DE102019205083A1 - Method for predicting a manual parking process and supporting a driver - Google Patents

Abstract

The invention relates to a method for predicting a manual parking process and supporting a driver of an ego vehicle (1) with the following steps: - Acquisition (S1) of surroundings data by means of at least one surroundings detection device; - Generation (S2) of a virtual map based on the surroundings data - Detection (S3) of parking spaces in the vehicle surroundings; - Entering (S4) the parking spaces in the virtual map as parking boxes (2); - Determining (S5) a value of an approach of the ego vehicle (1) to a parking box (2 ); - Determination (S6) of the current speed of the ego vehicle (1); - Determination (S7) of an imminent manual parking process when a certain threshold value of the approach value is exceeded and when a current speed is below a specified threshold value; - Output (S8 ) a note to support the driver.

Description

The invention relates to a method for predicting a manual parking process and supporting a driver.

Assistive or automated parking systems are known from the prior art which, with the aid of the environment sensors, detect free parking spaces after driving past and, if necessary, offer the driver to park the vehicle in a partially or fully automated manner.

Furthermore, systems are known which recognize the parking space even during a manual parking maneuver started by the driver and offer the driver to terminate the parking process partially or fully automatically. With the existing systems, the driver must at least partially drive into the parking space so that the automated function can be offered. Reason: The ultrasonic sensors installed today only have a short range, which is why it is possible to differentiate between "manual parking", i.e. the detection that the driver is driving into a parking space, and "driver is maneuvering", e.g. to turn, only after measuring the bounding objects.

It is accordingly an object of the invention to provide a method which ensures a more reliable and earlier recognition of a manual parking process.

This object is achieved by the subject matter of independent claim 1. Further advantageous configurations and embodiments are the subject of the subclaims.

The first considerations were that in existing systems, the parking maneuver started manually by the driver is usually recognized by a heuristic based on data from vehicle sensors (e.g. steering movements, current driving speed, etc.) and ultrasound environment sensors (e.g. distance measurement to neighboring objects) based. With existing systems, the driver usually has to drive very deep into the parking space so that “manual parking” can be recognized. This is also due to the short range of the sensors currently in use.

If the dynamic driving values match the profile of a parking maneuver and the distances to neighboring objects are sufficiently dimensioned, the current maneuver is recognized as a parking maneuver. This evaluation requires a constant load on the control units and can lead to incorrect identifications.

If a virtual map is generated from the surrounding data, it can also be used to easily and reliably identify a manual parking maneuver. If the card is required for the parking system anyway, the additional computational effort for recognizing a manual parking maneuver is very low.

• - Erfassen von Umfelddaten mittels zumindest einer Umfelderfassungsvorrichtung;
• - Generieren einer virtuellen Karte basierend auf den Umfelddaten;
• - Erkennung von Parklücken in dem Fahrzeugumfeld;
• - Eintragen der Parklücken in die virtuelle Karte als Parkboxen;
• - Ermitteln eines Wertes einer Annäherung des Ego-Fahrzeugs an eine Parkbox;
• - Ermitteln der aktuellen Geschwindigkeit des Ego-Fahrzeugs;
• - Feststellen eines bevorstehenden manuellen Einparkvorgangs bei Überschreitung eines bestimmten Schwellenwertes des Wertes Annäherung sowie bei Vorliegen einer aktuellen Geschwindigkeit unterhalb eines festgelegten Schwellenwertes;
• - Ausgeben eines Hinweises zur Unterstützung des Fahrers.

According to the invention, a method for predicting a manual parking process and supporting a driver of an ego vehicle with the following steps is therefore proposed:

• - Detection of environment data by means of at least one environment detection device;
• - Generating a virtual map based on the environment data;
• - Detection of parking spaces in the vicinity of the vehicle;
• - Entering the parking spaces in the virtual card as parking boxes;
• - determining a value of an approach of the ego vehicle to a parking box;
• - Determining the current speed of the ego vehicle;
• Detection of an imminent manual parking process when a certain threshold value of the approach value is exceeded and when a current speed is below a specified threshold value;
• - Output of a note to support the driver.

This method is particularly advantageous because the mapping of the parking spaces means that there must be little or no overlap between the vehicle and the parking space in order to detect a manual parking process. Thus, the value of the approximation can be a small percentage of the coverage of the parking space with the vehicle, or a distance from the parking space or the parking box when the ego vehicle is moving in the direction of the parking space. This is particularly advantageous in scenarios when the parking space is in front of the vehicle, or the parking space was detected while driving past and the vehicle is approaching the parking space in reverse.

The threshold value for the approach is accordingly, for example, a small percentage of the coverage, for example 5-10%, which is determined, for example, by comparing the current vehicle position with the mapped parking space in the virtual map. Alternatively or cumulatively, the threshold value can be a certain distance between Vehicle and parking space, for example, less than or equal to 1 meter. It would also be conceivable to engage reverse gear and / or to define a steering angle as a value for the approach when parking spaces are detected on the side.

The surrounding data include, for example, objects and markings that limit the freely maneuverable space for the vehicle. These can be, for example, parked vehicles, curbs, street lamps or parking lot markings. These data are entered in the virtual card. The position of free parking spaces between the recognized objects or the markings, the dimensions of which offer sufficient space for the ego vehicle, are then determined and entered in the virtual map as parking boxes. The virtual map also contains the relative position of the objects to one another and to the current position of the ego vehicle.

The advice to assist the driver can be, for example, an indication of whether the parking space is of sufficient size. Since the size of the vehicle is known and the parking spaces are measured when they are recognized and entered in the virtual map, this information can be made available to the driver at an early stage as support. Another possible indication would be, for example, the indication of an optimal steering angle depending on the position of the vehicle.

In a preferred embodiment, the surroundings data are recorded by means of at least one radar sensor and / or an ultrasonic sensor and / or a camera. With the help of the camera, for example, parking lot markings can be recognized. This has the advantage that a parking lot can already be deduced from this without any delimiting objects having been detected by means of the ultrasonic sensor and / or radar sensor. The ultrasonic sensor is preferably a sensor with a range of 7 meters. Furthermore, the size of the parking space can be determined when parking markings are at least partially recognized.

In a further preferred embodiment, a valid parking end position is determined after the manual parking process has been determined. A valid parking end position is determined, for example, by checking whether the vehicle is in this position within two parking lot markings.

Furthermore, after the manual parking process has been determined, at least one collision-free trajectory between a current position and a potential parking end position is preferably determined. The trajectory can consist of several partial trains and contain changes of direction.

Particularly preferably, after a manual parking process has been determined and a parking end position and a possible trajectory have been determined, the driver of the ego vehicle is informed about the feasibility of a partially or fully autonomous parking maneuver. A partially autonomous parking maneuver would, for example, include autonomous steering, that is, following the planned trajectory. Accelerating and decelerating and, for example, engaging reverse gear would be the task of the driver in this embodiment. The fully autonomous parking maneuver includes the autonomous execution of all activities until the end of parking position is reached.

The partially autonomous or fully autonomous parking maneuver can preferably be confirmed by the driver. The driver can thus decide whether he would like to make use of the support offered.

The recognized parking space is particularly preferably classified based on the alignment with the ego vehicle. The parking space is classified in particular as a parallel, inclined or vertical parking space. Such a classification is advantageous because a trajectory can be planned more reliably based on this information, since it is so clear how the vehicle must enter the parking space.

Further advantageous configurations and embodiments emerge from the drawings.

• 1: eine schematische beispielhafte Darstellung einer virtuellen Karte gemäß einer Ausführungsform der Erfindung;
• 2: ein schematisches Ablaufdiagramm eines Verfahrens gemäß einer Ausführungsform der Erfindung.

Show in it:

• 1 : a schematic exemplary representation of a virtual map according to an embodiment of the invention;
• 2 : a schematic flow diagram of a method according to an embodiment of the invention.

In 1 a schematic exemplary representation of a virtual map according to an embodiment of the invention is shown. There is a vehicle in this map 1 shown which in one as a parking box 2 enters the parking space entered in the virtual map. The parking box 2 shows parking lot markings on both sides 3 on. Like in the 1 further shown, has the vehicle 1 in this case already a certain excess 4th with the parking box 2 on. At this point it would be due to the excess coverage 4th a manual parking process can be detected at the appropriate speed. However, there is an overlap 4th not absolutely necessary to infer a manual parking process. Depending on the embodiment, a certain distance from the parking space would also be sufficient, for example. There are also several objects on the map 5a-5c entered, which are located around the parking space and can possibly influence the trajectory of the parking process. The objects 5a-5c are not specified in more detail and can therefore be other vehicles, trees or other static or moving objects.

2 shows a schematic flow diagram of a method according to an embodiment of the invention. In step S1 environment data are recorded by means of at least one environment detection device. In step S2 a virtual map is generated based on the surrounding data. In a further step S3 parking spaces are detected in the area around the vehicle, which then step S4 as parking boxes 2 entered in the virtual card. After that, in step S5 a value for an approach of the ego vehicle 1 to a parking box 2 determined. Furthermore, the current speed of the ego vehicle is determined at the same time or after the approach is determined 1 in one step S6 determined. If a certain threshold value of the approach is exceeded and a current speed is below a specified threshold value, step S7 an imminent manual parking process can be determined. Finally, in step S8 a message can be issued to the driver, who assists him in the parking process.

List of reference symbols

1

Ego vehicle

2

Parking box

3

Parking lot markings

4th

Overlap

5a-5c

object

S1-S8

Procedural steps

Claims (8)

Method for predicting a manual parking process and supporting a driver of an ego vehicle (1) with the following steps: - Detection (S1) of environment data by means of at least one environment detection device; - Generating (S2) a virtual map based on the environment data; - Detection (S3) of parking spaces in the vehicle environment; - Entering (S4) the parking spaces in the virtual card as parking boxes (2); - Determining (S5) a value of an approach of the ego vehicle (1) to a parking box (2); - Determining (S6) the current speed of the ego vehicle (1); - Establishing (S7) an imminent manual parking process when a specific threshold value of the approach value is exceeded and when a current speed is below a specified threshold value; - Output (S8) a note to support the driver. Procedure according to Claim 1 , characterized in that the environment data are recorded by means of at least one radar sensor and / or an ultrasonic sensor and / or a camera. Procedure according to Claim 1 , characterized in that after the manual parking process has been determined, at least one possible valid parking end position is determined. Method according to one of the preceding claims, characterized in that after the manual parking process has been determined, at least one collision-free trajectory between a current position and a potential parking end position is furthermore determined. Procedure according to the Claims 3 and 4th , characterized in that after a manual parking process has been determined and a parking end position and a possible trajectory have been determined, a message is issued to the driver of the ego vehicle (1) about the feasibility of a partially or fully autonomous parking maneuver. Procedure according to Claim 5 , characterized in that the partially autonomous or fully autonomous parking maneuver can be confirmed by the driver. Method according to one of the preceding claims, characterized in that the recognized parking space is classified based on the orientation to the ego vehicle. Procedure according to Claim 7 , characterized in that the parking space is classified as a parallel, diagonal or vertical parking space.

Images