WO2018212290A1

WO2018212290A1 - Information processing device, control method, program and storage medium

Info

Publication number: WO2018212290A1
Application number: PCT/JP2018/019150
Authority: WO
Inventors: 岩井　智昭; 多史藤谷; 加藤　正浩
Original assignee: パイオニア株式会社
Priority date: 2017-05-19
Filing date: 2018-05-17
Publication date: 2018-11-22

Abstract

A landmark information acquisition unit 15 of an in-vehicle device 1 acquires landmark information stored in a feature-enhanced map DB 10 of a server device 7. A landmark characteristic detection/location measurement unit 17 acquires an output signal from an external sensor 12 such as a lidar 2 for measuring a landmark Ltag, which is a peripheral physical object. In addition, a reliability determination unit 18 outputs, to a vehicle location estimation unit 19, information about the reliability R of the landmark measurement location calculated from the landmark measurement data WLD on the basis of the landmark information and the landmark measurement data WLD extracted from the output signal.

Description

Information processing apparatus, control method, program, and storage medium

The present invention relates to a technique for measuring a position.

2. Description of the Related Art Conventionally, a technique for detecting features installed around a vehicle using a radar or a camera and calibrating the position of the vehicle based on the detection result is known. For example, Patent Document 1 discloses a technique for estimating a self-position by collating the output of a measurement sensor with the position information of a feature registered in advance on a map. Patent Document 2 discloses a vehicle position estimation technique using a Kalman filter.

JP 2013-257742 A JP 2017-72422 A

When measuring the position of a feature (landmark) that serves as a reference for position estimation from measurement data obtained by an external sensor such as a lidar, it is a landmark because it is occluded by other vehicles traveling around the vehicle or trees on the road edge. In some cases, only a part of the area can be measured, and an accurate landmark position cannot be measured. In this case, if position estimation is performed using an erroneous landmark measurement position as it is, an incorrect position estimation result is obtained.

The present invention has been made to solve the above-described problems, and provides an information processing apparatus capable of outputting information related to the measurement position of a landmark in consideration of the possibility of occlusion. The main purpose.

The invention according to claim 1 is an information processing apparatus, wherein a first acquisition unit that acquires feature information stored in a storage unit and an output signal of a measurement unit that measures surrounding features are acquired. And an output unit that outputs information related to the reliability of the measurement position of the feature generated from the output signal based on the feature information and the output signal.

The invention according to claim 7 is a control method executed by the information processing apparatus, the first obtaining step for obtaining the feature information stored in the storage unit, and the output of the measuring unit for measuring surrounding features. A second acquisition step of acquiring a signal; and an output step of outputting information on the reliability of the measurement position of the feature generated from the output signal based on the feature information and the output signal. Have.

The invention according to claim 8 is a program executed by a computer, and acquires an output signal of a first acquisition unit that acquires feature information stored in a storage unit and a measurement unit that measures surrounding features. The computer is caused to function as an output unit that outputs information on the reliability of the measurement position of the feature generated from the output signal based on the second acquisition unit, the feature information, and the output signal. .

It is a schematic block diagram of a driving assistance system. It is the figure which showed the position of the vehicle on the two-dimensional coordinate. The functional block block diagram of a vehicle equipment is shown. Indicates the landmark measurement position according to the presence or absence of occlusion. The distribution of measurement points by a lidar or the like on the measurement surface of the landmark is shown. The positional relationship of landmarks with reference to the vehicle coordinate system is shown. It is a flowchart of the own vehicle position estimation process.

According to a preferred embodiment of the present invention, the information processing apparatus acquires a first acquisition unit that acquires feature information stored in the storage unit, and an output signal of a measurement unit that measures surrounding features. And an output unit that outputs information related to the reliability of the measurement position of the feature generated from the output signal based on the feature information and the output signal. According to this aspect, the information processing apparatus can preferably output information on the reliability of the measurement position of the feature based on the output signal of the measurement unit that measures the surrounding feature.

In one aspect of the information processing apparatus, the feature information includes information indicating a size of the feature, and the output unit includes the size of the feature specified from the feature information, The reliability is determined based on the size of the feature specified from the output signal. According to this aspect, the information processing apparatus can suitably execute the determination of the reliability reflecting the degree of occlusion with respect to the feature.

In another aspect of the information processing apparatus, the output unit includes a horizontal width of the feature, a width in the height direction of the feature, an area of the feature, or the ground measured by the measurement unit. The reliability is determined using at least one of the number of points of the object as an index indicating the size of the feature. According to this aspect, the information processing apparatus can suitably execute the determination of the reliability reflecting the degree of occlusion with respect to the feature.

In another aspect of the information processing apparatus, the feature information includes direction information indicating a direction of the feature, and the output unit is based on the direction of the feature specified from the direction information. The reliability is determined. In a preferred example in this case, the output unit determines the reliability in the first direction based on the size of the feature in the first direction, and the feature in the second direction intersecting with the first direction. The reliability in the second direction may be determined based on the magnitude of. Thereby, the information processing apparatus can accurately determine and output the reliability related to the measurement position of the feature in each of the first direction and the second direction.

In another aspect of the information processing apparatus, the information processing apparatus predicts the feature based on a prediction unit that predicts a self-position, a measurement distance by the measurement unit to the feature, and the feature information. And a correction unit that corrects the self-position based on a difference from the predicted distance until the correction unit decreases the gain for correcting the self-position based on the difference as the reliability decreases. . According to this aspect, the information processing apparatus suitably corrects the self-position according to the reliability of the measurement position of the feature serving as a reference for position estimation, and suitably suppresses a decrease in position estimation accuracy when the occlusion occurs. Can do.

According to another preferred embodiment of the present invention, a control method executed by the information processing apparatus, the first acquisition step of acquiring the feature information stored in the storage unit, and the surrounding features are measured. Based on the second acquisition step of acquiring the output signal of the measurement unit, the feature information, and the output signal, an output that outputs information on the reliability of the measurement position of the feature generated from the output signal And a process. By executing this control method, the information processing apparatus can suitably output reliability information for the measurement position of the feature.

According to another preferred embodiment of the present invention, a program executed by a computer includes a first acquisition unit that acquires feature information stored in a storage unit, and a measurement unit that measures surrounding features. As an output unit that outputs information on the reliability of the measurement position of the feature generated from the output signal based on the second acquisition unit that acquires an output signal, the feature information, and the output signal Make the computer work. By executing this program, the computer can suitably output reliability information for the measurement position of the feature. Preferably, the program is stored in a storage medium.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. For convenience, a character with “^” or “-” attached on an arbitrary symbol is represented as “A ^{^} ” or “A ⁻ ” (“A” is an arbitrary character) in this specification.

[Schematic configuration]
FIG. 1 is a schematic configuration diagram of a driving support system according to the present embodiment. The driving support system shown in FIG. 1 is mounted on a vehicle and controls an on-vehicle device 1 that performs control related to driving support of the vehicle, and an external sensor such as a lidar (Lider: Light Detection and Ranging or Laser Illuminated Detection And Ranging) 2. And an internal sensor such as a gyro sensor 3, a vehicle speed sensor 4, and a satellite positioning sensor 5.

The in-vehicle device 1 is electrically connected to an external sensor such as a lidar 2 and an internal sensor such as a gyro sensor 3, a vehicle speed sensor 4, and a satellite positioning sensor 5, and the in-vehicle device 1 is mounted based on these outputs. The position of the vehicle (also referred to as “own vehicle position”) is estimated. And the vehicle equipment 1 performs automatic driving | operation control etc. of a vehicle based on the estimation result of the own vehicle position. In the present embodiment, the in-vehicle device 1 calculates the measurement position of the landmark Ltag based on the measurement data obtained by an external sensor such as the lidar 2 with respect to the feature (which is also referred to as “landmark Ltag”) used as a reference in position estimation. . In this case, the in-vehicle device 1 generates information on the reliability of the measurement position of the landmark Ltag (also referred to as “reliability R”). The landmark Ltag is, for example, a feature such as a kilometer post, a 100 m post, a delineator, a traffic infrastructure facility (for example, a sign, a direction signboard, a signal), a power pole, a streetlight, and the like periodically arranged along the road.

The lidar 2 emits a pulse laser in a predetermined angle range in the horizontal direction and the vertical direction, thereby discretely measuring the distance to an object existing in the outside world, and a three-dimensional point indicating the position of the object Generate group information. In this case, the lidar 2 includes an irradiation unit that emits laser light while changing the irradiation direction, a light receiving unit that receives reflected light (scattered light) of the irradiated laser light, and scan data based on a light reception signal output by the light receiving unit. Output unit. The scan data is generated based on the irradiation direction corresponding to the laser beam received by the light receiving unit and the response delay time of the laser beam specified based on the above-described received light signal. The lidar 2 is an example of the “measurement unit” in the present invention.

The server device 7 stores an advanced map DB (Data Base) 10 for distribution to the vehicle-mounted device 1 or the like. In the advanced map DB 10, landmark information that is information related to the feature serving as the landmark Ltag is stored in addition to the road data. The landmark information is used for the vehicle position estimation process and the reliability R determination process performed by the in-vehicle device 1.

The configuration of the driving support system shown in FIG. 1 is an example, and the configuration of the driving support system to which the present invention can be applied is not limited to the configuration shown in FIG. For example, instead of having the in-vehicle device 1 in the driving support system, the electronic control device of the vehicle may execute the processing of the in-vehicle device 1. In another example, the in-vehicle device 1 may store information corresponding to the advanced map DB 10 by itself instead of acquiring landmark information from the server device 7.

[Outline of the vehicle position estimation process]
First, the outline | summary of the own vehicle position estimation process in a present Example is demonstrated. The vehicle position estimation reflecting the reliability R will be described later.

FIG. 2 is a diagram showing the position of the vehicle-mounted device 1 mounted on the vehicle on a two-dimensional coordinate. As shown in the figure, there is a vehicle equipped with the vehicle-mounted device 1 in the map coordinate system (X _m , Y _m ), and the vehicle coordinate system (X _v , Y _v ) is defined based on the position of the vehicle. Specifically, the traveling direction of the vehicle is defined as “X _v axis” in the vehicle coordinate system, and the direction perpendicular thereto is defined as “Y _v axis” in the vehicle coordinate system.

In FIG. 2, there is a flat feature that becomes the landmark Ltag on the left side of the vehicle. The position of the landmark in the map coordinate system (also referred to as “landmark map position”) is included as part of the landmark information registered in the advanced map DB 10. In FIG. 2, the landmark map position is (mx _m , my _m ), and the provisional vehicle position (“predicted vehicle position”) in the map coordinate system predicted based on the output of the internal sensor such as the gyro sensor 3 is used. Is also given by (x ⁻ _m , y ⁻ _m ), and the provisional vehicle azimuth (also called “predicted vehicle azimuth”) in the map coordinate system is given by “Ψ ⁻ _m ”. .

Here, the in-vehicle device 1 uses the coordinates (L ⁻ x _v , L ⁻ y _v ) of the vehicle coordinate system of the provisional predicted position of the landmark (also referred to as “landmark predicted position”) as the landmark map position. Using (mx _m , my _m ), the predicted host vehicle position (x ⁻ _m , y ⁻ _m ), and the predicted host vehicle azimuth angle Ψ ⁻ _m , the following calculation (1) is performed.

Further, the in-vehicle device 1 determines a range where the landmark is predicted to exist (also referred to as “landmark predicted range WL”) based on the landmark predicted position. The landmark prediction range WL is set to a range within a predetermined distance from the landmark prediction position, for example. In addition, when the error information of the predicted vehicle position can be acquired, the in-vehicle device 1 may change the size of the landmark prediction range WL according to the size of the error indicated by the error information. In this case, the in-vehicle device 1 can prevent a detection failure of the landmark Ltag by increasing the landmark prediction range WL as the error in the predicted vehicle position increases, and the landmark as the error in the predicted vehicle position decreases. By reducing the prediction range WL, it is possible to reduce the influence of noise and reduce the processing amount. Similarly, the in-vehicle device 1 may change the size of the landmark prediction range WL according to the measurement accuracy of the external sensor (for example, the angular resolution of the lidar 2).

Next, the vehicle-mounted device 1 extracts the scan data measured as the feature points within the set landmark prediction range WL as data obtained by measuring the landmarks (also referred to as “landmark measurement data WLD”). For example, when the vehicle-mounted device 1 uses the lidar 2 as a landmark Ltag to measure a sign or signboard having a higher retroreflectivity than other features as a landmark Ltag, the reflection intensity is more than a predetermined degree from the point cloud data of the measured measurement points. Data of measurement points to be extracted as landmark measurement data WLD. The vehicle unit 1 calculates the center of gravity (average) in the vehicle coordinate system of a coordinate indicated by each measurement point landmarks measurement data WLD, as landmarks measurement position (Lx v, _Ly _v). In the example of FIG. 2, there is a difference between the predicted landmark position and the landmark measurement position due to an error in the predicted vehicle position.

The in-vehicle device 1 calculates a vehicle position (also referred to as “estimated vehicle position”) obtained by correcting the predicted vehicle position based on the difference between the landmark predicted position and the landmark measurement position. For example, when the in-vehicle device 1 performs vehicle position estimation using the extended Kalman filter, if the reference time (that is, the current time) is “t” and the Kalman gain is “K (t)”, the following equation (2) ) To calculate the state variable vector “X ^{^} (t) = (x ^{^} _m (t), y ^{^} _m (t), Ψ ^{^} _m (t)) ^T " indicating the estimated host vehicle position.

^{^{_{Here, X - (t) = (}}} x - m (t), y - m (t), Ψ - m (t)) T, Z (t) = (Lx v, Ly v) T, Z - ( t) = (L ⁻ x _v , L ⁻ y _v ) ^T. Then, the in-vehicle device 1 calculates a predicted own vehicle position, and a measurement update step that calculates the estimated own vehicle position by correcting the predicted own vehicle position calculated in the immediately preceding prediction step based on the equation (2). Execute alternately. The state estimation filter used in these steps can use various filters developed to perform Bayesian estimation, and is not limited to the extended Kalman filter, and may be, for example, an unscented Kalman filter, a particle filter, or the like. .

Here, when occlusion occurs due to vehicles traveling around the vehicle or trees on the road edge, and only a part of the landmark Ltag can be measured, the landmark measurement position cannot be measured accurately. And if position estimation based on Formula (2) etc. is performed using an incorrect landmark measurement position, the accuracy of the estimated own vehicle position will be reduced. Considering the above, the in-vehicle device 1 estimates the degree of occlusion with respect to the landmark Ltag based on the landmark information registered in the map DB 10 and the landmark measurement data WLD, and the reliability R according to the degree of occlusion. Determine. The method for determining the reliability R will be described in the section [Determining the reliability].

[Block configuration]
FIG. 3 shows a functional block configuration diagram of the in-vehicle device 1. The in-vehicle device 1 mainly includes a host vehicle position prediction unit 13, a communication unit 14, a landmark information acquisition unit 15, a landmark position / attribute prediction unit 16, a landmark feature detection / position measurement unit 17, A reliability determination unit 18 and a vehicle position estimation unit 19 are provided. The own vehicle position prediction unit 13, the landmark information acquisition unit 15, the landmark position / attribute prediction unit 16, the landmark feature detection / position measurement unit 17, the reliability determination unit 18, and the own vehicle position estimation unit 19 Actually, it is realized by a computer such as a CPU executing a program prepared in advance.

The own vehicle position prediction unit 13 is based on the output of the internal sensor 11 including the gyro sensor 3, the vehicle speed sensor 4, and the satellite positioning sensor 5, and uses the GNSS (Global Navigation Satellite System) / IMU (Inertia Measurement Unit) combined navigation. And the predicted vehicle position is supplied to the landmark position / attribute prediction unit 16. The own vehicle position prediction unit 13 is an example of the “prediction unit” in the present invention.

The communication unit 14 is a communication unit for wirelessly communicating with the server device 7. The landmark information acquisition unit 15 receives landmark information related to the landmark Ltag existing around the vehicle from the server device 7 via the communication unit 14. Here, as shown in FIG. 3, the landmark information includes at least position information indicating the landmark map position and size information for each feature serving as the landmark Ltag. The size information is information indicating the size of the feature. For example, the size information may be information indicating the area of the surface formed on the feature. Information on the width and length of the surface formed on the feature may be used. It may be shown. The landmark information acquisition unit 15 is an example of the “first acquisition unit” in the present invention. The landmark information is an example of “feature information” in the present invention.

Based on the landmark map position indicated by the position information of the landmark information and the predicted vehicle position acquired from the vehicle position prediction unit 13, the landmark position / attribute prediction unit 16 calculates the landmark by the above-described equation (1). A predicted position is calculated and supplied to the reliability determination unit 18. Further, the landmark position / attribute prediction unit 16 determines a landmark prediction range WL based on the landmark prediction position, and supplies the landmark prediction range WL to the landmark feature detection / position measurement unit 17. Further, the landmark position / attribute prediction unit 16 specifies an attribute (for example, a width) related to the size of the landmark Ltag based on the landmark information, and supplies the specified attribute to the reliability determination unit 18.

The landmark feature detection / position measurement unit 17 acquires measurement data from the external sensor 12. Here, the external sensor 12 is a sensor that detects an object around the vehicle, and includes the lidar 2 and the stereo camera 6. Then, the landmark feature detection / position measurement unit 17 calculates the land from the measurement data based on the landmark prediction range WL supplied from the landmark position / attribute prediction unit 16 and the measurement data acquired from the external sensor 12. Mark measurement data WLD is extracted. The landmark feature detection / position measurement unit 17 calculates the landmark measurement position based on the landmark measurement data WLD. Then, the landmark feature detection / position measurement unit 17 supplies the extracted landmark measurement data WLD and the calculated landmark measurement position information to the reliability determination unit 18. The landmark feature detection / position measurement unit 17 is an example of the “second acquisition unit” in the present invention.

The reliability determination unit 18 is based on the attribute relating to the size of the landmark Ltag supplied from the landmark position / attribute prediction unit 16 and the landmark measurement data WLD supplied from the landmark feature detection / position measurement unit 17. The reliability R is determined. The own vehicle position estimation unit 19 calculates an estimated own vehicle position based on the predicted landmark position, the landmark measurement position, the reliability R, and the like. In addition, the determination method of the reliability R by the reliability determination part 18 and the calculation method of the estimated own vehicle position using the reliability R are mentioned later. The reliability determination unit 18 is an example of an “output unit” in the present invention, and the own vehicle position estimation unit 19 is an example of a “correction unit” in the present invention.

[Judgment of reliability]
Next, a method for determining the reliability R will be specifically described.

(1) Reliability Determination Based on Width The in- vehicle device 1 measures the width in the horizontal direction (that is, the horizontal direction) indicated by the size information of the landmark information with respect to the landmark Ltag (also referred to as “map width”) and the measured land. The reliability R is determined based on the horizontal width (also referred to as “measurement lateral width”) of the mark Ltag. Thereby, the vehicle equipment 1 sets suitably the reliability R according to the degree of occlusion.

FIG. 4A shows the landmark measurement position (Lx _v , Ly _v ) when no occlusion has occurred in the landmark Ltag, and FIG. 4B shows that the occlusion of the landmark Ltag is caused by another vehicle. The landmark measurement position (Lx _v , Ly _v ) is shown. Here, “W _M ” represents the map width with respect to the landmark Ltag, and “W _L ” represents the measured width of the landmark Ltag measured by the lidar 2 or the like.

In the example of FIG. 4A, since no occlusion has occurred, the entire front surface of the landmark Ltag is a measurement target, and the map width W _M and the measurement width W _L are substantially the same. In this case, the landmark measurement position (Lx _v , Ly _v ) matches the landmark map position indicated by the landmark information. Therefore, the in-vehicle device 1 can preferably improve the position estimation accuracy by performing the vehicle position estimation using the landmark measurement position calculated in the case of FIG.

On the other hand, in the example of FIG. 4 (B), and occlusion is generated by another vehicle, only a part of the landmark Ltag is a measurement target, than the map width W _M measurement width W _L is shorter. In this case, the landmark measurement position (Lx _v , Ly _v ) is biased to the right of the landmark Ltag where no occlusion has occurred, and is a position that matches the landmark map position (ie, the land in FIG. 4A). It will deviate from the mark measurement position. Therefore, when the vehicle-mounted device 1 performs its own vehicle position estimation using the landmark measurement position calculated in the case of FIG. 4B, the position estimation accuracy decreases.

Thus, if the occlusion has occurred, rather than the map width W _M shorter measurement width W _L, the more the degree of occlusion is high, the width measurement for map width W _M W _L ratio "W _L / W _M "becomes lower. Considering the above, the in-vehicle device 1 determines the reliability R to be a value from 0 to 1 based on the ratio “W _L / W _M ” of the measured width W _L to the map width W _M.

In the first setting example of the reliability R in this case, the in-vehicle device 1 determines that the ratio of the measured width W _L to the map width W _M is equal to or greater than a predetermined threshold “Wth” (for example, 0.8), that is, “W _{When L} / W _M ≧ Wth ”, it is determined that the reliability R is“ 1 ”, and the ratio of the measured width W _L to the map width W _M is less than the threshold value Wth, that is,“ W _L / W _M < In the case of “Wth”, it is determined that the reliability R is “0”. The threshold value Wth is set in advance based on, for example, experiments in consideration of the degree of deterioration in position estimation accuracy.

In the second setting example, the in-vehicle device 1 is set so that the reliability R approaches 1 as the ratio of the measurement width W _L to the map width W _M increases. For example, the in-vehicle device 1 sets the reliability R based on the following equation (3).

In this case, when the map width W _M and the measured width W _L are equal, the reliability R is "1", when the map width W _M relative measurement width W _L is half, the reliability R is "0.5" It becomes.

In the third setting example, the in-vehicle device 1 determines the reliability R based on the combination of the first example and the second example. For example, the vehicle-mounted device 1 is set to "0" the reliability R when the ratio of the measured width W _L for map width W _M is less than the threshold value Wth, measuring the width W _L of the ratio threshold Wth for map width W _M In the above case, the reliability R is set based on the above equation (3).

In addition, the in-vehicle device 1 can be trusted by further considering the vertical width of the landmark Ltag when the vertical width (that is, the width in the height direction) of the landmark landmark Ltag can be acquired from the size information of the landmark information. The degree R may be determined. For example, the in-vehicle device 1 has a vertical width indicated by the landmark information (also referred to as “map vertical width H _M ”) and a vertical width of the measured landmark Ltag (also referred to as “measurement vertical width H _L ”). The lower the ratio “H _L / H _M ” is, the lower the reliability R is set. For example, the in-vehicle device 1 sets the reliability R based on the following equation (4).

In Formula (4), as a result, the vehicle-mounted device 1 has the measurement area “W _L × H _L ” of the measurement surface with respect to the area “W _M × H _M ” of the measurement surface of the landmark Ltag based on the landmark information. The reliability R is determined based on the ratio.

5A to 5C show the distribution of measurement points by the lidar 2 and the like on the measurement target surface of the landmark Ltag. In the example of FIG. 5 (A), because the occlusion has not occurred, the map width W _M and the measured width W _L, map vertical width H _M and the measured vertical width H _L are equal, respectively. Therefore, in this case, the reliability R based on Expression (4) is “1”. In the example of FIG. 5B, occlusion occurs in the width direction of the landmark Ltag, and the measured width W _L is about half of the map width W _M. Therefore, in this case, the reliability R based on the formula (4) is a value (about 0.5) corresponding to the ratio of the measured width W _L to the map width W _M. In the example of FIG. 5C, occlusion occurs in the horizontal width direction and the vertical width direction of the landmark Ltag, the measured horizontal width W _L is about half of the map horizontal width W _M , and the measured vertical width H _L is the map. It is about half of the vertical width H _M. Therefore, in this case, the reliability R based on Equation (4) is about 0.25.

(2) Reliability Determination Based on Number of Measurement Points The in- vehicle device 1 may determine the reliability R based on the number of measurement points measured by the external sensor 12 such as the lidar 2.

In this case, the in-vehicle device 1 first generates occlusion based on the size of the landmark Ltag (for example, the area of the surface to be measured) indicated by the landmark information, the angular resolution of the lidar 2, etc., the distance to the landmark Ltag, and the like. If not, the number of measurement points for the landmark Ltag (also referred to as “ideal point group number N”) is calculated. Further, the in-vehicle device 1 calculates the number of measurement points indicated by the landmark measurement data WLD (also referred to as “measurement point group number n”). And the vehicle equipment 1 sets the reliability R low, so that the ratio "n / N" of the number n of measurement point groups with respect to the number N of ideal point groups is low.

(3) Reliability Determination Based on Direction Information When the landmark information includes the direction information indicating the direction of the landmark Ltag, the vehicle-mounted device 1 may determine the reliability R by further considering the direction information. .

FIG. 6A shows the positional relationship of the landmark Ltag with reference to the vehicle coordinate system. Also, FIG. 6 (B), a normal angle "theta _L" indicating the normal direction of the landmark Ltag that specified on the basis of the landmark measurement data WLD, width in _{X v-axis} of the measuring width _{W L} _{"W LX"} indicating the width _{"W LY"} in _{Y v-axis} of the measurement width _{W L} and respectively. Here, normal angle theta _L indicates the normal direction of angle landmarks Ltag relative to the X _v axis.

In this case, the in-vehicle device 1 first calculates a plane (measurement plane) formed by the measurement point indicated by the landmark measurement data WLD based on regression analysis or the like, and sets a normal angle θ _L indicating the normal direction of the plane. Identify. The in-vehicle device 1 calculates the measurement width W _L that is the width of the measurement plane of the landmark Ltag based on the landmark measurement data WLD, and calculates the widths W _LX and W _LY based on the following formula (5), respectively. Calculate geometrically.

The in-vehicle device 1 also includes the map width W _M indicated by the size information of the landmark information, the azimuth angle “θ _M ” of the landmark Ltag indicated by the azimuth information of the landmark information, and the predicted own vehicle azimuth angle Ψ ⁻ _m . the basis is calculated based on equation (6) below and the width _{"W MY"} in _{Y v-axis} as the width _{"W MX"} in _{X v-axis} of the map width _{W M.}

The vehicle unit 1, the reliability R relative to the X _v-axis, as well as determination based on a map width W _MX and measuring the width W _LX, Y _v reliability R relative to the axis, a map width W _A determination is made based on _MY and the measured lateral width _WLY . For example, the in-vehicle device 1 determines the ratio “W _LX / W _MX ” of the measured width W _LX to the map width W _{MX as} the reliability R of the _Xv axis, and the ratio “W of the measured width W _LY to the map width W _MY _LY / _{W MY} "the judges that the reliability R of _{Y v-axis.}

According to this example, the vehicle-mounted unit 1, it is possible to calculate the respective reliability R in the X _v-axis and Y _v axes, such as in vehicle position estimation processing, the reliability R is higher direction of correction The amount can be made relatively larger than the correction amount in the direction of low reliability R, and the position estimation accuracy can be suitably improved.

The reliability R obtained in this way is output in association with the calculation result of the landmark measurement position, and is used for the vehicle position estimation. For example, in the vehicle position estimation, weighting is performed based on the reliability R when the vehicle position is estimated based on the landmark measurement position. Thereby, the landmark measurement position with low reliability is not used for the vehicle position estimation or is used with low weighting. Thus, it is possible to prevent the vehicle position estimation with low accuracy from being performed based on the landmark measurement position with low reliability.

[Processing flow]
FIG. 7 shows a flowchart executed by the vehicle-mounted device 1 in this embodiment.

First, the in-vehicle device 1 acquires a predicted host vehicle position based on the outputs of the internal sensors 11 such as the gyro sensor 3 and the vehicle speed sensor 4 (step S101). And the vehicle equipment 1 acquires the landmark information in which the positional information in the periphery of the prediction own vehicle position is included from the advanced map DB 10 (step S102).

And the vehicle equipment 1 determines the landmark prediction range WL based on the landmark map position specified from the position information of the acquired landmark information and the predicted host vehicle position acquired in step S101 (step S103). . And the vehicle equipment 1 acquires measurement data from the external sensors 12 such as the lidar 2 (step S104). And the vehicle equipment 1 extracts the landmark measurement data WLD satisfying a predetermined condition from the measurement data within the landmark prediction range WL, and calculates the landmark measurement position from the landmark measurement data WLD (step S105).

Next, the in-vehicle device 1 determines the reliability R based on the landmark measurement data WLD and the size information included in the landmark information (step S106). In this case, the in-vehicle device 1 executes at least one of the determination methods of the reliability R described in the section “Decision of reliability” to determine the reliability R.

Then, the vehicle-mounted device 1 estimates its own vehicle position based on the landmark measurement position calculated in step S105, the landmark predicted position calculated based on the equation (1), and the reliability R determined in step S106. (Step S107).

For example, when performing vehicle position estimation using an extended Kalman filter based on Expression (2), the in-vehicle device 1 sets the reliability R of the X _v axis to “R _X ” and the reliability R of the Y _v axis to “R”. _If “ _Y ”, the Kalman gain K (t) is set as shown in the following equation (7).

Incidentally, when calculating the reliability R without particular distinction between the _{X v-axis} and _{Y v} axes, sets the Kalman gain K (t) based on equation (7) as _"R X _{= R} Y = R" Good. In this way, by adjusting the Kalman gain K (t) based on the reliability R (R _X , R _Y ), the in-vehicle device 1 uses the landmark measurement position with the low reliability R and uses the low-accuracy vehicle. It is possible to suitably prevent the position estimation.

As described above, the landmark information acquisition unit 15 of the vehicle-mounted device 1 according to the present embodiment acquires landmark information stored in the advanced map DB 10 of the server device 7. The landmark feature detection / position measurement unit 17 acquires an output signal of the external sensor 12 such as the lidar 2 that measures the landmark Ltag that is a surrounding feature. Then, the reliability determination unit 18 determines the reliability R of the landmark measurement position calculated from the landmark measurement data WLD based on the landmark information and the landmark measurement data WLD extracted from the output signal. Information is output to the vehicle position estimation unit 19. By this aspect, the vehicle equipment 1 can improve the position estimation accuracy suitably.

DESCRIPTION OF SYMBOLS 1 In-vehicle apparatus 2 Lidar 3 Gyro sensor 4 Vehicle speed sensor 5 Satellite positioning sensor 7 Server apparatus 10 Sophisticated map DB

Claims

A first acquisition unit that acquires feature information stored in the storage unit;
A second acquisition unit that acquires an output signal of a measurement unit that measures surrounding features;
Based on the feature information and the output signal, an output unit that outputs information on the reliability of the measurement position of the feature generated from the output signal;
An information processing apparatus comprising:
The feature information includes information indicating the size of the feature,
The information processing according to claim 1, wherein the output unit determines the reliability based on a size of the feature specified from the feature information and a size of the feature specified from the output signal. apparatus.
The output unit includes at least one of a horizontal width of the feature, a width in the height direction of the feature, an area of the feature, or the number of point clouds of the feature measured by the measurement unit, The information processing apparatus according to claim 2, wherein the reliability is determined using an index indicating a size of the feature.
The feature information includes direction information indicating the direction of the feature,
The information processing apparatus according to any one of claims 1 to 3, wherein the output unit determines the reliability based on a direction of the feature specified from the orientation information.
The output unit determines the reliability in the first direction based on the size of the feature in the first direction, and the second based on the size of the feature in the second direction intersecting with the first direction. The information processing apparatus according to claim 4, wherein reliability in the direction is determined.
A prediction unit for predicting self-position;
A correction unit that corrects the self-position based on a difference between a measurement distance by the measurement unit to the feature and a predicted distance to the feature predicted based on the feature information;
6. The information processing apparatus according to claim 1, wherein the correction unit reduces the gain for correcting the self-position based on the difference as the reliability is lower.
A control method executed by the information processing apparatus,
A first acquisition step of acquiring feature information stored in the storage unit;
A second acquisition step of acquiring an output signal of a measurement unit that measures surrounding features;
An output step of outputting information on the reliability of the measurement position of the feature generated from the output signal based on the feature information and the output signal;
A control method.
A program executed by a computer,
A first acquisition unit that acquires feature information stored in the storage unit;
A second acquisition unit that acquires an output signal of a measurement unit that measures surrounding features;
The program which makes the said computer function as an output part which outputs the information regarding the reliability of the measurement position of the said feature produced | generated from the said output signal based on the said feature information and the said output signal.
A storage medium storing the program according to claim 8.