Method for determining a parking area layout

Abstract

The present disclosure provides a method of determining a parking area layout. The method comprises: determining, as a vehicle (200) moves in a parking area, environmental information relating to the vehicle's surroundings from sensing data provided by the vehicle's sensing system; extracting parameters associated with parking spaces identified from the environmental information; determining, based on at least the extracted parameters, layout models for the parking area, each layout model comprising a base parking space template; assigning (S9) to each determined layout model a confidence score indicative of how well the respective base parking space template matches the parameters associated with the parking spaces identified from the environmental information; and selecting (S10) as a current layout model for determining a layout for the parking area, the layout model for the parking area having the highest confidence score, wherein the selected layout model is updated over time based on the confidence scores assigned to it generated from sensing data collected as the vehicle moves.

Description

Technical Field

[0001] This disclosure relates to a computer-implemented method for determining the layout of a parking area. The parking area may be a parking lot or a zone of a parking lot. Such a method may be used, for example, by a vehicle in an automatic or partially automatic manner to determine the parking area layout and control the movement of the vehicle within the parking area based on the parking area layout. Background Technology

[0002] In the automotive industry, parking systems have been developed to control vehicles to perform parking maneuvers automatically or partially automatically. These systems rely on multiple sensors, such as vision sensors, lidar sensors, radar sensors, Global Navigation Satellite Systems (GNSS), and a perception stack, to gather information about the vehicle's surroundings and detect vacant or unoccupied parking spaces, also known as parking areas. Once a vacant parking space is detected and confirmed, the parking system begins planning the parking maneuver and then at least partially controls the vehicle to execute the planned parking maneuver.

[0003] Before planning a parking maneuver, the parking system accumulates occupancy evidence from the perception stack. This occupancy evidence is post-processed to detect gaps between occupied spaces and classify the detected gaps as parking spaces. After detecting a parking space, the occupancy status of that space is tracked, and if the parking space is unoccupied, a path is planned to execute the parking maneuver to park the vehicle in the unoccupied space.

[0004] Existing parking systems are effective in controlling vehicles to perform automatic or partially automatic parking maneuvers. However, they suffer from long waiting times. The algorithms implemented for planning parking maneuvers must wait to confirm that detected gaps are unoccupied and can accommodate the vehicle, which can typically only be performed after the gap has been traversed. As a result, the total execution time of parking maneuvers controlled by the parking system is significantly increased.

[0005] This disclosure allows for improvements to the situation. In particular, there is a need for ease of vehicle maneuverability in parking areas. Summary of the Invention

[0006] This disclosure relates to a computer-implemented method for determining the layout of a parking area, the method comprising the following steps:

[0007] When the vehicle moves in the parking area, environmental information related to the vehicle's surrounding environment is determined based on sensing data provided by the vehicle's sensing system.

[0008] The layout for the parking area is determined by performing the following steps:

[0009] Extract one or more parameters associated with the parking space identified from the environmental information;

[0010] One or more layout models for the parking area are determined based on at least one or more extracted parameters, each layout model including an underlying parking space template;

[0011] Assign (S9) a confidence score to each determined layout model, the confidence score indicating how well the corresponding underlying parking space template matches the one or more parameters associated with the parking space identified from the environmental information; and

[0012] The layout model with the highest confidence score for the parking area is selected as the current layout model for determining the layout for the parking area.

[0013] The selected layout model is updated over time based on a confidence score assigned to it, generated from sensor data collected during vehicle movement.

[0014] The determined parking area layout can be used to control one or more functions of the vehicle. For example, the determined parking area layout can be used by the vehicle's path planning module to better plan the vehicle's trajectory within the parking area. The vehicle can be controlled to move along the planned trajectory automatically or at least partially automatically. This allows for faster, smoother, and safer maneuvering of the vehicle within the parking area.

[0015] Advantageously, the determined layout for the parking area is used to control the vehicle's driving trajectory and / or advanced driver assistance system (ADAS) control functions.

[0016] For example, the method may also include the step of controlling the vehicle to move automatically or partially automatically along a driving trajectory planned using the determined parking area layout.

[0017] In an implementation, the method may further include the following steps: assigning an occupancy status score indicating the probability of occupancy of the parking space to each of the plurality of parking space units in the layout of the parking area, and updating the occupancy status score based on the environmental information.

[0018] In an implementation, unless the environmental information provides an indication of occupancy for the parking space unit, a score indicating an unoccupied state may be assigned by default to each of the multiple parking space units in the layout of the parking area.

[0019] Advantageously, the method may further include the steps of identifying unoccupied parking spaces based on the occupancy status score assigned to each of the plurality of parking space units, and determining a parking maneuver to park the vehicle in the identified unoccupied parking space. The method may also include the step of controlling the vehicle to perform the predetermined parking maneuver in an automatic or partially automatic manner.

[0020] Identifying unoccupied parking spaces by using a parking area layout constructed based on a layout model allows for a reduction in parking maneuver waiting time. In other words, it significantly reduces the execution time of parking maneuvers performed by the vehicle in an automatic or partially automatic manner. Unoccupied parking space detection can be performed earlier than in existing technologies because it is no longer necessary to pass through unoccupied parking spaces to initiate the parking maneuver. Furthermore, parking maneuvers can be significantly simplified by planning a smoother and simpler trajectory for the vehicle.

[0021] Another advantage of this method is that it does not require specific sensors and / or specific sensing devices.

[0022] In an implementation, the extracted parameters associated with the parking space identified from the environmental information include at least one of the following elements:

[0023] The position of the identified parking space relative to the vehicle;

[0024] One or more dimensions of the identified parking spaces;

[0025] The location of the identified parking space.

[0026] In one implementation, updating the layout of the parking area over time includes generating a first parking space unit representing parking spaces identified from the environmental information over time, and predicting a second parking space unit corresponding to a foreseen parking space based on the parking spaces identified from the environmental information and the current layout model.

[0027] By adding predicted or extrapolated parking spaces based on a layout model and already identified parking spaces, parking area layouts can extend beyond the identified or detected parking spaces. This allows for earlier estimation of parking space locations, which can significantly reduce the complexity of parking maneuvers by planning simpler trajectories. There is no guarantee that a predicted or anticipated parking space exists and is not occupied. However, the parking space may be reconfirmed or unconfirmed later, once environmental information about the parking space is obtained.

[0028] Advantageously, the step of predicting the second parking space unit will be set by or added to the layout for the parking area by parking space units outside the sensing area covered by the sensing system of the vehicle.

[0029] This disclosure also relates to a computer program including instructions that, when executed by a computer, cause the computer to perform a previously defined method; a computing system including means adapted to perform the steps of the method; and a carrier including the computing system. Attached Figure Description

[0030] Other features, objects, and advantages of this disclosure will become clearer from reading the detailed description of non-limiting embodiments with reference to the accompanying drawings.

[0031] Figure 1 The image schematically illustrates a vehicle according to an embodiment, which includes a calculation system for determining the layout of parking areas.

[0032] Figures 2A to 2B A flowchart illustrating a method for determining the layout of a parking area according to an embodiment is shown.

[0033] Figure 3 Examples of using Figures 2A to 2B An example of a parking area layout determined by this method. Detailed Implementation

[0034] Figure 1 A schematic functional block diagram of vehicle 200 is shown.

[0035] The vehicle 200 includes a sensing system 300 responsible for sensing data in the environment surrounding the vehicle 200. The sensing system 300 may include multiple sensors, preferably of different types, such as a radar sensor 301, a vision sensor 302, an ultrasonic sensor 303, and / or a lidar sensor 304. The vision sensor 302 may, for example, include one or more forward vision sensors 302a and one or more surround vision sensors 302b.

[0036] Optionally, the sensing system 300 may also include a semantic layer 305, a semantic function, or a semantic module, which is responsible for determining semantic data based on the data provided by sensors 301 to 304. The semantic data includes semantic information related to the environment surrounding the vehicle 200; in other words, information in semantic form. This may involve, for example, traffic signs, traffic regulations, prohibited parking spaces, etc. The semantic layer 305 can be implemented using both software and hardware.

[0037] The vehicle 200 may also include a perception stack 400 or a perception module. The perception stack 400 is responsible for using sensing data provided by the sensing system 300 to determine environmental information related to the surrounding environment of the vehicle 200. It can implement algorithms and / or applications and / or functions for processing sensing data from the sensing system 300 to determine environmental information related to the surrounding environment of the vehicle 200. The perception stack 400 may have multiple perception layers, each implementing corresponding algorithms and / or applications and / or functions for processing sensing data to determine various information about the surrounding environment of the vehicle 200. For example, the perception stack 400 may include:

[0038] Occupancy grid map generator 401, which is used to generate an occupancy grid map of the surrounding environment of vehicle 200 based on sensing data; and

[0039] Object tracker 402 is used to identify and track dynamic objects in the surrounding environment of vehicle 200 based on sensing data.

[0040] The perception stack 400 can be implemented using both software and hardware. The perception layer can use algorithms such as machine learning algorithms.

[0041] In this disclosure, vehicle 200 also includes system 500 for tracking parking areas. A parking area may be a parking lot or a section or portion of a parking lot. Parking area tracking system 500 has the function of determining the layout for the parking area. Optionally, system 500 may have another function of determining unoccupied or vacant parking spaces, also known as parking spaces, which are locations designated for parking within the parking area.

[0042] A parking area layout is a schematic representation of a parking area, such as a geometric representation. It includes parking space units representing individual parking spaces within the parking area. Each parking space unit may have a predetermined geometric shape, such as a U-shape or rectangle. The parking area layout may also include one or more aisle units representing passageways or traffic lanes, designated for movement within the parking area or zone.

[0043] The parking area tracking system 500 may include a parking area layout tracker 501 and a parking space tracker 502.

[0044] The parking area layout tracker 501 is responsible for determining and updating or tracking the layout of the parking areas over time, as described later in this method.

[0045] It should be noted that different parking spaces can be used in different parking areas or zones. For example, perpendicular parking can be used in the main parking area, and parallel parking can be used as a supplement in the limited parking area. The parking area layout tracker 501 can use the specific arrangement of parking spaces to determine the layout for each parking area or zone.

[0046] The parking space unit tracker 502 is responsible for identifying and updating or tracking the parking area layout of parking space units over time, as described later in this method.

[0047] The parking area tracking system 500 may also include or be connected to a database 550 of a predetermined parking lot layout model. Most parking lots are designed or arranged according to known parking layouts. There are at least three main types or categories of parking: parallel, perpendicular, and angled. In the parallel type, each parking space is parallel to the direction in which the vehicle approaches the parking space. In the perpendicular type, each parking space is perpendicular to the direction in which the vehicle approaches the parking space. In the angled type, each parking space forms an angle A with the direction in which the vehicle approaches the parking space, determined by... Figure 1 The arrows in the diagram indicate this. Depending on the angle, angled parking spaces can have various variations. For acute angle A, the most common angles used in angled parking spaces are 30°, 45°, and 60°.

[0048] Figure 1 Five different predetermined parking lot layout models of different types are shown, including a parallel type, a perpendicular type, and three angular types for three corresponding acute angles A of 30°, 45°, and 60°. For each parking type, the database 550 may also include multiple variations defined by different parking space dimensions. Parking space dimensions include, for example, parking space width and parking space length. Each parking lot layout model can be associated with a parking space template. In other words, each parking layout model may include an underlying parking space template. Each parking space template can be characterized by multiple features, including at least one of the following:

[0049] Parking angle (e.g., 0°, 30°, 45°, 60°, or 90°);

[0050] Parking space width;

[0051] Parking space length.

[0052] The length of a parking space can be the available length of the parking space; in other words, it is the effective length that can be used for parking vehicles.

[0053] Optionally, the parking area tracking system 500 may also include a semantic processing module 503, which has the function of collecting semantic information related to the surrounding environment of the vehicle 200 from sensors 301 to 304 and / or from the semantic layer 305 and / or from the perception stack 400, and the function of assigning specific semantic information to parking spaces and optionally to layouts. For example, if a special parking space such as a disabled parking space is sensed, the semantic processing module 503 assigns the information "disabled parking space" to that parking space. The parking space tracker 502 can be configured to refuse parking based on the information "disabled parking space", even if the parking space is vacant or unoccupied, because the vehicle 200 is not authorized to park in such a special parking space.

[0054] The parking area tracking system 500 can be implemented using both software and hardware.

[0055] Vehicle 200 can be an autonomous vehicle or at least a partially automated vehicle.

[0056] In an embodiment, vehicle 200 may also include an automatic parking system 600 configured to at least partially control the movement of vehicle 200 within a parking area and / or perform parking actions to stop at a parking position. The automatic parking system 600 may use information or data provided at the output of the parking area tracking system 500, as explained in the methods described later. The automatic parking system 600 may be implemented using both software and hardware.

[0057] The perception stack 400, the parking area tracking system 500, and / or the automatic parking system 600 can be implemented on the same processor or controller, or on different processors or controllers.

[0058] Now refer to Figures 2A to 2B and Figure 3 This describes a computer-implemented method for determining a parking layout for a parking area, according to an implementation method. The parking area can specify a parking lot or a zone or sub-zone of a parking lot.

[0059] In step S0, we consider the vehicle 200 entering the parking area 700 and moving within the parking area 700.

[0060] In step S1, the sensing system 300 of the vehicle 200 uses sensors 301 to 304 to collect information related to the surrounding environment of the vehicle 200 in real time. In other words, the sensing system 300 uses the sensors 301 to 304 of the vehicle 200 to monitor and sense the surrounding environment or surrounding area of ​​the vehicle in real time. Step S1 can be performed while the vehicle 200 is moving in the parking area 700.

[0061] In step S2, the sensing system 300 outputs sensing data. The sensing data may include raw or unprocessed data and / or preprocessed data provided by sensors 301 to 304. Optionally, the sensing data may include semantic data related to the surrounding environment of the vehicle 200 provided by the semantic layer 305. Optionally, one or more sensors 301 to 304 may have the ability to determine semantic information based on the sensing data. The sensing data contains information related to the surrounding environment of the vehicle 200.

[0062] At least a portion of the sensed data can be received as input by the sensing stack 400 or the sensing module. In step S3, the sensing data can be processed by the sensing stack 400 to determine environmental information related to the surrounding environment of the vehicle 200.

[0063] In this implementation, in step S3, the occupancy grid map generator 401 can generate an occupancy grid map related to the environment surrounding the vehicle 200. The occupancy grid map represents the area surrounding the vehicle 200 as a field of uniformly spaced binary random variables, each binary random variable representing the presence of an obstacle at that location within that area.

[0064] In step S3, the object tracker 402 can identify and track dynamic objects in the environment surrounding the vehicle 200.

[0065] The perception stack 400 continuously determines and updates environmental information in real time based on the sensing data continuously provided by the sensing system 300.

[0066] In receiving step S4, the parking area tracking system 500 obtains or receives environmental information related to the surrounding environment of the vehicle 200 provided by the perception stack 400.

[0067] Alternatively or concurrently, in receiving step S4, the parking area tracking system 500 may directly receive at least a portion of the sensing data from the sensing system 300 and process the received sensing data to determine environmental information related to the environment surrounding the vehicle 200.

[0068] In the implementation, in step S5, the semantic processing layer 503 of the parking area tracking system 500 can collect semantic information from sensors 301 to 304 and / or semantic layer 305 and / or perception stack 400, and provide it to the parking area layout tracker 501 and parking space tracker 502.

[0069] In step S6, the parking area layout tracker 501 can identify or detect parking spaces in the environment surrounding the vehicle 200 based on environmental information obtained from the perception stack 400 and / or the semantic processing layer 503 and / or the sensing system 300. The tracker 501 can extract one or more parameters associated with the parking spaces identified from the environmental information through data analysis. For example, the parking area layout tracker 501 can identify or detect pavement markings demarcating parking spaces and / or parked vehicles and / or occupied and / or unoccupied parking spaces based on the obtained environmental information. For each identified parking space, the extracted parameters associated with the identified parking space may include at least a portion of the following features: parking space position relative to the vehicle 200, parking space width, parking space length, parking orientation (e.g., parallel, perpendicular, or angular), and / or parking angle. Some of the extracted parameters may not be very accurate and are quantified by a range of values.

[0070] Then, in step S7, the parking area layout tracker 501 determines one or more parking layout models for the parking area based on one or more extracted parameters. As previously mentioned, each layout model includes an underlying parking space template in the database 550. The extracted parking space parameters associated with one or more parking spaces identified in step S6 can be used as a filter or selection criterion for obtaining one or more parking layout models from the database 550. Selecting one or more parking layout models from the database 550 can therefore be performed by selecting only parking layout models whose underlying parking space template at least partially matches the extracted parameters. Parking layout models from the database 550 whose parking space templates do not match any parameters extracted from the environmental information can be excluded from the selection.

[0071] Optionally, in step S8, the parking area layout tracker 501 can obtain map information related to the current location of the vehicle 200. In step S7, this map information can be used as an additional filter or selection criterion to select one or more parking layout models from the database 550. This map information can be received from the vehicle 200 via an online map platform connected to a mobile communication network and / or from a map application installed on the vehicle 200.

[0072] Optionally, one or more parking layout models for the parking area can be further determined based on semantic information related to the vehicle's surrounding environment.

[0073] In step S9, the parking area layout tracker 501 assigns confidence scores to each parking layout model determined in step S7. The confidence score assigned to each parking layout model indicates how well the corresponding underlying parking space template matches one or more parameters extracted and associated with the parking space identified from the environmental information. In other words, the confidence score assigned to each parking layout model quantifies the correspondence, i.e., the match, between the parking space template of that parking layout model and one or more extracted parameters. The calculation of the confidence score can be performed by a computational algorithm, such as one based on an algorithm called "intersection over union," which gives a similarity value or coefficient from 0 to 1 and, for object detection in an image, measures the similarity between two sample sets or objects. The computational algorithm can measure the similarity between the parking space template and the identified parking space in terms of orientation and / or size. For each parking layout model, a similarity coefficient can be calculated for each identified or detected parking space. An overall score, which gives the confidence score assigned to that parking layout model, can then be calculated by combining the scores obtained for the multiple identified parking spaces.

[0074] The method further includes step S10, performed by parking layout tracker 501: selecting and updating the layout model of the parking area over time based on confidence scores assigned to one or more parking layout models. In step S10, parking layout tracker 501 selects the layout model of the parking area with the highest confidence score as the current layout model. The selected layout model is updated over time by parking layout tracker 501 based on a confidence score assigned to that layout model, generated from sensing data collected during vehicle movement.

[0075] In step S10, the parking layout tracker 501 uses the current parking layout model to construct or determine the parking area layout. In some embodiments, all features of the parking area layout can be derived from the current parking layout model. In other embodiments, some features of the parking area layout can be derived from parameters of the extracted and identified parking spaces. For example, the parking area layout can be based on a basic parking layout model, such as... Figure 1 One of the five models shown, but with parking space width and length determined based on the extracted parameters.

[0076] Step S10 can be executed in real time and continuously.

[0077] Steps S1 to S10 are executed in real time and continuously. As the vehicle 200 moves within the parking area 700, its sensing system 300 continuously collects information about the vehicle 200's surrounding environment. Thus, the parking area tracking system 500 acquires more and more environmental information related to the vehicle 200's surroundings over time. As time progresses, more and more parking spaces can be identified in step S6, and the extracted parking space parameters become increasingly accurate. In step S9, the confidence scores assigned to one or more parking layout models are updated over time and become increasingly reliable. This allows the parking area layout tracker 501 to converge to the correct parking layout model over time.

[0078] Figure 3 An example representing a parking area layout 800 includes multiple parking space units 801, 802… representing parking spaces in parking area 700. Figure 3 In the diagram, the parking area layout is overlaid on the occupancy grid. In this implementation, the parking space unit may include:

[0079] - The first parking space unit represents the parking space identified by the parking area layout tracker 501 based on the acquired environmental information, and

[0080] - The second parking space unit corresponds to the parking area layout tracker 501, which anticipates, predicts, or extrapolates parking spaces based on parking spaces identified from the obtained environmental information and the current parking layout model.

[0081] The prediction or extrapolation of the second parking space unit can be adjusted using environmental information and / or map information (such as road curves) and / or the presence of buildings in the area, and / or the presence of walls at a certain distance of 200 from the vehicle.

[0082] Therefore, step S10 may include:

[0083] - Generate the first parking space unit representing the parking spaces identified from the obtained environmental information;

[0084] -Predict the anticipated or extrapolated second parking space unit based on the parking spaces already identified from the obtained environmental information and the current parking layout model.

[0085] Predicting a second parking space unit may include adding or setting a parking space unit located outside the sensing area covered by the sensing system 300 of the vehicle 200 in the parking area layout 800.

[0086] The determined parking area layout can be used to control one or more functions of vehicle 200. In some embodiments, the determined parking area layout can be used to control the driving trajectory of vehicle 200 and / or advanced driver assistance system (ADAS) control functions. For example, parking area layout 800 can be used by automatic parking system 600 and / or automatic driving system of vehicle 200 to control vehicle 200 to move in parking area 700 automatically or partially automatically. Therefore, the method may include the step of controlling one or more functions of vehicle 200 moving in parking area 700 automatically or partially automatically by using parking area layout 800. This allows for facilitating and ensuring the movement of vehicle 200 in parking area 700.

[0087] In an implementation, the method may further include step S11: assigning an occupancy status score to each parking space unit 801, 802... of the parking area layout 800, and updating the occupancy status score based on the obtained environmental information. Step S11 can be executed continuously in real time by the parking space tracker 502. The occupancy status score of each parking space unit 801, 802... of the parking area layout 800 indicates the occupancy probability of the corresponding parking space in the parking area 700.

[0088] In a particular implementation, a score indicating an unoccupied state may be assigned by default to each parking space unit 801, 802... in the parking area layout 800, unless the obtained environmental information provides an indication or evidence of occupancy for the corresponding parking space. An occupancy indication may be an indication that a vehicle is parked in a parking space, or that an object is detected in a parking space, or that part of a vehicle is detected in a parking space.

[0089] The occupancy grid map generated by generator 401 can also provide evidence about vacant or unoccupied spaces in a specific area. Parking space tracker 502 can rely on such information to determine whether the area can be considered unoccupied, unknown, or occupied, and assign an occupancy status score to the parking space unit.

[0090] In optional step S12, the parking space tracker 502 may also use the output of the semantic processing module 503 to assign special information to the parking space units of the parking layout 800, such as information that the parking space is for disabled persons.

[0091] In step S13, the parking space tracker 502 may optionally classify the parking space units 801, 802... of the parking area layout 800 into different categories based on the scores assigned to the parking space units and, optionally, special information received from the semantic processing module 503. The different categories may include at least a portion of the following: unoccupied parking spaces, occupied parking spaces, unknown status, and / or special parking spaces, such as parking spaces for disabled persons.

[0092] Based on continuously acquired environmental information, steps S11 and optional steps S12 and S13 are executed in real time and continuously. Therefore, based on continuously acquired environmental information, the occupancy status scores of parking space units 801, 802... in the parking area layout 800 are updated in real time and continuously.

[0093] The automatic parking system 600 can be activated by the user, for example, when the vehicle 200 enters the parking area 700.

[0094] In optional step S14, the automatic parking system 600 can control the vehicle 200 to move in the parking area automatically or at least partially automatically based on the parking area layout 800 to search for unoccupied parking spaces suitable for parking.

[0095] The method may further include step S15, performed by parking space tracker 502: identifying at least one unoccupied parking space in parking area 700 suitable for parking, based on the occupancy status score of each parking space unit among a plurality of parking space units allocated to parking area layout 800 and / or based on the classification results of the parking space units performed in step S13. This can be performed by comparing the occupancy status score of each parking space unit with a predetermined threshold and / or based on the parking space category generated by classification.

[0096] In step S16, after identifying or detecting at least one unoccupied parking space suitable for parking in the parking area 700 in step S13, the automatic parking system 600 can set or plan parking maneuvers to park the vehicle 200 in the unoccupied parking space. If multiple unoccupied parking spaces suitable for parking have been detected, the automatic parking system 600 can select one of them, such as the unoccupied parking space closest to the vehicle 200.

[0097] Then, in step S17, the automatic parking system 600 can control the vehicle 200 to perform the parking maneuver set or planned in step S16 in an automatic or partially automatic manner to park the vehicle 200 in an unoccupied parking space. The parking maneuver may include at least the movement of the vehicle 200, which includes approaching an unoccupied parking space intended as a target parking space. As the vehicle 200 approaches the target parking space, the parking area layout and occupancy status score of the parking space unit are updated in real time and continuously. It is possible that the target parking space used for the parking maneuver may ultimately be identified as an occupied parking space. In this case, the automatic parking system 600 repeats steps S13 and S15 to park the vehicle 200 in another parking space.

[0098] The parking area layout can be displayed on the display device of vehicle 200. It can be overlaid on a map representing the environment of vehicle 200. The parking area layout displayed on the display device can be used by the driver to facilitate maneuvering in the parking area 700.

[0099] The method may also include the step of generating a map based on the parking area layout. This map can be displayed on the vehicle's display device. Parking space units in the parking area layout can be categorized into three classes based on their occupancy status scores: unoccupied, occupied, and unknown. These three categories can be represented by different colors on the display device.

[0100] This disclosure also relates to:

[0101] A computer program including instructions that, when executed by a computer, cause the computer to perform a previously defined method;

[0102] A computer-readable data medium having a computer program as defined above stored thereon;

[0103] A computing system having means suitable for performing the steps of the aforementioned method;

[0104] A vehicle that includes the aforementioned computing system.

Claims

1. A computer-implemented method for determining the layout of a parking area (700), the method comprising the steps of: When the vehicle (200) moves in the parking area (700), environmental information related to the surrounding environment of the vehicle (200) is determined (S3, S4, S5) based on sensing data provided by the vehicle's sensing system. The step of determining the environmental information includes at least one of the following: generating an occupancy grid map of the vehicle's environment, performing object tracking in the vehicle's environment, obtaining semantic information of the vehicle's environment, and performing semantic classification to divide parking spaces into different categories. The layout (800) for the parking area (700) is determined by performing the following steps: Data analysis extracts (S6) one or more parameters associated with the parking space identified from the environmental information, wherein the one or more parameters include at least one of the following elements: the position of the identified parking space relative to the vehicle, one or more dimensions of the identified parking space, and the orientation of the identified parking space. At least based on one or more extracted parameters, one or more layout models for the parking area are determined (S7), each layout model including an underlying parking space template, wherein one or more extracted parameters are used to select one or more predefined parking area layout models obtained from the database (550) to determine (S7) the one or more layout models for the parking area; Assign (S9) a confidence score to each of the determined layout models, the confidence score indicating how well the corresponding underlying parking space template matches the one or more parameters associated with the parking space identified from the environmental information; The layout model with the highest confidence score for the parking area (S10) is selected as the current layout model for determining the layout (800) for the parking area (700); and The selected layout model is updated over time based on the confidence score assigned to the layout model and generated from the sensing data collected while the vehicle is moving.

2. The computer-implemented method according to claim 1, wherein, The layout (800) determined for the parking area (700) is used to control one or more functions of the vehicle.

3. The computer-implemented method according to claim 1 or 2, wherein, The layout (800) determined for the parking area (700) is used to control the driving trajectory of the vehicle and / or the advanced driver assistance system (ADAS) control functions.

4. The computer-implemented method according to claim 1, the method further comprising the step (S11): assigning an occupancy status score indicating the occupancy probability of the parking space to each of the plurality of parking space units (801, 802...) in the layout (800) for the parking area (700), and updating the occupancy status score based on the environmental information.

5. The computer-implemented method according to claim 4, wherein, Unless the environmental information provides an indication of the occupancy of the parking space unit, by default, a score indicating an unoccupied state is assigned to each of the multiple parking space units in the layout (800) for the parking area (700).

6. The computer-implemented method according to claim 4, the method further comprising the step (S13): identifying unoccupied parking spaces based on the occupancy status score assigned to each of the plurality of parking space units, and the step (S15): determining a parking maneuver to park the vehicle (200) in the identified unoccupied parking space.

7. The computer-implemented method according to claim 1, wherein, The layout (800) for the parking area (700) is updated over time (S10) by generating a first parking space unit representing parking spaces identified from the environmental information, and predicting a second parking space unit corresponding to a foreseen parking space based on the parking spaces identified from the environmental information and the current layout model.

8. The computer-implemented method according to claim 7, wherein, The step of predicting the second parking space unit includes setting the parking space unit outside the sensing area covered by the sensing system of the vehicle within the layout (800) for the parking area (700).

9. The computer-implemented method according to claim 1, wherein, The step (S7) of determining one or more layout models includes obtaining the one or more layout models from a database (550) that stores multiple predefined layout models.

10. The computer-implemented method according to claim 1, wherein, The step of determining (S7) one or more layout models for the parking area is also based on semantic information related to the surrounding environment of the vehicle (200).

11. A computer-readable data medium having a computer program stored thereon, which, when executed by a computer, causes the computer to perform the method according to any one of claims 1 to 10.

12. A computing system comprising means for performing the steps of the method according to any one of claims 1 to 10.

13. A vehicle comprising the computing system according to claim 12.

Images