Use of a radar sensor with a waveguide antenna array for a method for determining an estimated ego velocity value and an estimated angle value of targets
Patent Information
- Application Number
- EP2023715517
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2022-05-24
- Filing Date
- 2023-03-29
- Publication Date
- 2025-04-09
AI Technical Summary
Conventional radar systems in the automotive sector face challenges in accurately determining the airspeed and angle estimates of targets, especially moving targets, due to assumptions of stationary environments and the need for additional sensors like IMU or computationally intensive autofocus algorithms, leading to incorrect and blurry SAR images.
A radar sensor with a waveguide antenna array employing digital beamforming and MIMO technology to differentiate between stationary and moving targets by calculating individual airspeed estimates and angle estimates directly from measurements, without requiring additional sensors, using a combination of digital beamforming, Doppler effect evaluation, and clustering of airspeed estimates to identify moving targets.
Enables accurate airspeed and angle estimation for both stationary and moving targets, reducing computational effort and eliminating the need for external sensors, while minimizing antenna channels and installation space, thus improving dynamic range and clarity in target detection.
Smart Images
Figure 1.1
Abstract
Description
[0001] Description
[0002] title
[0003] Use of a radar sensor with a waveguide antenna array for a method for determining an airspeed estimate and an angle estimate of targets
[0004] The present invention lies in the field of radar sensors and relates in particular to the fields of multiple input multiple output (MIMO) and synthetic aperture radar (SAR).
[0005] State of the art
[0006] Radar systems for measuring the distance, relative speed, and angle of objects are increasingly being used in motor vehicles for safety and comfort functions. Today, multiple-input-multiple-output (MIMO) radar, i.e., radar with multiple transmission channels for transmitting and receiving radar signals, is particularly popular for this purpose. Recently, synthetic aperture radar (SAR) has become popular. The principle of the synthetic aperture allows for particularly precise angle measurements when the radar sensor is moving. The synthetic aperture takes advantage of the fact that the transmitting and receiving antennas are in different positions at the time of each measurement due to the radar sensor's own movement. The measurements are then processed to create a synthetic antenna aperture. For evaluation purposes, this can be equated with a large antenna aperture along the driving trajectory.This achieves a large synthetic aperture, which would be impractical or even impossible with a real antenna aperture due to the large number of antenna elements required. SAR enables higher resolutions in angle measurements than with real antenna apertures.
[0007] To evaluate the measured radar signals as a synthetic aperture, the radar environment is typically assumed to be stationary. In addition, the radar sensor's own motion, and thus the positions at which the individual measurements were taken, must be known. The radar's trajectory is incorporated into the SAR evaluation algorithm and forms the basis for calculating the SAR image. Depending on the evaluation algorithm, an estimated airspeed may be sufficient for calculating the SAR image instead of the more precise trajectory. The trajectory is typically assumed to be linear; more complex trajectories, however, cannot be mapped.
[0008] Today's automotive radar systems typically employ frequency-modulated continuous wave (FMCW) radar with rapidly rising ramps—so-called fast-chirp modulation—in which several linear frequency ramps with the same steep gradient are successively passed through. Mixing the instantaneous transmitted signal with the received signal produces a low-frequency signal (called the beat frequency), whose frequency is proportional to the distance. The system is typically designed so that the component of the beat frequency caused by the Doppler frequency is negligible. The distance information obtained is largely unambiguous. Furthermore, a Doppler shift can be determined by observing the temporal evolution of the phase of the complex distance signal across the ramps, and from this, the relative speed can be calculated. The distance and relative speed are determined independently of each other.Typically, a two-dimensional Fourier transform is used for this purpose.
[0009] Conventional SAR analysis assumes stationary targets. Moving targets that do not meet this assumption result in an erroneous, angularly shifted, and blurred image in the SAR image. However, moving targets are also of interest in motor vehicles (e.g., to avoid collisions with them). Two approaches are known to estimate the radar sensor's trajectory or velocity: One approach involves using an external sensor—e.g., an inertial measurement unit (IMU) or an odometry sensor. The other approach involves using highly computationally intensive autofocus algorithms that are not applicable for real-time processing. Disclosure of the Invention
[0010] The subject matter of the invention is the use of a radar sensor with a waveguide antenna array, described in detail below, for a method for determining an estimated airspeed and an estimated angle of several targets in the environment, the method steps of which are described in detail below.
[0011] The waveguide antenna array of the radar sensor has at least two groups of antenna units, with antenna elements in each antenna unit arranged next to one another in a first direction. In a first group, the antenna units are arranged offset from one another in a second direction perpendicular to the first direction. In a second group, the antenna units are arranged offset from one another in the first direction. This creates a multiple-input-multiple-output (MIMO) radar sensor.
[0012] The radar sensor uses digital beamforming (DBF), allowing the antenna elements of an antenna unit to jointly receive and evaluate a radar signal. The first group, with the antenna units facing the second direction, is used primarily for measuring the azimuth angle, while the second group, with the antenna units facing the first direction, is used primarily for measuring the elevation angle.
[0013] In the second group, the antenna units can be arranged offset from each other only in the first direction. This reduces the computational effort required for digital beamforming. Optionally, the antenna units in the second group can also be offset from each other in the second direction. This enables measurement in the second direction for the second group, and the radar sensor is designed for multiple-input-multiple-output (MIMO).
[0014] Preferably, the array groups are alternately assigned to either the transmitting or receiving side. According to the MIMO principle, transmitting and receiving antennas are essentially interchangeable. A distinction is made below between stationary targets (also referred to as fixed targets) and moving targets (also referred to as moving targets). Stationary targets are objects in the environment that are not moving, such as buildings, trees, infrastructure on and along roads, etc. Moving targets are objects in the environment that are moving, such as other vehicles, pedestrians, other road users, etc.
[0015] The radar sensor moves, transmitting a plurality of measurement signals and receiving signals reflected from targets. The radar sensor is therefore a synthetic aperture radar sensor. A relative velocity is determined for each target from the transmitted and received signals. To determine the relative velocity, the Doppler effect in the measurement signals is evaluated, and in particular, a Doppler shift is determined. Furthermore, the distance between the radar sensor and the target is also determined from the transmitted and received signals. This can be done, for example, using Fourier processing. To identify the targets in the measurements, detection with a constant false alarm rate (CFAR) is performed.
[0016] A rough angle estimate is then performed. For each target, an angle estimate is estimated that characterizes the target angle between the direction of the radar sensor's own speed – i.e., the direction in which the radar sensor is moving ("forward direction") – and the respective target. The angle estimate can be performed, for example, using digital beamforming. For this purpose, the radar sensor has at least one additional receive channel and / or at least one additional transmit channel. Preferably, the angle estimate directly indicates the target angle. However, the target angle can also be derived indirectly from the angle estimate through a transformation or from mathematical relationships. Since the angle estimates are subsequently further processed, the angle estimate can be significantly less accurate than conventional angle measurement.
[0017] By backprojection, an individual airspeed estimate for the radar sensor is calculated separately for each target using the relative speed and the angle estimate. This means that for each target, the measured or estimated values are used to obtain an individual airspeed estimate for the radar sensor. Typically, a multitude of individual airspeed estimates are determined, which are therefore dependent on the target's speed. For stationary targets, the individual airspeed estimates are close to each other, since the relative speed between the target and the radar sensor is proportional to the radar sensor's airspeed and the target angle.For moving targets, however, the individual airspeed estimates differ further, as the relative speed depends not only on the radar sensor's own speed and the target angle, but also on the target's speed. In addition, in a typical situation, there are significantly more stationary targets in the vicinity than moving targets with the same relative speed to the radar sensor, and the moving targets typically have different speeds from each other.
[0018] This enables classification and distribution of individual airspeed estimates, particularly through clustering. For this purpose, a range is defined for the individual airspeed estimates, which allows a distinction to be made between stationary targets and moving targets. Individual airspeed estimates that lie within a predefined range are assigned to stationary targets. Individual airspeed estimates that lie outside the range are assigned to moving targets. This allows moving targets to be identified (MTI - Moving Target Indication). The calculated individual airspeed estimates can be incorporated into a histogram, for example, for classification purposes.
[0019] The individual airspeed estimates are then evaluated separately depending on their assignment. For the stationary targets, a combined airspeed estimate is determined from the individual airspeed estimates assigned to the stationary targets. The combined airspeed estimate can be regarded as the actual airspeed of the radar sensor, since this was essentially calculated only from the stationary targets (autofocus). In addition, a corrected angle estimate is calculated for the stationary targets using the combined airspeed estimate and the respective measured relative speed. The corrected angle estimate can be regarded as the actual angle of the target relative to the radar sensor.
[0020] The method also enables the determination of the airspeed estimate and the angle estimate directly from the measurements, without the need for additional sensors such as IMUs or odometry sensors. The odometry sensors typically used in vehicles are usually located too far from the radar sensor and also perform too few measurements per time interval.
[0021] By using the radar sensor with waveguide antenna array for the described method, a waveguide antenna with MIMO and the SAR concept are combined. The same antenna units or
[0022] Antenna elements are used for both MIMO and SAR. As a result, the number of radar sensor channels can be reduced, thus minimizing the antenna area. Reducing the number of channels is particularly advantageous for waveguide antennas, as these typically require a large installation space due to their three-dimensional structure and are complex to manufacture.
[0023] The use of waveguide antennas allows for a great deal of freedom in the arrangement of the antenna array elements. This allows for a simple, advantageous λ / 2 arrangement of the antenna units, which is advantageous both in terms of dynamics and unambiguous angle estimation.
[0024] To ensure that the received, downconverted, and baseband-sampled radar signals of the antenna array are compatible with the combination of MIMO and SAR, unambiguous sampling of the Doppler frequency must be enabled. For the unambiguously detectable range of the radial component of the relative velocity v rel,r,u applies: V rel,r,max and v rel,r,min the upper or lower limit of this range. f D,u denotes the bandwidth of the maximum clearly detectable Doppler frequency, c denotes the wave propagation velocity, f c the carrier frequency for the radar signal and T ctc the duration between two frequency ramps (chirp-to-chirp) of the same transmitting antenna.
[0025] In time division multiplexing, the duration between two frequency ramps T ctcWith a fixed bandwidth and fixed ramp slope or ramp duration, the number of transmitters increases, so that the number of transmitters in conventional FMCW MIMO radar sensors based on time division multiplexing is limited. Preferably, the present invention performs multiplexing in the frequency dimension (FDM, Frequency Division Multiplexing) or in the code dimension (CDM, Code Division Multiplexing). For SAR, there are no such limitations regarding the number of transmitters.
[0026] Advantageously, the predeterminable range used for classifying and dividing the individual airspeed estimates is an error tolerance range for the measurements. This is determined from the error for the relative speed measurement and the error for the angle estimation. This ensures that the division is carried out specifically within the error limits for the measurements, thus providing the greatest possible selectivity.
[0027] To determine the combined airspeed estimate, an average speed value of the individual airspeed estimates can be calculated. This can be done using conventional averaging, such as an arithmetic mean, a weighted averaging (e.g., with weights dependent on the signal-to-noise ratio), determining the maximum in the histogram, calculating a median, etc.
[0028] Preferably, a respective angle estimate and a respective speed estimate are also determined for the moving targets. However, since the speed of the moving target is unknown, the evaluation described above would lead to an erroneous angle estimate. The angle estimate resulting from the above-mentioned angle estimation for this same moving target can be used as the angle estimate for the respective moving target. Although this does not achieve an improved angle estimate, it does avoid the erroneous angle estimate. Furthermore, a radial speed estimate for the respective moving target can be determined from the relative speed measured by Doppler shift. For this purpose, the combined airspeed estimate determined above is assumed to be the airspeed of the radar and, weighted with the cosine of the target angle, is subtracted from the relative speed.
[0029] The movements of both the radar and the moving targets are assumed to be two-dimensional in a plane. However, the target measured by the sensor may be at a different elevation relative to the plane. This can occur, for example, when only part of an object is detected. In this case, an elevation angle between the plane and the target can be determined for the target. Preferably, the elevation angle for each target is taken into account when determining the radar sensor's individual airspeed estimate using the relative speed and the estimated angle.
[0030] Preferably, the radar sensor is a chirp sequence radar, which functions as a frequency-modulated continuous wave radar and emits chirp signals with rapidly rising ramps. This allows the distance to be easily measured in a conventional manner. Furthermore, the Doppler effect, in particular the Doppler shift, can be determined from the temporal evolution of the phase of the complex distance signal across the ramps, allowing the relative velocity to be measured.
[0031] Known methods can be used to determine the relative velocity using the Doppler effect. Keystone processing based on the chirp-Z transform (CZT) is preferred, as it is particularly computationally efficient and can compensate for any migration effects that may occur.
[0032] Short description of the drawings
[0033] Embodiments of the invention are illustrated in the drawings and explained in more detail in the following description. Figure 1 shows an isometric view of a waveguide antenna of a radar sensor used for the method according to the invention.
[0034] Figure 2 shows a schematic representation of a traffic situation, with a vehicle with the radar sensor and various targets and corresponding angles and relative speeds.
[0035] Figure 3 shows a flowchart of an embodiment of the method.
[0036] Figure 4a shows a diagram of the distribution of individual airspeed estimates of the radar sensor for different targets.
[0037] Figure 4b shows a histogram for the distribution from Figure 4a.
[0038] Figure 5 shows a location diagram of a trajectory generated using an embodiment of the method according to the invention and a trajectory generated using an odometry sensor
[0039] Embodiments of the invention
[0040] Figure 1 shows a waveguide antenna 100 of a radar sensor S (not shown in detail here). The waveguide antenna 100 has a waveguide antenna array consisting of several antenna elements 101. Several antenna elements 101, twelve in this example, are arranged in a column in a first direction R1 and together form an antenna unit 102 (in Figure 1, an antenna unit is marked with a frame as an example). In this example, the first direction R1 corresponds to the vertical in a global reference system. The antenna elements 101 of an antenna unit 102 transmit and receive radar signals together. In Figure 1, the phase centers 103 are marked for each antenna unit 102. The antenna units 102 of the waveguide antenna array are divided into two groups 104, 105.In this example, the first group 104 comprises eight antenna units 102, each having twelve antenna elements 101 arranged in columns in the first direction R1. In the first group 104, the antenna units 102 are offset from one another in the second direction R2. In this example, the second direction R2 corresponds to one of the horizontals in a global reference system and, in this example, runs in the direction of the airspeed v, as shown in Figure 2. ego of a vehicle F, which has the radar sensor S. In general, the second direction R2 can also be at an angle to the vehicle's own speed v egoof the vehicle F. This only swivels the antenna lobe. The second direction R2 is associated with the azimuth angles α1, α2, α3, and the first group 104 of antenna units 102 serves to measure the azimuth angles α1, α2, α3. The second group 105, in this example, comprises three antenna units 102, each of which has twelve antenna elements 101 arranged in columns in the first direction R1. In the second group 105, the antenna units 102 are offset from one another in both the first direction R1 and the second direction R2. The second group 105 of antenna units 102 serves to measure the elevation angles Φ iand for measuring the azimuth angles α1, α2, α3. In the state shown here, the first group 104 of antenna units 102 is assigned to the receiver side RX and the second group 105 of antenna units 102 is assigned to the transmitter side TX. The radar signals received by the first group 104 are processed using digital beamforming. However, the assignment can also be reversed, so that the first group 104 is assigned to the transmitter side TX and the second group 105 is assigned to the receiver side RX. The waveguide antenna 100 is thus designed for MIMO. In other embodiments not shown, the antenna units 102 in the second group 105 can be offset from one another only in the first direction R1. This simplifies two-dimensional digital beamforming. The radar sensor S with the described waveguide antenna 100 or the described waveguide antenna array is used for the method described below.
[0041] Figure 2 shows a schematic representation of a traffic situation, with a vehicle F, which has a radar sensor S with a waveguide antenna 100 described above, and several other vehicles, which are designated as targets Z1 to Z3. Typically, there are other targets in the surrounding area that are not shown here, such as buildings, the road infrastructure, i.e. traffic signs, crash barriers and the like, or the road itself. The vehicle F and thus also the radar sensor S move at an intrinsic speed v ego along a straight line. From the radar sensor S, the azimuth angle α1, α2, α3 between the direction of the airspeed v ego and the direction of the respective target Z1, Z2, Z3. In addition, the relative velocity v rel,1 , v rel,2 , v rel,3of each target Z1, Z2, Z3 relative to the radar sensor S. If one of the targets, for example target Z1, is a stationary target, ie it is not moving, the corresponding relative speed v rel,1 as a projection of the airspeed v ego of the radar sensor S to the corresponding azimuth angle ai. The projections are shown in Figure 1 for all three targets Z1, Z2, Z3. For a moving target, for example, target Z2, which moves with an unknown speed, the speed of target Z2 is part of the relative speed v rel,2 and the measured relative velocity v rel,2 deviates from the projection.
[0042] Figure 3 shows a flowchart of an embodiment of the method according to the invention. A plurality of targets, generally designated i here, are examined. While the vehicle F and the radar sensor S are moving, measurements 1 are performed. The measurements 1 are carried out using frequency-modulated continuous wave radar modulation (FMCW), in which chirp signals with rapidly rising linear frequency ramps of the same gradient are output at predetermined time intervals. The reflected signals are recorded and processed as received signals. Mixing the instantaneous transmitted signal with the received signal produces a low-frequency beat signal whose frequency is proportional to the distance to the target i. The measurements 1 are carried out such that the Doppler effect or the Doppler shift in the beat frequency is negligible or is taken into account in the evaluation.
[0043] Subsequently, keystone processing 2 is carried out. This involves estimating the Doppler shift or Doppler frequency by determining the temporal evolution of the phase of the complex measurement signals across the frequency ramps and compensating for the corresponding linear distance change (migration) for each estimated value. This results in relative velocities for each target i. Subsequently, a distance estimation is performed using conventional Fourier processing 3, in particular a fast Fourier transform (FFT) from the time domain to the frequency domain. The generated two-dimensional spectra (distance and relative velocity) of the individual transmit-receive channel combinations are non-coherently averaged 4. For this purpose, the magnitude of each of these spectra is calculated, and these magnitudes, or their squared magnitudes, are then summed. To identify the targets in the measurements, detection with a constant false alarm rate (CFAR) 5 is performed.
[0044] Furthermore, an angle estimation 6 is carried out, in which azimuth angle estimates for the targets. The azimuth angle estimate represents the azimuth angle between a measuring axis of the radar sensor S and the target i and thus also reflects the installation situation of the radar sensor S. Since the installation situation is known, the azimuth angle estimate can be converted into an estimate for the azimuth angle α i between the direction of the airspeed v ego and the direction of the target i. For the case shown in Figure 2, the measuring axis is perpendicular to the direction of the airspeed ν ego . This gives the following relationship: θ i = 90 - α i Digital beamforming is used for the angle estimation 6. Simultaneous measurements are carried out across the multiple antenna units 102 on the waveguide antenna array and a phase difference is calculated, from which the azimuth angle estimate is then calculated. The influence of the airspeed v egoof the radar sensor S is negligible for this type of angle estimation, so that the azimuth angle estimates independent of the airspeed v ego of the radar sensor S. In the angle estimation 6, an elevation angle Φ i between the plane in which the vehicle is moving and the height at which the target i is detected.
[0045] For each target i, the relative speed The azimuth angle estimate and, if applicable, the elevation angle are known. for each target i separately an individual airspeed estimate calculated according to formula 2 7: Figure 4a shows individual airspeed estimates calculated in this way for some targets i in a diagram. Figure 4b shows a histogram in which for a large number of different individual airspeed estimates, the determined number n In both figures it can be seen that the individual airspeed estimates cluster in an area B. In a In a typical traffic situation, there are significantly more stationary targets than moving targets with the same relative speed radial to the radar sensor S.
[0046] With reference to Figure 3, a clustering 8 is carried out in which the individual airspeed estimates that are within the area B are assigned to stationary targets and the individual airspeed estimates that are located outside of area B are assigned to moving targets. This way, the moving targets are identified (MTI - Moving Target Indication) and separated from the stationary targets. Area B is defined by the errors in measurement 1 and angle estimation 6 and represents an error tolerance range.
[0047] The individual airspeed estimates that represent the stationary assigned to targets, ie located within area B, are averaged 9 to obtain a combined airspeed estimate Different types of averaging can be performed, for example, a classical averaging, such as an arithmetic mean, a weighted averaging, e.g., with weights dependent on the signal-to-noise ratio, determining the maximum in the histogram, forming a median, etc. Since the combined airspeed estimate was calculated without the moving targets, it can be used as an estimate for the actual airspeed of the Radar sensor S. This achieves autofocus. For each stationary target, an angle calculation 10 is also carried out from the relative speed for the stationary target determined by keystone processing 2 using the Doppler effect and the calculated Individual airspeed estimate for the stationary target using Formula 3: (Formula 3)
[0048] Consequently, a corrected angle estimate is calculated, which is known as actual azimuth angle of the target with respect to the radar sensor S.
[0049] For moving targets, however, the angle calculation 10 described above would lead to an incorrect angle estimate, since the velocity component of the moving target is unknown and therefore cannot be taken into account. Consequently, for a moving target, the azimuth angle estimate determined during angle estimation 6 is adopted 11.
[0050] Although this does not result in an improved angle estimate, it does prevent the erroneous angle estimate. Finally, a radial velocity estimate of the moving target can be calculated 12 by subtracting the relative velocity determined by keystone processing 2 using the Doppler shift from the velocity of the stationary targets by averaging 9. determined combined airspeed estimate weighted with subtracted from the cosine of the azimuth of that target.
[0051] Figure 5 shows a comparison between an odometry trajectory T o , which was determined conventionally using an odometry sensor, and a trajectory Tv generated using an embodiment of the method according to the invention. It can be seen that the two trajectories match very well and thus the autofocus using the method according to the invention delivers precise results.
Claims
Claims 1. Use of a radar sensor (S) with a waveguide antenna array, comprising at least two groups (104, 105) of antenna units (102) comprising multiple antenna elements (101), wherein antenna elements (101) in each antenna unit (102) are arranged side by side in a first direction (R1), wherein in a first group (104) the antenna units (102) are arranged offset from one another in a second direction (R2) perpendicular to the first direction (R1), and wherein in a second group (105) the antenna units (102) are arranged offset from one another in the first direction (R1), for a method for determining an intrinsic velocity estimate and an angle estimate of goals (i) with the following steps: Using the radar sensor (S), measure (1) a distance (A) between the radar sensor (S) and the respective target (i), and use the radar sensor (S) to measure (1) a relative velocity of the respective target (i). Doppler effect; Angle estimation (6) of each angle estimate , which characterizes the angle between the direction of the radar sensor's own velocity (S) and the respective target (i); Determining (7) an individual self-speed estimate of the radar sensor (S) using the relative speed and the angle estimate for each target (i); Classifying and dividing (8) the individual velocity estimates relating to stationary targets, whose individual- Own velocity estimates within a within a predefinable area (B) and concerning moving targets, whose individual velocity estimates lie outside the area (B); Determining (9) a combined propulsion estimate from the individual velocity estimates of the stationary targets; and Determining (10) a corrected angle estimate for stationary targets using the combined proper velocity estimate and the respective measured relative velocity Use of the radar sensor for the method according to claim 1, wherein in the second group the antenna units are additionally arranged offset from one another in the second direction. Use of the radar sensor for the method according to claim 1 or 2, wherein the array groups are alternately assigned to either the transmitting side or the receiving side. Use of the radar sensor for the method according to any one of the preceding claims, wherein the predefinable range (B) is an error tolerance range consisting of the error for the measurement (1) of the relative velocity and the error for the angle estimation. (6) is determined. Use of the radar sensor for the method according to one of the preceding claims, wherein an averaged velocity value for the stationary targets is determined as a combined self-speed estimate by weighted or unweighted averaging. Use of the radar sensor for the method according to one of the preceding claims, wherein for a moving target the angle estimate resulting from the angle estimation (6) is used as the angle- An estimated value for the moving target is taken (11). Use of the radar sensor for the method according to one of the preceding claims, wherein a velocity estimate for a moving target is derived from the relative velocity of the target, which is obtained with The Doppler effect is used to measure the value (12). Use of the radar sensor for the method according to one of the preceding claims, wherein when determining (7) the individual self-speed estimate of the synthetic aperture radar sensor (S) using the relative velocity and the angle estimate for each target (i) an elevation angle (Φ i ) is taken into account. Use of the radar sensor for the method according to one of the preceding claims, wherein the radar sensor (S) is a chirp sequence radar. Use of the radar sensor for the method according to one of the preceding claims, wherein the determination of the relative velocity using the Doppler effect by means of a Keystone processing (2) is performed.