Vehicle control device, object detection method, and vehicle control system

Abstract

For example, the present invention provides an object detection device and object detection method that can determine an object to be stationary even in situations where an object would otherwise be mistakenly identified as a moving object moving back and forth due to angular errors. [Solution] For example, in the circumferential direction, it is determined whether the object (detected object) has moved in both clockwise and counterclockwise directions (right / left), and a confidence level is added or subtracted for each of these directions. Then, using the result of the added or subtracted confidence level, it is determined whether the object (detected object) is moving (including the direction of movement) or stationary.

Description

[Technical Field]

[0001] The present invention relates to an object detection device and an object detection method that can distinguish between stationary and moving objects, for example, even in situations where angular accuracy deteriorates. [Background technology]

[0002] Recent vehicles are required to have improved collision safety performance for pedestrians. In particular, the Agreement Regulations apply as national laws, making their installation mandatory. Among these, Agreement Regulation No. 159 (UNR159) requires warning output even when the vehicle and the object being warned about are stationary (both ground speed and relative speed are 0 km / h). Previously, false detection and false warnings were suppressed by not recognizing stationary objects as objects (see Figure 1(a)), but with the addition of the aforementioned regulations, detection of stationary objects has become mandatory. For example, radar has high accuracy in distance and speed in the line of sight, but it cannot measure speed in the circumferential direction, and its angular accuracy is worse than its distance accuracy. Therefore, when stationary objects are used as the target for detection (warning judgment), it is not possible to distinguish between moving objects (pedestrians crossing) and stationary objects due to angular errors, and stationary objects may be mistakenly judged as moving objects (crossing objects) (see Figure 1(b)). In particular, when multiple stationary objects are equidistant, angular accuracy tends to deteriorate due to the precise position of the detection points and multipath.

[0003] Patent Document 1 describes a conventional technique for determining whether an object with a velocity of 0 m / s in the line of sight is stationary (a wall) or moving.

[0004] Patent Document 1 describes a system that, when a wall is present on the side of the vehicle while it is moving, can determine whether the wall is a moving object or not by comparing the movement history with the current detection point if the wall is mistakenly identified as a moving object. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2022-133623 [Overview of the project] [Problems that the invention aims to solve]

[0006] However, Patent Document 1 assumes that detection points occur in addition to the target, and is therefore not applicable when only one detection point is output for an object. In situations where multiple stationary objects are equidistant and only one detection point is detected, a large angular error may occur, potentially leading to the stationary object being mistakenly identified as a moving object that moves back and forth.

[0007] The present invention has been made in view of the above problems, and aims to provide an object detection device and object detection method that can determine an object to be stationary even in situations where, for example, an object would be mistakenly identified as a moving object reciprocating from side to side due to angular errors. [Means for solving the problem]

[0008] To solve the above problems, the object detection device according to the present invention includes: a ground speed acquisition unit that acquires the relative position and relative speed between the vehicle and an object to be detected around the vehicle, detected by an object recognition sensor mounted on the vehicle, and acquires the ground speed of the object to be detected based on the relative position, the relative speed, the vehicle's speed, and the vehicle's yaw rate; a determination unit that acquires a first direction confidence score, which is an index indicating the movement of the object to be detected in a predetermined first direction, based on the ground speed, acquires a second direction confidence score, which is an index indicating the movement of the object to be detected in a second direction opposite to the first direction, based on the ground speed, and determines whether the object to be detected is stationary or moving based on the first direction confidence score and the second direction confidence score.

[0009] Moreover, the object detection method according to the present invention is an object detection method executed by a control unit mounted on a vehicle. The control unit acquires the relative position and relative speed between a detection target around the vehicle and the vehicle detected by an object recognition sensor mounted on the vehicle, and based on the relative position, the relative speed, the speed of the vehicle, and the yaw rate of the vehicle, acquires the ground speed of the detection target. Based on the ground speed, a first direction reliability, which is an index indicating the movement of the detection target with respect to a predetermined first direction, is acquired. Based on the ground speed, a second direction reliability, which is an index indicating the movement of the detection target with respect to a second direction that is the opposite direction of the first direction, is acquired. Based on the first direction reliability and the second direction reliability, it is determined whether the detection target is a stationary object or a moving object.

Advantages of the Invention

[0010] According to the present invention, for example, even in a situation where it is determined to be a moving object that reciprocates left and right due to an angle error, it can be determined to be a stationary object.

[0011] Problems, configurations, and effects other than those described above will be clarified by the description of the following embodiments.

Brief Description of the Drawings

[0012] [Figure 1] Explanation diagram of the conventional concept for stationary objects and problems in the case of detection (warning determination) target. [Figure 2] System configuration example of the object detection device in Example 1. [Figure 3] Calculation image diagram of the ground speed in the line-of-sight direction and the ground speed in the circumferential direction. [Figure 4] Flowchart of the line-of-sight direction movement determination unit 13 in Example 1. [Figure 5] Flowchart of the circumferential direction movement determination unit 14 in Example 1. [Figure 6] Truth table of stationary object and moving direction determination using the movement direction reliability. [Figure 7] Explanation diagram of the simulation environment in Example 1. [Figure 8] Parameters used in the simulation (conventional method). [Figure 9] Parameters used in the simulation (Example 1). [Figure 10] A diagram illustrating the simulation results. [Figure 11] An explanatory diagram illustrating the concept of error using a monocular camera in Example 2. [Figure 12] Flowchart of the circumferential movement determination unit 14 in Example 2. [Figure 13] Flowchart of the gaze direction movement determination unit 13 in Example 2. [Figure 14] Diagram illustrating the factors that prevent the application of Example 3 during high-speed driving. [Modes for carrying out the invention]

[0013] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0014] Conventionally, as shown in Figure 1(a), when an object recognition sensor 21 mounted on a vehicle (own vehicle) 1 recognizes (detects) objects 2 present around the vehicle 1, false detections and false alarms were suppressed by not recognizing stationary objects as objects. As a result, stationary moving objects (pedestrians crossing the road, etc.) 2a that are the target of alarms could not be detected. For example, the radar 21a used as the object recognition sensor 21 has poor angular accuracy compared to its distance accuracy. Therefore, when a stationary object is used as the target of detection (alarm judgment) using radar 21a as the object recognition sensor 21, as shown in Figure 1(b), it may be impossible to distinguish between a moving object (pedestrians crossing the road, etc.) 2a and an angular error, and a stationary object (telephone pole, sign, etc.) 2b may be mistakenly judged as a moving object (crossing object). The present invention applies the embodiments described below to address these problems.

[0015] [Example 1] The following describes the vehicle control device 10, which is an object detection device according to Embodiment 1 of the present invention, with reference to Figures 2 to 10.

[0016] Figure 2 is a functional block diagram of the driver assistance system installed in the vehicle 1 of this embodiment. This driver assistance system has an object recognition sensor 21, a vehicle speed sensor 22, a steering angle sensor 23, and a yaw rate sensor 24 on the input side of the vehicle control device 10, and a warning device 31 on the output side. Each of these is assumed to be connected to communicate via CAN (Controller Area Network) or the like.

[0017] The object recognition sensor 21 is a sensor that acquires information about objects 2 (for example, moving objects 2a such as other vehicles or pedestrians) around the vehicle 1, and is, for example, a radar 21a or a monocular camera 21b. If the object recognition sensor 21 is a monocular camera 21b, the relative position P of object 2 can be obtained from each captured data. TR (x TR ,y TR The relative velocity V of object 2 is measured sequentially, and the displacement of object 2 in each image is calculated from the displacement of object 2. TR This allows for the calculation of the relative position and relative velocity of object 2. Furthermore, if the object recognition sensor 21 is a radar 21a or the like, the relative position and relative velocity of object 2 can be measured directly.

[0018] In this embodiment 1, we will explain the control process when the object recognition sensor 21 is a radar 21a.

[0019] The vehicle speed sensor 22 detects the vehicle speed V S This is a sensor that detects the vehicle speed V, for example, based on the rotational speed of the tires output by the wheel speed sensor, assuming that the vehicle 1 is moving in a straight line. S Calculate.

[0020] The steering angle sensor 23 is a sensor that acquires the steering angle (amount) of the vehicle 1, and is, for example, an angle sensor attached to the steering wheel.

[0021] The yaw rate sensor 24 is a sensor that acquires the yaw rate ω of the vehicle 1, and is, for example, an acceleration sensor that detects the acceleration around the vertical axis of the vehicle 1. Note that the yaw rate ω can also be calculated using the vehicle speed sensor 22 and the steering angle sensor 23, so the yaw rate sensor 24 may be omitted.

[0022] The warning device 31 is a device that, when the vehicle control device 10 outputs a warning request, warns the driver of the possibility of collision with object 2 via a display, LED illumination, steering control, voice notification, etc., and prompts the driver to take appropriate evasive action. In Figure 2, the warning device 31 prompts the driver to take evasive action, but the system may also be configured so that the vehicle control device 10 directly controls the braking and steering of the vehicle 1 when there is a possibility of collision.

[0023] <Vehicle control device 10> Next, the system configuration of the vehicle control device 10 in this embodiment will be described. As shown in Figure 2, the vehicle control device 10, which is an object detection device, includes a parameter storage unit 11, an object ground speed calculation unit 12, a gaze direction movement determination unit 13, a circumferential movement determination unit 14, and an alarm unit 15. The object ground speed calculation unit 12, gaze direction movement determination unit 13, circumferential movement determination unit 14, and alarm unit 15 periodically execute processing, and when predetermined conditions are met, an alarm request is created and output to the alarm device 31. Specifically, the vehicle control device 10 is an ECU (Electronic Control Unit) equipped with hardware such as a CPU or other arithmetic unit, a memory device such as semiconductor memory, and a communication device. The arithmetic unit executes a program loaded into the memory device to realize each function such as the gaze direction movement determination unit 13, but in the following description, such well-known technologies will be omitted as appropriate.

[0024] The parameter storage unit 11 is a storage device that stores vehicle parameters and warning parameters. The vehicle parameters are mainly parameters used by the circumferential movement determination unit 14, and are, for example, parameters related to the specifications of the host vehicle 1 such as the movement threshold value and reliability threshold value in the circumferential direction, wheelbase, gear ratio for converting the steering angle to the tire angle, and the like. The warning parameters are mainly parameters used by the warning unit 15, and are parameters such as the TTC threshold value for comparison with the time to collision (Time To Collision, hereinafter referred to as "TTC") of the warning candidate, and the like.

[0025] The object ground speed calculation unit 12 can obtain the ground speed from the vehicle speed of the vehicle 1 and the object relative speed of the object recognition sensor 21. However, the method for calculating the ground speed is not particularly limited, and a known ground speed calculation method can be used. Also, the relative speed may be calculated using the object relative position of the object recognition sensor 21, and the ground speed may be calculated in combination with the aforementioned vehicle information (the vehicle speed and yaw rate of the vehicle 1).

[0026] As shown in FIG. 3, the line-of-sight direction movement determination unit 13 calculates the ground speed in the line-of-sight direction of the object recognition sensor 21 using V TA from the object ground speed V calculated by the object ground speed calculation unit 12, and uses V rad =V TA cosθ, and is a process of determining whether the object is a stationary object or a moving object using the ground speed in the line-of-sight direction of the object recognition sensor 21. The line-of-sight direction movement determination unit 13 may be implemented in combination with the circumferential movement determination unit 14. In the first embodiment, the process of the line-of-sight direction movement determination unit using the radar 21a will be described. In the second embodiment described later, the process of the line-of-sight direction movement determination unit using the monocular camera 21b will be described. Also, in the third embodiment, the reason for using the conventional stationary object determination result during high-speed driving will be described.

[0027] As shown in FIG. 3, the circumferential movement determination unit 14 calculates the object ground speed V calculated by the object ground speed calculation unit 12, and from V TA from, V cir =V TAThis process calculates the ground velocity of the object recognition sensor 21 in the circumferential direction using sinθ, and then uses the ground velocity of the object recognition sensor 21 in the circumferential direction to determine whether the object is stationary or moving. Example 1 describes the processing of the circumferential movement determination unit using radar 21a. Example 2, described later, describes the processing of the circumferential movement determination unit using monocular camera 21b. Furthermore, Example 3 explains the reason for using the conventional stationary object determination result when traveling at high speed.

[0028] The alarm unit 15 uses the object recognition sensor 21, vehicle speed sensor 22, steering angle sensor 23, yaw rate sensor 24, parameter storage unit 11, object ground speed calculation unit 12, gaze direction movement determination unit 13, and circumferential direction movement determination unit 14 to calculate the possibility of collision with object 2 and the TTC, and outputs an alarm request to alert the driver. The alarm unit 15 may output an alarm using the alarm device 31, or the ECU itself may output an alarm. In addition, the alarm unit 15 may control the vehicle in addition to issuing alarms, such as by controlling the brakes or steering.

[0029] Next, the gaze direction movement determination unit 13 will be explained using the flowchart in Figure 4.

[0030] Step S1: First, the line of sight direction movement determination unit 13 determines the object ground velocity V calculated by the object ground velocity calculation unit 12, as shown in Figure 3. TA From, V rad =V TA V is the line-of-sight component of the object's velocity relative to the ground, using cosθ. rad Calculate.

[0031] Step S2: Next, V is the line-of-sight component of the object's velocity relative to the ground calculated in step S1. rad This is used to determine whether an object is stationary or moving. Specifically, the line-of-sight component V of the object's velocity relative to the ground. rad A determination is made as to whether or not it is within the range from the first threshold to the second threshold. The line-of-sight component V of the object's velocity relative to the ground. radWhen the value falls within the range of the first threshold to the second threshold, the object can be determined to be not moving in the line of sight direction. The line of sight component is considered equivalent to a stationary object, and step S3 is performed. Line of sight component V of object velocity relative to ground rad When the value is outside the range from the first threshold to the second threshold, it is determined that the line-of-sight component is moving, and step S4 is performed.

[0032] Step S3: Since the line-of-sight component was determined to be equivalent to a stationary object in step S2, if the object is also stationary in the circumferential direction, then the object can be determined to be stationary. Therefore, the circumferential movement determination unit 14 is executed to determine whether the object is stationary or moving.

[0033] Step S4: In step S3, the line-of-sight component is determined to correspond to a moving object. Since the line-of-sight velocity component of radar 21a is the result of direct measurement, the circumferential movement determination unit 14 is not executed, and the object is determined to be a moving object.

[0034] Next, the circumferential movement determination unit 14 will be explained using the flowchart in Figure 5.

[0035] Step S11: First, the line of sight direction movement determination unit 13 determines the object ground velocity V calculated by the object ground velocity calculation unit 12, as shown in Figure 3. TA From, V cir =V TA V is the circumferential component of the object's velocity relative to the ground, using sinθ. cir Calculate.

[0036] Step S12: Next, the circumferential component V of the object's velocity relative to the ground calculated in step S11. cir This is used to determine whether the object is moving in the first (counterclockwise or left) direction of the circumferential component. That is, the circumferential component V of the object's velocity relative to the ground. cir The determination is made as to whether the value is within the range from the first threshold to the second threshold (within the range of movement in the first direction). The circumferential component V of the object's velocity relative to the ground. cirIf it is within the range of movement in the first direction, step S13 is performed to add the confidence in the first direction. Circumferential component V of the object's velocity relative to the ground. cir If the object is outside the range of movement in the first direction, it is determined to be a stationary object, and step S14 is performed to subtract the confidence level in the first direction. Multiple confidence thresholds and confidence addition / subtraction values ​​may be used with a table or the like for adding or subtracting the confidence level, which is an indicator of the movement of an object in the first direction.

[0037] Step S13: Based on the judgment result in step S12, it is determined that the object is moving in the first direction (counterclockwise or left), and the confidence score for the first direction is added.

[0038] Step S14: Based on the judgment result in step S12, the object is determined to be stationary, and the confidence level in the first direction is subtracted.

[0039] Step S15: Next, the circumferential component V of the object's velocity relative to the ground calculated in step S11. cir This is used to determine whether the object is moving in the second (clockwise or to the right) direction of the circumferential component. That is, the circumferential component V of the object's velocity relative to the ground. cir The determination is made as to whether or not it is within the range of the third to fourth threshold (within the range of movement in the second direction). The circumferential component V of the object's velocity relative to the ground. cir If it is within the range of movement in the second direction, step S16 is performed to add the confidence in the second direction. Circumferential component V of the object's velocity relative to the ground. cir If the object is outside the range of movement in the second direction, it is determined to be a stationary object, and step S17 is performed to subtract the confidence level in the second direction. For adding or subtracting the confidence level in the second direction, which is an indicator of the movement of an object in the second direction which is the opposite direction to the first direction, multiple confidence thresholds and confidence addition / subtraction values ​​may be used using a table or the like.

[0040] Step S16: Based on the judgment result in step S15, it is determined that the object is moving in the second direction (clockwise or to the right), and the confidence score for the second direction is added.

[0041] Step S17: Based on the judgment result in step S15, the object is determined to be stationary, and the second direction confidence is subtracted.

[0042] Step S18: Based on step S12, the first-direction confidence and second-direction confidence obtained by performing addition and subtraction in steps S13, S14, and steps S15, S16, S17 are used to determine whether the object is stationary or moving, based on the truth table in Figure 6. If only the first-direction confidence is above the threshold (the second-direction confidence is below the threshold), it is determined that the object is moving in the first direction (in other words, the direction of movement of the moving object is the first direction). If only the second-direction confidence is above the threshold (the first-direction confidence is below the threshold), it is determined that the object is moving in the second direction (in other words, the direction of movement of the moving object is the second direction). If both the first-direction confidence and the second-direction confidence are below the threshold, the object is not moving and is therefore determined to be stationary. If both the first-direction confidence and the second-direction confidence are above the threshold, it is determined that the object is diverging and is therefore determined to be stationary.

[0043] The threshold for first-way confidence and the threshold for second-way confidence may be the same value or different values.

[0044] <Simulation Results> This section describes the results of applying this embodiment to the situation when the problem occurred. In the scenario, as shown in Figure 7, there is a stationary object (wall, etc.) 2c and a stationary object (structure) 2b whose detection point is the same, and both the vehicle 1 and object 2 are stationary. Measurement data (circumferential component V of object-to-ground velocity) cir We compared the results obtained by applying the parameters shown in Figure 8, which represent the conventional method, with the results obtained by applying the parameters shown in Figure 9, which were set in this embodiment 1.

[0045] The results of the application are shown in Figure 10. With the conventional method, the confidence level exceeds the threshold, resulting in the stationary object being mistakenly identified as a moving object. On the other hand, when the settings of this embodiment 1 are applied, the confidence level falls below the threshold, and the object is detected as stationary.

[0046] Therefore, by using this embodiment, stationary objects can be recognized as stationary objects without being misidentified as moving objects, thereby improving the accuracy of separating moving and stationary objects. This allows for the determination of potential collisions with appropriate objects and the implementation of warnings and control measures.

[0047] [Example 2] Next, using Figures 11 to 13, we will describe the vehicle control device 10, which is an object detection device according to Embodiment 2 of the present invention. Embodiment 2 will explain the difference when the object recognition sensor 21 used in Embodiment 1 is changed from radar 21a to monocular camera 21b.

[0048] As shown in Figure 11, radar 21a has high accuracy in the line of sight direction but poor accuracy in the circumferential direction (angle errors are likely to occur). On the other hand, monocular camera 21b has poor accuracy in the line of sight direction (distance errors are likely to occur) but high accuracy in the circumferential direction. Therefore, this embodiment is applied to the line of sight direction where accuracy is likely to deteriorate.

[0049] Therefore, in the vehicle control device 10, the processing order in which the object-to-ground velocity calculation unit 12, line-of-sight movement determination unit 13, and circumferential movement determination unit 14 were performed in the first embodiment, is changed in the second embodiment to the object-to-ground velocity calculation unit 12, circumferential movement determination unit 14, and line-of-sight movement determination unit 13.

[0050] Furthermore, the circumferential movement determination unit 14 uses the same method as the line-of-sight movement determination unit 13 of Embodiment 1 applied to the circumferential direction, as shown in Figure 12. Also, the line-of-sight movement determination unit 13 uses the same method as the circumferential movement determination unit 14 of Embodiment 1 applied to the line-of-sight direction, as shown in Figure 13. By performing the process shown in Figure 13, the line-of-sight movement determination unit 13 can determine whether an object is moving in the first (approaching) direction in the line-of-sight direction (in other words, the direction of movement of the moving object is the first (approaching) direction), whether an object is moving in the second (separating) direction in the line-of-sight direction (in other words, the direction of movement of the moving object is the second (separating) direction), or whether an object is stationary.

[0051] Therefore, in this embodiment as well, stationary objects are recognized as stationary objects without being misidentified as moving objects, thereby improving the accuracy of separating moving and stationary objects. This allows for the determination of potential collisions with appropriate objects and the implementation of warnings and control measures.

[0052] [Example 3] Next, Embodiment 3 of the present invention will be described using Figure 14. Embodiment 3 explains why the circumferential movement determination unit 14 (or the gaze direction movement determination unit 13, which is the determination unit in Embodiment 1) is performed only at low speed (vehicle speed is below a predetermined value) or when stopped.

[0053] As shown in Figure 14, we investigated the cases when vehicle 1 is moving forward at high speed, when object 2 is in front of vehicle 1, when object 2 is to the side of vehicle 1, and when the detection target is a vertically long object 2 such as a wall.

[0054] As shown in Figure 14(a), when object 2 is sufficiently far away and in front of vehicle 1, it is possible to determine that it is a stationary object using a known method that uses line-of-sight velocity, so it was determined that this method is not necessary.

[0055] As shown in Figure 14(b), when object 2 is located to the side of vehicle 1, the change in angle when passing to the side is significantly larger than the angle error, which is the problem of this invention. Therefore, it was determined that the effectiveness of this invention is difficult to confirm.

[0056] As shown in Figure 14(c), when a stationary object, including a wall, is located directly to the side, known methods for detecting walls, including those described in Patent Document 1, can be applied. However, since stationary objects can be identified from surrounding detection points, it was determined that application of these methods was unnecessary.

[0057] As vehicle speed decreases, the influence of angular errors due to noise and multiple stationary objects becomes greater. Also, when vehicle speed is sufficiently high, displacement is more influenced by positional changes than by angular changes. Therefore, this embodiment effectively reduces processing time by determining whether an object (detection target) is stationary or moving only when the speed of vehicle 1 is below a predetermined value, especially when vehicle 1 is stationary, and only when the difference in line-of-sight velocity of stationary objects due to angle is small, and by not making a determination when the influence of positional changes is large.

[0058] [Summary of Examples 1-3] As explained above, the object detection device (vehicle control device 10) of this embodiment is The relative position and relative speed between the vehicle and the detected object in the surrounding area, as detected by the object recognition sensor 21 (radar 21a or monocular camera 21b) mounted on the vehicle, are acquired. A ground speed acquisition unit (object ground speed calculation unit 12) acquires the ground speed of the object to be detected based on the relative position, the relative speed, the vehicle speed, and the vehicle yaw rate, Based on the ground velocity, a first-direction confidence score is obtained, which is an index indicating the movement of the detected object in a predetermined first direction (leftward in the circumferential direction centered on the object recognition sensor 21, or approaching in the line of sight direction of the object recognition sensor 21). Based on the ground velocity, a second direction confidence score is obtained, which is an index indicating the movement of the detected object in a second direction (to the right in the circumferential direction centered on the object recognition sensor 21, or in the direction of separation in the line of sight of the object recognition sensor 21), which is the opposite direction to the first direction. The system includes a determination unit (circumferential movement determination unit 14, or line-of-sight movement determination unit 13) that determines whether the object to be detected is stationary or moving based on the first directional reliability and the second directional reliability.

[0059] Furthermore, the object detection method in this embodiment is an object detection method executed by a control unit (vehicle control device 10) mounted on the vehicle, The aforementioned control unit The object recognition sensor 21 (radar 21a or monocular camera 21b) mounted on the vehicle detects the relative position and relative speed of the detected object in the vicinity of the vehicle and the vehicle, and the relative position and relative speed of the vehicle are acquired. Based on the relative position, relative speed, vehicle speed, and vehicle yaw rate, the ground speed of the object to be detected is obtained. Based on the ground velocity, a first-direction confidence score is obtained, which is an index indicating the movement of the detected object in a predetermined first direction (leftward in the circumferential direction centered on the object recognition sensor 21, or approaching in the line of sight direction of the object recognition sensor 21). Based on the ground velocity, a second direction confidence score is obtained, which is an index indicating the movement of the detected object in a second direction (to the right in the circumferential direction centered on the object recognition sensor 21, or in the direction of separation in the line of sight of the object recognition sensor 21), which is the opposite direction to the first direction. Based on the first directional reliability and the second directional reliability, it is determined whether the object to be detected is stationary or moving.

[0060] In other words, for example, in the circumferential direction, it is determined whether the object (detected object) has moved in both clockwise and counterclockwise directions (right / left), and a confidence level is added or subtracted for each of these directions. Then, using the result of the added or subtracted confidence level, it is determined whether the object (detected object) is moving (including the direction of movement) or stationary.

[0061] According to this embodiment, for example, even in situations where an object would be mistakenly identified as a moving object reciprocating from side to side due to an angular error, it can be identified as a stationary object.

[0062] In other words, in situations where a large angular error occurs, making a stationary object appear to be moving, it is possible to prevent misidentification as a crossing object by using confidence levels decomposed in the direction of movement (for example, according to confidence levels separately set for the left and right directions) to determine whether the detected object is stationary or moving.

[0063] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are explained in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described.

[0064] Furthermore, each of the above configurations, functions, processing units, and processing means may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. Alternatively, each of the above configurations and functions may be implemented in software by having the processor interpret and execute programs that implement each function. Information such as programs, tables, and files that implement each function can be stored in memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD. [Explanation of symbols]

[0065] 1. Vehicle (Own vehicle) 10. Vehicle control device (object detection device) 11 Parameter storage section 12. Object-to-Ground Velocity Calculation Unit (Ground Velocity Acquisition Unit) 13. Eye-line movement determination unit (determination unit in Example 2) 14. Circumferential movement determination unit (determination unit of Example 1) 15 Alarm section 21 Object Recognition Sensor 21a Radar 21b Monocular camera 22 Vehicle speed sensor 23. Steering angle sensor 24 Yaw rate sensor 31 Alarm device 2 objects 2a Moving objects (pedestrians, etc.) 2b Stationary objects (telephone poles, signs, etc.) 2c Stationary object (wall, etc.) V TA Velocity of an object relative to the ground V rad The line-of-sight component of the object's velocity relative to the ground. V cir Circumferential component of object-to-ground velocity

Claims

1. The relative position and relative speed between the vehicle and the detected object in the vehicle's vicinity, as detected by the object recognition sensor mounted on the vehicle, are acquired. A ground speed acquisition unit that acquires the ground speed of the object to be detected based on the relative position, the relative speed, the vehicle speed, and the vehicle's yaw rate, Based on the ground velocity, a first-direction confidence index is obtained, which is an index indicating the movement of the object to be detected in a predetermined first direction. Based on the ground velocity, a second direction confidence index is obtained, which is an index indicating the movement of the object to be detected in a second direction, which is the opposite direction to the first direction. An object detection device comprising: a determination unit that determines whether the object to be detected is stationary or moving based on the first directional reliability and the second directional reliability.

2. An object detection method performed by a control unit mounted on a vehicle, The aforementioned control unit The relative position and relative speed between the detected object in the vicinity of the vehicle and the vehicle, as detected by the object recognition sensor mounted on the vehicle, are acquired. Based on the relative position, relative speed, vehicle speed, and vehicle yaw rate, the ground speed of the object to be detected is obtained. Based on the ground velocity, a first-direction confidence index is obtained, which is an index indicating the movement of the object to be detected in a predetermined first direction. Based on the ground velocity, a second direction confidence index is obtained, which is an index indicating the movement of the object to be detected in a second direction, which is the opposite direction to the first direction. An object detection method that determines whether the object to be detected is stationary or moving based on the first directional reliability and the second directional reliability.

Images