METHOD AND A DEVICE FOR CLASSIFYING AN OBJECT, IN PARTICULAR IN THE SURROUNDINGS OF A MOTOR VEHICLE

Abstract

A method is provided for classifying an object, in particular in the surroundings of a motor vehicle, using an ultrasonic sensor system, the ultrasonic sensor system including a plurality of spatially distributed ultrasonic sensors. A plurality of measurements are carried out continuously during a measurement, an ultrasonic signal being emitted by one of the ultrasonic sensors, a signal being received by at least one of the ultrasonic sensors which includes a plurality of reflected echo signals, so-called multiple echoes, and the received echo signals being associated with an object. A plurality of features may be determined from the received echo signals. The object is classified as a function of a combination of at least two of these features, in particular as a pedestrian.

Claims

1-12. (canceled)

13. A method for classifying an object in surroundings of a motor vehicle, using an ultrasonic sensor system, the ultrasonic sensor system including a plurality of spatially distributed ultrasonic sensors, the method comprising the following steps: carrying out a plurality of measurements, during each of the measurements, an ultrasonic signal being emitted by one of the ultrasonic sensors, a signal being received by at least one of the ultrasonic sensors, the signal including a plurality of reflected echo signals, the received echo signals being associated with an object; determining a plurality of features being determined from the received echo signals, including at least two of the following eight features: a first feature, which represents a frequency with respect to a number of the measurements carried out with which a respective distance exceeds a predefined distance threshold value, each respective distance corresponding to a distance between a temporally first received echo signal of a measurement of the measurements and a temporally last received echo signal of the measurement; a second feature, which represents a variance of the distances; a third feature which represents a distribution of a number of the received echo signals per measurement via the ultrasonic sensors; a fourth feature which represents an alignment of the ultrasonic sensors with respect to the object; a fifth feature, which represents a variance of first object distances over multiple measurements, an associated first object distance being calculated for each determined temporally first received echo signal in an approach of the object being taken into consideration in the determination of the variance; a sixth feature, which represents a correlation of the received echo signal with the sent ultrasonic signal; a seventh feature, which represents an amplitude of an echo signal; an eighth feature, which represents a distribution of reflection points, each reflection point indicating a measured object position; and classifying the object as a function of a combination of at least two of the eight features, as a pedestrian.

14. The method as recited in claim 13, wherein the object is classified as a pedestrian when at least the frequency according to the first feature exceeds a frequency threshold value and the distribution according to the third feature has the result that only one of the ultrasonic sensors or only a few of the ultrasonic sensors which are adjacent to one another, have received a plurality of echo signals per measurement.

15. The method as recited in claim 14, wherein the object is classified as a pedestrian when, in addition, the variance according to the second feature exceeds a first variance threshold value.

16. The method as recited in claim 14, wherein according to the fourth feature, only those ultrasonic sensors are taken into consideration in the determination of the distance which are aligned on the object.

17. The method as recited in claim 14, wherein according to the fourth feature, only that ultrasonic sensor being taken into consideration which has a main measuring direction which is aligned best on the object.

18. The method as recited in claim 14, wherein the object is classified as a pedestrian when, in addition, the variance according to the fifth feature exceeds a second variance threshold value.

19. The method as recited in claim 14, wherein the emitted ultrasonic signal has a certain frequency profile and a frequency profile is determined for at least one received echo signal and according to the sixth feature, a correlation value of the echo signal with the sent ultrasonic signal is calculated.

20. The method as recited in claim 19, wherein a classification of the object as a pedestrian is excluded when the correlation value is greater than a certain correlation threshold value for a certain number of measurements.

21. The method as recited in claim 14, wherein according to the seventh feature, an amplitude is determined for at least one received echo signal, and a classification of the object as a pedestrian is excluded when the amplitude of the at least one received echo signal is greater than a certain amplitude threshold value.

22. The method as recited in claim 14, wherein the object is classified as a pedestrian when, in addition, the spatial distribution of the reflection points according to the eighth feature has a characteristic shape, the characteristic shape including an accumulation of reflection points at a probable object position and a scattering in a lateral direction.

23. The method as recited in claim 13, wherein an optimized combination of the features and/or the threshold value for classifying the object as a pedestrian, is determined beforehand using a machine learning method.

24. A device configured to classify an object in surroundings of a motor vehicle, the device comprising: an ultrasonic sensor system, the ultrasonic sensor system including a plurality of spatially distributed ultrasonic sensors situated at a body of a motor vehicle; and an evaluation unit configured to: carry out a plurality of measurements, during each of the measurements, an ultrasonic signal being emitted by one of the ultrasonic sensors, a signal being received by at least one of the ultrasonic sensors, which includes a plurality of reflected echo signals, the received echo signals being associated with an object; determine a plurality of features being determined from the received echo signals, including at least two of the following eight features: a first feature, which represents a frequency with respect to a number of the measurements carried out with which a respective distance exceeds a predefined distance threshold value, each respective distance corresponding to a distance between a temporally first received echo signal of a measurement of the measurements and a temporally last received echo signal of the measurement; a second feature, which represents a variance of the distance; a third feature which represents a distribution of a number of the received echo signals per measurement via the ultrasonic sensors; a fourth feature which represents an alignment of the ultrasonic sensors with respect to the object; a fifth feature, which represents a variance of first object distances over multiple measurements, an associated first object distance being calculated for each determined temporally first received echo signal in an approach of the object being taken into consideration in the determination of the variance; a sixth feature, which represents a correlation of the received echo signal with the sent ultrasonic signal; a seventh feature, which represents an amplitude of an echo signal; an eighth feature, which represents a distribution of reflection points, each reflection point indicating a measured object position; and classify the object as a function of a combination of at least two of the eight features, as a pedestrian.

25. A motor vehicle, comprising: a device, including: an ultrasonic sensor system, the ultrasonic sensor system including a plurality of spatially distributed ultrasonic sensors situated at a body of a motor vehicle; and an evaluation unit configured to: carry out a plurality of measurements, during each of the measurements, an ultrasonic signal being emitted by one of the ultrasonic sensors, a signal being received by at least one of the ultrasonic sensors, which includes a plurality of reflected echo signals, the received echo signals being associated with an object; determine a plurality of features being determined from the received echo signals, including at least two of the following eight features: a first feature, which represents a frequency with respect to a number of the measurements carried out with which a respective distance exceeds a predefined distance threshold value, each respective distance corresponding to a distance between a temporally first received echo signal of a measurement of the measurements and a temporally last received echo signal of the measurement; a second feature, which represents a variance of the distance; a third feature which represents a distribution of a number of the received echo signals per measurement via the ultrasonic sensors; a fourth feature which represents an alignment of the ultrasonic sensors with respect to the object; a fifth feature, which represents a variance of first object distances over multiple measurements, an associated first object distance being calculated for each determined temporally first received echo signal in an approach of the object being taken into consideration in the determination of the variance; a sixth feature, which represents a correlation of the received echo signal with the sent ultrasonic signal; a seventh feature, which represents an amplitude of an echo signal; an eighth feature, which represents a distribution of reflection points, each reflection point indicating a measured object position; and classify the object as a function of a combination of at least two of the eight features, as a pedestrian.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Specific embodiments of the present invention are described in detail with reference to the figures.

[0045] FIG. 1 shows a device according to one possible exemplary embodiment of the present invention upon the detection of a pedestrian.

[0046] FIG. 2 shows distance data detected during multiple temporally successive measurements by way of example according to an example embodiment of the present invention.

[0047] FIG. 3 shows distance data detected during multiple temporally successive measurements by way of example according to the present invention in consideration of multiple echoes, for each measurement a distance d.sub.i being determined in accordance with a distance between a temporally first received echo signal and a temporally last received echo signal of a measurement.

[0048] FIG. 4 schematically shows a motor vehicle designed according to the present invention and a pedestrian, and superimposed, a distribution of reflection points from a plurality of measurements.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

[0049] In the following description of the exemplary embodiments of the present invention, identical elements are identified by identical reference numerals, a repeated description of these elements possibly being omitted. The figures only schematically show the subject matter of the present invention.

[0050] The front part of a motor vehicle 1 is schematically shown in FIG. 1. Motor vehicle 1 includes a device 10 for classifying an object 70 in the surroundings of motor vehicle 1. Device 10 includes an ultrasonic sensor system, the ultrasonic sensor system including four ultrasonic sensors 12.1, 12.2, 12.3, and 12.4 situated along the front of motor vehicle 1. Device 10 additionally includes an evaluation unit 11, which is designed to evaluate the measurement data of ultrasonic sensors 12.1 through 12.4 and to classify object 70 based thereon. The evaluation unit is designed to activate each of ultrasonic sensors 12.1 through 12.4 so that ultrasonic sensors 12.1 through 12.4 send ultrasonic signals and to receive ultrasonic signals reflected at object 70 and associate the received signals with object 70.

[0051] In the illustrated example, object 70 is a pedestrian 80.

[0052] To classify object 70 as a pedestrian 80, various features of the received echo signals are determined.

[0053] Due to the characteristic structure and shape of a pedestrian 80, the probability is high that at least one of ultrasonic sensors 12.1 through 12.4 will receive a signal which includes a plurality of reflected echo signals. Thus, for example, a hand 82 and a foot 84 of pedestrian 80 reflect the sent ultrasonic signal of ultrasonic sensor 12.3. For example, foot 84 may have a lesser distance 24 to ultrasonic sensor 12.3 than distance 22 of hand 82. Ultrasonic sensor 12.3 will thus receive at least two echo signals in one measurement. Assuming foot 84 has the least distance to ultrasonic sensor 12.3 of all reflecting points of pedestrian 80 and hand 82 has the greatest distance to ultrasonic sensor 12.3 of all reflecting points of pedestrian 80, the echo signal which was reflected from foot 84 is thus received as the temporally first echo signal and the echo signal which was reflected from hand 82 is received as the temporally last echo signal.

[0054] A distance d between the temporally first received echo signal and the temporally last received echo signal may be determined from the echo signals. For this purpose, for example, initially a propagation time difference of these echo signals is determined, from which a spatial distance d may be calculated in a conventional way if the speed of sound of the ultrasonic signals is known. It may be observed over a plurality of measurements how particular determined distance d behaves. For example, if a distance d is determined in a certain minimum portion of the measurements, which exceeds a certain threshold value, this fact may thus be used as an indicator that object 70 is a pedestrian. In addition, great variations of distance d are caused by movement, for example, of the arms and legs, and/or different alignment of a pedestrian 80 in relation to the ultrasonic sensors. If a variance of distance d is thus observed, for example, if this variance exceeds a certain threshold value, this may thus be used as a further indicator that object 70 is a pedestrian.

[0055] In FIG. 3, by way of example in a diagram 200 for a plurality of measurements i, temporally first received echo signal 210 and temporally last received echo signal 220 and distance d.sub.i determined therefrom are shown in each case. Measurement time t is plotted on the x axis and measured distance s to the ultrasonic sensor is plotted on the y axis.

[0056] It may be compared for each measurement whether distance d exceeds a predefined distance threshold value and a frequency with respect to the number of measurements carried out, in which distance d exceeds the predefined distance threshold value, may thus be determined.

[0057] As a further feature, the number of the received echo signals per measurement over the ultrasonic sensors is determined. In the situation shown in FIG. 1, the two middle ultrasonic sensors 12.2 and 12.3 will receive stronger echo signals due to their proximity and alignment with respect to pedestrian 80 than the two outer ultrasonic sensors 12.1 and 12.4. It is accordingly more probable that middle ultrasonic sensors 12.2 and 12.3 will receive multiple echoes. The distribution of the number of the received echo signals per measurement will thus have an accumulation in ultrasonic sensors 12.2 and 12.3 arranged adjacent to one another.

[0058] Also because of the complex structure of a pedestrian 80, in a sequence of measurements, a distance determined from a particular temporally first received echo signal between pedestrian 80 and the ultrasonic sensor may also vary, in particular in comparison to a geometrically simple, stable reflector such as a curbstone. This is shown by way of example in FIG. 2 in a measurement diagram 100. Measurement time t is plotted on the x axis and measured distance s to the ultrasonic sensor is plotted on the y axis. A distance 90 between reflecting object 70 and the measuring ultrasonic sensor is determined at each measurement time with the aid of the temporally first received echo signal. Overall, a relative, linear approach between the measuring ultrasonic sensor and reflecting object 70 is visible, as shown by straight line 95. With known vehicle velocity and the assumption of a stationary object, straight line 95 may be calculated. If the movement state of object 70 is not known, straight line 95 may be determined from measured values 90 (for example by typical fit methods). To determine the fifth feature, a variance of measured values 90 representing the particular object distance is determined over multiple measurements with respect to straight line 95, by which an approach of object 70 to vehicle 1 or the measuring ultrasonic sensor is taken into consideration. If the variance thus determined according to the fifth feature exceeds a second variance threshold, this may be assessed as a further indication that reflecting object 70 is a pedestrian 80.

[0059] FIG. 4 schematically shows in a possible example how the eighth feature according to the present invention may be determined, which represents a distribution of reflection points 60. FIG. 4 schematically shows the front of a vehicle 1, which includes an ultrasonic sensor system 12 including four ultrasonic sensors 12.1, 12.2, 12.3, and 12.4. An object 70 is located ahead of vehicle 1. By way of a plurality of measurements with the aid of ultrasonic sensors, coordinates of reflection points 60 are determined, for example, with the aid of trilateration, each of which indicates a measured position of object 70. A spatial distribution of the reflection points is formed, as indicated by the coordinate system in FIG. 4. It has been shown that in the case of a pedestrian 80, a characteristic distribution of reflection points 60 results, which includes, on the one hand, an accumulation at a probable object position, and a certain scattering of reflection points 60 in the lateral direction, thus in a direction perpendicular to the main measurement direction.