JP6996882B2 - Map data structure of data for autonomous driving support system, autonomous driving support method, and autonomous driving - Google Patents

Description

The present invention relates to an automatic driving support system, an automatic driving support method, a map data structure of data for automatic driving, and the like.

In recent years, a system for automatically driving a vehicle has been developed for the purpose of reducing the driving load of the driver.

For example, in Patent Document 1, a manually driven vehicle is detected by communicating with a peripheral vehicle, and when the manually driven vehicle exists, the driver of the own vehicle is alerted more than when the manually driven vehicle does not exist. A system that changes the mode of automatic operation is described so as to increase the frequency.

Japanese Unexamined Patent Publication No. 2017-30748

By the way, in the automatic driving system as described in Patent Document 1, a highly accurate map is required. As a highly accurate map, for example, it is conceivable to use a public survey map for an automatic driving system, but since the public survey map has a high production cost and there is no association between features, it can be used for an automatic driving system. Not suitable.

Therefore, in view of the above circumstances, it is an object of the present invention to make it possible to provide map data suitable for automatic driving.

The automatic driving support system according to one embodiment of the present invention includes a storage device that stores map data including lane markings, an external sensor that detects peripheral information of the vehicle, peripheral information detected by the external sensor, and map data. Based on, it is equipped with a control unit that controls the running of the vehicle while calculating the relative positional relationship between the vehicle and the lane marking, and the map data is generated by connecting a series of dots through the center of the lane. It comprises a lane network and a plurality of lane marking data associated with the lane network, and among the plurality of lane marking data, any two lane marking data located within a certain range from any reference point on the lane network. With respect to the error in the distance between positions in the horizontal cross section of, the standard deviation of the distribution of the error is within 12.5 cm.

In the present specification and the like, the “part” does not simply mean a physical configuration, but also includes a case where the function of the configuration is realized by software. Further, the function of one configuration may be realized by two or more physical configurations, or the function of two or more configurations may be realized by one physical configuration.

According to the present invention, it is possible to provide map data suitable for automatic driving.

It is a block diagram which shows an example of the structure of an automatic driving support system. It is a figure which shows typically the structure of the map data in one Embodiment of this invention. It is a figure which shows an example of the structure of MMS. It is a schematic diagram which shows that the map data in one Embodiment of this invention can be used for automatic driving.

[Embodiment]
Hereinafter, one of the embodiments of the present invention will be described in detail. It should be noted that the following embodiments are examples for explaining the present invention, and the present invention is not limited to the embodiments thereof. Further, the present invention can be modified in various ways as long as it does not deviate from the gist thereof. Further, those skilled in the art can adopt an embodiment in which each element described below is replaced with an equal one, and such an embodiment is also included in the scope of the present invention. Further, in the following, in order to facilitate understanding, an embodiment in which the present invention is realized by using an information processing apparatus will be described as an example, but as described above, the present invention is not limited thereto.

<1. System overview>
FIG. 1 shows an example of the configuration of the automatic driving support system 1 according to the present embodiment. As shown in FIG. 1, the automatic driving support system 1 is configured by connecting an operation control unit (ECU) 11, a detection unit 12, a storage device 13, and an actuator 14 to each other. The automatic driving support system 1 is mounted on a vehicle such as a car or a train to support the driver's driving. At this time, the operation control unit 11 can communicate with the external system via the network to configure the map data 130 stored in the storage device 13. The storage device 13 may be configured not to be mounted on the vehicle. In this case, the operation control unit 11 executes automatic operation by communicating with the storage device 13 via the network.

The detection unit 12 has a function of acquiring information around the vehicle necessary for automatic driving, and includes, for example, an in-vehicle camera 121, a GPS positioning unit 122, and a sensor 123.

The in-vehicle camera 121 is provided on the back side of the windshield of the vehicle, for example, and photographs the front of the vehicle. Images and moving images taken by the vehicle-mounted camera 121 are transmitted to the operation control unit 11.

The GPS positioning unit 122 receives a signal from a GPS satellite and thereby detects position information (for example, latitude and longitude information) of its own vehicle. The GPS positioning unit 122 transmits the detected position information to the operation control unit 11.

The sensor 123 detects the traveling state of the vehicle and is not shown in the figure, but is composed of, for example, a vehicle speed sensor, an acceleration sensor, a yaw rate sensor, and the like. The vehicle speed sensor is a detector that detects the speed of the vehicle. The acceleration sensor is, for example, a detector that detects the acceleration of the vehicle in the front-rear direction. The yaw rate sensor is a detector that detects the rotational angular velocity around the vertical axis of the center of gravity of the vehicle. The information detected by these vehicle speed sensors, acceleration sensors, and yaw rate sensors is transmitted to the operation control unit 11.

The actuator 14 is provided for controlling the traveling of the vehicle, and is not shown in the figure, but is composed of, for example, an accelerator actuator, a brake actuator, a steering actuator, and the like. The accelerator actuator controls the driving force of the vehicle by controlling the throttle opening degree according to the control signal from the driving control unit 11. The brake actuator controls the braking force on the wheels of the vehicle by controlling both depressions of the brake pedal in response to a control signal from the operation control unit 11. The steering actuator controls the driving of the steering assist motor of the electric power steering system in response to the control signal from the operation control unit 11 to control the steering action of the vehicle.

The map data 130 according to the present embodiment is stored in the storage device 13. The map data 130 includes, for example, three-dimensional position information of a road, information on the shape of a road (for example, curvature of a curve), information on intersections and branch points, information on three-dimensional positions of features such as signs and traffic lights, and the like. Is remembered.

The data structure of the map data 130 according to the present embodiment will be described with reference to FIG. FIG. 2A is a feature data A1 (in the example of FIG. 2A, it is a sign installed on the side of the road), and FIG. 2A is a feature data A2 (in the example of FIG. 2A, it is a signboard on the road). It is a diagram schematically showing the relationship between the feature data AL and AR (which is a lane marking in the example of FIG. 2A) and the lane network A3. As shown in FIG. 2A, in the map data 130, the feature data A1 and A2, the lane marking data AL, and AR are all stored in association with the lane network A3.

First, a method of generating the lane network A3 will be described. The lane network A3 is generated in the section where the lane is defined by the lane marking. The lane network A3 is generated by acquiring the latitude / longitude and altitude of a row of points (hereinafter referred to as “constituent points”) indicating the center of positions separated by lane marking data AL and AR on the road. It is preferable that the constituent points are acquired at intervals of several meters. Further, it is preferable that the start point and the end point of the section in which the lane is defined are constituent points.

Next, the correspondence between the local product data A1, A2, AL, AR and the lane network A3 will be described. First, the lane marking data AL and AR will be described with reference to FIG. 2 (B). As shown in FIG. 2B, among the constituent points constituting the lane network A3, a line segment consisting of two consecutive points (points P1 and P2 in the example of FIG. 2B) from the midpoint to the normal direction. The most recent lane marking at is associated with the lane network A3. On the other hand, feature data A1 and A2 such as signboards and signs are associated with the nearest lane network A3 in the horizontal crossing direction. Examples of features associated with the lane network A3 and stored include road shoulder edges, road markings, road signs, traffic lights, tunnel boundaries, intersection areas, and the like. Since the feature data is stored in association with the lane network A3, the vehicle traveling on the lane network A3 can infer the presence or absence of the feature, the type and position of the feature. It is possible to reduce the load of the image matching process between the image taken by the in-vehicle camera 121 and the feature data, and as a result, the process of estimating the current position of the own vehicle can be efficiently performed. In addition, the vehicle traveling on the lane network A3 has the presence or absence of features that need to be grasped in order to control by automatic driving such as stop lines, intersections, and curve entrances in front of the vehicle in the traveling direction. It is possible to grasp the relative distance to the feature in advance (so-called look-ahead processing of the feature), and it is possible to efficiently grasp the feature by the in-vehicle camera 121. In this way, by using the map data 130 based on the lane network, it becomes possible to perform more accurate automatic driving.

Next, the accuracy of the map data 130 will be described. The accuracy of the map data 130 according to the present embodiment is in the horizontal cross section of any two lane marking data located within a certain range (for example, a range of 10 m before and after in the road) from an arbitrary reference point on the lane network A3. It is defined by the error of the distance between the lane marking data (hereinafter, also referred to as "relative error"). Specifically, the map data 130 according to the present embodiment has an accuracy of 4σ of the relative error distribution within about 50 cm (that is, the standard deviation σ is 12.5 cm). In other words, the probability that the error (relative error) of the distance between the lane marking data A and B on the map data with respect to the actual distance between the lane markings A and B is within 50 cm is 99% or more, 12 It means that the probability of being within .5 cm is 68% or more. By defining the map data 130 according to the present embodiment with such a relative error, more accurate automatic driving becomes possible in the automatic driving support system 1 using the map data 130.

Returning to FIG. 1, the continuation of the configuration of the automatic driving support system 1 will be described.
The operation control unit 11 performs automatic operation with reference to the map data 130 stored in the storage device 13. First, the operation control unit 11 performs map matching on the map data 130. At this time, the driving control unit 11 first detects the latitude and longitude (initial coordinates) of the current position of the own vehicle based on the position information transmitted from the GPS positioning unit 122.

Next, the driving control unit 11 detects the traveling state such as the traveling direction of the own vehicle based on the information detected by the sensor 123, and obtains the traveling locus (estimated locus) of the vehicle from the initial coordinates of the own vehicle position. create. The operation control unit 11 performs map matching based on the initial coordinates and the information of the estimated locus. Further, the operation control unit 11 calculates the relative positional relationship with the feature data by image map matching between the image taken by the in-vehicle camera 121 and the feature data, and estimates the current position of the own vehicle. conduct.

Further, the operation control unit 11 reads out the map data within a predetermined range from the current position of the own vehicle obtained by the above-mentioned image map matching from the storage device 13. Then, the operation control unit 11 refers to the read map data, and when there is a feature such as a stop line, an intersection, a curve entrance, etc. that needs to be controlled by automatic operation in front of the traveling direction of the own vehicle, the operation control unit 11 refers to the read-out map data. By calculating the distance from the own vehicle to the target feature based on the relationship between the positioned position of the own vehicle and the position of the target feature stored in the map data 130, what kind of feature can be obtained. It is possible to know in advance how far ahead it will exist (look-ahead processing of features). The operation control unit 11 actually detects the feature based on the image taken by the in-vehicle camera 121, controls the actuator 14, and sets the lane network in consideration of the calculated distance to the feature. Autonomous driving is executed so that the vehicle runs without protruding from the lane markings (so-called lanes) on both sides along the line.
The operation control unit 11 can be configured by using various conventional techniques.

Next, a method of creating the map data 130 according to the present embodiment will be described.
The map data 130 according to the present embodiment is created by performing a survey using an existing MMS (Mobile Mapping System). The MMS is a system that travels in the measurement area with the measurement vehicle 2 shown in FIG. 3 and measures the three-dimensional coordinate values of the features around the traveled road. As shown in FIG. 3, the measurement vehicle 2 includes a GNSS receiver 21, an IMU (Inertial Measurement Unit) 22, a laser scanner 23, and a camera 24. These devices are attached to a top plate 20 provided on the top of the measuring vehicle 2. Further, an odometer 25 is attached to the wheel of the measuring vehicle 2.

The GNSS receiver 21 has an antenna that receives a positioning signal from a GNSS (Global Navigation Satellite Systems) satellite. The GNSS receiver 21 calculates the pseudo distance from the satellite, the phase of the carrier wave carrying the positioning signal, and the three-dimensional coordinate value based on the positioning signal received by the antenna.

The IMU 22 has a gyro sensor that measures the angular velocity in the three axial directions and an acceleration sensor that measures the acceleration. The IMU 22 acquires the attitude data of its own vehicle by these sensors.

The laser scanner 23 emits laser light while changing the radiation angle in the width direction of the measurement vehicle 2, and receives the laser light reflected by the feature located at the radiation destination. The laser scanner 23 measures the time from when the laser beam is emitted to when the laser beam is received, and calculates the distance to the feature.

The camera 24 takes an image of the front of the measuring vehicle 2. Further, the odometer 25 measures the mileage of the measuring vehicle 2.

Conventionally, GCP (Ground Control Points) is specified based on the positioning data of the GNSS receiver 21, and the data measured by the measuring vehicle 2 is corrected by using the GCP to correct the absolute error of the map data. It improved the accuracy and used it as a public survey map. On the other hand, in the automatic driving system, the map data is used for the purpose of grasping the relative positional relationship between the vehicle and the feature, whether it is the estimation process of the current position of the own vehicle or the look-ahead processing of the feature. Therefore, the accuracy of the relative distance between the features in the map data is more important than the accuracy of the absolute error. The map data 130 according to the present embodiment is created so that the relative error, not the absolute error, falls within a predetermined value (standard deviation σ is about 12.5 cm or less). Therefore, it is possible to omit the process of making corrections based on the observation data received by the GNSS receiver 21, thereby ensuring the accuracy of automatic operation while reducing the cost of creating map data. There is.

Next, with reference to FIG. 4, if the standard deviation σ of the map data is about 12.5 cm or less, the reason why the vehicle can be automatically driven without protruding the lane markings (that is, lanes) on both sides will be described. Under the Road Law, the maximum limit of vehicles that can pass on the road is specified as 2.5 m in width W1 and 12 m in length L1. Further, in the Road Law (Road Structure Ordinance), the minimum value of the width W2 of the expressway (distance between the actual lane markings AL and AR in FIG. 4A) is defined as 3.5 m.

It is assumed that the relative error between the lane marking data AL and AR on the map data is +50 cm (25 cm larger to the left and right of the lane network A3 as the center). In this case, on the map data, a vehicle traveling inside at a predetermined distance (for example, 50 cm) from the positions of the lane marking data AL and AR has an error in the estimation process of the current position of the own vehicle, and the error is caused. If the operation control unit 11 controls the actuator 14 without correction, the vehicle may actually travel 25 cm closer to the lane marking line (for example, the lane marking AR side) on one side. However, when the vehicle is traveling on a straight road (FIG. 4 (A)), the road width W2 (3.5 m) has a margin of 50 cm on each side of the maximum vehicle width W1 (2.5 m). .. Therefore, even if the vehicle is 25 cm closer to the lane marking AR side, it can actually travel without protruding from the lane markings AL and AR. On the other hand, when the vehicle is approaching a curve (FIG. 4 (B)), the center (point 6 m from the tip) M in the length direction is only a predetermined width x cm from the rear end portion B of the vehicle. , The lane markings located inside the curve (the lane marking AR in FIG. 4B) will be closer to the side. Even in this case, in order for the side surface N of the vehicle to travel on the curve without protruding from the lane marking AR, the radius of curvature r of the curve and the above x must satisfy the following equation (1).

(Number 1)
600 ² + (r-x) ² = r ² ... Equation (1)

Here, according to the applicant's investigation on the curvature of the curve of the expressway nationwide, the radius of curvature of the curve of the expressway was 89 m or more. Therefore, if r is 72 m and the above equation (1) is solved for x, x becomes about 25.0 cm. As described above, the road width W2 (3.5 m) with respect to the maximum vehicle width W1 (2.5 m) has a margin of 50 cm on each side. Therefore, about 25 cm, which is the value obtained by subtracting the width of 25.0 cm at which the center M of the vehicle moves toward the inside of the curve on the curve with the minimum curvature from the margin of 50 cm, does not protrude from the lane marking even while traveling on the curve. It can be seen that it is the maximum value of the relative error allowed on one side centering on the lane network A3 for traveling (when converted to both sides, it is about 50 cm).

Since the map data according to the present embodiment has an accuracy that the relative error distribution 4σ is within 50 cm (that is, the standard deviation σ is 12.5 cm), 99% or more of the relative error is the above-mentioned lane network A3. It will be within 25 cm, which is the maximum value allowed on one side centered on. Therefore, it can be determined that the map data according to the present embodiment has sufficient accuracy for automatically driving the vehicle without extending the lane markings (that is, lanes) on both sides.

As described above, when the standard deviation σ of the relative error distribution is within 12.5 cm in the map data according to the present embodiment, the accuracy required for automatic operation can be guaranteed while reducing the generation cost. ..

Each of the embodiments described above is for facilitating the understanding of the present invention, and is not for limiting the interpretation of the present invention. The present invention can be modified / improved without departing from the spirit thereof, and the present invention also includes an equivalent thereof. Further, each embodiment is an example, and it goes without saying that partial substitutions or combinations of the configurations shown in different embodiments are possible, and these are also included in the scope of the present invention as long as the features of the present invention are included. ..

1 System 11 Operation control unit 12 Detection unit 13 Storage device 14 Actuator 130 Map data

Claims (2)

A storage device that stores map data including lane markings,
An external sensor that detects information around the vehicle and
A control unit that controls the running of the vehicle while calculating the relative positional relationship between the vehicle and the lane marking based on the peripheral information detected by the external sensor and the map data.
Equipped with
The map data is
A lane network created by connecting a sequence of dots through the center of the lane,
Multiple lane marking data associated with the lane network and
Equipped with
Of the plurality of lane marking data, the error of the distance between the positions in the horizontal cross section of any two lane marking data located within a certain range from the arbitrary reference point on the lane network is the distribution of the error. Standard deviation is within 12.5 cm,
Automatic driving support system. A step of detecting peripheral information of the vehicle by an external sensor provided in the vehicle,
Based on the map data including the lane markings stored in the storage device and the peripheral information detected by the external sensor, the vehicle travels while calculating the relative positional relationship between the vehicle and the lane markings. Steps to control and
It is an automatic driving support method for vehicles that execute
The map data is
A lane network created by connecting a sequence of dots through the center of the lane,
Multiple lane marking data associated with the lane network and
Equipped with
Of the plurality of lane marking data, the error of the distance between the positions in the horizontal cross section of any two lane marking data located within a certain range from the arbitrary reference point on the lane network is the distribution of the error. Standard deviation is within 12.5 cm,
How to support automatic driving of vehicles.

Images