WO2020021596A1 - Vehicle position estimation device and vehicle position estimation method - Google Patents

Abstract

In this vehicle position estimation device 6, a vehicle position estimation unit 61 calculates first estimated vehicle position information indicating the position and orientation of a vehicle on a map on the basis of absolute positioning information for the vehicle obtained through satellite positioning and vehicle sensor information obtained from a vehicle sensor of the vehicle. A road feature position/orientation detection unit 62 selects a road feature to be detected on the basis of the first estimated vehicle position information and at least one from among road feature information included in map data and imaging sensor information obtained from an imaging sensor of the vehicle. Further, the road feature position/orientation detection unit 62 detects the relative position and relative orientation of the selected road feature in relation to the vehicle from optical distance measurement information obtained from an optical distance measurement sensor of the vehicle and calculates second estimated vehicle position information indicating the position and orientation of the vehicle on the map on the basis of the detected relative position and relative orientation of the road feature.

Description

Vehicle position estimating apparatus and vehicle position estimating method

The present invention relates to a vehicle position estimating device for estimating a current position of a vehicle.

For example, in Patent Document 1 below, a road sign is detected from image data obtained by photographing the periphery of a vehicle, and a road sign is detected based on a relative position of the vehicle with respect to the detected road sign and a position of the road sign included in the map information. Thus, a technique for estimating the position of a vehicle on a road has been disclosed.

JP 2016-176679 A

(4) An imaging sensor such as a monocular camera has an advantage that it can recognize the type of an imaging target such as a road sign and is relatively inexpensive. However, the position detection using the image sensor has a detection error in the traveling direction of the vehicle of about several meters, and it is difficult to detect the position with high accuracy. Therefore, the position detection by the image sensor is not suitable for applications that need to detect the position and direction of the vehicle with high accuracy, such as automatic driving of vehicles and preventive safety technology.

On the other hand, optical ranging sensors such as LiDAR (Light Detection and Ranging) and stereo cameras have high position detection accuracy. However, in order to detect the position of a road sign or the like, it is necessary to process a large amount of point cloud information including information on a distance to a detection target and information on reflection intensity or luminance / color. (Central Processing Unit) is required. In addition, since the processing time required for position detection is also required in units of several seconds, it is not suitable for applications requiring immediate responsiveness, such as automatic driving of vehicles and preventive safety technology.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a vehicle position estimating apparatus having both high-accuracy position detection and immediate responsiveness.

The vehicle position estimating device according to the present invention is a first vehicle position estimating device that indicates a position and an azimuth of the own vehicle on a map based on absolute positioning information of the own vehicle obtained by satellite positioning and vehicle sensor information obtained from a vehicle sensor of the own vehicle. A vehicle position estimating unit for calculating the estimated own vehicle position information, one of the first estimated own vehicle position information, information on road features included in the map data, and image sensor information obtained from the image sensor of the own vehicle. Or, based on both, select a road feature to be detected, and determine the relative position and relative orientation of the selected road feature from the own vehicle from the optical ranging information obtained from the optical ranging sensor of the own vehicle. Based on the detected relative position and relative orientation of the road feature detected from the optical ranging information, the road feature position and second estimated vehicle position information indicating the position and orientation of the vehicle on the map are calculated. An orientation detection unit. Than it is.

According to the present invention, when obtaining the second estimated own-vehicle position information, the range in which the detection processing is performed on the optical ranging information includes the position of the road feature estimated from the first estimated own-vehicle position information. , The time required for the detection process can be reduced. As a result, a vehicle position estimating device having both high-accuracy position detection and immediate responsiveness can be realized.

The objects, features, aspects, and advantages of the present invention will become more apparent from the following detailed description and the accompanying drawings.

1 is a block diagram illustrating a configuration of a vehicle position estimation system according to an embodiment of the present invention. It is a figure showing an example of a detectable range of an optical distance measuring sensor device. It is a figure which shows the example of derivation of the detectable area of a road feature. It is a figure which shows the example of the target range of the detection processing of a road feature. 4 is a flowchart illustrating an operation of the vehicle position estimation device according to the embodiment of the present invention. FIG. 2 is a diagram illustrating an example of a hardware configuration of a vehicle position estimation device. FIG. 2 is a diagram illustrating an example of a hardware configuration of a vehicle position estimation device.

FIG. 1 is a block diagram showing a configuration of a vehicle position estimation system according to an embodiment of the present invention. As shown in FIG. 1, the vehicle position estimation system includes a satellite positioning device 1, a vehicle sensor information output device 2, an imaging sensor device 3, an optical ranging sensor device 4, a high-accuracy map database 5, and a vehicle position estimation device 6. ing. Hereinafter, a vehicle equipped with the vehicle position estimation system is referred to as “own vehicle”.

The satellite positioning device 1 calculates the absolute position (latitude, longitude, altitude) and absolute azimuth of the own vehicle by satellite positioning based on a positioning signal transmitted by a Global Positioning System (GNSS) satellite such as a GPS (Global Positioning System) satellite. Then, the calculated information on the absolute position and the absolute direction of the own vehicle is output as “absolute positioning information”.

The vehicle sensor information output device 2 outputs “vehicle sensor information” obtained from vehicle sensors such as a vehicle speed sensor, a gyro, a steering angle sensor, and an air pressure sensor mounted on the own vehicle. It is assumed that the vehicle sensor information includes at least one of travel speed, rotation angle, steering angle, and air pressure.

The imaging sensor device 3 includes an imaging device (camera sensor) such as a monocular camera, for example, and detects a feature around the road where the own vehicle is located (hereinafter, referred to as a “road feature”) from an image of the vicinity of the own vehicle. Detect position, shape, type, etc. Road features include not only three-dimensional features such as traffic road signs and traffic lights, but also planar features such as lane markings and stop lines drawn on the road surface. In addition, the position of the road feature detected by the imaging sensor device 3 is a relative position from the own vehicle. The imaging sensor device 3 outputs information on the detected position, shape, type, and the like of the road feature and information on its detection accuracy as “imaging sensor information”.

The optical ranging sensor device 4 includes an optical ranging sensor such as a LiDAR or a stereo camera, and includes “optical ranging information including information on a distance to an obstacle existing around the own vehicle and reflection intensity or luminance / color. Is output. For example, in LiDAR, an object that reflects light is detected as an obstacle. In the stereo camera, an object whose distance can be detected from a parallax image (images captured from a plurality of different viewpoints) is detected as an obstacle.

Here, it is assumed that the detectable range of the optical distance measuring sensor device 4 is preset in the vehicle position estimating device 6 as a parameter. Since the range changes according to the performance of the optical distance measuring sensor device 4, it is preferable that the range is determined based on an experiment or the like.

FIG. 2 shows an example of the detectable range of the optical distance measuring sensor device 4. In FIG. 2, the X axis is set in the traveling direction (head direction) of the host vehicle, the Y axis is set in the lateral direction of the host vehicle, and the Z axis is set in the height direction with the host vehicle position as the origin. The XY plane with Z = 0 corresponds to the road surface of the road. The own vehicle position as the origin corresponds to the own vehicle position estimated by the vehicle position estimating device 6, and is, for example, the center of gravity position, the center position, or the head position of the own vehicle. The detectable range of the optical ranging sensor device 4 is defined as a space defined by a detectable distance L1, a detectable angle Θ1 on the XY plane, and a detectable angle Θ2 on the ZY plane, as shown in FIG. it can. In a general optical distance measuring sensor, the detectable distance is considered to be the same on the XY plane and the ZX plane, but different detectable distances may be set on the XY plane and the ZX plane.

Returning to FIG. 1, the high-precision map database 5 is a database in which data of high-precision maps is stored. The high-precision map includes information indicating the lane shape of the road, such as the position, shape, and type of lane markings of the road, the position of the stop line, the position, shape, type of traffic road signs and traffic signals installed around the road, Detailed information on road features, such as direction information, is included.

The high-accuracy map database 5 may not be mounted on the own vehicle, and may be, for example, a server that distributes high-accuracy map data to the vehicle position estimating device 6 by communication. In FIG. 1, the high-accuracy map database 5 is depicted as a different block from the vehicle position estimating device 6, but the high-accurate map database 5 may be provided inside the vehicle position estimating device 6. Further, the high-accuracy map database 5 and the vehicle position estimating device 6 may be provided in a navigation system of the own vehicle.

The vehicle position estimating device 6 includes a vehicle position estimating unit 61 and a road feature position / azimuth detecting unit 62, as shown in FIG.

The vehicle position estimating unit 61 includes the absolute positioning information (the absolute position and the absolute azimuth of the own vehicle) output from the satellite positioning device 1 and the vehicle sensor information (the traveling speed and the rotation angle of the own vehicle) output from the vehicle sensor information output device 2. , Steering angle, and air pressure) and the position and orientation of the vehicle on the high-accuracy map based on the high-accuracy map data stored in the high-accuracy map database 5. Is output as “first estimated vehicle position information”. Note that the absolute positioning information and the vehicle sensor information are input to the vehicle position estimating unit 61 at regular intervals, and the vehicle position estimating unit 61 outputs the first estimated own vehicle position information at regular intervals.

The road feature position / azimuth detecting unit 62 outputs the first estimated own vehicle position information (the position and the azimuth of the own vehicle on the high-accuracy map estimated by the vehicle position estimating unit 61) output by the vehicle position estimating unit 61, Based on the high-accuracy map stored in the high-accuracy map database 5 and at least one of the image sensor information (the position, shape, type, and the like of the road feature detected by the image sensor 3) output by the image sensor 3 Then, a road feature to be detected is selected from the optical ranging information output by the optical ranging sensor device 4. In addition, the road feature position / azimuth detecting unit 62 detects the relative position and relative orientation of the selected road feature from the own vehicle from the optical ranging information output by the optical ranging sensor device 4. Further, the road feature position / azimuth detecting unit 62 detects the position of the vehicle on the high-precision map based on the relative position and relative orientation of the road feature detected from the imaging sensor information and the data of the high-precision map. And the direction is estimated. The information on the position and orientation of the vehicle on the high-accuracy map estimated by the road feature position / azimuth detecting unit 62 is output from the road feature position / azimuth detecting unit 62 as “second estimated own vehicle position information”. You.

Here, the road feature position / azimuth detection unit 62 can detect the relative position and relative orientation from the own vehicle at higher speed and with higher accuracy as the road feature to be detected, that is, the optical distance measurement information. Select the one with higher density and accuracy. Specifically, a road feature having a short distance from the own vehicle (a short detection distance) and a large size (a large detectable area) when viewed from the own vehicle is selected as a detection target. For example, by performing a road feature detection process based on the first estimated own vehicle position information and the high-precision map or the imaging sensor information, a 10-cm-diameter cylindrical pole and a 30-cm-radius disk-shaped road sign are automatically detected. If detected at the same distance from the car, the road feature position / azimuth detecting unit 62 selects a road sign of both as a detection target.

Since the size (detectable area) of the road feature viewed from the own vehicle changes depending on the relationship between the direction of the own vehicle (the detection direction of the optical ranging sensor device 4) and the direction of the surface of the road feature, the road feature is determined. The position / direction detection unit 62 may calculate the size of the road feature based on the positional relationship between the vehicle and the road feature. For example, as shown in FIG. 3, when the angle (detection angle) between the surface of a rectangular road feature having an area S and the detection direction of the optical distance measuring sensor device 4 is Θa, the road feature can be detected. The area is S · sinΘa.

When detecting the relative position and relative orientation of the road feature selected as the detection target, the road feature position / azimuth detecting unit 62 outputs the optical ranging information (point group information) output by the optical ranging sensor device 4. Among them, the detection processing is performed on the optical ranging information within the range in which the road feature is estimated to be located. The position of the road feature is known at the stage of selecting the road feature to be detected. By limiting the range of the optical ranging information to be subjected to the road feature detection processing in this way, the calculation load required for the detection processing can be reduced, and the position and orientation of the road feature can be detected at high speed.

The range of the optical distance measurement information to be subjected to the road feature detection processing includes an area having a width corresponding to a predetermined margin at the outer edge of the range corresponding to the estimated position and shape of the road feature. The range is added. For example, when the road feature selected as the detection target is a rectangular road sign as shown in FIG. 4, the range of the optical ranging information to be subjected to the detection processing is a radius R including the shape of the road sign. 4 is defined as a range obtained by adding a region having a margin m to the outer edge of the circle, that is, a range of a circle having a radius R + m shown in FIG.

The allowance m is preferably a value equal to or larger than the estimation error e of the first estimated own vehicle position information. The reason is that the range where the road feature is estimated to be located is obtained as the position (relative position) of the road feature based on the own vehicle position on the high-accuracy map indicated by the first estimated own vehicle position information. For this reason, the same error as the error generated in the first estimated own-vehicle position information also occurs in the position of the road feature position. The information of the estimation error e may be included in the first estimated own-vehicle position information, or may be a value previously obtained from an experiment or the like.

In addition, for the detection of road features using optical ranging information, an optimal detection model is applied according to the shape information (circle, rectangle, cylinder, etc.) included in the road feature information of the high-precision map. I do. For example, when the shape information of a road feature indicates a cylinder, the cylinder may be detected using a Cylinder model of PCL (Point @ Cloud @ Library). By doing so, the road feature detection process can be performed specifically for the process of extracting a specific shape, so that the load on the road feature detection processing is reduced, and the position and orientation of the road feature are rapidly increased. Can be detected.

Since the position of the own vehicle indicated by the second estimated own vehicle position information calculated by the road feature position / azimuth detecting unit 62 is obtained based on the position of the road feature detected from the optical ranging information, the accuracy is high. . Further, the road feature position / azimuth detecting unit 62 restricts the range of the optical distance measurement information to be subjected to the road feature detection processing when obtaining the second estimated own vehicle position information. Can be obtained in a short time. Therefore, the vehicle position estimating device 6 of the present embodiment has both high-accuracy position detection and immediate responsiveness. Also, by limiting the range of the optical ranging information to be subjected to the road feature detection processing, the calculation load for the detection processing is reduced. Therefore, the effect that the position of the own vehicle can be estimated with high accuracy can be obtained.

However, even if the range of the optical ranging information to be subjected to the road feature detection processing is limited, the detection processing may take a certain amount of time (200 to 300 milliseconds may be required). . Therefore, especially when the own vehicle is running, a difference may occur between the position and the azimuth of the own vehicle indicated by the second estimated own vehicle position information and the current position and the azimuth of the own vehicle.

Therefore, in the present embodiment, the vehicle position estimating unit 61 obtains an error in the first estimated own vehicle position information using the second estimated own vehicle position information, and based on the error, determines the current (latest) By correcting the estimated own vehicle position information of No. 1, a more accurate position and orientation of the own vehicle are obtained. Specifically, the vehicle position estimating unit 61 acquires from the road feature position / azimuth detecting unit 62 the second estimated own vehicle position information and the information on the acquisition time of the optical ranging information used for the calculation, The difference between the first estimated own vehicle position information at the same time and the second estimated own vehicle position information acquired from the road feature position / azimuth detecting unit 62 is obtained, and the obtained difference is used as the first estimated own vehicle position information. It is regarded as an error of the vehicle position information. Then, the vehicle position estimating unit 61 calculates the more accurate position and orientation of the own vehicle by adding the error to the latest first estimated own vehicle position information.

In order for the vehicle position estimating unit 61 to perform the above-described correction, the vehicle position estimating unit 61 uses the first estimated own-vehicle position information from the present time to a certain time (200 to 300 ms or more). Must be held together with the calculation time.

FIG. 5 is a flowchart showing the operation of the vehicle position estimation device 6. Hereinafter, with reference to FIG. 5, a process for the vehicle position estimating device 6 to estimate the own vehicle position (own vehicle position estimating process) will be described. Note that the flow of FIG. 5 is repeatedly executed after the activation of the vehicle position estimating device 6.

First, the vehicle position estimating unit 61 determines the position and orientation of the vehicle on the high-precision map based on the absolute positioning information output by the satellite positioning device 1 and the vehicle sensor information output by the vehicle sensor information output device 2. First estimated vehicle position information including the information is calculated (step S100).

Next, the road feature position / azimuth detecting unit 62 is in a detectable range of the optical distance measuring sensor device 4 among the road features existing around the own vehicle position indicated by the first estimated own vehicle position information. Information on road features is acquired from the high-accuracy map database 5 (step S101).

Then, the road feature position / azimuth detecting unit 62 calculates the first estimated own vehicle position information calculated in step S100, the road feature information acquired in step S101, and the imaging sensor information output by the imaging sensor device 3. A road feature to be detected is selected based on one or both of the above (step S102).

Subsequently, the road feature position / azimuth detecting unit 62 performs a detection process on the optical ranging information (point group information) output from the optical ranging sensor device 4, so that the road feature selected in step S <b> 102 can be used. The relative position and relative azimuth from the car are detected (step S103). This detection process is performed on the optical ranging information output from the optical ranging sensor device 4 within the range where the road feature is estimated to be located.

Further, the road feature position / azimuth detection unit 62 performs a high-precision map based on the information on the road feature obtained in step S101 and the relative position and relative orientation of the road feature detected from the own vehicle in step S103. The second estimated own vehicle position information including the information on the position and the direction of the upper own vehicle is calculated (step S104).

Lastly, the vehicle position estimating unit 61 calculates the second estimated own vehicle position information calculated in step S104 and the first estimated own vehicle position information corresponding to the acquisition time of the optical ranging information used for the calculation. Is obtained as an error of the first estimated own vehicle position information, and the error is added to the latest first estimated own vehicle position information, thereby correcting the latest first estimated own vehicle position information. (Step S105). Based on the corrected first estimated own-vehicle position information, the current position and orientation of the own-vehicle can be obtained.

FIGS. 6 and 7 are diagrams illustrating examples of the hardware configuration of the vehicle position estimating device 6, respectively. Each function of the components of the vehicle position estimation device 6 shown in FIG. 1 is realized by, for example, a processing circuit 70 shown in FIG. That is, the vehicle position estimating device 6 uses the first absolute position information of the own vehicle obtained by the satellite positioning and the vehicle sensor information obtained from the own vehicle sensor to indicate the position and the direction of the own vehicle on the map. Estimated own vehicle position information is calculated, and based on the first estimated own vehicle position information and one or both of the road feature information included in the map data and the image sensor information obtained from the own vehicle image sensor. Selecting the road feature to be detected, detecting the relative position and relative orientation of the selected road feature from the own vehicle from optical ranging information obtained from the optical ranging sensor of the own vehicle, and performing optical ranging. A processing circuit 70 is provided for calculating second estimated vehicle position information indicating the position and direction of the vehicle on the map based on the relative position and direction of the road feature detected from the information. The processing circuit 70 may be dedicated hardware, or a processor that executes a program stored in a memory (a central processing unit (CPU: Central Processing Unit), a processing device, an arithmetic device, a microprocessor, a microcomputer, It may be configured using a DSP (also called Digital Signal Processor).

When the processing circuit 70 is dedicated hardware, the processing circuit 70 includes, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), and an FPGA (Field-Programmable). Gate Array) or a combination of these. Each function of the components of the vehicle position estimating device 6 may be realized by an individual processing circuit, or the functions may be realized by one processing circuit.

FIG. 7 shows an example of a hardware configuration of the vehicle position estimating device 6 when the processing circuit 70 is configured using the processor 71 that executes a program. In this case, the functions of the components of the vehicle position estimation device 6 are realized by software or the like (software, firmware, or a combination of software and firmware). Software and the like are described as programs and stored in the memory 72. The processor 71 implements the function of each unit by reading and executing the program stored in the memory 72. That is, when the vehicle position estimating device 6 is executed by the processor 71, the vehicle position estimating device 6 uses the vehicle positioning information on the map based on the absolute positioning information of the vehicle obtained by satellite positioning and the vehicle sensor information obtained from the vehicle sensor of the vehicle. Calculating the first estimated own-vehicle position information indicating the position and direction of the vehicle, the first estimated own-vehicle position information, the information on road features included in the map data, and the image sensor of the own vehicle. Based on one or both of the imaging sensor information, a process of selecting a road feature to be detected, and the relative position and relative orientation of the selected road feature from the own vehicle from the optical ranging sensor of the own vehicle. A second estimated own-vehicle position indicating a position and an azimuth of the own vehicle on a map based on a process of detecting from the obtained optical ranging information and a relative position and a relative azimuth of the road feature detected from the optical ranging information. Calculate information Comprising a process of the memory 72 for storing a program that will but executed consequently. In other words, it can be said that this program causes a computer to execute the procedure and method of operation of the components of the vehicle position estimation device 6.

Here, the memory 72 is, for example, a non-volatile or non-volatile memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), and an EEPROM (Electrically Erasable Programmable Read Only Memory). Volatile semiconductor memory, HDD (Hard Disk Drive), magnetic disk, flexible disk, optical disk, compact disk, mini disk, DVD (Digital Versatile Disc) and its drive device, or any storage medium used in the future. You may.

The configuration in which the functions of the components of the vehicle position estimation device 6 are realized by one of hardware and software has been described above. However, the present invention is not limited to this, and some components of the vehicle position estimating device 6 may be realized by dedicated hardware, and another part of the components may be realized by software or the like. For example, for some components, the function is realized by a processing circuit 70 as dedicated hardware, and for other components, the processing circuit 70 as a processor 71 executes a program stored in a memory 72. The function can be realized by reading and executing.

As described above, the vehicle position estimating apparatus 6 can realize the above-described functions by hardware, software, or the like, or a combination thereof.

In the present invention, the embodiments can be appropriately modified and omitted within the scope of the present invention.

Although the present invention has been described in detail, the above description is illustrative in all aspects, and the present invention is not limited thereto. It is understood that innumerable modifications that are not illustrated can be assumed without departing from the scope of the present invention.

1) satellite positioning device, 2) vehicle sensor information output device, 3) imaging sensor device, 4) optical ranging sensor device, 5) high-accuracy map database, 6) vehicle position estimating device, 61) vehicle position estimating section, 62) road feature position / direction detection , 70 processing circuit, 71 processor, 72 memory.

Claims (4)

Based on absolute positioning information of the vehicle obtained by satellite positioning and vehicle sensor information obtained from a vehicle sensor of the vehicle, first estimated vehicle position information indicating the position and direction of the vehicle on a map is calculated. A vehicle position estimating unit,
A road to be detected based on the first estimated vehicle position information and one or both of road feature information included in the map data and image sensor information obtained from an image sensor of the vehicle. Selecting a feature, detecting the relative position and relative orientation of the selected road feature from the own vehicle from optical ranging information obtained from the optical ranging sensor of the own vehicle, and detecting the relative location and relative orientation from the optical ranging information. Based on the detected relative position and relative orientation of the road feature, a road feature position / azimuth detection unit that calculates second estimated own vehicle position information indicating the position and orientation of the own vehicle on the map,
A vehicle position estimating device comprising: The road feature position / azimuth detection unit sets the range of the optical ranging information to be the target of the selected road feature detection processing, by using the first estimated own vehicle position information, the information of the road feature, and The vehicle position estimating device according to claim 1, wherein the vehicle position estimating device is limited to a range in which the road feature is estimated to be located based on one or both of the imaging sensor information. The vehicle position estimating unit is configured to calculate the second estimated own vehicle position information and the first estimated own vehicle position corresponding to an acquisition time of the optical ranging information used for calculating the second estimated own vehicle position information. Correcting the latest first estimated own-vehicle position information based on the difference with the information;
The vehicle position estimating device according to claim 1. The vehicle position estimating unit of the vehicle position estimating device, based on the absolute positioning information of the vehicle obtained by satellite positioning and vehicle sensor information obtained from the vehicle sensor of the vehicle, determines the position and orientation of the vehicle on a map. Calculating the first estimated vehicle position information shown in FIG.
The road feature position / azimuth detecting unit of the vehicle position estimating device detects the first estimated own vehicle position information, the information of the road feature included in the map data, and the image obtained from the image sensor of the own vehicle. Based on one or both of the sensor information, select a road feature to be detected,
The road feature position / azimuth detecting unit detects a relative position and a relative orientation of the selected road feature from the own vehicle from optical ranging information obtained from an optical ranging sensor of the own vehicle,
A second estimation indicating the position and orientation of the vehicle on the map based on the relative position and orientation of the road feature detected by the road feature position / azimuth detection unit from the optical ranging information. Calculate own vehicle position information,
Vehicle position estimation method.

Images