Method for ascertaining misalignment of a radar sensor

Abstract

A method for detecting misalignment of a radar sensor positioned on a vehicle. A Doppler spectrum for the radiation emitted and received by the radar sensor is ascertained. For at least one frequency bin of the Doppler spectrum, an angle of incidence is determined in at least a subinterval. The determined angle of incidence is compared to the angle of incidence expected for the frequency bin. A misalignment of the radar sensor is detected as a function of the difference of the measured angle of incidence from the expected angle of incidence.

Claims

1. A method for detecting misalignment of a radar sensor positioned on a vehicle, the method comprising the following steps: controlling a transmitting antenna of the radar sensor to emit radiation at a first time that a receiving antenna of the radar sensor receives at a second time after the first time; ascertaining a Doppler spectrum for the radiation emitted and received by the radar sensor; determining, for at least one frequency bin of the Doppler spectrum, an angle of incidence in at least a subinterval of a predefined frequency interval; comparing the determined angle of incidence to an angle of incidence expected for the frequency bin; and detecting a misalignment of the radar sensor as a function of a difference of the determined angle of incidence from the expected angle of incidence.

2. The method as recited in claim 1, wherein angles of incidence are determined for a plurality of frequency bins of the Doppler spectrum, the determined angles of incidence are compared to expected angles of incidence of the frequency bins, and a difference between the determined angles of incidence and the expected angles of incidence are ascertained as a measure of the misalignment of the radar sensor.

3. The method as recited in claim 2, wherein the radar sensor has a plurality of transmitting and/or receiving antennae, and the angles of incidence of the frequency bins are determined with the aid of the transmitting and/or receiving antennae.

4. The method as recited in claim 2, wherein both a quality of the determined angles of incidence and absolute values of the determined angles of incidence are taken into account for ascertaining an average difference between the determined angles of incidence and the expected angles of incidence.

5. The method as recited in claim 4, wherein in the case of the average difference, the differences of the ascertained angles of incidence from the expected angles of incidence of the frequency bins, which have a greater angle of incidence relative to the direction of travel, are weighted more heavily than the differences of the ascertained angles of incidence from the expected angles of incidence of the frequency bins, which have a lower angle of incidence.

6. The method as recited in claim 4, wherein in the case of the average difference, the differences of the ascertained angles of incidence from the expected angles of incidence of the frequency bins, which are above a predefined limiting value, are not considered.

7. The method as recited in claim 1, wherein the subinterval has a frequency spectrum, in which Doppler frequencies are contained that are generated by a speed range between 0 and a speed of motion of the vehicle.

8. The method as recited in claim 1, wherein the subinterval is determined as a function of a moving direction of the vehicle, a nominal alignment, and a beam angle of the radar sensor.

9. The method as recited in claim 8, wherein the moving direction of the vehicle is ascertained by at least one sensor on the vehicle, and/or interim deviations from straight-ahead travel are compensated for by averaging over the time.

10. The method as recited in claim 1, wherein the angle of incidence determined for the frequency bin is dominated by power reflected by at least one stationary object, the stationary object having a speed relative to the radar sensor that corresponds to the frequency bin.

11. The method as recited in claim 1, wherein the radar sensor emits radiation having a constant transmitting frequency.

12. A non-transitory storage medium on which is stored a computer program for detecting misalignment of a radar sensor positioned on a vehicle, the computer program, when executed by a computer, causing the computer to perform the following steps: ascertaining a Doppler spectrum for radiation emitted and received by the radar sensor; determining, for at least one frequency bin of the Doppler spectrum, an angle of incidence in at least a subinterval of a predefined frequency interval; comparing the determined angle of incidence to an angle of incidence expected for the frequency bin; and detecting a misalignment of the radar sensor as a function of a difference of the determined angle of incidence from the expected angle of incidence.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows a schematic representation of a vehicle having a sensor system, in the case of straight-ahead travel of the vehicle.

(2) FIG. 2 shows a schematic representation of a vehicle having a sensor system, where a radar sensor is mounted so as to be incorrectly rotated.

(3) FIG. 3 shows a schematic representation of a corner sensor of the sensor system, in the case of straight-ahead travel.

(4) FIG. 4 shows a schematic representation of a vehicle having a sensor system, in the case of straight-ahead travel with objects positioned symmetrically.

(5) FIG. 5 shows a schematic representation of a corner sensor of the sensor system, where a radar sensor is mounted so as to be incorrectly rotated and objects are positioned symmetrically.

(6) FIG. 6 shows a schematic representation of a radar sensor.

(7) FIG. 7 shows a graph, in which the received power of the reflected and received radar signal is plotted versus relative speed v.sub.α/v.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

(8) In the figures, identical structural elements have, in each instance, the same reference numerals.

(9) FIG. 1 shows a schematic representation of a vehicle 1 having a sensor system 2, in the case of straight-ahead travel of vehicle 1. According to the exemplary embodiment, sensor system 2 includes a first radar sensor 4, which is positioned on a vehicle front end. A second radar sensor 6 is configured as a corner sensor and is positioned at a transition between the vehicle front-end and a right side of vehicle 1 in direction of travel F.

(10) In addition, sensor system 2 includes an acceleration sensor 8 on the vehicle. Sensors 4, 6, 8 of sensor system 2 are coupled to a control unit 10 so as to be able to transmit data. Through this, the control unit may read out sensors 4, 6, 8, evaluate the measured values of the sensors, and implement a method for ascertaining misalignment of at least one of radar sensors 4, 6.

(11) For the sake of clarity, only scanning range A of first radar sensor 4 is shown. Two objects SO1, SO2, which are positioned asymmetrically relative to vehicle 1, are situated in scanning range A.

(12) From the point of view of vehicle 1, stationary objects SO1, SO2 move towards vehicle 1 at a speed v.sub.ego. For objects SO1, SO2 having a lateral (but also vertical) offset, relative speed v.sub.r1, v.sub.r2 decreases by the cosine of angle φ.sub.1, φ.sub.2 to direction of travel F. All in all, relative speeds v.sub.r1, v.sub.r2 are therefore distributed on the interval [−v.sub.ego; 0].

(13) For a radar sensor 4 oriented exactly forwards, in the case of straight-ahead travel, the angle to the direction of travel, that is, to the moving direction of vehicle 1, corresponds directly to the observation/receiving angle:
φ=φ.sub.motion=φ.sub.observer=φ.sub.1 and/or φ.sub.2

(14) For the two azimuth angles φ.sub.1, φ.sub.2, which belong to a particular relative speed, that is, Doppler frequency, due to the cosine effect, the following applies in this case:
φ.sub.1,φ.sub.2=+/−arccos(−v.sub.r/v.sub.ego)

(15) In FIG. 2, a schematic representation of a vehicle 1 having a sensor system 2 is shown, in which case a radar sensor 4 is positioned so as to be incorrectly rotated.

(16) A horizontal misalignment of radar sensor 4 results mathematically in an angular deviation Δφ:
φ.sub.observer1,2=φ.sub.motion1,2+Δφ

(17) However, such an angular deviation is produced for a deviation from straight-ahead travel, as well.

(18) FIG. 3 shows a schematic representation of a corner sensor 6 of sensor system 2, in the case of straight-ahead travel of vehicle 1. In addition, scanning range A of corner sensor 6 is shown.

(19) Due to an installation position of corner sensor 6 rotated with respect to direction of travel F, a high power of stationary objects SO1, SO2 may even be received at large angles to direction of travel φ.sub.motion since the corresponding observation angles φ.sub.observer lie in the major lobe of the antenna.

(20) A further object SO1′ is positioned in back of an object SO1, the power reflected back being superposed with the reflected-back power of object SO1. Due to this, the two ascertained powers fall into the same frequency bin in the Doppler spectrum, since they have the same relative speed with respect to sensor 6.

(21) FIG. 4 illustrates a schematic representation of a vehicle 1 having a sensor system 2, in the case of straight-ahead travel with objects positioned symmetrically. Due to this, the two objects SO1, SO2 are positioned at a symmetric angle relative to vehicle 1. The following relation results from this:
|φ.sub.observer1,2|=|φ.sub.motion1,2|=|φ.sub.1,2|

(22) FIG. 5 shows a schematic representation of a front sensor 4 of sensor system 2, where a radar sensor 4 is positioned so as to be incorrectly rotated and objects are positioned symmetrically.

(23) Angles to the direction of travel φ.sub.motion1,2=+/−arccos (−v.sub.r/v.sub.ego) are known beforehand for each analyzed frequency bin and stored in a data storage unit of the control unit. Therefore, the known angles of the frequency bins may be used for detecting misalignment of a radar sensor. To that end, the known angle of a frequency bin is compared to a measured angle of the same frequency bin. An angular deviation Δφ may be determined from the comparison. Angular deviation Δφ may be identical for a plurality of frequency bins or for each analyzed frequency bin.

(24) In this connection, a least squares method would provide a solution for combining the values of the individual frequency bins. For example, the following factors may be taken into account in the weighting of these errors: quality of the estimate of the observation angles in each frequency bin; weighting frequency bins having large angles φ to direction of travel F more heavily, since in the case of large angles φ, a small change in angle Δφ already produces a relatively large change in relative speed v.sub.r1, v.sub.r2, and therefore, in the Doppler frequency (cf. cos′(φ)=−sin(φ)); in the case of large differences, the angle of incidence may originate from a moving object and consequently be ignored by the control unit.

(25) For example, values of the quality of the estimate of the observation angles for each frequency bin may be stored in a data storage unit of the control unit. In addition, an angle value, which indicates as of when an angle φ to direction of travel F is categorized as large and weighted more heavily, may be stored in the data storage unit. Furthermore, an angle value of a large deviation for the angle of incidence may be stored, which indicates that the received signal originates from a moving object and is consequently ignored by the control unit.

(26) In all of these methods, the measured values of the individual frequency bins may also be averaged over time.

(27) If the two angles of incidence φ.sub.observer1,2 of stationary objects SO1, SO2, in a frequency bin, may be determined, previous knowledge of independent speed v.sub.ego is not absolutely necessary. Both independent speed v.sub.ego of vehicle 1 and the misalignment, that is, direction of travel F, may be determined directly from difference Δφ and/or the average value of the two observation angles φ.sub.observer1,2.

(28) If φ.sub.observer1<0<φ.sub.observer2, then:
v.sub.ego=−cos((φ.sub.observer2−α.sub.observer1)/2)/v.sub.r
Δφ(φ.sub.observer2+φ.sub.observer1)/2

(29) The movement of vehicle 1 is normally not with respect to sensors 4 and 6, but with respect to the midpoint of a rear axle. However, the movement of any other point on the rigid body, vehicle 1, may be derived from this. In addition, in the case of straight-ahead travel, the movement vector of all points on vehicle 1 is the same.

(30) Misalignment of the radar sensors may also be detected erroneously, if utilized speed v.sub.ego of the vehicle differs from the actual speed. Deviation Δφ from the ideal angle with regard to moving direction φ is largest in the vicinity of negative independent speed v.sub.ego. Therefore, the frequency bins in the Doppler spectrum contain the most information regarding an error in the estimation of the independent speed.

(31) FIG. 6 shows a schematic representation of a block diagram for a radar sensor. A high-frequency oscillator 18 is provided, whose frequency is controllable. High-frequency oscillator 18 generates a transmitted signal, which reaches an antenna 22 via a mixer 20 and is then emitted by the antenna as a radar lobe. The radar echoes generated by objects in the surrounding area of the vehicle are received by antenna 22 and are mixed, in mixer 20, with a component of the transmitted signal generated by high-frequency oscillator 18 at the receiving time. Thus, an intermediate frequency signal 24, which is evaluated further in an evaluation unit 26, is obtained by beating. The frequency of the transmitted signal generated by high-frequency oscillator 18 is modulated and forms a series of CW R-signal ramps having a constant frequency. Thus, it is a CW radar transmission signal. If an elongated object, which is situated at a distance in front of the object, is located, then the interval of the ramps determines the frequency difference of the signals, which are mixed with each other in mixer 20 and produce the frequency of intermediate frequency signal 24. If the vehicle traveling ahead moves relative to the radar sensor, then the frequency difference is a function of a Doppler shift, which, for its part, is a function of the relative speed.

(32) Intermediate frequency signal 24 is initially sampled and digitized as a time signal and then converted to a Fourier spectrum, using, for example, a rapid Fourier transform. In this spectrum, each located object is characterized in the form of a peak at a particular frequency, which is a function of the distance and of the relative speed of the object. If the same object is now located once on the rising ramp and then once more on the falling ramp a little later, then the frequencies of the two peaks may be added. Since the ramps have opposite slopes, the components that are a function of distance cancel each other out in this instance. Thus, only the Doppler component that is a function of the relative speed remains. Conversely, if the frequencies of the two peaks are now subtracted, the components that are a function of speed cancel each other out, and a pure distance component is obtained, which allows the distance of the object to be determined. Normally, more than two modulation sequences or at least two ramps are used, which differ in their slope. This facilitates the assignment of the peaks contained in the spectra to the respective objects, if two or more objects are present.

(33) Evaluation unit 26 is used for analyzing the spectrum of intermediate frequency signal 24. A discrete spectrum of the intermediate frequency signal is used and analyzed for a digital analysis. Accordingly, the frequency axis is split up into a finite number of frequency bins, and the spectrum is a discrete function, which indicates the power apportioned to each frequency bin.

(34) FIG. 7 shows a graph, in which received power P, in decibels (dB), of the reflected radar signal received is plotted versus relative speed v.sub.α/v. In this context, angle α determines an angular difference between the direction of travel of the vehicle and the connecting direction to the object. The independent speed of the vehicle is denoted by v. A relative speed of the vehicle with respect to the object is yielded from the following formula: v.sub.α=−v.Math.cos(α). The quotient of measured relative speed v.sub.α=−v.Math.cos(α) of the object and independent speed v of the vehicle is indicated on the x-axis. Accordingly, the power reaches a maximum at v.sub.α/v=−1. The signal falls abruptly to zero for relative speeds that are even smaller (greater by magnitude).

(35) When angle α approaches the limiting value of 0, the slope of the cosine function becomes smaller and smaller, with the effect that the relative speeds obtained for different angles α become more and more similar to each other, and that therefore, the radar echoes from an increasing number of scattering centers fall into the same frequency bin. The result of this is that the power of the frequency bins increases markedly as the limiting value of −1 is approached more and more. Therefore, the abrupt decrease in the power at a relative speed v.sub.α/v=−1 for α=0 is so prominent and readily detected in the spectrum.