METHOD FOR DETECTING ANGLE MEASURING ERRORS IN A RADAR SENSOR

Abstract

A method for detecting angle measuring errors in an angular-resolution radar sensor for motor vehicles. For stationary radar targets, in each instance, the radial velocity and at least one locating angle are measured, and with the aid of the measured locating angle, an expected value of the radial velocity is calculated and compared to the measured value. Measurements of the radial velocities and the locating angles for one or more stationary targets are taken. For each of these targets, an individual indicator value is calculated, which indicates the difference of the measured radial velocity from the expected radial velocity. The individual indicator values obtained are subjected to angle-dependent scaling to compensate for the angular dependence of distortive angle errors. An indicator of the angle measuring error is calculated from the scaled, individual indicator values.

Claims

1-6. (canceled)

7. A method for detecting angle measuring errors in an angular-resolution radar sensor for a motor vehicle, the method comprising the following steps: measuring a respective radial velocity and at least one respective locating angle for each of stationary radar targets; calculating, for each of the stationary radar targets, a respective expected value of the radial velocity using the respective measured locating angle, and comparing the respective expected value to the respective measured radial velocity; calculating, for each of the stationary radar targets, an individual indicator value, which indicates a difference of the respective measured radial velocity from the respective expected radial velocity; subjecting the individual indicator values to angle-dependent scaling to compensate for angular dependence of distortive angle errors; and calculating an indicator of the angle measuring error from the scaled individual indicator values.

8. The method as recited in claim 7, wherein the method is used for a FMCW radar, in which a frequency of a radar signal in consecutive measuring intervals is modulated in a ramp-shaped manner, wherein the calculation of the individual indicator values is carried out based on measurement results which are obtained within the same measuring interval.

9. The method as recited in claim 8, where after the angle-dependent scaling, the individual indicator values are combined to form an effective value, effective values obtained in consecutive measuring intervals are subjected to temporal filtering, and the indicator is calculated taking into consideration a result of the filtering.

10. The method as recited in claim 7, where the angle-dependent scaling takes place in a two-dimensional angle space.

11. The method as recited in claim 7, where a misalignment error of the radar sensor is detected and corrected using the respective measured radial velocities and the respective locating angles, and the individual indicator values are calculated based on the angle measurements that are corrected by the misalignment errors.

12. A radar sensor for a motor vehicle, comprising: a transmitting and receiving unit; and a control and evaluation device configured to detect angle measuring errors in an angular-resolution radar sensor of the motor vehicles, the control and evaluation device configured to: measure a respective radial velocity and at least one respective locating angle for each of stationary radar targets; calculate, for each of the stationary radar targets, a respective expected value of the radial velocity using the respective measured locating angle, and compare the respective expected value to the respective measured radial velocity; calculate, for each of the stationary radar targets, an individual indicator value, which indicates a difference of the respective measured radial velocity from the respective expected radial velocity; subject the individual indicator values to angle-dependent scaling to compensate for angular dependence of distortive angle errors; and calculate an indicator of the angle measuring error from the scaled individual indicator values.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0019] FIG. 1 shows a sketch for illustrating distortive angle errors at a radar sensor.

[0020] FIG. 2 shows a sketch for illustrating a misalignment error.

[0021] FIGS. 3 and 4 show sketches for illustrating the dependence of the radial velocity of a radar target on the locating angle.

[0022] FIG. 5 shows a diagram for illustrating the angle relationships for a radar target in a spherical coordinate system.

[0023] FIG. 6 shows a diagram for illustrating the angle relationships for a radar target in a conical coordinate system.

[0024] FIG. 7 shows a flow chart for illustrating important steps of a method according to an example embodiment of the present invention.

[0025] FIG. 8 shows a graph for illustrating a scaling and limiting function.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

[0026] A horizontal section of a radar sensor 10 having a housing 12 is shown schematically in FIG. 1; the housing containing a MIMO antenna array 14 and being delimited on a transmitting and receiving side by a radome 16. A control and evaluation device 18 is connected to antenna array 14, the control and evaluation device being used to control the functions of the radar sensor and to determine, with the aid of the received radar echoes, the distances r, relative velocities V_r (radial velocities), azimuth angles φ, and elevation angles α (referred to in summary as the locating angles) of radar targets 20 situated in the locating range. The radar beams 22 reflected by four radar targets 20 and received again by antenna array 14 are drawn in schematically.

[0027] As an example, it is assumed that a coating 24, e.g., an icy crust, at whose upper surface radar beams 22 are refracted, is situated on radome 16, which means that during the angle measurement (in this case, azimuth), a distortive angle measuring error Δφ occurs. It is apparent that radar beams 22 are refracted by film 24 to different degrees and in different directions, which means that the magnitude and algebraic sign of distortive angle measuring errors Δφ are a function of the position of specific radar target 20 relative to radar sensor 10.

[0028] Radar sensor 10 is installed in the front end of a motor vehicle and is used, in particular, for tracking vehicles traveling ahead, as well as other obstacles in the near field of the vehicle. In this context, in the normal case, the radar sensor is aligned in such a manner, that its optical axis coincides with the x-axis, which indicates the forward direction or direction of travel of the motor vehicle.

[0029] By way of comparison, FIG. 2 shows a situation, in which no distortive angle error occurs, but radar sensor 10 is not correctly aligned, which means that its optical axis 26 deviates from the x-axis in the azimuth. The result is that azimuth angles φ measured for the different radar targets 20 have a misalignment error δφ. However, unlike distortive angle measuring error Δφ, misalignment error δφ has the same algebraic sign and the same magnitude for all targets 20.

[0030] In the following, a method is described, by which the presence of such angle measuring errors, in particular, the distortive angle measuring errors shown in FIG. 1, may be detected reliably.

[0031] A motor vehicle 28, which moves past a stationary radar target 20, for example, a traffic sign standing on the side of the road, is represented in a sketch in FIG. 3. Specific velocity V of the motor vehicle is drawn in as a vector. A vector V_rel=−V indicates the relative velocity of radar target 20 relative to motor vehicle 28. In the following, it is assumed, for the sake of simplicity, that the moving direction of the antenna of the radar sensor installed in the motor vehicle coincides with the moving direction of the rear axle of the vehicle. In general, the direction of the actual, specific velocity of antenna 14 may deviate, however, from the x-axis of the coordinate system as a function of the mounting location of radar 10 in vehicle 10, due to pitching motions, rolling motions, and rotary motions about the vertical vehicle axis (yaw). This must be appropriately taken into account by using actual velocity V of the antenna at its mounting location and correspondingly corrected angle measuring values ((φ, α) and/or (α, β) are the angles between the actual moving direction of the antenna array and the respective target), or the evaluation is limited to driving situations having negligible pitching, rolling and yawing motions.

[0032] Radar target 20 is located by the radar sensor 10 installed in the front end of motor vehicle 28 (and not shown in FIG. 3). In the situation shown in FIG. 3, a relatively small locating angle (azimuth angle) φ is measured for this target. Vector V_r may be broken down into a radial component along the line of sight between the radar sensor and the radar target, and a transverse component perpendicular to it. The magnitude of the radial component is radial velocity V_r=cos(φ)*V, where V is the magnitude of the specific velocity of the motor vehicle, that is, of the antenna, above ground and, simultaneously, the magnitude of relative velocity V_rel.

[0033] FIG. 4 shows the situation at a later time, when azimuth angle φ has increased and radial velocity V_r has correspondingly decreased in relation to specific velocity V.

[0034] If it is known that radar target 20 is a stationary target, and if specific velocity V of the vehicle, that is, in particular, of the antenna array at the respective mounting location, is additionally known, for example, on the basis of direct measurement by wheel speed sensors on the vehicle, on the basis of the yaw rate, etc., then V_r may be calculated according to the formula V_r=cos(φ)*V indicted above. On the other hand, V_r may also be measured directly, on the basis of the Doppler effect, with the aid of radar sensor 10. A comparison of the measured to the calculated value enables a check as to whether the measurement of the azimuth angle φ was correct.

[0035] In FIGS. 3 and 4, only two spatial dimensions are taken into account. When all three spatial dimensions are considered, then radial velocity V_r is also a function of elevation angle α of radar target 20, namely, according to the formula:

V_r=cos(α).Math.cos(φ).Math.V.  (1)

[0036] FIG. 5 shows radar target 20 in a three-dimensional Cartesian coordinate system having axes x, y, and z. In spherical coordinates, the position of radar target 20 is given by radius r, azimuth angle φ and elevation angle α. For the sake of simplicity, vectorial specific velocity V of the motor vehicle or antenna array is represented as parallel to the x-axis in FIG. 5 and FIG. 6. In addition, a possible azimuthal angle measuring error φ_e and a possible elevation angle measuring error α_e are drawn in.

[0037] The following equations are valid for converting spherical coordinates to Cartesian coordinates:

x=r.Math.cos(φ).Math.cos(α)

y=r.Math.sin(φ).Math.cos(α)

z=r.Math.sin(α)

[0038] Alternatively, as shown in FIG. 6, conical coordinates (r, β, α) may also be used in place of the spherical coordinates shown in FIG. 5. In conical coordinates, elevation angle α has the same meaning as in spherical coordinates. It indicates the angle between the location vector of radar target 20 and the xy-plane. However, in conical coordinates, azimuth angle φ is replaced by angle β, which indicates the angle between the location vector of the radar target and the xz-plane. Thus, the following applies for conversion to Cartesian coordinates:

x=r.Math.(cos.sup.2(α).Math.sin.sup.2(β)).sup.1/2

y=r.Math.sin(β)

z=r.Math.sin(α)

[0039] An example of a possible angle measuring error β_e is drawn in, as well.

[0040] In principle, angle measuring errors φ_e, α_e, β_e may be misalignment errors and/or distortive errors. Methods for detecting misalignment errors as such are conventional. In order to detect distortive errors, as well, for example, the method represented as a flow chart in FIG. 7 may be executed.

[0041] The number of the stationary targets located in a given measuring cycle is denoted by R_m (m is an index, which indicates the measuring cycle). Criteria for distinguishing between stationary and moving targets are conventional and include, in particular, the comparison of the measured relative velocity of the target to the specific velocity of the vehicle. In step S1, a subset P_m is selected from set R_m, the subset being intended to be used to check for distortive measuring errors. The number N m of selected targets should be so large, that a certain degree of compensation for statistical fluctuations is attained. In addition, the selected targets should be distributed as uniformly as possible over as large a solid angle as possible.

[0042] In step S2, the state of motion of reference motor vehicle 28 is estimated, for example, on the basis of signals of wheel speed sensors. Thus, in the coordinate system according to FIG. 5 or FIG. 6, an estimate of vector V is obtained, which indicates the specific motion of the vehicle and, consequently, of the antenna of the radar sensor installed in the vehicle. The result of estimating step S2 may simultaneously form the basis of the identification of stationary targets in step S1 of the next measuring cycle.

[0043] In a further step S3, the validity of the targets selected in step S1 is preferably checked once more. In particular, in view of the specific motion of the vehicle and, in particular, of the antenna, which was determined in step S2. In this context, criteria include, for example, a minimum specific velocity of the vehicle and/or of the radar sensor above ground, the acceleration and yaw rate of the reference vehicle, the number of elements (targets) in P_m, and the spread of the angle measuring data.

[0044] In a further optional step S4, the data, which indicate the specific motion of the vehicle, are verified and possibly updated in light of the radar data obtained in the current and, in some instances, preceding measuring cycles.

[0045] In the example considered here, it should also be assumed that independently of the check for distortive angle errors, a check for misalignment errors also takes place, optionally, on the basis of measurement data for the targets selected in step S1.

[0046] In step S5, the measurement data for the locating angles (e.g., φ and α) are then corrected in view of the detected sensor misalignment, which means that the subsequent check for distortive errors may take place on the basis of more accurate angle measuring data.

[0047] In step S6, an indicator value q_p, which constitutes a measure of the deviation of calculated radial velocity V_r from the radial velocity actually measured on the basis of the Doppler effect, is then calculated for each individual target in set P_m (the targets are identified with the aid of an index p). In this context, the starting point is equation (1). However, it is useful to distinguish between the approach and the moving-away of the radar target, by allowing V_r to take on negative values when the target approaches. Then, the following applies in spherical coordinates:

−V_r/V=cos(α).Math.cos(φ)=cos(α−α_e)−cos(φ−φ_e)  (2)

[0048] where α and φ are the measured values possibly containing errors and α_e and φ_e are the angle measuring errors.

[0049] The following applies analogously in conical coordinates:

−V_r/V=(1−sin.sup.2(β)−sin.sup.2(α)).sup.1/2=(cos.sup.2(β−β_e)−sin.sup.2(α−α_e)).sup.1/2  (3)

[0050] If α_p and φ_p are the measured locating angles of the target having index p and V_r_p is the measured radial velocity of this target, then a suitable indicator value q_p is given, for example, by:

q_p=(−V_r_p/V)−cos.sup.2(α_p)−cos(φ_p)  (4)

[0051] or in conical coordinates:

q_p=(−V_R_P/V)−(cos.sup.2(α_p)−sin.sup.2(β_p)  (5)

[0052] However, different definitions of the indicator values are also possible, for example:

q_p=(−V_r_p/V)−cos.sup.2(α_p).Math.cos.sup.2(φ_p)  (6)

[0053] and

q_p=(−V_r_p/V)−cos.sup.2(α_p)+sin.sup.2(β_p)  (7)

[0054] Since the distortive angle measuring errors are a function of the angle, as was explained with the aid of FIG. 1, the indicator values obtained in step S6 will also be, theoretically, a function of the angle; that is, a different indicator value is obtained, in principle, for each target in P_m. Thus, the sum or the mean of the indicator values will generally be a function of the angular distribution of the targets, as well. In this instance, the indicator values may also take on different algebraic signs, and depending on the angular distribution of the targets, the mean of the indicator values may possibly lie close to zero and simulate a correct measurement, although in reality, a distortive measuring error is present.

[0055] Nevertheless, in order for a meaningful indicator of the presence of distortive errors to be obtained, angle-dependent scaling of the indicator values is carried out in step S7. To that end, (in the case of spherical coordinates,) an arbitrary scaling function F(α, φ) or, in the case of conical coordinates, F(α, β) is defined, which represents the angular dependence of the distortive angle errors at least approximately. The scaling function may be, in turn, a function of the gradient G(α, φ)=−sin (α+φ):

F(α,φ)=f[G(α,φ)]=f[sin(α+φ)]  (8)

[0056] In the case of conical coordinates, a scaling function F(α, β) is formed, which may be, for example, a function F(α, β)=f[G(α,β)] of the gradient G(α, β):

G(α,β)=−[sin(2α)+sin(2β)]*(2*[cos(2α)+cos(2β)]).sup.−1/2  (9)

[0057] An effective value Q_m substantially independent of the angle is then calculated from indicator values q_p for the individual targets, for example, according to the following formula:

Q_m=(Σ.sub.p,|q_p*F(α,φ)|.sup.2)/N_m)).sup.1/2  (10)

where the summation sign means a summation over all of the targets in P_m. In an optional step S9, the effective values obtained in consecutive measuring cycles in, in each instance, step S8 are then subjected to temporal filtering, in order to attain a higher stability with respect to statistical fluctuations. A filtered effective value Q_filt is obtained as a result of the filtering. Finally, in step S10, this filtered value is scaled, using a scaling factor F_scal, and limited by upper and lower limiting values Q_min and Q_max, so that in the end, an indicator value I is obtained, which varies linearly between 0 and 1 in accordance with the function shown in FIG. 8. In step S10, this indicator value I is then outputted to other modules of a driver assistance system, and in these modules, it permits an evaluation of the accuracy and reliability of the results of the angle measurements.

[0058] The information used for calculating indicator value I is independent of the phase information obtained in the receiving channels of antenna array 14 and forms a measure, which characterizes the angle errors and is independent of the classical angle estimation. In particular, angle errors or angular blindness of the radar sensor may also be detected, if the quality of the angle estimate is so high, that an error would not be deduced from the quality.

[0059] Assuming that elevation angle α is error-free, a correction value, which, apart from ambiguity in the algebraic sign, indicates azimuthal angle measuring error φ_e (in spherical coordinates) and β_e (in conical coordinates), may also be derived from equation (2) or (3) (by solving for φ or β) Conversely, a correction value for the elevation angle may be derived, assuming that the azimuth angle is error-free.