Vehicle collision detection device

Abstract

To provide a vehicle collision detection device capable of accurately determining collisions with vulnerable road users.SOLUTION: A processor 152 of an ECU 150 includes: an object probability calculation unit 152b that calculates a first probability indicating the likelihood that an object in front of a vehicle is a pedestrian or a bicycle; a collision probability calculation unit 152c that calculates a second probability indicating the probability of the vehicle colliding with the object; a collision prediction probability calculation unit 152a that calculates a collision prediction probability indicating the probability of colliding with the pedestrian or the bicycle based on the first probability and the second probability; a collision determination threshold acquisition unit 152d that acquires a collision determination threshold for the vehicle's deceleration based on the collision prediction probability; and a collision determination unit 152e that determines whether the vehicle collided with the pedestrian or the bicycle based on the detected value of the vehicle's deceleration and the collision determination threshold.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a vehicle collision determination device. [Background technology]

[0002] It is known that after detecting a pedestrian or bicycle as a collision target, if the deceleration (front Gf) set according to the vehicle's speed within the collision prediction time range exceeds a judgment threshold, it is determined that a collision has occurred with a vulnerable road user (see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-169016 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in the above patent document determines that a collision with a vulnerable road user has occurred when the deceleration exceeds a determination threshold, without taking into consideration the likelihood that the object being hit is a pedestrian, a bicycle, or other vulnerable road user. As a result, for example, based on the deceleration detected when traveling on a rough road, it may be erroneously determined that a collision with a vulnerable road user has occurred even when no collision with a vulnerable road user has occurred.

[0005] SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide a vehicle collision determination device that can accurately determine a collision with a vulnerable road user. [Means for solving the problem]

[0006] The gist of the present disclosure is as follows.

[0007] (1) an object probability calculation unit that calculates a first probability that an object in front of the vehicle is a pedestrian or a bicycle; a collision probability calculation unit that calculates a second probability that the vehicle will collide with the object; a collision prediction accuracy calculation unit that calculates a collision prediction accuracy, which is the accuracy of a collision with a pedestrian or a bicycle, based on the first accuracy and the second accuracy; a collision determination threshold acquisition unit that acquires a collision determination threshold of a vehicle deceleration based on the collision prediction accuracy; a collision determination unit that determines whether the vehicle has collided with a pedestrian or a bicycle based on the detected value of the deceleration of the vehicle and the collision determination threshold; A vehicle collision determination device comprising:

[0008] (2) the collision determination threshold acquisition unit acquires a map that defines the collision determination threshold according to the vehicle speed based on the collision prediction accuracy; The vehicle collision determination device described in (1) above, wherein the collision determination unit applies the detected values ​​of vehicle speed and vehicle deceleration to the map, and determines that the vehicle has collided with a pedestrian or bicycle if the vehicle deceleration is greater than the collision determination threshold.

[0009] (3) The vehicle collision determination device according to (2) above, wherein the collision determination threshold acquisition unit acquires a map in which the collision determination threshold in a low vehicle speed range is smaller as the collision prediction accuracy is higher.

[0010] (4) A vehicle collision determination device as described in (2) or (3) above, wherein the collision determination threshold acquisition unit acquires a map in which the collision determination threshold in the medium vehicle speed range or the high vehicle speed range is larger as the collision prediction accuracy is lower. [Effects of the Invention]

[0011] According to the present invention, a vehicle collision determination device is provided that can accurately determine a collision with a vulnerable road user. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a schematic configuration diagram of a vehicle driving assistance system according to one embodiment; [Figure 2]FIG. 2 is a schematic diagram showing functional blocks of a processor of an ECU. [Figure 3] 4 is a flowchart showing a process performed by a processor of the ECU at each predetermined control period. [Figure 4] 10 is a timing chart for explaining an example in which a collision prediction accuracy calculation unit calculates a level of collision prediction accuracy using a control state of PCS control and a collision prediction time TTC. [Figure 5] 10 is a diagram showing an example of a map acquired by a collision determination threshold acquisition unit based on a collision prediction probability Lv1. FIG. [Figure 6] 10 is a diagram showing an example of a map acquired by a collision determination threshold acquisition unit based on a collision prediction probability Lv2. FIG. [Figure 7] 10 is a diagram showing an example of a map acquired by a collision determination threshold acquisition unit based on a collision prediction probability level Lv3. FIG. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, several embodiments of the present invention will be described with reference to the drawings. However, these descriptions are intended to be merely examples of preferred embodiments of the present invention and are not intended to limit the present invention to such specific embodiments.

[0014] 1 is a schematic configuration diagram of a vehicle cruise assistance system 1000 according to one embodiment. The cruise assistance system 1000 is mounted on a vehicle such as an automobile, and includes an on-board camera 110, a periphery monitoring sensor 120, a vehicle speed sensor 130, an electronic control unit (ECU, hereinafter referred to as ECU) 150, a G sensor 160, and a warning device 170. The on-board camera 110, the periphery monitoring sensor 120, the vehicle speed sensor 130, the ECU 150, the G sensor 160, and the warning device 170 are each connected to each other so as to be able to communicate with each other via an in-vehicle network that complies with a standard such as a Controller Area Network (CAN).

[0015] The vehicle-mounted camera 110 has a two-dimensional detector configured with an array of photoelectric conversion elements, such as a CCD or C-MOS, that are sensitive to visible light, and an imaging optical system that forms an image of the area to be photographed on the two-dimensional detector. The vehicle-mounted camera 110 photographs the surroundings of the vehicle (for example, the area in front of the vehicle) and generates an image representing the environment around the vehicle. The vehicle-mounted camera 110 may include a front camera, two side cameras (left and right), and a rear camera. The vehicle-mounted camera 110 photographs at a predetermined photographing interval (for example, 1 / 30 to 1 / 10 seconds). The vehicle-mounted camera 110 may be configured as a stereo camera and may be configured to obtain the distance to each structure on the image from the parallax between the left and right images. Each time the vehicle-mounted camera 110 generates an image, it outputs the generated image to the ECU 150 via the in-vehicle network.

[0016] The periphery monitoring sensor 120 is a sensor for monitoring the periphery of the vehicle. The periphery monitoring sensor 120 includes, for example, sensors such as Lidar (Light Detection and Ranging) and Radar. The radar includes a front side radar sensor located inside the front bumper and a rear side radar sensor located inside the rear bumper. The vehicle speed sensor 130 is a sensor for detecting the vehicle speed V of the vehicle.

[0017] The ECU 150 is one aspect of the collision determination device according to the present disclosure. The ECU 150 includes a processor 152, a memory 154, and a communication interface 156. The processor 152 includes one or more central processing units (CPUs) and their peripheral circuits. The processor 152 may further include other arithmetic circuits such as a logic unit, a numerical calculation unit, or a graphics processing unit. The memory 154 includes, for example, a volatile semiconductor memory and a non-volatile semiconductor memory, and stores data related to the processing according to this embodiment. The communication interface 156 includes an interface circuit for connecting the ECU 150 to an in-vehicle network.

[0018] The G sensor 160 outputs a detection signal that indicates the deceleration (front Gf) of the vehicle that occurs when the vehicle collides with an object in front of the vehicle.

[0019] The warning device 170 includes a display device and a speaker. The display device is configured, for example, by a liquid crystal display (LCD) and is provided near the meter panel or the dashboard, and displays and outputs a warning in response to an instruction from the ECU 150. The speaker outputs the warning as sound in response to an instruction from the ECU 150.

[0020] After predicting a collision with a vulnerable road user such as a pedestrian or a bicycle, the driving assistance system 1000 determines that a collision with a vulnerable road user has occurred when a certain level of impact (front Gf) is applied to the front of the vehicle. The driving assistance system 1000 then reports the collision with the vulnerable road user to an emergency notification service center. Examples of emergency notification services include HELPNET, an Automatic Collision Notification (ACN) system, and an Advanced Automatic Collision Notification (AACN) system. In this case, the driving assistance system 1000 changes the determination threshold for the front Gf based on the accuracy of the prediction of a collision with a vulnerable road user. This makes it possible to increase the number of reports of collisions with vulnerable road users without increasing unnecessary reports to the emergency notification service center.

[0021] FIG. 2 is a schematic diagram showing functional blocks of the processor 152 of the ECU 150 for implementing the above-described processing. The processor 152 includes a collision prediction accuracy calculation unit 152a, a collision determination threshold acquisition unit 152d, a collision determination unit 152e, and a PCS control unit 152f. The collision prediction accuracy calculation unit 152a includes an object accuracy calculation unit 152b and a collision accuracy calculation unit 152c. Each of these units included in the processor 152 is a functional module implemented by, for example, a computer program running on the processor 152. In other words, the functional blocks of the processor 152 are configured by the processor 152 and a program (software) for operating the processor 152. The program may be stored in the memory 154 included in the ECU 150 or in an externally connected recording medium. Alternatively, each of these units included in the processor 152 may be a dedicated arithmetic circuit provided in the processor 152.

[0022] The collision prediction accuracy calculation unit 152a calculates the accuracy of a collision with a vulnerable road user (collision prediction accuracy). Specifically, the object accuracy calculation unit 152b calculates a first accuracy that an object ahead of the vehicle is a vulnerable road user. Furthermore, the collision accuracy calculation unit 152c calculates a second accuracy that the vehicle will collide with the object. Then, the collision prediction accuracy calculation unit 152a calculates the collision prediction accuracy based on the first accuracy and the second accuracy.

[0023] The object accuracy calculation unit 152b detects the presence of an object ahead of the vehicle based on an image showing the area ahead of the vehicle generated by the on-board camera 110. In this case, the object is detected from the image by, for example, template matching between a template image and the image generated by the on-board camera 110, or by inputting the image generated by the on-board camera 110 into a classifier that has been machine-learned for object detection.

[0024] The object accuracy calculation unit 152b may use, as the classifier, a segmentation classifier that is trained in advance to output, for each pixel of an input image, the accuracy that the pixel represents an object for each type of object that may be represented at that pixel, and to identify the object with the highest accuracy as the object represented.The object accuracy calculation unit 152b may use, as such a classifier, a deep neural network (DNN) having a convolutional neural network (CNN) architecture for segmentation, such as a fully convolutional network (FCN).

[0025] Then, the object accuracy calculation unit 152b calculates a first accuracy that the object is a vulnerable road user based on the object detected from the image. The first accuracy becomes higher the more the object detected from the image resembles a pedestrian or a bicycle. The object accuracy calculation unit 152b may calculate the first accuracy taking into consideration the size of the object, the duration of detection of the object, the presence and number of other objects near the object, etc.

[0026] The collision probability calculation unit 152c calculates a collision prediction time TTC. The collision prediction time TTC is a predicted time from the current time until the vehicle collides with an object, and is calculated based on the distance d between the object and the vehicle at the current time and the relative speed Vr of the object to be collided with the vehicle. Specifically, TTC = d / Vr. The collision prediction time TTC is used as an index representing the likelihood that the host vehicle will collide with the object, and the smaller its value, the higher the likelihood that the vehicle will collide with the object. The collision prediction time TTC changes from moment to moment. The collision probability calculation unit 152c then calculates a second probability based on the position of the object obtained from the image generated by the in-vehicle camera 110, the collision prediction time TTC, whether or not the driver of the vehicle has performed an evasive operation against the object, and the like. The second probability has a higher value as the object's position is closer to the vehicle, the collision prediction time TTC is shorter, and the driver has not performed an evasive operation against the object.

[0027] The collision determination threshold acquisition unit 152d acquires a collision determination threshold Gfref for the deceleration of the vehicle based on the collision prediction accuracy calculated by the collision prediction accuracy calculation unit 152a. The collision determination threshold acquisition unit 152d may acquire a map that defines the collision determination threshold Gfref according to the vehicle speed based on the collision prediction accuracy. The collision determination threshold acquisition unit 152d may acquire a map in which the collision determination threshold Gfref in the low vehicle speed range is smaller as the collision prediction accuracy is higher. Furthermore, the collision determination threshold acquisition unit 152d may acquire a map in which the collision determination threshold Gfref in the medium vehicle speed range or the high vehicle speed range is larger as the collision prediction accuracy is lower. Note that these maps may be stored in the memory 154 of the ECU 150.

[0028] The collision determination unit 152e determines whether the vehicle has collided with a vulnerable road user based on the detected value of the vehicle deceleration and the collision determination threshold Gfref. The collision determination unit 152e may apply the detected values ​​of the vehicle speed and the vehicle deceleration to the map acquired by the collision determination threshold acquisition unit 152d, and determine that the vehicle has collided with a vulnerable road user if the vehicle deceleration is greater than the collision determination threshold Gfref.

[0029] The PCS control unit 152f performs collision avoidance support control (pre-crash safety control, hereinafter referred to as PCS control). The PCS control includes output of an alarm sound or alarm display from the warning device 170, and control of pre-crash brake assist or pre-crash brake.

[0030] Fig. 3 is a flowchart showing the processing performed by the processor 152 of the ECU 150 at each predetermined control cycle. In Fig. 3, the processing other than steps S11' and S15' may be performed in the same manner as the processing described in the aforementioned Patent Document 1. As in the aforementioned Patent Document 1, in the processing of Fig. 3, if the collision prediction timing determined from the collision prediction time TTC of the object coincides with the timing at which the front Gf exceeds the collision determination threshold Gfref (actual detection timing), it is estimated that the vehicle has collided with the object.

[0031] First, the object accuracy calculation unit 152b determines whether an object in front of the vehicle has been detected from the image generated by the in-vehicle camera 110 (step S11). If an object is detected, the collision prediction accuracy calculation unit 152a calculates the accuracy of a collision with a vulnerable road user, i.e., the collision prediction accuracy (step S11'). On the other hand, if an object is not detected in step S11, the process returns to step S11.

[0032] After step S11', the collision prediction time TTC is calculated (step S12). The collision prediction time TTC is constantly calculated by the collision probability calculation unit 152c. Next, it is determined whether the collision prediction time TTC is equal to or less than a threshold value TTCref (step S13). If the collision prediction time TTC is equal to or less than the threshold value TTCref, the timer starts measuring the timer value t (step S14). On the other hand, if the collision prediction time TTC exceeds the threshold value TTCref, the process returns to step S11.

[0033] After step S14, the vehicle speed V detected by the vehicle speed sensor 130 and the front Gf detected by the G sensor 160 are acquired (step S15). Next, the collision determination threshold acquisition unit 152d acquires a map corresponding to the probability of a collision with a vulnerable road user, i.e., the collision prediction probability, from among a plurality of collision determination threshold maps that define the collision determination threshold Gfref according to the vehicle speed V (step S15').

[0034] Next, with reference to the map acquired in step S15', a collision determination threshold Gfref is set according to the vehicle speed V (step S16). Next, it is determined whether the front Gf detected by the G sensor 160 is greater than the collision determination threshold Gfref (step S17). If the front Gf is greater than the collision determination threshold Gfref, it is determined whether the timer value t is within an object collision prediction allowable period (step S19). The object collision prediction allowable period is a period between TTC-α and TTC+α, where α is an error allowable value. On the other hand, if the front Gf is equal to or less than the collision determination threshold Gfref, it is determined whether the timer value t is greater than the determination end time (TTC+α) (step S18). If the timer value t is equal to or less than the determination end time (TTC+α), the process returns to step S15.

[0035] In step S19, if the timer value t is within the object collision prediction allowable period, i.e., if the collision prediction timing determined from the collision prediction time TTC and the actual detection timing at which the front Gf exceeds the collision determination threshold Gfref are the same, it is guaranteed that the front Gf is caused by a collision with an object. Therefore, the collision determination unit 152e determines that the vehicle has collided with a vulnerable road user (step S20). Next, the collision occurrence information is notified to the emergency call service center (step S21). After step S21, the process ends.

[0036] On the other hand, if the timer value t is not within the object collision prediction allowable period in step S19, the collision determination unit 152e determines that the collision is not between the vehicle and a vulnerable road user (step S22).

[0037] In addition, in step S18, if the timer value t exceeds the determination end time (TTC+α) without the front Gf becoming larger than the collision determination threshold Gfref, the collision determination unit 152e determines that the vehicle has not collided with a vulnerable road user (step S22). After step S22, the process ends.

[0038] 4 is a timing chart illustrating an example in which the collision prediction accuracy calculation unit 152a calculates the level of collision prediction accuracy using the control state of PCS control and the collision prediction time TTC. FIG. 4 shows how, at a timing before the collision prediction time (TTC=0), the alarm sound (alarm (ALM)), pre-crash brake assist (PBA), and pre-crash brake (PB) are activated in this order as the collision prediction time TTC approaches 0. In this example, the object accuracy calculation unit 152b determines that the first accuracy that the object is a vulnerable road user is equal to or greater than a certain level, and the object is highly likely to be a vulnerable road user. Therefore, the level of the collision prediction accuracy is determined mainly by the second accuracy of a collision with the object calculated by the collision accuracy calculation unit 152c.

[0039] In the example shown in Figure 4, the higher the probability that the PCS will activate, the higher the collision prediction probability. The probability that the PCS will activate is highest when the pre-crash brake (PB) is turned on, followed by when the pre-crash brake assist (PBA) is turned on, and lowest when the alarm (ALM) is turned on. In addition, the probability that the PCS will activate is higher as the collision prediction time TTC is shorter.

[0040] 4, the pre-crash brake (PB) is turned on and activated when the collision prediction time TTC reaches the collision determination threshold TTCref_Lv1. When the pre-crash brake (PB) is turned on, the collision prediction accuracy calculation unit 152a calculates Lv1 as the collision prediction accuracy.

[0041] The pre-crash brake assist (PBA) assists the braking force when the driver depresses the brake pedal. Even if the pre-crash brake assist (PBA) is turned on, no assistance is provided unless the driver depresses the brake pedal. When the pre-crash brake assist (PBA) is turned on and the collision prediction time TTC reaches the collision determination threshold TTCref_Lv2, the collision prediction accuracy calculation unit 152a calculates Lv2 as the collision prediction accuracy.

[0042] Furthermore, when the alarm sound (ALM) is turned on and the collision prediction time TTC reaches the collision determination threshold value TTCref_Lv3, the collision prediction accuracy calculation unit 152a calculates Lv3 as the collision prediction accuracy.

[0043] As described above, the collision prediction accuracy calculation unit 152a can calculate the collision prediction accuracy levels Lv1, Lv2, and Lv3 based on the control state of the PCS and the collision prediction time TTC. Note that there is a relationship of Lv1>Lv2>Lv3.

[0044] 5 to 7 are diagrams showing examples of maps acquired by the collision determination threshold acquisition unit 152d based on collision prediction accuracy levels Lv1, Lv2, and Lv3. All of the maps shown in Fig. 5 to 7 define the collision determination threshold value Gfref for the front Gf according to the vehicle speed V. Of these, the map shown in Fig. 6 is a map acquired when the collision prediction accuracy is Lv2, and the collision determination threshold value Gfref does not change according to the vehicle speed V but is a constant value.

[0045] In the map shown in Figure 6, in order to avoid the front Gf detected by the G sensor 160 being mistakenly determined as a collision with a vulnerable road user when driving on a rough road or in the event of a minor collision, the collision determination threshold value Gfref is set to a value large enough to prevent such a misjudgment from occurring.

[0046] On the other hand, the map shown in Fig. 5 is a map obtained when the collision prediction accuracy is Level 1. In the map shown in Fig. 5, the collision determination threshold Gfref in the map shown in Fig. 6 is lowered so that collisions with vulnerable road users can be detected more frequently in areas with low vehicle speeds and low front Gf when the frequency of collisions with vulnerable road users is high. As a result, in the map shown in Fig. 5, the area in which a collision with a vulnerable road user is determined to be a collision is expanded in the hatched area R1 compared to the map in Fig. 6.

[0047] Because Figure 5 is a map acquired when the collision prediction accuracy is lower than that of Figure 6, if the collision detection threshold Gfref is lowered as in Figure 6, the front Gf detected when driving on rough roads or in a minor collision is more likely to be mistaken for a collision with a vulnerable road user, increasing the probability of detecting unnecessary information other than a collision with a vulnerable road user. Because region R1 is an area where collisions with vulnerable road users occur particularly frequently, the map of Figure 6 makes it possible to determine more collisions with vulnerable road users that could not be detected by the map of Figure 5 when the collision prediction accuracy is high.

[0048] The map shown in Fig. 7 is obtained when the collision prediction accuracy is Level 3. In the map shown in Fig. 7, the collision determination threshold Gfref is increased in the medium or high vehicle speed range and in the high front Gf range. As a result, the area in which a collision with a vulnerable road user is determined to have occurred is reduced by the hatched area R2.

[0049] If the map shown in Fig. 5 is used when the collision prediction accuracy is low, even if the front Gf exceeds the collision judgment threshold Gfref in the medium or high vehicle speed range, the probability that the object is a vulnerable road user is low, and there are many cases where the front Gf is erroneously judged as a collision with a vulnerable road user due to driving on a rough road, etc. According to the map in Fig. 7, by increasing the collision judgment threshold Gfref in the medium or high vehicle speed range when the collision prediction accuracy is low, it is possible to prevent the front Gf detected when driving on a rough road or in a minor collision from being erroneously judged as a collision with a vulnerable road user.

[0050] In addition, when the collision prediction accuracy is low, especially in the low-vehicle speed area R3, the risk of collision with vulnerable road users is low. For this reason, the map shown in Fig. 7 does not determine whether a collision with a vulnerable road user will occur even in the low vehicle speed region R3.

[0051] In addition, when the collision prediction accuracy is low and the front Gf exceeds the collision determination threshold Gfref shown in Fig. 7 in the medium or high vehicle speed range, it is estimated that the impact of a collision with a vulnerable road user will be large. According to the map in Fig. 7, it is possible to detect such a large-impact collision with a vulnerable road user while preventing the front Gf detected when driving on rough roads, etc., from being erroneously determined as a collision with a vulnerable road user.

[0052] 5 to 7, the maps reduce the likelihood of a collision with a vulnerable road user being mistakenly determined as such due to the front Gf when driving on rough roads, thereby reducing the frequency of unnecessary notifications to emergency call centers. This also makes it possible to improve the detection rate of collisions with vulnerable road users in areas with low vehicle speeds and low front Gf, and in areas with medium and high vehicle speeds and high front Gf, where the frequency of collisions with vulnerable road users is high, thereby improving the accuracy of notifications to emergency call centers.

[0053] As described above, according to this embodiment, by changing the collision judgment threshold Gfref based on the collision prediction probability, which is the probability of colliding with a vulnerable road user, erroneous judgments of collisions with vulnerable road users are suppressed. [Explanation of symbols]

[0054] 110 In-car camera 120 Perimeter monitoring sensor 130 Vehicle speed sensor 150 Electronic Control Unit (ECU) 152 processors 152a Collision prediction accuracy calculation unit 152b Object accuracy calculation unit 152c Collision probability calculation unit 152d Collision detection threshold acquisition unit 152e Collision detection section 152f PCS control unit 154 memory 156 Communication Interface 160 G sensor 170 Warning device 1000 Driving Assistance System

Claims

1. an object probability calculation unit that calculates a first probability that an object in front of the vehicle is a pedestrian or a bicycle; a collision probability calculation unit that calculates a second probability that the vehicle will collide with the object; a collision prediction accuracy calculation unit that calculates a collision prediction accuracy, which is the accuracy of a collision with a pedestrian or a bicycle, based on the first accuracy and the second accuracy; a collision determination threshold acquisition unit that acquires a collision determination threshold of a vehicle deceleration based on the collision prediction accuracy; a collision determination unit that determines whether the vehicle has collided with a pedestrian or a bicycle based on the detected value of the deceleration of the vehicle and the collision determination threshold; A vehicle collision determination device comprising:

2. the collision determination threshold acquisition unit acquires a map that defines the collision determination threshold according to a vehicle speed based on the collision prediction accuracy; 2. The vehicle collision determination device according to claim 1, wherein the collision determination unit applies the detected values ​​of the vehicle speed and the vehicle deceleration to the map, and determines that the vehicle has collided with a pedestrian or a bicycle if the vehicle deceleration is greater than the collision determination threshold.

3. The vehicle collision determination device according to claim 2 , wherein the collision determination threshold acquisition unit acquires a map in which the collision determination threshold in a low vehicle speed range is smaller as the collision prediction probability is higher.

4. The vehicle collision determination device according to claim 2 or 3, wherein the collision determination threshold acquisition unit acquires a map in which the collision determination threshold in a medium vehicle speed range or a high vehicle speed range is larger as the collision prediction accuracy is lower.

Images