DE102024201954A1 - Method for mapping a parking space - Google Patents

Abstract

The invention relates to a method for mapping a parking space, comprising at least one vehicle (2) which has at least one vehicle sensor system (4), and a controller (28) which is signal-coupled to the at least one vehicle (2) and has a digital map (31), wherein a sensor signal (19) is generated by the vehicle sensor system (4), wherein the sensor signal (19) is examined for at least one secondary parking space indicator, wherein, if at least one secondary parking space indicator is present, a parking space hypothesis (23a) with a confidence measure (23b) is generated, wherein the parking space hypothesis (23a) and the confidence measure (23b) are transmitted to the controller (28), and wherein, depending on the confidence measure (23b), the parking space hypothesis (23a) is mapped as a parking space (37) in the digital map (31).

Description

The invention relates to a method for mapping a parking space, comprising at least one vehicle having at least one vehicle sensor system, and a controller having a digital map and signal-coupled to the at least one vehicle. The invention further relates to a device for implementing such a method.

A "parking stall" is understood here and below to mean, in particular, a device or facility specifically designed for temporarily parking a vehicle, in particular a motor vehicle. The parking stall may also have additional features, such as anchoring devices, charging devices for electric vehicles, or integrated security systems. A parking stall within the meaning of this application is understood to mean, in particular, a parking space or parking space.

A "parking space" is understood here and below to mean a specifically designated or marked (parking/parking) area intended for the temporary parking of vehicles, especially motor vehicles. This area can be located outdoors or in covered structures such as parking garages or underground car parks. Parking spaces can have various features tailored to the needs of different user groups and vehicle types, such as designations for small cars, trucks, disabled parking spaces, or charging stations for electric vehicles. They can be designed both as part of public infrastructure and as private facilities.

A “parking space” is understood here and below to mean, in particular, a parking space or parking area, which is designed as a gap or a distance, i.e. as a free space on a public road next to a neighboring vehicle or between two neighboring vehicles, which is large enough to allow a motor vehicle to be positioned therein and which allows the motor vehicle to be parked temporarily or permanently.

"Mapping" or "mapping" here and below refers specifically to the process of capturing, analyzing, and displaying spatial data and information on a two-dimensional or three-dimensional surface of a digital map. In modern applications such as autonomous vehicles, such mapping is a key element for navigation and understanding the environment, providing a basis for decision-making and route planning. Mapping can also integrate advanced aspects such as real-time data processing, machine learning, and artificial intelligence to produce adaptive and dynamic maps that adapt to changing conditions.

Motor vehicles, particularly road vehicles such as automobiles, are frequently used to transport or convey people and freight. After use, the motor vehicle is usually parked or left in a parking space, i.e., in a parking lot or parking area, such as a parking garage. Parking or leaving a motor vehicle is understood here and below to mean, in particular, a stationary traffic process of the motor vehicle, in particular a deliberate interruption in the journey, which is not caused by the traffic situation or an order. In particular, this means an interruption in the journey during which the motor vehicle user leaves the motor vehicle.

From the DE 10 2017 204 455 A1 , the DE 10 2019 201 893 A1 , the DE 10 2015 209 147 A1 and the DE 10 2014 217 567 A1 For example, it is known to create digital maps for parking spaces (e.g., a map for a parking garage) from the sensor data of a vehicle sensor system. This involves entering or mapping parking spaces, especially parking spaces or parking spaces, into a digital map in order to later make them available to useful functions (e.g., a navigation function).

There are various approaches to mapping parking spaces. For example, camera systems, particularly side-view cameras in the vehicle's side mirrors of an area-view or top-view system, can be used as vehicle sensor systems to directly detect parking space markings and use them to determine the parking position. However, the necessary image processing algorithms are sometimes very computationally intensive and therefore not feasible in many vehicles. Furthermore, this approach is often not applicable if there are no parking space markings in a parking space at all, or if they are not in the form taken into account in the image processing process.

The invention is therefore based on the object of providing a particularly suitable method for mapping a parking space. In particular, the largest possible coverage of parking spaces should be ensured. The invention is also based on the object of particularly suitable device for carrying out the process.

With regard to the method, the object is achieved according to the invention with the features of claim 1, and with regard to the device with the features of claim 10. Advantageous refinements and developments are the subject of the dependent claims (subclaims). The advantages and refinements cited with regard to the method are also transferable to the device, and vice versa. Where method steps are described below, advantageous refinements for the device result, in particular, from the fact that the device is designed to carry out one or more of these method steps.

The method according to the invention is intended for mapping a parking space, i.e. for entering a parking space into a digital map, and is suitable and configured for this purpose.

The method according to the invention can be carried out by a single (motor) vehicle, wherein the vehicle has at least one vehicle sensor system. Preferably, however, a vehicle fleet comprising a number of fleet or swarm vehicles is used as a swarm. The method is thus particularly designed for swarm mapping. "Swarm mapping" is understood here and below as a process in which information from a multitude of sources (swarm) is collected and combined to create a comprehensive and detailed map of a specific area. This approach utilizes the principles of swarming, i.e., decentralized data acquisition and processing, to achieve high data coverage and accuracy. The method enables large amounts of data to be collected quickly and efficiently by combining the capabilities and range of many small, often autonomously operating units.

The following explanations refer specifically to swarm mapping with a number of swarm or fleet vehicles. However, the explanations are also applicable to a single vehicle that is not part of a swarm or fleet. In particular, the explanations regarding a fleet vehicle are also applicable to a single vehicle.

A "number of fleet vehicles" is understood here and below to mean, in particular, several fleet vehicles, i.e., at least two, preferably a plurality of fleet vehicles. Each of the fleet vehicles has at least one vehicle sensor system to provide sensor data or sensor signals for mapping. The vehicle sensor system comprises a vehicle sensor for acquiring and generating the sensor data/signals. Preferably, the vehicle sensor system also has an evaluation unit for signal processing or evaluation of the sensor data. The evaluation unit preferably generates parking position or map signals or data based on the sensor signals.

The fleet vehicles are signal-linked to a common controller. The fleet vehicles are linked to the controller via a wireless communication connection, for example, via Wi-Fi. The controller is external to the fleet vehicles. The controller is integrated into a network connected to the fleet vehicles, for example, as software or an algorithm on a server or in a data cloud.

The controller maps a global digital map based on (swarm) data. The (swarm) data is map or parking space data generated by one source (the vehicle) or multiple sources (the swarm), particularly by the evaluation units of the vehicle's sensor systems. The swarm data is generated primarily via crowdsourcing by a large number of fleet vehicles.

The digital map includes at least one parking space, i.e., a spatial area with at least one parking bay, which is to be mapped during the process. The process steps described below are carried out, for example, when a fleet vehicle is located in such a parking space. Whether a fleet vehicle is located in a parking space to be mapped can be determined, for example, using navigation data (e.g., via GPS).

The vehicle sensor systems of the fleet vehicles are, in particular, not designed for the direct or immediate detection of a parking space. In other words, the at least one vehicle sensor system is not intended or configured to detect a primary parking space indicator, such as a parking space marking or a parking space sign delimiting the parking space. Since the fleet vehicle, according to the method, does not detect or cannot detect any primary parking space indicators, direct parking space detection or determination for mapping the digital map is therefore not possible or intended. According to the invention, an indirect parking space determination based on secondary parking position indicators in the sensor signals or sensor data.

A "secondary parking space indicator" is understood here and below to mean, in particular, supplementary or indirect information regarding the presence of a parking space. Such a secondary parking space indicator is a heuristic that includes clues that indicate the presence of a parking space with a sufficient degree of probability (confidence). What probability is considered sufficient and how high the actual probability is is initially irrelevant. This can be determined, for example, from past driving or swarm data or from corresponding experiments or trials. Different secondary parking space indicators may arise for different motor vehicles, parking space and environmental conditions, or application scenarios. Examples of secondary parking space indicators could be the presence of vehicles in neighboring parking areas, time-dependent parking restrictions, seasonal occupancy patterns, or data on traffic flow near parking spaces recorded by sensors.

According to the method, a sensor signal is generated by the vehicle sensor system, in particular the vehicle sensor, and is examined, in particular by the evaluation unit, for the presence of at least one secondary parking space indicator. If at least one such parking space indicator is detected, i.e., if a possible parking space can be identified based on the sensor data, the evaluation unit generates a parking space hypothesis with a confidence measure.

A "parking space hypothesis" is understood here and below to mean, in particular, a theoretical assumption or prediction regarding the position and/or orientation of a parking space in the parking area. The parking space hypothesis is based on the sensor signal of the vehicle sensor system or on at least one secondary parking space indicator, possibly with the aid of mathematical or physical models.

A “confidence measure” is understood here and below in particular to mean a measure of confidence, i.e. a measure of trust or certainty in the accuracy or reliability of the parking space hypothesis. The confidence measure is, for example, a probability with which the parking space hypothesis is considered correct or valid. The confidence measure can be expressed quantitatively, for example as a probability value between 0 and 1 or in percent, and serves as an indicator of the reliability of the parking space hypothesis. The confidence measure depends on the secondary parking space indicator identified in the sensor signals. In other words, the confidence measure is determined depending on the at least one identified secondary parking space indicator. It is conceivable, for example, that different values for the assigned confidence measure are stored for different secondary parking space indicators.

According to the method, the parking space hypothesis and the confidence measure are transmitted to the controller. In other words, the fleet vehicle, or the vehicle sensor system, in particular its evaluation unit, transmits the parking space hypothesis and the confidence measure to the vehicle-external controller. Depending on the confidence measure, the parking space hypothesis is mapped by the controller as a parking space on the digital map. This implements a particularly suitable method for mapping a parking space. According to the invention, parking spaces are thus not recorded directly, but are indirectly estimated as parking space hypotheses based on secondary parking space indicators. This allows fleet vehicles that are not equipped with a vehicle sensor system for directly recording parking spaces or primary parking space indicators to participate in the swarm mapping. This increases the number of fleet vehicles that can be used for parking space mapping, thus ensuring particularly extensive coverage of parking spaces.

In an advantageous embodiment, a parking space to be mapped is determined based on a number of parking space hypotheses and mapped on the digital map. In other words, the controller collects various parking space hypotheses for the parking area, particularly from different fleet vehicles, and "merges" the individual parking spaces or parking space hypotheses into effectively mappable parking spaces.

The controller records all directly detected parking spaces (from fleet vehicles with corresponding vehicle sensor systems) and indirectly generated parking space hypotheses. If all these parking spaces and parking space hypotheses are plotted on a map, different situations are obtained for each actual parking space due to the different sensor properties (calibration, measurement principle, possible accuracy, 2D vs. 3D, inaccuracies in self-motion estimation, etc.).

If at least one directly identified parking space can be assigned to a location, this parking space weighs more than possible additional parking space hypotheses at the same location that could also be associated. In the global digital Therefore, only the directly recognized parking space is preferably entered on the map as the parking space to be mapped.

If no directly identified parking space can be assigned to a location, but at least one indirectly identified parking space hypothesis can be assigned, the parking space hypothesis with the highest confidence level is selected, for example, depending on the confidence measure of the associable parking space hypotheses, and entered into the digital map as the parking space to be mapped. In an alternative development, it is possible, for example, to perform a weighted clustering of the parking space hypotheses present at a location to determine the parking space to be mapped. The weighted center points and orientations of the parking space hypotheses are used to determine an effective parking space center point and an effective parking space orientation for the parking space to be mapped. A standard parking space can serve as the parking space dimensions, or alternatively, weighted side length means of the parking space hypotheses involved. The weighting is preferably based on the confidence measures. In other words, during clustering, parking space hypotheses with a higher confidence level are given greater weight than those with a low confidence level.

In a preferred embodiment, if at least one secondary parking space index is present in the sensor signals, a bounding box is generated as a parking space hypothesis. A "bounding box" is understood here and below to mean, in particular, a rectangle that minimally surrounds the parking space or a minimal bounding box, i.e., a rectangular framework, which is used to define and delimit the position and dimensions of a parking space within a given space. The bounding box is subsequently used to mark the area around the parking space in the digital map. The box is defined by the coordinates of its corners or by the position of its center, as well as its height and width (in two-dimensional spaces) or depth (in three-dimensional spaces). The bounding box can therefore be a two-dimensional (2D) or three-dimensional (3D) construct.

Instead of a bounding box, a zero-dimensional (OD) or one-dimensional (1D) construct, i.e., a point or a line, can alternatively be generated and used as a parking space hypothesis. In such one- or zero-dimensional parking space hypotheses, for example, it is assumed that a parking space with a standardized dimension is located or centered around the 1D or 0D construct, with the orientation of the parking space resulting from another convention, such as the overall context.

As an indirect method for detecting parking spaces or parking space hypotheses from vehicle sensor system data, one conceivable embodiment could, for example, involve the vehicle sensor system generating a sensor signal characterizing the vehicle's own movement, with the vehicle's own movement and the parking of the (fleet) vehicle being used as secondary parking space indicators. In other words, a so-called "park-in event," i.e., the vehicle parking, is used as a secondary parking space indicator. This is based on the assumption that a vehicle user parks the vehicle in a permitted parking space.

The vehicle's own motion is determined in particular from a set of odometry trajectories (odometry traces) and the respective vehicle pose. The term "odometry" (also called hodometry) refers to a method of estimating the position and orientation (attitude estimation) of a vehicle based on data from its propulsion system (e.g., the number of wheel revolutions). A "vehicle pose" is understood here and below to mean, in particular, the spatial orientation and position of the vehicle in relation to its surroundings. This includes the specification of the vehicle's attitude in three dimensions: the position (x, y, z coordinates) and the orientation, which is described by angles such as yaw, pitch, and roll.

In a practical implementation, the parking of a (fleet) vehicle is detected by the vehicle remaining stationary for a specified minimum period and/or by the vehicle being switched off. Alternatively, the parking of the vehicle can also be detected by locking a vehicle door or by the fact that no seat is occupied within the vehicle.

The vehicle thus records its own (odometry) trajectory all the way to the parking location. This trajectory is translated into a global reference system, and the bounding box of the stationary vehicle (standard parking space) is entered into the global digital map as a parking space hypothesis. This creates a parking space map for a vehicle fleet over time. By evaluating several such parking space hypotheses in a parking area, it is possible, for example, to use a clustering process to determine an overall parking space hypothesis for the parking space to be mapped from a multitude of such individual parking space hypotheses. The confidence of this overall hypothesis is based on the number of individual hypotheses involved, as well as possibly the length of the trajectory. from the entrance of the parking area to the parking space in order to take into account the increasing uncertainty of the self-motion estimation.

Confidence should preferably not depend solely on the length of the trajectory to the parking bay. Other criteria for determining confidence could include, for example, the number and/or magnitude and/or extent of turns to the parking bay, as many ego-motion measurement systems have difficulty estimating the yaw angle, resulting in corresponding inaccuracies and uncertainties in the yaw angle. Another criterion could be the number of floor changes (via ramps, spindles, etc.), as many ego-motion measurement systems have difficulty estimating or measuring the pitch angle, resulting in corresponding inaccuracies. Furthermore, weather influences could also be relevant, particularly in open-sky parking spaces, as they influence the road friction coefficient and can thus also lead to greater uncertainty.

In an additional or alternative embodiment, the at least one vehicle sensor system comprises a (vehicle) environmental sensor for detecting the vehicle's surroundings. In other words, the at least one vehicle sensor is designed as an environmental sensor. During operation, the vehicle or environmental sensor examines the vehicle's surroundings for parked vehicles, i.e., vehicles that are not the (fleet) vehicle itself. The presence of parked vehicles in the vehicle's surroundings is used as a secondary parking space indicator, assuming that the vehicles are parked or left in permitted parking spaces.

An "environmental sensor" is understood here and below to mean, in particular, a vehicle sensor system designed and configured to collect and process information about the immediate and extended surroundings of the vehicle. Such an environmental sensor captures a variety of environmental data, including, but not limited to, distances to other objects (such as vehicles, pedestrians, and obstacles), road conditions, weather conditions, lighting conditions, and traffic signals. Typical technologies used in such sensors include radar, LIDAR (Light Detection and Ranging), ultrasound, infrared, and optical camera systems.

The environmental sensor can be, for example, an optical sensor, in particular a camera system, such as a front camera of the vehicle configured as a multifunction camera (MFK). This allows parked vehicles and their precise location to be detected, for example, using semantic segmentation or other learning methods. The detected vehicles can be entered as bounding boxes in the controller's global digital map. Provided attributes such as age (number of sightings over time), distance of the vehicle from the camera, etc., can be used to calculate a respective confidence measure.

In a preferred application, a non-camera-based sensor is used as an environmental or vehicle sensor.

The environmental sensor, for example, is a radar sensor. The radar data are, for example, individual measurement or radar points during a radar distance measurement. For evaluation, the radar points are analyzed, for example, using density-based clustering. The secondary parking space indicator of a parked vehicle is characterized, for example, by a vehicle cluster, i.e., by a specific density or spatial distribution of the radar points. For the radar point clusters, a length and width are determined, whereby only those whose length and width correspond to a vehicle are selected as possible vehicle clusters (e.g., a length between 3 and 5 meters and a width between 1.5 and 2 meters). By expanding the bounding box of the detected vehicle or the vehicle cluster to a standard parking space dimension, parking space hypotheses can be generated and mapped. The confidence of these hypotheses can be based on the confidence values and other attributes of the underlying radar data (raw targets, clusters) as well as the geometric properties of the vehicle detection (e.g., properties of the determined eigenvectors / principal components).

The environmental sensor can alternatively be implemented as an ultrasonic sensor. Similar to the radar analysis described above, parked vehicles can also be detected from ultrasonic data and further processed into parking position hypotheses, which can then be mapped. The characteristic geometric mapping properties of vehicle fronts in the point clouds of the sensor data can be exploited.

Analogous to radar and ultrasonic sensors, a LIDAR sensor can also be used as an environmental sensor, whereby the evaluation of the LIDAR data is carried out in accordance with the above statements.

The device according to the invention is used for mapping a parking space, in particular in a predetermined parking space, and is suitable and configured for this purpose. The device comprises at least one vehicle, in particular a vehicle fleet comprising a number of fleet vehicles, wherein the vehicle or each fleet vehicle comprises at least one vehicle sensor system. The device further comprises a controller, signal-coupled to the at least one vehicle, having a digital map of the parking space to be mapped.

The device is generally configured—in terms of programming and/or circuitry—to carry out the method according to the invention described above. The device is thus specifically configured to generate a sensor signal using the vehicle sensor system, wherein the vehicle or the evaluation unit of the vehicle sensor system examines the sensor signal for the presence of a secondary parking space index. If a secondary parking space index is present, the vehicle or the evaluation unit generates a parking space hypothesis with a confidence level and transmits or sends this to the controller. Depending on the confidence level, the controller maps the parking space hypothesis as a parking space on the digital map. This creates a particularly suitable device that ensures particularly large-area coverage of parking spaces.

In a suitable embodiment, the at least one vehicle sensor system of the or each vehicle comprises an ultrasonic sensor, a radar sensor or a LIDAR sensor.

• 1 ein Flottenfahrzeug mit einem Fahrzeugsensorsystem,
• 2 ein Blockdiagramm für eine Vorrichtung zum Kartieren eines Parkstands, und
• 3 eine perspektivischer Darstellung von Sensordaten des Flottenfahrzeugs zur Bestimmung einer Parkstandhypothese.

Exemplary embodiments of the invention are explained in more detail below with reference to a drawing. In schematic and simplified representations, the drawings show:

• 1 a fleet vehicle with a vehicle sensor system,
• 2 a block diagram of a device for mapping a parking space, and
• 3 a perspective view of sensor data from the fleet vehicle to determine a parking position hypothesis.

Corresponding parts and sizes are always provided with the same reference symbols in all figures.

The 1 schematically shows a (motor) vehicle as a fleet vehicle 2 with a vehicle sensor system 4. The vehicle sensor system 4 is designed as a vehicle sensor system by means of which the environment of the fleet vehicle 2 is monitored.

The vehicle sensor system 4 has, for example, a front radar 6 and two rear radars 8. In the 1 The radar areas 10, 12 of the front radar 6 and the rear radar 8 are also shown schematically. The vehicle sensor system 4 further comprises a vehicle camera 14 for optically detecting the front vehicle environment. The sensor area 16 of the vehicle camera 14 is in the 1 shown in dashed lines. In addition, the vehicle sensor system 4 has twelve environmental sensors 18, with six environmental sensors 18 arranged at the front and six at the rear. The environmental sensors 18 are designed in particular as ultrasonic sensors. The environmental or ultrasonic sensors 18 are in the 1 provided with reference symbols for illustrative purposes only.

The vehicle camera 14 is designed as a multifunction camera (MFK) with an integrated controller 20, for example, an EYE-Q5 chip. The vehicle camera 14 or the controller 20 is coupled to an evaluation unit 22. The evaluation unit 22 is also coupled to the radar and ultrasonic sensors 6, 8, 18. The evaluation unit 22 acquires sensor data or sensor signals 19 from the vehicle sensors 6, 8, 14, 18 and examines them for secondary parking position indicators. The evaluation unit 22 is provided and configured to determine a parking position hypothesis 23a and a confidence measure 23b if the presence of a secondary parking position indicator can be inferred from the sensor data.

The evaluation unit 22 is coupled to a server 24 (In Car Application Server, ICAS). The server 24 is further coupled to a transmitting and receiving unit (Online Connectivity Unit, OCU) 26, by means of which the parking position hypothesis and the confidence measure are transmitted to a controller 28 ( 2 ) can be transmitted or sent. The controller 28 is designed, for example, as a data cloud, which is signal-linked to the fleet vehicle 2 or to the vehicle sensor system 4 by means of a wireless signal connection.

The fleet vehicle 2 is preferably part of a vehicle fleet 30 ( 2 ), i.e. a swarm of vehicles. The fleet vehicles 30 are each signal-linked to the controller 28. The following is based on the 2 A method for mapping a parking space is explained in more detail. The method is designed for swarm mapping.

The controller 28 maps the parking position hypotheses generated by the vehicle fleet 30 and confidence measures as swarm data in a global digital map 31.

According to the method, a sensor signal is generated by the vehicle sensor system 4 in a method step 32. Measurement or sensor data is thus acquired by the vehicle sensors 6, 8, 14, 18 and sent to the evaluation unit 22. In a method step 34, the sensor signals are examined by the evaluation unit 22 for the presence of at least one secondary parking space index. If at least one such parking space index is acquired, i.e., if a possible parking space can be identified based on the sensor data, a corresponding parking space hypothesis 23a with a confidence measure 23b is generated by the evaluation unit 22 in a method step 36. This is sent by the fleet vehicle 2 or by the fleet vehicles of the vehicle fleet 30 as swarm data to the controller 28.

The controller 28 maps the parking space hypothesis 23a as a parking space 37 in the digital map 31 depending on the confidence measure 23b. To this end, the controller 28 collects the swarm data in a method step 38 and, from a number of parking space hypotheses 23a, determines a parking space 37 to be mapped for a specific location, which is then mapped in the digital map 31 in a method step 40. This map 31 can then be made available to the vehicle fleet 30 or the fleet vehicle 2 so that driver assistance functions of the respective vehicles can use this parking space information. For this purpose, the digital map with the mapped parking spaces 37 is transmitted to the server 24 and stored locally in the respective fleet vehicles 2.

Using the vehicle camera 14, parking spaces can be directly identified based on primary parking space indicators, for example. Such directly detected parking spaces are preferably also transmitted to the controller 28 as swarm data. The controller 28 thus records all directly detected parking spaces and indirectly generated parking space hypotheses of the vehicle fleet 30.

If at least one directly identified parking space exists at a location, i.e., in a geometric area of the map, this parking space carries more weight than possible additional parking space hypotheses at the same location that could also be associated. Thus, in process step 40, only the directly identified parking space is entered in the global digital map as parking space 37 to be mapped.

If no directly identified parking space can be assigned to a location, but at least one indirectly identified parking space hypothesis 23a can be assigned, the parking space hypothesis 23a with the highest confidence measure 23b is selected, for example, depending on the confidence measure 23b of the associable parking space hypotheses, and entered into the digital map 31 in method step 40 as the parking space 37 to be mapped. For example, it is possible to perform a weighted clustering of the parking space hypotheses 23a present at a location in method step 38 to determine the parking space 31 to be mapped. The weighting is preferably performed based on the respective confidence measures 23b.

The 3 shows, for example, a schematic representation of the sensor data of one or more of the vehicle sensors 6, 8, 18. The fleet vehicle 2 moves in a parking space as a parking area, where, for example, ten vehicles are parked in the parking space, each in a parking space. As they drive past, the vehicle sensors 6, 8, 18 record the parked vehicles as measurement point clouds or vehicle clusters 42, i.e., as a spatial density distribution of individual measurement points. A bounding box 44 is placed around each vehicle cluster 42 as a parking space hypothesis, where 3 a bounding box 44 is shown merely as an example.

In the illustrated embodiment, the evaluation unit 22 recognizes, for example, based on the bounding boxes 44 that the parking space hypotheses are parallel parking spaces. From the clear distance between two adjacent bounding boxes 44, further parking space hypotheses for free parking spaces can thus be derived. For this purpose, further bounding boxes 46 are generated in the free space between two bounding boxes 44 of the parked vehicles as parking space hypotheses 23a. The secondary parking space indicator is a free space between two parking space hypotheses 23a or bounding boxes 44, 46, which is sufficiently large for another parking space. A confidence measure 23b is assigned to each parking space hypotheses 23a, and these are transmitted to the controller 28 as swarm data.

The invention is not limited to the exemplary embodiments described above. Rather, other variants of the invention can also be derived therefrom by those skilled in the art without departing from the scope of the invention. In particular, all individual features described in connection with the exemplary embodiments can also be combined with one another in other ways without departing from the scope of the invention.

For example, it is conceivable that a vehicle's own movement and the parking of fleet vehicle 2 are used as secondary parking position indicators. In other words, a so-called "park-in event," i.e., a parking of fleet vehicle 2, is used as a secondary parking position indicator.

List of reference symbols

2

vehicle, fleet vehicle

4

Vehicle sensor system

6

Front radar, vehicle sensor

8

Rear radar, vehicle sensor

10

Radar range

12

Radar range

14

Vehicle camera, vehicle sensor

16

Sensor range

18

Ultrasonic sensor, environmental sensor, vehicle sensor

19

Sensor signal, sensor data

20

Controller

22

Evaluation unit

23a

Parking space hypothesis

23b

Confidence measure

24

server

26

Transmitting and receiving unit

28

Controller

30

Vehicle fleet

31

digital map

32 ... 36

Process step

37

Parking space

38, 40

Process step

42

Vehicle cluster

44

Bounding Box

46

Bounding Box

QUOTES CONTAINED IN THE DESCRIPTION

This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 10 2017 204 455 A1 [0007]
• DE 10 2019 201 893 A1 [0007]
• DE 10 2015 209 147 A1 [0007]
• DE 10 2014 217 567 A1 [0007]

Claims (10)

A method for mapping a parking space, comprising at least one vehicle (2) having at least one vehicle sensor system (4), and a controller (28) having a digital map (31) and signal-coupled to the at least one vehicle (2), - wherein a sensor signal (19) is generated by the vehicle sensor system (4), - wherein the sensor signal (19) is examined for at least one secondary parking space indicator, - wherein, if at least one secondary parking space indicator is present, a parking space hypothesis (23a) having a confidence measure (23b) is generated, - wherein the parking space hypothesis (23a) and the confidence measure (23b) are transmitted to the controller (28), and - wherein, depending on the confidence measure (23b), the parking space hypothesis (23a) is mapped as a parking space (37) in the digital map (31). Procedure according to Claim 1 , characterized in that a parking space (37) to be mapped is determined on the basis of a number of parking space hypotheses (23a) and is mapped in the digital map (31). Procedure according to Claim 2 , characterized in that in order to determine the parking space (37) to be mapped, a clustering of the parking space hypotheses (23a) weighted on the basis of the confidence measures (23b) is determined. Method according to one of the Claims 1 until 3 , characterized in that if at least one secondary parking space index is present, a bounding box (44, 46) is generated as a parking space hypothesis (23a). Method according to one of the Claims 1 until 4 , characterized in that a sensor signal (19) characterizing the vehicle's own movement is generated by the vehicle sensor system (4), wherein the vehicle's own movement and a parking of the vehicle (2) are used as secondary parking position indicators. Procedure according to Claim 5 , characterized in that a parking of the vehicle (2) is detected by a vehicle standstill for a predetermined minimum period of time and/or by switching off the vehicle. Method according to one of the Claims 1 until 6 , characterized in that the vehicle sensor system (4) has an environmental sensor (18) for detecting a vehicle environment of the vehicle (2), wherein the vehicle sensor system (4) examines the vehicle environment for parked vehicles, and wherein a presence of a parked vehicle is used as a secondary parking position indicator. Procedure according to Claim 7 , characterized in that a non-camera-based sensor is used as the environmental sensor (18). Device for mapping a parking space, comprising at least one vehicle (2) which has at least one vehicle sensor system (4), and a controller (28) which is signal-coupled to the at least one vehicle (2) and has a digital map (31), wherein the vehicle sensor system (4) generates a sensor signal (19) and examines the sensor signal (19) for the presence of a secondary parking space index, wherein the vehicle sensor system (4) generates a parking space hypothesis (23a) with a confidence measure (23b) if a secondary parking space index is present, wherein the vehicle sensor system (4) transmits the parking space hypothesis (23a) and the confidence measure (23b) to the controller (28), and wherein the controller (28) maps the parking space hypothesis (23a) as a parking space (37) in the digital map (31) depending on the confidence measure (23b). Device according to Claim 9 , characterized in that the at least one vehicle sensor system (4) comprises an ultrasonic sensor, a radar sensor or a LIDAR sensor.