DE102017006780A1 - Method for radar-based determination of a height of an object - Google Patents

Abstract

The invention relates to a method for radar-based determination of a height (ht) of an object (O) in a vehicle environment, wherein the vehicle surroundings are detected by means of at least one radar sensor (2) arranged on a vehicle (1). According to the invention, multipath propagation of radar beams of the radar sensor (2) is considered in the determination of the height (ht) of the object (O), wherein a high-resolution spectral analysis is performed and a high-resolution method for determining a path difference (Δr) in multipath propagation using a power density spectrum is used becomes.

Description

The invention relates to a method for radar-based determination of a height of an object according to the features of the preamble of claim 1.

From the prior art, as in the DE 10 2015 009 382 A1 describes a method for radar-based determination of a height of an object in a vehicle environment, wherein the vehicle surroundings are detected by means of at least one radar sensor arranged on a vehicle. The height of the object is determined by means of an evaluation of a shift in a Doppler frequency between a radar sensor emitted by the radar sensor and a radar signal reflected by the object. The altitude is evaluated from a temporal Doppler frequency response determined in consideration of a vehicle approach speed to the object based on a comparison with stored pattern Doppler frequency characteristics or a ratio of a radial velocity component of the approach speed of the vehicle to the object and a longitudinal velocity component of the approach speed taking into account a distance of the vehicle to the object determined.

The invention is based on the object to provide a comparison with the prior art improved method for radar-based determination of a height of an object.

The object is achieved by a method for radar-based determination of a height of an object having the features of claim 1.

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

In a method for radar-based determination of a height of an object in a vehicle environment, the vehicle surroundings are detected by means of at least one radar sensor arranged on a vehicle. According to the invention the height of the object is determined by means of an evaluation of a path difference of radar beams, d. H. by means of a path difference, the radar beams or radar waves in a multipath propagation scenario. Thus, when determining the height of the object, multipath propagation of the radar beams of the radar sensor is taken into account, performing a high-resolution spectral analysis and using a high-resolution method for determining a path difference in multipath propagation based on a power density spectrum.

The method according to the invention makes it possible to determine the height of the object in a particularly reliable and simple manner by means of the at least one radar sensor. As a result, an increase in reliability of driver assistance systems for autonomous or semi-autonomous longitudinal and / or lateral control of a vehicle is achieved, since elevated and non-elevated objects can be reliably distinguished from one another from one lane. Thus, erroneous controls of the vehicle, for example, to avoid collisions with non-raised objects or objects with low height, such as curbs avoided. Furthermore, a robustness of radar-based driver assistance systems, in particular an increase in the robustness of a localization of the vehicle based on the radar data, is increased since the method no longer provides only two-dimensional results, but three-dimensional results, in particular point clouds.

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

Showing:

1 schematically a perspective view of a section of a first vehicle environment,

2 schematically a perspective view of a section of a second vehicle environment,

3 schematically a perspective view of a section of a third vehicle environment,

4 1 is a schematic perspective view of a detail of a fourth vehicle environment;

5 2 is a schematic perspective view of a section of a fifth vehicle environment;

6 schematically a perspective view of a detail of a sixth vehicle environment,

7 schematically a perspective view of a section of a seventh vehicle environment,

8th schematically a side view of a vehicle and a detection range of a vehicle disposed on the radar sensor,

9 schematically usable for determining a height of an object geometric relationships,

10 schematically a representation of a high-resolution spectral analysis,

11 a schematic representation of a high-resolution spectral analysis,

12 schematically a power density spectrum by means of fast Fourier transformation,

13 schematically a high-resolution power density spectrum, and

14 schematically a flow of a method for radar-based determination of a height of an object.

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

In the 1 to 7 different vehicle environments are shown with different objects O, the object O in 1 is a crossable speed bump, the objects O in 2 are not traversable boundaries, the object O in 3 a curb is, the object O in 4 a parking lock is the objects O in 5 Underrun support beams and piping in a parking garage are the object O in 6 an accessible bridge is and the object O in 7 a cover for a gutter is.

8th shows a side view of a vehicle 1 and a detection area E of one on the vehicle 1 arranged radar sensor 2 ,

The vehicle 1 includes in a possible embodiment in a manner not shown a driver assistance system to assist a driver of the vehicle 1 when parking the same. Such a driver assistance system is designed, for example, as a self-learning system for the highly automated start-up and shutdown of frequently used, non-measurable parking spaces, such as a private garage.

For such assistance of the driver, it is necessary that objects O existing in the vehicle environment be recognized and evaluated as an obstacle or not as an obstacle depending on their height h _t .

From the vehicle 1 Over or under-run objects O, such as a curb, a ceiling element, a manhole cover or manhole cover, can without damage to the vehicle 1 , especially for its tires, are run over but are very difficult to identify without additional information about the height _{ht of} the object O. The objects O are by means of the at least one on the vehicle 1 arranged radar sensor 2 recorded and for example in occupancy grids, also referred to as occupancy grids represented. Without additional information about the height _{ht of} the respective object O, this is difficult to identify and is therefore marked as an obstacle in the occupancy grid. During operation of the driver assistance system, it may be due to these objects marked as obstacle, but over or unterfahrbaren objects O to false positive braking or a wrong lateral guidance of the vehicle 1 come to avoid collisions with the objects O or avoid the objects O.

A height evaluation or determination of the height h _t of such objects only by means of a O by the radar sensor 2 supplied amplitude is not sufficient and can often lead to wrong decisions. Also problematic is a height assessment of potential obstacles by means of camera sensors, since objects O can have different characteristics, for example with regard to their shape and color.

To during the autonomous or semi-autonomous operation of the vehicle 1 recognizing a respective stationary object O designed as an obstacle on a travel path and classifying it as passable, accessible or not passable, is determined by means of the radar sensor 2 detected data performs a radar-based determination of the height h _{t of} the respective object O in the vehicle environment by means of an evaluation of a path difference Δr in a multipath propagation scenario of radar beams, as in 8th shown. A high-resolution spectral analysis is performed and a high-resolution method for determining the path difference Δr for multipath propagation using a power density spectrum (PSD) is used.

H for determining the height _t of the object O is the path difference .DELTA.R, ie, a difference between a direct path AB between the radar sensor 2 and the object O and an indirect path ACB between the radar sensor 2 and the object O in combination with a mounting height h _{s of} the radar sensor 2 and a horizontal object distance D of the radar sensor 2 used to object O, as in 9 shown.

This in 8th shown object O is raised from a road surface and has the height h _t in the direction of a y-axis y of a coordinate system and is in the direction of an x-axis x of the coordinate system from the vehicle 1 and thus of its radar sensor 2 spaced. It reflects the radar signals of the radar sensor 2 and generates Doppler frequencies in the received radar signal. The radar sensor 2 radiated radar beams travel along two separate paths AB, ACB, ie along the direct path AB and along the indirect path ACB, before reaching a point target B at the object O. The energy of the radar beams reflected from the object O, more precisely from its point target B, moves on the same respective path AB, ACB to the radar sensor 2 back.

Geometric relationships between the radar sensor 2 and the object O are in 9 shown. Shown is the known installation height h _{s of} the radar sensor 2 , The height h _t of the object O, shown here as a height h of the target point _t B to the object O, which is mirrored on the x-x axis for the mirrored target point B ', a radar sensor by means of the 2 detectable radial object distance R of the radar sensor 2 to the object O along the direct path AB, the horizontal object distance D of the radar sensor 2 to the object O and an angle β of the indirect path ACB to the x-axis x, for example to a roadway.

This results in the following formula:

It follows:

To determine the path difference Δr between the direct path AB between the radar sensor 2 and the object O and the indirect path ACB between the radar sensor 2 and the object O uses the high-resolution spectral analysis. A modulated continuous wave radar is used, also referred to as FMCW (Frequency Modulated Continuous Wave) radar. In this case corresponds to a target distance, ie the radial object distance R of the radar sensor 2 to the object O along the direct path AB, to a main excursion HA at a first frequency in the frequency domain.

In multipath propagation, a secondary noise SA is also obtained at a second frequency near the main peak HA. This secondary shock SA is the result of the propagation and reflection of the radar waves along the indirect path ACB, so that its second frequency is always higher than the first frequency of the main swing HA. Im in 10 shown distance range in the path-r diagram corresponds to a distance between the main deflection HA and the secondary shock SA to the path difference Δr, as in 10 shown.

A distance resolution c of the path difference Δr of the multipath propagation is limited by a radar bandwidth RB: Δr = c / 2RB (3)

The path difference Δr of the multipath propagation is small for small objects O, ie for objects O of small height h _t . With such low objects O, the path difference Δr of the multipath propagation can be smaller than the distance resolution c or radar bandwidth RB of the radar sensor 2 : Δr <RB (4)

If this is the case, the main rash HA and the secondary ram SA coincide and form a common rash, as in 11 shown, so that then the distance between the main deflection HA and secondary shock SA can no longer be determined, whereby the height h _{t of} the object O can no longer be determined. The same problem occurs with objects O at a great distance from the radar sensor 2 on, because even then the path difference .DELTA.r of multipath propagation may be smaller than the distance resolution c or radar bandwidth RB of the radar sensor 2 ,

Therefore, it is expedient to use high-resolution methods for determining the path difference Δr in the case of multipath propagation using the power density spectrum, more particularly high-resolution spectral analysis methods, for example RELAX, MODE, MODE-RELAX, ESPRIT. By using them, the main swing HA and the secondary swing SA can be determined very accurately, so that they can also be used for small objects O and for far from the radar sensor 2 remote objects O can be separated from each other. As a result, the path difference Δr of the multipath propagation and thus the height h _{t of} the object O can be determined. The 12 and 13 show, in each case in a frequency f deflection AS diagram, a comparison of a fast Fourier transformation generated power density spectrum FFTPSD ( 12 ) and a high-resolution power density spectrum HRPSD ( 13 ).

To determine the height _{ht of} the object O, it is expedient to proceed as in a flowchart in FIG 14 shown. In a first step S1, a time signal is detected. In a second step S2, a window is laid over this time signal, also referred to as windowing. In a third step S3, a fast Fourier transformation, in particular a coarse fast Fourier transformation, is performed. Objects O are detected by means of a so-called adaptive CFAR method. The detected objects O are divided into subbands in a fourth step S4 in the frequency domain by means of a bandpass filter. In a fifth step S5, these subbands are transformed by means of inverse fast Fourier transformation into the time domain, ie into a time signal. In a sixth step S6, a high-resolution spectral analysis algorithm is applied to the time signal, for example RELAX, Mode or ESPRIT. In a seventh step S7, two deflections HA, SA are detected by using local maxima in the obtained high-resolution power density spectrum. The distance between the two deflections HA, SA is determined and the multi-path model for determining the height h _t of the object O in an eighth step S8 supplied. The result of this method is the determined height _{ht of} the object O. If the detection of the two excursions HA, SA fails in the seventh step S7, this seventh step S7 is initially followed by an intermediate step ZS, in which a so-called Gaussian Mixture Fit is performed , Two Gaussian curves are applied to the high-resolution power density spectrum. Then, the two excursions HA, SA are extracted corresponding to the mean of the Gaussian curves. The distance between the two deflections HA, SA is determined and the multi-path model for determining the height h _t of the object O in the eighth step S8 supplied, so that is the result of the method, the height h _t determined in the object O as well.

LIST OF REFERENCE NUMBERS

1

vehicle

2

radar sensor

FROM

direct path

ACB

indirect path

AS

rash

B

point target

B

mirrored point target

D

horizontal object distance

e

detection range

f

frequency

FFTPSD

Power density spectrum generated by fast Fourier transform

HRPSD

Power density spectrum generated by a high resolution process

HA

main rash

SA

secondary rash

h _s

installation height

h _t

Height of the object

O

object

R

radial object distance

r

path

x

X axis

y

y-axis

β

angle

.delta..sub.R

path difference

S1

first step

S2

second step

S3

Third step

S4

fourth step

S5

fifth step

S6

sixth step

S7

seventh step

S8

eighth step

ZS

intermediate step

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 102015009382 A1 [0002]

Claims (3)

Method for the radar-based determination of a height ( _ht ) of an object (O) in a vehicle environment, wherein the vehicle surroundings are determined by means of at least one vehicle ( 1 ) arranged radar sensor ( 2 ) is detected, characterized in that in determining the height (h _t ) of the object (O) a multipath propagation of radar beams of the radar sensor ( 2 ), wherein a high-resolution spectral analysis is performed and a high-resolution method for determining a path difference (Δr) in multipath propagation using a power density spectrum is used. A method according to claim 1, characterized in that for determining the height (h _t ) of the object (O) the path difference (Δr) between a direct path (AB) between the radar sensor ( 2 ) and the object (O) and an indirect path (ACB) between the radar sensor ( 2 ) and the object (O) in combination with a mounting height (h _s ) of the radar sensor ( 2 ) and a horizontal object distance (D) of the radar sensor ( 2 ) is used for the object (O). A method according to claim 2, characterized in that in the high-resolution spectral analysis of the path difference (Δr) from a distance between a frequency of a direct path (AB) corresponding main swing (HA) and a frequency of the indirect path (ACB) corresponding secondary noise (SA) is determined.

Images