JP2021157427A - Driving assistance method and driving assistance device - Google Patents

Abstract

To prevent a pedestrian from being avoided by sudden speed reduction by executing an avoidance preparing action beforehand even if there is the pedestrian who blind crosses a crosswalk in a scene that an own vehicle gets closer to the crosswalk.SOLUTION: The present invention relates to a driving assistance method including: determining the presence/absence of a crosswalk WR within a predetermined distance ahead of an own vehicle A; recognizing an aspect of an own lane traffic light AS which is installed correspondingly to the crosswalk WR and an own lane AL, when the presence of the crosswalk WR is determined; recognizing an aspect of a counter lane traffic light BS which is installed correspondingly to the crosswalk WR and a counter lane BL; predicting that there is a possibility for a pedestrian C to cross the crosswalk WR when the aspect of the own lane traffic light AS is recognized to indicate that forward traveling is allowed and the aspect of the counter lane traffic light BS is recognized to indicate that forward traveling is inhibited; and executing an avoidance preparing action of the own vehicle A to avoid the pedestrian C in the case where it is predicted that it is possible for the pedestrian C to cross the crosswalk WR.SELECTED DRAWING: Figure 1

Description

The present disclosure relates to a driving support method and a driving support device in a scene in which the own vehicle approaches toward a pedestrian crossing.

Conventionally, as a route generation method for an autonomous driving vehicle that avoids pedestrians on a pedestrian crossing, information on the surroundings of the own vehicle is acquired, and it is determined whether or not the behavior prediction target person on the road is visually recognizing the traffic light. It is described that the behavior prediction target person is predicted to move in the transverse direction when it is determined that the vehicle is moving (see, for example, Patent Document 1).

JP-A-2018-205965

However, there are cases where the behavior prediction target person (pedestrian) on the road does not visually recognize the traffic light, for example, recognizes that the vehicle on the roadway has stopped and crosses the pedestrian crossing. In the case of this parting-off crossing, there is a problem that it is not possible to predict that a pedestrian will move in the crossing direction depending on whether or not the traffic light is visually recognized.

This disclosure focuses on the above-mentioned problems, and in a scene where the own vehicle approaches the pedestrian crossing, even if there is a pedestrian who crosses the pedestrian crossing, the avoidance preparation operation is executed in advance to suddenly decelerate. The purpose is to prevent pedestrian avoidance.

In order to achieve the above object, the present disclosure is a driving support method by a controller that assists the driving of the own vehicle when the own vehicle approaches the pedestrian crossing, and the pedestrian crossing exists within a predetermined distance in front of the own vehicle. Decide whether to do it or not. When the existence of the pedestrian crossing is determined, the traffic light installed corresponding to the pedestrian crossing and the display of the own lane signal corresponding to the own lane in which the own vehicle is traveling is recognized. It is a traffic light installed corresponding to a pedestrian crossing, and recognizes the indication of the oncoming lane traffic light corresponding to the oncoming lane in which the oncoming vehicle travels toward the own vehicle. When it is recognized that the display of the own lane traffic light indicates that the vehicle can proceed and the display of the oncoming lane traffic light indicates that the vehicle cannot proceed, it is predicted that a pedestrian may cross the pedestrian crossing. When it is predicted that a pedestrian may cross the pedestrian crossing, the avoidance preparation action of the own vehicle to avoid the pedestrian is executed.

Since the above-mentioned problem-solving means is adopted, in a scene where the own vehicle approaches the pedestrian crossing, even if there is a pedestrian who crosses the crosswalk in advance, the avoidance preparation operation is executed in advance to prevent pedestrian avoidance due to sudden deceleration. be able to.

FIG. 5 is a block configuration diagram showing a detailed configuration of an automatic driving control system configuration and an action prediction unit of an autonomous driving vehicle to which the driving support method and the driving support device of the first embodiment are applied. It is a pedestrian crossing prediction explanatory diagram which shows an example of the pedestrian crossing approach scene of own vehicle which needs the pedestrian crossing prediction. It is a task explanatory diagram which shows the problem when a pedestrian crosses a pedestrian crossing in the example of the pedestrian crossing approach scene in which the own vehicle approaches a pedestrian crossing by the pedestrian crossing prediction by the background technique. It is a driving support action explanatory diagram which shows the avoidance preparation operation in advance in the example of the pedestrian crossing approach scene in which the own vehicle approaches a pedestrian crossing by the pedestrian crossing prediction in this disclosure. It is a flowchart which shows the flow of the driving support processing based on the action prediction executed by the automatic driving control system. It is a flowchart which shows the flow of the pedestrian crossing prediction processing in the pedestrian crossing approach scene that the own vehicle approaches toward a pedestrian crossing executed by a behavior prediction unit.

Hereinafter, a mode for implementing the driving support method and the driving support device according to the present disclosure will be described with reference to the first embodiment shown in the drawings.

When the automatic driving mode is selected, the driving support method and the driving support device in the first embodiment are automatic driving vehicles (driving support vehicles) in which the vehicle motion is controlled by the speed and the steering angle so as to travel along the generated target route. It is applied to one example). Hereinafter, the configuration of the first embodiment will be described separately for "automatic driving control system configuration" and "detailed configuration of the behavior prediction unit".

[Automatic driving control system configuration (Fig. 1)]
As shown in FIG. 1, the autonomous driving vehicle (own vehicle) includes an object detection device 1, an object detection integration / tracking unit 2, a vehicle position estimation device 3, a map storage device 4, and a vehicle position estimation unit 5 in the map. The behavior prediction unit 6, the vehicle route generation unit 7, and the vehicle control unit 8 are installed as an automatic driving control system.

The object detection device 1 uses an in-vehicle sensor that detects an object such as a laser radar, a millimeter-wave radar, a rider, and a camera, and uses an in-vehicle sensor (for example, another vehicle, a motorcycle, a bicycle, a pedestrian, an obstacle) existing around the own vehicle. Etc.) position, posture, size, speed, etc. are detected. The detection result represents, for example, a two-dimensional position, posture, size, speed, etc. of an object in a zenith map in which the own vehicle is viewed from the air.

The object detection integration / tracking unit 2 outputs one two-dimensional position, posture, size, speed, etc. for each object based on the object detection result obtained from the object detection device 1. That is, based on the results of a plurality of object positions, postures, sizes, and velocities obtained from a plurality of in-vehicle sensors, the error characteristics of each in-vehicle sensor are taken into consideration by sensor fusion, etc., and the error such as the object position is the smallest. Calculate one rational two-dimensional position or the like. Furthermore, the identity verification (association) of the object between different times is performed for the object position, posture, size, velocity, etc. output at different times, and the velocity of the object is based on the association. Estimate the information.

The own vehicle position estimation device 3 uses sensors that measure absolute positions such as GPS (abbreviation of "Global Positioning System") and odometry to determine the absolute position of the own vehicle, that is, the position, posture, speed, etc. of the own vehicle with respect to a certain reference point. measure.

The map storage device 4 holds high-precision map information including information on the road shape and the traveling lane, and the own vehicle and other vehicles such as the absolute position of the lane, the lane connection relationship, and the relative positional relationship from the high-precision map information. Acquire road map information about the road on which such as travels.

The own vehicle position estimation unit 5 in the map is based on the absolute position of the own vehicle obtained by the own vehicle position estimation device 3 and the road map information around the own vehicle obtained by the map storage device 4 in the road map. Estimate the position of your vehicle in. That is, information such as which lane the own vehicle is traveling in which direction is acquired.

The behavior prediction unit 6 of the own vehicle is based on the object position information obtained by the object detection integration / tracking unit 2 and the position information of the own vehicle in the map obtained by the own vehicle position estimation unit 5 in the map. Performs behavior prediction processing for surrounding objects (for example, other vehicles, motorcycles, bicycles, pedestrians, etc.). In the first embodiment, among various behavior predictions of objects existing around the own vehicle, when the own vehicle approaches the pedestrian crossing, there is a possibility that a pedestrian waiting for a signal at the pedestrian crossing shifts to the pedestrian crossing behavior. Performs pedestrian crossing prediction processing as to whether or not there is. The details will be described later in "Detailed configuration of behavior prediction unit".

The own vehicle route generation unit 7 generates a target route in advance based on the input of the destination by the driver, and also generates a target speed profile representing the target speed when the own vehicle travels along the target route. Has been done. Therefore, when it is predicted by the behavior prediction unit 6 that a pedestrian may cross the pedestrian crossing, at least one of the target route and the target speed profile generated in advance is selected according to the avoidance preparation action. Fix it.

The vehicle control unit 8 has a drive actuator, a braking actuator, and a steering angle so that the traveling speed and steering angle of the own vehicle are set according to the target route and the target speed profile generated (including correction) by the own vehicle route generation unit 7. Vehicle motion control is performed to output a control command to the actuator. Therefore, if it is predicted that a pedestrian may cross the pedestrian crossing and the mode of the avoidance preparation motion is the lane change motion or the deceleration motion, the target route and the target speed profile modified according to the mode are adopted. The avoidance preparation operation is executed by the control output along with it. When the mode of the avoidance preparation operation based on the pedestrian crossing prediction is the brake pressurization application, the target path and the target speed profile generated in advance are maintained as they are, and the brake actuator is instructed to apply the brake pressurization. The output will execute the avoidance preparation operation.

Here, as the avoidance preparatory movement based on the pedestrian crossing prediction, there are the modes described in the following (a) to (d), and any one mode or a combination mode of two or more is used as the avoidance preparatory movement. Select as. In addition, for example, the mode selection of the avoidance preparation operation may be changed based on the driver request, the situation judgment, or the like.
(a) If the road on which the vehicle is traveling has multiple lanes and the vehicle is traveling in a lane other than the lane farthest from the sidewalk, the lane in which the vehicle is traveling is changed to the lane away from the sidewalk. Change operation.
(b) A constant deceleration operation that decelerates the vehicle speed of the own vehicle at a predetermined constant deceleration.
(c) The brake pressurization operation is performed to apply the brake pressurization immediately before the hydraulic braking force is generated to the hydraulic brake device of the own vehicle.
(d) If the vehicle speed at the time when the own vehicle starts the avoidance preparation operation is equal to or higher than the predetermined vehicle speed, the deceleration operation is performed to decelerate from the starting vehicle speed to the target vehicle speed in the predetermined pedestrian crossing reach area.
Further, the avoidance preparation operation may include a notification operation for displaying or sounding a warning to the occupants of the vehicle at the same time as or separately from the above (a) to (d).

[Detailed configuration of behavior prediction unit (Figs. 1 and 2)]
As shown in FIG. 1, the behavior prediction unit 6 includes a pedestrian crossing presence determination unit 601, a own lane signal state recognition unit 602, an oncoming lane signal state recognition unit 603, a pedestrian crossing detection unit 604, and a pedestrian crossing prediction unit. Has 605. It also has a pedestrian face orientation detection unit 606, a pedestrian number measurement unit 607, a time difference type traffic light detection unit 608, an intersection size detection unit 609, and a pedestrian crossing likelihood determination unit 610.

Hereinafter, as shown in FIG. 2, the own vehicle is A, the oncoming vehicle is B, the pedestrian is C, the target route of the own vehicle A is TL, the intersection is CP, the own lane in which the own vehicle A travels is AL, and the oncoming vehicle. The oncoming lane in which B travels is called BL, the own lane signal is called AS, the oncoming lane signal is called BS, and the pedestrian crossing is called WR.

The pedestrian crossing existence determination unit 601 determines whether or not the pedestrian crossing WR exists within a predetermined distance in front of the own vehicle A. Here, the "predetermined distance in front of the own vehicle A" means that the own vehicle A approaches the pedestrian crossing WR by slow deceleration and is reduced to a vehicle speed at which the pedestrian C can avoid the pedestrian C in the vicinity of the pedestrian crossing WR. This is the distance that can be prepared for avoidance. Therefore, the "predetermined distance" is set, for example, in a distance range of about 50 m to 100 m, the higher the vehicle speed of the own vehicle A, the longer the distance, and the lower the vehicle speed, the shorter the distance.

When the existence of the pedestrian crossing WR is determined, the own lane signal state recognition unit 602 is a traffic light installed corresponding to the pedestrian crossing WR and corresponds to the own lane AL in which the own vehicle A is traveling. Recognize the indication of the lane traffic light AS. Here, the display of the own lane signal AS is recognized based on the in-vehicle camera image information of the own vehicle A. If the display information cannot be recognized from the own lane signal AS based on the in-vehicle camera image information, the own vehicle A can acquire the display information of the own lane signal AS by road-to-vehicle communication or vehicle-to-vehicle communication. It is also possible to recognize the indication of the lane traffic light AS.

The oncoming lane signal state recognition unit 603 is a traffic light installed corresponding to the pedestrian crossing WR, and recognizes the display of the oncoming lane signal BS corresponding to the oncoming lane BL in which the oncoming vehicle B heading for the own vehicle A travels. do.

Here, the display of the oncoming lane signal BS is often not recognized by the in-vehicle camera image information of the own vehicle A as compared with the own lane signal AS. Therefore, when the display of the oncoming lane traffic light BS cannot be recognized by the in-vehicle camera image information, any of the following recognition methods is used.
(a) When it is detected that the oncoming vehicle B in the oncoming lane BL is stopped on the lane (see FIG. 2) or decelerated with respect to the stop line, the indication of the oncoming lane traffic light BS indicates that the vehicle is prohibited from proceeding. Recognize that. However, if the front of the oncoming vehicle B in the oncoming lane BL is congested and the oncoming vehicle B passes through the intersection CP and cannot enter ahead of the oncoming lane traffic light BS, the oncoming lane traffic light due to the behavior of the oncoming vehicle B By prohibiting the process itself of recognizing the display of the BS, it is prevented that the oncoming vehicle B is not detected as being stopped or decelerating with respect to the stop line on the lane.
(b) Based on the communication information acquired from the outside by the own vehicle A through road-to-vehicle communication or vehicle-to-vehicle communication, the oncoming lane signal BS recognizes whether the oncoming lane signal BS indicates progress or prohibition. In road-to-vehicle communication, for example, a signal indicating the indication of the oncoming lane signal BS transmitted from a transmitter provided in an infrastructure such as a traffic light is received by a receiver provided in the own vehicle A and received. The indication of the oncoming lane traffic light BS is recognized based on the signal. As such an information communication system, for example, an information communication system called I2V (Invisible-to-Visible) that visualizes what cannot be seen from the own vehicle A is used. In the inter-vehicle communication, the information is acquired by receiving the display information of the oncoming lane traffic light BS transmitted from the oncoming vehicle B that is stopped or decelerating with respect to the stop line on the lane.

The pedestrian detection unit 604 near the pedestrian crossing detects whether or not the pedestrian C exists in the vicinity of the pedestrian crossing WR based on the input information from the object detection integration / tracking unit 2. In particular, as a pedestrian C existing in the vicinity of the pedestrian crossing WR, as shown in FIG. 2, the pillar of the own lane traffic light AS is easier to observe than the display of the own lane traffic light AS. Detects pedestrian C, which is located near the vehicle and has a high possibility of crossing the pedestrian crossing.

When the pedestrian crossing prediction unit 605 recognizes that the display of the own lane signal AS indicates that the progress is possible and the display of the oncoming lane signal BS indicates that the progress is prohibited, the pedestrian C crosses the pedestrian crossing WS. The basic prediction is to predict that there is a possibility. However, even when the display conditions of the own lane signal AS and the oncoming lane signal BS are satisfied, if the presence of the pedestrian C is not detected in the vicinity of the pedestrian crossing WS, the pedestrian who is the target of the behavior prediction. Since C itself does not exist, it is predicted that there is no possibility that pedestrian C will cross the pedestrian crossing WS. That is, in the first embodiment, when the display conditions of the own lane traffic light AS and the oncoming lane traffic light BS are satisfied and the presence of the pedestrian C is detected in the vicinity of the pedestrian crossing WS, the pedestrian C is the pedestrian crossing WR. The prediction base is to predict that there is a possibility of crossing. In addition, the pedestrian crossing likelihood determination unit 610, which will be described later, determines the prediction that the pedestrian C may cross the pedestrian crossing WR when the pedestrian crossing likelihood condition is satisfied.

When the presence of the pedestrian C is detected in the vicinity of the pedestrian crossing WR, the pedestrian face orientation detection unit 606 detects the face orientation of the pedestrian C as a crossing determination item. Then, if the face direction of the pedestrian C is not looking in the direction of the own vehicle A, the pedestrian crossing likelihood is set to a higher value than when looking in the direction of the own vehicle A. Here, as the "face orientation detection of the pedestrian C", for example, the vertical direction of the face is recognized based on the shape of the triangle connecting both eyes and the mouth of the pedestrian C's face, and the pedestrian C's face is detected. A well-known detection method that recognizes the left-right direction of the face based on the distance between the eyes and the eyebrows is used (see Japanese Patent Application Laid-Open No. 2008-186247, etc.), but the face orientation is not limited to this. May be detected.

When the presence of the pedestrian C is detected in the vicinity of the pedestrian crossing WR, the pedestrian number measuring unit 607 measures the number of pedestrians C as a crossing determination item. Then, when the number of pedestrians C is equal to or greater than the predetermined number, the pedestrian crossing likelihood is set to a value higher than when the number of pedestrians C is less than the predetermined number. Here, the "number of pedestrians C" is measured based on the input information from the object detection integration / tracking unit 2. The "predetermined number of people" is set to, for example, two people from the viewpoint that when there are two or more pedestrians C near the boundary between the sidewalk and the pedestrian crossing WR, the possibility of crossing the parting is high.

The time difference type traffic light detection unit 608 detects whether or not the own lane signal AS and the oncoming lane signal BS are time difference type traffic lights as a crossing determination item. When the own lane signal AS and the oncoming lane signal BS are staggered traffic lights, the pedestrian crossing likelihood is set to a higher value than when the pedestrian crossing likelihood is not a staggered traffic light. Here, information on whether or not the own lane traffic light AS and the oncoming lane traffic light BS are staggered traffic lights is acquired by, for example, high-precision map information.

The intersection size detection unit 609 detects the size of the intersection CP when the place having the pedestrian crossing WR is the intersection CP as a crossing determination item. Then, when the size of the intersection CP is smaller than the predetermined size, the pedestrian crossing likelihood is set to a value higher than when the size of the intersection CP is equal to or larger than the predetermined size. Here, the "predetermined size of the intersection CP" is set from the viewpoint of the allowable sudden deceleration and the legal maximum speed of the road on which the own vehicle A travels.

The pedestrian crossing likelihood determination unit 610 is used to determine the high possibility that the pedestrian C will cross the road when the display condition of the own lane traffic light AS and the oncoming lane traffic light BS and the pedestrian existence condition are satisfied. Detects crossing judgment items (pedestrian face, number of pedestrians, staggered traffic lights, intersection size). Then, if the pedestrian crossing likelihood set based on the detection result of the crossing determination item is equal to or higher than a predetermined value, it is predicted that the pedestrian C may cross the pedestrian crossing WR, and the avoidance preparation operation is executed. ..

Here, the "pedestrian crossing likelihood" is, for example, the total sum of the pedestrian crossing likelihoods set based on the detection results of each of the pedestrian face orientation, the number of pedestrians, the staggered traffic light, and the intersection size. Let it be a value. For the "predetermined value of pedestrian crossing likelihood", the relationship between the magnitude of the pedestrian crossing likelihood and the frequency of occurrence of parting crossing is determined by experiments, etc., and the possibility of parting crossing by pedestrian C increases. Set to the cross-likelihood threshold.

Next, the driving assistance technology in the pedestrian crossing approach scene will be described. Then, the operation of the first embodiment will be described separately by "driving support processing action based on behavior prediction" and "pedestrian crossing prediction processing action in a pedestrian crossing approach scene".

[Driving assistance technology in pedestrian crossing approach scenes (Figs. 3 and 4)]
In the background technology of driving support in a pedestrian crossing approach scene, as a route generation method of an automatically driving vehicle that avoids pedestrians on the pedestrian crossing, information on the surroundings of the own vehicle is acquired, and a person whose behavior is predicted on the road visually recognizes a traffic light. It is predicted that the behavior prediction target person moves in the transverse direction when it is determined whether or not the vehicle is doing so (see Japanese Patent Application Laid-Open No. 2018-205965).

However, in the background technology, the behavior prediction target person on the road does not visually recognize the traffic light, and for example, the pedestrian C recognizes that the vehicle on the roadway has stopped and crosses the pedestrian crossing. In some cases. In the case of this parting-off crossing, there is a problem that it is not possible to predict that a pedestrian will move in the crossing direction depending on whether or not the traffic light is visually recognized.

In particular, as shown in FIG. 3, when the own lane signal AS and the oncoming lane signal BS are staggered traffic lights, it is assumed that the own vehicle A approaches the pedestrian crossing WR. In this pedestrian crossing approach scene, when the own lane signal AS remains blue and the oncoming lane signal BS switches from blue to red and the oncoming vehicle B on the oncoming lane BL stops, the pedestrian C is on the oncoming lane BL. It is often the case that the oncoming vehicle B is recognized as having stopped and crosses the road.

In this case, since the background technology cannot predict the parting crossing of the pedestrian C as described above, the traveling of the own vehicle A does not decelerate based on, for example, a pre-generated target speed profile. It will be a run approaching the pedestrian crossing WR. Therefore, the detection timing for detecting the pedestrian C who is crossing the pedestrian crossing in the own vehicle A is the timing when the own vehicle A has already approached the position immediately before the pedestrian crossing WR. As a result, in order to avoid the pedestrian C by decelerating the own vehicle A, as shown by the arrow X in FIG. 3, the position of the own vehicle A and the position of the pedestrian crossing WR at the detection timing of the pedestrian C are short. You will be forced to slow down suddenly using the distance.

As a result of verifying the solution method for the above background technology
(A) In the scene where the own vehicle A approaches the pedestrian crossing WR, if the indication of the oncoming lane signal BS indicates that the progress is prohibited when the indication of the own lane signal AS indicates that the vehicle can proceed, a pedestrian There is a high possibility that C will cross the pedestrian crossing WR.
(B) When pedestrian C predicts that he will cross the pedestrian crossing WR, when the avoidance preparation operation for avoiding pedestrian C is started at the predicted timing, there is a long distance between the position of the own vehicle A and the position of the pedestrian crossing WR. It is possible to secure a margin for avoidance preparation.
I focused on that point.

Based on the above points of interest, the driving support method of the present disclosure determines whether or not the pedestrian crossing WR exists within a predetermined distance ahead of the own vehicle A. When the existence of the pedestrian crossing WR is determined, the traffic light AS installed corresponding to the pedestrian crossing WR and corresponding to the own lane AL in which the own vehicle A is traveling is recognized. .. When it is recognized that the display of the own lane signal AS indicates that the vehicle can proceed, it is a signal installed corresponding to the pedestrian crossing WR and the oncoming vehicle B travels toward the own vehicle A. Recognize the indication of the oncoming lane traffic light BS corresponding to the lane BL. When it is recognized that the display of the own lane traffic light AS indicates that the vehicle can proceed and the display of the oncoming lane traffic light BS indicates that the vehicle cannot proceed, it is predicted that the pedestrian C may cross the pedestrian crossing WR. When it is predicted that the pedestrian C may cross the pedestrian crossing WR, the problem-solving means of executing the avoidance preparation operation of the own vehicle A to avoid the pedestrian C is adopted.

That is, when it is recognized that the display of the own lane signal AS indicates that the vehicle can proceed and the display of the oncoming lane signal BS indicates that the progress is prohibited, it is predicted that the pedestrian C may cross the pedestrian crossing WR. NS. In the case of this pedestrian crossing prediction, since the signal display conditions of the own lane signal AS and the oncoming lane signal BS are determined, it is necessary to determine whether or not the pedestrian C is visually recognizing the surrounding objects. No. In addition, since the pedestrian crossing prediction is based on the determination of the traffic light display condition, as shown in FIG. 4, the vehicle travels at a position farther from the pedestrian crossing WR as compared with the case where the determination is made by visually recognizing surrounding objects by the pedestrian C. It is possible to predict the parting crossing of the pedestrian C in the own vehicle A.

Then, when it is predicted that the pedestrian C may cross the pedestrian crossing WR, the avoidance preparation operation of the own vehicle A for avoiding the pedestrian C is executed. Therefore, when the pedestrian crossing prediction is made by the own vehicle A traveling at a position away from the pedestrian crossing WR, as shown in FIG. 4, the own vehicle A at the pedestrian crossing prediction timing of the pedestrian C A sufficiently long avoidance preparation margin distance L is secured between the position and the position of the pedestrian crossing WR. Therefore, for example, when the pedestrian C is avoided by decelerating the own vehicle A, the pedestrian C is avoided by the slow deceleration using the long avoidance preparation margin distance L as shown by the arrow Y in FIG. Becomes possible.

In this way, in the scene where the own vehicle A approaches the pedestrian crossing WR, it is predicted that the pedestrian C may cross the pedestrian crossing WR based on the determination of the traffic light display condition, and the pedestrian crossing is predicted. The avoidance preparation operation is started at the predicted timing. As a result, in the scene where the own vehicle A approaches the pedestrian crossing WR, even if there is a pedestrian C who crosses the pedestrian crossing, the pedestrian avoidance due to sudden deceleration can be prevented by executing the avoidance preparation operation in advance. can.

By the way, according to the inventor's experiment, in the background technology, when pedestrians are detected to cross the intersection 20m before the intersection and decelerated to 10km / h using 10m, deceleration = 0.59G sudden deceleration. There has occurred. On the other hand, in the present disclosure, it is predicted that a pedestrian may cross the road 50 m before the intersection, and if the vehicle decelerates to 10 km / h using 40 m, the deceleration will be slowed down by 0.15 G. rice field. In other words, the deceleration of the own vehicle decreased from (0.59G) to (0.15G), and a 75% deceleration improvement effect was confirmed.

[Driving support processing action based on behavior prediction (Fig. 5)]
FIG. 5 is a flowchart showing the flow of driving support processing based on the behavior prediction executed by the automatic driving control system (FIG. 1). Hereinafter, the driving support processing action based on the behavior prediction will be described with reference to FIG.

In step S1, when the process is started, the object detection device 1 uses the in-vehicle sensor to detect the object, and the position, posture, size, speed, etc. of the objects (oncoming vehicle B, pedestrian C) existing around the own vehicle are used. The object detection process for detecting the above is performed.

In the next step S2, the object detection integration / tracking unit 2 outputs one two-dimensional position, posture, size, speed, etc. for each object based on the object detection result obtained by the object detection process. Object detection integration / tracking processing is performed.

In the next step S3, the own vehicle position estimation device 3 measures the absolute position of the own vehicle A, that is, the position, posture, speed, etc. of the own vehicle A with respect to a certain reference point by a sensor that measures the absolute position such as GPS or odometry. The vehicle position information acquisition process is performed.

In the next step S4, the map storage device 4 performs a map information acquisition process for acquiring road map information regarding the road on which the own vehicle A, the oncoming vehicle B, or the like travels, based on the high-precision map information.

In the next step S5, the own vehicle position estimation unit 5 in the map estimates the position of the own vehicle A in the road map based on the absolute position of the own vehicle A and the road map information around the own vehicle. Vehicle position estimation processing is performed.

In the next step S6, the behavior prediction unit 6 performs a behavior prediction process for predicting the behavior of an object existing around the own vehicle A according to the information obtained up to the own vehicle position estimation process in the map. Specifically, in the behavior prediction process, when the own vehicle A approaches the pedestrian crossing WR, pedestrian C waiting for a signal at the pedestrian crossing WR may shift to the pedestrian crossing behavior. Perform pedestrian crossing prediction processing (FIG. 6).

In the next step S7, the own vehicle route generation unit 7 performs the own vehicle route generation process for generating the own vehicle route according to the result of the action prediction process. Specifically, in the own vehicle route generation process, when the prediction information that the pedestrian C may cross the pedestrian crossing WR is input by the pedestrian crossing prediction process, the target route and the target speed generated in advance are input. At least one of the profiles is modified according to the mode of avoidance preparation operation.

In the next step S8, the vehicle control unit 8 outputs a control command to each actuator so that the traveling speed and steering angle of the own vehicle according to the target route and the target speed profile generated in the own vehicle route generation process are obtained. Performs motion control processing and proceeds to the end of processing. Specifically, when it is predicted that the pedestrian C may cross the pedestrian crossing WR, the avoidance preparation operation of the own vehicle A to avoid the pedestrian C is executed.

That is, when the mode of the avoidance preparation operation based on the pedestrian crossing prediction is the lane change operation or the deceleration operation, the avoidance preparation operation is executed by the control output along the target route and the target speed profile modified according to the mode. Further, when the mode of the avoidance preparation operation based on the pedestrian crossing prediction is the brake pressurization application, the target path and the target speed profile generated in advance are maintained as they are, and the command output of the brake pressurization application to the braking actuator is maintained. Executes the avoidance preparation operation. Hereinafter, the characteristics of each mode of avoidance preparation operation based on the pedestrian crossing prediction will be described.

(a) If the road (own lane AL) on which the own vehicle A travels is in multiple lanes and the vehicle is traveling in a lane other than the lane farthest from the sidewalk, the avoidance preparation operation is performed in the driving lane of the own vehicle A. Is a lane change operation for changing lanes to the lane away from the sidewalk. Therefore, even if the pedestrian C who is crossing the sidewalk crosses the lane on the sidewalk side, it is possible to avoid crossing the course of the own vehicle A.

(b) The avoidance preparation operation is a constant deceleration operation in which the vehicle speed of the own vehicle A is decelerated at a predetermined constant deceleration. Therefore, by decelerating without changing the deceleration of the own vehicle A, it is possible to execute the avoidance preparation operation without deteriorating the ride quality of the occupant.

(c) The avoidance preparation operation is a brake pressurization operation that applies the brake pressurization immediately before the hydraulic braking force is generated to the hydraulic brake device of the own vehicle A. Therefore, even if there is a pedestrian C crossing the parting line, deceleration can be started at the moment of observing the parting crossing, so that the idle section until the hydraulic braking force is generated in the own vehicle A is reduced, and the vehicle suddenly decelerates. The degree of can be suppressed.

(d) When the vehicle speed when the own vehicle A starts the avoidance preparation operation is equal to or higher than the predetermined vehicle speed, the avoidance preparation operation is a deceleration operation for decelerating from the start vehicle speed to the target vehicle speed in the predetermined pedestrian crossing reach area. .. Therefore, when there is a pedestrian C who crosses the parting, the speed is reduced to such an extent that the deceleration does not occur suddenly, so that the pedestrian C who crosses the parting can be avoided without deteriorating the ride quality of the occupant.

[Pedestrian crossing prediction processing action in a pedestrian crossing approach scene (Fig. 6)]
FIG. 6 is a flowchart showing a flow of pedestrian crossing prediction processing in a pedestrian crossing approach scene in which the own vehicle A approaches the pedestrian crossing WR executed by the behavior prediction unit 6. Hereinafter, the pedestrian crossing prediction processing action in the pedestrian crossing approach scene will be described with reference to FIG.

In step S601, following the start of the action prediction, the pedestrian crossing existence determination unit 601 determines whether or not the pedestrian crossing WR exists within a predetermined distance ahead of the own vehicle A. If YES (there is a pedestrian crossing WR), the process proceeds to step S602, and if NO (there is no pedestrian crossing WR), the determination in step S601 is repeated.

In step S602, following the determination of YES in S601, the own lane signal state recognition unit 602 corresponds to the own lane AL in which the own vehicle A is traveling, which is a traffic light installed corresponding to the pedestrian crossing WR. Recognize the display of the own lane traffic light AS.

In the next step S603, following S602, it is determined whether or not the display of the own lane traffic light AS is a display indicating that the vehicle can proceed. If YES (the display of the own lane signal AS can proceed), the process proceeds to step S604, and if NO (the display of the own lane signal AS cannot proceed), the process proceeds to step S611.

In step S604, following the determination of YES in S603, the oncoming lane signal state recognition unit 603 is a traffic light installed corresponding to the pedestrian crossing WR, and the oncoming vehicle B heading for the own vehicle A travels on the opposite side. Recognize the indication of the oncoming lane traffic light BS corresponding to the lane BL.

In the next step S605, following S604, it is determined whether or not the indication of the oncoming lane traffic light BS is an indication indicating prohibition of progress. If YES (the display of the oncoming lane signal BS is prohibited), the process proceeds to step S606, and if NO (the display of the oncoming lane signal BS can proceed), the process proceeds to step S611.

In step S606, following the determination of YES in S605, the pedestrian detection unit 604 near the pedestrian crossing detects the pedestrian C existing in the vicinity of the pedestrian crossing WR, and proceeds to step S607.

In the next step S607, following S606, it is determined whether or not the pedestrian C exists in the vicinity of the pedestrian crossing WR. If YES (with pedestrian C), the process proceeds to step S608, and if NO (without pedestrian C), the process proceeds to step S611.

In step S608, following the determination of YES in S607, the crossing determination items (pedestrian face orientation, number of pedestrians, staggered traffic light, intersection size) used for determining the high possibility that the pedestrian C will cross the road are detected. Then, the process proceeds to step S609.

In the next step S609, following S608, the pedestrian crossing is based on the sum of the pedestrian crossing likelihoods set based on the detection results of the pedestrian face, the number of pedestrians, the staggered traffic light, and the intersection size. It is determined whether or not the likelihood is equal to or higher than a predetermined value. If YES (pedestrian crossing likelihood ≥ predetermined value), the process proceeds to step S610, and if NO (pedestrian crossing likelihood <predetermined value), the process proceeds to step S611.

In step S610, following the determination of YES in S609, it is predicted that the pedestrian C may cross the pedestrian crossing WR, and the action prediction ends.

In step S611, following the determination of NO in any of S603, S605, S607, and S609, it is predicted that the pedestrian C is unlikely to cross the pedestrian crossing WR, and the action prediction ends.

As described above, in the pedestrian crossing prediction process in the pedestrian crossing approach scene, when the pedestrian crossing WR exists within a predetermined distance in front of the own vehicle A, all the conditions in S603, S605, S607, and S609 are satisfied. Then, it is predicted that pedestrian C may cross the pedestrian crossing WR in step S610. On the other hand, in the pedestrian crossing prediction process in the pedestrian crossing approach scene, when the pedestrian crossing WR exists within a predetermined distance in front of the own vehicle A, at least one of the conditions in S603, S605, S607, and S609 is not satisfied. If so, it is predicted that the process proceeds to step S611 and there is no possibility that the pedestrian C crosses the pedestrian crossing WR. Hereinafter, the characteristics of the pedestrian crossing prediction process in the pedestrian crossing approach scene in the first embodiment will be described.

In the first embodiment, it is detected whether or not the pedestrian C is present in the vicinity of the pedestrian crossing WR (S606). When the display conditions of the own lane traffic light AS and the oncoming lane traffic light BS are satisfied, when the presence of pedestrian C is detected in the vicinity of the pedestrian crossing WR (YES in S607), the pedestrian C sets the pedestrian crossing WR. Predicting that there is a possibility of crossing (S610), the avoidance preparation operation is executed.

That is, the pedestrian existence condition in which the pedestrian C exists in the vicinity of the pedestrian crossing WR is added to the signal display condition. As a result, even if the display conditions of the own lane traffic light AS and the oncoming lane traffic light BS are satisfied, if the presence of the pedestrian C is not detected in the vicinity of the pedestrian crossing WR (NO in S607), the pedestrian It is predicted that C is unlikely to cross the pedestrian crossing WR (S611), and the avoidance preparation operation is not executed. Therefore, since the avoidance preparation operation is executed only when there is a pedestrian C who may cross the parting line, it is possible to avoid executing the avoidance preparation operation when there is no pedestrian C.

For example, if there is a pedestrian C waiting for a signal near the pillar of the own lane traffic light AS, the indication of the own lane traffic light AS cannot be observed or is difficult to observe, but the indication of the oncoming lane traffic light BS is Easy to observe. The pedestrian C in such a place has a high possibility of crossing the pedestrian crossing WR by recognizing that the oncoming vehicle B on the oncoming lane BL has stopped. This is because there are few pedestrians C recognized as staggered traffic lights by the own lane traffic light AS and the oncoming lane traffic light BS, and the indication of the oncoming lane traffic light BS in the range that one can see changes from being able to proceed to being prohibited. This is because when the oncoming vehicle B stops, it is easy to mistakenly determine that the own lane traffic light AS is also prohibited from proceeding.

In the first embodiment, when recognizing the display of the oncoming lane traffic light BS, when it is detected that the oncoming vehicle B of the oncoming lane BL is stopped or decelerating with respect to the stop line on the lane, the current oncoming lane traffic light BS is displayed. Recognize that the indication is an indication of prohibition of progress.

That is, it can be inferred that the oncoming vehicle B in the oncoming lane BL is stopped or decelerated with respect to the stop line because the oncoming lane signal BS is prohibited from proceeding. Therefore, when the display of the oncoming lane signal BS cannot be seen from the own vehicle A side, it is possible to indirectly detect that the display of the oncoming lane signal BS corresponding to the oncoming lane BL indicates that the vehicle is prohibited from proceeding. It will be possible. Therefore, when the display of the oncoming lane traffic light BS cannot be directly detected by the vehicle-mounted sensor, it is detected that the display of the oncoming lane traffic light BS indicates that the progress is prohibited only by the vehicle-mounted sensor that detects the behavior of the oncoming vehicle B. be able to.

In the first embodiment, when recognizing the indication of the oncoming lane traffic light BS, the oncoming lane traffic light BS indicates the progress or progress based on the communication information acquired from the outside by the own vehicle A by the road-to-vehicle communication or the inter-vehicle communication. Recognize whether it is a sign indicating prohibition.

That is, if a road-to-vehicle communication or vehicle-to-vehicle communication system is established, it is possible to acquire the display information of the oncoming lane traffic light BS corresponding to the current oncoming lane BL from the cloud or another vehicle at the request of the own vehicle A. It is possible. Therefore, when the display of the oncoming lane signal BS cannot be directly detected by the in-vehicle sensor, the display of the oncoming lane signal BS indicates that the oncoming lane signal BS can proceed or the progress is prohibited based on the information acquired by the own vehicle A from the outside. It is possible to detect whether it is shown.

In the first embodiment, when the display condition and the pedestrian existence condition of the own lane traffic light AS and the oncoming lane traffic light BS are satisfied, the crossing judgment item used for determining the high possibility that the pedestrian C crosses the road is detected. do. If the pedestrian crossing likelihood set based on the detection result of the crossing determination item is equal to or higher than a predetermined value, it is predicted that the pedestrian C may cross the pedestrian crossing WR, and the avoidance preparation operation is executed.

That is, even when the display condition of the own lane traffic light AS and the oncoming lane traffic light BS and the pedestrian existence condition are satisfied, it can be determined that the pedestrian C is unlikely to cross the road. Therefore, only when the crossing judgment item used for determining the high possibility that the pedestrian C crosses the crossing is detected and the pedestrian crossing likelihood set based on the detection result of the crossing judgment item is equal to or more than a predetermined value. It is predicted that pedestrian C may cross the pedestrian crossing WR. Therefore, in the case where the pedestrian C is unlikely to cross the parting line, while suppressing the execution of the avoidance preparation operation, when the pedestrian C is likely to cross the parting line, the pedestrian C is said to have the parting crossing. The avoidance preparation action can be executed based on the prediction of.

In the first embodiment, the face orientation of the pedestrian C is detected as a crossing determination item. If the pedestrian C's face is not looking in the direction of the own vehicle A, the pedestrian crossing likelihood is set to a higher value than when looking in the direction of the own vehicle A.

That is, when the face direction of the pedestrian C does not look in the direction of the own vehicle A, the own vehicle A travels closer than when the face direction of the pedestrian C looks in the direction of the own vehicle A. Not being able to recognize what you are doing increases the chances of crossing the line. Therefore, when the face direction of the pedestrian C is looking in the direction of the own vehicle A, the execution of the avoidance preparation operation is suppressed, and when the face direction of the pedestrian C is not looking in the direction of the own vehicle A, the pedestrian The avoidance preparation operation can be executed based on the prediction that C has a parting crossing.

In the first embodiment, when the presence of the pedestrian C is detected in the vicinity of the pedestrian crossing WR, the number of pedestrians C is measured as a crossing determination item. When the number of pedestrians C is equal to or greater than the predetermined number, the pedestrian crossing likelihood is set to a value higher than when the number of pedestrians C is less than the predetermined number.

That is, when the number of pedestrians C is greater than or equal to the predetermined number, the probability of crossing the parting is proportional to the number of pedestrians C as compared with the case where the number of pedestrians C is less than the predetermined number. The possibility of doing it increases. Therefore, based on the prediction that the pedestrian C has a parting crossing when the number of pedestrians C is more than the predetermined number while suppressing the execution of the avoidance preparation operation when the number of pedestrians C is less than the predetermined number. It is possible to execute the avoidance preparation operation.

In the first embodiment, as a crossing determination item, it is detected whether or not the own lane traffic light AS and the oncoming lane traffic light BS are staggered traffic lights. When the own lane signal AS and the oncoming lane signal BS are staggered traffic lights, the pedestrian crossing likelihood is set to a higher value than when the pedestrian crossing likelihood is not a staggered traffic light.

That is, when the own lane traffic light AS and the oncoming lane traffic light BS are not staggered traffic lights, the display of the own lane traffic light AS and the oncoming lane traffic light BS are compared with the case where the own lane traffic light AS and the oncoming lane traffic light BS are staggered traffic lights. There is a time lag in the manifestation of, which increases the possibility of crossing the deadline. Therefore, while suppressing the execution of the avoidance preparation operation when the own lane signal AS and the oncoming lane signal BS are not the time difference type traffic lights, when the own lane signal is a time difference type traffic light, the avoidance preparation is based on the prediction that the pedestrian C has a parting crossing. The operation can be executed.

In the first embodiment, the size of the intersection CP is detected when the place having the pedestrian crossing WR is the intersection CP as a crossing determination item. When the size of the intersection CP is smaller than the predetermined size, the pedestrian crossing likelihood is set to a higher value than when the size of the intersection CP is the predetermined size or more.

That is, when the size of the intersection CP having the pedestrian crossing WR is smaller than the predetermined size, there is a possibility that the pedestrian C will cross the intersection because the crossing distance by the pedestrian C is shorter than when the size of the intersection CP is larger than the predetermined size. Increase. Therefore, when the size of the intersection CP having the pedestrian crossing WR is larger than the predetermined size, the execution of the avoidance preparation operation is suppressed, and when the size of the intersection CP is smaller than the predetermined size, it is predicted that the pedestrian C has a parting crossing. Based on this, the avoidance preparation operation can be executed.

As described above, the driving support method and the driving support device in the autonomous driving vehicle of the first embodiment have the effects listed below.

(1) This is a driving support method using a controller (automatic driving control system) that assists the driving of the own vehicle A when the own vehicle A approaches the pedestrian crossing WR.
It is determined whether or not the pedestrian crossing WR exists within a predetermined distance in front of the own vehicle A (S601).
When the existence of the pedestrian crossing WR is determined, the traffic light AS installed corresponding to the pedestrian crossing WR and corresponding to the own lane AL in which the own vehicle A is traveling is recognized. (S602),
It is a traffic light installed corresponding to the pedestrian crossing WR, and recognizes the indication of the oncoming lane signal BS corresponding to the oncoming lane BL in which the oncoming vehicle B travels toward the own vehicle A (S603, S604).
When recognizing that the display of the own lane signal AS indicates that the vehicle can proceed and the display of the oncoming lane signal BS indicates that the progress is prohibited, it is predicted that the pedestrian C may cross the pedestrian crossing WR (S610). ),
When it is predicted that the pedestrian C may cross the pedestrian crossing WR, the avoidance preparation operation of the own vehicle A to avoid the pedestrian C is executed (FIG. 6).
Therefore, in the scene where the own vehicle A approaches the pedestrian crossing WR, even if there is a pedestrian C who crosses the pedestrian crossing, the avoidance preparation operation is executed in advance to prevent pedestrian avoidance due to sudden deceleration. A method can be provided.

(2) Detecting whether or not pedestrian C exists in the vicinity of the pedestrian crossing WR (S606),
When the display conditions of the own lane traffic light AS and the oncoming lane traffic light BS are satisfied, if the presence of the pedestrian C is detected in the vicinity of the pedestrian crossing WR, the pedestrian C may cross the pedestrian crossing WR. Is predicted, and the avoidance preparation operation is executed (Fig. 6).
Therefore, since the avoidance preparation operation is executed only when there is a pedestrian C who may cross the parting line, it is possible to avoid executing the avoidance preparation operation when there is no pedestrian C.

(3) When recognizing the indication of the oncoming lane traffic light BS, when it is detected that the oncoming vehicle B of the oncoming lane BL is stopped or decelerating with respect to the stop line, the indication of the oncoming lane traffic light BS is displayed. Recognize that it is a manifestation indicating prohibition of progress (Fig. 6).
Therefore, when the display of the oncoming lane traffic light BS cannot be directly detected by the vehicle-mounted sensor, it is detected that the display of the oncoming lane traffic light BS indicates that the progress is prohibited only by the vehicle-mounted sensor that detects the behavior of the oncoming vehicle B. be able to.

(4) When recognizing the indication of the oncoming lane traffic light BS, the oncoming lane traffic light BS indicates the progress or prohibits the progress based on the communication information acquired from the outside by the own vehicle A by the road-to-vehicle communication or the inter-vehicle communication. Recognize whether it is a manifestation (Fig. 6).
Therefore, when the display of the oncoming lane signal BS cannot be directly detected by the in-vehicle sensor, the display of the oncoming lane signal BS indicates that the oncoming lane signal BS can proceed or the progress is prohibited based on the information acquired by the own vehicle A from the outside. It is possible to detect whether it is shown.

(5) When the display conditions of the own lane traffic light AS and the oncoming lane traffic light BS and the pedestrian existence condition are satisfied, the crossing judgment items used for determining the high possibility that the pedestrian C will cross the road are detected ( S608),
If the pedestrian crossing likelihood set based on the detection result of the crossing judgment item is equal to or higher than a predetermined value (YES in S609), it is predicted that the pedestrian C may cross the pedestrian crossing WR (S610). The avoidance preparation operation is executed (Fig. 6).
Therefore, in the case where the pedestrian C is unlikely to cross the parting line, while suppressing the execution of the avoidance preparation operation, when the pedestrian C is likely to cross the parting line, the pedestrian C is said to have the parting crossing. The avoidance preparation action can be executed based on the prediction of.

(6) When the presence of pedestrian C is detected in the vicinity of the pedestrian crossing WR, the face orientation of pedestrian C is detected as a crossing judgment item.
If the pedestrian C's face is not looking in the direction of the own vehicle A, the pedestrian crossing likelihood is set to a higher value than when looking in the direction of the own vehicle A (FIG. 6).
Therefore, when the face direction of the pedestrian C is looking in the direction of the own vehicle A, the execution of the avoidance preparation operation is suppressed, and when the face direction of the pedestrian C is not looking in the direction of the own vehicle A, the pedestrian The avoidance preparation operation can be executed based on the prediction that C has a parting crossing.

(7) When the presence of pedestrian C is detected in the vicinity of the pedestrian crossing WR, the number of pedestrian C is measured as a crossing judgment item.
When the number of pedestrians C is equal to or greater than the predetermined number, the pedestrian crossing likelihood is set to a higher value than when the number of pedestrians is less than the predetermined number (FIG. 6).
Therefore, based on the prediction that the pedestrian C has a parting crossing when the number of pedestrians C is more than the predetermined number while suppressing the execution of the avoidance preparation operation when the number of pedestrians C is less than the predetermined number. It is possible to execute the avoidance preparation operation.

(8) As a crossing judgment item, it is detected whether or not the own lane traffic light AS and the oncoming lane traffic light BS are staggered traffic lights.
When the own lane signal AS and the oncoming lane signal BS are staggered traffic lights, the pedestrian crossing likelihood is set to a higher value than when the pedestrian crossing likelihood is not a staggered traffic light (FIG. 6).
Therefore, while suppressing the execution of the avoidance preparation operation when the own lane signal AS and the oncoming lane signal BS are not the time difference type traffic lights, when the own lane signal is a time difference type traffic light, the avoidance preparation is based on the prediction that the pedestrian C has a parting crossing. The operation can be executed.

(9) As a crossing judgment item, the size of the intersection CP is detected when the place where the pedestrian crossing WR is located is the intersection CP.
When the size of the intersection CP is smaller than the predetermined size, the pedestrian crossing likelihood is set to a higher value than when the size of the intersection CP is the predetermined size or more (FIG. 6).
Therefore, when the size of the intersection CP having the pedestrian crossing WR is larger than the predetermined size, the execution of the avoidance preparation operation is suppressed, and when the size of the intersection CP is smaller than the predetermined size, it is predicted that the pedestrian C has a parting crossing. Based on this, the avoidance preparation operation can be executed.

(10) In the avoidance preparation operation, when the road on which the own vehicle A travels is in multiple lanes and the vehicle is traveling in a lane other than the lane farthest from the sidewalk, the side away from the sidewalk in the traveling lane of the own vehicle A. The lane change operation is performed to change lanes to the above lane (Fig. 5).
Therefore, even if the pedestrian C who is crossing the sidewalk crosses the lane on the sidewalk side, it is possible to avoid crossing the course of the own vehicle A.

(11) The avoidance preparation operation is a constant deceleration operation in which the vehicle speed of the own vehicle A is decelerated at a predetermined constant deceleration (Fig. 5).
Therefore, by decelerating without changing the deceleration of the own vehicle A, it is possible to execute the avoidance preparation operation without deteriorating the ride quality of the occupant.

(12) The avoidance preparation operation is a brake pressurization operation in which the brake pressurization immediately before the hydraulic braking force is generated to the hydraulic brake device of the own vehicle A is applied (FIG. 5).
Therefore, even if there is a pedestrian C crossing the parting line, deceleration can be started at the moment of observing the parting crossing, so that the idle section until the hydraulic braking force is generated in the own vehicle A is reduced, and the vehicle suddenly decelerates. The degree of can be suppressed.

(13) The avoidance preparation operation is a deceleration operation that decelerates from the start vehicle speed to the target vehicle speed in the predetermined pedestrian crossing reach area when the vehicle speed when the own vehicle A starts the avoidance preparation operation is equal to or higher than the predetermined vehicle speed. (Fig. 5).
Therefore, when there is a pedestrian C who crosses the parting, the speed is reduced to such an extent that the deceleration does not occur suddenly, so that the pedestrian C who crosses the parting can be avoided without deteriorating the ride quality of the occupant.

(14) A driving support device provided with a controller (automatic driving control system) that assists the driving of the own vehicle A when the own vehicle A approaches the pedestrian crossing WR.
The controller (automatic driving control system)
A pedestrian crossing existence determination unit 601 that determines whether or not a pedestrian crossing WR exists within a predetermined distance in front of the own vehicle A, and
When the existence of the pedestrian crossing WR is determined, the traffic light AS installed corresponding to the pedestrian crossing WR and corresponding to the own lane AL in which the own vehicle A is traveling is recognized. Own lane signal state recognition unit 602 and
With the oncoming lane signal state recognition unit 603, which is a traffic light installed corresponding to the pedestrian crossing WR and recognizes the display of the oncoming lane signal BS corresponding to the oncoming lane BL in which the oncoming vehicle B heading for the own vehicle A travels. ,
A pedestrian who predicts that pedestrian C may cross the pedestrian crossing WR when he / she recognizes that the display of the own lane traffic light AS indicates that the vehicle can proceed and the display of the oncoming lane traffic light BS indicates that the vehicle cannot proceed. Crosswalk prediction unit 605 and
When it is predicted that the pedestrian C may cross the pedestrian crossing WR, the vehicle control unit 8 that executes the avoidance preparation operation of the own vehicle to avoid the pedestrian C is provided (FIG. 1).
Therefore, in the scene where the own vehicle A approaches the pedestrian crossing WR, even if there is a pedestrian C who crosses the pedestrian crossing, the avoidance preparation operation is executed in advance to prevent pedestrian avoidance due to sudden deceleration. Equipment can be provided.

The driving support method and the driving support device of the present disclosure have been described above based on the first embodiment. However, the specific configuration is not limited to the first embodiment, and design changes and additions are permitted as long as the gist of the invention according to each claim is not deviated from the claims.

In the first embodiment, an example is shown in which, when the driver inputs a destination, the own vehicle route generation unit 7 generates in advance a target route TL and a target speed profile as a planned travel route connecting the current location of the own vehicle to the destination. rice field. However, as the own vehicle route, even if the target route is not generated in advance, if the travel route for each road connecting the current location of the own vehicle to the destination is determined, the lane unit is based on the behavior prediction of surrounding objects. The planned travel route may be calculated. Further, in the case of a driving support vehicle, there is a case where the own vehicle route generation unit 7 that generates a target route and a target speed profile is not provided. However, for example, in the case of a driving support vehicle based on preceding vehicle follow-up control, it is possible to respond to an avoidance preparation operation based on a pedestrian crossing prediction by controlling a driving lane change operation, a deceleration operation, or the like.

In the first embodiment, an example is shown in which the driving support method and the driving support device of the present disclosure are applied to an autonomous driving vehicle in which the vehicle motion is controlled so as to travel along a target route. However, the driving support method and the driving support device of the present disclosure are not limited to the autonomous driving vehicle, but are provided with an auto cruise function, a lane keeping function, etc., and support at least one of steering operation / accelerator operation / brake operation. It can also be applied to driving support vehicles. Further, even in a manually driven vehicle in which the driver performs steering operation / accelerator operation / braking operation, it is possible to perform an avoidance preparation operation by notifying the driver of a warning based on a pedestrian crossing prediction. .. Further, as the vehicle to which the driving support method and the driving support device of the present disclosure are applied, it can be applied to all kinds of vehicles such as engine vehicles, hybrid vehicles, electric vehicles and the like.

1 Object detection device 2 Object detection integration / tracking unit 3 Own vehicle position estimation device 4 Map storage device 5 Own vehicle position estimation unit in map 6 Behavior prediction unit 601 Pedestrian crossing existence determination unit 602 Own lane signal State recognition unit 603 Oncoming lane signal State recognition unit 604 Pedestrian crossing detection unit 605 Pedestrian crossing prediction unit 606 Pedestrian face orientation detection unit 607 Pedestrian number measurement unit 608 Staggered traffic light detection unit 609 Crossing size detection unit 610 Pedestrian crossing likelihood determination unit 7 Own vehicle route generation unit 8 Vehicle control unit A Own vehicle B Oncoming vehicle C Pedestrian TL Target route (planned travel route of own vehicle A)
CP intersection AL own lane BL oncoming lane AS own lane traffic light BS oncoming lane traffic light WR pedestrian crossing

Claims (14)

It is a driving support method by a controller that assists the driving of the own vehicle when the own vehicle approaches the pedestrian crossing.
Judging whether or not the pedestrian crossing exists within a predetermined distance in front of the own vehicle,
When the existence of the pedestrian crossing is determined, the traffic light installed corresponding to the pedestrian crossing recognizes the display of the own lane signal corresponding to the own lane in which the own vehicle is traveling.
It is a traffic light installed corresponding to the pedestrian crossing, and recognizes the indication of the oncoming lane traffic light corresponding to the oncoming lane in which the oncoming vehicle travels toward the own vehicle.
Recognizing that the display of the own lane traffic light indicates that the vehicle can proceed and that the display of the oncoming lane traffic light indicates that the vehicle cannot proceed, it is predicted that a pedestrian may cross the pedestrian crossing.
A driving support method comprising executing an avoidance preparation operation of the own vehicle to avoid the pedestrian when it is predicted that the pedestrian may cross the pedestrian crossing. In the driving support method according to claim 1,
Detecting whether or not there are pedestrians in the vicinity of the pedestrian crossing,
When the display conditions of the own lane traffic light and the oncoming lane traffic light are satisfied, if the presence of the pedestrian is detected in the vicinity of the pedestrian crossing, the pedestrian may cross the pedestrian crossing. A driving support method characterized by predicting that and executing the avoidance preparation operation. In the driving support method according to claim 1 or 2.
When recognizing the indication of the oncoming lane traffic light, when it is detected that the oncoming vehicle in the oncoming lane is stopped or decelerating with respect to the stop lane, the indication of the oncoming lane traffic light indicates that the vehicle is prohibited from proceeding. A driving support method characterized by recognizing that it is a manifestation. In the driving support method according to claim 1 or 2.
When recognizing the indication of the oncoming lane traffic light, the indication indicating that the oncoming lane traffic light is progressing or the indication indicating prohibition of progress based on the communication information acquired from the outside by the own vehicle by road-to-vehicle communication or vehicle-to-vehicle communication. A driving support method characterized by recognizing a vehicle. In the driving support method according to any one of claims 2 to 4,
When the display condition and the pedestrian existence condition of the own lane traffic light and the oncoming lane traffic light are satisfied, the crossing determination item used for determining the high possibility that the pedestrian will cross the road is detected.
If the pedestrian crossing likelihood set based on the detection result of the crossing determination item is equal to or higher than a predetermined value, it is predicted that the pedestrian may cross the pedestrian crossing, and the avoidance preparation operation is executed. A driving support method characterized by that. In the driving support method according to claim 5,
When the presence of a pedestrian is detected in the vicinity of the pedestrian crossing, the face orientation of the pedestrian is detected as the crossing determination item.
A driving support method characterized in that when the pedestrian's face direction does not look in the direction of the own vehicle, the pedestrian crossing likelihood is set to a higher value than when looking in the direction of the own vehicle. .. In the driving support method according to claim 5,
When the presence of a pedestrian is detected in the vicinity of the pedestrian crossing, the number of the pedestrian is measured as the crossing determination item.
A driving support method characterized in that when the number of pedestrians is equal to or greater than a predetermined number, the pedestrian crossing likelihood is set to a value higher than when the number of pedestrians is less than the predetermined number. In the driving support method according to claim 5,
As the crossing determination item, it is detected whether or not the own lane traffic light and the oncoming lane traffic light are staggered traffic lights.
A driving support method characterized in that when the own lane signal and the oncoming lane signal are staggered traffic lights, the pedestrian crossing likelihood is set to a higher value than when the pedestrian crossing likelihood is not the staggered traffic light. In the driving support method according to claim 5,
As the crossing determination item, when the place having the pedestrian crossing is an intersection, the size of the intersection is detected.
A driving support method characterized in that when the size of the intersection is smaller than the predetermined size, the pedestrian crossing likelihood is set to a higher value than when the size of the intersection is the predetermined size or more. In the driving support method according to any one of claims 1 to 9,
In the avoidance preparation operation, when the road on which the own vehicle travels is a plurality of lanes and the vehicle is traveling in a lane other than the lane farthest from the sidewalk, the driving lane of the own vehicle is on the side away from the sidewalk. A driving support method characterized by a lane change operation for changing lanes to a lane. In the driving support method according to any one of claims 1 to 9,
The avoidance preparation operation is a driving support method characterized in that the vehicle speed of the own vehicle is decelerated at a predetermined constant deceleration. In the driving support method according to any one of claims 1 to 9,
The avoidance preparation operation is a driving support method characterized in that the avoidance preparation operation is a brake pressurization operation in which a brake pressurization is applied immediately before a hydraulic braking force is generated to the hydraulic brake device of the own vehicle. In the driving support method according to any one of claims 1 to 9,
When the vehicle speed at the time when the own vehicle starts the avoidance preparation operation is equal to or higher than a predetermined vehicle speed, the avoidance preparation operation is a deceleration operation for decelerating from the start vehicle speed to a target vehicle speed in a predetermined pedestrian crossing reach area. Characteristic driving support method. It is a driving support device equipped with a controller that assists the driving of the own vehicle when the own vehicle approaches the pedestrian crossing.
The controller
A pedestrian crossing existence determination unit that determines whether or not a pedestrian crossing exists within a predetermined distance in front of the own vehicle, and a pedestrian crossing existence determination unit.
When the existence of the pedestrian crossing is determined, the own lane that is a traffic light installed corresponding to the pedestrian crossing and recognizes the indication of the own lane signal corresponding to the own lane in which the own vehicle is traveling. Signal state recognition unit and
An oncoming lane signal state recognition unit that recognizes the display of an oncoming lane signal corresponding to the oncoming lane in which the oncoming vehicle heading for the own vehicle is traveling, which is a traffic light installed corresponding to the pedestrian crossing.
When recognizing that the display of the own lane traffic light indicates that the vehicle can proceed and the display of the oncoming lane traffic light indicates that the vehicle is prohibited from traveling, it is predicted that the pedestrian may cross the pedestrian crossing. Prediction department and
A driving support device including a vehicle control unit that executes an avoidance preparation operation of the own vehicle to avoid the pedestrian when it is predicted that the pedestrian may cross the pedestrian crossing.

Images