DE102017203057A1 - Method for object classification with polarimetric radar data and suitable device therefor - Google Patents

Abstract

Described is a method for object classification, which comprises the following steps for providing an elliptically or circularly polarized transmission signal which is sent to the object to be classified, generating a first radar image from the copolar polarized reflection signal and generating a second radar image from the cross-polarized reflection signal and comparing the first radar image with the second radar image.

Description

The present invention relates to a method for object classification with polarimetric radar data and suitable apparatus therefor.

It is well known to use radar with linear polarized signals for object classification. The results to be achieved in this case, for example, in recorded radar images are not unique or ambiguous with respect to different objects.

It is therefore an object of the present invention to provide a method of object classification and apparatus therefor which reduces the disadvantages of the prior art. It is a further object of the present invention to increase the assignment in object classification and, if appropriate, to provide data that can be used for a further diverse application.

The above-mentioned problems are solved procedurally with the features of claim 1 and device technology with the features of claim 12.

It turns out that by providing an elliptically or circularly polarized transmission signal, which is transmitted to the object to be classified, correspondingly different reflection signals are used to generate different radar images, which can then be compared. With this measure it is achieved that prominent object areas can be distinguished and thus an improved object classification can be brought about.

The method according to the application and the apparatus according to the application can thus be used in future radar sensors, which can be used in particular in highly automated and autonomous driving.

For this purpose, polarimetric radar sensors are necessary, which are particularly characterized by the fact that they can generate significantly more target information with them compared to currently used radars with linearly polarized signals. This is due to the fact that one can generate independent radar images for the co- and cross-polarization and a higher target detection probability in circular polarization is present.

In the method according to the invention is the evaluation of polarimetric radar data with respect to the pattern recognition for the classification of different objects, as well as the detection of so-called "ghost targets". These are caused by multipath propagation, side lobes and periodically recurring main lobes (so-called grating praise).

The application according physical principle shows 1 , A circular or elliptically polarized wave is transmitted and depending on the structure of the target one receives a mainly cross-polarized wave or a mainly copolarized wave. With an odd number of reflections at the target, the polarization direction rotates, and with an even number of reflections, the same polarization is returned. For example, if you send with a left circular wave, then the cross polarization is a right circular wave and the copolarization is a left circular wave. A description of this principle and the construction of a polarimetric radar sensor is given in [1].

For this principle to be implemented, one needs a transmitter which transmits at least one left circular or right circular or elliptical polarization. The receiver must be fully polarimetric. This means that circular, elliptical and linear polarizations can be received. This can be realized by receiving left circular and right circular portions of the received signal. All polarizations can then be represented by the ratio of left and right circular components. Another way to receive full polarimetry is to receive vertically and horizontally linearly polarized portions of the received signal. So that all polarizations can be represented, the amount and the phase of the vertically linearly and horizontally linearly polarized portions of the received signal must be evaluated here.

2 shows the result of a circular polarimetric measurement using the example of a car. This gives two independent radar images, a copolar radar image and a cross-polar radar image. In the copolar radar image even reflections are shown at the target, mainly double reflections. The crosspolar radar image shows odd reflections at the target, mainly single reflections. In both cases, local maxima of the amplitudes of the back-reflected signal are represented by circles whose diameter is in proportion to their amplitude.

Fig. 3a and 3b show a polarimetric radar image showing both the co-polar and cross-polar local maxima 2 include. However, in addition, for each local maxima, the absolute value ratio between the co-polarized and cross-polarized signal component (referred to below as "ratio") is now in the form of different symbols shown. This creates a polarimetric pattern on different areas of the car that can be used to classify objects.

• - Anzahl der kopolaren Maxima, hierbei können auch bestimmte Störabstände (Abstand zum Rauschniveau) einbezogen werden
• - Anzahl der kreuzpolaren Maxima, hierbei können auch bestimmte Störabstände (Abstand zum Rauschniveau) einbezogen werden
• - Durchschnittliches Betragsverhältnis zwischen Ko- und Kreuzpolarisation
• - Maximales Betragsverhältnis zwischen Ko- und Kreuzpolarisation
• - Minimales Betragsverhältnis zwischen Ko- und Kreuzpolarisation
• - Phasenverhältnisse zwischen Ko- und Kreuzpolarisation
• - Besonders charakteristische Eigenschaften

For pattern recognition for different areas of the goals, the following properties of the local maxima are evaluated:

• - Number of copolar maxima, in this case, certain disturbance intervals (distance to the noise level) can be included
• - Number of cross-polar maxima, in this case, certain disturbance intervals (distance to the noise level) can be included
• - Average amount ratio between co- and cross-polarization
• - Maximum amount ratio between co- and cross-polarization
• - Minimum amount ratio between co- and cross-polarization
• - Phase relationships between co- and cross polarization
• - Particularly characteristic properties

The latter is e.g. the first reflection on the car in the area of the front license plate. This has a ratio of less -20 dB.

The pattern classification here distinguishes different types of objects such as e.g. Cars, pedestrians, cyclists, trucks and motorcyclists and road construction targets such as e.g. Gullies, barriers, crash barriers bridges and tunnels.

• - Die charakteristische Reflektion an der Front mit dem Ratio kleiner als -20 dB verschiebt sich zur rechten Seite des PKWs
• - Im Bereich der Front treten Reflektionen auf, mit Ratiowerten zwischen 10dB und 15dB, als auch eine Reflektion mit einem Ratio zwischen 15dB und 20 dB. (Dies ist in den stärken kopolaren Eigenschaften begründet. Diese entstehen dadurch, dass es durch das schräg stehende Fahrzeug mehr Doppelreflektionen im Bereich der Front insbesondere im Bereich des Autogrills gibt)
• - Es ergibt sich eine Veränderung des polarimetrischen Musters vor allem im Bereich der Front, des Lenkrads und im hinteren Bereich des Fahrzeuges

4 shows the car measured frontal and at an angle of -20 °. With the angular deviation of -20 °, the following change with respect to the car measured from the front can be observed:

• - The characteristic reflection at the front with the ratio smaller than -20 dB shifts to the right side of the car
• - At the front, reflections occur, with ratiowers between 10dB and 15dB, as well as a reflection with a ratio between 15dB and 20dB. (This is due to the strong copolar properties, which are caused by the fact that there are more double reflections in the area of the front, especially near the autogrills, due to the vehicle at an angle)
• - There is a change in the polarimetric pattern, especially in the area of the front, the steering wheel and the rear of the vehicle

Attachment of these properties can be determined at what angle the vehicle is to the sensor.

• - Frontbereich
• - Vordere Radkästen
• - Lenkradbereich
• - Außenspiegel

5 shows areas of particular importance in pattern recognition and classification of a car measured from the front. These are:

• - Front area
• - Front wheel arches
• - Steering wheel area
• - Exterior mirrors

Characteristic here is accordingly 3 , a strong reflection in the area of the front license plate with a very low ratio (strong single reflection).

• - Frontbereich
• - die zum Sensor ausgerichteten Radkästen
• - der zum Sensor ausgerichtete Außenspiegel
• - der zum Sensor ausgerichtete vordere Türspalt
• - die zum Sensor ausgerichtete hintere Fahrzeugecke

6 shows areas of particular importance in pattern recognition and classification of an obliquely detected car:

• - Front area
• - The wheel arches aligned with the sensor
• - The exterior mirror facing the sensor
• - The front door gap aligned with the sensor
• - The rear vehicle corner aligned to the sensor

Particularly characteristic here is the detection of the contour of the vehicle as an L-shape, the exact position detection of the wheel arches, as well as the relative appearance (compared to the other measurement positions) of many signals with a high ratio (double reflections).

• - der zum Sensor ausgerichtete Querberiech
• - die zum Sensor ausgerichteten Radkästen
• - die zum Sensor ausgerichteten Fahrzeugecken
• - die zum Sensor ausgerichteten Vordertür

7 shows areas of particular importance in pattern recognition and classification of a transiently detected car:

• - The aligned to the sensor Querberiech
• - The wheel arches aligned with the sensor
• - The vehicle corners aligned to the sensor
• - The sensor-oriented front door

Particularly characteristic here is the detection of strong reflections with a very low ratio (strong single reflections) in the area of the front door.

• - Hintere Fahrzeugbereich
• - die zum Sensor ausgerichteten Fahrzeugecken
• - die zum Sensor ausgerichteten Vordertür

8th shows areas of particular importance in pattern recognition and classification of a car detected from the rear:

• - Rear vehicle area
• - The vehicle corners aligned to the sensor
• - The sensor-oriented front door

Particularly characteristic here is the detection of a strong reflection with a very low ratio (strong single reflections) on the outer contour of the stern, as well as a polarimetric pattern, which originates from the interior of the car.

Typical polarimetric patterns with the described properties of object classes or object subclasses are always assigned to different angle and distance ranges and serve as the basis for a classification algorithm.

Furthermore, there are advantages to using circular polarization in ghost target detection caused by multipath propagation or sidelobes or annoying main recurring lobes. The latter two are particularly pronounced when strong targets are detected at high angular rates.

9 shows by way of example the situation with a multipath propagation due to an additional reflection on an object. This creates a mirrored target that is at a different angle. However, the additional reflection also causes a rotation of the polarization properties, so that the analysis of the changed polarization pattern allows identification of multipath propagation.

10 shows the same situation as in 9 but refers exclusively to targets with a relative speed to the radar sensor and the environment. If, at targets within a range and an angle gate, a local maximum has different velocities in the co-polar signal and in the cross-polar signal, then it is a "ghost target". These can be caused by multipath propagation as well as side lobes or recurring main lobes.

Advantageous further developments are subject matters of the dependent claims.

In summary shows 1 a principle of a polarimetric radar, including an odd number of reflections, reflecting a cross-polarized wave, and an even number of reflections, reflecting the copolarized wave. 2 shows a co-and Kreuzpolares radar image of a car detected head-on and represents the local maxima (LM) for the copolar and cross-polar radar image, with circles mark their positions. The diameters of the circles are in proportion to the amplitude of the LMs. 3a shows a polarimetric radar image of a car detected head-on. The table according to Fig. 3b and 4 represents the ratio between the signal magnitude of the co-polar and cross-polar local maximum (LM). The size of the symbols are in proportion to the amplitudes of the LMs. Further shows 4 a polarimetric radar image of a front-facing and a car at an angular deviation of -20 ° located car. 5 shows pattern areas for the classification of a car detected head-on. 6 shows pattern areas for classification of an obliquely detected car. 7 shows pattern areas for classification of a transversely detected car. 8th shows pattern areas for classification of a car detected from the rear. 9 Figure 12 shows detection of multipath propagation using polarimetric information for stationary and moving targets and, inter alia, that the pattern has rotated polarization properties at a mirrored target. 10 Figure 12 shows detection of "ghost targets" using polarimetric information on moving targets, including a target having a relative velocity to the sensor and the environment, and a "ghost target," which is a "ghost target" if both polarizations are in the same location (identical receiving gate) have different speeds.

LIST OF REFERENCE NUMBERS

(1)

front area

(2)

Front wheel arch

(3)

steering wheel area

(4)

Mirrors

(5)

Door gap driver's door

(6)

Rear wheel arch

(7)

Rear corner of the vehicle

(8th)

cross section

(9)

In front of the door

(10)

Front corner of the vehicle

(11)

Rear vehicle area

(12)

Vehicle interior

Claims (17)

Method for object classification, comprising the following steps: a. Providing an elliptically or circularly polarized transmission signal which is sent to the object to be classified. b. Generating a first radar image from the copolar polarized reflection signal and generating a second radar image from the cross-polarized reflection signal c. Comparing the first radar image with the second radar image Method according to Claim 1 wherein the generating of the first and second radar image from left or right circular signal components takes place. Method according to Claim 1 wherein the generating of the first and second radar image from linear horizontal and linear vertical signal components takes place. Method according to one of Claims 1 to 3 wherein the generating of the first and second radar image takes place simultaneously, preferably by means of the transmission signal. Method according to Claim 1 to 4 wherein the comparison is based on the signal properties of the local maxima of the reflection signals. Method according to Claim 5 wherein the comparison is based on the signal properties of individual target areas of the object. Method according to one of Claims 5 or 6 in which the signal properties comprise at least one of the following criteria: number of copolar local maxima number of crosspolar local maxima magnitude ratio between co- and cross polarization, preferably average magnitude ratio thereof maximum limit ratio between co- and cross polarization minimum ratio between co- and cross polarization Cross polarization - phase relationship between co- and cross polarization - location of local maxima with high or low magnitude ratio between co- and cross polarization - velocity evaluation Method according to one of Claims 1 to 7 , where similarity patterns are created for object classification. Method according to one of Claims 1 to 8th , wherein further sub-object classes are provided which are divided at least in a radar sensor emitting the transmission signal, in - distance to the radar sensor - angular alignment to the radar sensor - object orientation to the radar sensor - relative speed to the radar sensor Method according to one of Claims 1 to 9 , wherein the object classification refers to the position and orientation recognition of the object, preferably a car. Method according to one of Claims 1 to 10 In the case of frontal object determination, at least the following areas of a car are evaluated. Device for determining an object classification, in particular using the method according to one of Claims 1 to 11 , characterized in that a. Means for providing an elliptically or circularly polarized transmission signal which is transmitted to the object to be classified, b. Means for generating a first radar image from the copolar polarized reflection signal and generating a second radar image from the cross-polarized reflection signal and c. Means are provided for comparing the first radar image with the second radar image Device after Claim 12 , wherein means for generating the first and second radar image from left or right circular signal components are provided. Device after Claim 12 , wherein means for generating the first and second radar image of linear horizontal and linear vertical signal components are provided. Device according to one of Claims 12 to 14 in which the generation of the first and second radar image takes place simultaneously, preferably by means of the transmission signal. Device according to one of Claims 12 to 15 in which the comparison is based on the signal properties of the local maxima of the reflection signals. Device after Claim 16 in which the comparison is based on the signal properties of individual target areas of the object.

Images