Parking Support Method and Parking Support Device

Abstract

This parking assistance method involves: storing, as learned object positions, a plurality of positions of objects detected around a target parking position when parking an own vehicle at a target parking position (S2); when the own vehicle travels in the vicinity of the target parking position after storage of the learned object positions, counting, for a plurality of learned object positions, the number of times the learned object positions match with a surrounding object position that is the position of an object detected around the own vehicle, and providing a higher reliability to a learned object position in which the number of times of matching with the surrounding object position is large as compared with a learned object position in which the number of times of matching with the surrounding object position is small (S12, S25); and making a comparison between a learned object position exhibiting at least a predetermined reliability threshold value among the learned object positions and the position of an object detected around the own vehicle, and calculating a relative position of the own vehicle with respect to the target parking position (S22).

Description

Technical Field

[0001] The present invention relates to a parking assistance method and a parking assistance device.

Background Art

[0002] As a technology related to parking assistance of a host vehicle to a target parking position, Patent Document 1 below is known. In Patent Document 1, an object around the target parking position is detected and stored, and based on the position of the object detected around the host vehicle during automatic parking and the position of the stored object, the relative position of the host vehicle with respect to the target parking position is calculated, and the host vehicle is automatically moved to the target parking position based on the relative position.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, when the stored object is a movable object, there is a risk that the calculation accuracy of the relative position of the host vehicle with respect to the target parking position may decrease due to a change in the position of the object after the object is stored. An object of the present invention is to suppress a decrease in the calculation accuracy of the relative position due to a change in the stored object position in parking assistance for calculating the relative position of a host vehicle with respect to a target parking position based on the position of an object around the target parking position stored in advance.

Means for Solving the Problems

[0005] In the parking assistance method according to one aspect of the present invention, when parking the host vehicle at a target parking position, the positions of a plurality of targets detected around the target parking position are stored as learned target positions. After storing the learned target positions, when the host vehicle travels near the target parking position, the number of times the learned target positions match the surrounding target positions, which are the positions of the targets detected around the host vehicle, is counted for each of the plurality of learned target positions. A higher reliability is assigned to the learned target positions with a larger number of matches with the surrounding target positions compared to the learned target positions with a smaller number of matches with the surrounding target positions. The relative position of the host vehicle with respect to the target parking position is calculated by comparing the learned target positions with a reliability equal to or higher than a predetermined reliability threshold among the learned target positions and the positions of the targets detected around the host vehicle. A target travel trajectory from the current position of the host vehicle to the target parking position is calculated based on the calculated relative position, and parking assistance control is executed to assist the movement of the host vehicle along the calculated target travel trajectory.

Advantages of the Invention

[0006] According to the present invention, in parking assistance for calculating the relative position of the host vehicle with respect to the target parking position based on the positions of the targets around the target parking position memorized in advance, it is possible to suppress a decrease in the calculation accuracy of the relative position due to a change in the memorized target positions. The objects and advantages of the present invention are realized and achieved by using the elements and combinations thereof shown in the claims. It should be understood that both the foregoing general description and the following detailed description are merely illustrative and explanatory and are not intended to limit the present invention like the claims.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Mode for Carrying Out the Invention

[0008] (Configuration) FIG. 1 is a diagram showing an example of the schematic configuration of the parking support device according to the embodiment. The host vehicle 1 includes a parking support device 10 that supports the parking of the host vehicle 1 at a target parking position. The parking support device 10 calculates a target travel trajectory from the current position of the host vehicle 1 to the target parking position, and supports the host vehicle to travel along the target travel trajectory. The parking support by the parking support device 10 includes various forms of supporting the host vehicle 1 to travel along the target travel trajectory. For example, the parking of the host vehicle 1 may be supported by performing an automatic driving that controls the host vehicle to travel to the target parking position along the target travel trajectory of the host vehicle 1. Note that the automatic driving that controls the host vehicle 1 to travel to the target parking position along the target travel trajectory of the host vehicle 1 means controlling all or part of the steering angle, driving force, and braking force of the host vehicle, and automatically performing all or part of the travel along the target travel trajectory of the host vehicle 1 to assist the parking operation of the occupant. Also, for example, the parking of the host vehicle 1 may be supported by displaying the target travel trajectory and the current position of the host vehicle 1 on a display device visible to the occupant of the host vehicle 1.

[0009] The parking support device 10 includes a positioning device 11, an object sensor 12, a vehicle sensor 13, a human machine interface (HMI) 15, an actuator 16, and a controller 17. The positioning device 11 measures the current position of the host vehicle 1. The positioning device 11 may include, for example, a global navigation satellite system (GNSS) receiver. The GNSS receiver may be, for example, a global positioning system (GPS) receiver or the like. The object sensor 12 detects an object within a predetermined distance range (for example, the detection area of the object sensor 12) from the host vehicle 1. The object sensor 12 detects the surrounding environment of the host vehicle 1, such as the relative position between the object existing around the host vehicle 1 and the host vehicle 1, the distance between the host vehicle 1 and the object, and the direction in which the object exists. The object sensor 12 may include, for example, a camera that captures the surrounding environment of the host vehicle 1. The camera may be, for example, a camera that captures the surroundings of the host vehicle 1 and generates a captured image that is converted into an aerial view (around view monitor image). The object sensor 12 may also include a distance measuring device such as a laser range finder (LRF), a radar, a LiDAR (Light Detection and Ranging), or a laser radar.

[0010] The vehicle sensor 13 detects various information (vehicle information) obtained from the host vehicle 1. The vehicle sensor 13 may include, for example, a vehicle speed sensor that detects the traveling speed (vehicle speed) of the host vehicle 1, a wheel speed sensor that detects the rotational speed of each tire provided on the host vehicle 1, a three-axis acceleration sensor that detects the acceleration (including deceleration) of the host vehicle 1 in three axial directions, a steering angle sensor that detects the steering angle of the steering wheel, a steering angle sensor that detects the steering angle of the steered wheels, a gyro sensor that detects the angular velocity generated in the host vehicle 1, and a yaw rate sensor that detects the yaw rate. The human machine interface 15 is an interface device that exchanges information between the parking support device 10 and the occupant. The human machine interface 15 includes a display device visible to the occupant of the host vehicle 1 (for example, a display screen of a navigation system or a display device provided near the meter in front of the driver's seat). The human machine interface 15 also includes an operator that receives an operation input from the occupant to the parking support device 10. For example, the operator may be a button, a switch, a lever, a dial, a keyboard, a touch panel, or the like.

[0011] The actuator 16 includes a steering actuator, an accelerator actuator, and a brake control actuator. The steering actuator controls the steering angle of the steering mechanism of the host vehicle 1 in response to a control signal from the controller 17. The accelerator actuator controls the accelerator opening of a drive device, which is an engine or a drive motor, in response to a control signal from the controller 17. The brake actuator operates the braking device in response to a control signal from the controller 17. The controller 17 is an electronic control unit that performs parking assistance control of the host vehicle 1. The controller 17 includes a processor 18 and peripheral components such as a storage device 19. The processor 18 may be, for example, a CPU or an MPU.

[0012] The storage device 19 may include a semiconductor storage device, a magnetic storage device, an optical storage device, or the like. The storage device 19 may include memories such as registers, cache memories, ROM and RAM used as main storage devices. The functions of the controller 17 described below are realized, for example, when the processor 18 executes a computer program stored in the storage device 19. Note that the controller 17 may be formed of dedicated hardware for executing each information process described below. For example, the controller 17 may include a functional logic circuit set in a general-purpose semiconductor integrated circuit. For example, the controller 17 may have a programmable logic device such as a field programmable gate array.

[0013] Next, an example of parking support control by the parking support device 10 will be described. Refer to FIG. 2. When starting to use the parking support by the parking support device 10, first, the relative positional relationship between the target position of the target object existing near the target parking position 2 and the target parking position 2 is stored in the storage device 19. The target parking position 2 is the target position where the host vehicle 1 is to be parked. The target object is a ground object that serves as a landmark for specifying the current position of the host vehicle 1. The target object may be, for example, road markings (lane boundary lines 3a, stop lines 3b, road signs, etc.), road boundaries (curbs 3c to 3e, guardrails, etc.), obstacles (houses 3f, fences 3g, objects 3h, etc.).

[0014] When storing the relative positional relationship between the target position of the target object and the target parking position 2 in the storage device 19, the operation mode of the parking support device 10 is set to the "target object learning mode". Then, the host vehicle 1 is parked at the target parking position 2 by manual driving. Note that the parking support device 10 may automatically set the operation mode to the "target object learning mode" when the host vehicle 1 is parked at the target parking position 2 by manual driving. When the host vehicle 1 is parked at the target parking position 2 by manual driving, the driver may select whether to set the operation mode of the parking support device 10 to the "target object learning mode". While the host vehicle 1 is being parked at the target parking position 2 by manual driving, the target position of the target object around the host vehicle 1 is detected by the object sensor 12. The object sensor 12 is a sensor that detects the target position existing in the detection area within a predetermined detection distance range from the object sensor 12.

[0015] For example, the parking support device 10 may detect edges and corner portions such as road markings (lane boundary lines 3a, stop lines 3b in the example of FIG. 2), road boundaries (ground contact portions of curbs 3c to 3e in the example of FIG. 2), and obstacles (ground contact portions of houses 3f, fences 3g, objects 3h in the example of FIG. 2) on the captured image obtained by photographing with the camera of the object sensor 12 as feature points, and use the positions of the feature points as the target positions of the target object. The parking assistance device 10 calculates the feature amounts of the detected feature points (for example, the shading or attributes of the feature points). For detecting the feature points and calculating the feature amounts, methods such as SIFT, SURF, ORB, BRIAK, KAZE, AKAZE, etc. can be used. Note that the parking assistance device 10 may detect the feature points of the point cloud information obtained by a laser range finder, radar, or LiDAR. In this specification, an example of detecting feature points from a captured image will be described. Note that detection of the feature amounts of these feature points is not essential, and at least the positions of the feature points may be detected.

[0016] The parking assistance device 10 obtains the relative positional relationship between the target position (the position of the feature point) detected by the object sensor 12 and the target parking position 2. For example, the parking assistance device 10 calculates the relative position of the feature point with respect to the host vehicle 1 based on the position of the feature point in the image and the camera information regarding the mounting state (mounting position, optical axis angle, and field angle) of the camera to the host vehicle 1. Next, the current position of the host vehicle 1 in the fixed coordinate system at the time when the object sensor 12 detects the target position is estimated, and based on the estimated current position and the relative position of the target with respect to the host vehicle 1, the target position in the fixed coordinate system is calculated. Here, the fixed coordinate system is a coordinate system with a specific point as the coordinate origin (for example, a map coordinate system). The current position of the host vehicle 1 in the fixed coordinate system may be estimated, for example, by a positioning device 11, odometry, or dead reckoning. The current position of the host vehicle 1 in the fixed coordinate system may also be estimated by map mapping between the target position detected by the object sensor 12 and known target positions or high-precision map information. Next, the parking assistance device 10 specifies the target parking position 2 in the fixed coordinate system. For example, the position of the host vehicle 1 when the host vehicle 1 is located at the target parking position 2 may be detected as the target parking position 2. When the positions of the target position and the target parking position 2 are determined in the fixed coordinate system, the relative positional relationship between the target position and the target parking position 2 is determined.

[0017] The parking assistance device 10 stores in the storage device 19 the relative positional relationship between the target position and the target parking position 2. For example, the storage device 19 may store the target position and the position of the target parking position 2 in a fixed coordinate system respectively. Alternatively, the target position of the target object in a relative coordinate system with the target parking position 2 as the coordinate origin may be calculated and stored in the storage device 19. Here, an example in which the storage device 19 stores the target position and the position of the target parking position 2 in a fixed coordinate system will be described. In the following description, the target object stored in the storage device 19 may be referred to as a "learned target object". Also, the target position of the learned target object may be referred to as the "learned target position". The round plots in FIG. 2 indicate the learned target positions stored in the storage device 19.

[0018] Next, a parking assistance method by the parking assistance device 10 will be described with reference to FIG. 3. To use the parking assistance, the operation mode of the parking assistance device 10 is set to the "parking assistance mode". In the parking assistance mode, the parking assistance device 10 reads out the target position of the learned target object stored in the storage device 19 and the position of the target parking position 2. The round plots in FIG. 3 indicate the target positions of the learned target objects read from the storage device 19 in the parking assistance mode. The parking assistance device 10 detects, by the object sensor 12, the relative position of the target objects around the host vehicle 1 with respect to the host vehicle 1 as the target positions of the target objects around the host vehicle 1. The method for detecting the target positions is the same as the detection method in the target object learning mode. The target positions detected by the object sensor 12 in the parking assistance mode are indicated by triangular plots. In the example of FIG. 3, the lane boundary lines 3a, the curbs 3c and 3e, and the corner portions of the wall 3g are detected as the target positions. In the parking assistance mode, the parking assistance device 10 compares each target position (triangular plot) detected by the object sensor 12 with the learned target position (circle plot) read out from the storage device 19 to associate identical target positions. For example, the parking assistance device 10 may determine that target positions having the same or similar feature values are the same target position. Alternatively, regardless of the feature value, the parking assistance device 10 may compare the relative positional relationship between the target positions (triangular plots) detected by the object sensor 12 with the relative positional relationship between the target positions (circle plots) of the learned target read out from the storage device 19 to associate the target positions (triangular plots) detected by the object sensor 12 with the target positions (circle plots) of the learned target read out from the storage device 19. Alternatively, the target positions (triangular plots) detected by the object sensor 12 may be associated with the target positions (circle plots) of the learned target read out from the storage device 19 using both the feature values and the relative positional relationship of the feature points described above. In the example of FIG. 3, lane boundary lines 3a, curbs 3c and 3e, and corners of a fence 3g are associated with each other.

[0019] The parking assistance device 10 calculates the relative position of the current position of the vehicle 1 relative to the target parking position 2 based on the relative positional relationship between the target position (triangular plot) detected during the parking assistance mode and the vehicle 1, and the relative positional relationship between the learned target position (circular plot) associated with these feature points (triangular plot) and the target parking position 2. For example, the parking assistance device 10 may calculate the position of the target parking position 2 on a relative coordinate system (hereinafter referred to as the "vehicle coordinate system") based on the current position of the vehicle 1. Alternatively, the parking assistance device 10 may calculate the current position of the vehicle 1 on a fixed coordinate system based on the relative positional relationship between each target position (triangle plot) detected in the parking assistance mode and the vehicle 1, and the learned target position (circle plot) on the fixed coordinate system. By determining the positions of the vehicle 1 and the target parking position 2 in the fixed coordinate system, the relative position of the current position of the vehicle 1 with respect to the target parking position 2 is determined. The parking support device 10 calculates a target travel trajectory from the current position of the host vehicle 1 to the target parking position 2 based on the relative position of the current position of the host vehicle 1 with respect to the target parking position 2. The parking support device 10 executes parking support control to support the movement of the host vehicle 1 along the calculated target travel trajectory. For example, it performs an automatic driving operation to control the host vehicle 1 to travel to the target parking position along the calculated target travel trajectory. Also, for example, it displays the target travel trajectory and the position of the host vehicle 1 on a display device visible to the user of the host vehicle 1.

[0020] However, when the learned target position is the target position of a movable object (for example, a bicycle placed around the target parking position 2, a potted plant planted in a flower pot, etc.), this object may move after the learned target position is memorized. In this case, the target position detected during parking support deviates from the learned target position stored in the storage device 19, or the target itself disappears, so that the target position is not detected during parking support. Therefore, if the relative position of the current position of the host vehicle 1 with respect to the target parking position 2 is calculated based on the learned target position including these, there is a risk that the calculation accuracy of the relative position will decrease.

[0021] Therefore, the parking support device 10 according to the embodiment detects the target positions of the targets around the host vehicle 1 when the host vehicle 1 has an opportunity to travel near the target parking position 2 after memorizing the learned target positions. The target positions detected when the host vehicle 1 has an opportunity to travel near the target parking position 2 after memorizing the learned target positions may be referred to as "surrounding target positions" in the following description. After the parking support device 10 stores the learned object positions, each time the host vehicle 1 has an opportunity to travel near the target parking position 2, the parking support device 10 counts the number of times the learned object positions stored in the storage device 19 match the surrounding object positions for each learned object position. Alternatively, after the parking support device 10 stores the learned object positions, each time the number of opportunities for the host vehicle 1 to travel near the target parking position 2 reaches a predetermined number (for example, every 2 times, every 3 times), the parking support device 10 may count the number of times the learned object positions stored in the storage device 19 match the surrounding object positions for each learned object position. Furthermore, after the parking support device 10 stores the learned object positions, each time the host vehicle 1 has an opportunity to travel near the target parking position 2 and in a predetermined specific scene (for example, each time when leaving the target parking position 2, or each time when entering the target parking position 2, etc.), the parking support device 10 may count the number of times the learned object positions stored in the storage device 19 match the surrounding object positions for each learned object position. That is, the parking support device 10 counts the number of times the learned object positions stored in the storage device 19 match the surrounding object positions for each learned object position at an arbitrary timing when the host vehicle 1 travels near the target parking position 2 after storing the learned object positions.

[0022] The parking support device 10 assigns a higher reliability to the learned object positions with a larger number of times of matching the surrounding object positions than to the learned object positions with a smaller number of times of matching the surrounding object positions. The parking support device 10 compares the learned object positions having a reliability equal to or higher than a predetermined reliability threshold among the learned object positions with the positions of the objects detected around the host vehicle 1, and calculates the relative position of the host vehicle 1 with respect to the target parking position 2. As a result, the positions of the fixed objects and the movable objects are stored as the learned object positions. When the movable object moves before the time of executing the parking support control, a higher reliability can be assigned to the learned object positions of the fixed objects (immovable objects or non-moving objects) than to the learned object positions of the movable object. As a result, only the learned object positions of the fixed objects among the learned object positions can be selected and used for calculating the relative position of the host vehicle 1 with respect to the target parking position 2, so that a decrease in the calculation accuracy of the relative position can be suppressed.

[0023] Refer to FIGS. 4(a) to 4( d ) to explain an example of a method for assigning reliability to learned object positions. First, in the object learning mode, when the parking support device 10 parks the host vehicle 1 at the target parking position 2, it detects the object positions of a plurality of objects around the host vehicle 1. The parking support device 10 stores the detected positions of the plurality of objects in the storage device 19 as learned object positions.

[0024] The round plots 30a to 30e in FIG. 4(a) indicate the learned object positions stored in the storage device 19. In the example of FIG. 4(a), the round plots 30a and 30b indicate the learned object positions of the curb, the round plot 30c indicates the learned object position of the lane boundary line, the round plot 30d indicates the learned object position of the house, and the round plot 30e indicates the learned object position of the movable object 3j (for example, a bicycle). Reliabilities are respectively assigned to the learned object positions 30a to 30e. Immediately after storing the learned object positions 30a to 30e, the reliabilities of the learned object positions 30a to 30e are all the same value (initial value). The numerical values "1" at the lower left of the learned object positions 30a to 30e in FIG. 4(a) indicate the reliabilities of the learned object positions 30a to 30e. The parking support device 10 stores the learned object positions and the reliabilities in association with each other in the storage device 19 as learned object data 20.

[0025] Refer to FIG. 4(b). After storing the learned object positions 30a to 30e, when the host vehicle 1 has an opportunity to travel near the target parking position 2, the parking support device 10 detects the object positions of the objects around the host vehicle 1 (that is, detects the surrounding object positions). The triangular plots 31a to 31e indicate the surrounding object positions. The triangular plots 31a and 31b indicate the surrounding object positions of the curb, the triangular plot 31c indicates the surrounding object position of the lane boundary line, the triangular plot 31d indicates the surrounding object position of the house, and the triangular plot 31e indicates the surrounding object position of the movable object 3j. The opportunity to detect the surrounding object positions 31a to 31e may be any opportunity as long as the host vehicle 1 travels near the target parking position 2 after memorizing the learned object positions 30a to 30e. In the example of Fig. 4(b), the surrounding object positions 31a to 31e are detected when the host vehicle 1 departs from the target parking position 2. The opportunity for the host vehicle 1 to depart from the target parking position 2 may be, for example, an opportunity for the driver to manually drive the host vehicle 1 to depart from the target parking position 2. Further, when the parking support device 10 has a function of departure support control, the opportunity to detect the surrounding object positions 31a to 31e may be during the execution of the departure support control. The departure support control may be, for example, control that calculates a target travel trajectory for the host vehicle 1 to depart from the target parking position 2, which is the current position of the host vehicle 1 during parking, based on the object positions detected around the host vehicle 1 and the learned object positions, and moves the host vehicle 1 along the target travel trajectory. Also, for example, the opportunity to detect the surrounding object positions 31a to 31e may be an opportunity for the parking support device 10 to park the host vehicle 1 at the target parking position 2 by parking support control after memorizing the learned object positions 30a to 30e.

[0026] Then, each time there is an opportunity for the host vehicle 1 to travel near the target parking position 2, the parking support device 10 counts the number of times the learned object positions 30a to 30e match the surrounding object positions 31a to 31e for each of the learned object positions 30a to 30e. For example, when the feature amounts of the learned object position and the surrounding object position are the same or similar, it may be determined that these object positions match. In the example of Fig. 4(b), the learned object positions 30a to 30d respectively match the surrounding object positions 31a to 31d. Therefore, the number of times the learned object positions 30a to 30d match the surrounding object positions is increased by one. On the other hand, since the movable object 3j has moved after the learned object position 30e is memorized, the learned object position 30e does not match the surrounding object position 31e. Therefore, the number of times the learned object position 30e matches the surrounding object position is not increased.

[0027] The parking assistance device 10 assigns a higher reliability to the learned object positions 30a to 30d that match the surrounding object positions more frequently than to the learned object position 30e that matches the surrounding object positions less frequently. For example, the parking assistance device 10 may assign reliability according to the number of times the learned object positions 30a to 30e match the surrounding object positions. The numerical values "2" at the lower left of the learned object positions 30a to 30d in Fig. 4(b) indicate the reliability of the learned object positions 30a to 30d, and the numerical value "1" at the lower left of the learned object position 30e indicates the reliability of the learned object position 30e. In this way, a higher reliability is assigned to the learned object positions 30a to 30d than to the learned object position 30e.

[0028] Referring to Fig. 4(c). When the parking assistance device 10 executes parking assistance control after storing the learned object positions 30a to 30e, it detects the object positions of the objects around the host vehicle 1 (i.e., the surrounding object positions). The rectangular plots 32a to 32e indicate the surrounding object positions detected when the parking assistance control is executed. The rectangular plots 32a and 32b indicate the surrounding object positions of the curb, the rectangular plot 32c indicates the surrounding object position of the lane boundary line, the rectangular plot 32d indicates the surrounding object position of the house, and the rectangular plot 32e indicates the surrounding object position of the movable object 3j.

[0029] The parking assistance device 10 selects only the learned object positions among the learned object positions 30a to 30e that have a reliability equal to or higher than a predetermined reliability threshold. Here, it is assumed that the reliability threshold is "2" for the sake of explanation. For this reason, only the learned object positions 30a to 30d of the fixed object are selected, and the learned object position 30e of the movable object 3j is not selected. The parking assistance device 10 calculates the relative position of the current position of the host vehicle 1 with respect to the target parking position 2 based on the selected learned object positions 30a to 30d and the surrounding object positions 32a to 32d corresponding to the learned object positions 30a to 30d. In this way, the parking assistance device 10 can select only the learned object positions 30a to 30d where fixed objects are detected and use them to calculate the relative position of the host vehicle 1 with respect to the target parking position 2, so that a decrease in the calculation accuracy of the relative position can be suppressed.

[0030] Refer to FIG. 4(d). After the parking assistance device 10 has completed the calculation of the relative position of the current position of the host vehicle 1 with respect to the target parking position 2 (for example, after the host vehicle 1 has moved to the target parking position 2 by parking assistance control), the number of times the learned object positions 30a to 30e match the surrounding object positions 32a to 32e is counted for each of the learned object positions 30a to 30e. In the example of FIG. 4(d), the learned object positions 30a to 30d respectively match the surrounding object positions 32a to 32d. For this reason, the number of times the learned object positions 30a to 30d match the surrounding object positions is increased by one. On the other hand, the movable object 3j has moved further after detecting the surrounding object position 31e, and the learned object position 30e does not match the surrounding object position 32e. For this reason, the number of times the learned object position 30e matches the surrounding object position is not increased.

[0031] The parking assistance device 10 updates the reliability according to the number of times the learned object positions 30a to 30e match the surrounding object positions 32a to 32e. As a result, the reliability of the learned object positions 30a to 30d has increased to "3". On the other hand, the reliability of the learned object position 30e remains "1". By repeating such detection of the surrounding object positions and updating of the reliability, a higher reliability can be given to the learned object positions of the fixed objects than to the learned object positions of the movable objects. For this reason, only the learned object positions of the fixed objects among the learned object positions can be selected and used to calculate the relative position of the host vehicle 1 with respect to the target parking position 2, so that a decrease in the calculation accuracy of the relative position can be suppressed.

[0032] Incidentally, after the parking support device 10 stores the learned object positions 30a to 30e, when the host vehicle 1 first travels near the target parking position 2, it may detect the surrounding object positions 31a to 31e and assign reliability levels to the learned object positions 30a to 30e. For example, when the host vehicle 1 first leaves the factory from the target parking position 2, reliability levels may be assigned to the learned object positions 30a to 30e. Thereby, after storing the learned object positions 30a to 30e, if a movable object 3j has moved between that time and when the host vehicle 1 first travels near the target parking position 2 thereafter, when parking at the first target parking position 2 after storing the learned object positions, the learned object position of the movable object 3j can be excluded to calculate the target travel trajectory.

[0033] Hereinafter, the functional configuration of the controller 17 will be described in more detail. Refer to FIG. 5. The controller 17 functions as an image conversion unit 40, a self-position calculation unit 41, a feature point detection unit 43, a learned object data generation unit 44, a reliability assignment unit 45, a relative position estimation unit 46, a target trajectory generation unit 47, a steering control unit 48, a vehicle speed control unit 49, and a support image generation unit 50. The image conversion unit 40 converts the captured image of the camera of the object sensor 12 into an aerial image (surround view monitor image) seen from a virtual viewpoint directly above the host vehicle 1 as shown in FIGS. 2 and 3. Hereinafter, the aerial image after conversion by the image conversion unit 40 may be referred to as the "surrounding image".

[0034] The self-position calculation unit 41 calculates the current position of the host vehicle 1 in the fixed coordinate system by dead reckoning or the like based on the vehicle information output from the vehicle sensor 13. The self-position calculation unit 41 may correct the calculated current position by map matching or the like between the object position detected by the object sensor 12, the known object position, and the high-precision map information. The feature point detection unit 43 detects the feature points of the targets around the host vehicle 1 from the surrounding image output from the image conversion unit 40, and calculates the feature amounts of the feature points. The feature point detection unit 43 outputs the detected feature points and their feature amounts, together with the current position of the host vehicle 1 received from the self-position calculation unit 41, to the learned target data generation unit 44 and the relative position estimation unit 46.

[0035] The learned target data generation unit 44 calculates the positions of the feature points on the fixed coordinate system based on the feature points output from the feature point detection unit 43 and the current position of the host vehicle 1. In the target learning mode, the learned target data generation unit 44 stores the positions of the feature points detected by the feature point detection unit 43 in the storage device 19 as the learned target positions. For example, the learned target data generation unit 44 sets the reliability of the learned target position to an initial value, associates the learned target position with the reliability, and stores them in the storage device 19 as the learned target data 20. On the other hand, when the operation mode of the parking support device 10 is not the target learning mode, the learned target data generation unit 44 outputs the feature point information including the positions and feature amount information of the feature points to the reliability assignment unit 45.

[0036] The reliability assignment unit 45 determines whether the host vehicle 1 is traveling near the target parking position 2. For example, the reliability assignment unit 45 may determine whether the host vehicle 1 is traveling near the target parking position 2 based on the current position of the host vehicle 1 calculated by the self-position calculation unit 41. For example, when the host vehicle 1 is being taken out of the target parking position 2, it may be determined that the host vehicle 1 is traveling near the target parking position 2. For example, when the distance between the current position of the host vehicle 1 and the target parking position 2 is within a predetermined value, and the host vehicle 1 is moving away from the target parking position 2 or the out-of-storage support function is in execution, it may be determined that the host vehicle 1 is being taken out of the target parking position 2. For example, when executing parking support control to support parking at the target parking position 2, it may be determined that the host vehicle 1 is traveling near the target parking position 2. For example, when the distance between the current position of the host vehicle 1 and the target parking position 2 is within a predetermined value and the operation mode of the parking support device 10 is the parking support mode, it may be determined that parking support control for supporting parking at the target parking position 2 is being executed.

[0037] When the host vehicle 1 is traveling near the target parking position 2, the reliability assignment unit 45 uses the position of the feature point output from the learned target data generation unit 44 as the surrounding target position. Each time the host vehicle 1 travels near the target parking position 2, the reliability assignment unit 45 counts, for each learned target position, the number of times the learned target position stored in the storage device 19 matches the surrounding target position, and updates the reliability of each learned target position according to the number of times the learned target position matches the surrounding target position.

[0038] Specifically, the reliability assignment unit 45 assigns reliability such that the reliability of the learned target position with a larger number of times of matching the surrounding target position is higher than the reliability of the learned target position with a smaller number of times of matching the surrounding target position. For example, a higher reliability may be assigned to a learned target position with a larger number of times of matching the surrounding target position. For example, the number of times the learned target position matches the surrounding target position may be assigned as the reliability.

[0039] The relative position estimation unit 46 reads out the learned target positions stored in the storage device 19 as the learned target data 20 and their reliabilities. Only the learned target positions having a reliability equal to or higher than a predetermined reliability threshold among the read learned target positions are selected. The predetermined reliability threshold may be, for example, a value obtained by subtracting a predetermined value from the highest reliability among the reliabilities of the learned target positions, may be the reliability of a predetermined rank from the highest value among the reliabilities of the learned target positions, or may be set based on the average value of the reliabilities of the learned target positions.

[0040] The relative position estimation unit 46 matches the selected learned target position with the target position detected during the parking assistance mode, and associates the target positions detected for the same object with each other. The relative position estimation unit 46 estimates the relative position of the current position of the vehicle 1 relative to the target parking position 2 based on the relative position relationship between the target positions detected during the parking assistance mode and the vehicle 1, and the relative position relationship between the learned target positions associated with these target positions and the target parking position 2.

[0041] For example, the target position detected during parking assistance mode is (x i ,y i ) and the target position (x i ,y i ) are the target positions of the learned targets associated with each of the mi ,y mi ) (i=1 to N). For example, the relative position estimation unit 46 calculates the affine transformation matrix M affine may be calculated.

number

[0042] Using the weighted least squares method, the column vector (a1, a2, a3, a4) is calculated as follows: T may be calculated.

number

[0043] The relative position estimation unit 46 calculates the position of the target parking position 2 (targetx) on the fixed coordinate system included in the parking learned target data 20 by the following equation: m ,targety m ) is converted into a position (targetx, targety) in the vehicle coordinate system.

number

[0044] The target trajectory generation unit 47 generates a target driving trajectory from the current position of the host vehicle 1 (i.e., the coordinate origin) in the vehicle coordinate system to the position (targetx, targety) of the target parking position 2 in the vehicle coordinate system. For calculating the target driving trajectory from the current position of the host vehicle 1 to the target parking position 2, well-known methods already generally adopted in automatic parking devices can be applied. As an example, for instance, it can be calculated by connecting the current position of the host vehicle 1 to the target parking position 2 with a clothoid curve. When the target driving trajectory includes a turning point, it can be calculated by connecting the current position of the host vehicle through the turning point to the target parking position 2 with a clothoid curve.

[0045] Further, the target trajectory generation unit 47 calculates a target vehicle speed profile in which the moving speed at each position on the target driving trajectory from the current position of the host vehicle to the target parking position 2 is set. For example, the target vehicle speed profile can be calculated based on a predetermined set speed, such that after accelerating from the current position of the host vehicle 1 to the set speed, the vehicle stops at the target parking position 2. When the target driving trajectory includes a turning point, a vehicle speed profile can be calculated such that the vehicle decelerates before the turning point and stops at the turning point, accelerates from the turning point to the set speed, decelerates before the target parking position 2, and stops at the target parking position 2. The set speed when calculating the speed profile may be set based on the curvature of the calculated target driving trajectory, such that the greater the curvature, the lower the speed.

[0046] The steering control unit 48 controls the steering actuator of the actuator 16 so that the host vehicle 1 travels along the target driving trajectory. Also, the vehicle speed control unit 49 controls the accelerator actuator and the brake actuator of the actuator 16 so that the vehicle speed of the host vehicle 1 changes according to the movement speed plan calculated by the target trajectory generation unit 47. Thereby, the host vehicle 1 is controlled to travel along the target driving trajectory. The support image generation unit 50 generates a parking support image representing the target driving trajectory calculated by the target trajectory generation unit 47 and the current position of the host vehicle 1. For example, the parking support image may be an image in which the target driving trajectory and the current position of the host vehicle 1 are superimposed on a bird's-eye view or an overhead view of the surroundings of the host vehicle 1 as seen from above. The support image generation unit 50 displays the parking support image on the display device of the human machine interface 15.

[0047] (Operation) FIG. 6 is a flowchart of an example of the process for storing the target position in the target learning mode. In step S1, while parking the host vehicle 1 manually at the target parking position 2, the feature point detection unit 43 detects the target position around the target parking position 2 from the surrounding image obtained by photographing the surroundings of the host vehicle. In step S2, the learned target data generation unit 44 stores the target position detected by the feature point detection unit 43 in the storage device 19 as the learned target position. At this time, the learned target data generation unit 44 associates the learned target position with the reliability and stores it in the storage device 19 as the learned target data 20. Then the process ends.

[0048] FIG. 7 is a flowchart of an example of the process for updating the reliability of the target position each time the host vehicle 1 has an opportunity to travel near the target parking position 2 after storing the learned target position. However, an example of the case where the host vehicle 1 parks at the target parking position 2 in the parking support mode among the opportunities for the host vehicle 1 to travel near the target parking position 2 is shown in FIG. 8. For example, the process in FIG. 7 is executed when the host vehicle 1 departs from the target parking position 2. In step S11, the feature point detection unit 43 detects the surrounding target position, which is the target position around the host vehicle 1, from the surrounding image obtained by photographing the surroundings of the host vehicle 1. In step S12, the reliability assignment unit 45 determines whether the learned object position stored in the storage device 19 matches the object position detected in step S11. For the learned object position that matches the object position detected in step S11, the reliability assignment unit 45 increments by one the number obtained by counting the number of times the learned object position matches the surrounding object position. For the learned object position that does not match the object position detected in step S11, the reliability assignment unit 45 does not increment the number obtained by counting the number of times the learned object position matches the surrounding object position. The reliability assignment unit 45 updates the reliability of each learned object position according to the number of times the learned object position matches the surrounding object position. Then the process ends.

[0049] Figure 8 is a flowchart of an example of the parking support control according to the embodiment. In step S21, the feature point detection unit 43 detects the surrounding object position, which is the object position around the host vehicle 1, from the surrounding image obtained by photographing the surroundings of the host vehicle. In step S22, the relative position estimation unit 46 estimates the relative position of the current position of the host vehicle 1 with respect to the target parking position 2 based on only the learned object positions having a reliability equal to or higher than a predetermined reliability threshold among the learned object positions stored in the storage device 19 as the learned object data 20 and the surrounding object position detected in step S21. In step S23, the target trajectory generation unit 47 generates a target travel trajectory for driving the host vehicle from the current position of the host vehicle 1 to the target parking position 2 and a target vehicle speed profile based on the relative position of the host vehicle 1 with respect to the target parking position 2.

[0050] In step S24, the steering control unit 48 controls the steering angle so that the host vehicle 1 travels along the target travel trajectory. The vehicle speed control unit 49 controls the moving speed of the host vehicle 1 according to the target vehicle speed profile. Thereby, the steering control unit 48 and the vehicle speed control unit 49 move the host vehicle 1 to the target parking position 2. In step S25, the reliability assignment unit 45 determines whether the learned target position stored in the storage device 19 matches the target position detected in step S21. For the learned target position that matches the target position detected in step S21, the reliability assignment unit 45 increments by one the number obtained by counting the number of times the learned target position matches the surrounding target position. For the learned target position that does not match the target position detected in step S21, the reliability assignment unit 45 does not increment the number obtained by counting the number of times the learned target position matches the surrounding target position. The reliability assignment unit 45 updates the reliability of each learned target position according to the number of times the learned target position matches the surrounding target position. Thereafter, the process ends.

[0051] (Effect of the embodiment) (1) When the controller 17 parks the host vehicle 1 at the target parking position 2, the controller 17 stores the positions of a plurality of targets detected around the target parking position 2 as learned target positions. After storing the learned target positions, when the host vehicle 1 travels near the target parking position 2, the controller 17 counts, for each of the plurality of learned target positions, the number of times the learned target position matches the surrounding target position, which is the position of the target detected around the host vehicle 1. The controller 17 assigns a higher reliability to the learned target position with a larger number of times of matching the surrounding target position than to the learned target position with a smaller number of times of matching the surrounding target position. The controller 17 compares the learned target positions with a reliability equal to or higher than a predetermined reliability threshold among the learned target positions with the positions of the targets detected around the host vehicle 1, calculates the relative position of the host vehicle 1 with respect to the target parking position 2, calculates the target travel trajectory from the current position of the host vehicle 1 to the target parking position 2 based on the calculated relative position, and executes parking support control to support the movement of the host vehicle 1 along the target travel trajectory.

[0052] As a result, the positions of the stationary object and the movable object are stored as the learned target positions. When the movable object moves before the parking assistance control is executed thereafter, a higher reliability can be given to the learned target position of the stationary object than to the learned target position of the movable object. Among the learned target positions, only the learned target position of the stationary object is selected and can be used to calculate the relative position of the host vehicle 1 with respect to the target parking position 2, so that a decrease in the calculation accuracy of the relative position can be suppressed.

[0053] (2) The opportunity for the host vehicle 1 to travel near the target parking position 2 after storing the learned target position may include the first opportunity for the host vehicle 1 to travel near the target parking position 2 after storing the learned target position. As a result, if the movable object has moved between storing the learned target position and then the first time the host vehicle 1 travels near the target parking position thereafter, the learned target position of the movable object can be excluded and the target travel trajectory can be calculated when parking at the first target parking position after storing the learned target position. (3) Also, the opportunity for the host vehicle 1 to travel near the target parking position 2 after storing the learned target position may include the time of executing the parking assistance control for assisting parking at the target parking position 2. Thereby, the reliability of the learned target position can be updated each time the parking assistance control is executed. (4) The opportunity for the host vehicle 1 to travel near the target parking position 2 after storing the learned target position may include the opportunity to take the host vehicle 1 out of the target parking position 2. Thereby, the reliability of the learned target position can be updated each time the host vehicle 1 is taken out of the target parking position 2.

[0054] (5) The predetermined reliability threshold value may be a value obtained by subtracting a predetermined value from the highest reliability among the reliabilities of the plurality of learned target positions, may be the reliability of a predetermined rank from the highest value among the reliabilities of the plurality of learned target positions, or may be set based on the average value of the reliabilities of the plurality of learned target positions. Thereby, the reliability threshold value can be set appropriately. ( 6The parking support control is, for example, control for moving the host vehicle 1 along a target travel trajectory from the current position of the host vehicle 1 to the target parking position 2. Thereby, parking support for controlling the host vehicle so that the host vehicle 1 travels to the target parking position 2 can be realized. ( 7 )The parking support control may be control for displaying the target travel trajectory and the position of the host vehicle 1 on a display device visible to the user of the host vehicle. Thereby, the occupant can visually recognize the target travel trajectory for traveling the host vehicle 1 to the target parking position 2.

[0055] All examples and conditional terms described herein are intended for educational purposes to assist the reader in understanding the present invention and the concepts provided by the inventors for the advancement of technology, and are to be construed without being limited to the above-described examples and conditions specifically described, and the configurations of the examples in this specification regarding demonstrating the superiority and inferiority of the present invention. Although the embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions, and modifications can be added thereto without departing from the spirit and scope of the present invention.

Explanation of Reference Numerals

[0056] 40…Image conversion unit, 41…Self-position calculation unit, 43…Feature point detection unit, 44…Learned object target data generation unit, 45…Reliability assignment unit, 46…Relative position estimation unit, 47…Target trajectory generation unit, 48…Steering control unit, 49…Vehicle speed control unit, 50…Support image generation unit

Claims

1. When the controller parks the host vehicle at a target parking position, it stores the positions of a plurality of targets detected around the target parking position as learned target positions, after storing the learned target positions, when the host vehicle travels near the target parking position, it counts the number of times the learned target positions match the surrounding target positions, which are the positions of targets detected around the host vehicle, for each of the plurality of learned target positions, assigns a higher reliability to the learned target positions where the number of matches with the surrounding target positions is greater than that of the learned target positions where the number of matches is less, compares the learned target positions among the learned target positions with a reliability equal to or higher than a predetermined reliability threshold value and the positions of targets detected around the host vehicle, and calculates the relative position of the host vehicle with respect to the target parking position, calculates a target travel trajectory from the current position of the host vehicle to the target parking position based on the calculated relative position, and executes parking support control to assist the movement of the host vehicle along the calculated target travel trajectory. A parking support method characterized by the above.

2. The opportunity for the host vehicle to travel near the target parking position after storing the learned target positions includes the first opportunity for the host vehicle to travel near the target parking position after storing the learned target positions. The parking support method according to Claim 1.

3. The opportunity for the host vehicle to travel near the target parking position after storing the learned target positions includes the execution time of the parking support control for assisting parking at the target parking position. The parking support method according to Claim 1 or 2.

4. The opportunity for the host vehicle to travel near the target parking position after storing the learned target positions includes the opportunity for the host vehicle to leave the target parking position. The parking support method according to any one of Claims 1 to 3.

5. The controller sets the predetermined reliability threshold value to a value obtained by subtracting a predetermined value from the highest reliability among the reliabilities of the plurality of learned target positions. The parking support method according to any one of Claims 1 to 4.

6. The controller sets the predetermined reliability threshold value based on the average value of the reliabilities of the plurality of learned target positions. The parking support method according to any one of Claims 1 to 4. ​ ​ ​ ​

7. The parking assistance method according to any one of claims 1 to 4, characterized in that the controller sets the predetermined reliability threshold to the reliability of a predetermined rank from the highest value among the reliabilities of the plurality of learned object positions.

8. The parking assistance control is a control for moving the host vehicle along the target travel trajectory from the current position of the host vehicle to the target parking position. The parking assistance method according to any one of claims 1 to 7.

9. The parking assistance control is a control for displaying the target travel trajectory and the position of the host vehicle on a display device visible to the user of the host vehicle. The parking assistance method according to any one of claims 1 to 8.

10. A sensor that detects an object position, which is the position of an object around the host vehicle, When parking the host vehicle at a target parking position, the positions of a plurality of objects detected around the target parking position are stored as learned object positions in a predetermined storage device. After storing the learned object positions, when the host vehicle travels near the target parking position, the number of times the learned object positions match the surrounding object positions, which are the positions of the objects detected around the host vehicle, is counted for each of the plurality of learned object positions. A higher reliability is assigned to the learned object positions with a larger number of matches with the surrounding object positions compared to the learned object positions with a smaller number of matches with the surrounding object positions. The learned object positions with a predetermined reliability threshold or higher among the learned object positions are compared with the object positions detected around the host vehicle to calculate the relative position of the host vehicle with respect to the target parking position. Based on the calculated relative position, a target travel trajectory from the current position of the host vehicle to the target parking position is calculated, and a controller that executes parking assistance control to assist the movement of the host vehicle along the calculated target travel trajectory. A parking assistance device, characterized by comprising the above.

Images