Method for compensating for transmitter-side and receiver-side interference effects

Abstract

The present invention relates to a method for compensating for interference effects on the transmitter and receiver sides in a radar sensor, comprising the steps of: setting (20, 30, 40) a desired signal power for a measured radar signal (x(n)), transmitting (23) a first radar signal (s1(n)) by a transmitter (1), receiving (24) the first radar signal (r1(n)) by a receiver (2) of the radar sensor, estimating (26) coefficients (w(n)) of an algorithm for correcting interference effects on the receiver side, transmitting (33) a second radar signal (s2(n)) by the transmitter (1), and receiving (34) the second radar signal (r2(n)) by the receiver (2) of the radar sensor, using the calculated coefficients (w(n)) for compensation, and calculating an inverse model for predistortion of the transmitter (1) using the calculated coefficients (w(n)) for the interference effects on the receiver side. Step (39) of training TIFF2026505462000025.tif1725 and the inverse model A step (41) of predistorting the transmitted measurement radar signal (s(n)) using TIFF2026505462000026.tif1725 and a step (46) of correcting the received measurement radar signal (r(n)) using an algorithm for correcting interference effects on the receiving side are performed.

Description

[Technical Field]

[0001] The present invention relates to a method for compensating for transmitter-side and receiver-side interference effects in a radar sensor. Furthermore, the present invention relates to a computer program, a machine-readable storage medium, an electronic control device and a radar sensor. [Background technology]

[0002] Today, radar sensors employ modulation types such as frequency-modulated continuous-wave radar (FMCW) or pulsed radar, as well as digital modulation types such as orthogonal frequency-division multiplexing (OFDM), pseudorandom radar (PN), and phase-modulated continuous-wave radar (PMCW). This allows for substantially greater flexibility during operation. For accurate evaluation, it is important that radar signals are transmitted and received without distortion. This is not always the case due to nonlinearities in components such as digital-to-analog converters, IQ modulators, mixers, and amplifiers. This is particularly true when the crest factor of the transmitted signal is large.

[0003] To counter the nonlinearity of analog components, the signal power is typically reduced on the transmitter side so that the peak value of the time signal has a defined reference relative to the 1 dB compression point of the transmitter amplifier. To operate the transmitter as linearly as possible, the transmit power is reduced by input backoff (IBO). For example, the transmit power is reduced by 10 dB with input backoff. Therefore, such compensation requires more hardware costs than, for example, FMCW, which can utilize the full power spectrum of the amplifier. As a result, traditionally, digitally modulated radar sensors have had to compromise between strong transmit power or long range and good signal or measurement quality.

[0004] Furthermore, in communications technology, methods for compensating for nonlinearities are known, both on the transmitting and receiving side: in the transmitter, a predistortion is applied to the transmitted radar signal, and in the receiver, an equalization of the received radar signal is performed.

[0005] The paper by L. Anttila, M. Valkama, and M. Renfors, "Circularity-Based I / Q Imbalance Compensation in Wideband Direct-Conversion Receivers," published in IEEE Transactions on Vehicular Technology, vol. 57, no. 4, pp. 2099-2113, July 2008, doi:10.1109 / TVT.2007.909269, describes an IQ demodulator algorithm that compensates for interference effects on the receiver side. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Anttila, Lauri, Mikko Valkama, and Markku Renfors. "Circularity-based I / Q imbalance compensation in wideband direct-conversion receivers." IEEE Transactions on Vehicular Technology 57.4 (2008): 2099-2113. [Non-patent document 2] Michev, Rossen, et al. "Adaptive compensation of hardware behavioral impairments in digitally modulated radars using ML-based models." IEEE Transactions on Microwave Theory and Techniques 72.1 (2023): 421-434. Summary of the Invention

[0007] Disclosure of the Invention In the method of the present application, a desired signal power of the radar signal is set. This signal power is preferably set with a fixed input backoff (IBO). Preferably, the input backoff can be reduced to set the desired signal power. In the method of the present application, component nonlinearities are no longer compensated for by the input backoff. In this case, the input backoff can be reduced, for example, to 0 dB or less, so that, for example, the transmitter can operate at maximum efficiency. In particular, the input backoff is implemented by a digital-to-analog converter for the radar sensor transmitter, which scales the amplitude of the digital signal value. Alternatively, the input backoff can be implemented by analog and / or digital circuits within or for the radar sensor transmitter. In this case, the signal power can be freely selected and is not limited by component nonlinearities, as is common in conventional methods.

[0008] The transmitter of the radar sensor transmits a first radar signal in a radar channel, and the receiver of the radar sensor receives the first radar signal. This first radar signal serves to calibrate the receiver's IQ demodulator. The first radar signal can be transmitted with the signal power described above and corresponds in particular to the second radar signal described below. In general, the first radar signal can be a test signal, which is independent of other radar signals and is transmitted with any amplitude and on any signal line. Therefore, the desired signal power can also be set only later.

[0009] The received first radar signal is downmixed to baseband or an intermediate frequency. Preferably, an IQ mixer is used for this purpose, although other simple mixers can also be used. The received first radar signal can then be preprocessed (calibration preprocessing), which consists in particular of time synchronization and, optionally, frequency synchronization. If necessary, filtering and / or averaging can also be performed. If orthogonal frequency division multiplexing (OFDM) is used, the cyclic prefix can also be removed. Thus, all known effects not due to the IQ mixer (IQ demodulator) are eliminated.

[0010] For the first radar signal, an algorithm for compensating for interference effects on the receiver side of the receiver is executed, and the coefficients of this algorithm are estimated. In particular, this algorithm can compensate for interference effects in an IQ demodulator, which is called IQ compensation or IQC for short. For this purpose, an algorithm with blind coefficient estimation (e.g., the algorithm described in the aforementioned paper by L. Anttila, M. Valkama, and M. Renfors, "Circularity-Based I / Q Imbalance Compensation in Wideband Direct-Conversion Receivers") can be preferably used. However, alternative algorithms can also be used. The calculated coefficients can also be used to equalize the radar signal on the receiver side. Additionally or alternatively, an algorithm for compensating for interference effects in a receiver amplifier can also be used.

[0011] The transmitter then transmits a second radar signal having the desired signal power as described above into the radar channel, and the receiver receives the second radar signal. The coefficients of the algorithm described above are applied when processing the received radar signal. Again, the received second radar signal is downmixed to baseband or an intermediate frequency, and optionally, the received second radar signal is preprocessed. Again, effects not belonging to the analog component, such as channel effects (e.g., time required, frequency shift due to Doppler effect, etc.), are compensated for.

[0012] Here, an inverse model for predistortion of the transmitted radar signal is learned or trained using the second radar signal and coefficients calculated for an algorithm that corrects for interference effects at the receiver. Predistortion of the radar signal (Digital Pre-Distortion, DPD) corrects distortions of the radar signal caused by interference effects at the transmitter. This inverse model can be, for example, a memory polynomial model, a parallel Hammerstein model, or other known nonlinear models, or can be implemented using a neural network. This model represents the inverse function (post-inverse function) of the transmitter. This inverse model is trained using the transmitted data and the received data corrected by the algorithm. Training this inverse model to predistort a subsequently transmitted radar signal using a previously transmitted radar signal is also called an "indirect learning architecture."

[0013] To derive the coefficients and inverse model of the algorithm as accurately as possible, the above-described steps can be repeated several times. This is advantageous, especially in the case of arbitrary transmission data, as it allows the coefficients to be estimated as accurately as possible. Furthermore, the above-described steps can be repeated under different conditions, thereby incorporating, for example, temperature dependence, aging, and environmental effects. As a result, the coefficients and inverse model of the algorithm can be obtained for different conditions. Furthermore, the above-described steps can be performed for various operating points, such as various bias voltages, various power and frequency ranges of the linear oscillator or amplifier, and various frequency ranges of the radar signal.

[0014] In summary, the correction of the receiver interference effect implemented by this algorithm is integrated into the predistortion feedback path. The correction of the transmitter interference effect is taken into account in the inverse model. As a result, the receiver interference effect is separated from the transmitter interference effect of the radar sensor, which is not calibrated itself.

[0015] In the measurement mode, the transmitter transmits a transmitted measurement radar signal having the above-mentioned signal power, the transmitted measurement radar signal is predistorted by the trained inverse model to compensate for the interference effects of the transmitter, and the measurement radar signal is received by the receiver and corrected using an algorithm for correcting the interference effects at the receiver, thereby compensating for the interference effects at the receiver.

[0016] As a result, interference effects on the receiver and transmitter are compensated for separately during measurements. This enables independent adaptive calibration of the radar sensor without additional components. Distributing signal processing between the transmit and receive paths offers several advantages: The amplifier power in the transmitter is no longer limiting. As a result, the available amplifier power in the transmitter can be better utilized. Furthermore, ghost targets that may be caused by interference effects are suppressed. Furthermore, radar measurements achieve an improved signal-to-noise ratio for the same transmit power. Alternatively, increased transmit power for the same amount of distortion can be achieved, which also improves the signal-to-noise ratio. This extends the radar sensor's range without compromising measurement quality and / or measurement dynamics. Furthermore, out-of-band transmissions can be reduced, thereby maintaining the spectral mask. Special radar channels or special scenarios are not required. Furthermore, frequency shifts between the local oscillators in the transmitter and receiver can be eliminated.

[0017] Preferably, a radar sensor receiver can be provided as the receiver. In this case, by utilizing the radar channel after reflection, the radar signal can be received by the receiver or obtained directly from crosstalk between the transmitter and receiver. When a radar sensor receiver is used, the components can be realized entirely within the radar sensor, and no additional external components are required. This allows compensation to be performed even during radar operation. Predistortion and equalization of the radar signal can be performed during an already established measurement cycle or in an additional measurement. Therefore, calibration can be repeated in the field during application, allowing the radar sensor to be adapted to current conditions. Alternatively, this can be performed during the initial characterization of the radar sensor.

[0018] Alternatively, an additional measurement receiver can be used to receive and measure the radar sensor's radar signal. This additional measurement receiver can be an external measurement receiver that can be connected to the radar sensor or to external evaluation equipment. Such an additional measurement receiver is preferably also employed during the initial characterization of the radar sensor. Alternatively, this additional measurement receiver can be an integrated circuit on a chip or circuit board. The measurement receiver can be supplied with decoupling of the transmitted signal at the transmitter output, for example, via a directional coupler, power detector, or mixer. Such an integrated measurement receiver can be used both for the initial characterization, during operation, or between measurements. In particular, this additional measurement receiver is part of a defined measurement structure, resulting in consistent results. This allows for fast compensation of multiple radar sensors via the same measurement receiver. The IQC coefficients are radar receiver-dependent and cannot be determined directly using the additional receiver. Measurements using the additional measurement receiver should serve to determine the transmitter's interference effects and thus the inverse model. In this mode, the calibration sequence must be changed. First, the coefficients of the inverse model for the predistortion must be estimated using the (preprocessed) received signal from the additional receiver. This inverse model can then be used to transmit the predistorted signal to determine the IQC coefficients from the received signal of the original radar receiver. If the inverse model is correctly estimated, the most recent received signal will be free of the interference effects of the transmitter.

[0019] Preferably, this inverse model can be derived using a probabilistic estimation method, such as least squares. Alternatively, the inverse model can be trained using machine learning with a neural network. In this case, distorted signals are used as inputs and ideal input signals are used as outputs (labels). Such neural networks are described in a yet-to-be-published paper by R. Michev, Y. Shu, D. Werbunat, J. Hasch, and C. Waldschmidt, "Adaptive Compensation of Hardware Impairments in Digitally Modulated Radars Using ML-Based Behavioral Models."

[0020] If the scene is known, e.g., ground truth data is provided for a particular scenario, if separate channel estimation has been performed (especially by other calibration procedures), or if the scene is derived by other sensors, e.g., cameras and lidars, then such available channel information can be used to compensate for interference effects. In particular, the channel information can improve the separation of interference effects at the transmitter and receiver. In particular, the channel information can be used during channel estimation in the inverse model. During calibration, residual signal components from targets in the channel can distort the coefficient calculation depending on the backscattering cross section and / or signal power of the interference. Therefore, more accurate channel estimation leads to better compensation.

[0021] Optionally, the method can be applied to a multiple-input multiple-output (MIMO) radar sensor by performing the above steps for each receiver and each transmitter to obtain coefficients for the IQC and the inverse model.

[0022] The computer program is provided to execute the steps of the method, in particular on a computing device or control device, making it possible to implement the method in conventional electronic control devices without modifying the structure of the device, and for this purpose the computer program is stored on a machine-readable storage medium.

[0023] By running this computer program on a conventional electronic control device, the electronic control device is provided with the equipment to perform receive side equalization of radar signals.

[0024] Furthermore, a radar sensor is provided that includes a transmitter and a receiver. The radar sensor further includes a pre-processing unit and a compensation unit. In particular, the radar sensor can include the above-mentioned electronic control device. In this case, the pre-processing unit and the compensation unit can be part of the electronic control device. The radar sensor is configured to perform the steps of the method in order to compensate for interference effects occurring during measurements.

[0025] Embodiments of the invention are illustrated in the drawings and are explained in more detail in the following description. [Brief explanation of the drawings]

[0026] [Figure 1] FIG. 1 is a block diagram of a radar sensor. [Figure 2] 1 is a flowchart of one embodiment of a method for a first portion of a calibration cycle according to the present invention. [Figure 3] 4 is a flowchart of one embodiment of a method for a second portion of a calibration cycle according to the present invention. [Figure 4] 3 is a flow chart of one embodiment of a method during a measurement cycle according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0027] FIG. 1 shows a radar sensor according to the present invention, which includes a transmitter 1 and a receiver 2. The transmitter 3 generates a radar signal x(n), which is output from the transmitter 1. After passing through a digital-to-analog converter (DAC) 4, an IQ modulator 5, and a power amplifier 6, the radar signal x(n) is output by a transmitting antenna 7 as a transmit signal s(n) into a radar channel H. The receiver 2 includes a receiving antenna 8, by which the radar signal is received from the radar channel H as a receive signal r(n). Furthermore, crosstalk CT may occur between the transmitter 1 and the receiver 2. The received radar signal r(n) passes through a low-noise amplifier 9, an IQ demodulator 10, and an analog-to-digital converter (ADC) 12. The IQ demodulator demodulates the received radar signal r(n) into an output signal y(n). The IQ modulator 5 and the IQ demodulator 10 are connected to each other via a local oscillator 11 for coordinated modulation and demodulation of the radar signal. The components of the digital-to-analog converter 4, the IQ modulator 5, the power amplifier 6, the low-noise amplifier 9, the IQ demodulator 10 and the analog-to-digital converter 12 typically have non-linearities. The output signal y(n) of the IQ demodulator 10 is fed to a signal evaluation unit 13 where it is evaluated.

[0028] 2 and 3 are flowcharts of the method according to the invention, divided into two parts during a calibration cycle. In the first part of FIG. 2, a digital transmit signal is first generated (19). Then, in the transmitter 3, a predefined test signal x1(n) is applied to the desired signal power for calibration via the input backoff (20). This test signal x1(n) is applied to the inverse model of the most recent calibration cycle.

number

[0029] The modulated test signal z1(n) passes through a power amplifier 6 and is then transmitted (23) from transmitter 1 into radar channel H as a first transmit signal s1(n). The effect of the power amplifier 6 is described by a memory polynomial model. Thus, the first transmit signal s1(n) has an equivalent baseband representation as follows:

number

[0030] Receiver 2 receives (24) a first radar signal transmitted from radar channel H. In its equivalent baseband representation, this received first radar signal r1(n) is the convolution of the first transmitted signal s1(n) with the channel impulse response h(n). r1(n)=s1(n)*h(n)

[0031] The received first radar signal r1(n) is downmixed (25) to baseband or an intermediate frequency by an IQ demodulator 10. This results in an output signal y1(n) of the demodulator 10 for the first radar signal, which can be expressed as a sum of convolutions with an impulse response. y1(n)=g 1,Rx (n)*r1(n)+g 2,Rx (n)*r1 * (n) In the formula, g 1,Rx (n) and g 2,Rx (n) is the coefficient of the impulse response of the main path and mirror path of the IQ modulator 5, and r1 * (n) is the complex conjugate of r1(n), and * denotes convolution. The mirroring is compensated for using a well-known algorithm for compensating for interference effects on the receiver side, namely "Circularity-Based I / Q Imbalance Compensation in Wideband Direct-Conversion Receivers" with blind coefficient estimation, as mentioned at the beginning. This algorithm is based on linear ("widely linear") filtering. y 1,comp (n)=y1(n)+w(n)*y1 * (n) In the formula, y 1,comp (n) is the compensated first signal, w(n) is the coefficient of this algorithm, and y1 * (n) is the complex conjugate of y1(n), and * denotes convolution. This operation compensates for the effects of the demodulator 10 on the output signal y1(n), and as a result, this operation is also called IQ compensation or IQC (IQ Compensation). Here, the coefficients w(n), i.e., the filter coefficients, of this algorithm are calculated from the statistical characteristics of the output signal y1(n) (26). Ideally, these coefficients w(n) are calculated based on the (radar) receiver coefficients g 1,Rx (n) and g 2,Rx Depends only on (n).

[0032] Following the first part, the second part of the calibration cycle in FIG. 3 is performed, in which a digital transmit signal is first generated (29), followed by setting in transmitter 3 via input backoff a desired signal power for the measurement for a predefined second signal x2(n) (30). This second signal x2(n) corresponds to the measurement radar signal x(n) used during the measurement and can be generated, for example, using orthogonal frequency division multiplexing (OFDM). However, other modalities can also be used, such as phase-modulated continuous wave radar (PMCW), pseudo-noise (PN), etc. This second signal x2(n) is then subjected to an inverse model of the most recent calibration cycle.

number

[0033] The modulated second signal z2(n) passes through a power amplifier 6 and is then transmitted (33) from the transmitter 1 into the radar channel H as a second transmit signal s2(n). The effect of the power amplifier 6 is described by a memory polynomial model. Thus, the equivalent baseband representation of the second transmit signal s2(n) is:

number

[0034] Receiver 2 receives (34) a second radar signal transmitted from radar channel H. In its equivalent baseband representation, this received second radar signal r2(n) is the convolution of the second transmitted signal s2(n) with the channel impulse response h(n). r2(n)=s2(n)*h(n)

[0035] The received second radar signal r2(n) is downmixed to baseband or an intermediate frequency and demodulated (35) by an IQ demodulator 10. Time synchronization, e.g., correlation and shifting, as well as amplitude normalization are performed, e.g., during crosstalk CT or in the case of a direct path in the case of an integrated receiver 2. This results in an output signal y2(n) of the demodulator 10 for the second radar signal, which can be expressed as a sum of convolutions with an impulse response: y2(n)=g 1,Rx (n)*r2(n)+g 2,Rx (n)*r2 * (n) In the formula, g 1,Rx (n) and g 2,Rx (n) is the coefficient of the impulse response of the main path and mirror path of the IQ modulator 5, and r2 * (n) is the complex conjugate of r2(n), and * denotes convolution.

[0036] To this downmixed output signal y2(n), the IQC coefficients w(n) calculated in the first part are now applied (36), from which a compensated second signal y 2,comp (n) is generated, which has the following format: y 2,comp (n)=y2(n)+w(n)*y2 * (n)

[0037] The channel effects or channel impulse response h(n) are estimated (37) for radar channel H. For example, in the case of a crosstalk CT or a direct path, such as in an integrated receiver 2, time synchronization, e.g., correlation and shift, and amplitude normalization are performed. In further embodiments, a predefined radar channel H can be used, or other sensors can be used to derive ground truth data. Following this, radar channel H is compensated (38) for channel effects, thereby equalizing it.

[0038] The IQC coefficients w(n) and the second radar signal y compensated using the IQC 2,comp (n) and the equalized channel H are the inverse models for the current calibration cycle.

number

number

[0039] For this purpose, a mathematical model, for example, a memory polynomial model, a parallel Hammerstein model, or other nonlinear model, can be used, and this mathematical model can be used to estimate the second signal.

number

number

[0040] Model

number

number

[0041] Or the inverse model

number

[0042] Inverse model

number

number

number

[0043] As will be discussed later, the inverse model

number

[0044] The calibration cycle can be repeated any number of times to account for different temperature conditions and aging effects.

[0045] 4 shows a flow chart of the method according to the invention during a measurement cycle. First, a digital transmit signal is generated (39). Then, in the transmitter 3, a desired signal power for the measurement is set (40) via an input backoff for a predefined measurement radar signal x(n). This measurement signal x(n) corresponds to the second signal x2(n) described above and can likewise be generated, for example, using orthogonal frequency division multiplexing (OFDM) or other methods, such as phase-modulated continuous wave radar (PMCW), pseudo-noise (PN). This measurement signal x(n) is then subjected to the above-mentioned calculated inverse model of the current calibration cycle.

number

[0046] The modulated measurement signal z(n) passes through a power amplifier 6 before being transmitted (43) by the transmitter 1 into the radar channel H as a measurement transmit signal s(n). The effect of this power amplifier 6 is described by a memory polynomial model, which gives the equivalent baseband representation of the measurement transmit signal s(n):

number

[0047] Receiver 2 receives (44) the measurement radar signal transmitted from radar channel H. In its equivalent baseband representation, this received measurement radar signal r(n) is the convolution of the measurement transmitted signal s(n) and the channel impulse response h(n). r(n)=s(n)*h(n)

[0048] The received measurement radar signal r(n) is downmixed to baseband or intermediate frequency and demodulated (45) by the IQ demodulator 10. In the case of a crosstalk CT or direct path in the integrated receiver 2, time synchronization, e.g., correlation and shift, as well as amplitude normalization, are performed. This results in the output signal y(n) of the demodulator 10 for the measurement radar signal, which can be expressed as a sum of convolutions with impulse responses: y(n)=g 1,Rx (n)*r(n)+g 2,Rx (n)*r * (n) In the formula, g 1,Rx (n) and g 2,Rx (n) are the coefficients of the impulse response of the main path and mirror path of the IQ modulator 5, and r * (n) is the complex conjugate of r(n), and * denotes convolution.

[0049] Now, the IQC coefficients w(n) calculated in the first part of the calibration cycle are applied (46) to this down-converted output signal y(n), from which the compensated measured radar signal ycomp (n) is generated, which has the following form: y comp (n)=y(n)+w(n)*y * (n)

[0050] Algorithm coefficient w opt If (n) is optimally estimated depending only on the receiver parameters, then the optimally compensated signal y comp.opt (n) is obtained. y comp.opt (n)=[g 1,Rx (n)+w opt (n)*g * 2,Rx (n)]*r(n)

[0051] and the compensated measured radar signal y comp (n) is fed (47) to a signal evaluation unit 47 to derive a radar image.

[0052] For radar signals using orthogonal frequency multiplexing, the radar channel H in the frequency domain of interest is comp (n) the received code C obtained from Rx (n, μ) is the ideal measurement signal x pd (n) is the ideal transmitted code C obtained from Tx It is calculated by dividing by (n, μ).

number

[0053] A radar image can then be derived by standard radar signal processing.

Claims

1. 1. A method for compensating for transmitter-side and receiver-side interference effects in a radar sensor, comprising: The transmitter (1) transmits a first radar signal (s １ (n)) by a receiver (2) of the radar sensor; １ (n)) (24); a step (26) of estimating coefficients (w(n)) of an algorithm for correcting the interference effects on the receiving side; The transmitter (1) transmits a second radar signal (s ２ (n)) by the receiver (2) of the radar sensor; ２ (34) receiving the calculated coefficients (w(n)), wherein the calculated coefficients (w(n)) are used for the compensation; Using the calculated coefficients (w(n)) for the interference effect at the receiver, Inverse model for predistortion of the transmitter (1) [Equation 1] (39) training the The inverse model [Equation 2] and correcting (46) the received measurement radar signal (r(n)) using the algorithm for correcting the receiver interference effects. A method characterized by:

2. The first radar signal (r １ 2. The method of claim 1, wherein said step (24) of receiving (n)) is performed by an additional measurement receiver.

3. The inverse model [Equation 3] 3. The method according to claim 1, wherein the order of calculation of the coefficients (w(n)) of the receiver side is changed.

4. 2. The method according to claim 1, characterized in that the algorithm for correcting interference effects on the receiver side is an algorithm for blind coefficient estimation for correcting interference effects of an IQ modulator (10) of the receiver (1).

5. The inverse model [Equation 4] 3. The method according to claim 1, further comprising a step (39) of learning the eigenvalues ​​of ...

6. The inverse model [Equation 5] 3. The method according to claim 1, further comprising a step (39) of learning the following using machine learning:

7. Method according to any one of claims 1 to 6, characterized in that channel information available in a known scene is used to compensate for said interference effects.

8. The first radar signal (s １ (n)) and the step (23) of transmitting the first radar signal (r １ (n)) of the second radar signal (s), (24) and (26) estimating coefficients (w(n)) of the algorithm for correcting interference effects on the receiver side, ２ (n)) and the step (33) of transmitting the second radar signal (r ２ 8. The method according to claim 1, wherein the steps of receiving (n)) and training (39) an inverse model are performed in multiple iterations.

9. A computer program arranged to carry out the steps of the method according to any one of claims 1 to 8.

10. A machine-readable storage medium storing the computer program according to claim 9.

11. An electronic control device equipped to equalize a radar signal using a method according to any one of claims 1 to 8.

12. 9. A radar sensor comprising a transmitter (1) and a receiver (2) and a signal processing unit (13), said radar sensor being equipped to compensate for interference effects using the method according to any one of claims 1 to 8.

Images