JP2018079929A - Drive support apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance accuracy of drive support for parking a vehicle.SOLUTION: A travel control controller extracts, from a pickup image picked by a camera, a symbol including a numeral positioned on a road surface, the numeral presumably disposed in a parking-spot frame, the numeral of size presumably attached to the parking-spot frame. The travel control controller extracts, upon extracting a plurality of symbols, a symbol group consisting of a plurality of symbols that is arranged in parallel with an interval of the adjacent symbols having predetermined distance, and directions of the respective symbols are oriented in a direction across a direction of arrangement of the symbols and in the same direction for each other. The travel control controller estimates a position of a parking-spot frame from positions of the individual symbols of the symbol group and controlling decrease in acceleration in accordance with an acceleration operator on the basis of the estimated parking-spot frame in a case where a set of numerals of adjacent symbols in the extracted symbol group are determined to be consecutive numbers.SELECTED DRAWING: Figure 5

Description

  The present invention relates to a driving support technology that supports driving of a vehicle for parking.

  In the prior art described in Patent Document 1, when it is determined that the host vehicle is in a parking lot off the road with reference to the map information, even if the accelerator is depressed, it is determined that it is a mistake in stepping on the accelerator. The acceleration of the host vehicle is suppressed by throttle control.

JP 2003-137001 A

However, when it is determined that the host vehicle is in the parking lot, even if the accelerator is depressed, it is determined that the step is wrong and the acceleration of the host vehicle is suppressed by throttle control. Even when the vehicle is moving with respect to the parking space where the vehicle is parked, that is, when the driver does not make a stepping error, the acceleration of the vehicle may be suppressed, and the driver feels uncomfortable. May give.
An object of the present invention is to provide appropriate driving support for a stepping error by appropriately determining whether or not the host vehicle makes a mistake in stepping on an accelerator.

  In order to solve the above problem, according to one embodiment of the present invention, a symbol including a number is extracted from a captured image including a road surface around the host vehicle. When a plurality of the above symbols are extracted and it is determined that the numbers of these symbols are sequential numbers, it is estimated from the symbols that there is a parking frame at a predetermined position. Furthermore, when it is determined that the estimated parking frame exists ahead of the traveling direction of the host vehicle, acceleration generated in the host vehicle is reduced according to the acceleration operator.

  According to the present invention, when the vehicle is moving toward the parking frame, if there is a high possibility of a stepping error, the acceleration of the host vehicle is reduced, thereby reducing the possibility of giving the driver a sense of discomfort. be able to.

It is a conceptual diagram which shows the structure of the vehicle which concerns on embodiment based on this invention. It is a figure for demonstrating the structure of the traveling control controller which concerns on embodiment based on this invention. It is a perspective view which shows the example of an imaging with a camera. It is a top view which shows the example of the imaging region around a vehicle. It is a figure which shows the structure of an ambient environment recognition information calculating part. It is a figure which shows the process of a parking frame line information processing part. It is a conceptual top view which shows the part to which overhead conversion is carried out among the picked-up images imaged with the camera. It is a conceptual top view which shows the state which converted the image of FIG. 7 into the bird's-eye view image. It is a figure which shows the continuous frame position of the bird's-eye view image to acquire. It is a figure explaining the detection of a radial line. It is a figure explaining a radial line. It is a figure explaining the determination of a radial line. It is a figure explaining the determination of the line of a solid object. It is a figure explaining the example of a process which detects the line corresponding to a level crossing related display line. It is a figure explaining the process of a symbol extraction part. It is a figure explaining the process of a symbol group extraction part. It is a figure explaining the process of a serial number determination part. It is a figure for demonstrating the process of an acceleration suppression operation condition judgment part. It is a figure explaining the distance of the own vehicle and a parking frame, and the own vehicle and a parking frame. It is a figure explaining the process of the acceleration suppression amount calculating part. It is a figure which shows the specific example of a process of step S630. It is a figure which shows the example of the 2nd acceleration suppression amount. It is a figure explaining the example of the 1st acceleration suppression amount. It is a figure explaining the process of the target throttle opening calculating part. It is a figure which shows the example of a time chart in 1st Embodiment. It is a figure which shows the example of a movement between frames in the bird's-eye view image of the line of a solid object. It is a schematic diagram which shows the level crossing related display line currently displayed on a level crossing. It is a figure which shows the illustration of the line corresponding to the crossing related display line detected from the bird's-eye view image. It is a schematic diagram of another example showing a crossing related display line displayed on a crossing. It is a figure which shows the example of the parking frame detected in this embodiment. It is a schematic diagram which shows the example of the parking frame estimated and extracted from a number sequence. It is a figure which shows the transition of the acceleration suppression amount according to the accelerator operation amount. It is a figure explaining how to obtain | require comprehensive reliability TLVL. It is a figure explaining the parking assistance control using the comprehensive reliability TLVL of FIG. It is a figure explaining the process of step S586 which concerns on 2nd Embodiment based on this invention. It is a figure explaining the own vehicle anticipation track frame line overlap amount. It is a figure explaining the own vehicle estimated track parking frame entrance duplication rate. It is a figure which shows the example of a time chart in 2nd Embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
“First Embodiment”
(Constitution)
The vehicle includes a braking device that generates a braking force and a driving device that generates a driving force.
As shown in FIG. 1, the braking device includes a brake device 12 provided on the wheel 11, a fluid pressure circuit 13 including a pipe connected to each brake device 12, and a brake controller 14. The brake controller 14 controls the braking force generated by each brake device 12 via the fluid pressure circuit 13 to a value corresponding to the braking force command value. The brake device 12 is not limited to a device that applies a braking force with fluid pressure, and may be an electric brake device or the like.

As shown in FIG. 1, the drive device includes an engine 15 as a drive source and an engine controller 16 that controls torque (drive force) generated by the engine 15. The drive source of the drive device is not limited to the engine 15 and may be an electric motor or a hybrid configuration in which the engine 15 and the motor are combined.
Each of the brake controller 14 and the engine controller 16 is configured to receive each command value of a braking command and a drive command (acceleration command value) from the travel control controller 10 that is a host controller. The brake controller 14 and the engine controller 16 constitute an acceleration / deceleration control device.

As shown in FIGS. 1 and 2, the vehicle includes an ambient environment recognition sensor 1, a wheel speed sensor 2, a steering angle sensor 3, a shift position sensor 4, a brake operation detection sensor 5, and an accelerator operation detection sensor. 6, a navigation device 7, and a wiper detection sensor 8. The vehicle also includes a travel controller 10.
The surrounding environment recognition sensor 1 recognizes obstacles and road surfaces around the host vehicle MM and outputs the recognized surrounding state to the travel controller 10. The ambient environment recognition sensor 1 of the present embodiment is composed of one or two or more cameras capable of capturing an image around the vehicle. The camera 1 is provided, for example, at the position of the side mirror, the front part, the rear part, the roof part, or the like of the vehicle. Each camera 1 captures a road surface around the vehicle and acquires a captured image at preset imaging interval times.

  In the present embodiment, as shown in FIG. 3, cameras as the surrounding environment recognition sensors 1 are respectively arranged at four locations, front, rear, left, and right. Then, as shown in FIG. 4 which is a plan view, the area around the vehicle is divided into four parts ARA1 to ARA4, and each area ARA1 to ARA4 is imaged by each camera 1. There may be overlapping portions in the imaging area of each camera 1. Moreover, you may provide the separate camera for imaging a distant vehicle (for example, to 100 m). Note that the captured image does not have to be an image of the entire area captured by the camera 1, and may be an image cut out from a video captured by the camera 1.

  Here, in this embodiment, as an example, an area ahead of the host vehicle MM in the traveling direction is an area for acquiring information for detecting a parking frame. In the following description, a case where a parking frame is detected based on a captured image obtained by capturing an area ARA1 in front of the host vehicle will be described as an example. When entering the parking frame by reversing the vehicle, the parking frame may be detected using the captured image of the area ARA2.

The wheel speed sensor 2 detects the wheel speed and outputs the detected wheel speed information to the travel controller 10. The wheel speed sensor 2 is constituted by a pulse generator such as a rotary encoder that measures wheel speed pulses, for example.
The steering angle sensor 3 detects the steering angle of the steering wheel 20 and outputs the detected steering angle information to the travel controller 10. The steering angle sensor 3 is provided on a steering shaft or the like. The steered wheel turning angle may be detected as steering angle information.

The shift position sensor 4 detects shift information of a shift position (drive instruction position, parking instruction position, neutral position, etc.), and outputs a detection signal to the travel control controller 10.
The brake operation detection sensor 5 detects whether or not the brake pedal 18 is operated and the operation amount. The detected brake pedal operation amount is output to the travel controller 10. The brake pedal 18 is an operation element for decelerating instructions operated by the driver.

The accelerator operation detection sensor 6 detects the operation amount of the accelerator pedal 19. The detected accelerator pedal operation amount is output to the travel controller 10. The accelerator pedal 19 is an acceleration operator for an acceleration instruction operated by the driver. The navigation device 7 includes a GPS receiver, a map database, a display monitor, and the like, and is a device that performs route search, route guidance, and the like. . The navigation device 7 can acquire information such as the type of road on which the host vehicle MM travels and the road width based on the current position of the host vehicle MM obtained through the GPS receiver and the road information stored in the map database. it can.

The wiper detection sensor 8 detects the operation of the wiper. The detected wiper operation information is output to the travel controller 10.
The information presenting device outputs an alarm or other presenting by voice or image in accordance with a control signal from the travel controller 10. The information presentation device includes, for example, a speaker that provides information to the driver by a buzzer sound or voice, and a display unit that provides information by displaying an image or text. As the display unit, for example, the display monitor of the navigation device 7 may be used.

  The travel control controller 10 is an electronic control unit including a CPU and CPU peripheral components such as a ROM and a RAM. The travel control controller 10 includes a parking driving support unit that performs driving support processing for parking. As shown in FIG. 2, the parking driving support unit among the processes of the travel controller 10 functionally includes an ambient environment recognition information calculation unit 10A, a vehicle speed calculation unit 10B, a steering angle calculation unit 10C, and a steering angular velocity calculation. 10D, shift position calculation unit 10E, brake pedal operation information calculation unit 10F, accelerator operation amount calculation unit 10G, accelerator operation speed calculation unit 10H, acceleration suppression operation condition determination unit 10I, acceleration suppression amount calculation unit 10J, and target throttle opening The processing of the degree calculation unit 10K is provided. These functions are composed of one or more programs.

The surrounding environment recognition information calculation unit 10 </ b> A recognizes the road surface environment around the vehicle based on the captured image captured by the surrounding environment recognition sensor 1. As shown in FIG. 5, the surrounding environment recognition information calculation unit 10 </ b> A includes a parking frame line information processing unit 110 and a parking symbol information processing unit 120.
The own vehicle vehicle speed calculation unit 10 </ b> B calculates the vehicle speed based on a signal from the wheel speed sensor 2.
The steering angle calculation unit 10 </ b> C calculates a steering angle based on a signal from the steering angle sensor 3.

The steering angular velocity calculation unit 10D calculates the steering angular velocity by differentiating the signal from the operation angle sensor.
The shift position calculation unit 10E determines the shift position based on the signal from the shift position sensor 4.
The brake pedal operation information calculation unit 10 </ b> F determines a brake operation amount based on a signal from the brake operation detection sensor 5.

The accelerator operation amount calculation unit 10G calculates the operation amount of the accelerator pedal 19 based on a signal from the accelerator operation detection sensor 6.
The accelerator operation speed calculation unit 10H calculates the operation speed of the accelerator pedal 19 by differentiating the signal from the accelerator operation detection sensor 6.
The acceleration suppression operation condition determination unit 10I determines the acceleration suppression operation condition for the braking / driving force control of the vehicle based on the road surface environment information from the surrounding environment recognition information calculation unit 10A.

Next, the process of the parking frame line information processing unit 110 will be described with reference to FIG. The parking frame line information processing part 110 performs the process shown in FIG. 6 for every preset sampling time.
The parking frame line information processing unit 110 acquires a captured image captured by the surrounding environment recognition sensor 1 in step S10. In the present embodiment, as shown in FIG. 7, a captured image obtained by capturing an area ARA1 in the traveling direction of the host vehicle is used.

Next, in step S20, the captured image acquired in step S10 is overhead converted to acquire an overhead image.
In addition, the acquisition of the bird's-eye view image is performed by, for example, cutting out a portion of the image that is a preset bird's-eye view area around the vehicle (see FIG. 7) from the acquired captured image, and performing bird's-eye conversion on the cut-out image. A bird's-eye view image as shown in FIG. In the captured image, an object in a distant area appears smaller, so even a parallel line appears as a non-parallel line as shown in FIG. By performing bird's-eye conversion, this is detected as parallel lines on the bird's-eye view image as shown in FIG. In the overhead view conversion, for each captured image captured by each camera, an overhead view image may be obtained by performing an overhead conversion of an image of the overhead view area portion that the camera is responsible for.

Here, the bird's-eye view image is an image of the road surface environment viewed from a virtual camera assumed to be installed at a position looking down from directly above. The overhead conversion process may employ a known conversion method such as geometric conversion. In the overhead view conversion, image coordinate conversion is performed in a direction in which the viewpoint of the image is directed from the top to the bottom.
Next, in step S30, a line existing on the overhead image is detected. Specifically, line extraction is performed for the overhead image portion of the road surface in the direction along the traveling direction of the host vehicle (hereinafter, also referred to as a road surface overhead image) in the overhead image acquired in step S20. Perform image processing. The image processing detects a line existing on the road surface overhead image by performing a known line detection process such as edge processing on the road surface overhead image. In this embodiment, the case where the front of the vehicle is the traveling direction of the vehicle is illustrated.

  Here, as shown in FIG. 9, the road surface ahead in the vehicle traveling direction is sequentially displayed on the road surface overhead image sequentially acquired as the host vehicle moves. Of the lines in the road surface overhead image that are sequentially acquired, tracking is performed for lines extending outside the window area of the image (lines estimated to extend in the vehicle traveling direction). That is, a line matching process between different road surface overhead images is performed, and it is determined whether or not the lines between sequentially acquired road surface overhead images are the same line.

Next, in step S40, an attribute assignment process is performed on the detected line.
The attribute assignment process will be described next.
For each detected line, it is determined whether it corresponds to the following attribute, and if applicable, the attribute is assigned to the detected line. An assignment determination is performed for each attribute.
An example of the attribute to be given is shown below.

・ Radial lines (lines extending radially from the light receiving part 1a of the camera)
-Solid object line (line indicating a solid object)
-Luminance symmetry (luminance symmetry at both ends in the width direction of the detected line)
・ Luminance difference with the road surface ・ Left and right luminance difference (Symmetry of luminance on the outside of both ends in the width direction of the detected line)
・ Fixed object (long) (Generation of end points where lines are broken due to dirt: long breaks)
・ Fixed object (short) (Generation of end points where lines are broken due to dirt: short breaks)
・ Fixed object (lens) (edge generated by adhering to the range)
Here, attribute processing of a radial line (also referred to as a radial line) and a three-dimensional object line will be described as processing of the attribute addition processing unit.

First, radial line attribute assignment processing will be described with reference to FIG.
Here, when it is determined that the bird's-eye view screen processed immediately before is the same as the bird's-eye view screen processed this time, the next radial line attribute assignment processing may not be performed. The case where the previous bird's-eye view screen and the current bird's-eye view screen are the same is a case where it can be considered that the real vehicle is stopped when the vehicle is stopped. That is, for example, it is possible to determine whether the bird's-eye view screen processed immediately before and the bird's-eye view screen processed this time are the same based on whether or not the vehicle movement is estimated to be at a substantially stopped speed.

In the process of assigning the attribute of the radial line, first, in the overhead view image, among the lines on the current road surface overhead view image, as shown in FIG. 11, the imaging center 1a (light receiving unit 1a) of the camera 1 that captured the image is displayed. A radial line L3 as the center is detected (step S41a).
At this time, in consideration of an imaging error, whether or not the line is a radial line is determined as follows. That is, as shown in FIG. 12, when the difference between the upper end point of the line on the road surface overhead image and the straight line connecting the imaging center is within the preset radial line determination threshold angle θ, the line is determined as the radial line L3. . An imaging error is caused by a change in vehicle behavior (pitching or the like).

Here, when the target line is a line detected for the first time, an on-threshold angle (for example, ± 5 degrees) is adopted as the radial line determination threshold angle θ, and in the case of a line that has already been detected, An off-threshold angle (for example, ± 7 degrees) wider than the on-threshold angle is employed as the radial line determination threshold angle θ.
Further, it is determined whether or not the road surface around the host vehicle is likely to be reflected, and in the case of a road surface state in which reflection is likely to occur, the radial line determination threshold angle θ is set to an angle larger than the initial value angle as described above. use. Alternatively, a small value is used as the threshold value of the count n in the continuous collation determination (step S65 described later). For example, an ON threshold angle (for example, ± 7 degrees) and an OFF threshold angle (for example, ± 8 degrees) are set.

Examples of the road surface state in which reflection is likely to occur include the following.
When the wiper is activated due to rain, snow, etc. (for example, it can be determined by a signal from the wiper detection sensor 8).
・ Μ is low (for example, it can be judged by slip amount).
Sunlight is incident on the lens (for example, it can be determined by the brightness of the image).

Next, it is determined whether the line L3 determined to be a radial line is a line detected for the first time in the current road surface overhead image (step SS41b). If the line L3 is detected for the first time this time, the line L3 has a radial line attribute FF. The attribute (1) is assigned (step S41c). Here, n of the radial line attribute FF (n) is counted up each time it is determined as a radial line.
On the other hand, if it is determined that the line L3 determined to be a radial line is also detected in the previous road surface overhead image, it is determined whether the radial line attribute FF (n) is assigned to the line L3 (step S41d). If it is determined that the value is assigned, the counter value n of the radial line attribute FF (n) is counted up. That is, the attribute of the radial line FF (n + 1) is given to the line (step S41e).

  Here, the processing of steps S41b to S41e shown in FIG. 10 may be performed in step S50 described later. In this case, in step S40, if the line is determined to be a radial line, the attribute of the radial line is always given. In step S50, it is determined whether the line has been determined to be a non-radial line in the past. When it is determined that the line is a radial line, the parking frame determination is performed by ignoring the attribute assignment information of the radial line. Alternatively, the attribute of the radial line is set to OFF. In step S55, it is determined whether or not the same line has been continuously determined as a radial line for a preset duration or longer.

In addition, for a line L3 that has been continuously determined as an attribute of the radial line (n), a difference between the positions of the line L3 in the two road surface overhead images in which the line L3 is reflected (a specific line shape) If it is determined that the line is moving as an edge of the solid object along the vehicle movement based on the difference in position) and the vehicle movement information, the attribute FR of the solid object line is also given to the line.
In addition, with respect to a line determined to be an attribute of the radial line (n) continuously, a difference between the positions of the line in the two road surface overhead images in which the line is reflected (at a specific linear position) If it is determined that the line is moving as a display on the road surface along the vehicle movement based on the difference) and the vehicle movement information, a process that does not regard the line as a radial line may be added.

Next, a solid object line attribute assigning process will be described with reference to FIG.
Based on the road surface bird's-eye view image acquired continuously, with regard to a line where at least the upper end or the lower end can be detected among the both ends of the line, it is selected as a point for tracking two points on the line specified with the detected end as a reference Then (step S42a), the respective movement amounts of the two points accompanying the movement of the vehicle are calculated (step S42b). The amount of movement may be the amount of displacement of the two points between the two overhead images acquired with the movement of the host vehicle. Then, it is determined that the relationship between the movement amounts of the two points is movement along the three-dimensional object accompanying the movement of the vehicle (step S42c), and when the determination condition is satisfied, the attribute of the three-dimensional object line is determined with respect to the line. FR = 1 is assigned (step S42d). The three-dimensional object is a stationary object such as a stationary vehicle or a wall. When the relationship between the movement amounts of two points to be tracked is, for example, a movement amount of a point closer to the host vehicle (lower point) is smaller than a movement amount of a relatively far point (upper point), a three-dimensional object The line is determined. The line of the three-dimensional object is usually a line of the edge portion of the three-dimensional object.

  Here, when a line to which the attribute of the solid object line is added is detected, the solid object line between the line to which the attribute FR of the solid object line is added and the light receiving unit 1a of the camera in the road surface overhead image is displayed. A line to which the attribute FR is not added is detected, and among the detected lines, there is a line whose distance from the line to which the attribute FR of the three-dimensional object is added is within a preset separation distance (for example, 1 m). If it is, the priority frame candidate attribute PRR is assigned to the line. It is determined whether or not there are a plurality of three-dimensional object lines within a preset distance, the position of the three-dimensional object is estimated from the plurality of three-dimensional object lines, and the estimated position of the three-dimensional object at the vehicle side is estimated. The priority frame candidate attribute PRR may be assigned to a line whose distance from the three-dimensional object is within a preset distance. This process may be performed in step S40, step S50, or the like.

Next, in step S50 shown in FIG. 6, as a pre-process for parking frame recognition, a line to which an attribute of a radial line FF (n) (n: for example, n = 3 or more) or a three-dimensional line attribute FR is given. Are excluded from the frame line candidates.
Next, in step S52, line interpolation processing is performed. This is a process of interpolating the blurring of the line. That is, interpolation processing is performed between a plurality of lines detected along the same virtual straight line, that is, lines that may be the same line.

The line interpolation processing will be described with reference to FIG.
First, it is determined whether the length of the space between two adjacent lines along the same virtual line (also referred to as a break length) is equal to or less than a preset interpolation length (step S110). If the length is less than the interpolation length, a process of interpolating the line and regarding it as one line of the two lines is performed (step S120). The interpolation length is set to a value shorter than the track width at the railroad crossing (including a margin for wheels passing through the track). The interpolation length is set to 20 cm, for example.

Further, the following process is performed between the lines where the break length exceeds the interpolation length.
First, it is determined whether the interruption length exceeds a preset maximum interpolation length (for example, 1 m) (step S130). If the maximum interpolation length is exceeded, no interpolation processing is performed. That is, it is processed as another line.
On the other hand, whether or not the following interpolation processing is appropriate is determined between the lines whose break length is longer than 20 cm and within the maximum interpolation length.

That is, it is determined whether at least one of the two lines constituting the blank is a line having a parking line level FLVL of 1 or more (step S140). If the parking line level FLVL is not 1 or more, no interpolation is performed.
Next, when a line having a parking line level FLVL of 1 or more, that is, a pair of frame line candidates is detected, a line on the side where both ends of the two lines constituting the blank are known ( Compared with the paired line that makes a pair with the line. The examination line is usually a line detected earlier.

  Then, the lengths of the study line and the pair line are approximated (step S150), and the discontinuity length between adjacent lines in the longitudinal direction on the pair line side approximates the discontinuity length on the study line side. If so (step S160), both lines are interpolated to perform processing that regards the two lines as one line, and information on non-frame line candidates is added to the line (step S170). . In the case where the lengths of the study line and the pair line can be regarded as substantially the same as the approximate length, for example, when the difference in length between the study line and the pair line is equal to or less than a preset threshold value, the approximation is determined. The case where the break length is approximate is a case where the break length can be regarded as substantially the same length. For example, when the difference between the break lengths is equal to or less than a preset threshold value, the approximation is determined.

Next, in step S55 in FIG. 6, a process is performed to determine whether or not the line is a frame line candidate line with respect to the line detected in the process of continuously acquiring the road surface overhead image.
Here, the determination is made based on whether or not a preset parking frame condition as shown below is satisfied. The process for determining whether or not the frame is a candidate for a frame line is performed, for example, on a line located in a parking frame presence determination area (for example, an area within a radius of 10 m centered on the own vehicle) set for the own vehicle.

  That is, in step S55, when all the following parking frame line conditions are satisfied, it is determined as a parking frame line candidate line. If it is determined that the line is a parking frame candidate line, “1 (n)” is set in the parking line level FLVL as the attribute information of the line in step S60. The initial value of the parking line level FLVL is “0”. Further, n of LVL1 (n) is n = 1 when it is first determined as a parking frame line candidate, and is incremented every time it is determined as a parking frame line candidate. Conversely, the line once determined to have the parking line level FLVL of “1 (n)” is counted down every time it is determined that the parking frame line condition is not satisfied.

"Parking frame conditions"
-The line is a line estimated to be a straight line.
The line width is a preset line range (a range of line thicknesses considered as parking frame lines, for example, 2 cm to 4 cm).
-There is a pair of lines in a preset separation range (a range that can be regarded as a pair of parking frames, for example, 1.5 m to 2.5 m).

The parallelism between the paired lines is within a preset allowable angle (for example, within 4 degrees).
When the corresponding end portions of the paired lines are detected, the shift amount in the extending direction of the lines at both end points is equal to or less than a preset shift amount (for example, 50 cm).
The difference in the line width dimension between the paired lines is not more than a preset value (for example, 8 mm).

-The length of the line is not more than a preset maximum length (for example, 9 m).
A line that does not have information on non-frame line candidates.
Here, the line detected as the line of the parking frame is a line that is estimated to be a line in the direction along the front-rear direction of the vehicle when the vehicle is parked (a line that exists on the side of the vehicle when parked). Since the line width is confirmed, the shape of the end of the line can be detected.

Here, when the target line has the priority frame candidate attribute PRR, it is determined whether the following parking frame line condition is relaxed or not. The relaxation is determined, for example, by widening the value set below. For example, the determination is made with the allowable angle of parallelism being within 6 degrees, for example.
Next, in step S65, it is determined whether or not the variable n is greater than or equal to a preset threshold value (for example, n = 3 or more at a sampling period of 100 msec) for a line having a parking line level FLVL of “1 (n)”. If the line satisfies the condition, the parking line level FLVL is set to 2 for a line equal to or greater than a preset threshold value in step S70. For example, the preset threshold value is set to a value that can be detected by a predetermined length (for example, 2 m) as the length of the line. Here, instead of determining based on the size of the counter n, it is determined whether or not the length of the line is detected to a predetermined length or more, and the length of the line is equal to or longer than a predetermined length (for example, 2 m). If the parking line level FLVL is “1 (n)” and the parking line level FLVL is estimated to have been detected, the parking line level FLVL may be changed to 2. Alternatively, when it is determined that the vehicle travel distance is the same straight line during the preset travel distance, the line with the parking line level FLVL of “1 (n)” is changed to the parking line level FLVL2. Also good.

When the length of the line is equal to or greater than a preset maximum length (for example, 9 m), the parking line level FLVL is forcibly changed to “0” regardless of the level of the parking line level FLVL.
Next, in step S75, when one of the both ends of the line having the parking line level FLVL of 2 is detected, the process proceeds to step S80, and the parking line level FLVL is changed to 3 in the step S80. To do.

Here, the detection of the end may be recognized as the end only when the end shape is a specific shape set in advance. Examples of specific shapes are simple line ends, U-shaped, T-shaped end shapes, and the like.
Next, in step S85, when both ends of the line having the parking line level FLVL of 3 are detected, the process proceeds to step S90, and the parking line level FLVL is changed to 4 in the step S90. .

As shown in FIG. 5, the parking symbol information processing unit 120 includes a symbol extraction unit 120A, a symbol group extraction unit 120B, a serial number determination unit 120C, and a parking frame estimation unit 120D.
Next, the processing of the symbol extraction unit 120A will be described with reference to FIG.
In step S210, the symbol extraction unit 120A periodically inputs the bird's-eye view image acquired by the above-described parking frame line information processing unit 110 through the processes in steps S10 and S20.

Next, in step S220, symbols in the area estimated to be on the road surface in the input overhead image are extracted.
Next, in step S230, only symbols having a size smaller than a preset size are extracted from the symbols extracted in step S220. The predetermined size or less is usually equal to or less than the maximum size of a symbol written on the parking frame, for example, 50 cm × 50 cm or less. You may prescribe | regulate the lower limit of the magnitude | size (the minimum magnitude | size of the symbol described in the parking frame). For example, only symbols having a size of 15 cm × 15 cm or more may be extracted.

  Next, in step S240, it is determined whether there is a number in the symbol extracted in step S230 by a known process such as pattern matching. If the number is not included, the process proceeds to step S270. In the present embodiment, as a symbol, a symbol including at least a number among numbers, characters (such as alphabets), and “−” is set as a symbol to be extracted. In the pattern matching process, for example, the direction of numbers (vertical direction, etc.) can be specified.

  In step S250, it is determined whether the number in the extracted symbol is not distorted in view of the aspect ratio (ratio of the width to the width of the number), or even if the aspect ratio is equal to or less than the reference ratio. For example, the reference ratio of aspect ratio is set to 0.7 or more and 2 or less. When the number is two or more digits, the aspect ratio may be determined for each number, or may be determined by the aspect ratio of the entire group of numbers.

This is to exclude traffic signs on the road surface. The traffic sign is set to have a large aspect ratio so that the driver can easily recognize it from a car traveling at a predetermined speed or higher, and is described as being vertically distorted in the overhead view image. This step can eliminate such traffic signs.
When the aspect ratio of all the numbers is equal to or less than the reference ratio, the process proceeds to step S260. When the aspect ratio of the numbers exceeds the reference ratio, the process proceeds to step S270.

In step S260, it is determined in step S250 whether there is a symbol that is not to be extracted around symbols whose aspect ratio is less than or equal to the reference ratio. If satisfied, the process proceeds to step S270. Otherwise, the current symbol is excluded from the extraction target, and the process proceeds to step S270. As a result, it is possible to eliminate certain symbols such as time markings.
The area around the above-mentioned symbol to be searched for a symbol that is not the object is, for example, a range that is 2 m in front of the symbol and 1 m in the width direction from the center position of the symbol (see M-AREA in FIG. 31). It may be within a circle having a radius of 1 m from the central position of the symbol, or may be an area from the symbol to a position separated by a predetermined distance. This surrounding area is a range that will not include the vertical line of the parking frame in the width direction, and the surrounding area is set based on the assumption that there will be no symbol above the symbol (number etc.) indicating the parking position. ing. Of the symbols in the surrounding area, it is preferable to exclude straight lines from symbols that are not extracted. By excluding the straight line, it is possible to exclude some lines of the parking frame. The surrounding area may be an area of only 2 m in front of the symbol.

Here, the symbol detected in the present embodiment is a symbol having a line width greater than or equal to a predetermined value. In particular, it is assumed that only the line width of numbers is detected, for example, from 0.7 cm to 3 cm.
In step S270, a flag indicating whether or not the extracted symbol is out, position information, and the like are added to the treated symbol. Thereby, the position of the symbol can be determined.
Next, the processing of the symbol group extraction unit 120B will be described.

  Each time the symbol extraction unit 120B extracts a symbol that allows the symbol extraction unit 120A to identify a parking frame (hereinafter also referred to as a parking instruction symbol candidate), the symbol group extraction unit 120B acquires information on the symbol, and two or more parking instruction symbols For the relationship between candidates, a process as shown in FIG. 16 is executed. The target parking instruction symbol candidate may be a symbol having a specific area, for example, a symbol existing in an area within a radius of 10 m centered on the host vehicle, or a time range previously set with reference to the present (for example, current The target of processing may be narrowed down for symbols detected within 10 minutes from the target.

  When the symbol group extraction unit 120B newly acquires information on parking instruction symbol candidates from the symbol extraction unit 120A as shown in FIG. 16, first, in step S310, the positions of the plurality of parking instruction symbol candidates extracted so far. With reference to the information, it is determined whether another parking instruction symbol candidate exists at a preset distance from the new parking instruction symbol candidate. The distance set in advance is a value corresponding to the width of the parking frame, and is set to a distance in the range of 1.5 m to 2.5 m, for example. Only the upper limit value may be set as the distance set in advance.

If there is no other parking instruction symbol candidate within the preset distance range, it is handled as one of the parking instruction symbol candidates extracted so far, and the process is terminated.
If another parking instruction symbol candidate exists within the preset distance range, the process proceeds to step S320. Here, other parking instruction symbol candidates that existed within a preset distance range are referred to as other parking instruction symbol candidates (others).

In step S320, the numbers of the other parking instruction symbol candidates (other) and the new parking instruction symbol candidate are in the same direction, and the direction intersects the parallel direction of the two parking instruction symbol candidates. Judge whether you are doing. The intersection angle is, for example, 30 degrees or more. The reason for the intersection is to eliminate the numerical strings juxtaposed vertically.
If the determination condition is satisfied, the process proceeds to step S330.

If the judgment conditions are not satisfied, the other parking instruction symbol candidates (others) are not pairs that form a symbol group, so that they are treated as one of the parking instruction symbol candidates extracted so far. Exit.
In step S330, a flag indicating that the other parking instruction symbol candidates (others) and the new parking instruction symbol candidate are a pair of symbol groups is set.

Here, when there are two or more other parking instruction symbol candidates within the preset distance range, the processes of step S320 to step S330 are performed by the number of other parking instruction symbol candidates. The maximum number of other parking instruction symbol candidates that are paired with new parking instruction symbol candidates is two.
Next, processing of the serial number determination unit 120C will be described.

  Here, the parking symbol level SLVL is included as information of each parking instruction symbol candidate, and the initial value of the parking symbol level SLVL is set to “0”. For example, when the ignition is turned on, initial processing is executed. Here, the process of reducing the parking symbol level SLVL set previously is not executed. That is, once the parking symbol level SLVL is set to “2” in relation to another parking instruction symbol candidate, and it is determined that “1” is set in the current process, the parking symbol level SLVL is set to “1”. Do not set to a reduced value. Here, in this embodiment, the case where the parking symbol level SLVL is 2 or more is assumed to have high serial numbering (considered as serial numbering).

If the symbol group extraction unit 120B determines that there is another parking instruction symbol candidate (other) in the pair in which the new parking instruction symbol candidate forms a symbol group, the serial number determination unit 120C, as shown in FIG. Process.
That is, the serial number determination unit 120C determines in step S410 the new parking instruction symbol candidate and the parking instruction symbol candidate (others) determined as a pair symbol group by the symbol group extraction unit 120B, and the new parking instruction symbol. When it is determined that the number portion with the candidate is a serial number, the process proceeds to step S420, and the parking symbol level SLVL = 2 is set for the two pairs of parking instruction symbol candidates.

For example, if the symbol group of the pair is as follows, it is determined as a serial number.
・ "A12" and "A13"
"B40-1" and "B40-2" (: Example when branch numbers are sequential)
Further, in step S430, the parking instruction symbol candidate (other) of the pair forms a pair with another parking instruction symbol candidate, and the numbers of the three parking instruction symbol candidates including the new parking instruction symbol candidate are If it is determined that the symbols are sequentially numbered in the order of symbol arrangement, the parking symbol level SLVL = 3 of the three parking instruction symbol candidates is set in step S440.

On the other hand, when neither of the above conditions is satisfied, the parking symbol level SLVL = 1 of the parking instruction symbol candidate of the pair is set.
Next, in step S450, based on the processing result of the parking frame line information processing unit, the intermediate between the new parking instruction symbol candidate and the parking instruction symbol candidate (other) determined as a pair symbol group by the symbol group extraction unit 120B. It is determined whether there is a parking frame line candidate extending in the same direction as the direction of the number in the new parking instruction symbol candidate at the position. If the determination condition is satisfied, “2” is set to the parking symbol level SLVL in step S460. This is because the middle line is highly likely to be a vertical line of the parking frame.

  Next, in step S470, based on the navigation information, the position of the new parking instruction symbol candidate and the parking instruction symbol candidate (other) determined as a pair symbol group by the symbol group extraction unit 120B is in the parking lot. If it is determined, the parking symbol level SLVL is set to “2” for the new parking instruction symbol candidate and the parking instruction symbol candidate (others) determined as a pair symbol group by the symbol group extraction unit 120B in step S480. Set. Judgment whether it is a parking lot is not limited to navigation information. For example, when a character indicating the entrance of a parking lot such as a parking lot is recognized, it may be recognized that the vehicle is inside the parking lot until it leaves the place.

The processes in steps S410 to S480 are performed in order for each other parking instruction symbol candidate (others) paired with a new parking instruction symbol candidate. There are a maximum of two other parking instruction symbol candidates (others).
Next, the process of the parking frame estimation unit 120D will be described.
The parking frame estimation unit 120D estimates a preset range from the position of the parking instruction symbol candidate symbol as a virtual parking frame for the parking instruction symbol candidate whose parking symbol level SLVL is set to 2 or more.

  Usually, the symbol for identifying the parking frame is written on the front side of the parking frame and is present at the central position in the width direction of the parking frame. Estimate the line. For example, as shown by a broken line in FIG. 31, it is estimated that there is a vertical line of the parking frame at the center position between the paired parking instruction symbol candidates. And the parking frame corresponding to each symbol is estimated from the estimated vertical line and the position and orientation of the symbol, and set as a virtual parking frame. For example, a vertical line having a length of 4 m, for example, is set as a virtual parking frame from the horizontal position (alignment direction) of the symbol position of the parking instruction symbol candidate. Further, a vertical line that is paired with the estimated vertical line is estimated around the center position of the symbol. And the position of a horizontal line is estimated so that the edge part of a pair of vertical line may be connected.

The processing of the processing of the parking frame estimation unit 120D may be executed when, for example, the parking symbol level SLVL is determined to be 2 or more in step S525 described later.
Next, processing of the acceleration suppression operation condition determination unit 10I will be described with reference to the drawings. The acceleration suppression operation condition determination unit 10I performs the process shown in FIG. 18 at every preset sampling time.

In step S510, the acceleration suppression operation condition determination unit 10I acquires frame line information having a parking line level FLVL of 1 or more as road surface environment recognition information calculated by the surrounding environment recognition information calculation unit 10A.
Next, in step S520, the presence or absence of a parking frame is determined based on the frame line information acquired in step S110. If there is frame line information with a parking line level FLVL of 2 or more, it is determined that there is a parking frame, and the process proceeds to step S530. On the other hand, since there is no frame line information having a parking line level FLVL of 2 or more, when it is determined that there is no highly reliable parking frame, it is determined that the acceleration suppression operation condition is not satisfied, and the process proceeds to step S525.

  Since there is no frame line information with a parking line level FLVL of 2 or higher, it is determined that there is no highly reliable parking frame, and when the process proceeds to step S525, the currently recognized symbol is a parking symbol level SLVL of “2” or higher. It is determined whether or not it is a parking instruction symbol candidate. If it is determined that the parking instruction symbol candidate is “2” or more, the same value as the parking symbol level SLVL is set to the parking line level FLVL of the parking frame line estimated from the symbol, and the virtual parking frame of the parking instruction symbol candidate Is regarded as a parking frame line, and the process proceeds to step S530. If the currently recognized symbol is not a parking instruction symbol candidate whose parking symbol level SLVL is “2” or higher, the process proceeds to step S590, and in step S590, an acceleration suppression operation condition determination result (= acceleration suppression operation condition not established) ) Is output to the acceleration limit value calculation unit.

Next, in step S530, the vehicle speed of the host vehicle MM is acquired from the host vehicle vehicle speed calculation unit 10B.
Next, in step S540, the host vehicle vehicle speed condition is determined based on the host vehicle speed acquired in step S530. For example, if the host vehicle speed is less than a preset value, the process proceeds to step S550. If the host vehicle speed is equal to or greater than the preset value, it is determined that the acceleration suppression operation condition is not established, and the process proceeds to step S590. In step S590, the acceleration suppression operation condition determination result (= acceleration suppression operation condition not established) is output to acceleration suppression amount calculation unit 10J. The preset value is, for example, 15 [km / h].

Next, in step S550, brake pedal operation information is acquired from the brake pedal operation information calculation unit 10F.
Next, in step S560, the brake pedal operation is determined based on the brake pedal operation information acquired in step S550. If it is determined that there is no brake pedal operation, the process proceeds to step S570. On the other hand, if it is determined that there is a brake pedal operation, it is determined that the acceleration suppression operation condition is not satisfied, and the process proceeds to step S590. In step S590, the acceleration suppression operation condition determination result (= acceleration suppression operation condition is not satisfied). Is output to the acceleration suppression amount calculation unit 10J.

In step S570, the accelerator operation amount is acquired from the accelerator operation amount calculation unit 10G.
Next, in step S580, the accelerator operation amount is determined based on the accelerator operation amount acquired in step S570. For example, when the accelerator operation amount is equal to or greater than a preset value, it is determined that the acceleration suppression operation condition is satisfied. On the other hand, if the accelerator pedal operation is less than the preset value, it is determined that the acceleration suppression operation condition is not established, and the process proceeds to step S590. In step S590, the acceleration suppression operation condition determination result is used as the acceleration suppression operation calculation unit 10J. Output to. Here, the preset value is set to an operation amount corresponding to 3% of the accelerator opening of the accelerator pedal 19, for example.

Next, in step S583, parking frame approach determination information is acquired. Here, in this embodiment, it is assumed that the parking frame approach determination is performed based on the steering angle, the angle between the host vehicle MM and the parking frame, and the distance between the host vehicle MM and the parking frame.
Specifically, in step S583, the steering angle is acquired from the steering angle calculation unit 10C. In step S583, the angle α between the host vehicle MM and the parking frame L0 and the distance D between the host vehicle MM and the parking frame L0 are acquired based on the host vehicle surrounding image calculated by the surrounding environment recognition information calculation unit 10A. Here, for example, as shown in FIG. 19, the angle α between the host vehicle MM and the parking frame L0 is parked in the front-rear direction straight line X (straight line extending in the traveling direction) X passing through the center of the vehicle and the parking frame L0. Is the absolute value of the angle of intersection between the frame line L1 of the parking frame L0 portion that is parallel or substantially parallel to the front-rear direction of the vehicle and the line on the parking frame L0 side that is an extension thereof. The distance D between the host vehicle MM and the parking frame L0 is, for example, the distance between the center point of the front end surface of the host vehicle and the center point of the entrance L2 of the parking frame L0 as shown in FIG. However, the distance D between the host vehicle MM and the parking frame L0 is a negative value after the front end surface of the host vehicle passes through the entrance L2 of the parking frame L0. The distance D between the host vehicle MM and the parking frame L0 may be set to zero after the front end surface of the host vehicle passes through the entrance L2 of the parking frame L0.

Here, the position on the own vehicle MM side for specifying the distance D need not be the center point of the front end surface of the own vehicle. The distance between the position set in advance in the host vehicle MM and the position set in advance at the entrance L2 may be D.
Thus, in step S583, the steering angle, the angle α between the host vehicle MM and the parking frame L0, and the distance D between the host vehicle MM and the parking frame L0 are acquired as the parking frame approach determination information.

  Next, in step S586, a parking frame approach determination is performed based on the parking frame approach determination information acquired in step S583. If it is determined that the parking frame has entered, it is determined that the acceleration suppression operation condition is satisfied. On the other hand, when it is not determined that the parking frame has entered, it is determined that the acceleration suppression operation condition is not satisfied. Thereafter, the process proceeds to step S590, and the acceleration suppression operation condition determination result is output to the acceleration suppression amount calculation unit 10J.

For example, the parking frame approach is determined as follows. That is, in step S586, it is determined that the parking frame has entered when all of the following three conditions (ac) are satisfied.
a: Within a preset time (for example, 20 [sec]) after the steering angle detected in step S583 becomes equal to or greater than a preset steering angle value (for example, 45 [deg]) b: With the host vehicle MM The angle α of the parking frame L0 is equal to or less than a preset setting angle (for example, 40 [deg]). C: A preset distance (for example, 3 [m]) where the distance D between the host vehicle MM and the parking frame L0 is set.
Here, although the case where a plurality of conditions are used for the parking frame approach determination is illustrated, the determination may be performed based on one or more of the above conditions. Moreover, you may determine whether it is approaching to the parking frame L0 by the state of the vehicle speed of the own vehicle MM.

Next, processing of the acceleration suppression amount calculation unit 10J will be described with reference to the drawings. The acceleration suppression amount calculation unit 10J performs the process shown in FIG. 20 at every preset sampling time.
In step S610, an acceleration suppression operation condition determination result is acquired from the acceleration suppression operation condition determination unit 10I.
Next, in step S620, acceleration suppression process selection information is acquired. In step S620, for example, the accelerator operation amount is acquired from the accelerator operation amount calculation unit 10G, the accelerator operation speed is acquired from the accelerator operation speed calculation unit 10H, and the acceleration suppression operation condition determination result is acquired from the acceleration suppression operation condition determination unit 10I.

  Next, in step S630, an acceleration suppression process is selected based on the acceleration suppression process selection information acquired in step S620. Specifically, when it is determined that the operating condition of the second acceleration suppression process is established, the process proceeds to step S640. If it is determined that the operating condition of the second acceleration suppression process is not satisfied and the operating condition of the first acceleration suppression process is satisfied, the process proceeds to step S650. Further, when the operating conditions of the second acceleration suppression process and the first acceleration suppression process are not satisfied, the process proceeds to step S660.

The determination of the processing in step S630, particularly the operating conditions of the second acceleration suppression process and the operating conditions of the first acceleration suppression process will be described with reference to FIG.
First, in step S631, it is determined whether or not the second acceleration suppression process was operating during the determination process in the previous control cycle. If the second acceleration suppression process is operating in the determination in the previous control cycle, the process proceeds to step S633. If the second acceleration suppression process has not been activated in the determination in the previous control cycle, the process proceeds to step S635.

  In step S633, it is determined whether or not the second acceleration suppression process has ended when the second acceleration suppression process has been operating last time. Specifically, if it is determined that the accelerator operation is being performed based on the accelerator operation amount acquired in step S620, it is determined that the second acceleration suppression operation is to be continued, and the process proceeds to step S640. On the other hand, if it is determined that the accelerator operation is not performed, the process proceeds to step S635 in order to perform the operation condition determination again.

In step S635, the operating condition of the first acceleration suppression process is determined. For example, when the acceleration suppression operation condition determination result acquired in step S610 determines that the condition is satisfied, the operation condition of the first acceleration suppression process is determined to be satisfied, and the process proceeds to step S637. On the other hand, when the acceleration suppression operation condition determination result determines that the condition is not satisfied, the process proceeds to step S660.
In step S637, the second acceleration suppression process operating condition is determined. For example, when all of the following conditions (d to f) are satisfied, it is determined that the second acceleration suppression process is to be performed, and the process proceeds to step S640. Otherwise, the process proceeds to step S650.

d: The acceleration suppression operation condition determination result acquired in step S610 is satisfied. e: The accelerator operation amount acquired in step S620 is equal to or greater than a preset operation amount (for example, the accelerator opening is 50 [%]). f: Accelerator operation speed Is a preset operation speed (eg, 200 [% / sec]) or more. In step S640 in FIG. 20, the second acceleration suppression amount is calculated based on the information acquired in step S620, and the process proceeds to step S670.

  The calculation method of the second acceleration suppression amount is performed as follows, for example. That is, based on the accelerator operation amount acquired in step S620, an acceleration suppression amount is calculated such that the acceleration suppression amount does not become larger than the preset set suppression amount, and the process proceeds to step S670. Specifically, as shown in FIG. 22, for an acceleration operation amount less than a preset value, a throttle opening corresponding to the acceleration operation is calculated, and an acceleration operation (accelerator operation) greater than a preset value is performed. On the other hand, the acceleration suppression amount is calculated so that the acceleration throttle opening (acceleration command value) does not exceed 10 [%] regardless of the acceleration operation. In FIG. 22, the solid line indicates the accelerator operation amount and the throttle opening in the normal state, that is, in a state where the suppression is not performed. A one-dot chain line indicates the relationship between the accelerator operation and the throttle opening when the second acceleration suppression is performed. That is, the difference between the solid line and the one-dot chain line in the detected accelerator operation amount becomes the second acceleration suppression amount.

  In step S650, the first acceleration suppression amount is calculated based on the information acquired in step S620, and the process proceeds to step S670. A method for calculating the first acceleration suppression amount will be described. Based on the accelerator operation amount acquired in step S620, the first acceleration suppression amount is calculated so that the throttle opening is increased according to the accelerator operation amount, and the process proceeds to step S670. Specifically, as shown in FIG. 23, the throttle opening (acceleration command value) is calculated to increase as the accelerator operation amount increases. Here, the first acceleration suppression amount is an acceleration suppression amount that is smaller than the second acceleration suppression amount relative to the accelerator operation amount and greater in acceleration. Calculate the acceleration suppression amount so that In FIG. 23, the solid line indicates the accelerator operation amount and the throttle opening in the normal state, that is, in the state where the suppression is not performed. A one-dot chain line indicates the relationship between the accelerator operation and the throttle opening when the second acceleration suppression is performed. That is, the difference between the solid line and the two-dot chain line in the detected accelerator operation amount is the first acceleration suppression amount.

Here, as shown in FIG. 23, the second acceleration suppression amount is larger than the first acceleration suppression amount, and as shown in FIGS. 22 and 23, both the first acceleration suppression amount and the second acceleration suppression amount are accelerators. It is set to increase as the operation amount increases.
In step S660, an acceleration suppression amount that does not suppress acceleration is calculated for the accelerator operation, and the process proceeds to step S670. In the present embodiment, the acceleration suppression amount for which acceleration suppression is not performed is set to zero.

In step S670, the acceleration suppression amount calculated in step S602 is output to the target throttle opening calculation unit 10K.
Next, the processing of the target throttle opening calculation unit 10K will be described with reference to the drawings. The target throttle opening calculation unit 10K performs the processing shown in FIG. 24 at every preset sampling time.

First, in step S710, an acceleration suppression operation condition determination result is acquired from the acceleration suppression operation condition determination unit 10I.
Next, in step S720, the accelerator operation amount is acquired from the accelerator operation amount calculation unit 10G.
Next, in step S730, an acceleration suppression amount is acquired from the acceleration suppression amount calculation unit 10J.

  Next, in step S740, the target throttle opening is calculated based on the acceleration suppression operation condition determination result acquired in step S710, the accelerator operation amount acquired in step S720, and the acceleration suppression amount acquired in step S730. For example, when the acceleration suppression operation condition is not satisfied, the throttle opening based on the normal accelerator operation amount without acceleration suppression is set as the target throttle opening. On the other hand, when the acceleration suppression operation condition is satisfied, the throttle opening based on the acceleration suppression amount is set as the target throttle opening.

For example, the target throttle opening degree θ * is obtained by the following equation.
θ * = θ1-Δθ
Here, θ1 indicates the throttle opening corresponding to the accelerator operation amount, and Δθ indicates the acceleration suppression amount.
Next, in step S750, the target throttle opening degree θ * calculated in step S340 is output to the engine controller 16.
The engine controller 16 controls the engine that is a drive source by controlling the throttle opening so that the acquired target throttle opening θ * is obtained.

(Operation other)
An example of a time chart according to the processing of the present embodiment is shown in FIG.
In this example, the parking frame approach operation detection processing is performed based on the angle α (condition b) between the host vehicle MM and the parking frame L0 and the distance D (condition c) between the host vehicle MM and the parking frame L0. It is an example which detects approach operation.

In the example shown in FIG. 25, a parking frame L0 having a preset certainty (confidence) is detected, or a parking frame is estimated from a symbol for a specific symbol group (t1), and the vehicle speed is set in advance. When the speed is equal to or lower than the speed (t2), the operation for entering the parking frame L0 is determined.
When detecting the parking frame L0 having the certainty (certainty factor) set in advance here, a line having a radial line attribute FF (n) as shown in FIG. 11 and a line having a solid object line attribute ( The parking frame determination is performed by excluding the edge line of the three-dimensional object.

  At this time, there may be a case where a nearby structure (a surface portion of a three-dimensional object) is reflected on a frozen road surface or a water pool, and the edge of the reflected object is reflected as a line on the overhead view screen. Since such a line reflected on the road surface is a light beam that always faces the vehicle even when the vehicle moves, it is normally recognized as a radial line centered on the light receiving unit 1a of the camera on the overhead view image. In addition, the edge line of a three-dimensional object extending vertically from the road surface is also recognized as a radial line on the overhead image.

Based on this, by removing the radial line from the parking frame candidate, erroneous recognition due to the reflection line is suppressed, and the accuracy of the parking frame determination is improved.
Similarly, by removing the edge line of the three-dimensional object from the parking frame candidates, erroneous recognition due to the reflection of the three-dimensional object is suppressed, and the accuracy of the parking frame determination is improved. Here, as shown in FIG. 26, the edge line of the three-dimensional object is different from the movement of the line drawn on the road surface in the movement of the apparent line accompanying the movement of the vehicle. This makes it possible to determine whether the line is a solid object.

In addition, the accuracy of parking frame determination can be improved by interpolating two lines that are halfway and broken in the middle with the same line.
At this time, even if the breaks between the lines are so far apart that it is difficult to think that they are due to fading, if the paired lines are approximately the same length and the distance between the breaks is also approximate, Interpolate and treat the two lines as the same line and treat them as non-frame candidates.

  Here, at the railroad crossing, as shown in FIG. 27, a pair of lines Ls divided at the track position are continuous. This is detected as a continuous parting line in a line L5 of the overhead image as shown in FIG. In the present embodiment, the line L5 is regarded as a single line, thereby avoiding individual lines from being individually handled as parking frame candidate lines. In addition, the attribute with the non-candidate of a parking frame is added with respect to this line. In addition, as shown in FIG. 29, there is a case where a railroad crossing obliquely crosses a track, but there is no problem because a pair of lines is detected as an intermittent line of an approximate length.

Furthermore, in this embodiment, even if the parking frame is blurred and the line of the frame itself is unknown, it is estimated that a specific parking frame is specified. The driving support is performed by estimating the parking frame position from the size of the symbols, the serial number of the symbols, and the arrangement direction of the symbols, and detecting the approach operation to the estimated parking frame L0. .
Here, the example detected as a parking frame by the parking frame determination of this embodiment is shown in FIG.

Moreover, the example at the time of estimating a parking frame from the symbol containing a number and extracting the said parking frame is shown in FIG. In FIG. 31, the broken lines indicate the positions of the estimated horizontal and vertical lines of the parking frame. As shown in FIG. 31, the parking frame is estimated according to the position of the target symbol and the arrangement of the symbol group.
Next, after time t2, in the example shown in FIG. 19, the distance D (condition c) between the host vehicle MM and the parking frame L0 is equal to or smaller than the preset distance (t3), and the angle α between the host vehicle MM and the parking frame L0. When (Condition b) is equal to or less than a preset angle (t4), it is determined that the operation is to enter the parking frame L0, and the acceleration suppression is activated.

  When the driver performs an accelerator operation in this acceleration suppression operating state, an acceleration command value (throttle opening) corresponding to the accelerator operation is suppressed. Further, when the acceleration operation amount becomes equal to or larger than the preset operation amount in the state where the acceleration suppression is performed (t5), the acceleration command value suppression amount is increased. In the present embodiment, as a result of performing acceleration suppression so as to suppress below a preset throttle opening, as shown in FIG. 32, the actual throttle opening is compared with before the accelerator operation amount exceeds the preset operation amount. Is suppressed small. As a result, acceleration suppression against erroneous operation of the accelerator pedal 19 by the driver is executed more effectively.

Here, FIG. 32 shows an example of the transition of the throttle opening (acceleration command amount) of the acceleration suppression control according to the operation amount of the accelerator pedal. In the example shown in FIG. 32, even if the acceleration suppression process shifts to the second acceleration suppression process, the accelerator pedal is returned, and the acceleration suppression amount of the first acceleration suppression process is equal to the acceleration suppression amount of the second acceleration suppression process. At that time, the process proceeds to the first acceleration suppression process.
As described above, it is possible to suppress acceleration by detecting an entry operation to the parking frame L0 having a certainty as a parking frame that is equal to or higher than a certainty factor set in advance, that is, detecting that the host vehicle MM enters the parking frame L0. Operating conditions. As a result, even if the host vehicle MM departs from the road and enters the parking lot, for example, acceleration suppression is not performed until an entry operation to the parking frame L0 is detected. I can do it. Furthermore, after performing the approach operation to the parking frame L0, it is possible to realize acceleration suppression with a high acceleration suppression effect when the accelerator pedal is erroneously operated by suppressing acceleration.

In addition, two-stage acceleration suppression is performed when an entry operation to the parking frame L0 is detected and when an acceleration operation is further performed to increase the possibility of an erroneous operation of the accelerator pedal. As a result, it is possible to perform acceleration suppression with a high acceleration suppression effect when the accelerator pedal is erroneously operated while suppressing a decrease in drivability.
Even in the state of entering the parking frame L0, the throttle opening is increased according to the acceleration operation amount, but acceleration is suppressed so that the throttle opening is smaller than usual. That is, by increasing the acceleration suppression amount as the acceleration operation amount increases, it becomes possible to perform acceleration suppression with a high acceleration suppression effect with little decrease in drivability. When the acceleration operation is small, the acceleration suppression amount is small, so that the drivability is small. When the acceleration operation is large, the acceleration suppression amount is large and the acceleration suppression effect is high.

Furthermore, when a large acceleration operation is performed to enter the second acceleration suppression state, acceleration suppression is performed so that the throttle opening does not increase beyond a preset value (a predetermined amount greater than the acceleration operation amount determined to be the first acceleration state). I do. As a result, it is possible to suppress acceleration that is not intended by the driver due to erroneous operation of the acceleration operation, and it is possible to perform acceleration suppression that is highly effective in avoiding or reducing accidents. Even if the second acceleration suppression state is entered, the first acceleration suppression state is entered if the acceleration operation amount becomes smaller than a preset value.

  As described above, when a large acceleration operation is performed while the vehicle is entering the parking frame L0, acceleration suppression having a larger suppression amount than the acceleration suppression amount by the first acceleration suppression process is performed as the second acceleration suppression process. Thus, it is possible to suppress unintended acceleration due to an erroneous operation of the driver's acceleration operation, and it is possible to perform acceleration suppression with a high effect of parking at a target parking position.

Further, by detecting the operation amount of the accelerator pedal 19 and the speed of the accelerator pedal operation as the acceleration operation amount, it is possible to more accurately distinguish between the erroneous operation of the acceleration operation and the normal operation, and there is little decrease in drivability, It becomes possible to realize acceleration suppression with a high acceleration suppression effect.
Further, the parking frame is determined from the vehicle speed of the host vehicle MM, the steering angle of the host vehicle MM, the angle α between the host vehicle MM and the parking frame L0, and the distance D between any point of the host vehicle MM and the entrance L2 of the parking frame L0. By detecting the approach operation to L0, it is possible to distinguish whether the driver is traveling past the parking frame L0 or is about to park in the parking frame L0 from the surrounding environment recognition processing, and more drivability Parking assistance with little decrease in traffic becomes possible.

  At this time, the angle α formed by the traveling direction of the host vehicle MM and the parking direction to the parking frame L0 is set to the angle α between the host vehicle MM and the parking frame L0, so that the operation of entering the detected parking frame L0 is performed. It is possible to detect the progress (the certainty of entering the parking frame). As a result, it is possible to accurately determine that the vehicle is going to be parked in the parking frame L0 based on the detected value, and parking assistance with less reduction in drivability is possible.

(Modification)
(1) In the above embodiment, when the parking line level FLVL is 2 or more, the parking frame is present. The parking line level FLVL may be 3 or more with a parking frame.
(2) Moreover, in the said embodiment, the start determination of acceleration suppression control is performed by binarization of parking frame approach presence or absence at step S586 as parking frame approach determination. On the other hand, in the case where there is a parking frame approach, the likelihood of entering the parking frame may be determined in a plurality of stages using the approach certainty factor ALVL indicating the certainty of entering the parking frame. And according to the approach certainty factor ALVL and the parking line level FLVL, you may change the content of the said acceleration suppression control.

For example, the approach certainty level ALVL when there is a parking frame approach is divided into two stages of “low” and “high”, and the combination of the certainty level ALVL and the parking line level FLVL, as shown in FIG. The total certainty TLVL of the parking assistance is calculated. And based on the comprehensive reliability TLVL of the parking assistance, you may make it control parking assistance like FIG.
In the parking assistance control shown in FIG. 34, when the total certainty TLVL is “extremely low”, acceleration suppression is started when the accelerator opening is 80% or more, and the accelerator is controlled by a preset amount (small amount). Do suppression. In addition, when the total certainty level TLVL is “low”, acceleration suppression is started when the accelerator opening is 80% or more, and a preset amount (the total certainty level TLVL is larger than “very low”). Only the accelerator suppression is performed and the accelerator suppression notification process to the driver is performed. When the total certainty level TLVL is “high”, acceleration suppression is started when the accelerator opening is 50% or more, and a preset amount (the total certainty level TLVL is larger than “very low”). Only the accelerator suppression is performed and the accelerator suppression notification process to the driver is performed. When the total certainty level TLVL is “extremely high”, acceleration suppression is started when the accelerator opening is 50% or more, and a preset amount (the total certainty level TLVL is larger than “high”). Only the accelerator suppression is performed and the accelerator suppression notification process to the driver is performed. The percentage of acceleration suppression execution conditions (accelerator opening) shown in FIG. 34 is an example.

(3) Moreover, in the said embodiment, acceleration suppression control was illustrated as braking / driving force control based on the detected parking frame. The braking / driving force control based on the detected parking frame is not limited to this. For example, braking / driving force control for performing guidance support to the detected parking frame may be used.
Moreover, it is not limited to the braking / driving force control for assisting the approach to the detected parking frame. The braking / driving force control based on the detected parking frame may be vehicle start control from the detected parking frame.

Further, in the above description, the line is detected based on the bird's-eye view image that has been subjected to the bird's-eye view conversion.
Here, the surrounding environment recognition information calculation unit 10A constitutes an imaging unit. Step S20 constitutes an overhead image acquisition unit. The acceleration suppression operation condition determination unit 10I, the acceleration suppression amount calculation unit 10J, and the target throttle opening calculation unit 10K constitute a braking / driving force control unit that performs acceleration suppression control. The navigation 7 constitutes a parking lot position detection unit. The acceleration suppression amount calculation unit 10J constitutes an acceleration suppression unit.

(Effect of this embodiment)
According to this embodiment, the following effects are produced.
(1) The symbol extraction unit 120A extracts a symbol including at least a number from a captured image including a road surface around the host vehicle. When a plurality of symbols are extracted by the symbol extraction unit, the sequence number determination unit determines whether the numbers of these symbols are serial numbers. A parking frame estimation part presumes that there is a parking frame in a predetermined position from the above-mentioned symbol, when it judges with a serial number judgment part being a serial number. And if it determines with the parking frame estimated by the said parking frame estimation part ahead of the advancing direction of the own vehicle, the acceleration which the own vehicle generates will be reduced according to an acceleration operator.
According to this configuration, since the parking frame is estimated from the specific symbol, the acceleration of the host vehicle is reduced when there is a high possibility of a stepping error when traveling toward the parking frame. Therefore, the possibility of giving the driver a sense of incongruity can be reduced.

  (2) The symbol extraction unit 120A is a symbol that includes a number and is located on the road surface from the captured image including the road surface around the host vehicle, and has a size within a range assumed to be described in the parking frame. Extract symbols. The symbol group extraction unit 120B has the plurality of extracted symbols arranged in parallel with a predetermined distance between the adjacent symbols, and the direction of each symbol intersects the arrangement direction of the symbols. A symbol group consisting of a plurality of symbols that are directed in the same direction is extracted. The serial number determination unit 120C determines whether the numbers of adjacent symbols in the symbol group extracted by the symbol group extraction unit 120B are serial numbers. The parking frame estimation unit 120D estimates the position of the parking frame from the position of each symbol in the symbol group for the symbol group determined by the serial number determination unit 120C as the serial number. When the travel controller 10 determines that there is a parking frame estimated by the parking frame estimation unit ahead of the traveling direction of the own vehicle based on the parking frame extracted by the parking frame estimation unit 120D, the own vehicle according to the acceleration operator Reduce the acceleration that occurs.

According to this configuration, it is possible to estimate the position of the parking frame even in a situation where the parking frame is not recognized from the line because the line constituting the parking frame is faded. As a result, the reliability of parking frame detection is improved, and the accuracy of driving assistance for parking can be improved.
If symbols including specific numbers are arranged at predetermined intervals, it can be estimated that the symbol indicates a parking frame number.
Moreover, according to this structure, even if it makes a mistake in stepping on a brake pedal and an accelerator pedal when heading for a parking frame, acceleration of the own vehicle is suppressed.

(3) The symbol extraction unit 120A extracts symbols based on the bird's-eye view image obtained by bird's-eye conversion, and extracts only symbols whose distortion is less than a preset reference ratio in which the aspect ratio of the numbers in the symbols is set.
According to this configuration, it is possible to exclude numbers in traffic regulations such as vehicle speed that are marked on the roadway. In general, the numbers marked on the roadway are set to have a large aspect ratio so that they can be easily seen from the traveling vehicle. Symbols including such numbers can be excluded.
Further, since the determination is made from the overhead view, the aspect ratio can be obtained with higher accuracy, and the acceleration suppression control of the vehicle is performed based on the symbols in the vicinity of the host vehicle.

(4) When there is a symbol that is not subject to extraction in the surrounding area that is set in advance around the extracted symbol and based on the symbol position, the symbol extraction unit 120A uses the extracted symbol as the symbol to be extracted. Remove from.
According to this configuration, for example, when there is a time sign written in the vicinity of a number in the school zone, such a symbol can be excluded, and the accuracy of symbol extraction is improved.

(5) When the serial number determination unit 120C detects a straight line extending in the same direction as the direction of the symbol in the middle position between two adjacent symbols in the symbol group extracted by the symbol group extraction unit 120B, the adjacent symbol Even if the numbers are not serial numbers, they are regarded as serial numbers.
If symbols including numbers of a predetermined size are arranged at a predetermined interval and there is a straight line extending in the same direction as the numbers of the symbols at the intermediate position between the symbols, the symbols specify the parking frame The possibility of a symbol is high. Therefore, according to this configuration, it is possible to estimate the parking frame even if the numbers of adjacent symbols are low in serial number. For example, the parking frame can be estimated even when the numbers arranged at a predetermined interval are “12”, “12-A”, and the number itself has a low serial number.

(6) If the serial number determination unit 120C determines that each symbol of the symbol group extracted by the symbol group extraction unit 120B is located in the parking lot, even if the numbers of adjacent symbols are low in serial number, Consider it a turn.
In the parking lot, when symbols including numbers of a predetermined size are arranged at predetermined intervals, the symbols are highly likely to be symbols that identify a parking frame. Therefore, according to this configuration, it is possible to estimate the parking frame even if the numbers of adjacent symbols are low in serial number.

  (7) The travel controller 10 converts the captured image captured by the imaging unit (camera) into a bird's-eye view image to obtain a bird's-eye view image. The travel controller 10 detects a line located on the road surface from a captured image captured by the imaging unit. The travel controller 10 extracts a radial line centered on the light receiving unit 1a of the imaging unit in the overhead view image. The travel controller 10 adds noise information (radial line attribute FF) indicating that it is not a parking frame candidate to a line corresponding to the extracted radial line among the lines detected by the line detection unit. The travel controller 10 extracts a parking frame based on the detected line on the road surface and noise information (radial line attribute FF). The travel controller 10 controls the braking / driving force of the vehicle based on the extracted parking frame.

  According to this configuration, in the bird's-eye view image, a radial line centered on the light receiving unit 1a of the imaging unit is excluded from the parking frame candidate lines. The line reflected from the road surface is recognized as a radial line in the overhead view image. For this reason, it can suppress that the line reflected on the road surface is determined as the line which comprises a parking frame by excluding the said line. As a result, the reliability of parking frame detection is improved, and the accuracy of driving assistance for parking can be improved.

(8) The travel controller 10 excludes a line that has been recognized as a radial line continuously for a preset time from the line candidates that constitute the parking frame.
According to this configuration, it is possible to detect a line corresponding to the reflected line more reliably. This improves the reliability of parking frame detection and improves the accuracy of driving support for parking.

(9) When the traveling controller 10 estimates that the road surface is easy to reflect, the traveling controller 10 relaxes the determination condition for detecting the radial line.
As a result, road reflection lines can be detected more reliably. This improves the reliability of parking frame detection and improves the accuracy of driving support for parking.
(10) The travel controller 10 extracts a radial line from the overhead image at every preset time, and when the line detected by the line detection unit is extracted by the radial line detection unit as not being a radial line, Add non-radial line information to the line. And even if the noise information judgment part added the noise information, the parking frame judgment part parks the line if the non-radial line information is added before the noise information is added. Parking frame determination processing is performed as a candidate frame line.
As a result, it is possible to prevent a line that is not a line reflected on the road surface from being mistaken as a line reflected on the road surface.

(11) The travel controller 10 detects a line located on the road surface from the overhead image.
As a result, line detection and reflection line detection are performed on the same screen, so that coordinate conversion or the like is not required for collation between the detected line and the line estimated to be reflected, and the collation process and the like are simplified.
In addition, since the line is determined in the overhead image, it is also easy to determine parallelism with other lines.
(12) The travel controller 10 recognizes the environment around the host vehicle based on detection information (image information captured by the camera) from the surrounding environment recognition sensor. The travel controller 10 detects the acceleration operation amount from the state of the acceleration operator (accelerator pedal) operated by the driver to give an acceleration instruction. The travel controller 10 detects the state of the host vehicle MM. The traveling controller 10 detects that the host vehicle MM enters the parking frame L0 based on the surrounding environment and the traveling state of the host vehicle MM. When the traveling controller 10 determines that the host vehicle MM enters the parking frame L0, the traveling controller 10 suppresses an acceleration command value (throttle opening) corresponding to the operation amount of the acceleration operator. The travel controller 10 increases the suppression of the acceleration command value when detecting an acceleration operation that is equal to or greater than a preset acceleration operation amount while the acceleration command value is being suppressed.

According to this configuration, when the driver performs an approach operation to the parking frame L0, detection of entry of the host vehicle MM into the parking frame L0 is set as an acceleration suppression operation condition. This makes it possible to perform acceleration suppression with a high acceleration suppression effect when the accelerator pedal is erroneously operated while suppressing a decrease in drivability.
In addition, by suppressing acceleration in two stages, when entering the parking frame L0 and when further accelerating operation is performed after the entering operation, suppressing acceleration when the accelerator pedal is erroneously operated while suppressing deterioration in drivability It is possible to suppress acceleration with higher effect.

(13) If the traveling controller 10 detects an acceleration operation equal to or greater than a preset acceleration operation amount while suppressing the acceleration command value, the travel control controller 10 suppresses the acceleration command value to be equal to or less than a preset upper limit acceleration command value.
According to this configuration, even if the driver performs a large acceleration operation, the acceleration command value by the acceleration operation is suppressed so that the acceleration command value does not become larger than a preset value. It is possible to suppress acceleration that is not intended by the driver. As a result, parking in the parking frame L0 can be further supported.

(14) The travel controller 10 detects at least one of the operation amount of the acceleration operator and the operation speed of the acceleration operator as the acceleration operation amount.
According to this configuration, the operation amount of the accelerator pedal 19 and the operation speed of the accelerator pedal operation are detected as the acceleration operation amount. As a result, it is possible to more accurately distinguish between the erroneous operation of the acceleration operation and the normal operation. As a result, it is possible to realize acceleration suppression with a high acceleration suppression effect with little reduction in drivability.

  (15) Upon detecting the parking frame L0, the traveling controller 10 detects the vehicle speed of the host vehicle MM or the steering angle of the host vehicle MM, the angle α between the host vehicle MM and the parking frame L0, and the entrance of the host vehicle MM and the parking frame L0. Based on the distance D of L2 and at least one information of the positional relationship between the predicted trajectory of the host vehicle MM and the parking frame L0, an approach operation to the parking frame L0 is detected, and the host vehicle MM is detected by the detected approach operation. Detects that the vehicle enters the parking frame L0.

  According to this configuration, the vehicle speed of the host vehicle MM, the steering angle of the host vehicle MM, the angle α between the host vehicle MM and the parking frame L0, the distance D between the host vehicle MM and the entrance L2 of the parking frame L0, and the prediction of the host vehicle MM By using at least one information of the positional relationship between the track and the parking frame L0, it can be determined whether the host vehicle MM is traveling past the detected parking frame L0 or is about to park in the parking frame L0. Can be distinguished. Thus, it becomes possible to detect the approach operation to the parking frame L0.

(16) The travel controller 10 sets the angle α between the traveling direction of the host vehicle MM and the parking direction to the parking frame L0 as the angle α between the host vehicle MM and the parking frame L0, and sets the distance between the host vehicle MM and the parking frame L0. An approach operation to the parking frame L0 is detected based on the angle α.
According to this configuration, the angle α formed between the traveling direction of the host vehicle MM and the parking direction to the parking frame L0 is set to the angle α between the host vehicle MM and the parking frame L0, whereby the detected parking frame L0 is detected. It is possible to detect the progress of the entry operation. Therefore, it is possible to accurately determine whether or not the host vehicle MM is about to park in the parking frame L0 based on the detected value. As a result, it is possible to realize driving support with little reduction in drivability and high acceleration suppression effect.

“Second Embodiment”
Next, a second embodiment will be described with reference to the drawings. Here, the same components as those in the first embodiment will be described with the same reference numerals.
(Constitution)
The basic configuration of this embodiment is the same as that of the first embodiment. However, this embodiment is an example in the case where the entry determination to the parking frame L0 is performed based on the predicted trajectory of the host vehicle MM, the entrance position of the parking frame L0, and the frame range.

That is, the processes of steps S583 and S586, particularly the process of step S586, in the acceleration suppression operation condition determination unit 10I are different. Other processes are the same as those in the first embodiment.
Next, differences in the configuration will be described.
In the acceleration suppression operation condition determination unit 10I, step S583 acquires the steering angle, the steering angular velocity, the vehicle speed of the host vehicle MM, the shift position, the parking frame line position, and the entrance position of the parking frame L0.

Next, the process of step S586 in the present embodiment will be described with reference to the drawings.
Step S186 of this embodiment consists of the process of S586A-S586D, as shown in FIG.
In step S586A, the predicted vehicle trajectory is calculated. For example, the host vehicle predicted trajectory is calculated based on the steering angle, the steering angular velocity, and the shift position acquired in step S580A. Here, there are various methods for calculating the predicted trajectory of the host vehicle, and the present embodiment does not particularly limit the method for calculating the predicted trajectory of the host vehicle. For example, the traveling direction of the host vehicle MM is specified at the shift position, and the expected trajectory of the host vehicle MM is obtained based on the direction of the steered wheels specified by the current steering angle and steering angular velocity.

  In step S586B, the predicted own vehicle trajectory frame line overlap rate is calculated based on the predicted own vehicle track calculated in step S586A and the parking frame line position acquired in step S580A. For example, as shown in FIG. 36, the ratio of the area S0 occupied by the predicted vehicle track S passing through the parking frame L0 to the area of the target parking frame L0 is calculated as the predicted vehicle frame track overlap rate. To do.

  In step S586C, the host vehicle predicted track parking frame entrance overlap rate is calculated based on the host vehicle predicted track calculated in step S586A and the entrance position of the parking frame L0 acquired in step S580A. For example, as shown in FIG. 37, the ratio of the length of the portion H that overlaps the host vehicle track in the length of one side of the frame line that becomes the entrance L2 of the parking frame L0 is calculated as the host vehicle expected track parking frame entrance. Calculate as the overlap rate.

The predicted trajectory is, for example, a range through which the rear wheel passes. It may be the range through which the front wheels pass.
In step S586D, the host vehicle parking frame approach determination is performed based on the host vehicle predicted track frame line overlap rate calculated in step S586B and the host vehicle predicted track parking frame entrance overlap rate calculated in step S586C.
For example, when the host vehicle MM enters the parking frame L0 when the host vehicle predicted track frame line overlap rate is equal to or greater than a preset value and the host vehicle predicted track parking frame entrance overlap rate is equal to or greater than a preset value. to decide. Specifically, it is determined that the host vehicle MM enters the parking frame L0 when the host vehicle predicted track frame line overlap rate is 40% or more and the host vehicle predicted track parking frame entrance overlap rate is 30% or more. . Here, it is good also as a structure which judges the own vehicle parking frame approach only by either the own vehicle estimated track frame line duplication rate or the own vehicle expected track parking frame entrance duplication rate.

Further, the approach certainty factor ALVL indicating the probability of entering the parking frame may be set in two or more stages based on the predicted vehicle track orbit frame line overlap rate.
In addition, the likelihood of entering the parking frame may be determined based on the degree of progress of how much the predicted trajectory in the center in the width direction between the left and right rear wheels enters the target parking frame.
Other configurations are the same as those of the first embodiment.

(About operation and others)
An example of a time chart according to the processing of this embodiment is shown in FIG.
In this example, the travel controller 10 detects an entry operation to the parking frame L0 based on the positional relationship between the predicted trajectory of the host vehicle MM and the parking frame L0.
In the example shown in FIG. 38, when a parking frame L0 having a preset certainty (certainty) is detected (t1) and the vehicle speed is equal to or lower than a preset set speed (t2), the parking frame L0 is moved to the parking frame L0. Determine the approach operation. In the example shown in FIG. 38, it is detected that the own vehicle predicted track frame line overlap rate is greater than or equal to a preset value (t3), and the own vehicle expected track parking frame entrance overlap rate is greater than or equal to a preset value. Then (t7), it will be determined as an approach operation to the parking frame L0, and an acceleration suppression operation state is set.

  When the driver performs an accelerator operation in this acceleration suppression operating state, an acceleration command value (throttle opening) corresponding to the accelerator operation is suppressed. Furthermore, when the acceleration operation amount is equal to or greater than the preset operation amount in the state where the acceleration suppression is performed (t8), the suppression amount of the acceleration command value is increased. In the present embodiment, as a result of performing acceleration suppression so as to suppress below a preset throttle opening, as shown in FIG. 32, the actual throttle opening is compared with before the accelerator operation amount exceeds the preset operation amount. Is suppressed small. As a result, acceleration suppression against erroneous operation of the accelerator pedal 19 by the driver is executed more effectively.

  In the present embodiment, the parking operation can be detected more accurately by performing the parking frame approach determination based on the own vehicle predicted track frame line duplication rate and the own vehicle expected track parking frame entrance duplication rate, and more drivability can be detected. It is possible to realize a support system with a small decrease in the amount.

(Effect of this embodiment)
According to this embodiment, in addition to the effect by 1st Embodiment, there exists the following effect.
(1) The travel controller 10 includes information on the steering angle of the host vehicle MM, the steering angular speed of the host vehicle MM, the vehicle speed of the host vehicle MM, and the shift position of the host vehicle MM, the frame line position of the parking frame L0, and the parking frame. Based on at least one piece of information on the entrance position of L0, the positional relationship between the predicted trajectory of the host vehicle MM and the parking frame L0 is detected, and parking is performed based on the positional relationship between the detected predicted trajectory of the host vehicle MM and the parking frame L0. An entry operation to the frame L0 is detected.

  By using information on the steering angle of the host vehicle MM, the steering angle of the host vehicle MM, the steering angular velocity of the host vehicle MM, the vehicle speed of the host vehicle MM, and the shift position of the host vehicle MM, the expected trajectory of the host vehicle MM is obtained. I can do it. Then, the positional relationship between the predicted trajectory of the host vehicle MM and the parking frame L0 is detected from the obtained predicted trajectory of the host vehicle MM and at least one information of the frame line position of the parking frame L0 and the entrance position of the parking frame L0. Thereby, the approach operation to the parking frame L0 of the host vehicle MM can be detected with higher accuracy.

(2) The travel controller 10 detects an entry operation to the parking frame L0 based on the degree of overlap between the predicted track of the host vehicle MM and the parking frame L0.
Accordingly, it can be detected that the host vehicle MM is approaching the parking frame L0 as the degree of overlap is larger, and therefore, the entry operation of the host vehicle MM to the parking frame L0 can be detected with higher accuracy.

(3) The travel controller 10 detects an entry operation to the parking frame L0 based on the degree of overlap between the predicted track of the host vehicle MM and the entrance L2 of the parking frame L0.
Based on the degree of overlap, it can be detected that the host vehicle MM is moving toward the parking frame L0. As a result, the entry operation of the host vehicle MM into the parking frame L0 can be detected with higher accuracy.

1 Ambient environment recognition sensor (imaging unit)
DESCRIPTION OF SYMBOLS 1a Light-receiving part 8 Wiper detection sensor 10 Travel controller 10A Ambient environment recognition information calculation part 110 Parking frame line information processing part 120 Parking symbol information processing part 120A Symbol extraction part 120B Symbol group extraction part 120C Serial number determination part 120D Parking frame estimation part 10B Own vehicle vehicle speed calculation unit 10C Steering angle calculation unit 10D Steering angular velocity calculation unit 10E Shift position calculation unit 10F Brake pedal operation information calculation unit 10G Accelerator operation amount calculation unit 10H Acceleration operation speed calculation unit 10I Acceleration suppression operation condition determination unit 10J Acceleration suppression Quantity calculation unit 10K Target throttle opening calculation unit ALVL Approach certainty level FLVL Parking line level SLVL Parking symbol level TLVL Total confidence level ARA1-4 Area FF used as overhead image Radial line attribute FR Solid object line attribute L0 Parking frame

In driving support that supports vehicle driving for parking, the accuracy of driving support for parking can be improved by improving the reliability of parking frame detection.
An object of this invention is to improve the reliability of parking frame detection in the condition where a parking frame is not recognized from the line which comprises a parking frame .

In order to solve the above problems, an aspect of the present invention provides an imaging unit that acquires a captured image including a road surface around the host vehicle, a line detection unit that detects a line located on the road surface from the captured image, and a line detection A determination unit that determines whether a line detected by the unit satisfies a predetermined parking frame line condition, a symbol extraction unit that extracts a symbol including at least a number from the captured image, and a plurality of symbols extracted by the symbol extraction unit In this case, a serial number determination unit that determines whether the numbers of these symbols are serial numbers and a line that satisfies the parking frame line condition are not detected, and the serial number determination unit determines that the serial numbers are serial numbers. A parking frame estimation unit that estimates that there is a parking frame at a predetermined position from a symbol including a serial number.

ADVANTAGE OF THE INVENTION According to this invention, the reliability of parking frame detection can be improved in the condition where a parking frame is not recognized from the line which comprises a parking frame .

Claims (4)

An acceleration operator that the driver operates to instruct acceleration;
An acceleration operation amount detector for detecting an acceleration operation amount of the acceleration operation element;
A braking / driving force control unit that causes the host vehicle to generate acceleration corresponding to the acceleration operation amount detected by the acceleration operation amount detection unit;
An imaging unit that acquires a captured image including a road surface around the vehicle;
A symbol extraction unit for extracting a symbol including at least a number from the captured image;
When a plurality of symbols are extracted by the symbol extraction unit, a serial number determination unit that determines whether the numbers of these symbols are serial numbers;
When it is determined that the serial number determination unit is a serial number, a parking frame estimation unit that estimates that there is a parking frame at a predetermined position from the symbol;
An acceleration suppression unit that reduces the acceleration controlled by the braking / driving force control unit when it is determined that the parking frame estimated by the parking frame estimation unit exists ahead of the traveling direction of the host vehicle;
A driving support apparatus comprising: An overhead image acquisition unit that converts the captured image acquired by the imaging unit into a bird's-eye view image;
The driving support according to claim 1, wherein the symbol extraction unit extracts only symbols whose aspect ratio of numbers in the symbols is equal to or less than a preset reference ratio based on the bird's-eye view image obtained by bird's-eye view conversion. apparatus.   The symbol extraction unit extracts a symbol that is not extracted by the symbol extraction unit other than a straight line in a surrounding area that is set in advance around the extracted symbol and based on the symbol position. The driving support device according to claim 1, wherein the symbol is not set as a determination target of the serial number determination unit. It has a parking lot position detection unit that detects the position of the parking lot,
When the serial number determination unit determines that the plurality of symbols extracted by the symbol extraction unit are located at the parking lot position based on the detection of the parking position detection unit, the numbers of these symbols are serial numbers. Even if it is not, it is regarded as a serial number, The driving assistance apparatus of any one of Claims 1-3 characterized by the above-mentioned.

Images