JP7353239B2 - Motion prediction device and motion prediction program - Google Patents

Description

The present invention relates to a motion prediction device and a motion prediction program.

Patent Documents 1 to 5 disclose techniques for avoiding collisions between vehicles and moving objects.

WO2012/033173 publication Japanese Patent Application Publication No. 2019-144900 JP 2019-168953 Publication WO2019/092591 publication Japanese Patent No. 5673127 FIG. 12 shows the operation (original operation) of a conventional motion prediction device.

Similar to the techniques disclosed in Patent Documents 1 to 5 described above, the conventional motion prediction device performs the following operations in order to avoid a collision between vehicle SR and pedestrian HK.

First, as shown in FIG. 12A, the current position of the pedestrian HK near the building TS and the crosswalk CW, the position of the building TS, and the position of the crosswalk CW are detected.

Next, as shown in FIG. 12(B), a plurality of reference points KJ101 to KJ105 representing the current position of the detected pedestrian HK are extracted.

Finally, as shown in FIG. 12(C), based on the multiple extracted reference points KJ101 to KJ105 and considering the position of the building TS and the position of the crosswalk CW, the pedestrian HK's Predict future movements of UG101 to UG105.

However, the conventional motion prediction device differs from the above operation described with reference to FIG. 12 because it is difficult to accurately measure the distance in the traveling direction SH of the vehicle SR. The following actions may occur.

FIG. 13 shows the operation (possible operation) of a conventional motion prediction device.

As shown in FIG. 13(A), when the pedestrian HK is near the building TS and the crosswalk CW, the pedestrian HK is at a position IT far from the vehicle SR in the traveling direction SH of the vehicle SR. , false positive.

Despite the above-described erroneous detection, as shown in FIG. 13(B), a plurality of reference points KJ201 to KJ205 representing the erroneously detected position IT are extracted in the same manner as in FIG. 12(B).

Finally, as shown in FIG. 13(C), similar to FIG. 12(C), based on the reference points KJ201 to KJ205 and considering the position of the building TS and the position of the crosswalk CW, HK's future movements UG201 to UG205 are predicted.

As described above, the conventional motion prediction device cannot accurately predict the plurality of future movements UG101 to UG105 of the pedestrian HK due to the difficulty in measuring the distance in the traveling direction SH of the vehicle SR. Something happened.

An object of the present invention is to provide a motion prediction device and a motion prediction program that can accurately predict multiple future movement situations of pedestrians.

In order to solve the above problems, a first motion prediction device and a motion prediction program according to the present invention detect the current position of a movable object existing in the direction of travel of a vehicle, and detect the future position of the movable object. detecting an obstacle that can restrict the movement of the movable object; detecting an environment in which the movable object is placed; and when it is detected that the movable object may be within the obstacle; Virtually moving the current position of the detected movable object in a direction approaching the vehicle and to a position outside the obstacle; and controlling the current position of the moved movable object and the obstacle. predicting a plurality of future movements of the movable object based on the position of the object and the position of the environment; and based on the predicted future movements of the movable object, the future movement of the movable object. and displaying the status of.

A second motion prediction device and a motion prediction program according to the present invention detect the current position of a movable object existing in the direction of travel of a vehicle, and detect obstacles that can restrict future movement of the movable object. detecting the environment in which the movable object is placed; and when it is detected that the movable object may be within the obstacle, detecting the current state of the detected movable object; Virtually moving the position in a direction approaching the vehicle and to a position outside the obstacle; and the current position of the moved movable object, the position of the obstacle, and the environment. predicting a plurality of future movements of the movable object based on the position of the movable object; and predicting a plurality of future movements of the movable object based on the positions of assisting the vehicle to avoid the intersection when the vehicle is about to make an intersection.

A third motion prediction device and a motion prediction program according to the present invention provide information about the current state of a movable object existing in the direction in which one vehicle is traveling and in the direction in which another vehicle facing the one vehicle is traveling. detecting the position of the movable object; detecting an obstacle that can restrict future movement of the movable object; detecting an environment in which the movable object is placed; and whether the movable object is within the obstacle. When a possibility is detected, the current position of the detected movable object is virtually moved in a direction approaching the one vehicle and to a position straddling the inside and outside of the obstacle. and predicting a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment; and displaying a future movement status of the movable object based on a plurality of future movements.

A fourth motion prediction device and a motion prediction program according to the present invention provide a fourth motion prediction device and a motion prediction program for detecting the current state of a movable object existing in a direction in which one vehicle is traveling and in a direction in which another vehicle facing the one vehicle is traveling. detecting the position of the movable object; detecting an obstacle that can restrict future movement of the movable object; detecting an environment in which the movable object is placed; and whether the movable object is within the obstacle. When a possibility is detected, the current position of the detected movable object is virtually moved in a direction approaching the one vehicle and to a position straddling the inside and outside of the obstacle. and predicting future movements of the plurality of movable objects based on the current position of the moved movable object, the position of the obstacle, and the position of the environment; When the progress of at least one vehicle among the progresses of the other vehicles and at least one future movement of the plurality of predicted future movements of the movable object will intersect with each other; and assisting the one vehicle to avoid the intersection.

A fifth motion prediction device and a motion prediction program according to the present invention are the first to fourth motion prediction devices and motion prediction programs described above, and are one of the predicted future movements of the movable object. The method further includes supplementing an alternative future movement in place of the one future movement when the future movement of the first one is restricted by the obstacle.

A sixth motion prediction device and motion prediction program according to the present invention are the fifth motion prediction device and motion prediction program described above, and as the complement, the predicted motion prediction device and the motion prediction program are When the length of a line segment connecting points reached by the animal's future movement exceeds a predetermined threshold, a point reached by the alternative future movement is generated on the line segment.

According to the first motion prediction device and motion prediction program according to the present invention, the current position of the detected movable object is virtually moved in a direction approaching the vehicle and to a position outside the obstacle. Furthermore, the future movement status of the movable object is displayed. This makes it possible to predict the future movement of the movable object more accurately and to recognize it more accurately than in the prior art in which the movement and display are not performed.

According to the second motion prediction device and motion prediction program according to the present invention, the current position of the detected movable object is virtually moved in a direction approaching the vehicle and to a position outside the obstacle. and further assist the vehicle in avoiding the intersection. This makes it possible to predict the future movement of the movable object more accurately and to more accurately support the avoidance of the intersection, compared to the conventional technology that does not perform the movement and support.

According to the third motion prediction device and motion prediction program according to the present invention, the current position of the detected movable object is set in a direction approaching the one vehicle and at a position straddling the inside and outside of the obstacle. , to virtually move the movable object, and further to display the future movement status of the movable object. This makes it possible to more accurately determine the future movement of the movable object from both the perspective of the one vehicle and the perspective of the other vehicle, compared to conventional technology that does not perform the movement and display. can be predicted and recognized more accurately.

According to the fourth motion prediction device and motion prediction program according to the present invention, the current position of the detected movable object is set in a direction approaching the one vehicle and at a position straddling the inside and outside of the obstacle. , virtually moving and further assisting said one vehicle to avoid said intersection. This makes it possible to more accurately determine the future movement of the movable object from both the perspective of the one vehicle and the perspective of the other vehicle, compared to conventional technology that does not perform the movement and support. It is possible to predict the intersection and more accurately support the avoidance of the intersection.

According to the fifth motion prediction device and motion prediction program according to the present invention, an alternative future movement is supplemented in place of the first future movement. As a result, the future movement of the movable object can be predicted more accurately than in the prior art that does not perform the interpolation.

According to the sixth motion prediction device and motion prediction program according to the present invention, it is possible to complement the alternative future movement by generating a destination point according to the alternative future movement on the line segment. .

4 illustrates relationships between vehicles, movable objects, obstacles, and environments in an embodiment. 1 shows a configuration of a motion prediction device UYS according to an embodiment. It is a functional block diagram of motion prediction device UYS of an embodiment. It is another functional block diagram of motion prediction device UYS of an embodiment. It is a flow chart showing operation of motion prediction device UYS of an embodiment. The operation of the motion prediction device UYS of the embodiment is shown. The content displayed by the motion prediction device UYS of the embodiment is shown. The operation of the motion prediction device UYS of modification example 1 is shown. 12 is a flowchart showing the operation of the motion prediction device UYS of Modification 2. The principle of operation of support by the motion prediction device UYS of modification 2 is shown. The content supported by the motion prediction device UYS of Modification 2 is shown. The operation (original operation) of a conventional motion prediction device is shown. The operation (possible operation) of a conventional motion prediction device is shown.

An embodiment of a motion prediction device according to the present invention will be described.

<Embodiment>
A motion prediction device according to an embodiment will be described.

<Positional relationships among vehicles, movable objects, obstacles, and environment>
FIG. 1 shows the relationship between a vehicle, a movable object, an obstacle, and an environment in an embodiment.

As shown in FIG. 1, vehicle SR(A) is traveling along traveling direction SH. Vehicle SR (B), which is an oncoming vehicle of vehicle SR (A), is traveling in the opposite direction to the traveling direction SH.

In order to predict the future movement of the pedestrian HK, which is an example of a "movable object," and the building TS, which is an example of an "obstacle" that can restrict the future movement of the pedestrian HK. It exists near a crosswalk CW, which is an example of a useful "environment." Here, in addition to the building TS, "obstacles" include, for example, roadside objects (e.g., plants), parked vehicles (especially large parked vehicles), and "environment" refers to crosswalks. In addition to CW, examples include road markings and traffic lights.

Vehicles SR (A) and SR (B) are equipped with cameras CM (A) and CM (B), respectively. A camera CM (TS) is installed in the building TS.

<Configuration of embodiment>
FIG. 2 shows the configuration of the motion prediction device UYS of the embodiment. It does not matter whether the motion prediction device UYS is installed inside the vehicle SR(A) or SR(B), installed on the roof of the building TS, or installed near the crosswalk CW. The motion prediction device UYS may also be disposed in a distributed manner inside the vehicles SR(A) and SR(B), on the rooftop of the building TS, and near the crosswalk CW.

The configuration of the motion prediction device UYS according to the embodiment will be described below with reference to FIG. 2.

The motion prediction device UYS of the embodiment includes an input section 1, a CPU 2 (Central Processing Unit), an output section 3, a storage medium 4, and a memory 5, as shown in FIG.

The input unit 1 includes, for example, a touch panel and a keyboard. CPU2 is an example of a processor, and is the core of a well-known computer that operates hardware according to software. The output unit 3 includes, for example, a touch panel and a liquid crystal monitor. The storage medium 4 includes, for example, a hard disk drive (HDD), a solid state drive (SSD), and a ROM (Read Only Memory). The memory 5 includes, for example, DRAM (Dynamic Random Access Memory) and SRAM (Static Random Access Memory).

The storage medium 4 stores the program PR. The program PR is a group of instructions that defines the content of processing to be executed by the motion prediction device UYS.

FIG. 3 is a functional block diagram of the motion prediction device UYS according to the embodiment.

As shown in FIG. 3, the motion prediction device UYS includes a detection section 11, a movement section 12, a prediction section 13, a complementation section 14, a display section 15, a support section 16, a control section 17, and a storage section. 18.

Regarding the relationship between the hardware configuration and the functional configuration in the motion prediction device UYS, in terms of the relationship between the hardware configuration and the functional configuration, on the hardware, the CPU 2 stores the information stored in the storage medium 4 (which implements some functions of the storage unit 18). The program PR is executed using the memory 5 (which realizes some other functions of the storage section 18), and the control section 17 controls the operations of the input section 1 and the output section 3 as necessary. By doing so, the functions of the detection unit 11, movement unit 12, prediction unit 13, complementation unit 14, display unit 15, and support unit 16 are realized. The functions of each part will be explained together with the operation of the motion prediction device UYS.

FIG. 4 is another functional block diagram of the motion prediction device UYS according to the embodiment.

In relation to the functional block diagram shown in FIG. 3, the motion prediction device UYS includes a sensing section 21, a map information section 22, an own vehicle state estimation section 23, and a movable object state generation unit, as shown in FIG. 24, a physical movement prediction unit 25, a spatial movement restriction unit 26, a physical state update unit 27, a movable object state complementation unit 28, an audiovisual stimulus generation unit 29, and a vehicle control generation unit 30. . The functions of each part will be described later.

<Operation of embodiment>
FIG. 5 is a flowchart showing the operation of the motion prediction device UYS of the embodiment.

FIG. 6 shows the operation of the motion prediction device UYS of the embodiment.

The operation of the motion prediction device UYS according to the embodiment will be described below with reference to FIGS. 5 and 6.

Step S11: In the motion prediction device UYS, the CPU 2 (shown in FIG. 2), as the detection unit 11 (shown in FIG. 3), detects the current position of the pedestrian HK as shown in FIG. 6(A). To detect. Thereby, similarly to the prior art described above with reference to FIG. 13, it is detected that the pedestrian HK is at a position IT far from the vehicle SR(A) along the traveling direction SH of the vehicle SR(A). Here, the detection unit 11 that detects the pedestrian HK corresponds to a “first detection unit”.

The CPU 2 serves as the detection unit 11 and also detects the position, size, etc. of the building TS. Here, the detection unit 11 that detects the building TS corresponds to a “second detection unit”.

The CPU 2, as the detection unit 11, further detects the position, length, width, etc. of the crosswalk CW. Here, the detection unit 11 that detects the crosswalk CW corresponds to a "third detection unit".

Step S12: When the position of the pedestrian HK, the position of the building TS, etc., the position of the crosswalk CW, etc. are detected in step S11, more specifically, when the position of the pedestrian HK, the position of the building TS, etc., and the position of the crosswalk CW are detected, Due to erroneous detections similar to "possible operation of the device", it is not possible to determine whether the pedestrian HK may be inside the building TS, regardless of whether the pedestrian HK is actually inside the building TS. When detected, the CPU 2, as the moving unit 12 (shown in FIG. 3), detects five reference points KJ11 to KJ15 representing the position IT of the pedestrian HK, as shown in FIG. 6(B). The vehicle SR(A) is virtually moved along a direction substantially opposite to the traveling direction SH. More precisely, the CPU 2 sets the five reference points KJ11 to KJ15 so that they are along the virtual straight line KT connecting the vehicle SR(A) and the pedestrian HK and close to the vehicle SR(A). Approximately 100% of pedestrian HK's whole body, in other words, is moved so that almost all five of the reference points (five reference points KJ21 to KJ25 described later) after movement are located outside the structure TS. let As a result, five new reference points KJ21 to KJ25 are generated.

Step S13: In step S12, when new five reference points KJ21 to KJ25 are generated by the virtual movement of the five reference points KJ11 to KJ15, the CPU 2 operates as the prediction unit 13 (shown in FIG. 3). , as shown in Fig. 6(C), predicts five future movements of pedestrian HK. More specifically, as shown in FIG. 6(C), the CPU 2 determines whether the person is walking based on the five reference points KJ21 to KJ25 and with reference to the position of the building TS, the position of the crosswalk CW, etc. Predict HK's 5 future moves UG11 to UG15.

Step S14: When the five future movements UG11 to UG15 of the pedestrian HK are predicted in step S13, the CPU 2 predicts that the five future movements UG11 to UG15 of the pedestrian HK are based on the presence of the building TS (for example, It is determined whether or not it is restricted by the walls, enclosures, etc. (not shown) of the building TS. When any of the five future movements UG11 to UG15 of the pedestrian HK is restricted due to the presence of the building TS, the CPU 2, as a supplementary unit 14 (shown in FIG. 3), determines an alternative future movement. Complement.

As shown in FIG. 6(D), none of the five future movements UG11 to UG15 of the pedestrian HK is restricted by the presence of the building TS. Therefore, the CPU 2 does not generate any alternative future movement, that is, does not complement it, as shown in FIG. 6(E).

Step S15: After the alternative future movement is not supplemented in step S14, the CPU 2 displays the vehicle on the display unit 15 (shown in FIG. 3) based on the five future movements UG11 to UG15 of the pedestrian HK. In addition to SR(A), SR(B), building TS, and crosswalk CW, the future movement status UJ of pedestrian HK is displayed on an output unit (shown in FIG. 2) such as a liquid crystal monitor.

FIG. 7 shows the content displayed by the motion prediction device UYS of the embodiment.

In the motion prediction device UYS, the CPU 2 displays the future movement situation UJ of the pedestrian HK based on the above-mentioned five future movements UG11 to UG15 of the pedestrian HK. Here, it is assumed that the position and speed of vehicle SR (A), the position and speed of vehicle SR (B), etc. are also detected separately from steps S11 to S14 described above. Under this assumption, the CPU 2 calculates the following cases based on the five future movements UG11 to UG15 of the pedestrian HK, the position and speed of the vehicle SR (A), the position and speed of the vehicle SR (B), etc. The future movement situation UJ of the pedestrian HK is displayed for each division.

(1) Display on vehicle SR (A) (1-1) When vehicle SR (A) is expected to be unable to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 7 (A))
(1-2) When pedestrian HK is expected not to cross the crosswalk CW (Figure 7 (B))
(1-3) When vehicle SR (A) is expected to be able to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 7 (C))

(2) Display on vehicle SR (B) (2-1) When vehicle SR (B) is expected to be unable to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 7 (D))
(2-2) When pedestrian HK is expected not to cross the crosswalk CW (Figure 7 (E))
(2-3) When vehicle SR (B) is expected to be able to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 7 (F))

<Effects of embodiment>
As described above, the motion prediction device UYS of the embodiment moves the five reference points KJ11 to KJ15 of the pedestrian HK along the virtual straight line KT so as to approach the vehicle SR(A). As a result, the future movement situation UJ of the pedestrian HK (the irregularly hatched area shown in Figures 7(A) to 7(F)) can be displayed with higher accuracy than before. , the future movement situation UJ of the pedestrian HK can be made to be recognized by the driver of the vehicle SR (A) and the driver of the vehicle SR (B) with higher accuracy than in the past.

<Function of embodiment (supplement)>
The functions of the sensing unit 21 to vehicle control generation unit 30 and the correspondence relationship between the sensing unit 21 to vehicle control generation unit 30 and the detection unit 11 to support unit 16 (shown in FIG. 3) shown in FIG. 4 are as follows. , are as follows.

<Sensing section 21>
The sensing unit 21 is, for example, a camera installed on the back of the room mirror of the vehicle SR (A) and the vehicle SR (B) facing forward (in the direction of travel), and detects the coordinates of a rectangular area where the pedestrian HK exists in the image. is calculated in real time. It is also desirable that the sensing unit 21 has an identification function that classifies the direction of the pedestrian HK into front, back, right, and left directions. The sensing unit 21 may further determine and track the identity using the luminance distribution of pixels within a rectangular area in consecutive frames. The sensing unit 21 has a 360-degree laser radar installed on the ceiling of the vehicle, measures the positions of surrounding objects as a point group, and creates an occupancy grid map by time integration.

The sensing unit 21 may detect the pedestrian HK using at least one of the forward radar and the surrounding radar, and may also generate a so-called occupancy grid map. The sensing unit 21 may also detect pedestrians HS using a combination of radar, camera, laser radar, etc., and may also generate an occupancy grid map.

The sensing unit 21 includes cameras CM (A), CM (B), and CM (TS) (shown in FIG. 1). The sensing section 21 mainly corresponds to the detection section 11 (shown in FIG. 3).

<Map information department 22>
The map information unit 22 uses an electronic map obtained by accumulating a laser point cloud measured in advance by another vehicle equipped with a high-precision positioning system, and information that may affect the movement of pedestrians HK. , provides road markings such as white lines that are manually tagged from camera images and satellite photographs, and a line segment map that describes obstacles using line segments. The map information section 22 describes each point group and line segment in an orthogonal plane geodetic system. The map information section 22 mainly corresponds to the detection section 11 (shown in FIG. 3).

<Vehicle state estimation unit 23>
The own vehicle state estimating unit 23 estimates the own vehicle position and attitude using the measurement point group of the laser radar.

The own vehicle state estimation unit 23 estimates the position and orientation by matching the detection result by the camera with the electronic map, and also uses GPS/IMU (GPS stands for Global Positioning System. IMU stands for Inertial Measurement System). The position and orientation may be estimated using a unit (inertial measurement device).

The own vehicle state estimating unit 23 mainly corresponds to detecting the positions, speeds, etc. of the vehicles SR(A) and SR(B) described above.

<Movable object state generation unit 24>
The movable object state generation unit 24 calculates the pedestrian position as an ellipsoid in a three-dimensional space based on the external and internal parameters of the camera and assuming that the height of the rectangular area of the pedestrian is an average height (1.7 m). Estimate. The movable object state generation unit 24 also sets a total of five points, including the center point, focal points on the major axis, and focal points on the minor axis, as the initial positions of the movement prediction candidates, of the ellipse obtained by cutting the ellipsoid in a cross section. The movable object state generation unit 24 converts the state of the pedestrian HK into a four-dimensional state variable consisting of a two-dimensional position and two-dimensional velocity on a cross section, and a covariance matrix of a 4×4 state variable. If an obstacle area of the line segment map or the occupancy exercise map exists on the line segment connecting each candidate point and the origin of the own vehicle camera, the movable object state generation unit 24 moves the candidate point to the intersection closest to the origin. . The movable object state generation unit 24 translates all candidate points until there are no intersections. At this time, the movable object state generation unit 24 may set a constant margin for the candidate point closest to the obstacle. When a set of candidate points before and after movement prediction is input, the movable object state generation unit 24 performs similar processing on a line segment connecting the candidate points before and after movement. The movable object state generating section 24 mainly corresponds to the moving section 12 (shown in FIG. 3).

<Physical movement prediction unit 25>
The physical movement prediction unit 25 evolves state variables over time for the five pedestrian movement prediction candidates already held, and predicts future positions in time series using a linear movement model. The time step is 0.25 seconds/step. The physical movement prediction unit 25 performs prediction up to 4 seconds ahead of 16 steps, which is the number of prediction steps. The physical state prediction unit 25 mainly corresponds to the prediction unit 13 (shown in FIG. 3).

<Spatial movement restriction unit 26>
For each pedestrian movement prediction candidate, if there is a line segment connecting the positions before and after the prediction and an obstacle area on the line segment map or the occupancy exercise map, the spatial movement restriction unit 26 sets Replace the nearest intersection as the predicted position. The spatial movement restriction section 26 mainly corresponds to the complementation section 14 (shown in FIG. 3).

<Physical state update unit 27>
The physical state updating unit 27 uses the Hungarian method to set the initial speed to 0 m/sec when there is no candidate point close to the already held predicted candidate point and when no pedestrian-oriented identification result is obtained. When obtained, the general walking speed in that direction is set to 1 m/sec, and the point is set as a new predicted candidate point. If there is a candidate point close to the prediction candidate point already held, the physical state update unit 27 updates the predicted point using a Kalman filter. The physical state update section 27 mainly corresponds to the complement section 14 (shown in FIG. 3).

<Movable object state complementation unit 28>
If the length of a line segment connecting candidate points that are focal points generated from the same ellipse exceeds a threshold, the movable object state complementing unit 28 generates a new candidate point on the line segment. At this time, if there is an intersection between this line segment and the line segment map or the occupancy exercise map, the movable object state complementing unit 28 changes the position of the generated candidate point according to the distance from the end point to the intersection point. Also good. The movable object state complementation section 28 mainly corresponds to the complementation section 14 (shown in FIG. 3).

<Audiovisual stimulus generation unit 29>
When the distance of the normal line from the prediction candidate point to the line segment in the traveling direction of the own vehicle is less than or equal to the threshold value, the audiovisual stimulus generation unit 29 issues an alarm sound from the vehicle interior speaker. The audiovisual stimulus generation unit 29 also causes the center console display to blink. At this time, the audiovisual stimulus generation unit 29 displays the front camera image to emphasize the rectangular area that is the target of the warning, or draws the internal processing state in an overhead view from directly above the own vehicle. Also good. The audiovisual stimulus generation section 29 mainly corresponds to the display section 15 (shown in FIG. 3).

<Vehicle control generation unit 30>
If the distance of the normal line from the prediction candidate point to the line segment in the vehicle's traveling direction is less than or equal to the threshold, the vehicle control generation unit 30 decelerates at a constant acceleration to notify the driver and to notify the driver that a collision has occurred. Take measures to reduce damage. The vehicle control generation section 30 mainly corresponds to the support section 16 (shown in FIG. 3).

<Modification 1>
A motion prediction device according to modification 1 will be described.

<Configuration and function of modification 1>
The motion prediction device UYS of Modification 1 has the same configuration (shown in FIG. 2) and function (shown in FIGS. 3 and 4) as the motion prediction device UYS of the embodiment.

<Operation of modification 1>
The operation of the motion prediction device UYS of Modification 1 is basically the same as the operation of the motion prediction device UYS of the embodiment (shown in FIG. 5). On the other hand, the operation of the motion prediction device UYS of the first modification differs from the operation of the motion prediction device UYS of the embodiment in the process of step S12 and the process of step S14.

FIG. 8 shows the operation of the motion prediction device UYS of the first modification. The operation of the motion prediction device UYS of Modification 1 will be described below with reference to the flowchart of FIG. 5 (flowchart showing the operation of the embodiment) and FIG. 8.

Before step S12: Similar to the embodiment, the CPU 2 detects the current position of the pedestrian HK, the position of the building TS, and the position of the crosswalk CW.

Step S12: Similarly to step S12 of the embodiment, the CPU 2, as a moving unit (shown in FIG. 3), selects five reference points representing the position IT of the pedestrian HK, as shown in FIG. 8(B). KJ31 to KJ35 are moved along the direction opposite to the traveling direction SH of vehicle SR(A).

More precisely, the CPU 2 moves the five reference points KJ31 to KJ35 along the virtual straight line KT connecting the vehicle SR(A) and the pedestrian HK, as well as the vehicle SR( Move it so that it is closer to A).

Here, unlike step S12 of the embodiment, the CPU 2 controls the movement of the five reference points KJ31 to KJ35 so that approximately 70 to 80% of the whole body of the pedestrian HK moves, in other words, the reference points after movement ( This is done so that three to four of the five reference points KJ41 to KJ45) to be described later are located outside the building TS. In summary, the CPU 2 moves the five reference points KJ31 to KJ35 to positions spanning the inside and outside of the building TS. As a result, five new reference points KJ41 to KJ45 are generated.

Step S13: Similar to step S13 of the embodiment, the CPU 2 predicts five future movements UG31 to UG35 of the pedestrian HK.

Step S14: The CPU 2, as the complementing unit 14 (shown in FIG. 3), as shown in FIG. It is determined that it is limited by the existence of TS. In other words, the CPU 2 determines that the pedestrian HK will not be able to reach the reaching point TT31 of the pedestrian HK indicated by the movement UG31, as shown in FIG. 8(D).

As a result, as shown in FIG. 8(D), the CPU 2 determines that if the pedestrian HK moves and performs UG31 despite the existence of the building TS, instead of reaching the destination point TT31, the pedestrian HK will reach the destination point TT31. A modified destination ST31 that would be possible is generated.

Further, as shown in FIG. 8(E), the CPU 2 determines whether the movement is located next to the corrected arrival point ST31 and, for example, the movement UG31, in other words, if the risk of collision is high for the vehicle SR (A). The destination point TT51 that the pedestrian HK would be able to reach using the destination point TT34 of the movement UG34, which is an approximate movement and is not restricted by the existence of the building TS. generate.

The CPU 2 further connects the reference point KJ41 and the corrected arrival point ST51 to generate an alternative movement UG51 to replace the movement UG31, that is, to complement it.

Step S15: Similar to step S15 of the embodiment, the CPU 2 displays the future movement situation UJ of the pedestrian HK, as shown in FIGS. 7(A) to 7(F).

<Effects of modification 1>
As described above, the motion prediction device UYS of the first modification differs from the motion prediction device UYS of the embodiment in that the five reference points KJ31 to KJ35 of the pedestrian HK are moved to positions straddling the inside and outside of the building TS. let By moving all five reference points KJ11 to KJ15 of pedestrian HK outside the building TS, vehicle SR (A) and Unlike the motion prediction device UYS of the embodiment, which emphasizes the risk of collision between vehicle SR (A) and pedestrian HK, and the danger of collision between vehicle SR (A) and pedestrian HK, and the risk of collision between vehicle SR (B) and It becomes possible to balance the risk of collision with a pedestrian HK.

Furthermore, as described above, the motion prediction device UYS of the first modification supplements the alternative motion UG51 in place of the motion UG31 that is restricted due to the presence of the building TS, and therefore predicts the future motion of the pedestrian HK. The situation UJ can be more accurately notified to the driver of vehicle SR (A) and the driver of vehicle SR (B).

<Modification 2>
A motion prediction device according to modification 2 will be described.

<Configuration and function of modification 2>
The configuration and functions of the motion prediction device UYS of Modification 2 are similar to the configuration (shown in FIG. 2) and functions (shown in FIGS. 3 and 4) of the motion prediction device UYS of the embodiment.

<Operation of modification 2>
The operation of the motion prediction device UYS of the second modification is basically the same as the operation of the motion prediction device UYS of the first modification (shown in FIG. 5), but step S15 of the first modification (shown in FIG. 5) ), step S25 (described later with reference to FIG. 9) is included.

FIG. 9 is a flowchart showing the operation of the motion prediction device UYS of the second modification. Hereinafter, the operation of the motion prediction device UYS of the second modification will be described with reference to the flowchart of FIG. 9 and FIG. 8 (which shows the operation of the first modification).

Step S21: Similar to step S11 of modification 1, the CPU 2 (shown in FIG. 2), as the detection unit 11 (shown in FIG. 3), detects the pedestrian HK as shown in FIG. 8(A). The position of the building TS, the position of the crosswalk CW, etc. are detected.

Step S22: Similar to step S12 of the first modification, the CPU 2, as the moving unit 12 (shown in FIG. 3), sets five criteria representing the position IT of the pedestrian HK, as shown in FIG. 8(B). Points KJ31 to KJ35 are moved to positions spanning the inside and outside of the building TS. As a result, five new reference points KJ41 to KJ45 are generated.

Step S23: Similar to step S13 of the first modification, the CPU 2, as the prediction unit 13 (shown in FIG. 3), sets the five reference points KJ41 to KJ45, the building TS, and Five future movements UG31 to UG35 of the pedestrian HK are predicted based on the position of the pedestrian crossing CW and the position of the crosswalk CW.

Step S24: Similar to step S14 of Modification 1, the CPU 2, as the complementing unit 14 (shown in FIG. 3), calculates the movement UG31 of the pedestrian HK from the building TS as shown in FIG. 8(D). Therefore, as shown in FIG. 8(E), an alternative motion UG51 is supplemented to replace the motion UG31.

Step S25: Different from step S15 of the first modification, the CPU 2, as the support unit 16 (shown in FIG. 3), uses the vehicle SR(A), This helps avoid the intersection between SR(B) and pedestrian HK.

Here, "avoiding the intersection" means, for example, stopping the progress of the vehicles SR (A) and SR (B) by decelerating the vehicles SR (A) and SR (B), and This refers to changing the direction of travel of vehicles SR(A) and SR(B) by steering vehicles SR(A) and SR(B).

FIG. 10 shows the principle of support operation by the motion prediction device UYS of the second modification.

(1) Simulation for supporting vehicle SR(A) In the motion prediction device UYS, as shown in FIG. The distance between the movements UG51 and UG32 to UG35 (shown in FIG. 8(E)), (b) the speed and deceleration of the vehicle SR (A), and (c) the five future movements of the pedestrian HK. Based on UG51 and UG32 to UG35, it is simulated whether the vehicle SR (A) can stop before the crosswalk CW when it decelerates at a deceleration of 0.3 G or less. In summary, it is simulated whether the progress of the vehicle SR(A) intersects with any of the five future movements UG51, UG32 to UG35.

(B) Simulation for supporting the vehicle SR (B) In the motion prediction device UYS, the CPU 2 performs the simulation in FIG. As shown in, for example, (a) the distance between the vehicle SR (B) and the movement of the pedestrian HK UG51, UG32 to UG35, (b) the speed and deceleration of the vehicle SR (B), and ( c) Five future movements of pedestrian HK Based on UG51, UG32 to UG35, whether vehicle SR (B) can stop before the crosswalk CW when it decelerates at a deceleration of 0.3G or less. to simulate. In summary, it is simulated whether the progress of the vehicle SR(A) intersects with any of the five future movements UG51, UG32 to UG35.

FIG. 11 shows the content supported by the motion prediction device UYS of the second modification.

In the motion prediction device UYS, the CPU 2 performs the following support (including display) in the following cases, which are classified according to the results of the simulation described above.

(1) Support for vehicle SR(A) (1-1) When vehicle SR(A) is expected to be unable to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 11 (A))
The driver of the vehicle SR(A) is alerted to stop the vehicle SR(A) from moving forward, or the vehicle SR(A) is forcibly decelerated.

(1-2) When pedestrian HK is expected not to cross the crosswalk CW (Figure 11(B))
The situation of pedestrians HK, buildings TS, crosswalks CW, etc. is displayed to the driver of vehicle SR (A).

(1-3) When vehicle SR (A) is expected to be able to stop before the crosswalk CW due to deceleration of 0.3 G or less (Figure 11 (C))
Similarly to (1-2) above, the status of pedestrians HK, buildings TS, crosswalks CW, etc. is displayed to the driver of vehicle SR (A).

(2) Support for vehicle SR (B) (2-1) When vehicle SR (B) is expected to be unable to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 11 (D))
Similarly to (1-1) above, the driver of vehicle SR (B) is alerted to stop the movement of vehicle SR (B), or vehicle SR (B) is forcibly decelerated. let

(2-2) When pedestrian HK is expected not to cross the crosswalk CW (Figure 11 (E))
Similar to (1-2) above, the situation of pedestrians HK, buildings TS, crosswalks CW, etc. is displayed to the driver of vehicle SR (B).

(2-3) When vehicle SR (B) is expected to be able to stop before the crosswalk CW due to deceleration of 0.3G or less (Figure 11 (F))
Unlike (1-3) above, and similar to (1-1) and (2-1) above, the driver of vehicle SR (B) is instructed to proceed with vehicle SR (B). Call attention to the vehicle to stop, or force the vehicle SR(B) to decelerate.

In (2-3) above, the driver of vehicle SR (A) can visually recognize the pedestrian HK, unlike in (1-3) above, when the driver of vehicle SR (B) , the pedestrian HK cannot be visually recognized due to the presence of the building TS. Therefore, in order to compensate for the low visibility of vehicle SR(B), instead of the display in (1-3) above, a warning is called and forced deceleration is performed.

<Effects of modification example 2>
As described above, the motion prediction device UYS of the second modification differs from the motion prediction device UYS of the embodiment and the motion prediction device UYS of the first modification in that, in addition to displaying the situation of the pedestrian HK, In order to stop the progress of SR (A) and vehicle SR (B), the driver of vehicle SR (A) and the driver of vehicle SR (B) are alerted, and the driver of vehicle SR (A) and vehicle SR (B) are alerted. (B) Forcibly decelerates. This makes it possible to more accurately prevent a collision between the vehicle SR(A) and the pedestrian HK, and a collision between the vehicle SR(B) and the pedestrian HK.

<Modification 3>
In step S22 of the second modification described above, instead of moving the reference points KJ31 to KJ35 to positions that span the inside and outside of the building TS, similarly to step S12 of the embodiment, they are moved outside the building TS. Instead of supplementing the alternative movement UG51 in step S24 of the second modification, there is no need to perform the supplementation, as in step S12 of the embodiment, and even if no supplementation is performed, It is possible to obtain the same effects as in the second modification described above.

<Supplementary explanation of processor and program>
In the embodiments described above, the processor refers to a processor in a broad sense. In addition to general-purpose processors (e.g., CPU: Central Processing Unit, etc.), dedicated processors (e.g., GPU: Graphics Processing Unit, ASIC: Application Specific Integrated Circuit, FPGA: Field Programmable Gate Array, programmable logic device, etc.) include.

In the embodiments described above, the operations of the processor may be realized by one processor, or may be realized by cooperation of multiple processors. Further, the order of each operation of the processor is not limited to the order in the above-described embodiments, and may be changed as appropriate.

In the embodiment described above, the program PR is stored (installed) in the storage medium 4 in advance, and instead of being stored in the storage medium 4, the program PR is stored in a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disc Read Only Memory), The information may be provided recorded on a recording medium such as a USB (Universal Serial Bus) memory, or may be downloaded from an external device via a network.

SR (A), SR (B) Vehicle HK Pedestrian TS Building CW Crosswalk UYS Motion prediction device 11 Detection section 12 Movement section 13 Prediction section 14 Completion section 15 Display section 16 Support section 17 Control section 18 Storage section

Claims (12)

a first detection unit that detects the current position of a movable object existing in the direction in which the vehicle is traveling;
a second detection unit that detects an obstacle that can restrict future movement of the movable object;
a third detection unit that detects the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the vehicle and to a position outside the obstacle. , a moving unit that moves virtually;
a prediction unit that predicts a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
a display unit that displays the status of the future movement of the movable object based on the predicted future movements of the movable object;
A motion prediction device including: a first detection unit that detects the current position of a movable object existing in the direction in which the vehicle is traveling;
a second detection unit that detects an obstacle that can restrict future movement of the movable object;
a third detection unit that detects the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the vehicle and to a position outside the obstacle. , a moving unit that moves virtually;
a prediction unit that predicts a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
When the progress of the vehicle and at least one future movement of the plurality of predicted future movements of the movable object will intersect with each other, assisting the vehicle to avoid the intersection. Support department and
A motion prediction device including: a first detection unit that detects the current position of a movable object existing in the traveling direction in which one vehicle is traveling and in the traveling direction in which another vehicle opposing the one vehicle is traveling;
a second detection unit that detects an obstacle that can restrict future movement of the movable object;
a third detection unit that detects the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the one vehicle and inside or outside the obstacle. a moving part that virtually moves to a straddling position;
a prediction unit that predicts a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
a display unit that displays the status of the future movement of the movable object based on the predicted future movements of the movable object;
A motion prediction device including: a first detection unit that detects the current position of a movable object existing in the traveling direction in which one vehicle is traveling and in the traveling direction in which another vehicle opposing the one vehicle is traveling;
a second detection unit that detects an obstacle that can restrict future movement of the movable object;
a third detection unit that detects the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the one vehicle and inside or outside the obstacle. a moving part that virtually moves to a straddling position;
a prediction unit that predicts a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
The progress of at least one of the progress of the one vehicle and the progress of the other vehicle intersects with at least one future movement of the plurality of predicted future movements of the movable object. a support unit that assists the one vehicle in avoiding the intersection when the one vehicle would do so;
A motion prediction device including: When one future movement of the plurality of predicted future movements of the movable object is restricted by the obstacle, a complementing unit complements an alternative future movement to replace the one future movement. The motion prediction device according to any one of claims 1 to 4, further comprising the following. If the length of a line segment connecting points reached by the predicted future movement of the movable object, which are focal points generated from the same ellipse, exceeds a predetermined threshold, the complement unit The motion prediction device according to claim 5, wherein the motion prediction device generates a destination according to the alternative future motion above. to the computer,
Detects the current position of a movable object in the direction of travel of the vehicle,
detecting an obstacle that can restrict future movement of the movable object;
detecting the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the vehicle and to a position outside the obstacle. , virtually moved;
predicting a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
Displaying the future movement status of the movable object based on the predicted future movements of the movable object;
A motion prediction program for executing processing. to the computer,
Detects the current position of a movable object in the direction of travel of the vehicle,
detecting an obstacle that can restrict future movement of the movable object;
detecting the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the vehicle and to a position outside the obstacle. , virtually moved;
predicting a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
When the progress of the vehicle and at least one future movement of the plurality of predicted future movements of the movable object will intersect with each other, assisting the vehicle to avoid the intersection. do,
A motion prediction program for executing processing. to the computer,
Detecting the current position of a movable object existing in the traveling direction in which one vehicle is traveling and in the traveling direction in which another vehicle opposing the one vehicle is traveling;
detecting an obstacle that can restrict future movement of the movable object;
detecting the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the one vehicle and inside or outside the obstacle. Virtually move it to a straddling position,
predicting a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
Displaying the future movement status of the movable object based on the predicted future movements of the movable object;
A motion prediction program for executing processing. to the computer,
Detecting the current position of a movable object existing in the traveling direction in which one vehicle is traveling and in the traveling direction in which another vehicle opposing the one vehicle is traveling;
detecting an obstacle that can restrict future movement of the movable object;
detecting the environment in which the movable object is placed;
When it is detected that the movable object may be within the obstacle, the current position of the detected movable object is moved in a direction approaching the one vehicle and inside or outside the obstacle. Virtually move it to a straddling position,
predicting a plurality of future movements of the movable object based on the current position of the moved movable object, the position of the obstacle, and the position of the environment;
The progress of at least one of the progress of the one vehicle and the progress of the other vehicle intersects with at least one future movement of the plurality of predicted future movements of the movable object. assisting the one vehicle to avoid the intersection when it would do so;
A motion prediction program for executing processing. When one future movement of the plurality of predicted future movements of the movable object is restricted by the obstacle, an alternative future movement is supplemented to replace the one future movement. The motion prediction program according to any one of claims 7 to 10. The interpolation is performed on the line segment when the length of the line segment connecting the points reached by the predicted future movement of the movable object, which are focal points generated from the same ellipse, exceeds a predetermined threshold. 12. The motion prediction program according to claim 11, wherein the program generates a destination based on the alternative future movement.