle-mounted detectors and increase the capacity of the detection data (point cloud data) obtained by the vehicle-mounted detectors, thereby increasing the processing load on the point cloud data. In particular, the capacity of the point cloud data will further increase when there are many objects on the road or beside the road.

[0013] Therefore, taking into account the above aspects, the external identification device is configured as follows in the embodiment.

[0014] <Outline> The external identification device of this embodiment intermittently irradiates the vehicle 101 traveling on the road RD with an illumination light, which is an example of an electromagnetic wave, in the direction of travel of the vehicle 101 via a lidar. Point cloud data is discretely acquired at different locations on the road RD. The illumination range of the illumination light from the lidar is set such that point cloud data from the previous frame acquired by the lidar based on the previous illumination and point cloud data from the next frame acquired by the lidar based on the current illumination are used to avoid creating blank intervals in the direction of travel of the road RD.

[0015] The detection point density within the illumination range is set higher for road surfaces far from vehicle 101 and lower for road surfaces close to vehicle 101. This reduces the total number of detection points used in the recognition process compared to setting a higher detection point density for all road surfaces within the illumination range. Therefore, the number of detection points used in the recognition process can be reduced without compromising the accuracy of recognizing the position (distance from vehicle 101) and size of objects identified from point cloud data. Furthermore, the number of laser elements in the lidar can be reduced, resulting in a smaller and less expensive lidar system.

[0016] A more detailed explanation of such external identification devices will be provided.

[0017] <Structure of Vehicle Control Device> Figure 2 This is a block diagram showing the main structural components of a vehicle control unit 100, including an external detection device. The vehicle control unit 100 includes a controller 10, a communication unit 1, a positioning unit 2, an internal sensor group 3, a camera 4, a lidar 5, and a driving actuator AC. Additionally, the vehicle control unit 100 includes an external detection device 50, which forms part of the vehicle control unit 100. The external detection device 50 identifies the external conditions around the vehicle based on detection data from onboard detectors such as the camera 4 and lidar 5.

[0018] Communication unit 1 communicates with various servers (not shown) via wireless communication networks, including the Internet and mobile phone networks, and periodically or at any time retrieves map information, driving history information, and traffic information from the servers. The network includes not only public wireless communication networks but also closed communication networks set up for each designated management area, such as wireless LAN, Wi-Fi (registered trademark), and Bluetooth (registered trademark). The retrieved map information is output to storage unit 12 to update the map information. The map information is correlated with a road slope map used for predicting road surface slope, which will be detailed later.

[0019] The positioning unit (GNSS unit) 2 has a positioning sensor that receives positioning signals transmitted from positioning satellites. Positioning satellites are artificial satellites such as GPS satellites and quasi-zenith satellites. The positioning unit 2 uses the positioning information received by the positioning sensor to determine the current position (latitude, longitude, and altitude) of the vehicle 101.

[0020] The internal sensor group 3 is a collective term for multiple sensors (internal sensors) that detect the driving status of the vehicle 101. For example, the internal sensor group 3 includes a vehicle speed sensor that detects the vehicle speed (driving speed), an acceleration sensor that detects the longitudinal acceleration and lateral acceleration (lateral acceleration) of the vehicle 101, a speed sensor that detects the rotational speed of the driving source, and a yaw rate sensor that detects the rotational angular velocity of the vehicle 101's center of gravity about its vertical axis. Sensors that detect the driver's driving operations in manual driving mode, such as operation of the accelerator pedal, operation of the brake pedal, and operation of the steering wheel, are also included in the internal sensor group 3.

[0021] Camera 4 has imaging elements such as CCD (charge-coupled device) and CMOS (complementary metal-oxide-semiconductor) to capture images of the surroundings (front, rear, and sides) of vehicle 101. LiDAR 5 receives the scattered light from the illumination light and measures the distance, position, and shape of objects from vehicle 101 to the surrounding area.

[0022] The actuator AC is a driving actuator used to control the movement of the vehicle 101. When the driving source is an engine, the actuator AC includes a throttle actuator for adjusting the opening of the engine's throttle valve (throttle opening). When the driving source is a drive motor, the drive motor is included in the actuator AC. The brake actuator that operates the braking system of the vehicle 101 and the steering actuator that drives the steering system are also included in the actuator AC.

[0023] The controller 10 is composed of an electronic control unit (ECU). More specifically, the controller 10 is configured as a computer having an arithmetic unit 11 such as a CPU (microprocessor), a storage unit 12 such as ROM (read-only memory) and RAM (random access memory), and other peripheral circuits (not shown) such as I / O interfaces. It should be noted that although multiple ECUs with different functions, such as an engine control ECU, a drive motor control ECU, and a braking device ECU, can be separately configured, Figure 2 For convenience, controller 10 will be referred to as a set of these ECUs.

[0024] Storage unit 12 is capable of storing high-precision, detailed map information (referred to as high-precision map information). This high-precision map information includes road location information, road shape (curvature, etc.) information, road slope information, location information of intersections and forks in the road, information on the number of lanes (driving lanes), lane width and the location information of each lane (the center position of the lane, information on the lane's boundary lines), location information of landmarks (traffic lights, signs, buildings, etc.) used as markers on the map, and road surface features such as unevenness. In addition to the two-dimensional map information described later, storage unit 12 can also store various control programs, thresholds used in the programs, and setting information for vehicle-mounted detectors such as the LiDAR 5 (illumination point information described later).

[0025] It should be noted that high-precision and detailed map information may not be required in the implementation, so detailed map information may not be stored in the storage unit 12.

[0026] The arithmetic unit 11 has a functional structure comprising an identification unit 111, a setting unit 112, a decision unit 113, a prediction unit 114, and a driving control unit 115. It should be noted that, as... Figure 2 As shown, the identification unit 111, setting unit 112, decision unit 113, and prediction unit 114 are included in the external environment identification device 50. As described above, the external environment identification device 50 identifies the external conditions around the vehicle based on detection data from vehicle-mounted detectors such as the camera 4 and the lidar 5. Details of the identification unit 111, setting unit 112, decision unit 113, and prediction unit 114 included in the external environment identification device 50 will be described later.

[0027] In automatic driving mode, the driving control unit 115 generates a target track based on the external conditions around the vehicle identified by the external recognition device 50, and controls the actuator AC to make the vehicle 101 travel along the target track. It should be noted that in manual driving mode, the driving control unit 115 controls the actuator AC based on driving commands (steering operations, etc.) from the driver obtained by the internal sensor group 3.

[0028] Further explanation of LiDAR 5.

[0029] <Detection Area> The lidar 5 is mounted facing forward of the vehicle 101 in such a way that the area to be observed during driving is included in the field of view (FOV). Since the lidar 5 receives light scattered by objects such as three-dimensional structures that have been illuminated, the FOV of the lidar 5 corresponds to the illumination range and detection area of ​​the illumination light. In other words, the irradiated point within the illumination range corresponds to the detection point within the detection area.

[0030] In this implementation, the term "three-dimensional objects" refers to road surface shapes, including unevenness, steps, and undulations; three-dimensional objects located on the road RD (such as equipment related to the road RD, including traffic signals, signs, ditches, walls, fences, guardrails, etc.); objects on the road RD (including other vehicles and road obstacles); and road markings on the road surface. Road markings include white lines (including lines of different colors such as yellow), curb lines, road studs, etc., and may also be called lane markings. Additionally, sometimes three-dimensional objects pre-defined as detection targets are referred to as detection objects.

[0031] <Example of a coordinate system> Figure 3A It is a graph that uses a three-dimensional coordinate system to represent the position of point cloud data in three-dimensional space. Figure 3A In the diagram, the positive x-axis corresponds to the direction of travel of the vehicle 101, the positive y-axis corresponds to the left side of the vehicle 101 in the horizontal direction, and the positive z-axis corresponds to the top in the vertical direction.

[0032] Furthermore, the x-axis component of the position of data P is called the depth distance X, the y-axis component of the position of data P is called the horizontal distance Y, and the z-axis component of the position of data P is called the height Z.

[0033] If the distance measured by lidar 5, in other words, the distance between lidar 5 and a point on the object being detected, is set as D, then the coordinates (X, Y, Z) representing the position of data P are calculated by the following formula.

[0034] X = D × cosθ × cos φ (1) Y = D × sinθ × cos φ (2) Z = D × sin φ (3) It should be noted that the angle θ This is called the horizontal projection angle. φ This is called the vertical projection angle. The horizontal projection angle... θ and plumb light angle φThe setting unit 112 sets the lidar 5.

[0035] Figure 3B This diagram illustrates the mapping of point cloud data from three-dimensional space to two-dimensional XZ space. In the implementation, to calculate the pavement slope of road RD, for each data point constituting the point cloud data, the data P in three-dimensional space is mapped to data P′ in XZ space. Through this mapping, the three-dimensional point cloud data is converted into two-dimensional point cloud data in XZ space. In XZ space, information representing horizontal distance Y is omitted, leaving only the depth distance X and height Z.

[0036] Next, the XZ space is divided into grids of a specified size (e.g., 50cm square), and the number of data P′ contained in each grid is counted. Figure 3C This is a schematic diagram representing point cloud data divided into grids. It should be noted that the actual number of grids in the data P′ is far greater than the number shown in the diagram.

[0037] according to Figure 3C This shows the position data (depth distance X) of each grid, the height Z of each grid, and the number of data P′ entering each grid. In this implementation, since the data of the three-dimensional object is separated and excluded beforehand, the grid data mainly consists of X and Z values ​​relative to the road surface. Therefore, by sequentially extracting the grid with the largest number of data P′ in the depth distance X direction, the following is obtained: Figure 3D The grid shown represents the height Z of the road surface, i.e., the road surface slope in the X direction of the depth distance.

[0038] When focusing on each grid, the following equation (4) holds true between the vertical projection angle α of the road surface point (corresponding to the above-mentioned illumination point) relative to that grid, the depth distance X of that road surface point, and the height Z of the road surface. In addition, the following equation (5) holds true between the distance DL from the lidar 5 to that road surface point, the depth distance X of that road surface point, and the height Z of the road surface.

[0039] tanα=Z / X(4) DL = (X² + Z²)¹ / ² (5) It should be noted that, in this embodiment, the pitch angle, roll angle, and yaw angle of the lidar 5 installed on the vehicle 101 are fixed. Furthermore, the road slope map described later can be generated from the road slope (the columns of the grid representing the road height Z mentioned above) obtained by the vehicle 101 or other vehicles similarly equipped with lidar 5, based on the road slope (columns of the grid representing the road height Z) obtained by each vehicle traveling along the driving route with the centerline of its vehicle width as its length.

[0040] As an example, each vehicle can also periodically or at any time send data representing the relationship between the elevation Z and depth X of the road surface along the acquired driving route to an external server device or the like via the communication unit 1. The external server device or the like associates the data representing the relationship between the average elevation Z and depth X of each road surface point on the driving route, which is sent by multiple vehicles, with the two-dimensional map information of the road RD containing the driving route to form a road surface slope map.

[0041] External server devices calculate the deviation of the height Z of each road point on the newly sent travel route from multiple vehicles from the height of each road point's height in the existing road slope map. For areas where the deviation exceeds a predetermined threshold depth X, the existing road slope map can be updated with the new average data.

[0042] In addition, external server devices can also assign data representing the relationship between the average height Z of each road surface point on the driving route and the depth distance X to the two-dimensional map information, as the height information of each one-dimensional road surface point for each route, instead of establishing a connection between the road slope map and the two-dimensional map information.

[0043] Each vehicle, including vehicle 101, periodically or at any time obtains the latest road slope map information (or the height information of each one-dimensional road point) from an external server device via communication unit 1 and stores it in storage unit 12.

[0044] <Projection angle and depth> Figure 4A This is a schematic diagram showing the vertical projection angle α (the angle of the illumination light relative to the horizontal direction) and depth X of the lidar 5. The external identification device 50 changes the illumination direction of the illumination light vertically by changing the projection angle α, thereby moving the position of the illumination point in the vertical direction.

[0045] exist Figure 4A For example, when illuminating a road RD at a location with a depth distance X2 of 10m, the road surface is illuminated at a projection angle α2. Similarly, when illuminating a road RD at a location with a depth distance X1 of 40m, the road surface is illuminated at a projection angle α1. Furthermore, when illuminating a road RD at a location with a depth distance X0 of 100m, the road surface is illuminated at a projection angle α0.

[0046] Generally speaking, the larger the projection angle relative to the road surface, the smaller the scattered light returning from the road surface to the lidar 5. Therefore, in many cases, the received level of the scattered light corresponding to the illumination light illuminating a location at a depth distance X0 is the lowest.

[0047] Figure 4B This is a schematic diagram showing the distance DL measured by the lidar 5. (Refer to the above text.) Figures 3A-3D The external identification device 50 uses the projection angle α set for the lidar 5, the distance DL (optical path length of the irradiated light) measured by the lidar 5, and the above formulas (4) and (5) to calculate the depth distance X up to the road surface point irradiated by the irradiated light and the height Z of the road surface point.

[0048] When a longer depth distance is desired than the current value, the external recognition device 50 sets the projection angle α upward; when a shorter depth distance is desired, the external recognition device 50 sets the projection angle α downward. For example, when changing the depth distance from 70m to 100m, the external recognition device 50 sets the projection angle α upward, ensuring that the light illuminates the 100m depth location. Furthermore, if the light does not reach the road RD, for example, if the road RD is downhill, the external recognition device 50 sets the projection angle α downward, ensuring that the light illuminates the road RD.

[0049] Figure 5A This is a schematic diagram illustrating an example of the relationship between the depth distance X and the vertical projection angle α. The horizontal axis represents the depth distance X (in meters), and the vertical axis represents the vertical projection angle α (in degrees). The projection angle α can also be called the vertical angle. Figure 5A As shown, when the depth distance X is to be shortened, the external recognition device 50 sets the projection angle α downward compared to the current position; when the depth distance X is to be increased, the external recognition device 50 sets the projection angle α upward compared to the current position. The symbol N will be explained below.

[0050] <FOV and depth distance> In the implementation, the depth distance corresponding to the lower end of the FOV of the lidar 5 is detected (e.g., Figure 4A (x2) to the depth distance corresponding to the upper end of the FOV (e.g.) Figure 4A The road surface condition (X0). The depth distance corresponding to the lower end of the FOV is called the first specified distance, and the depth distance corresponding to the upper end of the FOV is called the second specified distance.

[0051] Generally speaking, at close range, camera 4 has a better resolution than lidar 5, while lidar 5 has better distance measurement accuracy and relative speed measurement accuracy than camera 4. Therefore, given that camera 4 has a wider field of view in the vertical direction compared to lidar 5's field of view, the external identification device 50 can also use camera 4 to detect road conditions at positions lower than the lower end of lidar 5's field of view (in other words, closer to the road surface of vehicle 101).

[0052] <Number of points illuminated by the illumination light> The external identification device 50 calculates the position of the illumination point of the illumination light illuminating the lidar 5 within the FOV of the lidar 5. More specifically, the external identification device 50 calculates the illumination point based on the angular resolution calculated according to the minimum size (e.g., 15 cm both vertically and horizontally) of a three-dimensional object pre-designated as the detection object (which may also be the identification object) and the necessary depth distance (e.g., 100 m). The three-dimensional object may be, for example, a stone or concrete block on a road. The necessary depth distance corresponds to the braking distance of the vehicle 101, which varies according to the vehicle speed.

[0053] In this implementation, based on the idea that the road surface condition of the vehicle 101 in motion should detect a distance at least exceeding the braking distance, the value after adding a predetermined margin to the braking distance is called the necessary depth distance. The vehicle speed of the vehicle 101 is detected by the vehicle speed sensor of the internal sensor group 3. The relationship between vehicle speed and necessary depth distance is pre-stored in the storage unit 12. Figure 5A The symbol N in the figure represents the required depth distance at a vehicle speed of, for example, 100 km / h.

[0054] When illustrating the angular resolution for detecting a 15cm object at a distance of 100m in the direction of travel where the necessary depth distance is required, refer to the following text. Figure 5B As stated, 0.05 degrees are required in both the vertical and horizontal directions. It should be noted that when inspecting objects smaller than 15cm or when inspecting a 15cm object at a depth distance X greater than 100m, it is necessary to further increase the number of illumination points within the FOV by improving the angular resolution.

[0055] The external identification device 50 calculates the positions of the illumination points, for example, by arranging them in a grid pattern within the field of view (FOV), such that the vertical and horizontal spacing of the grid points corresponds to the angular resolution in the vertical and horizontal directions, respectively. When the vertical angular resolution is increased, the FOV is divided vertically by a number based on the angular resolution, narrowing the vertical grid spacing and increasing the number of illumination points. In other words, the spacing of the illumination points is made closer together. Conversely, when the vertical angular resolution is decreased, the FOV is divided vertically by a number based on the angular resolution, widening the vertical grid spacing and reducing the number of illumination points. In other words, the spacing of the illumination points is made larger. The same applies to the horizontal direction.

[0056] The external identification device 50 generates information indicating the position of the illumination point calculated based on the angular resolution (hereinafter referred to as illumination point information), and stores it in the storage unit 12 in a corresponding relationship with the position information indicating the current driving position of the vehicle 101.

[0057] <Angular resolution and depth> Figure 5B This is a schematic diagram illustrating an example of the relationship between depth distance X and vertical angular resolution, showing the angular resolution (also called the necessary angular resolution) required to identify a detection object of the aforementioned size (15cm in both directions). The horizontal axis represents depth distance X (in meters), and the vertical axis represents the vertical angular resolution (in degrees). Generally, the shorter the depth distance X (in other words, the closer the detection object is to the vehicle 101), the larger the viewing angle for the detection object, thus allowing detection even with a lower angular resolution. Conversely, the longer the depth distance X (in other words, the farther the detection object is from the vehicle 101), the smaller the viewing angle for the detection object, thus requiring a higher angular resolution for detection. Therefore, as... Figure 5B As illustrated, the shorter the depth distance X, the lower the angular resolution of the external recognition device 50 (the larger the value of the angular resolution), and the longer the depth distance X, the higher the angular resolution of the external recognition device 50 (the smaller the value of the angular resolution).

[0058] It should be noted that although the diagram is omitted, the relationship between the depth distance X and the angular resolution in the horizontal direction is the same.

[0059] Figure 5B The symbol N in the figure represents the required depth distance at a vehicle speed of, for example, 100 km / h.

[0060] When the vehicle 101 is driving in autonomous driving mode, the external identification device 50 controls the lidar 5 to illuminate a predetermined illumination point (detection point) within the field of view (FOV). As a result, the illumination light from the lidar 5 shines on the predetermined illumination point (detection point).

[0061] It should be noted that the illumination light of the lidar 5 can illuminate all the illumination points (detection points) arranged in a grid pattern within the FOV in a grating scanning manner, or it can illuminate the illumination light intermittently only to the specified illumination points (detection points), or it can illuminate in other ways.

[0062] Figure 6A This is a schematic diagram illustrating an example of the illumination point when the lidar 5 is illuminated by illumination light in a grating scanning manner. When the external identification device 50 is illuminating the lidar 5 with illumination light, it controls the illumination direction of the illumination light by setting the required angular resolution at a necessary depth distance N for the entire area within the field of view.

[0063] For example, when the necessary angular resolution for identifying a target at a location with a necessary depth distance N on road RD is 0.05 degrees in both the longitudinal (vertical) and lateral (horizontal) directions, the external identification device 50 controls the illumination direction of the light to shift at 0.05-degree intervals in both the longitudinal and lateral directions across the entire area within the FOV. That is, in Figure 6A In the grid, each black circle corresponds to an illumination point (detection point), and the intervals between the vertical and horizontal illumination points (detection points) correspond to an angular resolution of 0.05 degrees.

[0064] The actual number of irradiation points within the FOV is far greater than Figure 6A The number of black circles shown. For example, with the FOV of the lidar 5 being 120 degrees horizontally, 2400 black circles corresponding to the illumination point (detection point) are arranged at 0.05 degrees intervals horizontally. Similarly, with the FOV being 40 degrees vertically, 800 black circles corresponding to the illumination point (detection point) are arranged at 0.05 degrees intervals vertically.

[0065] Each time the FOV is scanned with one frame of illumination light, the external identification device 50 acquires and... Figure 6AThe detection data of the detection points corresponding to the illumination points are used to extract detection point data based on the angular resolution required for identifying the detection object. More specifically, for areas within the FOV where the depth distance X is shorter than the necessary depth distance N and the necessary angular resolution is 0.1 degrees instead of 0.05 degrees, data is extracted by widening the intervals of the longitudinal and lateral data beyond the 0.05-degree interval. Additionally, for areas within the FOV corresponding to the sky, since there is no road RD, data is extracted by widening the intervals of the longitudinal and lateral data. The intervals of the extracted detection points are as described later. Figure 6B The detection points, indicated by black circles, are spaced at the same intervals.

[0066] By extracting data from the detection points using the external identification device 50, the total number of detection data used in the identification process can be suppressed.

[0067] Figure 6B This is a schematic diagram illustrating an example of how the illumination light from the lidar 5 illuminates only the designated illumination points (detection points) arranged in a grid pattern within the field of view (FOV). When the external identification device 50 illuminates the lidar 5 with illumination light, it sets the interval between the illumination points (detection points) within the FOV to an interval corresponding to the required angular resolution, thereby controlling the direction of the illumination light.

[0068] For example, when the necessary angular resolution for identifying a detection object at a location with a necessary depth distance N on the road RD is 0.05 deg in both the longitudinal (vertical) and lateral (horizontal) directions, the external identification device 50 controls the illumination direction of the illumination light to shift at 0.05 deg intervals in both the longitudinal and lateral directions in the area corresponding to the necessary depth distance N (a strip-shaped area with a longer left-right direction).

[0069] Furthermore, for areas within the FOV where the depth distance X is shorter than the necessary depth distance N and a necessary angular resolution of 0.1 degrees is sufficient, the illumination direction is controlled by increasing the spacing between longitudinal and lateral detection points. For areas within the FOV corresponding to the sky, since there is no road RD, the illumination direction is also controlled by increasing the spacing between longitudinal and lateral detection points. As an example, at the beginning of illumination, the density distribution of illumination points (detection points) based on the flatness of the road surface or the state last measured is used to begin illumination.

[0070] By controlling the interval of the detection points through the external identification device 50 (in other words, controlling the interval (density) of the irradiation points during scanning), the total number of detection data used in the identification process can be suppressed.

[0071] It should be noted that the actual number of irradiation points within the FOV is far greater than...Figure 6B The diagram shows a large number of black circles.

[0072] Figure 7 It shows the... Figure 6B A diagram illustrating an example of the irradiation sequence when the irradiation point is illuminated by the irradiation light. Figure 7 In the diagram, the direction of the illumination light is controlled by the arrows, from the upper left to the lower right of the FOV. Additionally, the labels P1 to P3, written on the vertical arrows, indicate the spacing between the illumination points (detection points). P1 represents, for example, the spacing between illumination points (detection points) corresponding to an angular resolution of 0.05 degrees. P2 represents, for example, the spacing between illumination points (detection points) corresponding to an angular resolution of 0.1 degrees. P3 represents, for example, the spacing between illumination points (detection points) corresponding to an angular resolution of 0.2 degrees.

[0073] exist Figure 7 The diagram illustrates an example of switching the angular resolution to three levels, but it is not limited to three levels; it can also be configured to switch the angular resolution to two or more levels. For example, in addition to the intervals P1, P2, and P3 of the illumination points (detection points), an interval P4 corresponding to the angular resolution of 0.3 degrees can be added to switch to four levels.

[0074] <Structure of External Identification Device> Details of the external identification device 50 are explained.

[0075] As described above, the external identification device 50 includes an identification unit 111, a setting unit 112, a decision unit 113, a prediction unit 114, and a lidar 5.

[0076] <Identification Department> The recognition unit 111 uses the time-series detection data detected in the FOV of the lidar 5 to generate three-dimensional point cloud data.

[0077] Furthermore, the identification unit 111 identifies the road structure in the direction of travel of the road RD in which the vehicle 101 is traveling, and the objects to be detected on the road RD in the direction of travel, based on the detection data measured by the lidar 5. Road structure, for example, refers to straight roads, curves, intersections, tunnel entrances and exits, etc.

[0078] Furthermore, the recognition unit 111 can detect road markings, for example, by performing brightness filtering processing on data representing flat road surfaces. In this case, the recognition unit 111 can also determine that a road marking is present when the height of a road surface with brightness exceeding a predetermined threshold is approximately the same as the height of a road surface with brightness not exceeding the predetermined threshold.

[0079] <Identification of Road Structure> This describes an example of road structure recognition performed by the recognition unit 111. The recognition unit 111 identifies the curbs, walls, ditches, guardrails, or road markings of the road RD (road RD) in the generated point cloud data, representing the direction of travel (i.e., the road ahead), as the boundary lines RL and RB of the road RD. Figure 1A The road markings, indicated by boundary lines RL and RB, define the direction of travel. As described above, road markings include white lines (including lines of different colors), curb lines, road studs, etc., and these road marking markings define the driving lanes of road RD. In this embodiment, the boundary lines RL and RB of road RD defined by the above markings are referred to as road markings.

[0080] The identification unit 111 identifies the area enclosed by boundary lines RL and RB as the area corresponding to road RD. It should be noted that the method for identifying road RD is not limited to this, and other methods can also be used for identification.

[0081] Furthermore, the recognition unit 111 separates the generated point cloud data into point cloud data representing flat road surfaces and point cloud data representing three-dimensional objects, etc. For example, it identifies road surface shapes such as bumps, steps, and undulations with dimensions exceeding 15cm, and objects with dimensions exceeding 15cm in length and width, among the three-dimensional objects on the road surface in the direction of travel contained in the point cloud data, as detection objects. 15cm is an example of the size of the detection object, which can be appropriately changed.

[0082] <Settings Department> The setting unit 112 sets the vertical projection angle of the illumination light of the lidar 5. φ With the FOV of lidar 5 at 40 degrees in the vertical direction, the vertical projection angle is set in 0.05-degree intervals within the range of 0 to 40 degrees. φ Similarly, the setting unit 112 sets the horizontal projection angle of the illumination light for the lidar 5. θ With the FOV of LiDAR 5 at 120 degrees in the horizontal direction, the horizontal projection angle is set in 0.05-degree intervals within the range of 0 to 120 degrees. θ .

[0083] As described below, the setting unit 112 sets the number of illumination points within the field of view (FOV) of the lidar 5 based on the angular resolution determined by the determination unit 113. Figure 6A and Figure 6B The number of black circles corresponds to the density of irradiation points. As mentioned above, the vertical and horizontal spacing of the irradiation points (detection points) arranged in a grid pattern within the FOV corresponds to the angular resolution in the vertical and horizontal directions, respectively.

[0084] <Decision Department> The determination unit 113 determines the scanning angle resolution set by the setting unit 112. First, the determination unit 113 calculates the vertical projection angle α of each depth distance X and the distance DL to the road surface point at each depth distance X. Specifically, as shown in reference... Figure 3D As explained, the depth distance X is calculated based on the distance DL to the road surface point measured by the lidar 5 and the projection angle α set for the lidar 5 during measurement. The determination unit 113 calculates the relationship between the calculated depth distance X and the vertical angle. Figure 5A Additionally, the decision unit 113 calculates the relationship between the depth distance X and the distance DL. Furthermore, as... Figure 5B As illustrated, the decision unit 113 calculates the relationship between the depth distance X and the vertical angular resolution based on the size of the object being inspected and the depth distance X. Thus, the vertical angular resolution is calculated based on the size of the object being inspected and the distance DL, and the relationship between the depth distance X and the vertical angular resolution is calculated based on the distance DL and the depth distance X.

[0085] Next, the decision unit 113 determines the angular resolution in the vertical direction required to identify the detection object of the aforementioned size. For example, in Figure 5B In this study, for depth distances X with a vertical angular resolution less than 0.1 degrees, 0.05 degrees (less than 0.1 degrees) is determined as the necessary angular resolution. Similarly, for depth distances X with a vertical angular resolution of 0.1 degrees or more but less than 0.2 degrees, 0.1 degrees (less than 0.2 degrees) is determined as the necessary angular resolution. Likewise, for depth distances X with a vertical angular resolution of 0.2 degrees or more but less than 0.3 degrees, and for depth distances X with a vertical angular resolution of 0.3 degrees or more but less than 0.4 degrees, even smaller values ​​of 0.2 degrees and 0.3 degrees are determined as the necessary angular resolutions, respectively.

[0086] The necessary angular resolution of the determined vertical direction can reflect the interval of the vertical direction of the detection points when acquiring the 3D point cloud data of the next frame.

[0087] Furthermore, the decision unit 113 can also determine the necessary angular resolution in the horizontal direction for identifying the detected object based on the size and depth distance X of the detected object. The necessary angular resolution in the horizontal direction can also reflect the horizontal spacing of the detection points when acquiring the three-dimensional point cloud data of the next frame.

[0088] It should be noted that the necessary angular resolution in the horizontal direction can also be made consistent with the previously determined necessary angular resolution in the vertical direction. In other words, on the same horizontal line as the detection point where the necessary angular resolution in the vertical direction is determined to be 0.05 degrees, the necessary angular resolution in the horizontal direction is also determined to be 0.05 degrees. Similarly, on the same horizontal line as the detection point where the necessary angular resolution in the vertical direction is determined to be 0.1 degrees, the necessary angular resolution in the horizontal direction is also determined to be 0.1 degrees. Furthermore, regarding other necessary angular resolutions, on the same horizontal line as the detection point where the necessary angular resolution in the vertical direction is determined, the necessary angular resolution in the horizontal direction is determined to be the same value as the necessary angular resolution in the vertical direction.

[0089] <Forecasting Department> For example, when the road surface of road RD, where vehicle 101 is traveling, is flooded by rainwater, the lidar 5 may sometimes fail to receive scattered light up to the necessary depth distance N. The furthest depth distance X that the lidar 5 can detect in this situation is called the maximum depth distance L. The maximum depth distance can also be called the maximum road surface detection distance. It should be noted that when the road surface of road RD is downhill in the direction of travel, and when the vehicle 101 is traveling at a high speed and the necessary depth distance N is long, the lidar 5 may also sometimes fail to receive scattered light up to the necessary depth distance N.

[0090] When the required depth distance N calculated based on the vehicle speed of the vehicle 101 exceeds the maximum depth distance L (for example, when the required depth distance N is 115m and the maximum depth distance L is 80m), the prediction unit 114 uses the road slope map described above to predict the height Z (road slope) of the road surface from the maximum depth distance L to the required depth distance N.

[0091] Regarding an example of prediction, see [reference]. Figure 8A and Figure 8B Please provide an explanation. Figure 8A and Figure 8B This is a two-dimensional graph illustrating the relationship between the depth distance Xr and the road surface height Z along the driving route. The horizontal axis represents the depth distance Xr (in meters), and the vertical axis represents the road surface height Z (in meters). The scale of the horizontal axis is based on the current position of the vehicle 101, with negative values ​​indicating that it is closer to the vehicle 101 than the current position, and positive values ​​indicating that it is closer to the depth side than the current position.

[0092] The prediction unit 114 aligns the road surface height Z measured by the lidar 5 with the road surface slope map associated with the aforementioned map information. More specifically, for point cloud data represented based on the position of the vehicle 101 obtained using the positioning unit 2, in a two-dimensional graph of depth distance Xr and height Z along the driving route, the road surface slope map is relatively offset forward and backward on the Xr axis, thereby adjusting the position on the Xr axis in a manner that minimizes the deviation ΔZ between the measured result of the road surface height Z (measurement data based on point cloud data) and the road surface height Z based on the road surface slope map.

[0093] exist Figure 8A In the measurement data representing the road surface height Z measured by the LiDAR 5, the range from tens of meters negative to tens of meters positive (e.g., 10m) on the side of the vehicle 101 (represented by solid lines) represents the road surface height Z obtained based on point cloud data of past frames measured by the LiDAR 5 in a time sequence. Additionally, the range from tens of meters positive (e.g., 10m) of travel distance shown by the solid lines to the maximum depth distance L (represented by double lines) represents the road surface height Z obtained based on the point cloud data of the current frame (the newly acquired frame). Furthermore, the data represented by dashed lines represents the road surface height Z obtained from the road slope map.

[0094] Generally, when the current position of the vehicle 101 obtained by the positioning unit 2 contains errors, the measurement data on the driving route and the position of the road slope map are inconsistent. Therefore, the measured result of the road height Z (represented by solid or double lines) is inconsistent with the road height Z (represented by dashed lines) obtained from the road slope map, resulting in a deviation ΔZ in the height direction.

[0095] As an example of suppressing the deviation ΔZ, the prediction unit 114 shifts the road slope map data (represented by dashed lines) on the Xr axis in a manner consistent with the data (represented by double lines) in a specified range (e.g., 5m to 10m) of the road height Z (measurement data) obtained based on the point cloud data of the current frame, while searching for the position where the least sum of squares based on the magnitude of the deviation ΔZ is minimized, and shifts the road slope map relative to that position.

[0096] like Figure 8B As shown, after offsetting the position of the road slope map corresponding to the current position of the vehicle 101 on the Xr axis, the measured result of the road height Z (represented by solid or double lines) is consistent with the road height Z (represented by dashed lines) obtained using the road slope map, and the deviation ΔZ in the height direction is suppressed to below the specified value.

[0097] As explained above, by making Figure 8AThe road slope map data is offset in the front-rear direction on the depth distance Xr axis of the driving route, so that the position of the road slope map is consistent with the position of the vehicle 101. Figure 8B ).

[0098] <Generating Location Data> The external identification device 50 can map the data representing the position of the detected object, which is obtained from the time-series point cloud data measured in real time by the lidar 5, onto a two-dimensional map, such as an XY map, to generate a series of position data. In the XY space, the information representing the height Z is omitted, leaving the information of the depth distance X and the horizontal distance Y.

[0099] The recognition unit 111 acquires the position information of three-dimensional objects, etc., stored on the two-dimensional map in the storage unit 12, and calculates the relative position of the three-dimensional objects, etc., based on the moving speed and moving direction (e.g., azimuth angle) of the vehicle 101 by performing coordinate transformation with the position of the vehicle 101 as the center. Whenever point cloud data is acquired by measuring the LiDAR 5, the recognition unit 111 performs coordinate transformation based on the relative position of the three-dimensional objects, etc., based on the acquired point cloud data with the position of the vehicle 101 as the center, and records it on the two-dimensional map.

[0100] <Explanation of the flowchart> Figure 9 It is shown Figure 2 A flowchart illustrating an example of processing performed by the arithmetic unit 11 of the controller 10 according to a predetermined program. Figure 9 The process shown in the flowchart is repeated at a predetermined cycle, for example, during the operation of this vehicle 101 in autonomous driving mode.

[0101] First, in step S10, the computing unit 11 enables the lidar 5 to acquire three-dimensional point cloud data, and then proceeds to step S20.

[0102] In step S20, the arithmetic unit 11 calculates the road surface slope and maximum depth distance L in the direction of travel of road RD based on the point cloud data acquired by the lidar 5, and proceeds to step S30. The following will refer to... Figure 10 The process of step S20 is explained in detail.

[0103] In step S30, the prediction unit 114 of the calculation unit 11 determines whether the maximum depth distance L is shorter than the necessary depth distance N. If the maximum depth distance L is shorter than the necessary depth distance N, the calculation unit 11 determines that step S30 is affirmative (S30: Yes) and proceeds to step S40. If the maximum depth distance L is longer than the necessary depth distance N, the calculation unit 11 determines that step S30 is negative (S30: No) and proceeds to step S50.

[0104] In step S40, the prediction unit 114 of the calculation unit 11 predicts the road surface slope from the maximum depth distance L to the necessary depth distance N, and proceeds to step S50. An example of the predicted road surface slope is... Figure 8B As shown.

[0105] In step S50, the calculation unit 11 calculates the vertical projection angle α and the distance DL to the road surface point for each depth distance X, and proceeds to step S60. The relationship between the vertical angle and the depth distance X is as follows: Figure 5A As shown. Furthermore, the relationship between the depth distance X, the road surface height Z, and the distance DL from the road surface is as follows: Figure 4B As shown.

[0106] In step S60, the arithmetic unit 11 calculates the necessary angular resolution for each depth distance X, and proceeds to step S70. The necessary angular resolution is the angular resolution required to detect an object of a pre-specified size. The relationship between depth distance X and angular resolution is as follows: Figure 5B As shown.

[0107] In step S70, the calculation unit 11 determines the vertical angular resolution as the necessary angular resolution through the decision unit 113, and proceeds to step S80. In this embodiment, the vertical angular resolution is determined more preferentially than the horizontal angular resolution.

[0108] In step S80, the decision unit 113 of the calculation unit 11 determines the horizontal angular resolution as the necessary angular resolution and proceeds to step S90. By determining the horizontal angular resolution after determining the vertical angular resolution, it is easy to make the horizontal angular resolution consistent with the vertical angular resolution.

[0109] In step S90, the arithmetic unit 11 determines the coordinates of the detection point. More specifically, it determines the representation as follows: Figure 6B The coordinates of the detection points are illustrated by black circles. The recognition unit 111 identifies three-dimensional objects, etc., in the direction of travel of the road RD on which the vehicle 101 is traveling, based on the detection data detected by the detection points determined in step S90.

[0110] It should be noted that whenever point cloud data is acquired in step S10, the computation unit 11 generates a series of two-dimensional position data by mapping the relative positions of the three-dimensional objects, etc., based on the point cloud data onto a two-dimensional XY map. Then, the relative positions of the three-dimensional objects, etc., based on the point cloud data can be transformed into coordinates with the position of the vehicle 101 as the center and recorded on the two-dimensional map.

[0111] In step S100, the arithmetic unit 11 determines whether to end the process. If the vehicle 101 continues to drive in autonomous driving mode, the arithmetic unit 11 determines that step S100 is negative (S100: No), returns to step S10, and repeats the above process. By returning to step S10, the measurement of three-dimensional objects based on point cloud data is periodically repeated during the driving of the vehicle 101. On the other hand, if the vehicle 101 ends driving in autonomous driving mode, the arithmetic unit 11 determines that step S100 is positive (S100: Yes), and ends the process. Figure 9 The processing.

[0112] Figure 10 This is a detailed explanation of step S20 performed by the arithmetic unit 11. Figure 9 The flowchart shows the processing of the point cloud data of the detection points determined by the decision unit 113. Figure 10 The processing.

[0113] In step S210, the arithmetic unit 11 performs separation processing on the point cloud data, proceeding to step S220. More specifically, by detecting and separating data on three-dimensional objects such as those on the road RD from the point cloud data, point cloud data representing a flat road surface and point cloud data representing three-dimensional objects such as those on the road RD are obtained. Three-dimensional objects such as those on the road include obstacles, curbs, walls, ditches, guardrails, etc., located at the left and right ends of the road RD, as well as other vehicles such as two-wheeled vehicles in motion.

[0114] Here is an example of the separate processing. The calculation unit 11 transforms the relative position coordinates of the point cloud data into a position centered on the position of the vehicle 101, and represents it on a two-dimensional map (XY) corresponding to the depth direction and road width direction, such as an aerial view of the road RD, and then grids the two-dimensional map to a predetermined size. If the difference between the maximum and minimum values ​​of the data within each grid is less than a predetermined threshold, the calculation unit 11 determines that the data of that grid represents a flat road surface. On the other hand, if the difference between the maximum and minimum values ​​of the data within a grid is greater than a predetermined threshold, the calculation unit 11 determines that the data of that grid represents a three-dimensional object, etc.

[0115] It should be noted that other methods can also be used to determine whether point cloud data corresponds to road surface data or to three-dimensional objects.

[0116] In step S220, the arithmetic unit 11 determines whether the data to be processed is road surface data. If the data is grid data separated from road surface data, the arithmetic unit 11 determines that step S220 is affirmative (S220: Yes) and proceeds to step S230. On the other hand, if the data is grid data separated from three-dimensional objects or the like, the arithmetic unit 11 determines that step S220 is negative (S220: No) and proceeds to step S250.

[0117] Upon entering step S250, the recognition unit 111 of the arithmetic unit 11 performs coordinate transformation on the relative positions of three-dimensional objects, etc., based on the point cloud data of the grid, with the position of the vehicle 101 as the center, and records them on a two-dimensional map. Then, the process ends. Figure 10 The processing, enter Figure 9 Step S30.

[0118] Upon proceeding to step S230, the prediction unit 114 of the calculation unit 11 calculates the road surface slope of road RD. An example of the road surface slope calculation process is referred to... Figures 3A-3D As explained.

[0119] It should be noted that other methods can also be used to calculate road surface slope.

[0120] In step S240, the prediction unit 114 of the calculation unit 11 obtains the maximum depth distance L and ends. Figure 10 The processing, enter Figure 9 Step S30.

[0121] As described above, the maximum depth distance L is the farthest depth distance that can be detected by the lidar 5. The prediction unit 114 of the calculation unit 11 obtains the depth distance corresponding to the data of the grid farthest from the position of the vehicle 101 in the grid extracted during the calculation of road slope, and uses it as the maximum depth distance L.

[0122] The implementation method described above achieves the following effects.

[0123] (1) The external identification device 50 includes: a lidar 5, which acts as a vehicle-mounted detector, and scans and irradiates electromagnetic waves in a horizontal direction (a first direction) and a vertical direction (a second direction intersecting the first direction) within the field of view (FOV), acquiring point cloud data frame by frame, including three-dimensional position information of detection points on the surface of objects based on reflected waves from objects around the vehicle 101; an identification unit 111, which identifies the road surface of the road RD on which the vehicle 101 travels and three-dimensional objects on the road as road surface information based on the point cloud data of each frame; and a determination unit 113, which determines the detection required for the next frame of point cloud data based on the size of a predetermined three-dimensional object as the object of identification and the measurement distance from the vehicle 101 to the three-dimensional object based on the point cloud data. The interval between points; and the prediction unit 114, which is a slope prediction unit, predicts the slope of the road surface not identified by the identification unit 111 when the maximum depth distance L, which is the farthest distance of the road surface in the direction of travel of the vehicle 101 identified by the identification unit 111, is shorter than the necessary depth distance N, which is the necessary distance based on the speed of the vehicle 101. The determination unit 113, while the slope is predicted by the prediction unit 114, determines the interval between the detection points required for the next frame of point cloud data for the range from the maximum depth distance L to the necessary depth distance N, based on the size of the three-dimensional object and the estimated distance from the vehicle 101 to the three-dimensional object estimated based on the map information and the slope, according to the size of the three-dimensional object and the estimated distance from the vehicle 101 to the three-dimensional object estimated based on the map information and the slope.

[0124] Generally, the shorter the depth distance X, the larger the viewing angle for the object to be identified, so the object can be identified even with a lower angular resolution. Conversely, the longer the depth distance X, the smaller the viewing angle for the object to be identified, thus requiring a higher angular resolution for identification. In this embodiment, for each detection point, the LiDAR 5 acquires the depth distance X up to the road surface RD in the direction of travel, and the determination unit 113 determines the interval of detection points required for the identification unit 111 to identify the three-dimensional object at that depth distance X. Furthermore, for road surfaces where the LiDAR 5 cannot acquire the depth distance X, the prediction unit 114 predicts the road surface based on a road slope map, and the determination unit 113 determines the interval of detection points required for the identification unit 111 to identify the three-dimensional object at the predicted depth distance X on the road surface.

[0125] Because of this configuration, the determination unit 113 can suppress the total number of detection data used in the recognition process by appropriately controlling the interval of the detection points of the three-dimensional point cloud data used by the recognition unit 111 in the recognition process. In other words, the processing load of the computing unit 11 can be reduced without reducing the accuracy of the recognition of the position and size of objects such as those being detected by the external recognition device 50.

[0126] Furthermore, in the implementation, when the vehicle 101 is traveling on a road RD that is not recorded in the high-precision map information, a road RD that is being traveled for the first time without high-precision map information, or a road RD that has changed to a different form from the high-precision map information due to construction, etc., it is also possible to use the lidar 5 to obtain the depth distance X up to the road surface of the road RD in the direction of travel for each detection point, and determine the interval of the detection points required for the recognition unit 111 to recognize the three-dimensional object at each depth distance X.

[0127] (2) The external identification device 50 also includes: a positioning unit 2, which acts as a position detection unit to detect the position of the vehicle 101 based on information from the satellite; and a prediction unit 114, which acts as an adjustment unit to offset the position of the vehicle 101 detected by the positioning unit 2 relative to the position of the map information and the slope information.

[0128] Because of this configuration, it is assumed that there is an error in the current position of the vehicle 101 obtained by the positioning unit 2. As a result, the positions of the measurement data of the lidar 5 (data represented based on the position of the vehicle 101 detected by the positioning unit 2) and the data of the road slope map on the Xr axis are inconsistent. Figure 8A Under these conditions, it is also possible to correct an amount equivalent to the error on the Xr axis. Figure 8B Therefore, compared to the case where the prediction unit 114 does not have the function of an adjustment unit, it is possible to predict the slope of the road surface that could not be identified by the identification unit 111 with high accuracy.

[0129] (3) In the external identification device 50, the slope information is data based on the height Z of the road surface of the road RD measured by the lidar 5 of this vehicle 101 and / or other vehicles. It is data recorded by establishing a correspondence between the location information of the road and the measured height Z of the road surface for each specified interval. The prediction unit 114, as the adjustment unit, searches for an interval of a specified length (a length corresponding to the size of the grid) from the roads where the slope information records the height Z of the road surface. The difference between the interval of the specified length (a length corresponding to the size of the grid) and the height Z of the road surface of the comparison object interval is less than or equal to a specified value. The specified length (e.g., 10m) includes the maximum depth distance L of the road RD based on the point cloud data, and the position of the slope information is offset in a way that makes the searched interval overlap with the comparison object interval. More specifically, the position information of the slope information is updated. In the above search, the prediction unit 114, as the adjustment unit, searches for the interval of the specified length where the least square sum of the height Z of the road surface of each specified interval contained in the comparison object interval, represented by the point cloud data, and the height Z of the road surface of each specified interval based on the slope information is the smallest.

[0130] Because of this configuration, the measured height Z of the road surface is shifted relative to the position of the road slope map. Figure 8B (represented by solid or double lines) and the road surface height Z obtained from the road slope map (in... Figure 8B (In the image, indicated by a dashed line) The deviation ΔZ in the height direction is suppressed to below a specified value. Therefore, it is possible to predict the slope of the road surface that was not identified by the identification unit 111 with high accuracy.

[0131] (4) In the external identification device 50, the decision unit 113 also determines the interval of the detection points required for the point cloud data of the next frame as the scanning angle resolution of the illumination light. The further the scanning illumination destination of the illumination light in the field of view (FOV) of the lidar 5 is from the necessary depth distance N, the coarser the scanning angle resolution is set.

[0132] Because of this configuration, it is possible to ensure higher recognition accuracy in areas corresponding to the necessary depth distance N, while reducing recognition accuracy in areas further up in the sky, thereby suppressing the total amount of detection data used by the recognition unit 111 in the recognition process. In other words, without reducing the recognition accuracy of the position and size of objects, etc., which are the objects of recognition for the external recognition device 50, in the vertical direction, the processing load of the computing unit 11 can be reduced.

[0133] (5) In the external identification device 50, the closer the target of the illumination light in the field of view (FOV) of the lidar 5 is to the necessary depth distance N, the coarser the scanning angle resolution is set by the decision unit 113.

[0134] Because of this configuration, it is possible to prevent setting an excessive number of detection points beyond the necessary number for objects to be identified near the vehicle 101. That is, without reducing the accuracy of identifying the position and size of objects such as those being identified by the external identification device 50 in the horizontal direction, the processing load of the arithmetic unit 11 can be reduced.

[0135] The above-described embodiments can be modified in various ways. The following describes some modifications.

[0136] In the above embodiment, an example was described in which the external identification device 50 uses the lidar 5 to detect the road conditions in the direction of travel of the vehicle 101. Alternatively, for example, it may be configured to have a lidar 5 capable of detecting the field of view (FOV) around the vehicle 101 in 360 degrees, so that the lidar 5 can detect the road conditions around the entire area of ​​the vehicle 101.

[0137] The above description is merely an example, and the present invention is not limited to the above embodiments and modifications as long as it does not impair the characteristics of the invention. The above embodiments and modifications can also be combined arbitrarily.

[0138] Using this invention, the processing load for identifying the external conditions around a vehicle can be reduced.

[0139] The present invention has been described above in conjunction with preferred embodiments. Those skilled in the art should understand that various modifications and changes can be made without departing from the scope of the claims.

Claims

1. An external environment recognition device characterized by comprising: Possessing: an on-vehicle detector (5) that scans irradiation electromagnetic waves in a first direction within a field of view and a second direction intersecting the first direction, and acquires point cloud data including three-dimensional position information of detection points of an object surface based on reflected waves from an object in the surroundings of a host vehicle, frame by frame; an identification section (111) that identifies a road surface of a road on which the host vehicle travels and a three-dimensional object on the road as road surface information, from the point cloud data of each frame; a determination section (113) that determines an interval of the detection points required for the point cloud data of a next frame, from a size of a prescribed three-dimensional object determined in advance as an identification target, and a measured distance to the three-dimensional object based on the point cloud data; and a slope prediction section (114) that predicts a slope of the road surface in a direction of travel of the host vehicle that is not identified by the identification section (111), from slope information associated with map information that includes the road, in a case where a farthest distance of the road surface in the direction of travel of the host vehicle identified by the identification section (111) is shorter than a necessary distance based on a vehicle speed of the host vehicle, when the slope is predicted by the slope prediction section (114), the determination section (113) determines the interval of the detection points required for the point cloud data of the next frame, from the size of the prescribed three-dimensional object, and an estimated distance to the three-dimensional object estimated from the map information and the slope, for a range from the farthest distance to the necessary distance.

2. The outside recognition device according to claim 1, characterized by Further possessing: a position detection section (2) that detects a position of the host vehicle from information from a satellite; and an adjustment section that relatively shifts the position of the host vehicle detected by the position detection section (2) and a position of the slope information.

3. The outside world recognition device according to claim 2, wherein the slope information is data produced from heights of road surfaces of the road measured by the on-vehicle detector (5) of the host vehicle and / or other vehicles, and is data in which position information of the road and the measured heights of the road surfaces are associated and recorded, the adjustment section searches for an interval of a prescribed length from the road in which the heights of the road surfaces are recorded in the slope information, the interval of the prescribed length having a difference in height of the road surface from a comparison target interval that is a prescribed value or less, the prescribed length including the farthest distance of the road based on the point cloud data, and shifts the position of the slope information in a manner that overlaps the searched interval and the comparison target interval.

4. The outside world recognition device according to claim 3, wherein the slope information is data in which the position information of the road and the heights of the road surfaces of the road are associated and recorded for each prescribed interval, the adjustment section searches for the interval of the prescribed length, the interval of the prescribed length having a least sum of squares of the heights of the road surface of each prescribed interval included in the comparison target interval represented by the point cloud data and the heights of the road surface of each prescribed interval based on the slope information being the least.

5. The outside world recognition apparatus according to claim 1, characterized in that the decision unit (113) further decides the interval of the detection points required for the point cloud data of the next frame as a scanning angle resolution of the electromagnetic wave, and the scanning irradiation destination of the electromagnetic wave in the field of view of the vehicle-mounted detector (5) is farther than the necessary distance, the scanning angle resolution is set to be coarser.

6. The outside world recognition apparatus according to claim 5, characterized in that the scanning irradiation destination of the electromagnetic wave in the field of view of the vehicle-mounted detector (5) is closer than the necessary distance, the decision unit (113) sets the scanning angle resolution to be coarser.

7. The outside world recognition apparatus according to any one of claims 1 to 6, characterized in that the vehicle-mounted detector (5) is a laser radar.

Images