DE102021117613A1 - METHOD OF OPERATING A PARKING ASSISTANCE SYSTEM, COMPUTER PROGRAM PRODUCT AND PARKING ASSISTANCE SYSTEM - Google Patents

Abstract

A method for operating a parking assistance system (100) is proposed. The parking assistance system (100) comprises a number (102) of sensors (102A - 102F) arranged in a predetermined area (110), each for detecting a vehicle (200) present in the predetermined area (110) and for outputting a respective sensor signal are set up. The method comprises the steps: a) detecting (S1) a number of specific features (201-203) of the vehicle (200),b) determining (S2) a specific vehicle model (MSPEC) for the vehicle (200) on the basis of the specific detected Features (201-203),c) receiving (S3) a sensor signal from at least one sensor (102A-102F) of the number (102),d) determining (S4) a pose (POS) of the vehicle (200) as a function of the received signal Sensor signal using the determined specific vehicle model (MSPEC), unde) controlling (S5) of the vehicle (200) depending on the determined pose (POS).

Description

The present invention relates to a method for operating a parking assistance system, a computer program product and a parking assistance system.

Vehicles are known which have an automated parking function which is particularly suitable for automatically parking the vehicle in a suitable multi-storey car park or a suitable parking space. Such systems are referred to, for example, as automated valet parking systems. A distinction is made between two types. In a first type, the vehicle is self-driving, with the parking garage having, for example, suitable features for orienting the vehicle, such as ARUCO codes. In a second type, the vehicle can be controlled remotely, with the multi-storey car park having sensors and path planning means, for example, in order to control the vehicle. There can be various intermediate levels between these two types, in which the functions are distributed differently between the vehicle and the parking garage.

In a type 2 valet parking system, the vehicle to be controlled must be detectable at every location in the parking system by sensors arranged externally to the vehicle in such a way that the pose of the vehicle can be determined in order to ensure safe operation. This requires a very dense coverage of the parking system with expensive sensors, which is why such systems are very expensive and complex to provide. In addition, powerful hardware must be provided that can process the amounts of sensor data generated very quickly, ideally in real time. In addition, the mathematical problem of determining an unambiguous determination of the pose on the basis of a large number of sensor data, which is available, for example, as a point cloud, is very complex and can be susceptible to interference. Known algorithms for determining the pose are based, for example, on a general vehicle model that applies to a large number of vehicles. The general vehicle model therefore includes in particular general characteristics that apply to all or most vehicles. The more a vehicle deviates from this model, the more imprecisely its pose can be determined since the vehicle model already does not fit the vehicle well.

It is desirable to improve the detection of vehicles and the determination of a respective pose so that the parking system achieves more reliable and robust control with the same effort, or alternatively allows for the same reliability as a conventional parking system with less effort.

Against this background, an object of the present invention is to specify an improved operating method for a parking assistance system and a corresponding parking assistance system.

1. a) Erfassen einer Anzahl spezifischer Merkmale des Fahrzeugs,
2. b) Ermitteln eines spezifischen Fahrzeugmodells für das Fahrzeug auf Basis der erfassten spezifischen Merkmale,
3. c) Empfangen eines Sensorsignals von wenigstens einem Sensor,
4. d) Ermitteln einer Pose des Fahrzeugs in Abhängigkeit des empfangenen Sensorsignals unter Verwendung des ermittelten spezifischen Fahrzeugmodells, und
5. e) Steuern des Fahrzeugs in Abhängigkeit der ermittelten Pose.

According to a first aspect, a method for operating a parking assistance system is proposed. The parking assistance system includes a number of sensors arranged in a predetermined area, which are each set up to detect a vehicle present in the predetermined area and to output a respective sensor signal. The procedure includes the steps:

1. a) detection of a number of specific characteristics of the vehicle,
2. b) determining a specific vehicle model for the vehicle based on the detected specific features,
3. c) receiving a sensor signal from at least one sensor,
4. d) determining a pose of the vehicle as a function of the received sensor signal using the determined specific vehicle model, and
5. e) Controlling the vehicle depending on the determined pose.

This method has the advantage that the pose of the vehicle can be determined based on the specific features of the vehicle that individualize it and the specific vehicle model. This means that instead of a general vehicle model, which is generalized to the extent that it fits every vehicle, the specific vehicle model is used, which essentially only fits the respective vehicle. The pose can thus be reliably determined even on a smaller database, which is why savings in the number of sensors and/or their sensitivity are possible. On the other hand, with the same database, greater accuracy can be achieved when determining the pose compared to a conventional parking assistance system.

If one formulates the determination of the pose as an extreme value problem, in which an alignment of a three-dimensional vehicle model is adapted to acquired sensor data of the vehicle, which only includes part of the vehicle, such that a deviation of the vehicle model from the sensor data is minimal, then the Effect of the specific vehicle model compared to a general vehicle model that the "right" solution (global minimum) differs significantly from "wrong" solutions (local minima). It can also be said that determining the pose is clearer.

A further advantage of the method is that the existence of the vehicle in the predetermined area does not have to be checked again in each measurement interval. Rather, the parking assistance system “knows” that the individual vehicle is present at a specific position in the predetermined area. In particular when several vehicles are present in the predetermined area at the same time, conventional parking assistance systems have the further difficulty that different vehicles cannot be individually distinguished by the parking assistance system, since all vehicles are based on the same vehicle model.

The predetermined area includes the area in which the vehicles can be controlled by the parking assistance system. The predetermined area is in particular completely covered by the sensors, so that there are no blind spots in which the pose of the vehicle cannot be determined.

A respective sensor of the number is set up to detect a partial area or section of the predetermined area and to output a corresponding sensor signal. If a vehicle is in the respective partial area, then the sensor signal includes the vehicle and the pose of the vehicle can be determined on the basis of the sensor signal.

The specific characteristics of the vehicle can be recorded in different ways. The vehicle is preferably detected and/or measured by sensors, and the specific features of the vehicle are thus determined by the parking assistance system itself. Additionally and/or alternatively, the or some specific features of the vehicle can be received from another entity, such as from the vehicle itself or also from a server on the Internet or a database.

The specific features of the vehicle can be all possible features that can be detected by the sensors of the parking assistance system, that is, in particular, that are visible or detectable from the outside. The specific features include both general features that may be specific to a particular vehicle model and, in particular, include optional equipment that may distinguish a particular vehicle from other vehicles of the same model. Furthermore, individualizing features such as stickers, special rims, the number plate, or also damage and soiling and the like, which are subsequently applied by a user, can be incorporated into the specific vehicle model as specific features.

Further examples of specific features are equipment features of the vehicle, such as paint color, shape and/or color of elements of the vehicle, such as trim strips, exterior mirrors, rims, headlights, door handles, roof rails, grill, spoiler, antenna, number plate and the like. Furthermore, the shape of window openings, a course of gaps or a vehicle shape, such as four-door, coupe, convertible, SUV, small car, transporter, minibus and the like, can be used as a specific feature.

The specific vehicle model is determined on the basis of the detected specific features of the vehicle. That is, the specific vehicle model includes the specific features. On the basis of the specific vehicle model, vehicles that only differ in one specific feature can be distinguished from one another. The specific vehicle model may vary depending on how the pose of the vehicle is determined in use. If the pose is determined, for example, as explained above, by bringing the vehicle model into agreement with the vehicle as contained in the sensor signals, the vehicle model can comprise, for example, a digital three-dimensional model of the respective vehicle having the detected specific features . If the pose is determined using an artificial intelligence as a detector, such as an appropriately trained neural network, then the specific vehicle model may comprise a detector trained with the specific features.

For example, a vehicle has a yellow sticker on the right side of the vehicle as a specific feature. If the yellow sticker is included in the sensor signal, it is immediately determined that the right side of the vehicle is visible and not the left side of the vehicle.

The determined pose includes in particular a position of the vehicle in the predetermined area and an orientation of the vehicle. This information relates, for example, to a center point of a second axle (rear axle) of the vehicle, in particular when this axle is a non-steered axle.

Controlling the vehicles includes, for example, determining a respective trajectory for a respective vehicle. The trajectory leads, for example, from a current position of the vehicle to a target position. The target position is Special a specific parking space within the predetermined area, or is a transfer area in which a user takes over the vehicle from the parking assistance system. On the basis of the determined trajectory, the vehicle is controlled in such a way that it follows the determined trajectory. For this purpose, the parking assistance system transmits appropriate control commands to the vehicle, for example. A wireless communication technology, such as a WLAN, a mobile radio connection or the like, is preferably used to transmit the control commands. During this remote-controlled drive of the vehicle, the parking assistance system monitors the vehicle on the basis of the received sensor signals. In particular, the pose of the vehicle is continuously monitored on the basis of the sensor signals and, if necessary, control commands for correcting the pose are output to the vehicle. The controller is also based, for example, on a digital map of the predetermined area, which the parking assistance system uses to determine the trajectory in such a way that the vehicle arrives at the target point without a collision. In particular, the digital map can include all mobile and immobile obstacles in the predetermined area. Mobile obstacles, such as other vehicles or pedestrians, can also be detected by the sensors and determined by the parking assistance system and entered accordingly in the digital map.

In embodiments, the transportable parking assistance system includes a transmitter that is set up to transmit control commands to the vehicles.

• sensorisches Erfassen des Fahrzeugs zum Bereitstellen von spezifischen Sensordaten für das Fahrzeug.

According to one embodiment of the method, step a) further comprises:

• sensory detection of the vehicle to provide specific sensor data for the vehicle.

The specific sensor data that is provided here can be indicative of the specific features of the vehicle and/or they are suitable for uniquely identifying the vehicle, such as a number plate.

The sensory detection takes place in particular when the vehicle is at a transfer point at which a user of the vehicle transfers control of the vehicle to the parking assistance system. Certain sensors of the number of sensors can be used for sensory detection, and/or additional sensors specially provided for this purpose can be provided.

• Empfangen eines spezifischen Datensatzes für das Fahrzeug, welcher die spezifischen Merkmale umfasst, von dem Fahrzeug, und/oder
• Empfangen einer Fahrzeugkennung von dem Fahrzeug und Abrufen eines spezifischen Datensatzes für das Fahrzeug, welcher die spezifischen Merkmale umfasst, aus einer Datenbank in Abhängigkeit der empfangenen Fahrzeugkennung, und/oder
• Abrufen eines spezifischen Datensatzes für das Fahrzeug, welcher die spezifischen Merkmale umfasst, aus einer Datenbank in Abhängigkeit der spezifischen Sensordaten.

According to a further embodiment of the method, step a) also comprises:

• receiving from the vehicle a specific data set for the vehicle, which includes the specific features, and/or
• Receiving a vehicle identifier from the vehicle and retrieving a specific data set for the vehicle, which includes the specific features, from a database depending on the received vehicle identifier, and/or
• Retrieving a specific data set for the vehicle, which includes the specific features, from a database depending on the specific sensor data.

It should be noted that the specific data set is received or retrieved in particular in addition to a sensory detection of the specific features according to the embodiment described above.

The specific data set for the vehicle, which includes the specific features, can be transmitted from the vehicle to the parking assistance system. The specific data record can be created, for example, by the vehicle manufacturer during production of the vehicle and stored in a memory unit of the vehicle.

Alternatively, the specific data record can also be retrieved from a database on the basis of a unique vehicle identifier. The unique vehicle identifier is, for example, a chassis number that the vehicle transmits to the parking assistance system for this purpose. The vehicle can also be clearly identifiable on the basis of its number plate, with the number plate being determined on the basis of the specific sensor data, for example. In embodiments, the database from which the specific data record is retrieved is an internal database of the parking assistance system, or an official database, in particular a registration office, or a database of the vehicle manufacturer. In particular, the database can be reached via a network such as the Internet.

According to a further embodiment of the method, this includes determining the specific features as a function of the specific sensor data.

According to a further embodiment of the method, the sensory detection of the vehicle includes detecting the vehicle by means of a number of sensor devices from different directions, the number of sensor devices having at least one visual camera and/or at least one infrared camera and/or comprises at least one laser scanner and/or at least one lidar and/or at least one radar.

The sensor devices are specifically designed for this task. The sensor devices can form a subset of the number of sensors. The sensor devices are arranged in particular in a transfer area.

By capturing the vehicle from different directions, it can be ensured that all specific features of the vehicle are captured and thus a complete specific vehicle model can be determined.

According to a further embodiment of the method, at least one sensor device is rotated and/or moved relative to the vehicle during the sensory detection of the vehicle.

This embodiment is advantageous because it allows the number of sensor devices that are required to detect the vehicle from all directions to be reduced.

For example, a camera is moved along a rail to capture the vehicle from different angles (directions).

According to a further embodiment of the method, the vehicle is illuminated with different lighting patterns during the sensory detection.

This embodiment has the advantage that optical properties of the vehicle and its specific features are also included in the specific vehicle model, which is particularly advantageous in parking garages, which usually have artificial lighting that varies greatly spatially. It can also be said that different lighting situations that can occur in the predetermined area while the vehicle is being driven are simulated here. In addition, specific features, such as a certain shape of a hood or dents, can be determined in this way using the principles of ray optics.

A respective lighting pattern includes in particular a specific relative arrangement of a lighting source, the vehicle and the respective sensor device, as well as a specific configuration of the lighting device and the sensor device. The configuration includes, for example, a specific radiation characteristic of the lighting device, such as spatial and/or temporal modulation, the use of specific filters, such as color filters or polarization filters, and, for example, settings of the sensor device, such as sensitivity, an integration time, the use of specific filters and more like that.

• Trainieren eines Detektors zum Ermitteln der Pose des Fahrzeugs auf Basis des ermittelten spezifischen Fahrzeugmodells und/oder der spezifischen Merkmale.

According to a further embodiment of the method, step b) further comprises:

• training a detector to determine the pose of the vehicle based on the determined specific vehicle model and/or specific features.

The detector can also be called a filter. After training, the detector is set up to determine the pose of the vehicle on the basis of the sensor signals. The detector is designed in particular as an artificial intelligence, such as a neural network, in particular a CNN (convolutional neural network).

In particular, the detector is already pre-trained on the basis of general characteristics that vehicles typically have. This means that the detector already includes a general vehicle model, for example, and that it can be suitable for determining the pose of a vehicle even before training with the specific features. However, the confidence in the determined pose of the detector before this training is significantly lower than afterwards.

For example, the structure of a neural network is as follows. The neural network can be divided into three layers or sections, an input layer, one or more hidden layers, and an output layer. Each of the layers includes a plurality of individual neurons, it also being possible for different layers to have a different number of neurons. The amount of neurons in the input layer depends, for example, on the amount and format of the data to be processed by the neural network, which can also be referred to as the dimensionality of the data. For example, the amount of neurons in the output layer depends on the dimensionality of the output vector.

The input layer forms the input of the neural network, to which the input vector is fed. The neurons of the input layer accept the information contained in the input vector and pass it on weighted to the neurons of the first hidden layer. The neurons of the first hidden layer receive the weighted signals from the input layer and pass them on to the following hidden layer in a weighted manner. The signals are passed from layer to layer in this way passed through the neural network until it reaches the output layer. The output vector is tapped off at the neurons of the output layer. The amount of hidden layers is basically unlimited, however the computational power needed to operate the neural network scales with the amount of hidden layers. The more layers a neural network includes, the more complex tasks the neural network can solve, but the training effort, i.e. the size of the training data set, increases accordingly. It is therefore advantageous to use only as many hidden layers as are necessary for the task at hand.

In addition to the training data used, the topology of the neural network has a major influence on how well suited a respective neural network is to solving a specific task. A suitable topology depends on the available sensors and/or the type of input data. A suitable topology can be determined for specific sensors, for example by trying out a number of different topologies. By using the specific features as training data, the neural network is optimally adapted to determine the pose of the respective vehicle. The pre-trained neural network can be stored like a computer program and serves as a template for the specific neural network. In particular, the training of the neural network with the specific features takes place without a great expenditure of time, since the training data is limited. One can also say that a fine-tuning of the pre-trained neural network is carried out.

According to a further embodiment of the method, step b) includes determining a kinematic vehicle model for the vehicle.

In particular, the kinematic vehicle model allows a prediction of how the pose of the vehicle will change when certain control commands are executed. This prediction can be used to exclude from the solution space implausible poses that represent a possible solution based solely on the sensor signal.

The kinematic vehicle model is preferably determined on the basis of a specific data set for the vehicle. The specific data record includes, for example, information such as a turning circle for the vehicle and/or whether the vehicle has rear axle steering and the like. The specific data set can also include the kinematic vehicle model as a whole.

• Ermitteln einer zweiten Pose des Fahrzeugs in Abhängigkeit des empfangenen Sensorsignals unter Verwendung eines allgemeinen Fahrzeugmodells, und
• Plausibilisieren der ermittelten Pose in Abhängigkeit der ermittelten zweiten Pose.

According to a further embodiment of the method, this also includes:

• determining a second pose of the vehicle as a function of the received sensor signal using a general vehicle model, and
• Plausibility check of the determined pose depending on the determined second pose.

This embodiment serves to safeguard the determined pose. In particular, when using an artificial intelligence as a detector to determine the pose, which is trained live based on the specific features, it is not possible to check the robustness and reliability of the artificial intelligence before using it. In order to rule out that the determined pose is incorrect, a second pose is therefore additionally determined using a conventional method. The second pose is in particular less precise than the determined pose. A confidence interval is determined for the second pose, for example, within which the determined pose must lie in order to be checked for plausibility. If the determined pose is outside the confidence interval, then it is rejected as implausible and the control is based on the second pose.

• Bereitstellen einer Anzahl von Fahrzeugklasse-Modellen, wobei ein jeweiliges Fahrzeugklasse-Modell ein Fahrzeugmodell für eine jeweilige Fahrzeugklasse umfasst,
• Ermitteln einer Fahrzeugklasse, zu welcher das Fahrzeug gehört, und
• Auswählen des der Fahrzeugklasse entsprechenden Fahrzeugklasse-Modells, wobei das spezifische Fahrzeugmodell auf Basis des ausgewählten Fahrzeugklasse-Modells ermittelt wird.

According to a further embodiment of the method, this also includes:

• providing a number of vehicle class models, a respective vehicle class model comprising a vehicle model for a respective vehicle class,
• determining a vehicle class to which the vehicle belongs, and
• selecting the vehicle class model corresponding to the vehicle class, wherein the specific vehicle model is determined based on the selected vehicle class model.

In this embodiment, predefined vehicle models for different vehicle classes are already present. For example, determining the pose of a vehicle using the vehicle class model applicable to the vehicle is already more accurate than using a generic vehicle model. The specific vehicle model is based on the respective vehicle class model and refines this even further so that the pose can be determined even more reliably, precisely and unambiguously.

According to a further embodiment of the method, the determined specific vehicle model is stored in the predetermined area after completion of an autonomous driving of the vehicle controlled by the parking assistance system and is called up again when the vehicle registers with the parking assistance system at a later date.

This has the advantage that the specific vehicle model does not have to be determined again. This embodiment is particularly advantageous when the vehicle is parked temporarily, for example in a parking lot served by the parking assistance system. If the user of the vehicle requests this again, the specific vehicle model of the vehicle is called up and the autonomous driving of the vehicle can be carried out on the basis of this. The specific vehicle model does not have to be available permanently, which saves resources of the parking assistance system.

According to a second aspect, a computer program product is proposed which comprises instructions which, when the program is executed by a computer, cause the latter to execute the method according to the first aspect.

A computer program product, such as a computer program means, can be made available or supplied by a server in a network, for example, as a storage medium such as a memory card, USB stick, CD-ROM, DVD, or in the form of a downloadable file. This can be done, for example, in a wireless communication network by transferring a corresponding file with the computer program product or the computer program means.

• eine Erfassungseinheit zum Erfassen einer Anzahl spezifischer Merkmale des Fahrzeugs,
• eine Ermittlungseinheit zum Ermitteln eines spezifischen Fahrzeugmodells für das Fahrzeug auf Basis der erfassten spezifischen Merkmale,
• eine Empfangseinheit zum Empfangen eines Sensorsignals von wenigstens einem Sensor der Anzahl, wobei die Ermittlungseinheit zum Ermitteln einer Pose des Fahrzeugs in Abhängigkeit des empfangenen Sensorsignals unter Verwendung des ermittelten spezifischen Fahrzeugmodells eingerichtet ist, und
• eine Steuereinheit zum Steuern des Fahrzeugs in Abhängigkeit der ermittelten Pose.

According to a third aspect, a parking assistance system is proposed with a number of sensors arranged in a predetermined area for detecting a vehicle present in the predetermined area and for outputting a respective sensor signal. The parking assistance system also includes:

• a detection unit for detecting a number of specific features of the vehicle,
• a determination unit for determining a specific vehicle model for the vehicle on the basis of the detected specific features,
• a receiving unit for receiving a sensor signal from at least one sensor of the number, wherein the determination unit is set up for determining a pose of the vehicle as a function of the received sensor signal using the determined specific vehicle model, and
• a control unit for controlling the vehicle depending on the determined pose.

The embodiments and features described for the proposed method apply accordingly to the proposed parking assistance system. The proposed parking assistance system is preferably operated according to the method of the first aspect.

The respective unit of the parking assistance system can be implemented in terms of hardware and/or software. In the case of a hardware implementation, the respective unit can be embodied, for example, as a computer or as a microprocessor. In the case of a software implementation, the respective unit can be designed as a computer program product, as a function, as a routine, as an algorithm, as part of a program code or as an executable object.

According to one embodiment of the parking assistance system, the detection unit comprises a number of sensor devices, the sensor devices being arranged in a transfer area in which a user of the vehicle transfers the vehicle to the parking assistance system and being set up to detect the vehicle from different directions and to output respective specific sensor data are, wherein the number of sensor devices comprises at least one visual camera and/or at least one infrared camera and/or at least one laser scanner and/or at least one lidar and/or at least one radar, the detection unit for determining the specific features of the Vehicle is set up based on the specific sensor data.

Further possible implementations of the invention also include combinations of features or embodiments described above or below with regard to the exemplary embodiments that are not explicitly mentioned. The person skilled in the art will also add individual aspects as improvements or additions to the respective basic form of the invention.

• 1 zeigt eine schematische Darstellung eines Ausführungsbeispiels eines Parkassistenzsystems;
• 2 zeigt eine schematische Darstellung eines Übertragens eines spezifischen Datensatzes;
• 3 zeigt eine schematische Darstellung eines sensorischen Erfassens eines Fahrzeugs;
• 4 zeigt eine weitere schematische Darstellung eines sensorischen Erfassens eines Fahrzeugs;
• 5 zeigt ein schematisches Flussdiagramm eines Ermittelns eines spezifischen Fahrzeugmodells;
• 6 zeigt ein schematisches Blockdiagramm eines Ausführungsbeispiels für ein Parkassistenzsystem;
• 7 zeigt ein schematisches Blockschaltbild eines ersten Ausführungsbeispiels eines Verfahrens zum Betreiben eines Parkassistenzsystems; und
• 8 zeigt ein schematisches Blockschaltbild eines zweiten Ausführungsbeispiels eines Verfahrens zum Betreiben eines Parkassistenzsystems.

Further advantageous refinements and aspects of the invention are the subject matter of the dependent claims and of the exemplary embodiments of the invention described below. The invention is explained in more detail below on the basis of preferred embodiments with reference to the enclosed figures.

• 1 shows a schematic representation of an embodiment of a parking assistance system;
• 2 shows a schematic representation of a transmission of a specific data set;
• 3 shows a schematic representation of a sensory detection of a vehicle;
• 4 shows a further schematic representation of a sensory detection of a vehicle;
• 5 shows a schematic flowchart of determining a specific vehicle model;
• 6 shows a schematic block diagram of an exemplary embodiment of a parking assistance system;
• 7 shows a schematic block diagram of a first exemplary embodiment of a method for operating a parking assistance system; and
• 8th shows a schematic block diagram of a second exemplary embodiment of a method for operating a parking assistance system.

Elements that are the same or have the same function have been provided with the same reference symbols in the figures, unless otherwise stated.

1 shows a schematic representation of an exemplary embodiment of a parking assistance system 100. Parking assistance system 100 includes a number of sensors 102A-102F arranged in a predetermined area 110, each of which is set up to detect a partial area of predetermined area 110 and to output a corresponding sensor signal. The respective output sensor signal is indicative of a vehicle 200 that is in the respective detected sub-area. In this example, the predetermined area 110 is a parking area with a plurality of parking spaces (marked with “P”). The parking assistance system 100 includes a control device 120 which is set up to control vehicles 200 on the basis of the sensor signals from the sensors 102A-102F. In particular, it is a type 2 valet parking assistance system. The sensors 102A-102F are designed as cameras, for example.

In the example of 1 the parking assistance system 100 includes a transfer area 112 which is separated from the remaining part of the predetermined area 110 by a barrier 130 . The transfer area 112 is used in particular for a user of the vehicle 200 to leave the vehicle 200 in this area and transfer control of the vehicle 200 to the parking assistance system 100 (or vice versa receive the vehicle and take control of the parking assistance system). The transfer area 112 can also be referred to as a “drop-off zone” and/or “pick-up zone”. Two sensor devices 103A, 103B are arranged to the side of the transfer area 112 . In this example, these sensor devices 103A, 103B are used to detect vehicle 200 by sensors when control is transferred to parking assistance system 100 . The sensor devices 103A, 103B include, for example, cameras, lidar and/or radar.

The sensor devices 103A, 103B detect the vehicle 200 and provide corresponding specific sensor data that are specific to the vehicle 200 ready. Since the vehicle 200 is at a standstill, the vehicle 200 can be detected with a very high resolution. The control device 120 includes, in particular, a detection unit 122 (see 6 ), which determines specific features 201 - 203 of vehicle 200 on the basis of the specific sensor data. The specific features 201 - 203 are, for example, a sticker, damage, soiling or the like. On the basis of the recorded specific features 201-203, a determination unit 124 (see 6 ) of the control device 120 a specific vehicle model MSPEC (see 5 or 8th ) for the vehicle 200. The specific vehicle model MSPEC can be understood in particular as a detector that makes it possible to determine the pose POS of the vehicle 200 on the basis of the sensor signals.

As soon as parking assistance system 100 has taken control of vehicle 200, it determines, for example, a trajectory TR for vehicle 200, which leads vehicle 200 to a specific parking space in predetermined area 110. In order to steer the vehicle 200 safely and reliably to the parking space, the controller 120 continuously receives their respective sensor signals from the sensors 102A-102F and determines the pose POS of the vehicle 200 depending on the received sensor signals using the determined specific vehicle model MSPEC. The pose POS includes, in particular, the position and the orientation of the vehicle 200. The vehicle 200 can be controlled safely and reliably along the trajectory TR on the basis of the determined pose POS. In this context, the use of the specific vehicle model MSPEC is of particular advantage, since this enables the pose POS of the vehicle 200 to be determined more reliably, more precisely and more clearly than in the case of conventional parking assistance systems.

It should be noted that the sensor devices 103A, 103B do not necessarily have to be present. In specific embodiments of parking assistance system 100, these sensor devices 103A, 103B are therefore not present. Instead, the specific features 201 - 203 of the vehicle 200 can also be in the form of a specific data set SPEC (see 2 ) received by the vehicle 200 or retrieved from a database 300. Alternatively or additionally, specific features 201 - 203 can be detected by sensors 102A - 102F of parking assistance system 100 .

2 shows a schematic representation of a transmission of a specific data record SPEC. The specific data record SPEC can, for example, contain specific characteristics 201 - 203 (see 1 ) of the vehicle 200 include. These include, for example, information on special equipment for vehicle 200, dimensions of the vehicle, and the like. The specific data set SPEC can be stored by vehicle 200 itself and can be transmitted by it to control device 120 of parking assistance system 100 . As an alternative to this, the vehicle 200 can retrieve the specific data set SPEC from a database 300 which is provided by the vehicle manufacturer, for example. For this purpose, vehicle 200 transmits, for example, a unique vehicle identification ID, such as the chassis number, to database 300, which then transmits the specific data record SPEC to vehicle 200, which forwards it to control device 120. As an alternative to this, the vehicle 200 can also transmit the vehicle identifier ID to the control device 120 and this retrieves the specific data record SPEC accordingly from the database 300 .

In specific embodiments (not shown), the control device 120 determines the vehicle identifier ID itself on the basis of detected specific sensor data. The vehicle identifier ID can be determined here, for example, in the form of the license plate number of the vehicle 200, with the database 300 being provided by a licensing authority.

3 shows a schematic representation of a sensory detection of a vehicle 200 according to an embodiment. In this case, vehicle 200 is detected by two sensor devices 103A, 103B from different directions. The sensor devices 103A, 103B are, for example, cameras and radar. The sensor devices 103A, 103B are movably arranged here, for example they can be displaced along the paths shown in dashed lines. Vehicle 200 is detected by sensor devices 103A, 103B from at least a first position I and a second position II. Positions I, II form the endpoints of the paths shown. The sensor devices 103A, 103B can also detect the vehicle 200 from a number of positions that lie between the two positions I, II in order to detect the vehicle 200 with an even greater degree of detail. The detection from different directions ensures that the vehicle 200 is completely detected, i.e. that no specific feature 201 - 203 (see 1 ) of the vehicle 200 is overlooked. In particular, sensor devices 103A, 103B provide specific sensor data of vehicle 200 . The specific vehicle model MSPEC can thus be determined with the greatest possible accuracy. By arranging the sensor devices 103A, 103B in a displaceable manner, the number of sensor devices that are required for the complete detection of the vehicle 200 or its specific features 201-203 can be kept low.

4 shows a further schematic representation of a sensory detection of a vehicle 200 according to a further specific embodiment. In this example, the vehicle 200 is illuminated by a lighting means 104 while being sensed by the sensor device 103A. The lighting means 104 is set up to provide different lighting patterns. In this example, the different lighting patterns are achieved in that the lighting means 104 illuminates the vehicle 200 from different positions I, II. As a result, light beams R1-R4, which strike vehicle 200 from lighting means 104, are reflected and/or scattered in different ways in order to be detected by sensor device 103A.

For example, the light beams R1 and R3 (the light beam R3 is emitted by the lighting means 104 when it is in the second position II) are each reflected on the surface of the vehicle 200, so that the reflected beams RR1, RR3 with a comparatively high Intensity (represented by the width of the light beams) can be detected by the sensor device 103A. From this behavior it can be deduced, for example, that the surface of vehicle 200 is flat and smooth in the corresponding area and reflects the incident light. On the other hand, the light rays R2 and R4 (the light ray R4 is emitted by the lighting means 104 when it is at the second position II) are scattered at the specific feature 201, such as an antenna of the vehicle 200, for example. The scattered light beams SR2, SR4 are detected by the sensor device 103A with a comparatively low intensity. In contrast to the flat areas of the surface, which reflect the light directly, the feature 201 reflects some light towards the sensor device 103A even if the arrangement of the sensor device 103A, lighting means 104 and vehicle 200 with respect to the feature 201 does not direct reflection of light beam len in the direction of the sensor device 103A allows.

The lighting of the vehicle 200 with different lighting patterns during sensory detection is particularly useful because the lighting situation in multi-storey car parks, in which the parking assistance system 100 is used, for example, often varies greatly locally, so that the specific vehicle model MSPEC also has this aspect on the basis of detected measurement data can take into account. It can also be said that the specific vehicle model MSPEC in this embodiment additionally includes a lighting model of the vehicle 200 .

5 shows a schematic flowchart of determining a specific vehicle model MSPEC. In this example, a plurality of different vehicle class models M1-M3 are provided, each of the vehicle class models M1-M3 including a vehicle model for a respective vehicle class CLS. For example, in this example, vehicle class model M1 is that of a convertible with the top down, vehicle class model M2 is that of a small car, and vehicle class model M3 is that of an SUV. The respective vehicle class model M1-M3 is a vehicle model that is generic for the respective vehicle class CLS, which is specific to a general vehicle model for any vehicle, but is general to exactly one specific vehicle compared to a specific vehicle model MSPEC. In a first step, the vehicle class CLS, to which the vehicle to be controlled 200 (see 1 - 4 ) should be determined. This is done, for example, on the basis of a received specific data set SPEC (see 2 ) and/or on the basis of detected specific sensor data of the vehicle 200. On the basis of the determined vehicle class CLS, the corresponding vehicle class model M2 is selected, in this example it is a small car. In a second step, specific features 201-203 of vehicle 200 are received. These were particularly detected by sensors, such as based on the 3 and or 4 described. Additionally and/or alternatively, a specific data set SPEC with specific features for the vehicle 200 can be received. A specific vehicle model MSPEC is determined using the specific features 201-203 on the basis of the vehicle class model M2. The specific vehicle model MSPEC is particularly suitable for determining a pose POS (see 1 ) of vehicle 200 based on sensor signals received from sensors 102A-102F.

It should be noted that the use of vehicle class models M1 - M3 as described above is only optional. In embodiments, the specific vehicle model MSPEC can also be based on a generic vehicle model MGEN (see 8th ) can be determined on the basis of the specific characteristics 201 - 203.

6 shows a schematic block diagram of an embodiment of a parking assistance system 100. The parking assistance system 100 includes a number 102 of in a predetermined area 110 (see 1 ) arranged sensors 102A - 102F (see 1 ), each for detecting a vehicle 200 present in the predetermined area 110 (see 1 - 4 ) and are set up to output a respective sensor signal. The parking assistance system 100 also includes a control device 120, which has a detection unit 122 for detecting a number of specific features 201-203 (see 1 , 4 or 5 ) of the vehicle 200, a determination unit 124 for determining a specific vehicle model MSPEC (see 5 or 8th ) for the vehicle 200 on the basis of the detected specific features 201 - 203, a receiving unit 126 for receiving a sensor signal from at least one sensor 102A - 102F of the number 102, the determination unit 124 for determining a pose POS (see 1 ) of the vehicle 200 as a function of the received sensor signal using the determined specific vehicle model MSPEC, and a control unit 128 for controlling the vehicle 200 as a function of the determined pose POS. The parking assistance system 100 is in particular for operating with the basis of 7 or 8th set up the procedure described. The based on 2 - 5 Aspects described form further embodiments of the parking assistance system 100.

• Erfassen S1 einer Anzahl spezifischer Merkmale 210 - 203 (siehe 1 oder 5) des Fahrzeugs 200 (siehe 1 - 4),
• Ermitteln S2 eines spezifischen Fahrzeugmodells MSPEC (siehe 5 oder 8) für das Fahrzeug 200 auf Basis der erfassten spezifischen Merkmale 201 - 203,
• Empfangen S3 eines Sensorsignals von wenigstens einem Sensor 102A - 102F (siehe 1) der Anzahl 102 (siehe 6) ,
• Ermitteln S4 einer Pose POS (siehe 1) des Fahrzeugs 200 in Abhängigkeit des empfangenen Sensorsignals unter Verwendung des ermittelten spezifischen Fahrzeugmodells MSPEC, und
• Steuern S5 des Fahrzeugs 200 in Abhängigkeit der ermittelten Pose POS.

7 FIG. 1 shows a schematic block diagram of a first exemplary embodiment of a method for operating a parking assistance system 100, for example parking assistance system 100 in FIG 1 or 6 . The parking assistance system 100 is, for example, as shown in FIG 1 or 6 configured as described. The procedure includes the steps:

• Detection S1 of a number of specific characteristics 210 - 203 (see 1 or 5 ) of the vehicle 200 (see 1 - 4 ),
• Determine S2 of a specific vehicle model MSPEC (see 5 or 8th ) for the vehicle 200 based on the recorded specific features 201 - 203,
• Receiving S3 of a sensor signal from at least one sensor 102A - 102F (see 1 ) of the number 102 (see 6 ) ,
• Determination S4 of a pose POS (see 1 ) of the vehicle 200 as a function of the received sensor signal using the determined specific vehicle model MSPEC, and
• Controlling S5 of the vehicle 200 depending on the determined pose POS.

8th FIG. 1 shows a schematic block diagram of a second exemplary embodiment of a method for operating a parking assistance system 100, for example parking assistance system 100 in FIG 1 or 6 . The parking assistance system 100 is, for example, as shown in FIG 1 or 6 configured as described. In a first step S10, the vehicle to be controlled reaches 200 (see FIG 1 - 4 ) for example the transfer area 112 (see 1 ), in which control of the vehicle 200 is transferred from a user of the vehicle 200 to the parking assistance system 100. In the transfer area 112 there are in particular sensor devices 103A, 103B (see 1 , 3 or 4 ) that sense vehicle 200 and output specific sensor data. The specific sensor data contain in particular specific features 210 - 203 (see 1 or 5 ) of the vehicle 200. On the basis of the specific sensor data, for example general features relating to the respective vehicle type, such as the geometry of the vehicle, can be determined and provided S12A. Alternatively, this data can also be retrieved from a database. Furthermore, the specific features 201-203 are provided in a data record S12B which is suitable for converting a general vehicle model MGEN into a specific vehicle model MSPEC.

The general vehicle model MGEN can also be referred to as a template that is adapted according to the general characteristics of the vehicle 200 and the specific characteristics 201 - 203 of the vehicle 200 . Accordingly, the specific vehicle model MSPEC is determined and made available. The specific vehicle model MSPEC can in particular be a trained artificial intelligence that is based on determining a pose POS (see 1 ) of a vehicle 200 with the general features and the specific features 201 - 203 is trained. Since the specific vehicle model MSPEC is determined on the basis of the general vehicle model MGEN, the specific training is merely fine tuning, for example, which can be completed with a few iterations.

Vehicle 200 is then driven by parking assistance system 100 through predetermined area 110 (see FIG 1 ) controlled S13, wherein the vehicle 200 from the sensors 102A - 102F (see 1 ) is detected and the pose POS of the vehicle 100 is determined based on the corresponding sensor signals using the specific vehicle model MSPEC. The pose POS includes both the position and the alignment of the vehicle 200. Accordingly, the parking assistance system 100 can transmit suitable control commands to the vehicle 200, so that it moves along a determined trajectory TR (see 1 ) emotional.

In step S14, vehicle 200 reaches the controlled end position, for example a parking position. The specific vehicle model MSPEC is now stored in a permanent memory from a working memory of the parking assistance system 100 . If the vehicle 200 is requested again by the user S15, the parking assistance system 100 loads the stored specific vehicle model MSPEC from the permanent memory into the main memory, so that it is again available for determining the pose POS for driving from the parking lot to the transfer area 112 stands. In step S16, the vehicle 200 reaches the transfer area 112, where the user takes control of the vehicle 200 again. The specific vehicle model MSPEC can now be deleted or it can be stored permanently, which can be particularly advantageous for private parking garages.

Although the present invention has been described using exemplary embodiments, it can be modified in many ways.

reference list

100

parking assistance system

102

number of sensors

102A - 102F

sensors

103A, 103B

sensor devices

104

illuminants

110

predetermined area

112

transfer area

120

control device

122

registration unit

124

investigation unit

126

receiving unit

128

control unit

200

vehicle

201 - 203

features

I

position

II

position

cls

vehicle class

ID

vehicle identifier

M1 - M3

vehicle models

MGEN

generic vehicle model

MSPEC

specific vehicle model

POS

pose

R1-R4

rays of light

RR1, RR3

reflected light rays

TR

trajectory

S1 - S5

process steps

S10 - S16

process steps

S12A, S12B

process steps

SPEC

specific record

SR2, SR4

scattered rays of light

Claims (15)

Method for operating a parking assistance system (100), which comprises a number (102) of sensors (102A- 102F) arranged in a predetermined area (110), each for detecting a vehicle (200) present in the predetermined area (110) and are set up to output a respective sensor signal, the method comprising the steps: a) detecting (S1) a number of specific features (201-203) of the vehicle (200), b) determining (S2) a specific vehicle model (MSPEC) for the vehicle (200) on the basis of the recorded specific features (201-203), c) receiving (S3) a sensor signal from at least one sensor (102A - 102F) of the number (102), d) determining (S4) a pose (POS) of the vehicle (200) as a function of the received sensor signal using the determined specific vehicle model (MSPEC), and e) controlling (S5) the vehicle (200) depending on the determined pose (POS). procedure after claim 1 , characterized in that step a) comprises: sensory detection of the vehicle (200) for providing specific sensor data for the vehicle (200). procedure after claim 1 or 2 , characterized in that step a) comprises: receiving a specific data set (SPEC) for the vehicle (200) comprising the specific features (201-203) from the vehicle (200), or receiving a vehicle identifier from the vehicle (200) and retrieving a specific data record (SPEC) for the vehicle (200), which includes the specific features (201 - 203), from a database (300) depending on the received vehicle identifier, or retrieving a specific data record (SPEC) for the vehicle (200), which includes the specific features (201 - 203), from a database (300) depending on the specific sensor data. procedure after claim 2 or 3 , characterized by : determining the specific features (201 - 203) depending on the specific sensor data. Procedure according to one of claims 2 until 4 , characterized in that the sensory detection of the vehicle (200) includes a detection of the vehicle (200) by means of a number of sensor devices (103A, 103B) from different directions, wherein the number of sensor devices (103A, 103B) at least one visual camera and /or at least one infrared camera and/or at least one laser scanner and/or at least one lidar and/or at least one radar. procedure after claim 5 , characterized in that at least one sensor device (103A, 103B) is rotated and/or moved relative to the vehicle (200) during the sensory detection of the vehicle (200). Procedure according to one of claims 2 until 6 , characterized in that the vehicle (200) is illuminated with different lighting patterns during the sensory detection. Method according to one of the preceding claims, characterized in that step b) comprises: training a detector for determining the pose (POS) of the vehicle (200) on the basis of the determined specific vehicle model (MSPEC) and/or the specific features (201 - 203). Method according to one of the preceding claims, characterized in that step b) comprises determining a kinematic vehicle model for the vehicle (200). Method according to one of the preceding claims, characterized by : determining a second pose (POS) of the vehicle (200) as a function of the received sensor signal using a general driving zeugmodells (MGEN), and plausibility of the determined pose (POS) depending on the determined second pose. Method according to one of the preceding claims, characterized by : providing a number of vehicle class models (M1 - M3), a respective vehicle class model (M1 - M3) comprising a vehicle model for a respective vehicle class (CLS), determining a vehicle class (CLS ) to which the vehicle (200) belongs, and selecting the vehicle class model (M1-M3) corresponding to the vehicle class (CLS), wherein the specific vehicle model (MSPEC) is determined on the basis of the selected vehicle class model (M1-M3). . Method according to one of the preceding claims, characterized in that the determined specific vehicle model (MSPEC) is stored in the predetermined area (110) after completion of an autonomous drive of the vehicle (200) controlled by the parking assistance system (100) and when the Vehicle (200) to the parking assistance system (100) is retrieved again. Computer program product, comprising instructions which, when the program is executed by a computer, cause the latter to carry out the method according to one of Claims 1 - 12 to execute. Parking assistance system (100) with a number (102) of sensors (102A - 102F) arranged in a predetermined area (110) for detecting a vehicle (200) present in the predetermined area (110) and for outputting a respective sensor signal, and having: a detection unit (122) for detecting a number of specific features (201 - 203) of the vehicle (200), a determination unit (124) for determining a specific vehicle model (MSPEC) for the vehicle (200) on the basis of the detected specific features (201 - 203), a receiving unit (126) for receiving a sensor signal from at least one sensor (102A - 102F) of the number (102), wherein the determination unit (124) for determining a pose (POS) of the vehicle (200) as a function of the received sensor signal using the determined specific vehicle model (MSPEC) is set up, and a control unit (128) for controlling the vehicle (200) as a function of the determined pose (POS). parking assistance system Claim 14 , characterized in that the detection unit (122) comprises a number of sensor devices (103A, 103B), wherein the sensor devices (103A, 103B) in a transfer area (112) in which a user of the vehicle (200) the vehicle (200) to the parking assistance system (100) and are set up to detect the vehicle (200) from different directions and to output respective specific sensor data, the number of sensor devices (103A, 103B) having at least one visual camera and/or at least one infrared camera -Camera and/or at least one laser scanner and/or at least one lidar and/or at least one radar, wherein the detection unit (122) for determining the specific features (201-203) of the vehicle (200) on the basis of the specific sensor data is set up.

Images