Method for estimating a radar angle

Abstract

The invention relates to a method for estimating an angle using transmitted signals, which are received after being reflected on an object, of a radar sensor that is angle-resolving in at least one dimension and has a MIMO-capable antenna array, wherein in order to estimate the locating angle of a radar target, a cross-path model is used which also models reflections of transmitted and / or received signals on a reflecting surface. The invention is characterized in that the angle is estimated on the basis of a reduced cross-path model, and a model for a SIMO mode (48) of the radar sensor is combined with a model for a MISO mode (50).

Description

[0001] Description

[0002] title

[0003] Method for radar angle estimation

[0004] Description

[0005] The invention relates to a method for angle estimation based on signals transmitted and received after reflection from an object by a radar sensor which has an angle resolution in at least one dimension and which has an Ml MO-capable antenna array, wherein a cross-path model is used to estimate the location angle of a radar target, which also models reflections of transmitted and / or received signals from a reflecting surface.

[0006] In particular, the invention relates to methods for radar sensors used in driver assistance systems of motor vehicles for environmental detection.

[0007] State of the art

[0008] In driver assistance systems, in addition to the distance and relative speed of the located objects, the azimuth angle and the elevation angle are also important, as this angle information can be used to assign lanes and to make a statement about the relevance of the target (overridable / oncoming / underridable). Azimuth and elevation angles of the targets can be determined from amplitude and / or phase differences of transmitting and / or receiving antennas of an antenna array. To ensure the accuracy and discriminatory power of the

[0009] To improve angle estimation, the MIMO (multiple input multiple output) principle is often used for radar sensors. Unlike traditional SIMO (single input multiple output) radars, which use only one transmitting antenna and several receiving antennas, this method uses multiple transmitting antennas and multiple receiving antennas. Angle estimation involves comparing the received signals with a previously measured angle-dependent antenna pattern. If only one target is located in a (d,v) cell (distance / speed cell), the estimated angle is the position of the best match between the received signal and the antenna pattern.

[0010] In M. Wintermantel: "Radar system comprising overlapping transmitter and receiver antennas", as well as in US Pat. No. 8,436,763 B2, a MIMO radar sensor is described that uses the MIMO principle with code division multiplexing and two transmit antennas to improve azimuth angle estimation. The two transmit antennas are arranged at the left and right edges of the overall array, respectively, in order to achieve the largest possible virtual aperture. In the case of multipath propagation due to reflections from, for example,a guardrail or the road surface, four different paths of signal propagation occur: a path on which the signal propagates directly from the radar sensor to the object and back to the radar sensor without reflection; a path on which the signal propagates directly from the radar sensor to the object but is reflected on the way back to the radar sensor; a path on which the signal is reflected on the way to the object and on the way back directly to the radar sensor without reflection; and finally a path on which the signal is reflected on the way to the object and the radar echo is reflected again on the way back from the same surface. Signal models that depict this scenario are called cross-path models. If multipath propagation is ignored in the signal model, a MIMO angle estimation produces incorrect estimates with errors of several angular degrees.This can then lead to undesirable system behavior, such as side-lane interference or target loss. Multipath propagation also means that the virtual array model cannot be used in MIMO beamforming (beamforming both the transmit and receive beams). In F. Engels, P. Heidenreich, M. Wintermantel, L. Stacker, M. AI Kadi, and AM Zoubir, "Automotive Radars Signal Processing: Research Directions and Practical Challenges," in IEEE Journal of Selected Topics in Signal Processing, doi: 10.1109 / JSTSP.2021.3063666, beamforming with separate gratings for the transmit angle (direction-of-departure (DOD)) and the receive angle (direction-of-arrival (DOA)) is presented as a mitigation measure.

[0011] Furthermore, a method for a multi-mode radar sensor was proposed that uses the full multipath signal model. While this allows for robust angle estimation, it requires considerable computational effort and thus high computing power.

[0012] Disclosure of the invention

[0013] The object of the invention is to provide a method which, in the case of multipath propagation, enables an accurate and reliable angle estimation with reduced computational effort.

[0014] This object is achieved according to the invention in that the angle estimation is carried out on the basis of a reduced cross-path model that combines a model for a SIMO mode of the radar sensor with a model for a MISO mode.

[0015] In SIMO mode (single input multiple output), a signal is sent from a single transmitting antenna, and the signal received by multiple receiving antennas is evaluated in multiple receiving channels to calculate an angular spectrum in at least one dimension based on the phase and amplitude relationships and a cross-path model. In MISO mode (multiple input signal output), on the other hand, signals are sent in time, frequency, or code division multiplex from multiple transmitting antennas, and only the signals in a single receiving channel are evaluated to evaluate the associated radar echoes. In this way, again based on a cross-path model, another angular spectrum is obtained, the information content of which, however, differs from that in SIMO mode.

[0016] The information available in the two spectra is then combined. This achieves a similarly precise and reliable angle estimation as with the MIMO method. However, the computational effort is significantly reduced due to the limitation to a single transmitting antenna in SIMO mode and the limitation to a single receiving antenna in MISO mode. Advantageous embodiments and further developments of the invention are set out in the subclaims.

[0017] Within the framework of a maximum likelihood angle estimation (in azimuth and / or in elevation), the angle spectra obtained in SIMO mode and MISO mode can be non-coherently added to form a one- or two-dimensional sum spectrum.

[0018] The following example is explained in more detail using the drawing. It shows:

[0019] Fig. 1 is a schematic representation of the analog part of a radar system;

[0020] Fig. 2 is a diagram of an antenna array of the radar system; Fig. 3 is a diagram illustrating a multipath propagation scenario;

[0021] Fig. 4 is a block diagram of the method according to the invention; Fig. 5 is an example of an angular spectrum obtained in SIMO mode;

[0022] Fig. 6 an angular spectrum obtained in MISO mode for the same

[0023] Environment configuration as in Fig. 5;

[0024] Fig. 7 shows a sum spectrum formed from the spectra of Figs. 5 and 6; and Fig. 8 shows a comparison spectrum calculated for the same environmental constellation using a complete MIMO model.

[0025] An embodiment of a radar system with which the method according to the invention is carried out is explained with reference to Figures 1 to 4.

[0026] Fig. 1 shows schematically and simplified the structure of the analog part of the radar system.

[0027] A frequency modulation device 10 controls an RF oscillator 12, which generates sequences of signals in the form of frequency ramps for a plurality of transmitting antennas 14. An amplifier 16 is arranged in each of the multiple transmitting channels, which either blocks the signals or forwards them in amplified form to the associated antenna. The oscillator 12 and the amplifiers 16 are controlled by a multiplexing device 18, for example, according to a time and frequency division multiplexing scheme, so that each of the transmitting antennas 14 transmits a frequency-modulated signal in a specific frequency subband within specific time slots.

[0028] The transmitted signal, reflected by an object 24, is received by several receiving antennas 26 and, in each receiving channel, mixed with a portion of the signal from the RF oscillator 12 by a mixer 28, bringing the signal into a low-frequency range. An A / D conversion then occurs in the usual manner by an A / D converter 30. The digitized signals are then further processed in a digital evaluation stage 32.

[0029] As shown in Fig. 2, the transmitting antennas 14 form a transmitting array 34, and the receiving antennas 26 form a receiving array 36. In the example shown, both arrays are two-dimensional, so that in principle MIMO angle measurements would be possible in both azimuth and elevation.

[0030] In the receiving array 36, the receiving antennas 26 are arranged at equal intervals in an angular resolution direction y, e.g., in the azimuth direction. The distances between the individual receiving antennas are so large that a large aperture and correspondingly high angular resolution can be achieved with just a few antennas. However, the distances from antenna to antenna are greater than half the wavelength of the radar radiation, so the Nyquist uniqueness criterion is not met.

[0031] In the example shown here, the receiving antennas 26 are also arranged at equal distances in elevation (in the angular resolution direction z), and in this direction too, the antenna spacing is so large that non-unique subsampling occurs.

[0032] In this example, the transmit antennas 14 of the transmit array 42 are also arranged at equal intervals in azimuth and elevation, but the spacing is chosen to allow for a clear angle measurement. However, the aperture is significantly smaller than that of the receive array 44, resulting in lower angular resolution.

[0033] The equidistant arrangement of the antenna elements (in azimuth and elevation) facilitates the evaluation of the data, as it enables, for example, the use of a fast Fourier transformation (FFT) to calculate a two-dimensional angular spectrum.

[0034] In the evaluation stage 32, a two-dimensional spectrum in the dimensions of distance and relative velocity is first calculated in a known manner using Fourier transformation. Based on this spectrum, individual objects can then be identified and their distances and relative velocities determined. In the case of a single-target scenario, i.e., when there is only a single object in each distance / velocity cell, the well-known MIMO model, which will be briefly outlined below, can be used to estimate the angle in azimuth and elevation for each object. For simplicity, only the angle estimation in azimuth is considered, for which only the first row of transmit antennas 14 in the transmit array 34 and only the first row of receive antennas 26 in the receive array 36 needs to be used.Let xn denote the four-component vector whose components (xn,1 , xn,2 , xn,3 , xn,4 ) indicate the complex amplitudes of the signals transmitted by the nth transmitting antenna 14 and received by the four receiving antennas 26. If d is the distance from antenna element to antenna element, I is the wavelength of the radar radiation, and s = xn,1 is the complex amplitude of the signal received by the first receiving antenna (for example, the rightmost receiving antenna in Fig. 2), the following relationship holds due to the runlength differences between the signals reaching the various receiving antennas 14:

[0035] The superscript symbol "T" denotes transposition, since vectors are written here as row vectors but should be considered as column vectors. The vector a rxis called the receive control vector. This control vector specifies the geometric properties and wave propagation characteristics of the receive array.

[0036] Accordingly, a control vector atx can also be defined for the transmit array 34, which indicates the run length differences of the optical paths from the transmit antennas to the object 24.

[0037] For the entire MIMO antenna array, the control vector a(q) = atx(q) * arx(q) 000 is obtained.

[0038] The symbol * here means the Kronecker product.

[0039] The received signals form a vector x with Ntx . Nrx components (96 components in this example, since the number Ntx of transmitting antennas is 8 and the number

[0040] Nrx of the receiving antennas 12), and x(q) = sa(q) applies. Knowledge of the steering vector a(q) makes it possible to establish a relationship (unique under suitable conditions) between the angle q of the object and the received signals x, and to deduce the azimuth angle q of the object from the amplitude and phase relationships of the received signals. However, since in practice the received signals will be more or less noisy, the azimuth angle cannot be calculated exactly, but can only be estimated, for example using a deterministic maximum likelihood estimation.

[0041] If we generalize this principle to multi-objective estimations, the single angle q becomes a vector q whose components specify the angles of the different objectives, the control vector a becomes a control matrix A, and the relationship x = A q

[0042] In the case of two targets with the detection angles qi and q2: x = [ atx(qi) * arx(qi) atx(q2) * arx(q2) ]

[0043] Fig. 3 outlines a scenario characterized by multipath propagation.

[0044] The signal sent by the radar sensor 8 can propagate not only on a direct path 38 to the object 24, but also on an indirect path 40, which first leads to a reflecting surface 42, for example a guardrail, and is then deflected to the object 24. Likewise, the signal reflected by the object 24 can

[0045] Signal 44 propagates not only on a direct path 44 to the radar sensor 8, but also on an indirect path 46, on which the signal is also reflected at the surface 42, so that a mirror object 24' is simulated to the radar sensor. In the case of multipath propagation, the complete MIMO signal model is a cross-path model which has the following form: x = [ atx(qi) * arx(qi) atx(q2) * arx(q2) atx(q2) * arx(qi) atx(qi) * arx(q2) ] Although the last two terms can be combined into a single path due to the reciprocity of the cross-paths, ie the path combinations 38, 46 and 40, 44, the computational effort for calculating the signal vector x is nevertheless considerably greater than in the case of one-way propagation.

[0046] This effort can, however, be reduced by applying a method shown as a flow chart in Fig. 4. In a SIMO mode 48, only signals that were transmitted by a single one of the transmitting antennas 4 and received by the multiple receiving antennas 26 are evaluated. In parallel, in a MISO mode 50, only signals that were transmitted by the multiple transmitting antennas 14 and received by only a single receiving antenna 26 are evaluated. In the case of frequency or code division multiplexing, the frequencies or the coding of the signals can be used to distinguish which signal originates from which transmitting antenna. In the case of time division multiplexing, multiplexing can be suspended in SIMO mode 48, which simultaneously results in a desired shortening of the measurement cycle.

[0047] The cross-path model described above for all four paths reduces for SIMO to: x = [ arx(qi) * arx(q2) ] and MISO to: x = [ atx(qi) * atx(q2) ]

[0048] This results in a significant reduction in model complexity and computational load.

[0049] In a DML estimation stage 52, a spectrum is formed from the complex amplitudes obtained in SIMO mode and in MISO mode by calculating a DML estimation function q for each combination of detection angles θ1, θ2 2 (0i,02) is calculated. A sum spectrum 54 is then formed from these spectra, and for each target a global maximum in this sum spectrum is sought

[0050] As an example, Fig. 5 shows the DML estimator q 2 (0i,02) for a SIMO mode 48 obtained spectrum 56 The function values ​​are represented by contour lines for equal values ​​of q 2 Local maxima of the function are shown hatched.

[0051] The initial scenario (ground truth) is given by θ1 = -10°, θ2 = 10°, and an equal amplitude of the four paths. The distances of the four receive channels are [0, 1 , 5, 3, 4.5] , and the distances of the three transmit channels are [0, 1 , 2] X.

[0052] For the same scenario, Fig. 6 shows the DML estimator q 2 (0i,02) for the spectrum 58 obtained in MISO mode 50.

[0053] Fig.7 shows the estimation function for the corresponding sum spectrum 52.

[0054] It is clearly visible that the SIMO spectrum 48 is ambiguous, with several maxima having the same height as the global maximum. The MISO spectrum 50, on the other hand, allows angle determination only with very low precision, since the line of the global maximum is very broad due to the very small MISO aperture. In contrast, the sum spectrum 52 allows a clear and precise angle determination.

[0055] For comparison, Fig. 8 shows a spectrum 60 based on the complete cross-path model for MIMO, which requires significantly more effort to calculate. It can be seen that the informative value of the much easier-to-calculate spectrum 52 is very close to that of the complete model (spectrum 60).

Claims

Claims 1 . Method for angle estimation based on signals transmitted and received after reflection from an object (24) from a radar sensor (8) which has an angle resolution in at least one dimension and which has an Ml MO-capable antenna array (34, 36), wherein a cross-path model is used to estimate the location angle of a radar target, which cross-path model also models reflections of transmitted and / or received signals from a reflecting surface (42), characterized in that the angle estimation is carried out on the basis of a reduced cross-path model that combines a model for an S1 MO mode (48) of the radar sensor (8) with a model for an MISO mode (50).

2. Method according to claim 1, in which separate DML estimation functions are calculated for a spectrum obtained in SIMO mode (48) and a spectrum obtained in MISO mode (50), and then a sum spectrum (52) is formed by arithmetic averaging of the two DML estimation functions and a global maximum of the sum spectrum (52) is sought.

3. Radar sensor (8) having a transmitting and receiving device with a MIMO-capable antenna array (34, 36) and a digital evaluation stage (32), characterized in that a method according to claim 1 or 2 is implemented in the evaluation stage (32).

4. Radar sensor according to claim 3, comprising a transmitting array (34) and a receiving array (36), wherein a first of the transmitting and receiving arrays (34, 36) is designed for an unambiguous angle measurement and the other array is designed for an ambiguous angle measurement and has a larger aperture than the first.

5. Radar sensor according to claim 4, wherein all transmitting antennas (14) of the transmitting array (34) are arranged at equal intervals in at least one dimension.

6. Radar sensor according to claim 4 or 5, wherein all receiving antennas (26) of the receiving array (36) are arranged at equal intervals in at least one dimension.

7. Radar sensor according to one of claims 3 to 6, wherein at least one of the transmitting and receiving arrays (34, 36) is a two-dimensional array.

Images