DE102019127140A1 - Determination of a pitch angle of an environmental sensor of a vehicle - Google Patents

Abstract

The invention relates to a method for determining a pitch angle (α) of an environmental sensor (16) in a mounted state, in particular in a state mounted on a vehicle (10), comprising the steps of providing a two-dimensional information field (28) with sensor information and determining the Pitch angle (α) based on a vertical angular position of a detected vehicle (22) in the sensor information. The invention also relates to a driver assistance system (12) for a vehicle (10) with a control unit (14) and an environment sensor (16) which is designed to provide a two-dimensional information field (28) with sensor information, the control unit (14) being designed to carry out the above method for determining a pitch angle (α) of an environmental sensor (16). The invention also relates to a vehicle (10) with the above driver assistance system (12).

Description

The present invention relates to a method for determining a pitch angle of an environmental sensor on a vehicle. The present invention also relates to a driving assistance system which is designed to carry out such a method. The present invention further comprises a computer program for carrying out such a method.

In driving assistance systems, environmental sensors are used to detect the surroundings of a vehicle. This is a prerequisite, for example, in order to identify routes and to recognize obstacles in the area of the routes. At least in some of these driving support systems, it is important that the environmental sensors are aligned straight or in a known manner in order to be able to reliably evaluate sensor information from these environmental sensors.

Accordingly, it is necessary that the environmental sensors are precisely aligned when they are installed on the vehicle. Incorrect pitch angles occurring during assembly, for example due to incorrect assembly of the environmental sensors on the vehicle, lead to permanent restrictions in the detection of the surroundings during operation. In addition, due to component scatter and / or play during assembly, incorrect alignments of the environmental sensors can occur, which can also be problematic. In order to remove these restrictions, it may therefore be necessary to compensate for an existing or incorrect pitch angle. The pitch angle is usually understood to be an angle in the direction of travel of the vehicle in relation to a flat floor surface, for example the horizon. The pitch angle can assume positive or negative values depending on the orientation of the environment sensor on the vehicle, i.e. the environment sensor can see either too high or too deep.

If the environmental sensors are correctly installed, the vehicle can also have a pitch angle as a function of a current driving situation. This can lead to temporary restrictions in the recording of the environment during operation. To remove these restrictions, it may be necessary to compensate for an existing pitch angle.

In principle, there are various ways of compensating for the pitch angle. For example, the environmental sensors can be aligned to compensate for angles of inclination. After installing the environmental sensor, this can only be done with great effort. In principle, compensation is possible by means of an active, mechanical compensation, which, however, requires a great deal of constructive effort. Alternatively, pitch angles can be computationally compensated when processing the sensor information. In all cases it is necessary for the compensation to be able to determine the current pitch angle as precisely as possible.

DE 10 2017 106 484 A1 relates to a method for determining a pitch angle of an environmental sensor in a mounted state, in particular in a state mounted on a vehicle, comprising the steps of providing a two-dimensional information field with sensor information, and determining the pitch angle based on a vertical angular position of a detected vehicle in the sensor information.

In this context, for example, from the US 9 052 721 B1 Methods and systems for aligning data obtained with "Light Detection and Ranging" (LIDAR) are known. A computing device of a vehicle may be configured to compare a three-dimensional (3D) point cloud with a 3D reference point cloud in order to identify obstacles on a road. However, the 3D point cloud and the 3D reference point cloud can be misaligned. In order to align the 3D point cloud with the 3D reference point cloud, the computing device can be configured to determine a planar feature in the 3D point cloud of the road, as well as a corresponding planar feature in the 3D reference point cloud. Furthermore, the computing device can be configured to determine an image based on the comparison of the planar feature with the corresponding planar feature. The computing device can be configured to apply the transformation to align the 3D point cloud with the 3D reference point cloud.

On the basis of the above-mentioned prior art, the invention is therefore based on the object of providing a solution which can enable a simple and reliable determination of pitch angles.

The object is achieved according to the invention by the features of the independent claims. Advantageous refinements of the invention are given in the subclaims.

According to the invention, a method is thus specified for determining a pitch angle of an environmental sensor in a mounted state, in particular in a state mounted on a vehicle, with the step of providing a two-dimensional information field with sensor information, the method having a step of determining the pitch angle based on a comprises vertical angular position of a detected vehicle classified with respect to a vehicle class in the sensor information, and wherein the method comprises an examination of the detected vehicle with regard to a misclassification before the step of determining the pitch angle, the detected vehicle being used to determine the pitch angle only if there is no misclassification.

Such a method makes it possible in a particularly safe and reliable manner to detect a pitch angle of an environmental sensor. In the context of the present invention, the detection of a pitch angle should equally include that a deviation of the pitch angle from a basic position or a desired position is determined.

The method described here thus serves to detect a pitch angle of an environmental sensor or, in other words, to determine whether the environmental sensor is aligned in the vertical direction in the desired manner, in particular exactly horizontally, or whether the pitch angle is misaligned. In this case, the method can preferably be carried out when the environment sensor is already arranged or mounted on a vehicle.

The method comprises the step of providing a two-dimensional information field with sensor information. The two-dimensional information field is a field with sensor information which, like an image, provides the sensor information, for example in rows and columns. This can be, for example, two-dimensional image information. Alternatively, three-dimensional image information that additionally contains depth information of image points can also be provided in this way, so that a two-dimensional information field is to be understood as an at least two-dimensional information field. Such information can be generated with a stereo camera, for example. Similarly, depth information can be provided with a laser scanner, for example.

It is further provided that the method includes a step of determining the pitch angle based on an in particular distance-dependent vertical angular position of a detected vehicle classified with regard to a vehicle class in the sensor information.

The present method is thus based on the basic idea of determining the pitch angle by determining a position of the detected vehicle in the information field of the environmental sensor. Detected vehicles in the area are well suited for this, since the detected vehicles mostly drive on flat roads. Accordingly, it is assumed as a working assumption that both vehicles, i.e. the own vehicle on which the environmental sensor is attached, and the detected vehicle, are in the same horizontal plane. As a result, the pitch angle can be easily determined from the vertical angular position of the detected vehicle, as it is determined with the environment sensor, since the vertical angular orientation of the detected vehicle is known. This vertical angular alignment corresponds to a reference angle and can be defined as zero, for example.

In addition, detected vehicles are often easy to identify in the sensor information, since, as already stated, the subsurface corresponds at least approximately to a horizontal plane. It is also assumed that detected vehicles are typically located on a lane of a street or the like, so that the occurrence of a few bumps on the ground can be expected, and that larger bumps can be other vehicles. This makes it easier to process the sensor information in order to determine the pitch angle. It is not necessary for the method to determine the vertical angular alignment of the detected vehicle in detail. It is only necessary to determine a measure that results from the vertical angular alignment of the detected vehicle. With additional knowledge of the distance to the detected vehicle, the pitch angle can be determined particularly easily, for example. The pitch angle is preferably determined in relation to a desired orientation of the environmental sensor.

In an advantageous embodiment of the invention, the step of determining the pitch angle based on a vertical angular position of a detected vehicle in the sensor information comprises determining a last line in the sensor information, in which the detected vehicle is still contained, for a given distance. In other words, what is known as a last-hit distance is determined. The observation can take place in a line-by-line view from top to bottom or from bottom to top. On the basis of an assumed similar distribution of different vehicle dimensions, a measure for the vertical angular alignment of the vehicle can thus be determined without having to carry out a measurement.

In an advantageous embodiment of the invention, the step of determining a last line in the sensor information, in which the detected vehicle is still included, comprises determining a ratio of points in each line that are related to the detected vehicle belong to the maximum number of points in a line that belong to the detected vehicle, wherein a limit value of the ratio of the number of points in a line to a maximum number of points in a line indicates that the detected vehicle is still contained in a line or is no longer included. The point ratio is thus determined for all lines in the vertical direction. As a result, the last line, in which the detected vehicle is still contained, can be determined simply and independently of a specific measurement of the vehicle. As soon as the ratio falls below the limit value, the recorded vehicle is "lost", ie the line in which the vehicle is no longer contained has been found. The same applies to the last line in which the vehicle is still included. The ratio of the maximum points per line recorded by the vehicle compared to the maximum number of points in a line, in each case related to the recorded vehicle, is also referred to as "measured points". For example, based on a maximum number of points in a line that capture the rear of a captured vehicle, when a vertical angular position is reached, so few points can lie on the rear of the captured vehicle that the captured vehicle is lost. The maximum number of points in a line which, for example, cover the rear of a detected vehicle, can be easily determined, for example, based on a width of the bounding box. The maximum number of points in a line, as well as the respective number of points in a line which cover the vehicle, depends on a distance from the detected vehicle. By determining the ratio, it is easy to determine when a detected vehicle is lost, regardless of the distance.

In an advantageous embodiment of the invention, the method comprises a step of creating a histogram, the ratio of points in the line belonging to the detected vehicle to the maximum number of points in a line belonging to the detected vehicle in the histogram for each line belong, depending on the distance is entered. Due to the specific point distribution of the rows belonging to the vehicle, the rear or the front of the detected vehicle can be determined in detail in the sensor information, from which the determination of the vertical angular alignment can be carried out with a high degree of accuracy. The ratio is entered into the histogram for different distances, for example in a range from 10m to 100m.

Alternatively, in a simpler determination of the pitch angle, only the first line in the sensor information, in which the detected vehicle is no longer contained, can be entered for a given distance in the histogram. The method must then determine the first line in which the detected vehicle is no longer included.

In an advantageous embodiment of the invention, the method includes a step of providing a look-up table in which the greatest distance is entered for different pitch angles, in which a horizontal line still records a detected vehicle, ie the “last hit distance”. The use of a look-up table enables the pitch angle to be determined quickly and easily. The determination of a value in the look-up table requires only a small amount of computation.

In such a method, it can be important that the detected vehicles have a comparable height according to an assumption in order to be able to obtain uniform information from the vertical angular position even with a large number of the detected vehicles. Accordingly, the method is based, in particular, on the assumption that all the vehicles recorded can be subsumed into a common vehicle class and thus have a comparable amount.

In this regard, the described method includes, before the step of determining the pitch angle, an examination of the detected vehicle with regard to a misclassification, wherein the detected vehicle is only used to determine the pitch angle if there is no incorrect classification.

In other words, the method described here examines whether the detected vehicle or whether the detected vehicles can be subsumed into the vehicle class on which the method is based as the assumption. For example, the method can include the assumption that the detected vehicles are all cars. This is because the assumption can be made that the largest number of vehicles in traffic are cars, so that the method can be designed to be particularly reliable if it is assumed that the vehicles are cars. It has been found, however, that such a basic assumption can be disadvantageous such that the detected vehicles do not all have the same height and therefore a vertical angle estimate can possibly provide a limited robustness.

Because by using a vertical angle in the method described here, the result of the method is dependent on the height of the vehicles. If all vehicles have the same or comparable height, good results can be expected. However, if the heights of the vehicles differ greatly, the result can be to be disabled. Nowadays, however, there is no clear delimitation with regard to the height of the vehicles and, for example, sports cars have a different height than, for example, SUVs, off-road vehicles, vans or transporters, an algorithm for determining the pitch angle may provide results with limited reliability when recording different vehicles, if all vehicles are classified uniformly. A uniform classification should include the assignment of the vehicles to a vehicle class, the vehicle class being fundamentally selectable and, in particular, being able to have vehicles with a comparable height.

However, the described method safely overcomes such disadvantages with regard to different heights of vehicles. Pre-processing is thus made possible, which examines the detected vehicles with regard to a correct classification and, if the result is that a vehicle does not correspond to the desired classification, does not take it into account for the further method. For example, preprocessing can be made possible to the effect that it is detected whether a vehicle is likely or unlikely to be a car. Accordingly, only those vehicles can be used for the further process that are likely to be a car.

Following the above, a pitch angle or the deviation of a pitch angle can thus be detected particularly reliably by the method described here and used, for example, for further method steps of a driving assistance system. In particular, the environment sensor can be recalibrated based on the detected pitch angle, such as the pitch angle deviation, in order to ensure a high quality of driving assistance.

Provision can preferably be made for the detected vehicle to be examined with regard to a misclassification in such a way that the detected vehicle is classified as a classified vehicle or a non-classified vehicle. In this embodiment, a particularly simple examination with regard to a misclassification can take place, since the result of this examination is limited to two parameters. It can thus be determined whether the vehicle belongs to the selected classification with a predominant probability, or whether the vehicle does not belong to the classification with a predominant probability. For example, it can be determined whether a detected vehicle is a passenger car with a predominant probability, or whether the vehicle is not a passenger car with a predominant probability.

The simplicity of this evaluation makes it possible to use simple algorithms for this process that are not very error-prone and furthermore only require limited computing power.

It can furthermore be preferred that, during an examination of the detected vehicle with regard to a misclassification, two groups are formed which are defined by the distance between the vehicles and the environmental sensor, in particular with the misclassification being affirmed or denied depending on the group to which the vehicle belongs is. In this refinement, the examination with regard to a misclassification can be further simplified. For example, but not limited to this, if the vehicles are classified as passenger cars, for example, it can be assumed that if a vehicle is detected at a great distance whereas the majority of vehicles are detected in the close range, the vehicle present in the long range is most likely not Is a car, so it is misclassified and does not need to be used. On the other hand, if most of the vehicles are detected in a long range, then all vehicles should be classified as passenger cars and misclassification of individual vehicles can be excluded. In this case, it can already be assumed that there is no incorrect alignment of the environmental sensor.

The examination of the detected vehicle with regard to a misclassification can preferably include the application of a k-means algorithm. A k-means algorithm is to be understood as a method for vector quantization that is also used for cluster analysis. A previously known number of k groups is formed from a set of similar objects. Such an algorithm is well known to the person skilled in the art and can be created accordingly without any problems.

In particular in an embodiment in which two groups are formed during an examination of the detected vehicle with regard to a misclassification, which are defined by the distance between the vehicles and the environment sensor, two groups are thus formed, namely a short-range cluster and a long-range cluster. This allows the algorithm to be kept very simple. For example, with regard to the distance, a distance range can be mapped which can range from a distance 0, that is to say directly to the environmental sensor, up to the maximum detection distance of the environmental sensor. For a LIDAR, for example, this range can be 150 m, for example, but in no way restrictive. The boundary between the near area and the far area can in principle be selected and can be determined by a person skilled in the art. For example, the boundary between the near range and the far range can be half of the maximum range.

This refinement allows a reliable assessment of a misclassification in a simple manner and with little computing power.

After applying the k-means algorithm, the correctly classified vehicles can be used to determine the pitch angle or the deviation in the pitch angle.

With regard to the k-means algorithm, it can be provided that it is based on a distance-based and, with regard to a vertical angle, angle-dependent detectability of the vehicle. In this embodiment, the so-called “last hit distance” of vehicles can thus be recorded and used with regard to the detectability of a vehicle. This “last hit distance” describes the distance up to which the determined vehicle can still be detected by the environmental sensor from the environment sensor.

The step of identifying the detected vehicle in the map of the surroundings further advantageously comprises the detection of a rear or a front of the detected vehicle. The rear or the front of the detected vehicle can, for example, simply be determined in the map of the surroundings as the front or rear delimitation of the vehicle driving ahead or following. The detection of a rear or a front of the detected vehicle can be carried out, for example, by using a bounding box, whereby the rear or the front can be identified in planar coordinates (x, y). As a result, individual points in the sensor information, which are located on the side of the detected vehicle, for example, can be recognized in order to only consider points at the rear or the front of the detected vehicle.

In an advantageous embodiment of the invention, the step of determining the pitch angle based on a vertical angular position of a detected vehicle in the sensor information includes determining a horizontal angle of the detected vehicle in relation to a sensor direction and discarding the sensor information at a horizontal angle greater than a critical angle. The sensor direction relates to a direction in which the sensor delivers the sensor information, for example as the middle direction. For example, in the case of an environmental sensor attached to the vehicle and oriented in the direction of travel, the sensor direction corresponds to the direction of travel. The detected vehicle is therefore with a small horizontal angular deviation in front of or behind the own vehicle. In this way, for example, vehicles on a parallel lane can certainly be taken into account, while crossing vehicles are excluded, even if they cross at an angle other than 90 °. This observation ensures that the rear or the front of the detected vehicle have sufficient points in the sensor information to be able to reliably detect the vertical angular position of the detected vehicle.

In an advantageous embodiment of the invention, the step of determining the pitch angle based on a vertical angular position of a detected vehicle comprises determining a plurality of vertical angular positions of a detected vehicle, the pitch angle being determined as the mean value of the plurality of vertical angular positions. By averaging, the pitch angle can be determined with improved accuracy. In this way, in particular, pitch angles can be determined that result from the installation on the vehicle or from longer-lasting properties of the vehicle. This applies, for example, to the case that a vehicle has a pitch angle due to a partially incorrect tire air pressure, or due to a structural change. With a sufficiently fast sequence in determining the vertical angular alignment, however, an essentially dynamic pitch angle can also be recorded as the mean value.

In an advantageous embodiment of the invention, the step of determining the pitch angle based on a vertical angular position of a detected vehicle comprises determining a plurality of vertical angular positions of a vehicle traveling ahead or following at different distances from the environment sensor. At different distances from the detected vehicle, different vertical angular positions result with the same pitch angle. The pitch angle can then be determined based on the different vertical angular positions.

As a result, measurement errors can be reduced which, for example, only arise at certain distances.

With regard to further technical features and advantages of the method, reference is made to the description of the driving assistance system, the computer program product, the figures and the description of the figures, and vice versa.

The present invention also relates to a computer program, comprising commands that are used when the computer program is executed cause a computer to carry out a method described here.

The computer program or computer program product described here is therefore used to output control commands for a driving support system of the vehicle which is set up to at least partially, for example completely, execute a method described above. The computer program product can have programs or program parts that are loaded into the processor, e.g. of the driving assistance system, and are set up to carry out the above-mentioned method. On the basis of these programs or program parts, it can be made possible that the method described above can be carried out in order to output control commands for a vehicle.

By means of the computer program described here, it can thus be made possible in an effective and reliable and therefore particularly safe manner that a pitch angle of the environmental sensor is determined.

With regard to further technical features and advantages of the computer program, reference is made to the description of the method, the driving support system, the figures and the description of the figures, and vice versa.

A driving support system for a vehicle for supporting a driver of the vehicle is also described, the driving support system having at least one environment sensor and also a control unit which can be fed with data from the at least one environment sensor, the control unit being configured to at least partially execute a method as described in detail above.

The driving assistance system can be part of a vehicle, such as a car or a truck. For an environment monitoring or environment detection, the driving support system has at least at least one environment sensor, also referred to as an environment detection sensor. The environment sensor can preferably be part of the driving support system that also executes the method and can optionally be provided together with further environment sensors.

In an advantageous embodiment of the invention, the environment sensor is designed as a lidar, in particular as a laser scanner, as a radar or as a camera. The camera can be designed as a stereo camera, for example.

In an advantageous embodiment of the invention, the environment sensor is attached to the vehicle in the direction of travel or against the direction of travel. Accordingly, vehicles driving ahead or behind can be detected with the environment sensor in order to determine the pitch angle. The principles of the implementation of the method can easily be applied to vehicles traveling in front as well as to vehicles behind. In this case, the method is then preferably carried out based on the rear of vehicles traveling ahead using an environment sensor mounted in the direction of travel. In the case of an environment sensor mounted against the direction of travel, the method is carried out based on the front of the following vehicles. In principle, however, vehicles that are parked on the edge of the road, for example, or are on an opposing lane, can also be taken into account when determining the pitch angle.

A control unit such as a processor is also provided. The control unit is suitable for evaluating the sensor data supplied by the sensor or sensors and for acquiring, storing and processing the sensor data. In particular, the control unit is designed to carry out the method based on sensor data acquired by the environment sensor, as is described in detail elsewhere

This can be possible, for example, in that a computer program product for outputting control commands for the driving support system of the vehicle is set up to at least partially, for example completely, execute a method described above. The computer program product can have programs or program parts that are loaded into the processor, e.g. of the driving assistance system, and are set up to carry out the above-mentioned method. On the basis of these programs or program parts, it can be made possible that the method described above can be carried out in order to output control commands for a vehicle.

The driving support system described here can thus enable a pitch angle of the environmental sensor to be determined in an effective and reliable and therefore particularly safe manner.

With regard to further technical features and advantages of the driving assistance system, reference is made to the description of the method, the computer program product, the figures and the description of the figures, and vice versa.

• 1 eine schematische Darstellung eines Fahrzeugs mit einem Fahrunterstützungssystem gemäß einer ersten, bevorzugten Ausführungsform in Seitenansicht,
• 2 eine schematische Darstellung des Fahrzeugs aus 1 zusammen mit einem vorausfahrenden Fahrzeug, wobei das Fahrzeug einen Laserscanner aufweist, der mit einem ersten Nickwinkel ausgerichtet ist, in Seitenansicht, zusammen mit einer Rückansicht des vorausfahrenden Fahrzeugs,
• 3 eine schematische Darstellung des Fahrzeugs aus 2 mit dem Laserscanner, der mit einem zweiten Nickwinkel ausgerichtet ist, in Seitenansicht, zusammen mit einer Rückansicht des vorausfahrenden Fahrzeugs,
• 4 ein schematisches Diagramm zeigend die Erfassung eines Objekts mit verschiedenen Detektionsebenen eines Laserscanners;
• 5 ein schematisches Diagramm zeigend eine Verteilung von als last hit distance bezeichneten Distanzen bei einem fehlerhaften Nickwinkel; und
• 6 ein schematisches Diagramm zeigend eine Verteilung von als last hit distance bezeichneten Distanzen bei einem korrekten Nickwinkel.

The invention is described below with reference to the accompanying drawing preferred embodiments explained in more detail. The features shown can represent an aspect of the invention both individually and in combination. Features of various exemplary embodiments can be transferred from one exemplary embodiment to another.
It shows

• 1 a schematic representation of a vehicle with a driving support system according to a first, preferred embodiment in side view,
• 2 a schematic representation of the vehicle 1 together with a vehicle traveling ahead, the vehicle having a laser scanner which is aligned with a first pitch angle, in side view, together with a rear view of the vehicle traveling ahead,
• 3 a schematic representation of the vehicle 2 with the laser scanner, which is aligned with a second pitch angle, in side view, together with a rear view of the vehicle in front,
• 4th a schematic diagram showing the detection of an object with different detection levels of a laser scanner;
• 5 a schematic diagram showing a distribution of distances referred to as last hit distance at an incorrect pitch angle; and
• 6th a schematic diagram showing a distribution of distances referred to as last hit distance at a correct pitch angle.

The 1 shows a vehicle 10 according to a first, preferred embodiment. The vehicle 10 includes a driving assistance system 12th , which is described in detail below. The vehicle is accordingly 10 own vehicle. The vehicle 10 is on a level surface 20th positioned.

The driving support system 12th includes a control unit 14th and an environmental sensor 16 . The control unit 14th and the environmental sensor 16 are via a data bus 18th connected with each other.

The environmental sensor 16 is designed in this embodiment as a laser scanner. The laser scanner 16 represents a two-dimensional information field 28 with sensor information in the direction of travel 26th of the vehicle 10 ready. The two-dimensional information field 28 is a field with sensor information, which, like an image, shows the sensor information as individual points 30th provides that in lines 32 and columns 34 are arranged. With the laser scanner 16 are additional depth information of the individual points 30th included, that is, a distance of an object at the point of a reflection from the laser scanner 16 emitted laser beam 36 .

The control unit 14th is carried out, a method for determining a pitch angle α of the environmental sensor 16 on the vehicle 10 perform. The procedure for determining the pitch angle α of the environmental sensor 16 is described below. The procedure is carried out in accordance with the vehicle 10 and the driver assistance system 12th the first embodiment.

The procedure for determining the pitch angle α In this exemplary embodiment, includes the provision of a look-up table. The look-up table shows different pitch angles α the greatest distance 25th entered, in which in a horizontal line 32 of the information field 28 a vehicle in front 22nd is still recorded. For example, it can be assumed that the vehicle in front 22nd is a car.

This is in the 4th explained in more detail. In the 4th is the distance from the vehicle on the X-axis 10 or from the environmental sensor 16 in sensor direction 26th shown, whereas the Y-axis is the height above the ground 20th shows. The lines a , b , c , and d show different detection levels of the environmental sensor 16 . It can be seen that, as a function of a vertical angle of the detection plane, different detection of a schematically represented detected vehicle as a result of the distance 22nd is present. Correspondingly, there is also a different detection of the detected vehicle 22nd at different pitch angles α in front. The greatest distance of the vehicle 22nd to the environmental sensor within which the vehicle 22nd is still detected is also referred to as the last hit distance.

This is preferred for a large number of vehicles 22nd Measured over a longer period of time, such as for a period of 1 minute, and the data are preferably stored, such as in a FIFO memory of the driving support system and preferably in the look-up table.

This becomes a two-dimensional information field 30th with the sensor information in the direction of travel 26th from the laser scanner 16 provided. The direction of travel is accordingly 26th equal to a sensor direction of the laser scanner 16 .

It becomes a vehicle in front 22nd and identified in the sensor information. To do this, it is initially based on the sensor information from the environmental sensor 16 a Area map of the detected vehicle 10 created. In the environment map, all objects are marked in planar coordinates, ie in an x and a y direction. Then the stern appears in the map of the area 24 of the vehicle in front 22nd in this exemplary embodiment as the rear boundary of the vehicle in front 22nd certainly. A bounding box is used, creating the rear end 24 is identified in the planar coordinates. Such points are then derived from the sensor information based on the bounding box 30th of the lines 32 those in the bounding box and thus at the stern are taken into account 24 of the vehicle in front 22nd lie.

Now the pitch angle will be α based on a vertical angular position of the preceding vehicle 22nd determined in the sensor information.

First, a horizontal angle of the vehicle in front is used 22nd related to the sensor direction 26th certainly. Sensor information at a horizontal angle greater than a critical angle is discarded. When the horizontal angle of the vehicle in front 22nd is not greater than the critical angle is a last line in the sensor information 32 determines in which the vehicle in front 22nd is still included, for a given distance 25th the vehicle in front 22nd . The last line 32 in which the vehicle in front 22nd is still included, is used here as a measure of the vertical angular alignment of the vehicle in front 22nd results. In detail, there is a line-by-line view from bottom to top. For each line 32 a ratio of points 30th a respective line 32 leading to the vehicle in front 22nd belong to the maximum number of points 30th one line 32 leading to the vehicle in front 22nd belong, intended, for lines 32 in the area of the bounding box. If the ratio is below a limit, it means that the vehicle in front 22nd in one line 32 is no longer included. The previous line 32 , ie the line immediately below 32 , is accordingly the last line 32 in which the vehicle in front 22nd is still included.

As in the 2 and 3 is shown in detail, the points 30th on the stern 24 of the vehicle in front 22nd detected. A first pitch angle α who is in 2 is shown as the "correct" pitch angle α , d .H. assumed as the reference angle, with an angle of 0 °. The vehicle in front 22nd and the vehicle 10 with the laser scanner 16 are on the ground that forms a substantially horizontal plane.

For the reference angle, ie the "correct" pitch angle α , the 2 hit four by the laser scanner 16 emitted laser beams 36 as acquisition levels in four lines 32 completely on the stern 24 of the vehicle in front 22nd . So it all have four lines 32 the maximum number of points 30th leading to the vehicle in front 22nd belong.

In contrast, with a second pitch angle, which in 3 is shown, and that as a "faulty" pitch angle α with an angular deviation assumed, only two from the laser scanner 16 emitted laser beams 36 as acquisition levels in two lines 32 completely on the stern 24 of the vehicle in front 22nd . So it only covers two lines 32 the maximum number of points 30th leading to the vehicle in front 22nd belong. One line 32 lies on the stern 24 of the vehicle in front 22nd at the top. The vehicle in front 22nd is narrower there, so there are fewer points 30th to the vehicle in front 22nd belong than in the two lines below 32 which is the maximum number of points 30th on the detected vehicle 22nd exhibit. The in 3 topmost laser beam 36 does not hit the vehicle in front 22nd . It represents the first line accordingly 32 in which the vehicle in front 22nd is no longer included. The four in the 2 and 3 emitted laser beams 36 point based on the laser scanner 16 the same radiation angles and hit depending on the different pitch angles α different on and the vehicle in front 22nd .

It is also shown in a histogram for each line 32 the ratio of points 30th of line 30, which corresponds to the vehicle in front 22nd belong to the maximum number of points 30th one line 32 leading to the vehicle in front 22nd belong, depending on the distance 25th registered. The ratio will be for different distances 25th entered into the histogram, for example in a range from 10m to 100m.

The pitch angle can then be based on the histogram in the look-up table α be determined. To do this, the greatest distance is calculated from the histogram 25th determines in which a horizontal line 32 of the information field 28 a vehicle in front 22nd still recorded. The pitch angle α is looked up and taken from the look-up table with this information.

In order to improve the result of the method, it is provided in the method described here that the method is performed before the step of determining the pitch angle α and after the step of providing a two-dimensional information field 28 an examination of the detected vehicle with sensor information 22nd with regard to a misclassification, wherein for determining the pitch angle α the detected vehicle 22nd is only used if there is no misclassification. This pre-processing takes place as follows.

The following approximations or assumptions are made, since in reality there are a large number of measured vehicles 22nd has a different height. According to the first approximation, the vehicles 22nd divided into two groups: probably a car and unlikely a car. The vehicles 22nd that can be detected in a close range are classified as likely cars, the remaining vehicles 22nd as unlikely to be a car. As a further approximation, it is chosen that there are more normal vehicles in a conventional street scene 22nd , so cars are available than other vehicles 22nd .

Based on these approximations, the method can be carried out using the distribution of the distances of determined vehicles described above, described as last hit distance 22nd the pitch angle α determine, in particular estimate, and furthermore vehicles that have been classified as unlikely cars can be filtered out as preprocessing.

This preprocessing of the identified vehicles 22nd before determining the pitch angle α can for example be done using a k-means algorithm, as shown in FIGS 5 and 6th is shown by way of example. Since there are essentially only two clusters, namely the near area and the far area, the k-means algorithm can be implemented very easily. There can be two initial values, namely the value 0, the distance of 0 to the ambient sensor 16 and a value which corresponds to the maximum detection distance of the environmental sensor 16 corresponds to. This maximum detection distance can be around 150 m for a laser scanner. After applying the k-means algorithm, there can be two situations in particular, such as those in FIG 5 and 6th are shown.

In the 5 and 6th the upper table or the Y-axis describes the vector of the determined distances, which are described as the last hit distance, whereas the lower graph shows the distribution of these distances, with the X-axis the distance to the environmental sensor 16 should represent. The selected difference between as likely car and unlikely car is shown by the hatching of the bars.

According to 5 there is a situation in which the environmental sensor 16 with an incorrect pitch angle α is present, as was the case with reference to the 3 is described. According to 5 are the vehicles detected in a long range 22nd , displayed on the X-axis at the values 21 to 24, classified as improbable cars and used to determine the pitch angle α not used. This is a typical result in the presence of an incorrect pitch angle α . As described above, this can be done by using the vehicles classified as likely to be passenger cars 22nd can be verified in the procedure described above.

The 6th furthermore shows a typical result for an environmental sensor 16 with a correct pitch angle α . In particular, there is a distribution of the distances, which are described as the last hit distance, in this case in such a way that most vehicles 22nd are arranged in a remote area, such as in particular at the distances marked on the x-axis with the distance 8 to 24. Therefore all vehicles can 22nd for determining the pitch angle α can be used as described above.

List of reference symbols

10

vehicle

12th

Driving assistance system

14th

Control unit

16

Environmental sensor

18th

Data bus

20th

Underground

22nd

detected vehicle, vehicle in front

24

Rear

25th

distance

26th

Direction of travel, sensor direction

28

Information field

30th

Point

32

row

34

column

36

laser beam

α

Pitch angle

a

Acquisition level

b

Acquisition level

c

Acquisition level

d

Acquisition level

QUOTES INCLUDED IN THE DESCRIPTION

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

Patent literature cited

• DE 102017106484 A1 [0006]
• US 9052721 B1 [0007]

Claims (10)

Method for determining a pitch angle (α) of an environmental sensor (16) in a mounted state, in particular in a state mounted on a vehicle (10), with the step Providing a two-dimensional information field (28) with sensor information, wherein the method comprises a step of determining the pitch angle (α) based on a vertical angular position of a detected vehicle (22) classified with regard to a vehicle class in the sensor information, and wherein the method comprises an examination of the detected vehicle (22) with regard to a misclassification before the step of determining the pitch angle (α), the detected vehicle (22) only being used to determine the pitch angle (α) if there is no misclassification. Procedure according to Claim 1 characterized in that the detected vehicle (22) is examined with regard to a misclassification in such a way that the detected vehicle (22) is classified as a classified vehicle (22) or an unclassified vehicle (22). Procedure according to Claim 1 or 2 , characterized in that when the detected vehicle (22) is examined with regard to a misclassification, two groups are formed which are defined by the distance between the vehicles (22) and the environmental sensor (16). Method according to one of the Claims 1 to 3 , characterized in that the examination of the detected vehicle (22) with regard to a misclassification comprises the application of a k-means algorithm. Procedure according to Claim 4 characterized in that the k-means algorithm is based on a detectability of the vehicle (22) that is distance-based with regard to the vehicle (22) and angle-dependent with regard to a vertical angle. Method according to one of the Claims 1 to 5 characterized in that the step of determining the pitch angle (α) based on a vertical angular position of a detected vehicle (22) comprises determining a plurality of vertical angular positions of a detected vehicle (22) at different distances (25) from the environmental sensor (16) . Method according to one of the preceding claims, characterized in that the step of determining the pitch angle (α) based on a vertical angular position of a detected vehicle (22) in the sensor information is determining a last line (32) in the sensor information in which the detected Vehicle (22) is still included for a given distance (25). Computer program, comprising instructions which, when the computer program is executed by a computer, cause the latter, a method here according to one of the Claims 1 to 7th to execute. Driving support system for a vehicle (10) for supporting a driver of the vehicle (10), wherein the driving support system (12) has at least one environment sensor (16) and furthermore a control unit (14) which can be fed with data from the at least one environment sensor (16) , wherein the control unit (14) is configured to implement a method according to one of the Claims 1 to 7th at least in part. Driving support system (12) according to Claim 9 , characterized in that the environment sensor (16) is designed as a lidar, in particular as a laser scanner, as a radar, or as a camera.

Images