GB2516738A

GB2516738A - Assistance of a driver of a motor vehicle

Info

Publication number: GB2516738A
Application number: GB1409260.5A
Authority: GB
Inventors: Holger Mielenz; Christian Heigele
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2013-05-27
Filing date: 2014-05-23
Publication date: 2015-02-04
Anticipated expiration: 2034-05-23
Also published as: GB201409260D0; DE102013209764B4; GB2516738B; FR3005923B1; DE102013209764A1; FR3005923A1

Abstract

The present invention relates to a method for assisting a driver of a motor vehicle (10 see fig 2), comprising a step of obtaining information about an environment (40 see fig 2) of the vehicle (10) and comparing the obtained information with stored environment data. The environment is stored in individual grid tile cells 1, 2, identifying whether for the obtained information a stored cell 1, 2 is a starting cell 1 for a stored automatic driving manoeuvre in the stored environment data. A notice is provided to the driver that an automatic driving manoeuvre can be carried out if the driver wishes this. The invention is at least partly aimed at reducing environment sensing demands.

Description

Description Title

Assistance of a driver of a motor vehicle

Prior art

The present invention relates to a method for assisting a driver of a motor vehicle. Furthermore, the invention relates to a computer program product which can perform such a method. The invention also relates to a contro].

apparatus which can perform the method for assisting the driver of the motor vehicle, and also to a motor vehicle which comprises the control appar-atus.

From the prior art it is known that a motor vehicle can autonomously carry out driving manoeuvres. Thus, for example, it is known from DE 10 2010 030 208 Al that a vehicle can autonomously perform standard manocuvres by storing a previously followed trajectory. For this, the driver of the motor vehicle merely has to activate a previously stored standard manoeuvre. In order to offer standard driving manoeuvres, it is necessary for a position of the vehicle to be determined, so that it can be identified whether automatic driving manoeuvres exist for the current position of the vehicle. In other words, it has to be established whother a trajectory whose starting point corresponds to a current position of the vehicle has already been followed.

According to the prior art, the environment of the

trajectory already followed is stored in such a manner that significant elements of the environment are extracted from environment measurements. These are subsequently stored in list form. A very storage-efficient way of detecting and storing a partial region of the environment thus exists.

However, it is not possible to perform collision-avoiding braking procedures or avoidance procedures using these lists. The vehicle thus always has to carry out a detailed detection of the environment when an automatic driving manoeuvre is performed.

Disclosure of the invention

The method for assisting a driver of a motor vehicle according to the invention comprises the following steps: Firstly, information about an environment of the motor vehicle is obtained. This information can be obtained, for example, by means of detecting the environment by sensors.

Preferably, ultrasonic sensors or imaging sensors can he used as the sensors. The information thus obtained is subsequently compared with stored environment data, the environment to be stored being stored in individual cells.

The individual cells are, in particular, GridTiles. In a next step, it is identified whether for the obtained information a stored cell is present as a starting cell for a stored automatic driving manoeuvre in the stored environment data. If this is the case, in a last step, the driver of the motor vehicle is informed that an automatic driving manoeuvre can be carried out. The driver can then communicate whether or not he wishes the automatic driving manoeuvre to be carried out. The use of GridTiles here is known in principle from the prior art. However, in the present invention, it is possible that, to identify a position of the vehicle at which the latter can carry out an automatic driving manoeuvre, can bo checked merely using the cells. Thus, in particular, no complex localisation, for example by GPS, has to be carried out. It is merely necessary to check whether the detected environment is presented in a cell stored as a starting cell. Since each cell preferably presents on].y a limited region of the environment of the vehicle, such a search can be carried out in a very resource-saving manner. Nevertheless, with the method according to the invention, it can be safely and reliably identified whether an automatic driving manoeuvre exists for the current position of the vehicle. The safety and convenience for the driver is therefore markedly increased.

In particular, the direct environment of the starting position of a Homezone trajectory is modelled by a dedicated starting tile in the form of a GridTile (starting cell) -The identification and loading of a llomeZone scenario is then preferably realised by a simple list of such starting tiles. If one of these starting tiles is identified as a "starting point of a HomeZone trajectory" by the local environment data, the rest of the environment model can be loaded from permanent storage iteratively by the GridTile method. This results in significant speed and storage advantages, but nevertheless a detailed environment model is advantageously available, on which braking trajectories and/or avoidance trajectories can be planned.

Preferred developments of the invention are contained in the subclaims.

Preferably, a trajectory of the stored automatic driving manoeuvre is stored in at least one cell. For a navigation during the performance of the automatic driving manoeuvre, it is thus possible to have recourse to the stored cells.

Preferably, each cell has features which are used to identify whether the environment of the motor vehicle is stored in one of the cells. Such features can be, in particular, edges, shapes and/or objects. Alternatively or additionally, it is preferably provided that the features are used to establish how the motor vehicle is oriented relative to the respective cell. The method according to the invention thus allows a very quick identification of whether the vehicle is situated in an environment which is presented by a specific cell. Furthermore, by means of the method according to the invcntion, it can be established which crientation the motor vehic]. e has within the cell. A very accurate navigation within the individual cells is thus possible, Advantageously, the driver of the motor vehicle is shown an environment model before the request for confirmation of the performance of the automatic driving manoeuvre. The environment model, is preferably generated from the stored environment data. The driver can thus see that his motor vehicle correctly assesses the current environment, with the result that the driver can have greater confidence in his vehicle for correctly carrying out the automatic driving manoeuvre. The driver can thus advantageously see how the vehicle will move for carrying out the automatic driving manoeuvre, with the result that he can be sure that his vehicle will be safely and reliably guided. The driver's confidence is particularly preferably increased also by the fact that the environment model is also generated from such stored environment data which cannot be detected by sensors from the current position of the vehicle. He can thus be sure that the current situation has been correctly identified.

Particularly preferably, it is provided that the environment model shown to the driver is generated at least from those cells which are traversed on the performance of the automatic driving manoeuvre. If, as described above, the trajectory of the automatic driving manoeuvre has been stored in the cells, the trajectory can also be displayed to the driver. Likewise, it is particularly advantageously possible for the driver to be shown an environment model in which there are integrated also such cells which are located adjacent to the traversed cells. The driver can therefore be shown a comprehensive environment model, so that the driver's confidence in the vehicle can be markedly increased.

Advantageously, based on the stored environment data, a vehicle reaction is performed. Such vehicle reactions are, in particular, collisionavoiding vehicle reactions, such as, for example, braking manoeuvres or steering manoeuvres.

Since any obstacles can already be detected early by the method according to the invention, it is possible for the vehicle to avoid obstacles which cannot yet be detected by means of sensors of the vehicle. An optimal trajectory can therefore be planned.

Advantageously, a storage of the automatic driving manoeuvre and/or of the stored environment data takes place in such a manner that a preset trajectory is initially followed. That cell which represents a starting point of the trajectory, in which a starting point of the trajectory thus lies, is stored as a starting cell. By following the preset trajectory, the movement carried out can be stored, so that on carrying out the automatic driving manoeuvre, the same movement or a slightly modified movement is carried out once again. The storage of the environment is preferably effected by means of individual cells, so that a minimal storage space is required for the environment data.

This is preferably achieved in that each cell has only small dimensions, so that the entire environment data consists of a plurality of cells, a new cell always being added when the vehicle moves in a region which has hitherto not been covered by a cell.

The invention further relates to a computer program product which has a program code for carrying out the above-mentioned method. The carrying-out of the method takes place when the computer program product runs on a computer.

A control apparatus, in particular, is also to be regarded as a computer here.

The present invention also relates to a control apparatus for a motor vehicle which comprises a control device. The control device is connected to further components of the vehicle, so that the control apparatus can preferably cause a movement of the motor vehicle. Furthermore, it is preferably provided that the control device is connected to sensors of the vehicle, so that the control apparatus according to the invention can detect an environment of the motor vehicle. Finally, the control device is adapted such that this device carries out the aforementioned method according to the invention.

Lastly, the invention relates to a motor vehicle which comprises the above-mentioned control apparatus.

Brief description of the drawings

Exemplary embodiments of the invention are described in detail below with reference to the accompanying drawings.

In the drawings: Figure 1 is a schematic representation of the stored environment data, as used in a method according to a preferred exemplary embodiment of the invention, and Figure 2 shows a schematic representation of a motor vehicle having a control apparatus according to an exemplary embodiment of the inven Lion.

Exemplary embodiments of the invention Firstly1 a first exemplary embodiment of the invention is described without reference to the drawings, In this case, provision is made for realising a HomeZone function with a significantly more detailed stored environment model by using multiple starting tiles. In this case, for each HomeZone trajectory entered, an environment model having cells realised by GridTiles is created, a dedicated starting tile as the starting cell modelling the direct environment of the starting position starting point) of the HomeZone trajectory.

During the initial following, i.e. the teach-in procedure, of the HomeZone trajectory, a GridTile environment model is constructed and written to permanent storage together with the trajectory on completion of the teach-in procedure. The dedicated starting tile, which is connected to the rest of the environment model via its neighbourhood relations, is additionally stored in a separate list, All dedicated starting tiles for all taught-In HomeZone scenarios associated with the vehicle are stored in this list.

For the identification of a Homezone scenario, it is now only necessary to compare this finite list with the currently availablo local environment data. If one of these dedicated starting tiles Is "recognisod" by local environment data, this tile can, via the neighbourhood relations inert to it, iteratively load the environment model stored in the GridTile format in the permanent storage available in tho vehicle, including lhe associated HomoZone trajectory.

The advantage of this method lies in the great computing efficiency and storage efficiency, since only a small environment model in the form of an individual GridTile has to be compared with the local measurement data.

Nevertheless1 via such a realisation, an environment model can be loaded which has such a high degree of detailing that collision-prevenLing braking or reaction manoeuvres can be planned thereon. An additional advantage consists in that, by presentation of the environment model in a human-machine interface, greater confidence in the Homezone system can be produced in the user, since he sees clearly that the system has correctly identified the scenario. For this purpose, the HomeZone trajectory and the loaded environment model are simultaneously shown in an appropriate human-machine interface. In particular, the driver's confidence is increased by the fact that the GridTiles also contain information of the static environment which cannot be detected and displayed by means of onboard sensors at the current instant of the Homezone request.

B

Further details of the exemplary embodiment are: Provision is advantageously made to store environment features on the dedicated starting tiles in addition to the pure environment data. These features can accelerate a renewed redetection and thereby simplify the localisation. The combination of raw environment data and features enables a large increase in performance here.

-In addition, a repositioning owing to the virtual movement onto a new tile can be implemented. A more accurate guidance is thereby achieved.

-These two ideas can be combined if environment features for localisation are stored also on the following tiles, not only the starting tile.

The features on the starting tile require particular unambiguousness, since the 1-Jomezone function is to be available not only at its actually assigned position.

The features on the following tiles can be simpler.

Thus, for example, wall elements suffice here to enable orientation, whereas an individual wall element is not unambiguous enough for the determination of the starting position. According to the invention, the type and number of localisation features and/or supporting features on the starting tile and the following tiles can therefore differ.

The Homezone trajectory can, according to the invention, be stored, subdivided into partial sections, on the individual associated tiles. As a result, storage advantages and/or localisation advantages can be obtained.

Figure 1 shows a schematic representation of stored environment data. The data here are stored in individual cells 2, so that the stored environment data remain very compact and thus require little storage space. Also present is a starting cell I which is constructed identically to the cells 2, but which is additionally identifiable as a starting cell. The starting cell 1 is distinguished by the fact that a starting point 31 of a trajectory 3 lies within the starting cell t the trajectory 3 enabling an automatic driving manoeuvre to be carried out. The automatic driving manoeuvre is carried out in such a manner that a vehic. e 10 (cf. Figure 2) to be guided is guided along the trajectory 3 from the starting point 31 to an end point 32.

In this case, obstacles i which occur are already known to the vehicle 10, since they are stored inside the cells 2.

The obstacles 4 also serve for detecting an orientation of the vehtcle 10 relative to the cells 2.

Also present in tbe.starting cell 1 is an obstacle 5 which is used to identify whether the current environment of the vehicle 10 is represented by the starting cell 1. If this is the case, the driver is informed that his vehicle is situatcd at a position at which it is possible to carry out an automatic driving manoeuvre with the trajectory 3.

If the vehicle 10 is situated within a predefined region around the starting point 31, the environment model consisting of starting cell 1 and the cells 2 is displayed to the driver of the vehicle. The driver can thus see that an automatic driving manoeuvre, with which the vehicle 10 is brought from the starting point 31 along the trajectory 3 to the end point 32, is possible. The driver can therefore be sure that the current environment has been correctly identified and the vehicle will be safely and reliably guided to the end point 32. He can thus be extremely confident in the vehicle when carrying out the automatic driving manoeuvre.

Collision-avoiding interventions in the movement of the vehicle can be carried out based on the stored obstacles 4.

Since the obstacles 4 are already known, although the vehicle is situated at a position at which it can detect the obstacles 4 only to a limited degree or not at all using its sensors, colLision-avoiding movements can already be planned very early, so that an optimal trajectory 3 can be followed.

In order to be able to identify the starting cell 1 unambiguously, it is provided that the latter has detailed features with which the starting cell 1 is unambiguously identifiable. Only in this way is IL ensured that the vehicle 10 safely and reliably detects the environment presented in the starting cell 1. This degree of detaii.ing is, however, not required for the cells 2, so that the type and the number of the features of the cells 2 differ from those of the starting cell.

Figure 2 shows a vehicle 10 having a control apparatus 30 according to an exemplary embodiment of the invention. The control apparatus 30 is connected to a plurality of components of the vehicle 10, inter silo to four sensors 20. By means of the sensors 20 it is possible for the conLrol apparatus 30 to scan an environment 40 of the vehicle 10. Via a communication line 50 the control apparatus 30 is further connected to other components of the vehicle 10, so that the control apparatus 30 can control, in particular, driving movements of the vehicle 10. The control, apparatus 30 can therefore directly affect a movement of the vehicle 10.

The control apparatus 30 is adapted to carry out the above-described method according to the invention. For this purpose, the control apparatus uses the sensors 20 and/or acts on the vehicle Di via the communication line 50, in order to influence a movement of the vehicle 10.