Vehicle control device

Abstract

To execute deceleration control of a vehicle even when it is difficult for an image recognition device to recognize a pedestrian.SOLUTION: A control unit 11 configured to execute avoidance assistance control for decelerating or avoiding the vehicle when an object is recognized in front of the vehicle based on a first detection value and a second detection value for detecting an environment around the vehicle 1, the control unit performing a collation process of collating a first recognition result based on the first detection value for the object and a second recognition result based on the second detection value for the object, determining that the collation result of the object matches in the collation process, and determining that the object is a pedestrian requiring the avoidance assistance control, in addition, the vehicle control device 10 executes the avoidance assistance control when the number of times of matching-matching in the matching processing is equal to or larger than the threshold value, and decreases the threshold value for the number of times of matching-matching when the condition that the accuracy of recognition of the object is lower than that in the normal state is satisfied in the environment around the vehicle.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a vehicle control device for protecting pedestrians. [Background technology]

[0002] In recent years, a vehicle control device for protecting pedestrians has become known (see, for example, Patent Document 1). The vehicle control device described in Patent Document 1 is configured to detect a person present in front of the host vehicle, and decelerate the host vehicle at a second deceleration that is smaller than a preset first deceleration based on the speed of the host vehicle and the stopping distance from the host vehicle to a target position in front of the person, and stop the host vehicle at the target position.

[0003] This vehicle control device detects an object using a first detection value of the object detected by a radar device and a second detection value of the object recognized by an image recognition device. When an object is detected by the first detection value and the second detection value at the same time and in the same position, this vehicle control device determines that the same object has been detected, and by integrating these two detection values, the vehicle control device improves the object detection accuracy compared to a method of detecting an object using a single detection value. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-019301 Summary of the Invention [Problem to be solved by the invention]

[0005] According to the vehicle control device described in Patent Document 1, when a pedestrian is between the host vehicle and the oncoming vehicle in a backlit condition with the lights of the oncoming vehicle on and the host vehicle's lights on, it may be difficult for the image recognition device to recognize the pedestrian. Therefore, the vehicle control device described in Patent Document 1 may cause a delay in deceleration control of the host vehicle.

[0006] An object of the present invention is to provide a vehicle control device that can execute deceleration control of the vehicle even in a state where it is difficult for an image recognition device to recognize a pedestrian. [Means for solving the problem]

[0007] One aspect of the present invention is a vehicle control device that includes a control unit that executes avoidance assistance control to slow down or avoid the vehicle when an object is recognized in front of the vehicle based on first detection values ​​and second detection values ​​that detect the environment around the vehicle, wherein the control unit performs a matching process to match a first recognition result based on the first detection value for the object with a second recognition result based on the second detection value for the object, and executes the avoidance assistance control when the matching results for the object match in the matching process and the object is determined to be a pedestrian requiring the avoidance assistance control, and when the number of matching matches in the matching process is equal to or greater than a threshold, and when a condition is met in the environment around the vehicle that causes the accuracy of object recognition to decrease compared to normal conditions, the control unit reduces the threshold for the number of matching matches. [Effects of the Invention]

[0008] According to the present invention, it is possible to execute deceleration control of a vehicle even in a state where it is difficult for an image recognition device to recognize a pedestrian. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a block diagram showing a configuration of a vehicle control device according to an embodiment; [Figure 2]FIG. 10 is a diagram showing a state in which a pedestrian is present in front of the vehicle. [Figure 3] 10 is a flowchart showing the flow of a process for avoidance assistance control in a normal state. [Figure 4] FIG. 1 is a diagram illustrating an example of an environment in which pedestrian visibility is reduced. [Figure 5] 10 is a flowchart showing the flow of avoidance assist control processing executed when a condition is met in which the accuracy of object recognition decreases compared to a normal state in the environment around the vehicle. [Figure 6] 10 is a flowchart showing the flow of avoidance assist control processing executed when a condition is met in which the accuracy of object recognition decreases compared to a normal state in the environment around the vehicle. [Figure 7] 10 is a flowchart showing the flow of avoidance assist control processing executed when a condition is met in which the accuracy of object recognition decreases compared to a normal state in the environment around the vehicle. [Figure 8] 10 is a flowchart showing the flow of avoidance assist control processing executed when a condition is met in which the accuracy of object recognition decreases compared to a normal state in the environment around the vehicle. DETAILED DESCRIPTION OF THE INVENTION

[0010] 1, vehicle 1 includes a detection unit 2 that detects the surrounding environment. Vehicle 1 also includes a vehicle control device 10 that executes driving assistance control based on a detection value detected by detection unit 2. Vehicle control device 10 is configured to execute deceleration control to decelerate vehicle 1 when a pedestrian is detected based on the detection value.

[0011] The detection unit 2 is equipped with an exterior camera 2A that captures images of the environment around the vehicle 1. The exterior camera 2A is installed, for example, to capture an image range in front of the vehicle 1. The detection unit 2 may be equipped with one or more exterior cameras 2A that capture an image of a predetermined image range around the vehicle 1. The exterior camera 2A captures images of objects such as pedestrians that are present in front of the vehicle 1 within the image range. The exterior camera 2A may be used not only for deceleration control but also for other driving assistance controls. The exterior camera 2A may also be used as a drive recorder. The exterior camera 2A generates image data capturing an image of the environment around the vehicle 1 and outputs the image data to the vehicle control device 10.

[0012] The detection unit 2 includes a LIDAR device 2B that detects objects around the vehicle 1. The LIDAR device 2B acquires three-dimensional data of objects around the vehicle 1, for example, by irradiating a laser beam within a scanning range and receiving light reflected from the object. The LIDAR device 2B acquires three-dimensional data of the environment around the vehicle 1 within the scanning range of the laser beam.

[0013] The detection unit 2 includes a radar device 2C that detects objects around the vehicle 1. The radar device 2C detects objects around the vehicle 1 by irradiating, for example, millimeter-wave radar waves within a scanning range and measuring the reflected waves. The radar device 2C is configured to be able to measure the relative distance, relative speed, and relative angle to objects such as pedestrians and vehicles.

[0014] The detection unit 2 is provided with a position sensor 2D that measures the current position of the vehicle 1. The position sensor 2D is configured by, for example, a GPS (Global Positioning System) sensor or the like. The position sensor 2D may be used in, for example, a navigation device. The position sensor 2D outputs a measurement value to the vehicle control device 10.

[0015] The vehicle 1 includes an input / output unit 3 that outputs a display image and accepts input operations from the occupant. The input / output unit 3 is configured, for example, by a display device such as a liquid crystal display or an organic EL (Electro-Luminescence) display. The input / output unit 3 is configured by a touch panel. The input / output unit 3 accepts input operations by touching a displayed operation image. For example, the input / output unit 3 displays a display image showing notification content when driving assistance is executed. The input / output unit 3 may also be configured to display the display content of a navigation device provided in the vehicle 1.

[0016] The vehicle 1 is equipped with a drive unit 5 that serves as a drive source for traveling. The drive unit 5 may be configured with an internal combustion engine or an electric motor. The drive unit 5 may be configured with a hybrid device that combines an internal combustion engine and an electric motor. If the vehicle 1 is a manually driven vehicle, the drive unit 5 is controlled based on operation by the driver, and under predetermined conditions, the vehicle control device 10 executes driving assistance control that assists the driver's operation. If the vehicle 1 is an autonomous vehicle, the drive unit 5 is controlled by the vehicle control device 10.

[0017] The vehicle 1 is equipped with a braking unit 6 for decelerating the vehicle 1. The braking unit 6 is configured, for example, by a brake device. When the drive unit 5 is configured by an electric motor, the braking unit 6 may be configured by the drive unit 5. In this case, the drive unit 5 may be configured to decelerate the vehicle 1 by regenerating power based on the deceleration energy of the vehicle 1.

[0018] The vehicle 1 is equipped with a steering unit 7 that accepts steering operations by the driver. The steering unit 7 is configured to apply a steering angle to steering wheels provided on the vehicle 1 in accordance with the amount of steering operation by the driver. When the drive unit 5 is configured with an electric motor, the steering unit 7 may be configured by the drive unit 5. The steering unit 7 may be realized by controlling the drive wheels driven by the drive unit 5.

[0019] The vehicle control device 10 includes a control unit 11 that executes control related to the traveling of the vehicle 1, and a storage unit 12 that stores data and programs required for the control. The control unit 11 is configured with at least one hardware processor such as a CPU (Central Processing Unit). The storage unit 12 is configured with a non-transitory storage medium such as a hard disk drive (HDD) or a solid state disk (SSD). The storage unit 12 may also store map data used in the navigation device.

[0020] The storage unit 12 stores data of the detection values ​​output from the detection unit 2. The detection value data may be stored for a predetermined period and then updated with new detection value data. When the control unit 11 recognizes an object such as a pedestrian ahead of the vehicle based on the detection values ​​that detect the environment around the vehicle 1, the control unit 11 executes avoidance assistance control to decelerate the vehicle 1 or to avoid the object.

[0021] The control unit 11 executes avoidance assistance control based on detection values ​​by the detection unit 2, including, for example, image data captured by the outside camera 2A, measurement value data by the LIDAR device 2B, and measurement value data by the radar device 2C.

[0022] The control unit 11 recognizes the environment around the vehicle 1 based on, for example, a determination process using AI (Artificial Intelligence). The control unit 11 is configured to perform machine learning such as deep learning in advance using imaging data of the environment around the vehicle 1 as training data, and to extract an object present in front of the vehicle 1 based on a first detection value including imaging data captured by the exterior camera 2A. The control unit 11 performs a first determination process to determine whether or not an object exists based on the first detection value including imaging data.

[0023] The control unit 11 compares multiple pieces of image data of the same environment included in the imaging range in chronological order, and extracts objects included in the image data. The objects may include moving objects, stationary objects, moving objects, and non-moving objects. The control unit 11 recognizes the shape, movement, color, etc. of the extracted objects, and recognizes pedestrians present in front of the vehicle 1 from among the objects. The control unit 11 stores in the memory unit 12 the first recognition result of the pedestrian extracted by the first determination process based on the first detection value.

[0024] The control unit 11 may be configured to perform not only the determination process using AI but also Simultaneous Localization and Mapping (SLAM) to generate an environmental map of the current position of the vehicle 1 based on the image data and map data stored in the storage unit 12. The environmental map is a reproduction of the surrounding environment of the current position of the vehicle 1 using three-dimensional data.

[0025] The control unit 11 executes SLAM and recognizes the environment around the vehicle 1 based on the generated environmental map. The control unit 11 may recognize the environment around the vehicle 1, including objects present on the road, by combining the imaging data with the detection values ​​of the LIDAR device 2B and the radar device 2C. The control unit 11 may also be configured to perform analysis based on other methods as long as it can recognize the environment around the vehicle 1.

[0026] The control unit 11 is configured to be able to extract an object present ahead of the vehicle 1 based on, for example, a second detection value different from the first detection value, including LIDAR data from the LIDAR device 2B acquired by scanning the environment around the vehicle 1. The control unit 11 may also extract an object present ahead of the vehicle 1 based on the second detection value including radar data from the radar device 2C.

[0027] The control unit 11 executes a second determination process to determine whether or not an object exists ahead of the vehicle 1 based on the second detection values. In the second determination process, the control unit 11 extracts the object and determines whether or not the object is a pedestrian. The control unit 11 calculates, for example, the width, height, moving speed, shape, etc. of the extracted object. If the control unit 11 determines that the object is a pedestrian through the second determination process based on the second detection values, it extracts the pedestrian from the second detection values. The control unit 11 stores in the memory unit 12 a second recognition result of the pedestrian extracted through the second determination process based on the second detection values.

[0028] 2 shows a state in which a pedestrian H is present in front of the vehicle 1. The control unit 11 executes a matching process to match a first recognition result of a first determination process based on a first detection value with a second recognition result of a second determination process based on a second detection value different from the first detection value. The matching process is a process in which the control unit 11 determines whether multiple detection means of different types have shown the same recognition result for the same object, and if it is determined that the same recognition result has been shown for the same object, performs a matching determination.

[0029] The control unit 11 executes the matching process by assuming that an environment in which the first detection value and the second detection value are sufficiently obtained is a normal state in which the accuracy of object recognition can be ensured. The control unit 11 determines whether the first detection value and the second detection value simultaneously ensure the accuracy of object recognition and whether the matching process is possible. In determining the matching process, the control unit 11 determines, for example, whether the first detection value and the second detection value satisfy a first condition. The first condition is set to include a preset threshold value for the first detection value in the normal state and a preset threshold value for the second detection value in the normal state.

[0030] The control unit 11 determines that the first condition is satisfied when it compares the first detection value with the threshold and determines that the first detection value is in a state where the first determination process can be executed, and when it compares the second detection value with the threshold and determines that the second detection value is in a state where the second determination process can be executed. The control unit 11 determines that the first condition is not satisfied when it determines that a condition exists in the environment around the vehicle where the accuracy of object recognition is reduced compared to a normal state. For example, the control unit 11 determines that the first condition is not satisfied when it compares the first detection value with the threshold and determines that the first detection value is in a state where the first determination process cannot be executed, or when it compares the second detection value with the threshold and determines that the second detection value is in a state where the second determination process cannot be executed.

[0031] When the first detection value and the second detection value simultaneously satisfy a first condition that can ensure the accuracy of object recognition, the control unit 11 determines whether the object is a pedestrian H based on the first detection value and the second detection value. The control unit 11 executes a matching process using a first recognition result of the pedestrian H recognized by the first detection value and a second recognition result of the pedestrian H recognized by the second detection value. The control unit 11 determines that the object is a pedestrian H when, for example, the first recognition result and the second recognition result are recognized at the same time and calculated at the same position, and the pedestrian recognition accuracy based on the first recognition result is equal to or greater than a threshold, and the pedestrian recognition accuracy based on the second recognition result is equal to or greater than a threshold. When the control unit 11 determines that the object is a predetermined pedestrian H that requires avoidance assist control based on the determination result of the matching process, the control unit 11 executes avoidance assist control.

[0032] In the matching process, the control unit 11 measures the relative distance L between the pedestrian H and the vehicle 1 based on the second detection value detected by the LIDAR device 2B or the radar device 2C. The control unit 11 calculates the time to collision (TTC) until contact between the vehicle 1 and the pedestrian H based on the distance L and the speed V of the vehicle 1. If the control unit 11 determines that the TTC is equal to or less than a threshold, it determines that avoidance assist control is necessary for the pedestrian H. If the control unit 11 recognizes a specific pedestrian H that requires avoidance assist control, it starts avoidance assist control and controls the drive unit 5 and the brake unit 6 to decelerate the vehicle 1 and prevent the vehicle 1 from contacting the pedestrian H.

[0033] When the distance L is equal to or less than the threshold value, the control unit 11 controls the drive unit 5 and the braking unit 6 to decelerate the vehicle 1, and controls the steering unit 7 to cause the vehicle 1 to avoid the pedestrian H, thereby performing avoidance assistance control to prevent the vehicle 1 from coming into contact with the pedestrian H.

[0034] 3 shows the flow of processing of the avoidance assistance control method executed in the vehicle control device 10 in a normal state. The avoidance assistance control method is executed by the control unit 11 based on a computer program installed in a computer mounted on the vehicle control device 10. The control unit 11 executes the following processes.

[0035] The control unit 11 sets the initial value of the counter value P used for avoidance assistance control to 0 (step S100). The control unit 11 acquires the first detection value and the second detection value (step S102). The control unit 11 determines whether the first detection value and the second detection value simultaneously satisfy a first condition indicating that the accuracy of object recognition can be ensured (step S104). If the control unit 11 determines that the first detection value does not satisfy the first condition, it determines whether a predetermined condition exists (step S105). The state in which the first detection value does not satisfy the first condition refers to a state in which the environment affects the detection value, such as when there is backlighting, such as when there is an oncoming vehicle with its headlights on.

[0036] The conditions in which the environment affects the detection value may include specific backlight conditions, such as poor visibility due to weather conditions, backlight conditions due to sunlight conditions such as sunset or sunrise, and backlight conditions at the exit of a tunnel. The control unit 11 determines that a predetermined condition exists when it recognizes poor visibility due to weather conditions based on the detection value. The control unit 11 determines that a predetermined condition exists when it determines that a backlight condition exists due to sunlight conditions based on the detection value. The control unit 11 determines that a predetermined condition exists when it determines that a backlight condition exists at the exit of a tunnel based on the detection value. The control unit 11 determines that a predetermined condition exists when it determines that a backlight condition exists in a backlight condition where an oncoming vehicle with its headlights on is present and the vehicle 1 also has its headlights on, as described below.

[0037] The state in which the second detection value does not satisfy the first condition refers to a state in which the environment affects the detection value, such as when the second detection value deteriorates due to rainfall or when the second detection value deteriorates due to snow accumulation on the exterior of the LIDAR device 2B and the radar device 2C. When the control unit 11 determines that the second detection value has deteriorated based on the detection value, it determines that a predetermined condition exists. When the control unit 11 determines that the predetermined condition exists, it executes processing described below. When the first detection value and the second detection value satisfy the first condition, the control unit 11 executes a matching process to match the first pedestrian recognition result based on the first detection value with the second pedestrian recognition result based on the second detection value (step S106).

[0038] The control unit 11 determines whether or not a predetermined pedestrian requiring avoidance assistance control is present based on the determination result of the matching process and the TTC (step S108). If the control unit 11 determines that a predetermined pedestrian is present, it increments the counter value P by 1 (step S110). The control unit 11 determines whether or not the counter value P has reached a preset threshold N (N is a natural number) of the number of matchings (step S112). If the counter value P has not reached the threshold N of the number of matchings, the control unit 11 returns the process to step S102. If the counter value P has reached the threshold N of the number of matchings, the control unit 11 executes avoidance assistance control (step S114). By repeatedly executing the above process, the control unit 11 can improve the accuracy of recognizing the pedestrian H.

[0039] FIG. 4 shows an example of a predetermined condition in which a pedestrian H becomes difficult to recognize due to environmental influences. An oncoming vehicle T is present in front of the vehicle 1, and the oncoming vehicle T has its headlights on. The vehicle 1 has its headlights on. The pedestrian H is present between the vehicle 1 and the oncoming vehicle. The pedestrian H is present within the illumination range of the headlights of the vehicle 1 and also within the illumination range of the headlights of the oncoming vehicle T.

[0040] As shown in the figure, when the headlights of the oncoming vehicle T are illuminated, the first detection value of the vehicle 1 indicates a backlit state. When the pedestrian H is within the illumination range G1 of the headlights of the oncoming vehicle T, it may be difficult for the vehicle 1 to recognize the pedestrian H. In this state, when the pedestrian H is within the illumination range G2 of the headlights of the vehicle 1, it may be even more difficult for the vehicle 1 to recognize the pedestrian H.

[0041] The control unit 11 recognizes the pedestrian H based on a first detection value using image data captured by the exterior camera 2A. When there is an influence of a predetermined environmental condition such as backlight, it becomes difficult to recognize the pedestrian H in the image data, and the accuracy of the first recognition result by the control unit 11 decreases compared to the normal state.

[0042] The control unit 11 executes the following process when a condition is met in the environment around the vehicle 1 that causes the accuracy of object recognition to decrease compared to normal conditions. For example, when the accuracy of either the first recognition result or the second recognition result is decreased compared to normal conditions and matching processing is not possible, the control unit 11 determines whether a pedestrian requiring avoidance assistance control is present based on either the first recognition result or the second recognition result. When the accuracy of the first recognition result does not satisfy the first condition, the control unit 11 determines whether a pedestrian requiring avoidance assistance control is present using the second recognition result. Below, a method for determining whether a pedestrian requiring avoidance assistance control is present when a condition is met in the environment around the vehicle that causes the accuracy of object recognition to decrease compared to normal conditions will be described.

[0043] 5 shows an example of a process flow for determining whether a pedestrian requiring avoidance assistance control is present when a condition is met in which the accuracy of object recognition is reduced compared to normal conditions in the environment around the vehicle. In the illustrated example, when, for example, a matching process based on the first recognition result from the exterior camera 2A cannot be performed and an object can be recognized based on the second detection value detected by the LIDAR device 2B or the radar device 2C, the control unit 11 performs a second determination process using the second detection value. When the accuracy of object recognition by the first detector is reduced compared to normal conditions, the control unit 11 determines whether the object is a pedestrian requiring avoidance assistance control based on the second recognition result from the second detector without performing a matching process.

[0044] In step S104 (see FIG. 3), if the control unit 11 determines that the first determination process cannot be performed sufficiently based on the first detection value and that the accuracy of the first recognition result is lower than in a normal state (step S104: No), the control unit 11 determines that the first condition is not satisfied. If the control unit 11 determines that the first detection value does not satisfy the first condition, the control unit 11 determines whether a predetermined condition exists (step S105). If the control unit 11 determines that the predetermined condition does not exist, the control unit 11 returns the process to step S100 (see FIG. 3).

[0045] When it is determined that the predetermined condition exists, the control unit 11 determines whether the distance L1 between the oncoming vehicle T and the object is within the range of a specified distance S1 based on the second detection value (step S202). When the distance L1 between the oncoming vehicle T and the object (pedestrian H) is within the range of the specified distance S1 (step S202: Yes), the control unit 11 executes a second determination process using the second detection value and recognizes the object (step S204).

[0046] The control unit 11 determines whether the object is a pedestrian or not based on the determination result of the second determination process using the second detection value (step S206). If the object is a pedestrian H, the control unit 11 determines that the distance between the pedestrian H and the oncoming vehicle T is within a specified distance S1, the TTC is within a threshold, and avoidance assist control is necessary. If the control unit 11 determines that a predetermined pedestrian is present based on the determination result of the second determination process using the second detection value, it adds 1 to the counter value P (step S208). The control unit 11 determines whether the counter value P has reached a preset threshold N for the number of matching attempts (step S210).

[0047] If the counter value P has not reached the threshold N of the number of matching attempts, the control unit 11 returns the process to step S100 (see FIG. 3). If the counter value P has reached the threshold N of the number of matching attempts, the control unit 11 proceeds to step S114 and executes avoidance assistance control (see FIG. 3). According to the above process, even if a predetermined condition such as an environmental influence, such as a backlight condition, exists and the matching process cannot be performed based on the image capture data of the outside camera 2A, the control unit 11 can determine the presence of a predetermined pedestrian H by executing the second determination process based on the second detection value using the LIDAR device 2B or the radar device 2C.

[0048] 6 shows another example of a processing flow for determining whether or not a pedestrian requiring avoidance assist control is present when a condition is met in which the accuracy of object recognition is reduced compared to a normal state in the environment around the vehicle. In step S104 (see FIG. 3), if the control unit 11 determines that the first condition is not satisfied by the first determination processing based on the first detection value (step S104: No), the control unit 11 determines whether or not a predetermined condition exists (step S105). If the control unit 11 determines that the predetermined condition does not exist, the control unit 11 returns the processing to step S100 (see FIG. 3).

[0049] When the control unit 11 determines that the predetermined condition exists, it determines whether the distance L1 between the oncoming vehicle T and the pedestrian H is within the range of the specified distance S1 based on the second detection value (step S302). When the distance L1 between the oncoming vehicle T and the pedestrian H is within the range of the specified distance S1 based on the second recognition result of the second determination process, the control unit 11 executes a matching process to match the first recognition result based on the first detection value of the object with the second recognition result based on the second detection value of the object (step S304).

[0050] The control unit 11 determines whether or not a predetermined pedestrian requiring avoidance assist control is present based on the determination result of the matching process (step S306). If the control unit 11 determines that a predetermined pedestrian is present, it increments the counter value P by 1 (step S308). The control unit 11 decreases the threshold value for the number of matching attempts. For example, the control unit 11 determines whether or not the counter value P has reached a threshold value (NM) for the number of matching attempts obtained by subtracting a predetermined value M (M is a natural number smaller than N) from a preset threshold value N for the number of matching attempts (step S310).

[0051] If the counter value P has not reached the threshold value (NM) of the number of matching attempts, the control unit 11 returns the process to step S100 (see FIG. 3). If the counter value P has reached the threshold value (NM) of the number of matching attempts, the control unit 11 proceeds to step S114 and executes avoidance assistance control (see FIG. 3). According to the above process, even if the determination process based on the detection value does not satisfy the first condition due to the influence of the environment, such as backlight conditions, the control unit 11 can determine the presence of the predetermined pedestrian H and execute avoidance assistance control by reducing the threshold value for the number of matching attempts used in the matching process.

[0052] 7 shows another example of a processing flow for determining whether or not a pedestrian requiring avoidance assist control is present when a condition is met in which the accuracy of object recognition is reduced compared to a normal state in the environment around the vehicle. In step S104 (see FIG. 3), if the control unit 11 determines that the first determination process based on the first detection value does not satisfy the first condition (step S104: No), the control unit 11 determines whether or not a predetermined condition exists (step S105). If the control unit 11 determines that the predetermined condition does not exist, the control unit 11 returns the processing to step S100 (see FIG. 3).

[0053] When it is determined that the predetermined condition exists, the control unit 11 determines whether or not the distance L1 between the oncoming vehicle T and the pedestrian H is within the range of the specified distance S1 based on the second detection value (step S402). When the distance L1 between the oncoming vehicle T and the pedestrian H is within the range of the specified distance S1, the control unit 11 executes a matching process (step S404).

[0054] In step S404, the control unit 11 reduces the determination threshold for the pedestrian likelihood, which is an index for estimating a pedestrian, and executes the matching process. The pedestrian likelihood is the probability density that the object is estimated to be a pedestrian when the object is recognized using the detection value. When a condition is met in which the accuracy of object recognition based on the first detection value and / or the second detection value is reduced compared to the normal state, the control unit 11 reduces the threshold for the pedestrian likelihood and executes the matching process.

[0055] For example, when the accuracy of the first detection value cannot be ensured, the control unit 11 lowers the determination threshold of the first pedestrian likelihood used in the first determination process and executes the first determination process. The control unit 11 lowers the threshold to a threshold (A-α) obtained by subtracting a predetermined value α from a preset first threshold A of the first pedestrian likelihood used in the first determination process, and executes the first determination process. The control unit 11 executes the first determination process by lowering the determination threshold of the first pedestrian likelihood, and also executes the second determination process.

[0056] When the accuracy of the second detection value cannot be ensured, the control unit 11 executes the second determination process by lowering the judgment threshold of the second pedestrian likelihood used in the second determination process. The control unit 11 executes the second determination process by lowering the threshold to a threshold (B-β) obtained by subtracting a predetermined value β from a second threshold B of the second pedestrian likelihood that is set in advance and used in the second determination process. The control unit 11 executes the matching process by using the first recognition result of the first determination process and the second recognition result of the second determination process in which the judgment threshold has been reduced. When the accuracy of the first detection value and the second detection value cannot be ensured, the control unit 11 may execute the matching process by using the first recognition result in which the judgment threshold of the first pedestrian likelihood has been reduced and the second recognition result of the second determination process in which the judgment threshold of the second pedestrian likelihood has been reduced.

[0057] Based on the determination result of the matching process, the control unit 11 determines whether or not a predetermined pedestrian requiring avoidance assist control is present (step S406). If the control unit 11 determines that a predetermined pedestrian is present, it adds 1 to the counter value P (step S408). The control unit 11 determines whether or not the counter value P has reached a preset threshold N for the number of matching matches (step S410).

[0058] If the counter value P has not reached the threshold N of the number of matching attempts, the control unit 11 returns the process to step S100 (see FIG. 3). If the counter value P has reached the threshold N of the number of matching attempts, the control unit 11 advances the process to step S114 and executes avoidance assistance control (see FIG. 3).

[0059] In the matching process of step S404, instead of or in addition to the pedestrian determination using the pedestrian likelihood, the control unit 11 may perform pedestrian determination using the pedestrian speed at which the pedestrian H crosses the road. The control unit 11 determines whether the error between the pedestrian speed at which the pedestrian H crosses the road based on the first recognition result and the pedestrian speed at which the pedestrian H crosses the road based on the second recognition result is within a threshold in the matching process. If the error in the pedestrian speed in the matching process is within the threshold, the control unit 11 determines that the object is the same pedestrian. If sufficient detection values ​​are not obtained and the accuracy of object recognition is lower than in a normal state, the control unit 11 increases the threshold for the error in the pedestrian speed in the matching process compared to in a normal state and executes the matching process.

[0060] When a condition that causes the accuracy of object recognition to decrease compared to normal conditions is satisfied in the environment around the vehicle, and a pedestrian H1 is recognized near the road (see FIG. 4), the control unit 11 determines whether or not there is a possibility that the pedestrian H1 will cross the road. When determining whether the pedestrian H1 is likely to cross the road, the control unit 11 executes the determination process by lowering the threshold standard used when determining whether a pedestrian H (see FIG. 4) is crossing the road.

[0061] For example, when a condition is met in the environment around the vehicle that causes the accuracy of object recognition to decrease compared to normal conditions, the control unit 11 uses a threshold value (VH1-C1) obtained by subtracting a predetermined value C1 from a preset first threshold value VH1 in the first determination process using the first detection value. The control unit 11 determines whether the pedestrian H1 is likely to cross based on the comparison result between the crossing speed VH of the pedestrian H1 and the threshold value (VH1-C1).

[0062] In the second determination process using the second detection value, the control unit 11 uses a threshold value (VH2-C2) obtained by subtracting a predetermined value C2 from a preset second threshold value VH2. The control unit 11 determines whether or not the pedestrian H2 is capable of crossing based on the comparison result between the crossing speed VH of the pedestrian H2 and the threshold value (VH2-C2). The control unit 11 executes a matching process using the first recognition result of the first determination process in which the threshold value is reduced and the second recognition result of the second determination process in which the threshold value is reduced.

[0063] According to the above process, if the accuracy of object recognition in the environment around the vehicle is lower than normal, the threshold for the crossing speed of pedestrian H1 is lowered and the matching process is performed, thereby making it possible to determine whether or not there is a pedestrian H1 who may be crossing the road.

[0064] 8 shows another example of a processing flow for determining whether or not a pedestrian requiring avoidance assist control is present when a condition is satisfied in the environment around the vehicle that causes the accuracy of object recognition to decrease compared to a normal state. In step S104 (see FIG. 3), if control unit 11 determines that the first condition is satisfied (step S104: Yes), control unit 11 executes a matching process for matching a first recognition result based on a first detection value of the object with a second recognition result based on a second detection value of the object. If the matching results match in the matching process and control unit 11 determines that the object is a pedestrian requiring avoidance assist control, and if the position error between the position of the pedestrian recognized in the first recognition result and the position of the pedestrian recognized in the second recognition result is equal to or smaller than a threshold, control unit 11 executes avoidance assist control (see step S114).

[0065] If the control unit 11 determines that the first determination process based on the first detection value does not satisfy the first condition (step S104: No), it determines whether or not a predetermined condition exists (step S105). If the control unit 11 determines that the predetermined condition does not exist, it returns the process to step S100 (see FIG. 3).

[0066] If it is determined that the predetermined condition exists, the control unit 11 determines whether the distance L1 between the oncoming vehicle T and the pedestrian H is within the range of the specified distance S1 (step S502). If the distance L1 between the oncoming vehicle T and the pedestrian H is within the range of the specified distance S1, the control unit 11 increases the threshold value of the pedestrian position error compared to the normal state and executes the matching process (step S504).

[0067] Based on the determination result of the matching process, the control unit 11 determines whether the object is a predetermined pedestrian that requires avoidance assist control (step S506). If the control unit 11 determines that a predetermined pedestrian is present, it adds 1 to the counter value P (step S508). The control unit 11 determines whether the counter value P has reached a preset threshold N for the number of matching matches (step S510).

[0068] If the counter value P has not reached the threshold N of the number of matching attempts, the control unit 11 returns the process to step S100 (see FIG. 3). If the counter value P has reached the threshold N of the number of matching attempts, the control unit 11 proceeds to step S114 and executes avoidance assistance control (see FIG. 3). According to the above process, even if the accuracy of object recognition is reduced compared to normal conditions due to environmental influences such as backlighting, the allowable range (threshold) of position error is made larger than normal conditions, and the frequency of recognition of pedestrian H is increased.

[0069] As described above, according to the vehicle control device 10, even when the oncoming vehicle T is in a backlit state, such as when its headlights are on, and the headlights of the vehicle 1 are also on, making it difficult to process pedestrians based on the detection values, delays in the pedestrian determination process and the avoidance assistance control can be suppressed.

[0070] In the above-described embodiment, the computer program executed in each component of the vehicle control device 10 may be provided in a form recorded on a computer-readable, portable, non-transitory recording medium such as a semiconductor memory, a magnetic recording medium, or an optical recording medium. [Explanation of symbols]

[0071] 1 vehicle, 2 detection unit, 2A exterior camera, 2B lidar device, 2C radar device, 2D position sensor, 3 input / output unit, 5 drive unit, 6 braking unit, 7 steering unit, 10 vehicle control device, 11 control unit, 12 memory unit, H, H1 pedestrian, T oncoming vehicle

Claims

1. a control unit that executes avoidance assistance control to decelerate or avoid an object in a case where an object is recognized ahead of the vehicle based on a first detection value and a second detection value that detect an environment around the vehicle, The control unit performing a matching process of matching a first recognition result based on the first detection value of the object with a second recognition result based on the second detection value of the object; executing the avoidance assist control when a match is found in the matching process for the object and the object is determined to be a pedestrian requiring the avoidance assist control, and when the number of matching matches in the matching process is equal to or greater than a threshold; When a condition is satisfied in which the accuracy of the object recognition is reduced compared to a normal state in the environment around the vehicle, the threshold value for the number of matching times is reduced. Vehicle control device.

2. a control unit that executes avoidance assistance control to decelerate or avoid an object in a case where an object is recognized ahead of the vehicle based on a first detection value and a second detection value that detect an environment around the vehicle, The control unit performing a matching process of matching a first recognition result based on the first detection value of the object with a second recognition result based on the second detection value of the object; when the matching result of the object in the matching process matches and it is determined that the object is a pedestrian requiring the avoidance assist control, and when a speed error between a moving speed of the pedestrian recognized in the first recognition result and a moving speed of the pedestrian recognized in the second recognition result is equal to or less than a threshold, execute the avoidance assist control; increasing the threshold value for the speed error when a condition is met in which the accuracy of the object recognition is reduced compared to a normal state in the environment around the vehicle; Vehicle control device.

3. a control unit that executes avoidance assistance control to decelerate or avoid an object in a case where an object is recognized ahead of the vehicle based on a first detection value and a second detection value that detect an environment around the vehicle, The control unit performing a matching process of matching a first recognition result based on the first detection value of the object with a second recognition result based on the second detection value of the object; when a match is found in the matching process for the object and the object is determined to be a pedestrian requiring the avoidance assist control, and when a position error between the position of the pedestrian recognized in the first recognition result and the position of the pedestrian recognized in the second recognition result is equal to or smaller than a threshold, execute the avoidance assist control; increasing the threshold value for the position error when a condition is met in which the accuracy of the recognition of the object is reduced compared to a normal state in the environment around the vehicle; Vehicle control device.

4. a control unit that executes avoidance assistance control to decelerate or avoid an object in a case where an object is recognized ahead of the vehicle based on a first detection value and a second detection value that detect an environment around the vehicle, The control unit performing a matching process of matching a first recognition result based on the first detection value of the object with a second recognition result based on the second detection value of the object; when a match is found in the matching process for the object and the object is determined to be a pedestrian requiring the avoidance assist control, and when a pedestrian likelihood that is an index for estimating the pedestrian based on the first detection value and the second detection value is equal to or greater than a threshold, the avoidance assist control is executed; reducing the threshold value for the pedestrian likelihood when a condition is satisfied in which accuracy of recognition of the object based on the first detection value and / or the second detection value is reduced compared to a normal state in an environment around the vehicle; Vehicle control device.

5. a control unit that executes avoidance assistance control to decelerate or avoid an object in a case where an object is recognized in front of the vehicle based on detection values ​​of a first detector and a second detector that detect an environment around the vehicle, The control unit performing a matching process for matching a first recognition result of the object obtained by the first detector with a second recognition result of the object obtained by the second detector; When a match is found in the matching process and the object is determined to be a pedestrian requiring the avoidance support control, the avoidance support control is executed; when the accuracy of the recognition of the object by the first detector has decreased compared to a normal state in the environment around the vehicle, the verification process is not performed, and a determination is made as to whether or not the object is the pedestrian requiring the avoidance assist control is made based on the second recognition result by the second detector. Vehicle control device.

Images