KR20240051804A - Automatic Parking Apparatus and Method for Automatic Parking based on Parking Area Environment Recognition - Google Patents

Abstract

Disclosed is an automatic parking device and method based on parking area environment recognition. An automatic parking method based on parking area environment recognition according to an embodiment includes the steps of searching for a parking space, recognizing the environment of the parking space and determining whether rear parking is possible, and detecting at least one other vehicle located in the parking space. It may include a step of recognizing, and a step of setting a parking target position based on the result of recognizing whether rear parking is possible and at least one other vehicle.

Description

Automatic Parking Apparatus and Method for Automatic Parking based on Parking Area Environment Recognition}

The described embodiments relate to automated parking technology for autonomous vehicles.

Autonomous parking technology is a system that uses sensors mounted on the vehicle to search for parking areas and automatically park and exit the vehicle.

The parking assistance system provides information such as parking space and parking trajectory and functions such as steering control while the driver is in the vehicle, and has been commercialized by many automobile companies. The automatic parking system is a technology that automatically operates steering and brakes to perform parking without driver intervention, and is currently being commercialized by various companies. The autonomous valet parking system is a technology that allows a vehicle to find a space in a parking lot and park through autonomous driving without a driver on board, and is currently under development.

Meanwhile, there are various etiquettes that are generally used when parking a vehicle. For example, if the area where you plan to park is adjacent to a glass building or flower bed, it is widely recommended to park in front to avoid pollution from vehicle exhaust.

In addition, when getting in and out of the car after parking, it is recommended to park with a little more space on the driver's side than in the center of the available parking space to avoid the so-called door locking, where the car's door collides with an adjacent car, and to make it easier for drivers of adjacent cars to get on and off.

However, to date, automatic parking technology for self-driving vehicles has focused on safe parking, and as described above, generally accepted parking etiquette is not applied.

The purpose of the described embodiment is to comply with generally accepted parking etiquette in automatic parking of self-driving vehicles and to improve the convenience of getting on and off the driver.

An automatic parking method based on parking area environment recognition according to an embodiment includes the steps of searching for a parking space, recognizing the environment of the parking space and determining whether rear parking is possible, and detecting at least one other vehicle located in the parking space. It may include a step of recognizing, and a step of setting a parking target position based on the result of recognizing whether rear parking is possible and at least one other vehicle.

At this time, the step of determining whether rear parking is possible may include the step of recognizing properties around the available parking space, and if the properties around the available parking space are residential or commercial spaces, determining that front parking is possible. there is.

At this time, the step of recognizing the properties of the available parking space can be recognized based on a semantic segmentation model using the vehicle's sensor data as input.

At this time, the step of recognizing another vehicle may include recognizing at least one other vehicle located in a parking available area, and recognizing the location and direction of at least one other vehicle.

At this time, in the step of recognizing at least one other vehicle, in the case of parallel parking, vehicles in front and behind the parking area can be recognized, and in the case of diagonal or perpendicular parking, vehicles parked on the left and right of the parking area can be recognized. .

At this time, in the step of recognizing at least one other vehicle, the other vehicle may be recognized based on object detection technology using a 3D bounding box.

At this time, in the step of setting the parking target position, the position of each driver's seat can be predicted based on the direction of the own vehicle and other vehicles, and the parking target position can be calculated by considering the free space at the predicted driver's seat position.

An automatic parking device based on parking area environment recognition according to an embodiment includes a memory in which at least one program is recorded, and a processor that executes the program, wherein the program searches for available parking spaces and recognizes the environment of the available parking spaces. Thus, it is possible to determine whether rear parking is possible, recognize at least one other vehicle located in the available parking space, and set a parking target location based on the results of recognizing whether rear parking is possible and at least one other vehicle.

At this time, in determining whether rear parking is possible, the program may perform the steps of recognizing attributes around a parking space and, if it is a residential or commercial space, determining that front parking is possible.

At this time, when recognizing the attributes of a parking space, the program can recognize the vehicle's sensor data based on a semantic segmentation model as input.

At this time, when recognizing other vehicles, the program may perform the steps of recognizing at least one other vehicle located in a parking area and recognizing the location and direction of at least one other vehicle.

At this time, when recognizing at least one other vehicle, the program recognizes vehicles in front and behind the parking area in the case of parallel parking, and recognizes vehicles parked on the left and right of the parking area in the case of diagonal or perpendicular parking. You can.

At this time, when recognizing at least one other vehicle, the program may recognize the other vehicle based on object detection technology using a 3D bounding box.

At this time, when setting the parking target position, the program can predict the position of each driver's seat based on the parking direction of the own vehicle and other vehicles, and calculate the parking target position by considering the free space at the predicted driver's seat position.

An automatic parking method based on parking area environment recognition according to an embodiment includes the steps of searching for a parking space, recognizing the surrounding environment attributes of the available parking space, and rear parking possible depending on whether the surrounding environment attributes are residential or commercial space. determining whether rear parking is possible, recognizing at least one other vehicle located in the parking area, recognizing the location and direction of at least one recognized other vehicle, whether rear parking is possible and the direction of the own vehicle and the other vehicle It may include predicting the position of each driver's seat based on the step of predicting the position of each driver's seat, and calculating the parking target position by considering the free space of the predicted driver's seat position.

At this time, the step of recognizing the properties of the available parking space can be recognized based on a semantic segmentation model using the vehicle's sensor data as input.

At this time, in the step of recognizing at least one other vehicle, the other vehicle may be recognized based on object detection technology using a 3D bounding box.

The described embodiment can improve the driver's convenience of getting on and off while complying with generally accepted parking etiquette in automatic parking of self-driving vehicles.

1 is a schematic block diagram of an automatic parking device based on parking area environment recognition according to an embodiment.
Figure 2 is a flowchart for explaining an automatic parking method based on parking area environment recognition according to an embodiment.
Figure 3 is a flowchart for explaining the step of recognizing the surrounding environment of a parking space according to an embodiment.
Figure 4 is an example of environment recognition based on semantic segmentation.
Figure 5 is an exemplary front parking view according to an embodiment.
Figure 6 is an example diagram of rear parking according to an embodiment.
Figure 7 is a flowchart for explaining steps for recognizing another vehicle according to an embodiment.
Figure 8 is an example diagram of perpendicular autonomous parking.
Figure 9 is an example diagram of diagonal autonomous parking.
Figure 10 is an example diagram of parallel autonomous parking.
Figure 11 is an example of object detection using a 3D bounding box.
Figure 12 is an example of predicting the location and direction of another vehicle according to an embodiment.
Figure 13 is a flowchart for explaining steps for recognizing another vehicle according to an embodiment.
Figure 14 is an example of free space creation according to an embodiment.
Figure 15 is an example of setting a parking target position considering free space according to an embodiment.
Figure 16 is a diagram showing the configuration of a computer system according to an embodiment.

The advantages and features of the present invention and methods for achieving them will become clear by referring to the embodiments described in detail below along with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and will be implemented in various different forms. The present embodiments only serve to ensure that the disclosure of the present invention is complete and that common knowledge in the technical field to which the present invention pertains is not limited. It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout the specification.

Although terms such as “first” or “second” are used to describe various components, these components are not limited by the above terms. The above terms may be used only to distinguish one component from another component. Accordingly, the first component mentioned below may also be the second component within the technical spirit of the present invention.

The terms used in this specification are for describing embodiments and are not intended to limit the invention. As used herein, singular forms also include plural forms, unless specifically stated otherwise in the context. As used in the specification, “comprises” or “comprising” implies that the mentioned element or step does not exclude the presence or addition of one or more other elements or steps.

Unless otherwise defined, all terms used in this specification can be interpreted as meanings commonly understood by those skilled in the art to which the present invention pertains. Additionally, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless clearly specifically defined.

FIG. 1 is a schematic block diagram of an automatic parking device based on parking area environment recognition according to an embodiment, and FIG. 2 is a flowchart for explaining an automatic parking method based on parking area environment recognition according to an embodiment.

Referring to FIG. 1 , an automatic parking device 100 based on parking area environment recognition according to an embodiment may be installed in an autonomous vehicle and perform an operation of automatically parking and unparking the vehicle.

At this time, the autonomous vehicle may be equipped with a sensor unit 10 that acquires surrounding information. The sensor unit 10 may include, for example, a camera 11, Lidar 12, and Radar 13.

Here, the camera 141 and LIDAR 142 can acquire images of the environment around the vehicle, and the radar 143 can acquire the location and distance of objects in the environment around the vehicle.

The parking area environment recognition-based automatic parking device 100 acquires information on the surrounding environment using the sensor unit 10 installed in the autonomous vehicle and performs automatic parking based on the obtained information.

The automatic parking device 100 based on parking area environment recognition may include, in detail, an environment recognition unit 110, a surrounding vehicle recognition unit 120, and a parking plan establishment unit 130.

The environment recognition unit 110 recognizes the environment surrounding the vehicle to perform autonomous driving or autonomous parking. In particular, parking requires more precise recognition compared to general road driving.

Depending on the embodiment, the environment recognition unit 110 searches for available parking spaces, recognizes the environment of the searched available parking spaces, and determines whether rear parking is possible.

That is, referring to FIG. 2, when the environment recognition unit 110 reaches the driving end position, it searches for a parking space based on sensor data detected by the sensor unit 10 (S210).

If a parking space exists as a result of the search at S210 (S220), the environment recognition unit 110 recognizes the environment surrounding the parking space (S230). A detailed description of this will be provided later with reference to FIGS. 3 to 6.

Next, the surrounding vehicle recognition unit 120 recognizes at least one other vehicle located nearby in the available parking space (S240). A detailed description of this will be provided later with reference to FIGS. 7 to 12.

Finally, the parking plan establishment unit 130 determines the parking target location based on whether rear parking is possible output from the environment recognition unit 110 and at least one other vehicle recognition result output from the surrounding vehicle recognition unit 120. Set (S250). A detailed description of this will be provided later with reference to FIGS. 13 to 15.

Then, if parking is not possible at the set target location, the parking plan establishment unit 130 proceeds to S210.

Figure 3 is a flowchart for explaining the steps of recognizing the surrounding environment of a parking space according to an embodiment, Figure 4 is an example of environment recognition based on semantic segmentation, and Figure 5 is an example of front parking according to an embodiment. Figure 6 is an example diagram of rear parking according to an embodiment.

Referring to FIG. 3, the environment recognition unit 110 recognizes attributes around a parking space (S231).

At this time, as shown in FIG. 4, the environment recognition unit 110 identifies each of the objects in the parking space based on a semantic segmentation model that inputs the camera 11 or lidar 12 mounted on the vehicle. recognize

Here, semantic segmentation is a technology that classifies the input of the sensor unit 10 into a defined class on a pixel basis, and can be used to predict the properties of a specific area.

The environment recognition unit 110 determines that front parking is possible when the properties surrounding the parking space recognized in S231 are recognized as residential or commercial spaces (S232). For example, as shown in FIG. 5 , if it is a residential or commercial space where a flower bed material 52 or a glass material exists around the parking space 51, it may be determined that front parking 50 is possible.

On the other hand, the environment recognition unit 110 determines that rear parking 60 is possible when the properties surrounding the available parking space 61 recognized in S231 are not recognized as residential or commercial spaces (S232) (S234 ). For example, as shown in FIG. 6, if the area around the available parking space 62 is not a residential or commercial space, it may be determined that rear parking is possible.

FIG. 7 is a flowchart for explaining steps for recognizing other vehicles according to an embodiment, FIG. 8 is an example of perpendicular autonomous parking, FIG. 9 is an example of oblique autonomous parking, and FIG. 10 is an example of parallel autonomous parking. 11 is an example of object detection using a 3D bounding box, and FIG. 12 is an example of predicting the position and direction of another vehicle according to an embodiment.

Referring to FIG. 7, the surrounding vehicle recognition unit 120 recognizes the presence of at least one other vehicle located in the parking area (S241 to S243).

At this time, as shown in FIG. 8, when the parking area allows parallel parking (S241), the surrounding vehicle recognition unit 120 recognizes other vehicles located in front and behind the parking area (S242).

On the other hand, when the parking area allows perpendicular parking as shown in FIG. 9 or when the parking area allows diagonal parking as shown in FIG. 10 (S241), the surrounding vehicle recognition unit 120 detects the parking area on the left and right of the available parking area. Recognize other vehicles located in (S243).

At this time, the surrounding vehicle recognition unit 120 detects other vehicles located around the parking area based on object detection technology using a three-dimensional bounding box, as shown in FIG. 11. recognize

Here, object detection is a technology that finds a target object in the form of a 3D box from an input sensor, and aims to simultaneously find the position and direction that the object occupies in 3D space.

Then, the surrounding vehicle recognition unit 120 predicts the location and direction of at least one other vehicle (S244). For example, as shown in FIG. 12, the positions and parking directions of other vehicles 81 and 82 located on the left and right of a parking area parking at a right angle are recognized.

FIG. 13 is a flowchart for explaining steps for recognizing other vehicles according to an embodiment, FIG. 14 is an example of free space creation according to an embodiment, and FIG. 15 is an example of setting a parking target position considering free space according to an embodiment. It is also a degree.

Referring to FIG. 13, the parking plan establishment unit 130 creates a free space by considering the parking direction of the own vehicle and at least one other vehicle (S251).

That is, the position of the driver's seat of each vehicle is predicted based on the parking direction of the own vehicle determined in the above-described S230 and the parking direction of at least one other vehicle nearby. For example, by Korean standards, it may be on the left.

Then, the parking plan establishment unit 130 creates a free space based on the predicted driver's seat position. For example, as shown in FIG. 14, a free space 91 on the left side of the driver's seat of another vehicle and a free space 92 on the left side of the driver's seat of the own vehicle are virtually created.

Through this, the driver is encouraged to park in a wider space at the predicted driver's seat location, allowing for smooth entry and exit of drivers of their own vehicle and other vehicles.

Then, the parking plan establishment unit 130 sets the final parking target position by considering the created free space (S252). That is, as shown in FIG. 15, the parking target position is calculated with an additional virtual space for the predicted driver's seat direction. As a result, they are encouraged to park at a location slightly further away from the driver's seats of adjacent vehicles than at the exact center of the parking area.

Figure 16 is a diagram showing the configuration of a computer system according to an embodiment.

Devices according to embodiments of the present invention may be implemented in a computer system 1000 such as a computer-readable recording medium.

Computer system 1000 may include one or more processors 1010, memory 1030, user interface input device 1040, user interface output device 1050, and storage 1060 that communicate with each other via bus 1020. You can. Additionally, the computer system 1000 may further include a network interface 1070 connected to the network 1080. The processor 1010 may be a central processing unit or a semiconductor device that executes programs or processing instructions stored in the memory 1030 or storage 1060. The memory 1030 and storage 1060 may be storage media that includes at least one of volatile media, non-volatile media, removable media, non-removable media, communication media, and information transfer media. For example, memory 1030 may include ROM 1031 or RAM 1032.

Although embodiments of the present invention have been described above with reference to the attached drawings, those skilled in the art will understand that the present invention can be implemented in other specific forms without changing the technical idea or essential features. You will understand that it exists. Therefore, the embodiments described above should be understood in all respects as illustrative and not restrictive.

Claims (17)

Searching for available parking spaces;
Recognizing the environment of the available parking space and determining whether rear parking is possible;
Recognizing at least one other vehicle located in a parking space; and
An automatic parking method based on recognition of a parking area environment, comprising the step of setting a parking target position based on whether rear parking is possible and a result of recognizing at least one other vehicle. The method of claim 1, wherein the step of determining whether rear parking is possible is:
Recognizing properties around available parking spaces; and
An automatic parking method based on recognition of the parking area environment, including determining that front parking is possible in the case of a residential or commercial space. The method of claim 2, wherein the step of recognizing the attributes of the available parking space includes:
An automatic parking method based on parking area environment recognition that recognizes vehicle sensor data based on a semantic segmentation model. The method of claim 1, wherein the step of recognizing another vehicle includes:
Recognizing at least one other vehicle located in the parking area; and
An automatic parking method based on parking area environment recognition, comprising the step of recognizing the location and direction of at least one other vehicle. The method of claim 4, wherein the step of recognizing at least one other vehicle includes:
In the case of parallel parking, vehicles in front and behind the parking area are recognized,
In the case of diagonal or perpendicular parking, an automatic parking method based on parking area environment recognition that recognizes vehicles parked on the left and right of the available parking area. The method of claim 4, wherein the step of recognizing at least one other vehicle includes:
An automatic parking method based on parking area environment recognition that recognizes other vehicles based on object detection technology using a 3D bounding box. The method of claim 4, wherein setting the parking target position comprises:
An automatic parking method based on parking area environment recognition that predicts the position of each driver's seat based on the direction of the driver's seat and other vehicles, and calculates the parking target position by considering the free space at the predicted driver's seat position. a memory in which at least one program is recorded; and
Contains a processor that executes a program,
The program is,
Searches for parking space, recognizes the environment of the parking space, determines whether rear parking is possible, recognizes at least one other vehicle located in the parking space, recognizes whether rear parking is possible and at least one other vehicle. An automatic parking device based on parking area environment recognition that sets the parking target location based on the results. According to claim 8, the program:
In determining whether rear parking is possible,
Recognizing properties around available parking spaces; and
An automatic parking device based on parking area environmental awareness that takes steps to determine if front parking is available in residential or commercial spaces. The method of claim 9, wherein the program:
An automatic parking device based on parking area environment recognition that recognizes the properties of available parking spaces based on a semantic segmentation model using the vehicle's sensor data as input. According to claim 8, the program:
In recognizing other vehicles, recognizing at least one other vehicle located in a parking area; and
An automatic parking device based on parking area environment recognition, which performs the step of recognizing the location and direction of at least one other vehicle. The method of claim 11, wherein the program:
In recognizing at least one other vehicle, in the case of parallel parking, vehicles in front and behind the parking area are recognized, and in the case of diagonal or perpendicular parking, parking area environment recognition is performed to recognize vehicles parked on the left and right sides of the parking area. based automatic parking device. The method of claim 11, wherein the program:
An automatic parking device based on parking area environment recognition that recognizes at least one other vehicle based on object detection technology using a 3D bounding box. The method of claim 11, wherein the program:
In setting the parking target location, the location of each driver's seat is predicted based on the parking direction of the vehicle and other vehicles, and the parking target location is calculated by considering the free space at the predicted driver's seat location. Automatic parking area environment recognition-based Parking device. Searching for available parking spaces;
Recognizing surrounding environmental attributes of available parking spaces;
determining rear parking availability depending on whether the surrounding environmental attributes are residential or commercial spaces;
Recognizing at least one other vehicle located in the parking area;
Recognizing the location and direction of at least one other recognized vehicle;
Predicting the position of each driver's seat based on rear parking availability and the direction of the vehicle and other vehicles; and
An automatic parking method based on parking area environment recognition, including the step of calculating a parking target position by considering the free space of the predicted driver's seat position. The method of claim 15, wherein the step of recognizing attributes of available parking spaces includes:
An automatic parking method based on parking area environment recognition that recognizes vehicle sensor data based on a semantic segmentation model. The method of claim 15, wherein the step of recognizing at least one other vehicle includes:
An automatic parking method based on parking area environment recognition that recognizes other vehicles based on object detection technology using a 3D bounding box.

Images