DE102013021401A1 - Method for determining an installation orientation of at least one radar sensor arranged on a vehicle - Google Patents

Abstract

The invention relates to a method for determining an installation orientation of at least one radar sensor (2, 3) arranged on a vehicle (1), wherein a movement of the radar sensor (2, 3) and a yaw rate) of the vehicle are determined. According to the invention, only those measured values of the movement of the radar sensor (2, 3) and the yaw rate) of the vehicle (1) for determining the installation orientation of the radar sensor (2, 3) are automatically taken into account, which are detected during a straight-ahead driving or cornering without swimming, wherein In this case at least one parameter is determined directly from the measured values. The invention further relates to the use of a compensation calculation formed compensation straight line for determining a scaling error of the gyroscope and a use of the detected measurements for determining a speed offset when using multiple radar sensors (2, 3) based on the measured values and distances (d) between the radar sensors (2 , 3).

Description

The invention relates to a method for determining a mounting orientation of at least one radar sensor arranged on a vehicle, wherein a movement of the radar sensor and a yaw rate of the vehicle are determined.

From the DE 10 2010 015 723 A1 a method for detecting a movement of a road vehicle is known in which measured values for a relative movement between a vehicle-mounted receiving unit and objects in a vehicle environment in an angular range are detected as an angle. A velocity vector of the road vehicle relative to the objects of the vehicle environment is calculated by determining an angle function which is determined by means of a compensation calculation to the angle-dependent measured values relative to the relative movement. In this case, a yaw behavior of the road vehicle is taken into account.

The invention is based on the object to provide a comparison with the prior art improved method for determining a mounting orientation of at least one arranged on a vehicle radar sensor.

The object is achieved by a method having the features specified in claim 1.

Advantageous embodiments of the invention are the subject of the dependent claims.

In a method for determining an installation orientation of at least one radar sensor arranged on a vehicle, a movement of the radar sensor and a yaw rate of the vehicle are determined.

According to the invention, only such measured values of the movement of the radar sensor and of the yaw rate of the vehicle are taken into account for determining the installation orientation of the radar sensor, which are detected during a straight-ahead drive or cornering without swimming.

Swimming is understood to mean an unstable driving behavior of the vehicle. A float angle is the angle between a movement of the vehicle in the center of gravity and its vehicle longitudinal axis. At high lateral accelerations, the slip angle is used as a measure of controllability of vehicles. In general, it is assumed that with a slip angle of less than 5 °, the handling is still considered stable.

Thus, it is possible in a particularly advantageous manner, the installation orientation of an arbitrarily arranged radar sensor with high precision and online, d. H. during operation of the radar sensor and the vehicle. It also orientation changes during the lifetime of the radar sensor can be determined. Due to the use of a particularly large number of measurements and a direct use of the raw data, a defined scenario and reference system can be dispensed with and at the same time a great robustness of the method is achieved. In this case, the installation orientation is measured directly in contrast to methods known from the prior art and is not determined on the basis of a deviation from a predetermined coarse installation orientation. The high-precision installation orientation allows a very accurate calibration and adjustment and thus a high-precision so-called auto-alignment. Furthermore, the high-precision determination with an error of in particular less than 0.1 ° can also be used, for example, for calculating an intrinsic motion, in radar networks, in adaptive beamforming and in the detection of objects at great distances in order to obtain highly precise results.

The method is also suitable for determining the installation orientation during installation of the radar sensor. In this case, the required relative movement, for example, on a manufacturing system, for example on rails or a conveyor belt, which can be additionally integrated as a reference.

It is furthermore particularly advantageous that there is no direct dependence on an angular resolution of the sensor, because no exactly positioned reference objects have to be measured to determine the installation position. Furthermore, the installation orientation can be determined without knowledge of a gross installation orientation and is applicable for any orientation angle. Therefore, by means of the method as well an independent routine in radar networks, where the same radar is mounted in several places, feasible.

Embodiments of the invention are explained in more detail below with reference to drawings.

Showing:

1 schematically a vehicle with a vehicle-mounted radar sensor when cornering,

2 1 is a schematic diagram of a method for calibrating a radar sensor.

3 FIG. 2 schematically shows results of a determination of an installation orientation of two radar sensors arranged on a vehicle, FIG.

4 schematically a movement angle of a radar sensor in dependence on a factor of a yaw rate of the vehicle of a first scene,

5 schematically a profile of a yaw angle of the vehicle over several measurements of the first scene,

6 schematically a movement angle of a radar sensor in dependence on a factor of a yaw rate of the vehicle of a second scene,

7 2 schematically shows a profile of a yaw angle of the vehicle over a plurality of measurements of the second scene,

8th 3 schematically shows a movement angle of a radar sensor as a function of a factor of a yaw rate of the vehicle of a third scene;

9 2 schematically shows a course of a yaw angle of the vehicle over a plurality of measurements of the third scene,

10 schematically a vehicle with two arranged on the vehicle radar sensors in a straight ahead, and

11 schematically the vehicle according to 10 when cornering.

Corresponding parts are provided in all figures with the same reference numerals.

In 1 is a vehicle 1 with a radar sensor arranged on the vehicle 2 when cornering.

Here is the radar sensor 2 in a position on the vehicle 1 arranged, which starting from a reference point RP of the vehicle 1 is characterized by a length l and a width b, and has an installation orientation characterized by an installation angle β, which is determined by means of the method according to the invention and its embodiments.

During the movement, the vehicle stands out 1 by the vehicle parameters "vehicle speed v _ego " and "yaw rate φ. " out. The parameters "measuring speed v _measuring " and "moving angle γ _measuring " are measured. The movement angle γ _measurement represents the angle between the measurement speed v _{measurement of} the radar sensor 2 and the installation orientation β thereof. The movement of the radar sensor 1 is determined in particular according to P823938 (please add official file number!).

wobei Gleichung (2) für jeden Zeitschritt gilt.The yaw rate φ. results according to

where equation (2) holds for each time step.

mit φ ._Gyro als vom Gyroskop ermittelte Gierrate.The reference system is an in-vehicle 1 existing, but not shown gyroscope used, which by an offset calibration in the state of the vehicle 1 has no zero error. This results in only one scaling error k according to

with φ. _Gyro as yaw rate determined by the gyroscope.

Resolved according to the movement angle γ _measurement results

ergibt.Starting from equation (4), a linearization of the inverse sine is performed so that

results.

gilt dabei beispielsweise ein Linearisierungsfaktor ε < 0.01%.For

For example, a linearization factor ε <0.01% applies here.

Equation (5) corresponds in particular to a straight line equation "y = mx + t" with the slope m equal to the scaling error k of the gyroscope and a y-section t equal to the mounting angle β. At the same time, in all measurements which fulfill the small-angle criterion, the very small yaw rate φ occurs. no swimming on, leaving a rear axle of the vehicle 1 with the distance in length l the center of rotation of the vehicle 1 in the reference point RP.

bestimmt.That is, for each measurement that satisfies the small angle approximation, a y value is determined according to y = γ _measurement (7) and an x value according to

certainly.

After a certain number of measurements, a linear equation is determined by a linear compensation method, in particular by a so-called RANSAC algorithm or a least-square method, and thus the installation angle β and the scaling error k are determined.

2 shows a flowchart of the method for calibration of the radar sensor 1 ,

At a time t, in a first method step S1, the yaw rate φ is determined by means of the measured values of the gyroscope according to the previous description. determined. In parallel, the measurement of the movement of the radar sensor takes place in a second method step S2 as described 1 , ie the measurement of the measurement speed v _measurement and the movement angle γ _measurement .

In a third method step S3, it is checked whether the small angle approximation according to equation (6) is satisfied.

Subsequently, in a fourth method step, it is checked whether the vehicle speed v _Ego and consequently the measurement speed v _{meas are} greater than a predetermined setpoint speed, since the determination of the direction of movement is not possible at low speeds.

In a fifth method step S5, a quality criterion for the radar measurement is checked, for example, whether a ratio between moving and stationary objects in the vehicle environment is sufficient.

If all the conditions of method steps S3 to S5 are met, the x and y values are determined in accordance with equations (7) and (8) in a sixth method step S6 and stored in a seventh method step S7.

If there is a sufficiently large number of measurements in the memory, the ascertainment of the linear equation in the linear compensation method and thus the installation angle β and the scaling error k takes place in an eighth method step S8.

If one of the conditions of the method steps S3 to S5 is not met, the measurement is rejected and a new measurement is started at the measurement time t.

In addition, instead of considering each measurement individually, it is possible to consider sequences in which an integrated yaw rate of the gyroscope, which represents a change in orientation, does not exceed a predetermined limit. Thereafter, as described above, the x and y values become the integrated values of the gyroscope and the radar sensor 1 educated.

In particular, measurements with a higher yaw rate φ are used for a determination of a slope of the regression line. used. For this, once the fit orientation has been determined, the algorithm again may have a higher yaw rate limit φ. be executed. As a result, the slope can be better determined. However, it has to be considered that if yaw rates are too high φ. swimming is possible. This must be avoided.

If the exact scaling error k of the gyroscope is known, the value of the length l of the installation position can be determined, since this corresponds to the slope of the compensation straight line. That is, either the scaling error k or the length l from the slope of the regression line or the factor k * l, if both are unknown, can be determined.

In 3 are results of a determination of an installation orientation, ie the installation angle β, of two radar sensors arranged on a vehicle 2 . 3 shown.

These were used in a test with an installation angle β of about 40 ° outwards on the front of the vehicle 1 arranged. In six independent measurements, three of them at higher speed on a first terrain and three at lower speed on another terrain, the installation angle β was determined online, without any previous knowledge of the installation orientation being used in the algorithm. The results show that all measurements calculate the same installation orientation. The standard deviation is about 0.03 ° for both cases. In addition, the length l was determined on the basis of the slope of the regression line. The scaling error k of the gyroscope was previously determined.

4 schematically shows the movement angle γ _{measurement of} the radar sensor 1 as a function of a factor RF of the yaw rate φ. of the vehicle 1 a first scene, which was determined by means of the linear compensation calculation.

gebildet.The factor RF of the yaw rate φ. is doing according to

educated.

This results in an estimated installation angle β of 7.055 ° and an estimated scaling error k of 1.0256.

It can be seen that the measuring speed v _{meas has} no influence on the parameters.

In 5 is a curve of the integrated yaw rate φ., that is, a yaw angle φ of the vehicle 1 , shown over several measurements n of the first scene.

6 schematically shows the movement angle γ _{measurement of} the radar sensor 1 as a function of a factor RF of the yaw rate φ. of the vehicle 1 a second scene, which was determined by means of the linear compensation calculation.

This results in an estimated mounting angle β of 6.88 ° and an estimated scaling error k of 1.0203.

In 7 is a curve of the integrated yaw rate φ., that is, a yaw angle φ of the vehicle 1 , shown over several measurements n of the second scene.

8th schematically shows the movement angle γ _{measurement of} the radar sensor 1 as a function of a factor RF of the yaw rate φ. of the vehicle 1 a third scene, which was determined by means of the linear compensation calculation.

This results in an estimated mounting angle β of 7.10 ° and an estimated scaling error k of 1.0249.

In 9 is a curve of the integrated yaw rate φ., that is, a yaw angle φ of the vehicle 1 , shown over several measurements n of the first scene.

The 10 and 11 show a vehicle 1 with two at the vehicle 1 arranged radar sensors 2 . 3 when driving straight ahead and cornering.

After the mounting angle β of the radar sensors 2 . 3 according to the description of the 1 to 9 has been determined in the context of another online calibration, a speed offset of several radar sensors 2 . 3 be determined among each other. This is particularly important when determining the proper motion. In addition, the scaling error k of the gyroscope can also be determined.

In a straight ahead according to 10 is the determined measuring speed v _Mess1 v _{Mess2 of} the radar sensors 2 . 3 theoretically the same size everywhere. When cornering according to 11 the measuring speed v _Mess1 v _{Mess2 of} the two radar sensors arranged at a distance d from one another differ 2 . 3 by a differential _speed v _Diff from each other.

ermitteln.This is the yaw rate φ. of the vehicle 1 , which is the same for a rigid body at positions, according to

determine.

At a speed offset ε results for a straight ahead: ν ' _Mess1 = ν _Mess1 + ε (11) and

This results in an addition of a constant yaw rate error. When calculating the proper motion, the vehicle drifts 1 more and more with increasing time.

In contrast to the online calibration for the installation orientation, ie the installation angle β, all measurements can be used here since no model is used. The result is the following straight line equation: ν _diff = k · d · φ. - ε. (13)

The differential speed v _diff is determined on the basis of the determined measuring speeds v _Mess1 v _{Mess2 of} the radar sensors 2 . 3 determined and the yaw rate φ. by means of the gyroscope, which has a constant but unknown scaling error k. The distance d is directly on the vehicle 1 measured and the speed offset ε is the size sought.

For this purpose, the differential _speed v _diff over (d * φ.) applied and according to the determination of the installation angle β a robust straight line compensation is performed by means of the linear compensation method. The y-intercept corresponds exactly to the differential _velocity v _Diff and the slope of the regression line to the scaling error k of the gyroscope.

If the scaling error k is known, the distance d of the radar sensors can also be 2 . 3 be determined with high precision.

In the same sequences, which were also used for the determination of the installation angle β, the velocity offset of the two radar sensors 2 . 3 certainly. For example, the result is a velocity offset of 0.051 m / s with a very low standard deviation of 0.0028 m / s.

LIST OF REFERENCE NUMBERS

1

vehicle

2

radar sensor

3

radar sensor

b

width

d

distance

k

scaling errors

l

length

n

Measurement

RF

factor

RP

reference point

S1 to S9

step

t

Measuring time

v _Diff

differential speed

v _ego

vehicle speed

v _measurement , v _measurement1 , v _measurement2

measurement speed

β

Installation angle yaw angle

φ.

yaw rate

γ _Mess

moving angle

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been 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.

Cited patent literature

• DE 102010015723 A1 [0002]

Claims (6)

Method for determining an installation orientation of at least one vehicle ( 1 ) arranged radar sensor ( 2 . 3 ), wherein a movement of the radar sensor ( 2 . 3 ) and a yaw rate (φ.) be determined of the vehicle, characterized in that automatically only such measured values of the movement of the radar sensor ( 2 . 3 ) and the yaw rate (φ.) of the vehicle ( 1 ) for determining the installation orientation of the radar sensor ( 2 . 3 ), which are detected during a straight-ahead or cornering without swimming, in which case at least one parameter is determined directly from the measured values. A method according to claim 1, characterized in that a selection of the considered measured values is performed under an approximation for small angles. Method according to claim 1 or 2, characterized in that the yaw rate (φ.) detected by a gyroscope, wherein an offset calibration of the gyroscope during a stance phase of the vehicle ( 1 ) using gyroscopic measurements as reference values. Method according to one of the preceding claims, characterized in that the installation orientation is determined from the measured values in a linear compensation calculation. Use of a compensation line formed in a method according to claim 4 in the compensation calculation for determining a scaling error of the gyroscope. Use of the measured values recorded in a method according to one of the preceding claims for determining a speed offset when using a plurality of radar sensors ( 2 . 3 ) based on the measured values and distances (d) between the radar sensors ( 2 . 3 ).

Images