JP2019039767A - Position estimating device and automatic driving system - Google Patents

Abstract

Provided is a technique for obtaining position information with accuracy required for automatic driving without using an optical sensor. In S110 to S140, it is determined whether or not the road has a straight line shape. If the road is a straight line shape, line fitting is performed in S150, and if the road is non-linear, curve fitting is performed in S160. Line fitting uses an approximate trajectory that is an approximate line calculated from trajectory data measured during the comparison period and an approximate path that is an approximate line calculated from path data corresponding to the comparison period in the map data. Vector E is calculated. In curve fitting, an error vector E is calculated for each measurement value Xc using the point indicated by the latest measurement value Xc and the corresponding point on the map data. [Selection] Figure 3

Description

  The present disclosure relates to a technique for improving measurement accuracy of a vehicle position in a vehicle that performs automatic driving.

  In automatic driving, it is necessary to detect the position of a vehicle (hereinafter referred to as the host vehicle) where automatic driving is performed with high accuracy with an error of about several centimeters. GPS currently in widespread use has an error of about several meters to a few dozen meters. GPS is an abbreviation for Global Positioning System. For this reason, the positional information detected by GPS cannot be used for automatic driving as it is.

  The following Patent Document 1 discloses an azimuth angle between a feature such as a white line or a sign registered in a map database and a detected feature in an apparatus for estimating the orientation of the host vehicle, which is information necessary for automatic driving or the like. A technique for correcting the position of the host vehicle using the above error is described.

JP 2008-249639 A

However, as a result of detailed studies by the inventors, the following problems have been found from the prior art described in Patent Document 1.
In the prior art, landmarks such as white lines and signs are recognized by an optical sensor such as a camera or LIDAR, and the self-position is estimated by matching the recognition result with a map. LIDAR is an abbreviation for Light Detection and Ranging. That is, the optical sensor is an essential requirement. However, the optical sensor is susceptible to deterioration in detection accuracy due to the influence of weather such as rain or heavy fog, and is one of the factors that hinder the robustness of the system.

  Further, in the conventional technique, in a situation where the optical sensor cannot be used, switching from automatic operation to manual operation can be considered. However, in order for the driver to cope with the switching, a certain grace time is required, and during the grace time, there is a possibility that the automatic operation may be continued in a state where the detection accuracy is greatly reduced.

  It is also conceivable to use RTK-GPS or DGPS that does not require an optical sensor in order to improve the robustness of the device for estimating the vehicle position. RTK-GPS is an abbreviation for Real Time Kinematic GPS, and DGPS is an abbreviation for Differential GPS.

  However, although the RTK-GPS can estimate the position of the vehicle with an error of about several centimeters, it is very expensive and is difficult to mount on a mass-produced vehicle or the like. DGPS has a simple configuration and is inexpensive, and can estimate the position of the vehicle with an error of several meters. However, the accuracy required for automatic driving is insufficient.

  One aspect of the present disclosure provides a technique for obtaining position information with accuracy required for automatic driving without using an optical sensor.

A position estimation device according to one aspect of the present disclosure includes a position measurement unit (11), an error estimation unit (13, 13a), and a position correction unit (14).
The position measurement unit uses the information obtained from the satellite positioning system to measure the position of the vehicle that is the vehicle on which the device is mounted. The error estimation unit includes map data including route data that is a series of position data representing a central point in the lane width direction of each lane belonging to the road, and trajectory data that is a series of measurement results in the position measurement unit. A measurement error in the position measurement unit is estimated by comparison. The position correction unit corrects the measurement result of the position measurement unit using the measurement error estimated by the error estimation unit.

  The error estimation unit includes a line fitting unit (13: S150, 13a: S470), a curve fitting unit (13: S160, 13a: S480), and a switching unit (13: S110 to S140, 13a: S430 to S460). Prepare.

  The line fitting unit is calculated from the approximate trajectory that is an approximate straight line calculated from the trajectory data measured during the preset comparison period and the route data corresponding to the comparison period in the map data by the position measurement unit. The distance along the preset error direction with the approximate path that is an approximate line is calculated as a measurement error. The curve fitting unit calculates a distance between a point indicated by the latest measurement result in the position measurement unit and a point on the map data located in the error direction as a measurement error. The switching unit obtains information for determining the shape of the road on which the vehicle is traveling, and uses a line fitting unit when the road is a straight shape, and uses a curve fitting unit when the road is a non-linear shape. The measurement error is calculated.

  According to such a configuration, when the measurement result in the position measurement unit has a measurement error that shifts by a certain amount in a certain direction, the estimation result of the own vehicle position can be used for automatic driving. The accuracy can be improved.

  Note that the reference numerals in parentheses described in this column and in the claims indicate the correspondence with the specific means described in the embodiment described later as one aspect, and the technical scope of the present disclosure It is not limited.

It is a block diagram which shows the structure of an automatic driving | operation system. It is a block diagram which shows the structure of the position estimation apparatus in 1st Embodiment. It is a flowchart of the error estimation process in 1st Embodiment. It is a flowchart of line fitting. It is a flowchart of curve fitting. It is explanatory drawing regarding a line fitting. It is explanatory drawing regarding curve fitting. It is a graph which illustrates the locus data and route data by DGPS. It is a graph which illustrates the locus | trajectory data and path | route data which were correct | amended in the straight area. 6 is a graph illustrating corrected trajectory data and route data in a section including a non-linear section. It is explanatory drawing which shows the influence when a non-linear area is included in a comparison period. It is a block diagram which shows the structure of the position estimation apparatus in 2nd Embodiment. It is a flowchart of the error estimation process in 2nd Embodiment. It is a block diagram which shows the structure of the position estimation apparatus in 3rd Embodiment. It is a flowchart of the output selection process in 3rd Embodiment. It is a block diagram which shows the structure of the position estimation apparatus in 4th Embodiment. It is a flowchart of the period setting process in 5th Embodiment. It is explanatory drawing regarding a period setting process.

Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.
[1. First Embodiment]
[1-1. Constitution]
The automatic driving system 1 shown in FIG. 1 includes a position estimation device 10, an operation control device 20, and a controlled device group 30.

The position estimation device 10 estimates the position (hereinafter, the own vehicle position) of a vehicle (hereinafter, the own vehicle) on which the automatic driving system 1 is mounted. Details thereof will be described later.
The controlled device group 30 includes a plurality of actuators that drive an engine, a brake, a steering, and the like. Furthermore, to the controlled device group 30, a display device for notifying other vehicles (hereinafter referred to as other vehicles) traveling around the own vehicle about the behavior of the own vehicle, and for notifying the driver of the traveling state of the own vehicle. The display device or the acoustic device may be included.

  The driving control device 20 sets a travel route based on the own vehicle position estimated by the position estimation device 10 and the set destination, and is controlled according to the situation around the own vehicle, the state of the own vehicle, and the like. Automatic operation is realized by controlling the device group 30.

[1-1-1. Position estimation device]
As illustrated in FIG. 2, the position estimation device 10 includes a position measurement unit 11, a map database (hereinafter referred to as map DB) 12, an error estimation unit 13, and a position correction unit 14.

  The position measurement unit 11 includes a GPS receiver 111, a difference data receiver 112, and a DGPS calculation unit 113. GPS is an abbreviation for Global Positioning System. DGPS is an abbreviation for Differential GPS.

  The GPS receiver 111 receives signals from a plurality of GPS satellites and analyzes the received signals to generate GPS position data representing the position of the vehicle. This GPS position data has an error of about several meters to several tens of meters.

  The differential data receiver 112 receives differential data that is wirelessly transmitted from a fixed base station whose installation position is known. The difference data is information necessary for differential positioning (ie, DGPS), and is generated by a fixed base station processing signals received from a plurality of GPS satellites.

  The DGPS calculation unit 113 corrects the GPS position data output from the GPS receiver 111 using the difference data received by the difference data receiver 112, thereby generating DGPS position data (hereinafter, measured value) Xc. . The measured value Xc has an error of several meters or less.

  Here, a measurement error in DGPS will be described. The main causes of measurement error in DGPS are shift error and random error. The shift error is an error caused by the ionosphere or the troposphere that affects the propagation time of the signal from the satellite to the receiver, and the magnitude of the error changes slowly. That is, the measurement error can be considered to be constant when viewed in a limited period. On the other hand, the random error is an error that occurs randomly depending on the characteristics of the receiver, and the average level is sufficiently small with respect to the shift error. Then, as shown in FIG. 8, the DGPS measurement result (that is, the trajectory data) acquired during the period in which the measurement error can be assumed to be constant is shifted by a fixed amount in a fixed direction with respect to the correct route data. Will be. In FIG. 8, the solid line is locus data, and the dotted line is route data.

  Returning to FIG. 2, the map DB 12 stores high-precision map data that can be used for automatic driving. The map data includes at least route data representing the shape of the road and feature data representing the positions of a plurality of features.

  The route data is a series of position data representing the center position in the width direction of the lane. When the road includes a plurality of lanes, route data is prepared for each lane. The individual position data belonging to the route data includes a route direction φ representing the tangential direction of the route in addition to the X coordinate representing the position in the latitude direction and the Y coordinate representing the position in the longitude direction. The X coordinate and Y coordinate are the same in the following description.

  The feature data includes the X coordinate and Y coordinate of each point representing the outlines of a plurality of types of features set in advance. The features include, for example, signs or traffic lights provided on the side or above the road, indications such as white lines or pedestrian crossings drawn on the road surface, and tunnels, bridges, buildings, or steel towers visible from the road. Etc. are included.

  The error estimation unit 13 may include a microcomputer having a CPU 131 and a semiconductor memory (hereinafter, memory 132) such as a RAM or a ROM. Each function of the error estimation unit 13 may be realized by the CPU 131 executing a program stored in a non-transitional tangible recording medium. In this example, the memory 132 corresponds to a non-transitional tangible recording medium that stores a program. Also, by executing this program, a method corresponding to the program is executed. The error estimation unit 13 may include one microcomputer or a plurality of microcomputers.

  The method for realizing the function of the error estimation unit 13 is not limited to software, and part or all of the function may be realized using one or a plurality of hardware. For example, when the function is realized by an electronic circuit that is hardware, the electronic circuit may be realized by a digital circuit, an analog circuit, or a combination thereof.

  The error estimation unit 13 executes at least an error estimation process. The error estimation process is a process for estimating a measurement error included in the measurement value Xc based on the measurement value Xc of the vehicle position generated by the position measurement unit 11 and the map data stored in the map DB 12.

  Here, details of the error estimation processing executed by the error estimation unit 13 will be described with reference to the flowchart of FIG. The error estimation process is repeatedly executed at a preset cycle Tx. This period Tx may be set to the same length as a comparison period Ts described later.

  In S110, the error estimation unit 13 acquires a series of measurement values Xc of the vehicle position measured by the position measurement unit 11 during the comparison period Ts before the current time. Hereinafter, the acquired series of measured values Xc is referred to as trajectory data.

  In S120, the error estimation unit 13 acquires the route data C corresponding to the trajectory data acquired in S110 from the map DB 12. In addition, when there are a plurality of lanes on the running road, route data C for the lane specified by the lane specifying process that is separately executed is acquired. Note that the lane identification process is not a main part of the present disclosure, and thus detailed description thereof is omitted. Here, “corresponding to the trajectory data” means, for example, that the trajectory data exists within the range of the maximum measurement error assumed by the position measurement unit 11.

In addition, the route data C is expressed by equation (1) when it includes M pieces of position data C _{1 to} C _M. Further, assuming that m = 1, 2,..., M, the position data C _m is expressed by equation (2). However, _{x m} values of X-coordinate, the _{y m} value of Y-coordinate, the phi _m is the path direction.

Error estimator 13, in S130, based on the M position data C1～C _M belonging to route data C obtained in S120, the average rate Δφav path direction in the route data C, calculated according to equation (3) .

Here, as the average rate of change Δφav, a value obtained by simply adding the rate of change in the path direction between adjacent position data is used, but a value obtained by dividing the result of addition by M−1 may be used. .

Error estimator 13, in S140, Derutafaiav calculated in S130 is judged whether it is smaller than a preset threshold value phi _TH, the process proceeds to S150 if an affirmative decision, if a negative decision is made in S160 Transition. Note that the threshold φ _TH is set to an upper limit value based on an experiment or the like so that the road shape indicated by the route data C can be regarded as a straight line.

In S150, the error estimator 13 calculates a measurement error by line fitting suitable for the case where the road has a linear shape, and ends this processing.
In S160, the error estimator 13 calculates a measurement error by curve fitting suitable for the case where the road has a non-linear shape, and ends this processing.

That is, in the error estimation process, the process for calculating the measurement error is switched depending on whether the road is a straight line shape or a non-linear shape.
Next, the contents of the line fitting process performed by the error estimation unit 13 in S150 will be described using the flowchart shown in FIG. 4 and the explanatory diagram shown in FIG.

In S210, the error estimator 13 calculates an inclination and an intercept (a1, b1) of an approximate straight line (hereinafter referred to as an approximate trajectory) Y = a1 · X + b1 using each measurement value belonging to the trajectory data.
Error estimator 13, in S220, by using each position data C1～C _M belonging to route data C, the approximate straight line (hereinafter, the approximate path) is calculated Y = a2 · X + b2 the slope and intercept (a2, b2) .

In S230, the error estimator 13 determines whether or not the absolute value of the difference between the inclination a1 of the approximate trajectory and the inclination a2 of the approximate path is smaller than a preset threshold value _aTH. Goes to S240, and if a negative determination is made, goes to S250. Note that the threshold value a _TH is set to an upper limit value at which two approximate lines can be regarded as parallel.

  In S240, the error estimator 13 calculates the error vector E = (0, b1-b2) with the difference b1-b2 between the intercepts of the two approximate lines as the magnitude of the measurement error and the error direction in the Y-axis direction. This process is terminated. Here, although the Y-axis direction is the error direction, the error vector E may be calculated using the direction orthogonal to the approximate locus or approximate path as the error direction.

In S250, the error estimator 13 executes a preset exception process and ends the process.
The exception process is, for example, a process for notifying the driver that an error has occurred, an error vector in which the average distance between two approximate lines is the magnitude of the measurement error, and the direction orthogonal to the approximate locus is the direction of the measurement error A process for obtaining E, a process for obtaining the error vector E again in accordance with the procedure described above after shortening the candidate segment, and the like can be considered.

Next, the curve fitting executed by the error estimation unit 13 in S160 will be described using the flowchart shown in FIG. 5 and the explanatory diagram shown in FIG.
In S310, the error estimation unit 13 starts time measurement.

In S320, the error estimation unit 13 acquires the latest measurement value Xc from the position measurement unit 11.
In S330, the error estimation unit 13 extracts the position data Xs on the map data corresponding to the measurement value Xc acquired in S320. The position data Xs is a point on the lane that is located in the error direction as viewed from the point indicated by the measured value Xc and that the vehicle is traveling. The error direction may be the Y-axis direction or the X-axis direction, or may be a direction orthogonal to the tangential direction of the locus at the point indicated by the measurement value Xc.

In S340, the error estimation unit 13 calculates a difference between the measurement value Xc and the position data Xs as an error vector E.
In S350, the error estimation unit 13 determines whether or not the elapsed time from the start of measurement in S310 has reached the execution period Tx of the error estimation process. If a negative determination is made, the process returns to S320. If an affirmative determination is made, this process ends.

  Returning to FIG. 2, the position correction unit 14 corrects the measurement value Xc by subtracting the error vector E from the measurement value Xc output from the position measurement unit 11, and uses the correction result as the own vehicle position as the driving control device. 20 is output.

[1-2. Operation]
When it is determined that the route data corresponding to the trajectory data measured during the comparison time Ts before the current time is a straight route (hereinafter, a straight section), line fitting is performed. In line fitting, as shown in FIG. 6, an error vector E is calculated based on the locus data. This error vector E is used to correct the measurement value Xc output from the position measurement unit 11 until the next error estimation process is executed.

  When it is determined that the route data is a non-linear route (hereinafter referred to as a non-linear section), curve fitting is performed. In the curve fitting, as shown in FIG. 7, every time the measurement value Xc is output from the position measurement unit 11 until the next error estimation process is executed, the error vector E is real-time based on the measurement value Xc. Is calculated. That is, the error vector E calculated individually for each measured value Xc is used for correcting the measured value Xc.

  For example, FIG. 8 shows a trajectory based on route data obtained from map data and a measurement value output from the position measurement unit 11 when the vehicle travels on a first straight road, a curved road, and a second straight road. It is the result shown on data and a graph. The trajectory data is shifted by a certain amount in a certain direction. From the figure, it can be seen that the magnitude of the measurement error is about 5 m, and the shift direction is the upper left direction.

  FIG. 9 and FIG. 10 show the results of the trace data based on the corrected measurement value output from the position correction unit 14 being superimposed on the route data. FIG. 9 shows a case where the route is a straight section, and FIG. 10 shows a case where the route includes a non-linear section. In either case, the vehicle position is accurately estimated with an error of the cm order over the entire route.

[1-3. effect]
According to the first embodiment described in detail above, the following effects are obtained.
(1a) The error vector E (that is, measurement error) includes a shift error and a random error. In line fitting performed when the road has a straight line shape, a process of averaging past measurement results is performed. ing. For this reason, in the present embodiment, it is possible to estimate the error vector E (that is, the measurement error) with high accuracy in which the variation of the shift error and the random error is suppressed. Further, by using this error vector E, it is possible to estimate the own vehicle position with high accuracy.

  However, in the line fitting, as shown in FIG. 11, if a curve road is included in the comparison period Ts, the estimation accuracy for the shift error, and hence the estimation accuracy of the error vector E, decreases. However, in the present embodiment, when the road is a non-linear shape, switching to curve fitting is performed, and the error vector E is estimated for each measurement value Xc. Deterioration can be suppressed.

(1b) In the present embodiment, it is possible to easily realize the estimation of the position of the vehicle with the accuracy necessary for such automatic driving without using an optical sensor.
[2. Second Embodiment]
[2-1. Difference from the first embodiment]
Since the basic configuration of the second embodiment is the same as that of the first embodiment, differences will be described below. Note that the same reference numerals as those in the first embodiment indicate the same configuration, and the preceding description is referred to.

  In the present embodiment, as shown in FIG. 12, the position estimation device 10a includes a behavior sensor group 15 in addition to the configuration of the position estimation device 10 described in the first embodiment, and an error estimation unit 13a. This is different from the first embodiment in that a part of the error estimation process to be executed is different.

[2-2. Constitution]
The behavior sensor group 15 includes at least a gyro 151 and an odometer 152.
The gyro 151 detects the yaw rate Δψ applied to the own vehicle, and obtains the yaw angle ψ by integrating the yaw rate Δψ.

The odometer 152 detects the speed of the own vehicle (hereinafter, the own vehicle speed) v and the acceleration Δv of the own vehicle.
Hereinafter, ψ, Δψ, v, and Δv output from the behavior sensor group 15 are collectively referred to as behavior information. The behavior information is supplied to the error estimation unit 13a.

[2-3. processing]
An error estimation process executed by the error estimation unit 13a of this embodiment instead of the error estimation process described with reference to FIG. 3 of the first embodiment will be described with reference to the flowchart of FIG.

The error estimator 13a acquires behavior information from the behavior sensor group 15 in S410.
In S420, the error estimation unit 13a sets the length of the comparison period Ts based on the behavior information acquired in S410. Specifically, the comparison period Ts is set to be shorter, for example, as the absolute value | ψ | of the yaw angle is larger, as the absolute value | Δψ | of the yaw rate is larger, or as the vehicle speed v is larger.

In S430, the error estimation unit 13a acquires trajectory data that is a series of measured values of the vehicle position measured by the position measurement unit 11 during the comparison period Ts before the current time.
In S440, the error estimation unit 13a acquires route data C, which is a series of position data corresponding to the trajectory data acquired in S430, from the map DB 12 in the same manner as S120.

Error estimating unit 13a, in S450, similarly to S130, based on the M position data C1～C _M belonging to route data C obtained in S440, and calculates the average rate Δφav road direction in the path data.

In S460, the error estimation unit 13a determines whether or not Δφav calculated in S450 is less than a preset threshold φ _TH and whether the yaw rate Δψ acquired in S410 is less than a preset threshold ψ _TH. If it is determined and an affirmative determination is made, the process proceeds to S470, and if a negative determination is made, the process proceeds to S490. The threshold value THψ is set to the upper limit value of the yaw rate Δψ that can be considered that the host vehicle is traveling straight, based on experiments or the like.

In S470, the error estimator 13 calculates the error vector E by line fitting in the same manner as S150, and ends this processing.
In S480, the error estimator 13 calculates the error vector E by curve fitting in the same manner as S160, and ends this processing.

[2-4. effect]
According to 2nd Embodiment explained in full detail above, there exist the effect (1a)-(1b) of 1st Embodiment mentioned above, and also there exist the following effects.

  (2a) The length of the comparison period Ts is variably set according to the behavior information representing the behavior of the host vehicle, and the comparison period Ts is shortened as the possibility that the host vehicle is traveling in the non-linear section is higher. It is set. For this reason, it is possible to suppress a decrease in the estimation accuracy of the error vector E in a portion where the straight section changes to the non-linear section.

  (2b) The determination of whether to perform line fitting or curve fitting is performed using not only the average change rate Δφav in the road direction but also the yaw rate Δψ. For this reason, the approach from the straight section to the non-linear section can be detected early and the fitting method can be effectively switched.

[3. Third Embodiment]
[3-1. Difference from the first embodiment]
Since the basic configuration of the third embodiment is the same as that of the first embodiment, differences will be described below. Note that the same reference numerals as those in the first embodiment indicate the same configuration, and the preceding description is referred to.

  In the present embodiment, as illustrated in FIG. 14, the position estimation device 10 b includes a feature correction unit 16, an environment acquisition unit 17, and a selection unit in addition to the configuration of the position estimation device 10 described in the first embodiment. 18 is different from the first embodiment.

The feature correction unit 16 includes a periphery monitoring unit 161, a feature observation unit 162, and a correction calculation unit 163.
The surroundings monitoring unit 161 includes an optical sensor whose monitoring range is around the host vehicle. The optical sensor includes at least one of a camera and LIDAR. LIDAR is an abbreviation for Light Detection and Ranging. The camera may be either a monocular camera or a compound eye camera.

The feature observation unit 162 detects various features existing around the host vehicle and stored in the map DB 12 based on the output from the surroundings monitoring unit 161.
The correction calculation unit 163 acquires map position data, which is position data of features corresponding to the observation position data, which is position data of the features detected by the feature observation unit 162, from the map DB 12, and the observation position data and the map The own vehicle position corrected so as to eliminate the deviation from the position data is calculated. The host vehicle position X2 calculated by the correction calculation unit 163 is supplied to the selection unit 18. Note that the feature correction unit 16 uses the measurement value Xc of the position measurement unit 11 as the initial position when calculating the vehicle position X2. The technique of the feature correction unit 16 that corrects the vehicle position using the feature in this way is a technique described in Patent Document 1 of the prior art, and a detailed description thereof is omitted.

  The environment acquisition unit 17 acquires various information about the environment in which the host vehicle is traveling (hereinafter, environment information). The environmental information may be acquired from an external infrastructure by wireless communication, or may be acquired from map information and various sensors mounted on the vehicle. In the environmental information, whether or not the environment is such that the detection capability of the optical sensor can be maintained within the allowable range, and the frequency of the feature is such that the estimation accuracy of the vehicle position by the feature correcting unit 16 can be maintained within the allowable range. Information for determining whether or not the environment is detected and whether or not the environment can receive radio waves from GPS is included.

  Similar to the error estimator 13, the selector 18 includes a microcomputer and executes at least output selection processing. The output selection process is a process of selecting either the output X1 of the position correction unit 14 or the output X2 of the feature correction unit 16 as the output Xo of the position estimation device 10b.

[3-2. Output selection process]
The output selection process executed by the selection unit 18 will be described with reference to the flowchart of FIG.
This process is activated every time the vehicle position is supplied from either the position correction unit 14 or the feature correction unit 16.

In S510, the selection unit 18 acquires environment information from the environment acquisition unit 17.
In S520, the selection unit 18 determines whether or not the vehicle is in an environment where radio waves from GPS satellites can be received based on the acquired environment information. If the determination is affirmative, the selection unit 18 proceeds to S550, and the determination is negative Proceeds to S530. For example, based on the map data and the vehicle position, whether or not the vehicle is traveling through the tunnel determines that the environment is incapable of receiving radio waves if traveling through the tunnel.

  In S530, the selection unit 18 determines whether or not the field of view is good based on the acquired environment information. If the determination is affirmative, the process proceeds to S540. If the determination is negative, the process proceeds to S560. Whether or not the visibility is good is, for example, detecting whether or not the weather (such as rain, fog, or snow) is such that the detection range by the optical sensor is equal to or less than a preset threshold distance. You may judge by. The determination may be made by calculating the actual detection range by analyzing the detection result of the optical sensor.

  In S540, the selection unit 18 determines whether or not the environment in which the feature appears with such a frequency that the processing accuracy in the feature correction unit 16 is within the allowable range based on the acquired environment information. When it judges, it transfers to S550, and when negative, it transfers to S560. The above determination may be made based on, for example, whether or not the vehicle is traveling in an urban area. Specifically, whether or not it is an urban area may be determined from the vehicle position and map data, or may be determined by analyzing the detection result of the optical sensor.

  In S550, the selection unit 18 selects the output X2 from the feature correction unit 16, outputs the output X2 to the operation control device 20, and ends this processing. In other words, it is an environment in which radio waves from GPS cannot be received, or an environment in which visibility is good (that is, an environment in which the performance of the optical sensor can be fully exhibited) and a feature is detected frequently (that is, In the environment in which the capability of the feature correction unit 16 can be sufficiently exhibited, the selection unit 18 selects an estimation result in the feature correction unit 16.

  In S560, the selection unit 18 selects the output X1 from the position correction unit 14, outputs the output X1 to the operation control device 20, and ends this processing. In other words, it is an environment in which radio waves from GPS can be received, and visibility is poor (that is, an environment in which the performance of the optical sensor cannot be fully exhibited) or an environment in which features are detected less frequently (that is, features). In an environment where the ability of the correction unit 16 cannot be sufficiently exhibited, the selection unit 18 selects the estimation result in the position correction unit 14.

[3-3. effect]
According to 3rd Embodiment explained in full detail above, there exist the effect (1a)-(1b) of 1st Embodiment mentioned above, and also there exist the following effects.

  (3a) A method for estimating the position of the own vehicle using DGPS and map data and a method for estimating the position of the own vehicle using the feature are used together, and one of the methods is appropriately selected according to the environmental information. Yes. By using such different methods in combination, it is possible to compensate for each other's weak environments, so it is possible to more effectively suppress a decrease in the estimation accuracy of the vehicle position due to the influence of the surrounding environment.

[4. Fourth Embodiment]
[4-1. Difference from the first embodiment]
Since the basic configuration of the fourth embodiment is the same as that of the first embodiment, differences will be described below. Note that the same reference numerals as those in the first embodiment indicate the same configuration, and the preceding description is referred to.

  In the present embodiment, as illustrated in FIG. 16, the position estimation device 10 c includes the behavior sensor group 15 described in the second embodiment, the feature correction unit 16 and the environment acquisition unit 17 described in the third embodiment. And the point which is provided with both the selection parts 18 differs from 1st Embodiment.

[4-2. effect]
According to 4th Embodiment explained in full detail above, there exists the effect (1a)-(1b) (2a) (2b) (3a) of 1st-3rd embodiment mentioned above.

[5. Fifth Embodiment]
[5-1. Difference from the first embodiment]
Since the basic configuration of the fifth embodiment is the same as that of the first embodiment, differences will be described below. Note that the same reference numerals as those in the first embodiment indicate the same configuration, and the preceding description is referred to.

The present embodiment is different from the first embodiment in that the error estimation unit 13 executes a period setting process for variably setting the comparison period in addition to the error estimation process.
[5-2. Period setting process]
The period setting process executed by the error estimation unit 13 will be described with reference to the flowchart shown in FIG. 17 and the explanatory diagram of FIG. This process is executed, for example, every time the error estimation process ends.

In S610, the error estimating unit 13 initializes the comparison period Ts. Here, the lower limit value of the setting allowable range of the comparison period Ts is used as the initial value of the comparison period Ts.
In S620, the error estimation unit 13 acquires route data for the comparison period Ts in the traveling direction of the host vehicle from the position data on the map data corresponding to the current measurement value Xc.

In S630, the error estimation unit 13 calculates the curvature Kp of the route corresponding to the comparison period Ts based on the acquired route data.
In S640, the error estimator 13 acquires route data corresponding to the preset additional period ΔT after the comparison period Ts.

In S650, the error estimation unit 13 calculates the curvature Kc of the route corresponding to the additional period ΔT based on the acquired route data.
In S660, the error estimation unit 13 determines that the absolute value of the difference between the curvature Kp of the comparison period Ts and the curvature Kc of the addition period ΔT is less than a preset threshold value _KTH , and the setting allowable range of the comparison period Ts. It is determined whether it is less than the upper limit value Tmax. If an affirmative determination is made, it is determined that the route corresponding to the additional period ΔT can be regarded as a straight section, and the process proceeds to S670. On the other hand, if a negative determination is made, it is determined that the route corresponding to the additional period ΔT is approaching the non-linear section, and the process proceeds to S680.

In S670, the error estimation unit 13 updates the length of the comparison period Ts to a length obtained by adding the additional period ΔT to the comparison period Ts, and returns to S630.
In S680, the error estimation unit 13 determines the comparison period Ts as the current set value, and ends this process.

The error estimator 13 uses the determined comparison period Ts to start the next error estimation process, and acquire trajectory data and path data in the error estimation process.
[5-3. effect]
According to the fourth embodiment described above in detail, in addition to the effects (1a) to (1b) of the first embodiment described above, the following effects can be obtained.

  (5a) The comparison period Ts used after the present time is variably set so as to become a longer period as the straight line continues according to the shape of the route traveling after the present time. Therefore, a more stable error vector E can be obtained in the straight section. In addition, at the timing of reaching the non-linear section, an error caused by including the non-linear section in the comparison period Ts can be effectively suppressed.

[6. Other Embodiments]
As mentioned above, although embodiment of this indication was described, this indication is not limited to the above-mentioned embodiment, and can carry out various modifications.

(6a) In the above embodiment, GPS is used, but a satellite positioning system other than GPS may be used.
(6b) In the embodiment described above, the vehicle position obtained by differential positioning is corrected, but the present invention is not limited to this. For example, the technique of the present disclosure can be applied to any positioning method that can obtain a measurement value having a measurement error that is shifted by a certain amount in a certain direction.

  (6c) A plurality of functions of one constituent element in the embodiment may be realized by a plurality of constituent elements, or a single function of one constituent element may be realized by a plurality of constituent elements. . Further, a plurality of functions possessed by a plurality of constituent elements may be realized by one constituent element, or one function realized by a plurality of constituent elements may be realized by one constituent element. Moreover, you may abbreviate | omit a part of structure of the said embodiment. In addition, at least a part of the configuration of the above embodiment may be added to or replaced with the configuration of the other embodiment. In addition, all the aspects included in the technical idea specified from the wording described in the claims are embodiments of the present disclosure.

  (6d) In addition to the position estimation device described above and the automatic driving system including the position estimation device as components, a program for causing a computer to function as the position estimation device, and a non-transitional actual state of a semiconductor memory or the like in which the program is recorded The present disclosure can also be realized in various forms such as a typical recording medium and a position estimation method.

DESCRIPTION OF SYMBOLS 1 ... Automatic driving system, 10, 10a-10c ... Position estimation apparatus, 11 ... Position measurement part, 12 ... Map database, 13, 13a ... Error estimation part, 14 ... Position correction part, 15 ... Behavior sensor group, 16 ... Ground Object correction unit, 17 ... environment acquisition unit, 18 ... selection unit, 20 ... operation control device, 30 ... controlled device group, 111 ... GPS receiver, 112 ... differential data receiver, 113 ... DGPS calculation unit, 131 ... CPU , 132, memory, 151, gyro, 152, odometer, 161, periphery monitoring unit, 162, feature observation unit, 163, correction calculation unit.

Claims (12)

A position measurement unit (11) configured to measure the position of the host vehicle, which is a vehicle equipped with the device, using information obtained from the satellite positioning system;
The position measuring unit is obtained by comparing map data including route data that is a series of position data representing a center point in the width direction of each lane belonging to a road with trajectory data that is a series of measurement results in the position measuring unit. An error estimator (13, 13a) configured to estimate a measurement error at
A position correction unit (14) configured to correct the measurement result of the position measurement unit using the measurement error estimated by the error estimation unit;
With
The error estimator is
Calculated from the approximate trajectory that is an approximate straight line calculated from the trajectory data measured during a preset comparison period and the route data corresponding to the comparison period among the map data in the position measurement unit. A line fitting unit (13: S150, 13a: S470) configured to calculate a distance along a preset error direction between the approximate path which is an approximate line and the measurement error;
A curve fitting unit configured to calculate a distance between a point indicated by the latest measurement result in the position measurement unit and a point on the map data located in the error direction as the measurement error (13: S160, 13a: S480)
Obtain information for determining the shape of the road on which the host vehicle is traveling, and use the line fitting unit when the road has a straight shape, and use the curve fitting unit when the road has a non-linear shape. A switching unit (13: S110 to S140, 13a: S430 to S460) configured to calculate the measurement error;
A position estimation device comprising: The position estimation apparatus according to claim 1,
The error estimation unit is configured to set a direction orthogonal to the approximate locus as the error direction.
Position estimation device. The position estimation apparatus according to claim 1,
The error estimation unit is configured to set a Y-axis direction in a coordinate system used when calculating the approximate straight line as the error direction.
Position estimation device. The position estimation device according to any one of claims 1 to 3, wherein
The correction unit is configured to shift the measurement result in the position measurement unit in the error direction by the measurement error estimated by the error estimation unit.
Position estimation device. The position estimation device according to any one of claims 1 to 4, wherein:
The position measuring unit is configured to use differential positioning;
Position estimation device. The position estimation device according to any one of claims 1 to 5, wherein:
The error estimation unit is configured to set the length of the comparison period according to at least one of the speed, acceleration, yaw angle, and yaw rate of the host vehicle (13a: S410). S420)
Position estimation device. The position estimation device according to any one of claims 1 to 5, wherein:
The error estimation unit sets the comparison period so that the curvature of the path shape indicated by the path data associated with the comparison period is a value within an allowable range that can be regarded as a straight path. A period setting unit (13: S610 to S670) configured in
Position estimation device. The position estimation device according to any one of claims 1 to 7,
A feature correction unit (16) configured to obtain the position of the vehicle using a positional relationship between the feature detected using an optical sensor and the vehicle;
An environment acquisition unit (17) for acquiring environment information representing the traveling environment of the host vehicle;
A selection unit (18) configured to select an output from the position correction unit and the feature correction unit according to the environment information acquired by the environment acquisition unit;
Further comprising
Position estimation device. The position estimation device according to claim 8, wherein
The environment acquisition unit acquires at least information that affects the ability of the optical sensor to detect the feature as the environment information,
The selection unit determines whether or not the detection capability of the optical sensor is maintained within a preset allowable value based on the environment information. If the determination is positive, the feature correction unit And when the determination is negative, the position correction unit is selected (18: S520).
Position estimation device. The position estimation device according to claim 8 or 9, wherein
The environment acquisition unit acquires at least information that affects the appearance frequency of the feature as the environment information,
The selection unit determines whether the appearance frequency of the feature is maintained within a preset allowable value based on the environment information, and if the determination is affirmative, the feature correction unit is When the selection is made and a negative determination is made, the position correction unit is selected (18: S540).
Position estimation device. The position estimation device according to any one of claims 8 to 10,
The environment acquisition unit acquires at least information that affects the accuracy of the measurement result by the measurement unit,
The selection unit determines whether or not the accuracy of the measurement result by the position measurement unit is maintained within a preset allowable value based on the environment information. When the correction unit is selected and a negative determination is made, the feature correction unit is selected (18; S530).
Position estimation device. The position estimation device (10, 10a to 10c) according to any one of claims 1 to 11,
An operation control device (20) that executes one or more controls for each part of the vehicle performed to realize automatic driving using the vehicle position output by the position estimation device;
An automatic driving system comprising: