JP6806891B2 - Information processing equipment, control methods, programs and storage media - Google Patents

Description

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

Conventionally, there has been known a technique of detecting a feature installed around a vehicle by using a radar or a camera and calibrating the position of the own vehicle based on the detection result. 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. Further, Patent Document 2 discloses a vehicle position estimation technique using a Kalman filter.

Japanese Unexamined Patent Publication No. 2013-257742 JP-A-2017-72422

When measuring the position of a feature (landmark) that serves as a reference for position estimation from measurement data from an external sensor such as a rider, the landmark is due to occlusion by other vehicles traveling around the vehicle or trees on the side of the road. In some cases, only a part of the landmark can be measured and the exact landmark position cannot be measured. In this case, if the position is estimated by using the measurement position of the wrong landmark as it is, the wrong position estimation result will be obtained.

The present invention has been made to solve the above problems, and provides an information processing device capable of outputting information on landmark measurement positions in consideration of the possibility of occlusion. The main purpose is that.

The invention according to claim 1 is an information processing apparatus, which is based on a first acquisition unit that acquires a measurement position of the feature based on an output signal of a measurement unit that measures surrounding features, and a map information. Based on the second acquisition unit that acquires the predicted position of the feature, the difference between the predicted position and the measured position at the first time, and the difference between the predicted position and the measured position at the second time. The output unit includes an output unit that outputs information regarding the reliability of the measurement position, and the output unit includes the difference at the first time in the first direction and the output unit at the second time in the first direction. Based on the difference, the reliability in the first direction is determined, and the difference at the first time in the second direction intersecting with the first direction and the difference at the second time in the second direction. Based on the above, the reliability in the second direction is determined.

The invention according to claim 4 is a control method executed by an information processing apparatus, which includes a first acquisition step of acquiring a measurement position of the feature based on an output signal of a measurement unit that measures surrounding features. , The second acquisition step of acquiring the predicted position of the feature based on the map information, the difference between the predicted position and the measured position at the first time, and the predicted position and the measured position at the second time. based on the difference, the possess an output step of outputting information about the reliability of the measurement position, wherein the output step includes: the difference in the first time in the first direction, in the first direction The reliability in the first direction is determined based on the difference in the second time, and the difference in the first time in the second direction intersecting with the first direction and the difference in the second direction. The reliability in the second direction is determined based on the difference at the second time .

The invention according to claim 5 is a program executed by a computer, the first acquisition unit that acquires the measurement position of the feature based on the output signal of the measurement unit that measures the surrounding features, and map information. Based on, the second acquisition unit that acquires the predicted position of the feature, the difference between the predicted position and the measured position at the first time, and the difference between the predicted position and the measured position at the second time. Based on this, the computer is made to function as an output unit that outputs information on the reliability of the measurement position, and the output unit is the difference between the difference at the first time in the first direction and the first in the first direction. The reliability in the first direction is determined based on the difference at two times, the difference at the first time in the second direction intersecting with the first direction, and the second in the second direction. The reliability in the second direction is determined based on the difference in time .

It is a schematic block diagram of a driving support system. It is a figure which showed the position of a vehicle on 2D coordinates. The functional block configuration diagram of the in-vehicle device is shown. Indicates the landmark measurement position according to the presence or absence of occlusion. It is the figure which showed the landmark prediction position and the landmark measurement position in time series when occlusion did not occur. It is the figure which showed the landmark prediction position and the landmark measurement position in time series when the occlusion occurred. It is a flowchart of own vehicle position estimation processing.

According to a preferred embodiment of the present invention, the information processing apparatus is based on the output signal of the measurement unit that measures the surrounding features, the first acquisition unit that acquires the measurement position of the feature, and the map information. Based on the second acquisition unit that acquires the predicted position of the feature, the difference between the predicted position and the measured position at the first time, and the difference between the predicted position and the measured position at the second time. , An output unit for outputting information on the reliability of the measurement position. According to this aspect, the information processing apparatus can suitably output information on the reliability of the measurement position of the feature in consideration of the possibility of occurrence of occlusion.

In one aspect of the information processing apparatus, the output unit lowers the reliability as the amount of change in the difference in the second time with respect to the difference in the first time increases. In general, when occlusion occurs in a feature to be measured, the degree of occlusion changes with the passage of time, so that the difference between the predicted position of the feature and the measurement position of the feature changes. Therefore, according to this aspect, the information processing apparatus can accurately determine the reliability of the measurement position of the feature.

In another aspect of the information processing apparatus, the output unit is based on the difference at the first time in the first direction and the difference at the second time in the first direction. The reliability in the direction is determined, and the difference in the first time in the second direction intersecting with the first direction and the difference in the second time in the second direction are used in the second direction. The reliability is determined. According to this aspect, the information processing apparatus can accurately determine and output the reliability of the measurement position of the feature in each of the first direction and the second direction.

In another aspect of the information processing device, the information processing device is a prediction unit that predicts the self-position and a correction unit that corrects the self-position based on the difference between the predicted position and the measurement position at the current time. The lower the reliability of the correction unit, the smaller the gain for correcting the self-position by the difference. According to this aspect, the information processing apparatus appropriately corrects the self-position according to the reliability of the measurement position of the feature which is the reference of the position estimation, and preferably suppresses the deterioration of the position estimation accuracy when the occlusion occurs. Can be done.

According to another preferred embodiment of the present invention, it is a control method executed by the information processing apparatus, and the measurement position of the feature is acquired based on the output signal of the measurement unit that measures the surrounding feature. 1 acquisition step, a second acquisition step of acquiring the predicted position of the feature based on map information, a difference between the predicted position and the measured position at the first time, and the predicted position and the said at the second time. It has an output step of outputting information on the reliability of the measurement position based on the difference from the measurement position. By executing this control method, the information processing apparatus can suitably output information on the reliability of the measurement position of the feature.

According to another preferred embodiment of the present invention, a first acquisition unit that is a program executed by a computer and acquires a measurement position of the feature based on an output signal of the measurement unit that measures surrounding features. The second acquisition unit that acquires the predicted position of the feature based on the map information, the difference between the predicted position and the measured position at the first time, and the predicted position and the measured position at the second time. The computer functions as an output unit that outputs information on the reliability of the measurement position based on the difference between the above. By executing this program, the computer can suitably output information on the reliability of 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. Incidentally, "^" or on any symbol - a letter is attached, for convenience in this specification, "A ^{^"} or "A ^-", "" expressed as ( "A" is an arbitrary character).

[Outline configuration]
FIG. 1 is a schematic configuration diagram of a driving support system according to this embodiment. The driving support system shown in FIG. 1 includes an in-vehicle device 1 mounted on the vehicle and controlling the driving support of the vehicle, and an external sensor such as a lidar (Light Detection and Ranking) 2 or a Laser Illuminated Detection And Ringing (2). , A gyro sensor 3, a vehicle speed sensor 4, a satellite positioning sensor 5, and other internal sensors.

The on-board unit 1 is electrically connected to an external sensor such as a rider 2 and an internal sensor such as a gyro sensor 3, a vehicle speed sensor 4, and a satellite positioning sensor 5, and the on-board unit 1 is mounted based on these outputs. The position of the vehicle (also called "own vehicle position") is estimated. Then, the vehicle-mounted device 1 performs automatic driving control of the vehicle based on the estimation result of the position of the own vehicle. In this embodiment, the on-board unit 1 calculates the measurement position of the landmark Ltag based on the measurement data by the external sensor such as the rider 2 for the feature (also referred to as “landmark Ltag”) which is the reference in the position estimation. .. In this case, the vehicle-mounted device 1 generates information regarding the reliability (also referred to as “reliability R”) of the measurement position of the landmark Ltag. Landmark Ltag is, for example, features such as kilometer posts, 100m posts, delineators, transportation infrastructure facilities (for example, signs, direction signs, signals), utility poles, and street lights that are periodically arranged on the side of the road.

The rider 2 emits a pulsed laser to a predetermined angle range in the horizontal and vertical directions to discretely measure 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 rider 2 has an irradiation unit that irradiates the laser beam while changing the irradiation direction, a light receiving unit that receives the reflected light (scattered light) of the irradiated laser light, and scan data based on the light receiving signal output by the light receiving unit. It has an output unit that outputs. 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-mentioned received signal. The rider 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 in-vehicle device 1 and the like. In addition to the road data, the advanced map DB 10 stores landmark information which is information about a feature that becomes a landmark Ltag. The landmark information is used for the vehicle position estimation process and the reliability R determination process by the vehicle-mounted 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 is applicable is not limited to the configuration shown in FIG. For example, in the driving support system, instead of having the vehicle-mounted device 1, the electronic control device of the vehicle may execute the processing of the vehicle-mounted device 1. In another example, the vehicle-mounted device 1 may store the information corresponding to the advanced map DB 10 by itself instead of acquiring the landmark information from the server device 7.

[Overview of own vehicle position estimation process]
First, the outline of the own vehicle position estimation process in this embodiment will be described. The vehicle position estimation that reflects 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 two-dimensional coordinates. As shown in the figure, there is a vehicle equipped with the on-board unit 1 in the map coordinate system (X _m , Y _m ), and the vehicle coordinate system (X _v , Y _v ) is defined with reference to the position of the vehicle. Specifically, the traveling direction of the vehicle is defined as the " _Xv axis" of the vehicle coordinate system, and the direction perpendicular to it is defined as the " _Yv axis" of the vehicle coordinate system.

In FIG. 2, there is a flat plate-shaped feature that serves as a landmark Ltag on the left side of the vehicle. The position of the landmark Ltag in the map coordinate system (also referred to as “landmark map position”) is included as a part of the landmark information registered in the advanced map DB10. In FIG. 2, the landmark map position is (mx _m , my _m ), and the provisional vehicle position in the map coordinate system predicted based on the output of the internal sensor such as the gyro sensor 3 (“predicted vehicle position”). (Also called) is given by (x ⁻ _m , y ⁻ _m ), and the provisional vehicle orientation angle (also called “predicted vehicle orientation angle”) in the map coordinate system is given by “Ψ ⁻ _m ”. ..

Here, the on-board unit 1 sets the coordinates (L ^- x _v , L ^- y _v ) of the vehicle coordinate system of the tentative predicted position of the landmark (also referred to as "landmark predicted position") to the landmark map position. It is calculated by the following equation (1) using (mx _m , my _m ), the predicted vehicle position (x ⁻ _m , y ⁻ _m ), and the predicted vehicle azimuth Ψ ⁻ _m .

また、車載機１は、ランドマーク予測位置に基づいて、ランドマークが存在すると予測される範囲（「ランドマーク予測範囲ＷＬ」とも呼ぶ。）を決定する。ランドマーク予測範囲ＷＬは、例えば、ランドマーク予測位置から所定距離以内の範囲に設定される。なお、車載機１は、予測自車位置の誤差情報を取得可能な場合、当該誤差情報が示す誤差の大きさに応じて、ランドマーク予測範囲ＷＬの大きさを変更してもよい。この場合、車載機１は、予測自車位置の誤差が大きいほどランドマーク予測範囲ＷＬを大きくすることでランドマークＬｔａｇの検出漏れを防ぐことができ、予測自車位置の誤差が小さいほどランドマーク予測範囲ＷＬを小さくすることでノイズの影響の低減及び処理量低減を実現することができる。同様に、車載機１は、外界センサの計測精度（例えばライダ２の角度分解能）に応じてランドマーク予測範囲ＷＬの大きさを変更してもよい。

Further, the vehicle-mounted device 1 determines a range in which the landmark is predicted to exist (also referred to as “landmark prediction range WL”) based on the landmark prediction position. The landmark prediction range WL is set, for example, within a predetermined distance from the landmark prediction position. If the vehicle-mounted device 1 can acquire the error information of the predicted own vehicle position, the size of the landmark prediction range WL may be changed according to the magnitude of the error indicated by the error information. In this case, the on-board unit 1 can prevent omission of detection of the landmark Ltag by increasing the landmark prediction range WL as the error of the predicted own vehicle position is larger, and the smaller the error of the predicted own vehicle position is, the more the landmark is. By reducing the prediction range WL, it is possible to reduce the influence of noise and the amount of processing. Similarly, the vehicle-mounted 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 rider 2).

Next, the vehicle-mounted device 1 extracts scan data measured as feature points within the set landmark prediction range WL as data for measuring landmarks (also referred to as “landmark measurement data WLD”). For example, when the on-board unit 1 measures a sign or a signboard having higher retroreflectivity than other features by the rider 2 as a landmark Ltag, the reflection intensity is equal to or higher than a predetermined degree from the point group data of the measured measurement points. The data of the measurement points to be used is extracted as the landmark measurement data WLD. Then, the in-vehicle device 1 calculates the center of gravity (average) of the coordinates indicated by each measurement point of the landmark measurement data WLD in the vehicle coordinate system as the landmark measurement position (Lx _v , Ly _v ). In the example of FIG. 2, there is a deviation between the landmark predicted position and the landmark measurement position due to the error of the predicted own vehicle position.

Then, the in-vehicle device 1 calculates the own vehicle position (also referred to as "estimated own vehicle position") obtained by correcting the predicted own vehicle position based on the difference between the landmark predicted position and the landmark measurement position. For example, when the vehicle-mounted device 1 estimates its own vehicle position using the extended Kalman filter, assuming that the reference time (that is, the current time) is "t" and the Kalman gain is "K (t)", the following equation (2) ), The state variable vector "X ^{^} (t) = (x ^{^} _m (t), y ^{^} _m (t), Ψ ^{^} _m (t)) ^T " indicating the estimated vehicle position is calculated.

ここで、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となる。そして、車載機１は、予測自車位置を算出する予測ステップと、直前の予測ステップで算出した予測自車位置を式（２）に基づき補正して推定自車位置を算出する計測更新ステップとを交互に実行する。なお、これらのステップで用いる状態推定フィルタは、ベイズ推定を行うように開発された様々のフィルタが利用可能であり、拡張カルマンフィルタに限定されず、例えばアンセンテッドカルマンフィルタ、パーティクルフィルタなどであってもよい。

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 vehicle-mounted device 1 has a prediction step of calculating the predicted own vehicle position and a measurement update step of correcting the predicted own vehicle position calculated in the immediately preceding prediction step based on the equation (2) to calculate the estimated own vehicle position. Are executed alternately. As the state estimation filter used in these steps, various filters developed for Bayesian estimation can be used, and the filter is not limited to the extended Kalman filter, and may be, for example, an unsented Kalman filter or a particle filter. ..

Here, if occlusion occurs due to a vehicle traveling around the own vehicle or trees on the roadside and only a part of the landmark Ltag can be measured, the landmark measurement position cannot be accurately measured. Then, if the position is estimated based on the equation (2) or the like by using the incorrect landmark measurement position, the accuracy of the estimated own vehicle position is lowered. In consideration of the above, the on-board unit 1 determines the reliability R according to the degree of occlusion with respect to the landmark Ltag, and reflects the determined reliability R in the own vehicle position estimation process to reduce the position estimation accuracy. Suppress. The method of 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 vehicle-mounted device 1. The on-board unit 1 mainly includes the own vehicle position prediction unit 13, the communication unit 14, the landmark information acquisition unit 15, the landmark position prediction unit 16, the landmark feature detection / position measurement unit 17, and the reliability. It has a determination unit 18 and a vehicle position estimation unit 19. The own vehicle position prediction unit 13, the landmark information acquisition unit 15, the landmark position prediction unit 16, the landmark feature detection / position measurement unit 17, the reliability determination unit 18, and the own vehicle position estimation unit 19 are actually Is realized by executing a program prepared in advance by a computer such as a CPU.

The own vehicle position prediction unit 13 uses GNSS (Global Navigation Satellite System) / IMU (Inertial Measurement Unit) combined navigation 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. The own vehicle position is predicted, and the predicted own vehicle position is supplied to the landmark position 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 wireless communication with the server device 7. The landmark information acquisition unit 15 receives the landmark information regarding 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 the position information indicating the landmark map position for each feature that becomes the landmark Ltag.

The landmark position prediction unit 16 uses the above equation (1) to predict the landmark position 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. Is calculated and supplied to the reliability determination unit 18. Further, the landmark position prediction unit 16 determines the landmark prediction range WL based on the landmark prediction position and supplies it to the landmark feature detection / position measurement unit 17. The landmark position prediction unit 16 is an example of the “second acquisition unit” in the present invention.

The landmark feature detection / position measurement unit 17 acquires measurement data from the external sensor 12. Here, the outside world sensor 12 is a sensor that detects an object around the vehicle, and includes a rider 2 and a stereo camera 6. Then, the landmark feature detection / position measurement unit 17 measures the landmark from the measurement data based on the landmark prediction range WL supplied from the landmark position prediction unit 16 and the measurement data acquired from the external sensor 12. Extract the data WLD. Further, 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 calculated landmark measurement position information to the reliability determination unit 18. The landmark feature detection / position measurement unit 17 is an example of the “first acquisition unit” in the present invention.

The reliability determination unit 18 is the difference between the landmark prediction position supplied from the landmark position prediction unit 16 and the landmark measurement position supplied from the landmark feature detection / position measurement unit 17 at the current time t (““ also called difference D (t) ". the), _{X v-axis} of the vehicle coordinate system, is calculated for each of _{Y v} axes. Incidentally, the reliability determination unit 18, X _v-axis of the vehicle coordinate system, in addition to Y _v axes, may calculate the difference D (t) also Z _v coordinates in the height direction. Then, the reliability determination unit 18 calculates the amount of change in the above-mentioned difference D (t) with respect to the difference D (t-1) calculated at the previous time t-1 (also referred to as “difference change amount dD (t)”). , Calculated for each coordinate axis of the vehicle coordinate system. Then, the reliability determination unit 18 determines the reliability R based on the difference change amount dD (t). The determination of the reliability R based on the difference change amount dD (t) will be described later. The reliability determination unit 18 is an example of the “output unit” in the present invention.

The own vehicle position estimation unit 19 calculates the estimated own vehicle position based on the landmark predicted position, the landmark measurement position, the reliability R, and the like. The method of calculating the estimated own vehicle position using the reliability R will be described later. The own vehicle position estimation unit 19 is an example of the “correction unit” in the present invention.

[Judgment of reliability]
Generally, the vehicle-mounted device 1 considers that the larger the difference change amount dD (t) is, the lower the accuracy of the landmark measurement position is, and sets the reliability R according to the difference change amount dD (t).

First, the change in the landmark measurement position depending on the presence or absence of occlusion will be described with reference to FIG. FIG. 4A shows the landmark measurement positions (Lx _v , Ly _v ) when the landmark Ltag does not have occlusion, and FIG. 4B shows the landmark Ltag occlusion by another vehicle. Indicates the landmark measurement position (Lx _v , Ly _v ) when the mark is displayed.

In the example of FIG. 4A, since occlusion does not occur, the entire front surface of the landmark Ltag is the measurement target, and the landmark measurement positions (Lx _v , Ly _v ) are the landmark map positions indicated by the landmark information. Consistent with. Therefore, the vehicle-mounted device 1 can suitably improve the position estimation accuracy by estimating the position of the own vehicle using the landmark measurement position calculated in the case of FIG. 4A.

On the other hand, in the example of FIG. 4B, occlusion by another vehicle has occurred, and only a part of the landmark Ltag is the measurement target. As a result, the landmark measurement position (Lx _v , Ly _v ) is biased to the right side of the landmark Ltag where occlusion does not occur, and the position consistent with the landmark map position (that is, the land in FIG. 4 (A)). It will deviate from the mark measurement position). Therefore, when the vehicle-mounted device 1 estimates its own vehicle position using the landmark measurement position calculated in the case of FIG. 4B, the position estimation accuracy is lowered.

In this way, even when the same landmark Ltag is targeted for measurement, the landmark measurement position changes depending on the presence or absence of occlusion. On the other hand, when occlusion occurs, the relative positional relationship between the vehicle of the vehicle-mounted device 1 and the landmark Ltag and the obstacle that causes occlusion changes with time. Therefore, in this case, the degree of occlusion, the presence or absence of occlusion, and the like change with the passage of time, and the landmark measurement position changes with the passage of time.

Next, the relationship between the change in the landmark measurement position with the passage of time and the difference change amount dD (t) will be described with reference to FIGS. 5 and 6.

FIG. 5 is a diagram schematically showing a landmark predicted position and a landmark measurement position when no occlusion occurs at both time t-1 and time t, and FIG. 6 is a diagram showing occlusion at time t. It is the figure which showed the landmark prediction position and the landmark measurement position at the time of occurrence roughly. In subsequent, the difference D (t) of at _{X v} axis "Dx (t)", _{Y v} in axis difference D (t) the "Dy (t)", _{X v} difference variation in shaft dD (t ) to "DDX (t)", the difference amount of change in _{Y v} axis dD (t) is referred to as "ddY (t)". These values are calculated by the following equation (3) based on their respective definitions.

図５の例では、時刻ｔ−１と時刻ｔのいずれにおいてもオクルージョンが発生していないため、車載機１は、時刻ｔ−１と時刻ｔのいずれにおいても類似した計測点の点群を示すランドマーク計測データＷＬＤに基づきランドマーク計測位置を算出する。その結果、時刻ｔ−１でのランドマーク計測位置（Ｌｘ_ｖ（ｔ−１），Ｌｙ_ｖ（ｔ−１））とランドマーク予測位置（Ｌ⁻ｘ_ｖ（ｔ−１），Ｌ⁻ｙ_ｖ（ｔ−１））の差分（Ｄｘ（ｔ−１）、Ｄｙ（ｔ−１））と、時刻ｔでのランドマーク計測位置（Ｌｘ_ｖ（ｔ），Ｌｙ_ｖ（ｔ））とランドマーク予測位置（Ｌ⁻ｘ_ｖ（ｔ），Ｌ⁻ｙ_ｖ（ｔ））の差分（Ｄｘ（ｔ）、Ｄｙ（ｔ））はほぼ等しくなる。以上のことから、図５の例では、差分変化量ｄＤｘ（ｔ）、ｄＤｙ（ｔ）は、それぞれ０に近い値となる。

In the example of FIG. 5, since occlusion does not occur at both time t-1 and time t, the vehicle-mounted device 1 shows a point cloud of similar measurement points at both time t-1 and time t. The landmark measurement position is calculated based on the landmark measurement data WLD. As a result, the landmark measurement position (Lx _v (t-1), Ly _v (t-1)) and the landmark prediction position (L ^- x _v (t-1), L ^- y _v ) at time t-1. (T-1)) difference (Dx (t-1), Dy (t-1)), landmark measurement position at time t (Lx _v (t), Ly _v (t)) and landmark prediction The differences (Dx (t), Dy (t)) of the positions (L ^- x _v (t), L ^- y _v (t)) are almost equal. From the above, in the example of FIG. 5, the difference change amounts dDx (t) and dDy (t) are values close to 0, respectively.

On the other hand, in the example of FIG. 6, occlusion occurs when the landmark Ltag is measured at time t, and the landmark measurement position (Lx _v (t), Ly _v (t) at time t is caused by the occlusion. )) Is deviated from the landmark measurement position (Lx _v (t-1), Ly _v (t-1)) at time t-1 by a predetermined distance in the _Yv axis direction. In this case, the difference Dx in _{X v-axis} (t), Dx (t- 1) is substantially equal, but the differential variation DDX (t) becomes substantially zero, the difference in _{Y v} axis Dy (t) , Dy (t-1) are different, so the difference change amount dDy (t) is a value for the above-mentioned predetermined distance.

As described above, when the accuracy of the landmark measurement position is lowered due to occlusion, the difference change amount dD (t) corresponding to any axis of the vehicle coordinate system becomes large. In consideration of the above, the vehicle-mounted device 1 considers that the larger the difference change amount dD (t) is, the more the landmark measurement position cannot be measured correctly, and sets the reliability R according to the difference change amount dD (t).

In the examples of FIGS. 5 and 6, the vehicle-mounted device 1 calculated the difference change amounts dDx (t) and dDy (t) for each axis of the vehicle coordinate system, but the present invention is not limited to this, and for example, the following equation ( Based on 4), one difference change amount dD (t) may be calculated with reference to the linear distance.

また、式（３）、（４）では、水平方向の２次元座標系のみを考慮したが、高さ方向の座標をさらに勘案し、式（３）と同様に車両座標系の軸ごとに、又は、式（４）と同様に直線距離を基準として、差分変化量ｄＤ（ｔ）を算出してもよい。

Further, in the equations (3) and (4), only the two-dimensional coordinate system in the horizontal direction is considered, but the coordinates in the height direction are further taken into consideration, and as in the equation (3), for each axis of the vehicle coordinate system, Alternatively, the difference change amount dD (t) may be calculated based on the linear distance as in the equation (4).

Next, an example of setting the reliability R according to the difference change amount dD (t) will be described. In the following, the reliability R is a value from 0 to 1.

In the first setting example of the reliability R, the vehicle-mounted device 1 determines that the reliability R is "1" when the difference change amount dD (t) is equal to or less than a predetermined threshold value, and determines that the difference change amount dD (t). ) Is greater than a predetermined threshold value, it is determined that the reliability R is “0”. The above-mentioned threshold value is set in advance based on an experiment or the like in consideration of, for example, the degree of deterioration of the position estimation accuracy. When the on-board unit 1 calculates the difference change amount dD (t) for each axis of the vehicle coordinate system, when any one of the difference change amounts dD (t) for each axis of the vehicle coordinate system is larger than a predetermined threshold value. , The reliability R may be set to "0". According to the first setting example, the vehicle-mounted device 1 can preferably set the reliability R with respect to the landmark measurement value when occlusion occurs to 0, and suppress a decrease in position estimation accuracy.

When the on-board unit 1 calculates the difference change amount dD (t) for each axis of the vehicle coordinate system, the reliability R may be set for each axis of the vehicle coordinate system. For example, when the vehicle coordinate system and the two-dimensional coordinates, the vehicle-mounted device 1 calculates the reliability R of X _v-axis direction based on a difference variation DDX (t), based on the difference variation ddY (t) Y The reliability R in the _v- axis direction is calculated. In this case, 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 direction of the correction amount reliability R is high, The reliability R can be made relatively larger than the correction amount in the low direction, and the position estimation accuracy can be suitably improved. Similarly, in the following second and third setting examples, the reliability R may be set for each axis of the vehicle coordinate system.

In the second setting example, the vehicle-mounted device 1 is set so that the reliability R approaches 0 as the difference change amount dD (t) increases. In this case, the vehicle-mounted device 1 sets, for example, a value obtained by normalizing the difference change amount dD (t) to a value from 0 to 1 as the reliability R. In the third setting example, the vehicle-mounted 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 sets the reliability R to "0" when the difference change amount dD (t) is larger than the predetermined threshold value, and when the difference change amount dD (t) is equal to or less than the predetermined threshold value, the second The reliability R is set according to the difference change amount dD (t) based on the setting example of.

The reliability R thus obtained is output in association with the calculation result of the landmark measurement position, and is used for estimating the position of the own vehicle. For example, in the own vehicle position estimation, when the own vehicle position is estimated based on the landmark measurement position, weighting is performed based on the reliability R. As a result, the landmark measurement position with low reliability is not used for the own vehicle position estimation or is used with a low weight. In this way, 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 vehicle-mounted device 1 acquires the predicted own vehicle position based on the output of the internal sensor 11 such as the gyro sensor 3 and the vehicle speed sensor 4 (step S101). Then, the vehicle-mounted device 1 acquires landmark information including the position information around the predicted own vehicle position from the advanced map DB 10 (step S102).

Then, the in-vehicle device 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 own vehicle position acquired in step S101 (step S103). .. Then, the vehicle-mounted device 1 acquires measurement data from the external sensor 12 such as the rider 2 (step S104). Then, the vehicle-mounted device 1 extracts the landmark measurement data WLD that satisfies a predetermined condition from the measurement data in the landmark prediction range WL, and calculates the landmark measurement position from the landmark measurement data WLD (step S105).

Next, the vehicle-mounted device 1 calculates the difference D (t) between the landmark measurement position calculated in step S105 and the landmark predicted position calculated based on the equation (1) (step S106). Then, the vehicle-mounted device 1 calculates the difference change amount dD (t) based on the difference D (t-1) and the difference D (t) calculated at the previous time t-1 (step S107). In this case, the vehicle-mounted device 1 may calculate the difference change amount dD (t) for each axis of the vehicle coordinate system.

Then, the vehicle-mounted device 1 determines the reliability R based on the difference change amount dD (t) (step S108). In this case, the vehicle-mounted device 1 determines the reliability R based on the reliability R determination method described in the [Reliability determination] section. At this time, the vehicle-mounted device 1 may determine the reliability R for each coordinate axis of the vehicle coordinate system.

Then, the vehicle-mounted device 1 estimates the position of the own vehicle based on the landmark measurement position, the landmark prediction position, the reliability R determined in step S108, and the like (step S109).

For example, when the vehicle position is estimated using the extended Kalman filter based on the equation (2), the on-board unit 1 sets the reliability R of the _Xv axis to " _RX " and the reliability R of the _Yv axis to "R". _When " _Y " is set, the Kalman gain K (t) is set as shown in the following equation (5).

なお、Ｘ_ｖ軸とＹ_ｖ軸とで特に区別せずに信頼度Ｒを算出する場合には、「Ｒ_Ｘ＝Ｒ_Ｙ＝Ｒ」として式（５）に基づきカルマンゲインＫ（ｔ）を設定するとよい。このように、カルマンゲインＫ（ｔ）を信頼度Ｒ（Ｒ_Ｘ、Ｒ_Ｙ）に基づき調整することで、車載機１は、信頼度Ｒの低いランドマーク計測位置を用いて精度の低い自車位置推定を行うのを好適に防止することができる。

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 (5) as _"R X _{= R} Y = R" It is good to do. By adjusting the Kalman gain K (t) based on the reliability R ( _RX , _RY ) in this way, the on-board unit 1 uses the landmark measurement position having a low reliability R and has a low accuracy. It is possible to preferably prevent the position estimation.

As described above, the landmark feature detection / position measurement unit 17 of the vehicle-mounted device 1 according to the present embodiment is based on the output signal of the external sensor 12 such as the rider 2 that measures the landmark Ltag, which is a surrounding feature. , Calculate the landmark measurement position. In addition, the landmark position prediction unit 16 calculates the landmark prediction position based on the landmark information stored in the advanced map DB 10. Then, the reliability determination unit 18 sets the difference D (t-1) between the landmark predicted position and the landmark measurement position at the previous time t-1 and the landmark predicted position and the landmark measurement position at the current time t. The reliability R is determined based on the difference change amount dD (t) from the difference D (t), and the information regarding the determined reliability R is output to the own vehicle position estimation unit 19. According to this aspect, the vehicle-mounted device 1 can suitably improve the position estimation accuracy.

1 In-vehicle device 2 Rider 3 Gyro sensor 4 Vehicle speed sensor 5 Satellite positioning sensor 7 Server device 10 Advanced map DB

Claims (6)

Based on the output signal of the measurement unit that measures the surrounding features, the first acquisition unit that acquires the measurement position of the feature, and the first acquisition unit.
The second acquisition unit that acquires the predicted position of the feature based on the map information,
An output unit that outputs information on the reliability of the measurement position based on the difference between the predicted position and the measurement position at the first time and the difference between the predicted position and the measurement position at the second time.
Equipped with a,
The output unit determines the reliability in the first direction based on the difference in the first time in the first direction and the difference in the second time in the first direction.
The reliability in the second direction is determined based on the difference at the first time in the second direction intersecting with the first direction and the difference at the second time in the second direction. /> Information processing device. The output unit
The greater the amount of change in the difference in the first direction at the second time with respect to the difference in the first time in the first direction, the lower the reliability in the first direction .
The first aspect of claim 1, wherein the larger the amount of change in the difference in the second direction with respect to the difference in the first time in the second direction, the lower the reliability in the second direction. Information processing equipment. Predictor that predicts your position and
Further provided with a correction unit for correcting the self-position based on the difference between the predicted position and the measurement position at the current time.
The correction unit
The lower the reliability in the first direction, the smaller the gain for correcting the self-position by the difference at the current time in the first direction .
The information processing apparatus according to claim 1 or 2 , wherein the lower the reliability in the first direction, the smaller the gain for correcting the self-position due to the difference at the current time in the first direction . It is a control method executed by the information processing device.
The first acquisition step of acquiring the measurement position of the feature based on the output signal of the measurement unit that measures the surrounding features, and
The second acquisition process to acquire the predicted position of the feature based on the map information,
An output process that outputs information on the reliability of the measurement position based on the difference between the predicted position and the measurement position at the first time and the difference between the predicted position and the measurement position at the second time.
Have a,
The output step determines the reliability in the first direction based on the difference at the first time in the first direction and the difference at the second time in the first direction.
The reliability in the second direction is determined based on the difference at the first time in the second direction intersecting with the first direction and the difference at the second time in the second direction. /> Control method. A program that a computer runs
Based on the output signal of the measurement unit that measures the surrounding features, the first acquisition unit that acquires the measurement position of the feature, and the first acquisition unit.
The second acquisition unit that acquires the predicted position of the feature based on the map information,
As an output unit that outputs information on the reliability of the measurement position based on the difference between the predicted position and the measurement position at the first time and the difference between the predicted position and the measurement position at the second time. a computer to function,
The output unit determines the reliability in the first direction based on the difference in the first time in the first direction and the difference in the second time in the first direction.
The reliability in the second direction is determined based on the difference at the first time in the second direction intersecting with the first direction and the difference at the second time in the second direction. /> Program. A storage medium that stores the program according to claim 5 .

Images