In-vehicle processing device, vehicle control device, and self-position estimation method

Abstract

To highly accurately estimate the self-position by selecting an optimum point group while considering a property about observation accuracy of the point group.SOLUTION: An on-vehicle processing device comprises: a past map storage unit which stores a past-generated map generated in the past; a self-map generation unit which generates an in-travel map on the basis of sensor data acquired during the travel; a self-position estimation unit which estimates a self-position by collating the past-generated map with the in-travel map; and a data selection unit which selects data used by the self-position estimation unit from the in-travel map on the basis of the property of the data included in the in-travel map.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to an in-vehicle processing device that generates a map from the observation results of sensors and estimates its own position.

Background Art

[0002] In the application of an automatic driving system or a driving support system, detailed map information for estimating one's own position is important. However, on highways, detailed maps for driving support systems are being developed, but the development of detailed maps in residential areas such as general roads and around one's home is not progressing. In order to expand the application range of an automatic driving system or a driving support system, a technology that can generate a map by itself and estimate its own position is required.

[0003] As the background art in this technical field, there is the following prior art. Patent Document 1 (Japanese Patent Application Laid-Open No. 2019-144041) includes a first acquisition unit that acquires an external situation signal indicating the situation outside the moving body, a second acquisition unit that acquires map data including information indicating characteristics affected by the weather, a third acquisition unit that acquires weather information indicating the state of the weather, and a self-position estimation unit that estimates the self-position of the moving body based on the external situation signal, and the self-position estimation unit evaluates the accuracy of the external situation signal based on the map data and the weather information. A self-position estimation device characterized by this is described.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In the invention described in Patent Document 1, by estimating the self-position while considering the reliability of the observation point group based on weather information, high-precision self-position estimation is made possible. However, since the properties related to the observation accuracy of the camera point group and the occurrence of local adaptation during position estimation using the camera point group are not considered, there are problems such as incorrect self-position estimation and low self-position estimation accuracy. If the self-position estimation accuracy is low, it is difficult to utilize it for autonomous driving or driving support.

[0006] Therefore, an object of the present invention is to select an optimal point group while considering the properties related to the observation accuracy of the point group, and to estimate the self-position with high precision.

Means for Solving the Problems

[0007] A typical example of the invention disclosed in the present application is as follows. That is, based on a CPU, a past generated map generated in the past, and sensor data acquired by a sensor during the running of the host vehicle , including the first point cloud data A recording unit that records a running map, and The first point cloud data includes road surface point cloud data representing landmarks on the road surface and high point cloud data which is point cloud data at a position higher than the road surface. The CPU collates the past point group data included in the past generated map with The road surface point cloud data, and the past point cloud data and the high point cloud data and Execute a first collation, the Based on the result of the first collation, a first self-position estimation for estimating the self-position of the host vehicle is executed. When it is determined that the first self-position estimation is successful, The past point cloud data and from the first point group data The road surface selected as the second point cloud data Based on the result of a second collation that collates with point group data, the in-vehicle processing device is characterized in that it estimates the self-position of the host vehicle.

Effects of the Invention

[0008] According to one aspect of the present invention, the self-position can be estimated with high precision. Problems, configurations, and effects other than those described above will be clarified by the description of the following embodiments.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Mode for Carrying Out the Invention

[0010] Hereinafter, with reference to FIGS. 1 to 9, an embodiment of a map generation / self-position estimation device 100 according to the present invention will be described.

Embodiment

[0011] FIG. 1 is a block diagram showing the configuration of a map generation / self-position estimation device 100 according to Embodiment 1 of the present invention.

[0012] The map generation / self-position estimation device 100 is mounted on a vehicle 105 and includes cameras 121 to 124, a stereo camera 125, a sonar 126, an external environment sensor 127, a vehicle speed sensor 131, a steering angle sensor 132, an interface 180, and an in-vehicle processing device 101. The cameras 121 to 124, the stereo camera 125, the sonar 126, the external environment sensor 127, the vehicle speed sensor 131, the steering angle sensor 132, and the interface 180 are connected to the in-vehicle processing device 101 by signal lines and exchange various data with the in-vehicle processing device 101.

[0013] The in-vehicle processing device 101 includes a CPU 110, a ROM 111, a RAM 112, and a recording unit 113. It may be configured to execute all or part of the arithmetic processing using another arithmetic processing device such as an FPGA.

[0014] The CPU 110 is an arithmetic device that reads and executes various programs and parameters from the ROM 111, and operates as an execution unit of the in-vehicle processing device 101.

[0015] The RAM 112 is a readable and writable storage area, and operates as a main storage device of the in-vehicle processing device 101. The self-generated map 161 and the driving status 162, which will be described later, are stored in the RAM 112.

[0016] The ROM 111 is a read-only storage area, and stores a program that will be described later. This program is expanded in the RAM 112 and executed by the CPU 110. By reading and executing the program, the CPU 110 operates as an odometry estimation unit 141, a landmark detection unit 142, a point cloud generation unit 143, a position estimation determination unit 144, a driving status diagnosis unit 145, an integration determination unit 146, an integration diagnosis unit 140, a data selection unit 147, a self-map generation unit 148, and a self-position estimation unit 149. Also, the ROM 111 stores sensor parameters 150. The sensor parameters 150 include the positional and postural relationship with the vehicle 105 and information specific to each sensor for each of the cameras 121 to 124, the stereo camera 125, the sonar 126, and the external environment sensor 127. For example, the sensor parameters 150 of the cameras 121 to 124 and the stereo camera 125 include lens distortion coefficients, optical axis centers, focal lengths, the number of pixels of the imaging element, and dimensions, and the sensor parameters 150 of the external environment sensor 127 include information specific to each sensor.

[0017] The recording unit 113 is a non-volatile storage device, and operates as an auxiliary storage device of the in-vehicle processing device 101. The self-generated map 151 and the driving status 152 are stored in the recording unit 113.

[0018] Cameras 121 to 124 are attached around the vehicle 105, for example, and photograph the surroundings of the vehicle 105. The positional and postural relationship between the cameras 121 to 124 and the vehicle 105 is stored in the ROM 111 as sensor parameters 150. The attachment method of the cameras 121 to 124 is not limited to the method described above, and other attachment methods may be used. Also, the number of cameras 121 to 124 does not necessarily have to be four.

[0019] The stereo camera 125 is attached, for example, inside the windshield in the passenger compartment of the vehicle 105, and photographs the front of the vehicle 105. The positional and postural relationship between the stereo camera 125 and the vehicle 105 is stored in the ROM 111 as sensor parameters 150. The attachment method of the stereo camera 125 is not limited to the method described above, and other attachment methods may be used. The stereo camera 125 may be one in which two cameras are arranged side by side with a predetermined baseline length and integrated into one unit, or one in which two cameras are roughly attached and can be utilized as a stereo camera through calibration, or one in which two cameras are arranged vertically, or one in which three or more cameras are arranged side by side. In this embodiment, a stereo camera in which two cameras are arranged side by side with a predetermined baseline length and integrated into one unit will be described by way of assumption. The stereo camera 125 outputs the captured image and the parallax necessary for distance calculation to the in-vehicle processing device 101. The stereo camera 125 may output only the captured image to the in-vehicle processing device 101, and the CPU 110 may calculate the parallax.

[0020] The sonar 126 is mounted on the vehicle 105 in plural, for example, and observes the surroundings of the vehicle 105. The positional and postural relationship between the sonar 126 and the vehicle 105 is stored in the ROM 111 as sensor parameters 150.

[0021] The external environment sensor 127 is, for example, a LiDAR mounted on the vehicle 105, and observes the surroundings of the vehicle 105. The positional and postural relationship between the external environment sensor 127 and the vehicle 105 is stored in the ROM 111 as sensor parameters 150.

[0022] Cameras 121 to 124 and stereo camera 125 have lenses and imaging elements. Sensor parameters 150 are parameters indicating the characteristics of these cameras 121 to 124, 125, such as lens distortion coefficients, which are parameters indicating lens distortion, optical axis centers, focal lengths, internal parameters such as the number of pixels and dimensions of the imaging elements, position and orientation relationships indicating the mounting state of the sensors to the vehicle 105, relative relationships between the two cameras of the stereo camera 125, and information specific to each of the sonar 126 and other external sensors 127, and are stored in the ROM 111. The position and orientation relationships between each of the cameras 121 to 124, stereo camera 125, sonar 126, and other external sensors 127 and the vehicle 105 may be estimated by the CPU 110 in the in-vehicle processing device 101 using captured images, parallax, vehicle speed sensor 131, and steering angle sensor 132.

[0023] Cameras 121 to 124, stereo camera 125, sonar 126, and other external sensors 127 may be mounted in various numbers and combinations as long as at least one of cameras 121 to 124 and stereo camera 125 is included.

[0024] Each of the vehicle speed sensor 131 and the steering angle sensor 132 measures the vehicle speed and steering angle of the vehicle 105 on which the in-vehicle processing device 101 is mounted and outputs them to the in-vehicle processing device 101. The in-vehicle processing device 101 calculates the movement amount and movement direction of the vehicle 105 on which the in-vehicle processing device 101 is mounted by a known dead reckoning technique using the outputs of the vehicle speed sensor 131 and the steering angle sensor 132.

[0025] The interface 180 provides, for example, a GUI that accepts instruction inputs from the user. Also, other information may be input and output in other forms.

[0026] FIG. 2 is a functional block diagram showing the operation of the map generation / self-position estimation device 100, and shows the functional blocks executed by the CPU 110 and the data flow between the RAM 112, the ROM 111, and the recording unit 113. In FIG. 2, as functional blocks, the functions of the odometry estimation unit 141, the landmark detection unit 142, the point cloud generation unit 143, the integrated diagnosis unit 140, the data selection unit 147, the self-map generation unit 148, and the self-position estimation unit 149 are shown.

[0027] The self-map generation / self-position estimation unit 250 includes a sensor value acquisition unit 201, an odometry estimation unit 141, a landmark detection unit 142, a point cloud generation unit 143, an integrated diagnosis unit 140, a data selection unit 147, a self-map generation unit 148, a self-position estimation unit 149, and a recording unit 113. The integrated diagnosis unit 140 includes a position estimation determination unit 144, a driving state diagnosis unit 145, and an integrated determination unit 146.

[0028] Among the self-map generation / self-position estimation unit 250, the combination of the sensor value acquisition unit 201, the odometry estimation unit 141, the landmark detection unit 142, the point cloud generation unit 143, the self-map generation unit 148, and the recording unit 113 operates in the map generation mode 202, and the combination of the sensor value acquisition unit 201, the odometry estimation unit 141, the landmark detection unit 142, the point cloud generation unit 143, the integrated diagnosis unit 140, the data selection unit 147, the self-map generation unit 148, the self-position estimation unit 149, and the recording unit 113 operates in the self-position estimation mode 203.

[0029] In the map generation mode 202, a point cloud map is generated from the own observation results. In the map generation mode 202, the sensor value acquisition unit 201, the landmark detection unit 142, the point cloud generation unit 143, the odometry estimation unit 141, the self-map generation unit 148, and the recording unit 113 synthesize the sensor information observed at each time series while considering the vehicle movement, and self-generate a point cloud map. The details of each functional block will be described later. In the map generation mode 202, after the point cloud map is recorded in the recording unit 113 as the self-generated map 151, the functional blocks related to the subsequent self-position estimation are not executed.

[0030] In the self-position estimation mode 203, the self-generated map 151 is read from the recording unit 113, and the self-position is estimated on the self-generated map 151. In the self-position estimation unit 149, similar to the map recorded in the recording unit 113 in the past and the map generation mode 202, the sensor value acquisition unit 201, the landmark detection unit 142, the point cloud generation unit 143, the odometry estimation unit 141, the self-map generation unit 148, and the recording unit 113 are used to synthesize the sensor information observed at each time series while considering the vehicle movement, and the self-generated point cloud map is collated to estimate the self-position. The details of each functional block will be described later.

[0031] The sensor value acquisition unit 201 acquires the signals output from the sensors. Images are acquired from the cameras 121 to 124, and images and disparities are acquired from the stereo camera 125. The cameras 121 to 124 and the stereo camera 125 take pictures continuously at a high frequency, for example, 30 times per second. The sonar 126 and the external environment sensor 127 output the observation results at the frequencies determined by the respective sensors. The sensor value acquisition unit 201 receives these images and signals at a predetermined frequency and outputs them to the landmark detection unit 142 and the point cloud generation unit 143. The subsequent processes, that is, each functional block 142 to 149, may operate each time the observation result is received, or may operate at a predetermined cycle.

[0032] The odometry estimation unit 141 estimates the movement of the vehicle 105 using the speed and steering angle of the vehicle 105 transmitted from the vehicle speed sensor 131 and the steering angle sensor 132. For example, it may be estimated using known dead reckoning, may be estimated using known visual odometry technology using a camera, or may be estimated using a combination of a known Kalman filter or the like. The vehicle speed, steering angle, vehicle movement information, and vehicle movement trajectory (odometry) obtained by integrating the vehicle movement are stored in the recording unit 113 as the driving situation 162 and output to the self-map generation unit 148. Note that position data by GNSS may be input to the odometry estimation unit 141, and odometry based on absolute coordinates estimated from the vehicle speed, steering angle, and position data may be output. In this case, it is possible to select the map closest to the current position from among a plurality of maps, or to refer to information combining the generated map and the actual map.

[0033] The landmark detection unit 142 first detects landmarks on the road surface using the images input from the cameras 121 to 124 and the stereo camera 125 and acquired by the sensor value acquisition unit 201. Landmarks on the road surface are features on the road surface having features distinguishable by sensors, such as lane marks which are a type of road paint, stop lines, crosswalks, and other regulatory signs. In this embodiment, vehicles and humans as moving objects, and building walls and the like which are obstacles obstructing the travel of vehicles are not treated as landmarks on the road surface.

[0034] Based on the information input from the cameras 121 to 124 and the stereo camera 125, the landmark detection unit 142 detects landmarks on the road surface existing around the vehicle 105, that is, features on the road surface distinguishable by sensors. The landmark information may be obtained in terms of pixels or as objects in which pixels are grouped. By image recognition, the type of landmark (for example, lane marks, crosswalks, etc.) may or may not be identified.

[0035] Next, based on the obtained landmark information, the landmark detection unit 142 generates a point cloud representing the landmarks on the road surface. This point cloud may be two-dimensional or three-dimensional, but in this embodiment, it will be described as a three-dimensional point cloud.

[0036] For example, using the information regarding the cameras 121 to 124 included in the sensor parameters 150, the distance to the road surface at each pixel in the images acquired from the cameras 121 to 124 can be calculated with relatively high precision, and three-dimensional coordinates can be calculated from the calculated distances. Using the calculated three-dimensional coordinates, point cloud data representing the three-dimensional coordinates of the pixels where landmarks exist, or the grouped objects, is output to the self-map generation unit 148.

[0037] Further, for example, the distance can be calculated from the parallax value acquired by the stereo camera 125 and the sensor parameters 150 based on the principle of triangulation. Furthermore, the three-dimensional coordinates of the object reflected in the corresponding pixel can be calculated from the calculated distance. In this embodiment, in order to output a high-precision three-dimensional point cloud, the Uniqueness Ratio, which is a parameter during parallax calculation by stereo matching, is adjusted in advance so that a high-precision parallax value is output. High-precision three-dimensional coordinates are calculated using the output high-precision parallax value, and point cloud data representing the three-dimensional coordinates of the pixels where road surface landmarks exist or the grouped objects is output to the self-map generation unit 148 using the calculated three-dimensional coordinates.

[0038] The point cloud data obtained by the landmark detection unit 142 is hereinafter referred to as road surface point cloud in order to distinguish it from the point cloud described later. Also, the three-dimensional coordinates are represented by relative coordinate values from the sensor, similar to the observed values in each sensor coordinate system.

[0039] The point cloud generation unit 143 generates a point cloud representing a landmark with a height such as a building wall using the image and sensor information acquired by the sensor value acquisition unit 201. This point cloud may be two-dimensional or three-dimensional, but will be described as a three-dimensional point cloud in this embodiment.

[0040] For example, three-dimensional coordinates can be calculated from the images acquired by the cameras 121 to 124 using a known motion stereo method. In the motion stereo method, the posture variation of the camera is measured by the time-series movement of feature points, which are characteristic points of the image such as the corners of an object, and three-dimensional coordinates are calculated based on the principle of triangulation. Calculating three-dimensional coordinates by the motion stereo method is difficult to calculate with high precision for arbitrary pixels, and can only be calculated with high precision for points that are easy to track as feature points. Points that are easy to track as feature points can be selected by known techniques. Three-dimensional coordinates are measured for points that are easy to track as feature points, and points higher than the road surface are output as a point cloud to the self-map generation unit 148.

[0041] Also, for example, in the case of the stereo camera 125, as described above, the distance can be measured from the parallax value and the sensor parameters 150 stored in the ROM 111 based on the principle of triangulation. Further, from the distance information, the three-dimensional coordinates of the object reflected in the corresponding pixel can be measured. Similar to the above, in order to output a high-precision three-dimensional point cloud, the Uniqueness Ratio, which is a parameter during parallax calculation by stereo matching, is adjusted in advance so that a high-precision parallax value is output. The three-dimensional coordinates are calculated using this high-precision parallax value, and the point cloud at a position higher than the road surface is output to the self-map generation unit 148.

[0042] Also, for example, the sonar 126 is a sensor that directly observes the point cloud. Since it is assumed to be attached so as not to observe the road surface, all the acquired point clouds are output as point clouds to the self-map generation unit 148. Among the acquired point clouds, the points arranged linearly are grouped, and the points are interpolated with a linear point cloud so as to have a predetermined interval, and a label indicating that it constitutes a linear point cloud is assigned.

[0043] Also, for example, the external environment sensor 127 is a sensor that directly observes the point cloud, and among the acquired point clouds, the point cloud at a position higher than the road surface is output to the self-map generation unit 148.

[0044] The point cloud data obtained by the point cloud generation unit 143 is called a high point cloud in order to distinguish it from the above-described road surface point cloud. When comparing the road surface point cloud and the high point cloud based on the observations of the cameras 121 to 124 and the stereo camera 125, if the observation distances are the same, the road surface point cloud has the characteristic of higher precision. Also, this three-dimensional coordinate is represented by the relative coordinate value from the sensor, similar to the observed value in each sensor coordinate system.

[0045] The self-map generation unit 148 obtains point cloud information represented by the relative coordinate values of the sensors from both the landmark detection unit 142 and the point cloud generation unit 143, and converts it into world coordinates using the vehicle motion information of the driving situation 162 obtained from the odometry estimation unit 141 and the sensor parameters 150 stored in the ROM 111. The converted point cloud information is synthesized in time series to generate a point cloud map. The world coordinate value is a coordinate value based on a certain coordinate and a certain axis. For example, it may be defined such that the position where the in-vehicle processing device 101 is activated is the origin, the direction in which it first moves forward is the X-axis, and the directions orthogonal to the X-axis are the Y-axis and the Z-axis.

[0046] The map obtained by the self-map generation unit 148 is stored in the RAM 112 as the self-generated map 161. After the next time, the self-generated map 161 is read from the RAM 112, the point cloud newly obtained from the landmark detection unit 142 and the point cloud generation unit 143 is coordinate-transformed, and they are synthesized in time series using the motion information of the host vehicle.

[0047] In the map generation mode 202, the processing in the self-map generation unit 148 is completed, and it waits for the input at the next time. In the self-position estimation mode 203, the self-map generation unit 148 outputs the self-generated map 161 and the driving situation 162 to the position estimation determination unit 144 and the driving situation diagnosis unit 145. When there is an instruction for map recording via the interface 180, the generated self-generated map and driving situation are recorded in the recording unit 113 as the self-generated map 151 and the driving situation 152. The processing flow of the self-map generation unit 148 will be described later with reference to FIG. 7.

[0048] The integrated diagnosis unit 140 includes a position estimation determination unit 144, a driving situation diagnosis unit 145, and an integrated determination unit 146. The integrated diagnosis unit 140 integrates the determination result by the position estimation determination unit 144 and the diagnosis result by the driving situation diagnosis unit 145 in the integrated determination unit 146 to determine the data to be used for position estimation.

[0049] The position estimation determination unit 144 determines, based on the point cloud, whether the global position estimation has been successful for the estimation result of the self-position estimation unit 149 described later. Whether the global position estimation has been successful will be described in detail later with reference to FIG. 3, but it refers to a state where there is no large (e.g., meter level) position error in the self-position estimation result and no large deviation such as a landmark. Whether the global position estimation has been successful can be determined, for example, by indicators such as whether the sum of the distance errors of the corresponding point clouds at the time of collation is below a predetermined value, whether the maximum value of the distances of the corresponding point clouds at the time of collation is below a predetermined value, whether the difference between the vehicle movement amount difference and the self-position estimation difference is below a predetermined value, or a combination thereof. Based on this determination result, the position deviation may be corrected.

[0050] At the time of the first operation, since the position estimation determination unit 144 is in an uncollated state, it does not operate and determines that the global position estimation has failed. After that, it refers to the collation result of the subsequent self-position estimation unit 149 to determine whether the global position estimation has been successful, and transmits the determination result to the integrated determination unit 146.

[0051] The driving situation diagnosis unit 145 diagnoses the driving situation using the driving situation 162 at the time of generating the self-generated map 161 by the self-map generation unit 148 and the past driving situation 152 recorded in the recording unit 113. The driving situation is determined for each point cloud obtained from each sensor as to whether it is a driving situation that affects the reduction in accuracy, and the determination result is transmitted to the integrated determination unit 146. For example, the high point cloud obtained from the cameras 121 to 124 via the landmark detection unit 142 has the property that its accuracy decreases as the speed of the vehicle 105 increases, and it is determined whether the speed at the time of map generation is equal to or higher than a predetermined speed.

[0052] Also, for example, in the case of the stereo camera 125, depending on the relationship between the viewing angle of the stereo camera and the observation distance, if the driving situation 162 is a vehicle behavior that greatly changes the observation direction in a short time, such as a switching operation, the observation target has changed compared to when generating the map, and there is a possibility that it becomes noise during point cloud matching and the accuracy decreases. However, even in the case of a switching operation, if the movement of the vehicle is the same as that during map generation, there will be no significant difference between the past map generation and the current observed point cloud, so the decrease in accuracy is small.

[0053] Also, for example, in the case of the sonar 126, since the density of the point cloud changes when the speed changes, if the speed is different from that during map generation, it contributes to a decrease in accuracy. Not limited to this example, considering the properties of the point clouds obtained from each sensor, it is determined whether the vehicle behavior regarding those properties satisfies the conditions.

[0054] The integrated determination unit 146 comprehensively determines the state from the determination result of the position estimation determination unit 144 and the diagnosis result of the driving situation diagnosis unit 145, and determines the point cloud data to be selected for improving the accuracy. The determination method based on the determination result of the position estimation determination unit 144 and the diagnosis result of the driving situation diagnosis unit 145 will be described with reference to FIG. 5. The integrated determination unit 146 transmits the determination result to the data selection unit 147.

[0055] The integrated diagnosis unit 140 is a functional block that summarizes the diagnostic functions composed of the position estimation determination unit 144, the driving situation diagnosis unit 145, and the integrated determination unit 146. This configuration is an example, and for example, it may be composed only of the position estimation determination unit 144, only of the driving situation diagnosis unit 145, or may be composed of other diagnostic methods. An example of the determination elements when the integrated diagnosis unit 140 is composed only of the position estimation determination unit 144 will be described with reference to FIG. 6.

[0056] The data selection unit 147 selects the point cloud data for each sensor based on the diagnosis result of the integrated diagnosis unit 140 and transmits it to the self-position estimation unit 149.

[0057] The self-position estimation unit 149 collates the self-generated map 161 obtained from the self-map generation unit 148 in the current driving situation with the self-generated map 151 generated in the past and stored in the recording unit 113, and estimates the self-position on the self-generated map 151. The point cloud used for the collation is the point cloud transmitted from the data selection unit 147. For the collation, for example, the ICP (Iterative Closest Point) algorithm, which is a known point cloud matching technique, can be used. Thereby, the coordinate transformation amount from the currently generated self-generated map 161 during driving to the self-generated map 151 generated in the past can be calculated, and the self-position on the self-generated map 151 can be estimated from the position of the current vehicle in the coordinate-transformed self-generated map 161. The processing flow of the self-position estimation unit 149 will be described later with reference to FIG. 8.

[0058] With reference to FIG. 3, the success or failure of the global position estimation will be described. FIGS. 3(a) and 3(b) show the states before the global position estimation is successful, and FIG. 3(c) shows the state after the global position estimation is successful. For the point cloud 301 of the past self-generated map 151 recorded in the recording unit 113, for the point cloud 302 of the currently observed self-generated map 161 generated during driving, for example, the ICP algorithm or the like is for use collated to estimate the estimated self-position 303. As shown in FIG. 3(b), there is a state where neither the local position estimation in a narrow range nor the global position estimation in a wide range has been successful. Also, as shown in FIG. 3(a), due to factors such as the initial collation position where the ICP algorithm is started and the similar point cloud arrangement even after sliding, even if the local position estimation in a narrow range is successful, the global position estimation in a wide range may not be successful, deviating from the original point cloud position, and as a result, the self-position estimation may be incorrect. Also, although the ICP algorithm operates to minimize the total distance error between the corresponding point clouds, depending on the initial collation position where the ICP algorithm is started, the collation may fail overall, and the point cloud may be collated to the position with the smallest error among the surrounding positions, resulting in an incorrect self-position estimation.

[0059] When global position estimation is successful, although a position estimation error of a certain degree (for example, about several tens of centimeters) remains due to the use of a wide range of point clouds, i.e., distant point clouds, there is no error at the meter level, and as shown in Fig. 3(c), a position can be estimated correctly to a certain extent. The position estimation determination unit 144 determines whether global position estimation has been successful based on the state of the point cloud and the like, and outputs the state of whether global position estimation has been successful or not.

[0060] Fig. 4 is a diagram showing the mounting states, observation ranges, and divisions of cameras 121 to 124, stereo camera 125, and sonar 126.

[0061] The stereo camera 125 has a narrow angle of view and can observe relatively distant observation ranges 401 to 403. Cameras 121 to 124 and sonar 126 are attached around the vehicle 105 and observe a relatively nearby observation range 404. The attachment positions and observation ranges of the respective sensors are examples. Cameras 121 to 124 and stereo camera 125 observe road surface point clouds and elevation point clouds. The point clouds observed by cameras 121 to 124 and stereo camera 125 have the property of being highly accurate in the vicinity of the vehicle 105 and having low accuracy far from the vehicle 105 for both road surface point clouds and valid point clouds. Therefore, for improving the self-position estimation accuracy, self-position estimation using point clouds observed in the vicinity of the vehicle 105 is desirable. However, if the self-position is estimated only using point clouds observed near the vehicle 105, as shown in Figs. 3(a) and 3(b), global position estimation may fail.

[0062] On the other hand, when distant points are also used, the matching is corrected as a whole, and as shown in Fig. 3(c), the possibility of successful global position estimation increases. Therefore, it is desirable to use distant points before global position estimation and, after global position estimation, remove distant points with poor accuracy and perform self-position estimation using nearby points.

[0063] Since the stereo camera 125 can observe relatively far away, it is advisable to divide the point clouds into groups such as the near point cloud 401, the middle point cloud 402, and the far point cloud 403, and use the point clouds appropriately according to the determination result of the integration determination unit 146. This is just an example, and it doesn't have to be divided into three groups.

[0064] Also, for the road surface point cloud and the high point cloud, since the road surface point cloud is more accurate than the high point cloud for both the cameras 121 to 124 and the stereo camera 125, it is considered in the appropriate use during the determination by the integration determination unit 146.

[0065] For the stereo camera 125, it is advisable to broadly classify the observed point clouds into a near road surface point cloud, a middle road surface point cloud, a far road surface point cloud, a near high point cloud, a middle high point cloud, and a far effective point cloud. For the cameras 121 to 124, since the observation range is narrow, it is advisable to distinguish between the road surface point cloud and the high point cloud without dividing by distance. This is just an example, and the cameras 121 to 124 can also be divided by distance and used appropriately by the integration determination unit 146. In this case as well, both the road surface point cloud and the effective point cloud have the property that the closer they are to the vehicle 105, the higher the accuracy. The point cloud obtained from the sonar 126 is a high point cloud, but since line labels and point labels are assigned to the effective point cloud, it is advisable to use the line-labeled effective point cloud and the point-labeled effective point cloud appropriately by the integration determination unit 146.

[0066] FIG. 5 is a diagram showing an example of the determination elements of the integration determination unit 146.

[0067] Based on the success or failure of the global position estimation determined by the position estimation determination unit 144 and the driving state determined by the driving situation diagnosis unit 145, the point cloud to be selected is determined for each sensor. FIG. 5(a) shows the determination elements related to the stereo camera 125, FIG. 5(b) shows the determination elements related to the cameras 121 to 124, and FIGS. 5(c) and 5(d) show the determination elements related to the sonar 126.

[0068] In the determination regarding the stereo camera 125 shown in Fig. 5(a), accuracy degradation is likely to occur with movements such as a switching operation. However, if the driving situation 152 during map generation is the same as the current driving situation 162 during driving, accuracy degradation is less likely to occur. Therefore, it is distinguished when the driving trajectory difference between the driving situation 152 during map generation and the current driving situation 162 during driving is smaller than a predetermined value and when it is larger. Generally, the trajectory difference becomes smaller during straight driving and larger during driving with a mixture of straight and curved sections. The driving trajectory difference may be divided into sections at a predetermined distance or time of the driving trajectory, and the difference in trajectories may be determined for each divided section. Also, before global self-position estimation, in order not to misestimate the global position, a wide range of points are used, and after global self-position estimation, it is possible to perform a more accurate position estimation by using only high-precision points for local position estimation. In view of these, when the driving trajectory difference is small and before global position estimation, the near, medium, and far high-point groups and the near, medium, and far road surface point groups are selected. When the driving trajectory difference is small and after global position estimation, the near high-point group and the near road surface point group are selected. When the driving trajectory difference is large and before global position estimation, the near and medium high-point groups and the near and medium road surface point groups are selected. When the driving trajectory difference is large and after global position estimation, the near road surface point group is selected.

[0069] In the determination regarding the cameras 121 to 124 shown in Fig. 5(b), when the speed is high, the accuracy of the high-point group decreases due to the nature of motion stereo. The road surface point group is not affected by speed and is highly accurate compared to the high-point group. Also, before global self-position estimation, in order not to misestimate the global position, a wide range of points are used, and after global self-position estimation, it is possible to perform a more accurate position estimation by using only high-precision points for local position estimation. In view of these, when the speed is low and before global position estimation, the high-point group and the road surface point group are selected. When the speed is low and after global position estimation, only the road surface point group is selected. When the speed is high and before global position estimation, only the road surface point group is used. When the speed is high and after global position estimation, only the road surface point group is used.

[0070] In the determination regarding the sonar 126 shown in FIGS. 5(c) and (d), when the speed difference between the driving situation 152 during map generation and the current driving situation 162 is large, the density of the point cloud changes and the accuracy decreases. However, for lines, they are less affected by the change in density. Also, in the case of the sonar 126, the accuracy does not vary significantly depending on the observation distance. In view of these, before the overall position estimation with a small speed difference, points and lines are selected; after the overall position estimation with a small speed difference, points and lines are selected; before the overall position estimation with a large speed difference, lines are selected; before the overall position estimation with a large speed difference, lines are selected. Also, since the sonar 126 is a sensor that uses sound waves, the accuracy varies due to temperature changes. Therefore, before the overall position estimation with a small temperature difference, points and lines are selected; after the overall position estimation with a small temperature difference, points and lines are selected; before the overall position estimation with a large temperature difference, lines are selected; before the overall position estimation with a large temperature difference, lines are selected.

[0071] These determination methods are just examples, and other combinations and point cloud selections can also be set according to the usage. Also, the driving situation may include weather, sunlight conditions, etc. Also, the stage of classifying the determination conditions does not have to be two stages, but can be three or more multi - stages.

[0072] FIG. 6 is a diagram showing an example of the determination elements of the integrated diagnosis unit 140, and shows an example of the determination elements when the integrated diagnosis unit 140 is composed only of the position estimation determination unit 144.

[0073] When the integrated diagnosis unit 140 is composed only of the position estimation determination unit 144, for each sensor, the point cloud to be selected is determined based only on the success or failure of the overall position estimation determined by the position estimation determination unit 144. FIG. 6(a) shows the determination elements regarding the stereo camera 125, FIG. 6(b) shows the determination elements regarding the cameras 121 - 124, and FIG. 6(c) shows the determination elements regarding the sonar 126. In FIGS. 6(a), (b), and (c), a wide range of points are utilized before the overall position estimation, and only the points with high accuracy are utilized after the overall position estimation.

[0074] FIG. 7 is a flowchart showing the operation of the self-map generation unit 148. When the self-map generation unit 148 receives an execution command from the landmark detection unit 142 and the point cloud generation unit 143, it executes the following operations. The execution entity of each step described below is the CPU 110.

[0075] In the landmark acquisition step 701, the self-map generation unit 148 acquires, from the RAM 112, the road surface point cloud generated from the outputs of the cameras 121 to 124 and the stereo camera 125 by the landmark detection unit 142. The road surface point cloud is expressed in each sensor coordinate system. Proceed to the next step 702.

[0076] In the point cloud acquisition step 702, the self-map generation unit 148 acquires, from the RAM 112, the elevation point cloud of each sensor generated by the point cloud generation unit 143. The elevation point cloud is represented in each sensor coordinate system. Proceed to the next step 703.

[0077] In the coordinate conversion step 703, the self-map generation unit 148 converts the coordinate systems of the road surface point cloud acquired in the landmark acquisition step 701 and the elevation point cloud acquired in the point cloud acquisition step 702 into the coordinate system of the vehicle 105. The sensor parameters 150 for each sensor stored in the ROM 111 are used for the conversion. The point cloud after coordinate conversion is output to the map generation step 704, and proceed to the next step 704.

[0078] In the map generation step 704, using the vehicle motion information obtained from the odometry estimation unit 141, the self-map generation unit 148 converts the coordinate point cloud obtained in the coordinate conversion step 703 into world coordinates, and self-generates a point cloud map by a time-series synthesis with the point cloud map at the previous time stored in the RAM 112. The self-generated map, and the driving conditions such as the vehicle speed, steering angle, vehicle motion information, and vehicle motion trajectory obtained by integrating the vehicle motion from the odometry estimation unit 141 are output to the point cloud map output step 705, and proceed to the next step 705.

[0079] In the point cloud map output step 705, the point cloud map generated in the map generation step 704 is output to the RAM 112 as the self-generated map 161 or to the recording unit 113 as the self-generated map 151. Similarly, the driving situation obtained from the map generation step 704 is output to the RAM 112 as the driving situation 162 or to the recording unit 113 as the driving situation 152, and the process ends. In the case of the map generation mode 202, basically, the self-generated map 161 and the driving situation 162 are output to the RAM 112, and then the sensor value acquisition unit 201 acquires sensor values. When a map recording instruction is received from the user via the interface 180, the point cloud map generated in the map generation step 704 is output to the recording unit 113 as the self-generated map 151 and the driving situation 152, and the process ends. In the case of the self-position estimation mode 203, the position estimation determination unit 144 and the driving situation diagnosis unit 145 start the process.

[0080] FIG. 8 is a flowchart showing the operation of the self-position estimation unit 149. When the self-position estimation unit 149 receives an execution command from the data selection unit 147, it executes the following operations. The execution subject of each step described below is the CPU 110. This process is executed only in the self-position estimation mode 203.

[0081] In the past map information acquisition step 801, the self-generated map 151 stored in the recording unit 113 is acquired. The self-generated map 151 is a map obtained by past driving. Proceed to the next step 802.

[0082] In the current map information acquisition step 802, the selected point cloud of the self-generated map 161 observed during the current driving, which is determined by the integration determination unit 146 and selected by the data selection unit 147, is acquired from the RAM 112. Proceed to the next step 803.

[0083] In the map matching step 803, from the past self-generated map 151 obtained in the past map information acquisition step 801 and the self-generated map 161 observed during the current driving obtained in the current map information acquisition step 802, the self-position on the past self-generated map 151 is calculated by matching the point clouds selected by the integration determination unit 146 and the data selection unit 147 from the viewpoint of accuracy improvement. For the matching of the point clouds, for example, the ICP (Iterative Closest Point) algorithm, which is a known point cloud matching technique, can be used. In the ICP algorithm, the correspondence relationship between each point is calculated, and the process of minimizing the distance error in the calculated correspondence relationship is repeated to match the point clouds. By this matching of the maps, the coordinate transformation amount for converting the self-generated map 161 obtained from the self-map generation unit 148 onto the self-generated map 151 can be calculated, and the self-position on the self-generated map 151 can be obtained from the position of the current vehicle after coordinate transformation. The self-position on the self-generated map 151 is output to end the process shown in FIG. 8. After the end, the sensor value acquisition unit 201 starts the process.

Embodiment

[0084] Hereinafter, Embodiment 2 of the present invention will be described. In Embodiment 2, the differences from Embodiment 1 described above will be mainly described, and the same components and processes as those in Embodiment 1 are denoted by the same reference numerals, and their descriptions are omitted.

[0085] FIG. 9 is a block diagram showing the configuration of the vehicle 105 according to Embodiment 2 of the present invention.

[0086] Vehicle 105 includes a map generation and self-position estimation device 100 and a group of vehicle control devices 170 to 173 that control the automatic driving of vehicle 105. The map generation and self-position estimation device 100 includes cameras 121 to 124, a stereo camera 125, a sonar 126, an external environment sensor 127, a vehicle speed sensor 131, a steering angle sensor 132, an interface 180, a GNSS receiver 181, a communication device 182, and a display device 183. The cameras 121 to 124, the stereo camera 125, the sonar 126, the external environment sensor 127, the vehicle speed sensor 131, the steering angle sensor 132, the interface 180, the GNSS receiver 181, the communication device 182, the display device 183, and the group of vehicle control devices 170 to 173 are connected to an in-vehicle processing device 101 by signal lines and exchange various data with the in-vehicle processing device 101.

[0087] The GNSS receiver 181 receives signals transmitted from a plurality of satellites constituting a satellite navigation system (e.g., GPS) and calculates the position of the GNSS receiver 181, i.e., the latitude and longitude, by performing calculations using the received signals. Note that the accuracy of the latitude and longitude calculated by the GNSS receiver 181 may be low, and for example, errors on the order of several meters to 10 meters may be included. The GNSS receiver 181 outputs the calculated latitude and longitude to the in-vehicle processing device 101.

[0088] The communication device 182 is a communication device used for the in-vehicle processing device 101 and a device outside vehicle 105 to exchange information wirelessly. For example, when the user is outside vehicle 105, it communicates with a mobile terminal worn by the user to exchange information. The communication target of the communication device 182 is not limited to the user's mobile terminal and may also be a device that provides data related to automatic driving.

[0089] The display device 183 is, for example, a liquid crystal display and displays information output from the in-vehicle processing device 101.

[0090] The vehicle control device 170 controls at least one of the steering device 171, the drive device 172, and the braking device 173 based on information output from the in-vehicle processing device 101, for example, the current self-position on the self-generated map 151 output from the self-position estimation unit 149. The steering device 171 operates the steering of the vehicle 105. The drive device 172 applies a driving force to the vehicle 105. For example, the drive device 172 increases or decreases the driving force of the vehicle 105 by increasing or decreasing the target rotational speed of the engine or motor provided in the vehicle 105. The braking device 173 applies a braking force to the vehicle 105 to decelerate the vehicle 105.

[0091] In the vehicle 105, the in-vehicle processing device 101 may store the self-generated map 151 stored in the recording unit 113 in a server or the like via the communication device 182. Further, the in-vehicle processing device 101 may store the self-generated map 151 mounted on another vehicle 105 and stored in the server in the recording unit 113 via the communication device 182 and utilize it for self-position estimation. Furthermore, the in-vehicle processing device 101 attaches the reception position of each time of the GNSS receiver 181 to the self-generated map 161, and when estimating the self-position, may search for a map with a value close to the current position information by GNSS and select a map to be collated. When the GNSS signal cannot be received, the odometry estimation unit 141 may estimate the absolute position from the difference from the absolute position most recently observed by GNSS.

[0092] FIG. 10 is a diagram showing map sharing by the map generation / self-position estimation device 100.

[0093] The host vehicle 105 and the other vehicle 1103 each have a map generation and self-position estimation device 100 having a communication device 182. The host vehicle 105 uses the communication device 182 to record the self-generated map 151 and the driving situation 152 in an external recording device 1102 connected to a server 1101 on the cloud via a communication line. The other vehicle 1103 uses the communication device 182 to acquire the self-generated map 151 and the driving situation 152 generated by the vehicle 105 from the external recording device 1102 connected to the server 1101 on the cloud, and utilizes them as a map for the other vehicle 1103. Further, the host vehicle 105 may utilize the self-generated map 151 and the driving situation 152 generated by the other vehicle 1103.

[0094] As described above, the in-vehicle processing device 101 of the present embodiment includes a past map storage unit (recording unit 113) that stores a past generated map generated in the past, a self-map generation unit 148 that generates an on-the-go map based on sensor data acquired during travel, a self-position estimation unit 149 that estimates the self-position by comparing the past generated map and the on-the-go map, and a data selection unit 147 that selects data used by the self-position estimation unit 149 from the on-the-go map based on the characteristics of the data included in the on-the-go map. Therefore, while considering the accuracy of the point cloud generated from the image captured by the camera and the occurrence of local fitting during position estimation using the point cloud, an optimal point cloud is selected to estimate the self-position, and the self-position can be estimated with high accuracy.

[0095] Further, the self-position estimation unit 149 is capable of performing a first comparison (global comparison) with a wide comparison range between the past generated map and the on-the-go map and a second comparison (local comparison) with a narrow comparison range, and the data selection unit 147 selects different data for the first comparison and the second comparison. Therefore, the self-position can be estimated with high accuracy by the global comparison that reduces the error as a whole including distant points and the local comparison using points with high sensing accuracy.

[0096] Further, since the self-position estimation unit 149 performs the second comparison after the first comparison, it is possible to prevent an incorrect estimation of the self-position by performing the local comparison before the global comparison, and the self-position can be estimated with high accuracy.

[0097] Furthermore, an integrated determination unit 146 is provided to identify the type of data used for position estimation based on the driving situation and the estimated result of the self-position during map generation. Since the driving situation includes the driving situation during past map generation and the current driving situation, the estimation accuracy of the self-position can be improved.

[0098] In addition, the self-map generation unit 148 generates a running map converted into absolute coordinates estimated by GNSS and odometry, and is further connected via a communication line to a server 1101 that stores the map. In order to make it available for other vehicles 1103, the running map is transmitted to the server 1101, and the map generated by the other vehicle 1103 is used as a past generated map, so that maps can be shared among many vehicles and a wide range of maps can be used.

[0099] In addition, the self-position estimation unit 149 outputs the estimated self-position to the vehicle control device 170, and the vehicle control device 170 controls at least one of the steering device 171, the driving device 172, and the braking device 173 based on the self-position output from the in-vehicle processing device 101. Therefore, automatic driving and driving support can be realized based on the map generated by itself.

[0100] Note that the present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the appended claims. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to those having all the configurations described. Other aspects conceivable within the scope of the technical idea of the present invention are also included in the scope of the present invention. Also, a part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Also, the configuration of another embodiment may be added to the configuration of one embodiment. Also, for a part of the configuration of each embodiment, addition, deletion, or replacement with other configurations may be performed.

[0101] In addition, the map generation and self-position estimation device 100 is provided with an input / output interface (not shown), and when necessary, it may read a program from another device via the input / output interface and a usable medium. The medium is, for example, a removable storage medium attached to the input / output interface, or a communication medium, that is, a wired, wireless, optical, etc. network, or a carrier wave or digital signal propagating through the network.

[0102] In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, by means of an integrated circuit, or may be realized in software by a processor interpreting and executing a program for realizing each function. Also, part or all of the functions realized by the program may be realized by a hardware circuit or FPGA.

[0103] Information such as programs, tables, files, etc. for realizing each function can be stored in a storage device such as a memory, hard disk, SSD (Solid State Drive), or a recording medium such as an IC card, SD card, DVD.

[0104] Also, the control lines and information lines show those considered necessary for explanation, and do not necessarily show all the control lines and information lines required for implementation. In reality, it may be considered that almost all configurations are interconnected.

Explanation of Signs

[0105] 100…Map generation and self-position estimation device 101…In-vehicle processing device 105…Vehicle 110…CPU 111…ROM 112…RAM 113…Recording unit 121~124…Camera 125…Stereo camera 126…Sonar 127…Other external sensors 131…Vehicle speed sensor 132… Rudder Angle Sensor 140… Integrated Diagnosis Unit 141… Odometry Estimation Unit 142… Landmark Detection Unit 143… Point Cloud Generation Unit 144… Position Estimation Judgment Unit 145… Driving Condition Diagnosis Unit 146… Integrated Judgment Unit 147… Data Selection Unit 148… Self-Map Generation Unit 149… Self-Position Estimation Unit 150… Sensor Parameter 151, 161… Self-Generated Map 152, 162… Driving Condition 170… Vehicle Control Device 171… Steering Device 172… Driving Device 173… Braking Device 180… Interface 181… GNSS Receiver 182… Communication Device 183… Display Device 201… Sensor Value Acquisition Unit 250… Self-Position Estimation Unit 1101… Server 1102… External Recording Device 1103… Other Vehicles

Claims

1. A CPU, a recording unit that records a running map generated based on a past generated map generated in the past and sensor data acquired by a sensor during the running of the host vehicle and including first point cloud data, comprising: the first point cloud data includes road surface point cloud data representing landmarks on the road surface and high point cloud data that is point cloud at a position higher than the road surface, the CPU performs a first collation for collating the past point cloud data included in the past generated map with the road surface point cloud data, and the past point cloud data with the high point cloud data, performs a first self-position estimation for estimating the self-position of the host vehicle based on the result of the first collation, when it is determined that the first self-position estimation is successful, based on the result of a second collation for collating the past point cloud data with the road surface point cloud data selected as the second point cloud data from the past point cloud data and the first point cloud data, the vehicle-mounted processing device is characterized in that it estimates the self-position of the host vehicle.

2. A CPU, a running map generated based on a past generated map generated in the past and sensor data acquired by a sensor during the running of the host vehicle, and a recording unit that records the running situation at the time of generation of the past generated map, comprising: the CPU performs a first self-position estimation for estimating the self-position of the host vehicle based on the result of a first collation for collating the past point cloud data included in the past generated map with the first point cloud data included in the running map, when it is determined that the first self-position estimation is successful, based on the running situation at the time of generation of the past generated map and the running situation at the time of generation of the running map, the second point cloud data is selected from the first point cloud data, and based on the result of a second collation for collating the past point cloud data included in the past generated map with the second point cloud data, the vehicle-mounted processing device is characterized in that it estimates the self-position of the host vehicle.

3. The vehicle-mounted processing device according to claim 1 or 2, wherein when it is determined that the first self-position estimation is successful, the CPU selects, as the second point cloud data, point cloud data included in a second collation range that is narrower than the first collation range of the first collation.

4. The vehicle-mounted processing device according to claim 3, wherein the second collation range is within the first collation range.

5. The vehicle-mounted processing device according to claim 1 or 2, The in-vehicle processing device is characterized in that the CPU generates a running map converted into absolute coordinates estimated by GNSS and odometry.

6. The in-vehicle processing device according to claim 5, wherein the CPU is connected via a communication line to a server that stores a map, transmits the running map to the server for use by other vehicles, and receives a map generated by other vehicles from the server and uses it as the past generated map. The in-vehicle processing device is characterized by this.

7. The in-vehicle processing device according to claim 1 or 2, wherein the CPU outputs the estimated own position to a vehicle control device. The in-vehicle processing device is characterized by this.

8. A vehicle control device, characterized in that it controls at least one of a steering device, a driving device, and a braking device based on the own position output from the in-vehicle processing device according to claim 7.

9. An own position estimation method for an in-vehicle processing device to estimate its own position, wherein the in-vehicle processing device includes a CPU that executes predetermined arithmetic processing and a recording unit accessible by the CPU, the recording unit records a past generated map generated in the past and a running map generated based on sensor data acquired by a sensor during the running of the own vehicle and including first point cloud data, the first point cloud data includes a road surface point cloud representing a landmark on the road surface and a high point cloud that is a point cloud at a position higher than the road surface, the own position estimation method includes a step of the CPU performing a first collation of collating the past point cloud data included in the past generated map with the road surface point cloud data and the past point cloud data with the high point cloud, and based on the result of the first collation, performing a first own position estimation of estimating the own position of the own vehicle, and a step of the CPU estimating the own position of the own vehicle based on the result of a second collation of collating the past point cloud data with road surface point cloud data selected as second point cloud data from the past point cloud data and the first point cloud data when it is determined that the first own position estimation has been successful. The own position estimation method is characterized by including this.

Images