Method for self-calibration of a radar system
By performing a self-calibration method in the radar system, the problem of phase error changes in the radar system is solved through two-dimensional linear regression and compensation of phase correction values, and the accuracy and separability of angle estimation are improved.
Patent Information
- Application Number
- CN202380063122.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-09-02
- Filing Date
- 2023-06-27
- Publication Date
- 2025-04-11
AI Technical Summary
In vehicle automation, existing radar systems have caused phase error changes due to uneven temperature distribution and component aging, which affects the accuracy and separability of angle estimation.
A self-calibration method is proposed, by measuring multiple targets, processing distance information and Doppler information, performing two-dimensional linear regression to estimate the regression plane, calculating intra- and inter-group phase correction values, and compensating control vectors to achieve calibration of phase errors.
The phase error of the radar system is effectively calibrated, the dynamic range, accuracy and separability of angle estimation are improved, and the influence of the side beam and the broadening of the main beam are reduced.
Smart Images
Figure CN119816751A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for self-calibrating a radar system, the radar system having at least two antenna groups, the antenna groups being assigned at least one transmission channel and one reception channel. Background Art
[0002] Radar systems are used to measure the distance, relative speed, and azimuth and elevation of objects. Typically, an antenna group (also called an antenna array) is used for angle estimation, which can be used not only for transmission but also for reception. Through digital beamforming (English: digital beamforming), the incident angle of the reflected plane wave generated by the target in the far field is detected. In order to perform beamforming, the phase difference of the reflected wave is analyzed across multiple receiving channels. The received signals of the antenna group can be processed by transformation as if they were measured by a virtual receiver. Conventionally, this analysis is implemented by previously stored control vectors, which are used in different methods. For example, in the Bartlett-Strahlformer, the expected differential phases for different angles of incidence are stored in the control vector, and then these expected differential phases are associated. Alternatively, a model-based estimation can be performed, which is also based on the control vector.
[0003] Conventionally, for each individual sensor, in a one-time end-of-line calibration, a control vector (one or more angular profiles) is measured by angle and then stored in a non-volatile memory. Phase errors (phase shifts) are thus taken into account, which can be caused by various effects, such as production tolerances of antennas or input lines or interactions of electromagnetic waves with covers, housings or circuit boards.
[0004] In addition, the antenna group can divide the environment into angle units and detect these angle units separately from each other. As a result, the following objects can be separated from each other by their angular position: the distance and radial speed of the objects are the same. In this case, each radar beam of the radar system is a spatial filter.
[0005] For specific functions in vehicle automation, better angle measurement is required, not only in terms of accuracy, but also in terms of separability. To achieve this, antenna groups with large apertures are used. In order to control the corresponding transceivers, these are typically grouped and cascaded, thereby creating a transceiver hierarchy. Preferably, in order to perform a coherent measurement of the angle of incidence of the reflected wave, all transceivers are connected to a reference oscillator. Typically, a high-frequency line is used for this purpose, which distributes the signal of the reference oscillator to the transceivers. However, the larger the aperture, the longer the high-frequency line.
[0006] Nowadays, in the automotive sector, patch antennas are often used as radar antennas. The patch antenna is applied to the surface of a circuit board and is supplied with power by means of a strip line. As a result, the size of the circuit board increases in proportion to the size of the aperture. Therefore, in the case of large circuit boards, the probability of an uneven temperature distribution on the circuit board also increases, which is caused by internal and / or external influencing factors. As an example of an internal influencing factor, a strong heating of different component groups should be mentioned. Examples of external influencing factors are partial shading of the circuit board and heating due to adjacent structural groups and / or due to air flow.
[0007] Uneven temperature distribution, aging of components and other factors affect the measurement of the radar system in that they change the previously calibrated phase error. If the angle estimation is still performed with the help of the control vector stored at the beginning, the working capacity of the angle estimation is reduced. In particular, in the angle spectrum, the side beams increase, while the main beam is weakened and widened and the position of the main beam changes (English: beam-pointing error). As a result, the dynamic range, accuracy and separability of the angle estimation are reduced.
[0008] A self-calibration of a radar system is known from the article "Error analysis and self-calibration of a digital beamforming radar system" by M. Harter et al. (2015 IEEE MTT-S International Conference on Microwaves for Intelligent Mobility (ICMIM), 2015, pp. 1-4). It is assumed that the wavefront of the reflected wave on the antenna is flat. This assumption is only fully met when the target is in the far field. Typically, the limit for the far field is defined by the aperture length D and the wavelength λ using formula 1:
[0009]
[0010] Under the assumption that the target is located in the far field, a one-dimensional linear regression is performed from the received phases at the antenna elements of the receiving antenna group and the most likely course of the wavefront is thus estimated. The phase error to be corrected is then calculated from this. The phase error estimated from the target is applicable to the corresponding target angle without limiting the generality. It is irrelevant here whether the radar system has a hood or is arranged, for example, behind a bumper.
[0011] A further prerequisite for the self-calibration according to the above article is that the antenna elements are arranged equidistantly along one dimension. This arrangement includes the following: for calibration of the azimuth control vector, the antenna elements are arranged horizontally, and for calibration of the elevation control vector, the antenna elements are arranged vertically.
[0012] Generally speaking, self-calibration can be used for different modulation methods. Currently, typical transmission frequencies are 24 GHz or 77 GHz, and the maximum bandwidth that can be occupied is less than 4 GHz, in particular about 0.5 GHz.
[0013] Typically, current radar systems in the vehicle sector use FMCW modulation (frequency modulated continuous wave radar) with fast ramps (fast linear frequency modulation), in which multiple linear frequency ramps of the same slope are emitted one after the other. The mixing of the instantaneous transmitted signal with the received signal produces a low-frequency signal whose frequency (called the beat frequency) is proportional to the spacing. Usually, the system is designed in such a way that the component of the beat frequency caused by the Doppler frequency can be ignored. The spacing information obtained from the beat frequency is largely unambiguous and unique, and the Doppler shift can then be determined by observing the time development of the phase of the complex spacing signal across the ramp. The spacing determination and the speed determination usually occur independently of each other by means of a two-dimensional Fourier transform. The above-mentioned angle estimation is located downstream of the spacing estimation and the speed estimation. Summary of the invention
[0014] A method for self-calibrating a radar system is proposed, which has at least two antenna groups, to which at least one transmit channel and a receive channel are assigned. The geometry of the antenna groups, i.e. the positions of the transmit and receive antennas, is known. In addition, the transceiver hierarchy of the transceiver is known, which serves to form antenna groups, whose elements experience the same type of phase errors. These data are part of the specifications of the radar system.
[0015] Initially, a radar system measures a number of targets. To this end, a transmitting antenna group emits electromagnetic waves in the direction of the target. The electromagnetic waves reflected by the target are then received by a receiving antenna group and evaluated using digital beam forming. The measurement is performed in an angularly resolved manner, wherein the environment is divided into angular cells by the antenna group.
[0016] For each antenna group, the distance information and Doppler information are processed (also referred to as range-Doppler processing). Preferably, the processing is implemented by a two-dimensional fast Fourier transform, but it can also be implemented by other algorithms known per se. Next, detection of the target is performed in angle units. To this end, preferably, the following steps are performed: for each angle unit, the energy is calculated. In addition, a threshold value for the estimated noise energy for such an angle unit is estimated. The calculated energy is compared with the threshold value for the noise energy, and the calculated energy is thus discriminated relative to the estimated noise energy in each unit. If multiple adjacent angle units are greater than the threshold value for the estimated noise energy, the local maximum of the measured complex data is determined for each combination of transmitter and receiver. The reflection list is obtained by the above steps, in which the distance, relative speed and complex amplitude measured on the corresponding virtual channel are stored for each target.
[0017] For the respective channels, the amplitude differences are first compensated. Preferably, the following steps are performed for this purpose. The average received power of all signal amplitudes is calculated, wherein in particular formula 2 is used:
[0018]
[0019] is the average value of the amplitude, are the measured complex amplitudes. The summation is performed by transmitters m=1...M and receivers n=1...N.
[0020] Next, the deviation of the amplitude from the average received power is calculated for each channel. In particular, the deviation is calculated by the quotient according to Formula 3:
[0021]
[0022] This deviation is then used to compensate for the amplitude error for each channel. This step allows the receive function to be calibrated.
[0023] In the following, an intra-group calibration is performed for each antenna group. To this end, a two-dimensional linear regression is performed for each antenna group. In this case, a regression plane is estimated for which the mean square distance of the phase measurement values of the channels of the antenna group becomes minimum. In the case of a two-dimensional regression, the geometric data of the antenna group are used, which, as indicated at the beginning, are known as part of the specifications of the radar system. Now, the difference between the measured phase value and the regression plane is calculated for each channel. This results in an intra-group phase correction value for the corresponding antenna group. Now, this step is repeated for all antenna groups so that there is an intra-group phase correction value for all (virtual) channels.
[0024] If the phase error is not angle-dependent - for example when the phase field is due to aging of the radar system - the measured values of all targets can be averaged for each channel. This achieves an averaging gain. If the phase error is angle-dependent - for example because the radar system is arranged behind the bumper - each angle can be calibrated using its own target. Preferably, the intermediate angles are interpolated for which there is no own target. In this case, the angles for which there is a target should be close enough to each other that the sampling theorem is observed for this phase error. By means of a plurality of measurement cycles running in succession, targets can be detected in angle units that were previously only interpolated or for which no phase error estimation could be achieved.
[0025] In the following, an intergroup calibration is performed for at least two different antenna groups in order to determine the phase errors between the different antenna groups. As described above, a regression plane has been determined for each antenna group. Now, the spacing between the two regression planes of the two different antenna groups is calculated, wherein the modulo 2π is applied. Thus, intergroup phase correction values for the two antenna groups are obtained. If the regression planes are compensated with the intergroup phase correction values, all regression planes of the antenna groups are superimposed one on top of the other.
[0026] Finally, the control vector of each channel is compensated by means of the intra-group phase correction value and the inter-group phase correction value. Next, the angle estimation can be performed according to known methods.
[0027] Based on two-dimensional regression, this method can generally be applied to all geometries of antenna groups, and these antenna groups are neither restricted in terms of orientation nor in terms of spacing. That is, the antenna groups do not have to be arranged along horizontal lines or along vertical lines, but can be placed freely. In addition, the antenna elements do not have to be positioned equidistant from each other, but can be at any distance relative to each other. In addition, the phase errors for each channel do not have to be distributed in the same way, but different average phase errors may exist for each antenna group. This method can also be used for multiple-input multiple-output (MIMO) radar sensors. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Exemplary embodiments of the invention are shown in the drawings and are explained in more detail in the following description.
[0029] Figure 1 A schematic diagram of an antenna of a radar system is shown in which the method according to the invention is used.
[0030] Figure 2 A flow chart shows one specific embodiment of the method according to the invention. DETAILED DESCRIPTION
[0031] exist Figure 1, an antenna of a radar system RS, not further shown, is shown in which the method according to the invention is used. The antenna has a plurality of antenna elements Rx1, Rx2, Rx3, Rx4, Tx1, Tx2, Tx3 and Tx4, which emit radar signals as transmitting antenna elements Tx1, Tx2, Tx3 and Tx4 and receive reflected radar signals as receiving antenna elements Rx1, Rx2, Rx3 and Rx4. That is, this is a MIMO antenna (Multiple-Input-Multiple-Output). In addition, transceivers T1 and T2 are provided, to which antenna elements Rx1, Rx2, Rx3, Rx4, Tx1, Tx2, Tx3 and Tx4 are connected. The first system on chip SoC1 (also referred to as system-on-chip, SoC) consists of a first transceiver T1 and two transmitting antenna elements Tx1, Tx2 connected thereto and two receiving antenna elements Rx1, Rx2 connected thereto. The second system on chip SoC2 consists of a second transceiver T2 and two transmitting antenna elements Tx3, Tx4 connected thereto and two receiving antenna elements Rx3, Rx4 connected thereto. The transmitting antenna elements Tx1 and Tx2 together with the receiving antenna elements Rx1 and Rx2 of the first system on chip SoC1 form a first antenna group, and the transmitting antenna elements Tx3 and Tx4 together with the receiving antenna elements Rx3 and Rx4 of the second system on chip SoC2 form a second antenna group. The two systems on chip SoC1 and SoC2 are symmetrically constructed and designed identically. The geometry G of the antenna group is known and stored through the specifications of the radar system RS (see Figure 2 ). Furthermore, the transceiver hierarchy of the transceivers T1 and T2 of the radar system RS is also known.
[0032] Figure 2 A flow chart of an embodiment of the method according to the present invention is shown. Initially, the radar system RS performs measurement 10 of multiple targets in the environment by means of antenna elements Rx1, Rx2, Rx3, Rx4, Tx1, Tx2, Tx3, and Tx4. The measurement 10 is performed in an angle-resolved manner, wherein the environment is divided into angle units by antenna groups. Next, processing of distance information and Doppler information 11 is performed by means of a fast Fourier transform. Furthermore, detection of the target 12 is performed within the angle unit. For each angle unit, the energy is determined, and in addition a threshold value for noise energy is determined. In each unit, the energy determined is judged relative to a threshold value for the estimated noise energy. If multiple adjacent angle units are greater than the threshold value for the estimated noise energy, a local maximum of the measured complex data is determined for each combination of transmitter and receiver.
[0033] For self-calibration, the 20 amplitude difference is compensated for the corresponding channel. To this end, the average received power of all signal amplitudes is calculated according to Formula 2:
[0034]
[0035] is the average value of the amplitude, are the measured complex amplitudes. The summation is performed by transmitters m=1...M and receivers n=1...N.
[0036] Next, the deviation of the amplitude from the average received power is calculated for each channel according to Equation 3:
[0037]
[0038] Then, using this deviation, the amplitude error is compensated 20 for each channel.
[0039] For each antenna group of the system-on-chip SoC1, SoC2, a two-dimensional linear regression 21 is performed. Here, a regression plane is estimated respectively, for which the mean square distance of the phase measurement values of the channels of the antenna group becomes minimum. In the case of two-dimensional regression, the geometric data G of the antenna group are used. Now, the difference between the measured phase value and the regression plane is calculated for each channel. This results in an intra-group phase correction value for the corresponding antenna group. If the phase error is angle-dependent, each angle is calibrated 22 by its own target. Interpolation 23 is performed for the following intermediate angles: for which there is no own target. If the phase error has no angle correlation, the measured values of all targets are averaged for each channel 24. Now, this process is repeated for all antenna groups, so that there is an intra-group phase correction value K for all (virtual) channels. intra .
[0040] Furthermore, the distance between the two calculated regression planes of the two antenna groups of the first system on chip SoC1 and the second system on chip is calculated 25, wherein the modulo 2π is applied. Thus, the intergroup phase correction value K for the two antenna groups is obtained inter .
[0041] Finally, with the help of the intra-group phase correction value K intra and the inter-group phase correction value K inter The control vector of each channel is compensated 26.
[0042] After the self-calibration, further evaluation can be performed from the reflection list L, for example an angle estimation 13 .
Claims
1. A method for self-calibrating a radar system (RS), the radar system having at least two antenna groups, the at least two antenna groups being assigned at least one transmission channel and one reception channel, Features The following steps: - measuring (10) a plurality of targets by means of said radar system (RS); - for each antenna group, processing the range information and Doppler information (11) and detecting (12) the targets in order to obtain for each target a list of reflections (L) with complex amplitudes; - compensating (20) for amplitude differences for the corresponding channels; - estimating a regression plane by means of a two-dimensional linear regression (21), for which the mean square distance of the phase measurement values of the channels of the antenna group becomes minimum; - Calculate the difference between the measured phase value and the regression plane for each channel to obtain the intra-group phase correction value (K intra ); - Calculate (25) the distance between the two regression planes of different antenna groups by means of modulo 2π in order to obtain the inter-group phase correction value (K inter ); - Using the phase correction value within the group (K intra ) and the inter-group phase correction value (K inter ) compensates the control vector of each channel (26).
2. The method according to claim 1, characterized in that In calculating the intra-group phase correction value (K intra ), if there is no angular correlation of the phase error, the measured values of all targets are averaged for each channel (24).
3. The method according to claim 1, characterized in that In calculating the intra-group phase correction value (K intra ), if the phase error is angle dependent, each angle is calibrated with its own target (22).
4. The method according to claim 3, characterized in that Interpolation (23) is performed for intermediate angles for which no own target exists.
5. The method according to any one of the preceding claims, characterized in that The distance information and Doppler information (11) are processed by fast Fourier transform.
6. The method according to any one of the preceding claims, characterized in that The detection of the target is achieved through the following steps (12): - discriminating the energy in each angular cell relative to a threshold for the estimated noise energy; If a plurality of adjacent angular cells is greater than said threshold, a local maximum is determined.
7. The method according to any one of the preceding claims, characterized in that Compensation for the amplitude difference is achieved by the following steps (20): - Calculate the average received power; - calculating the deviation of the amplitude of each channel from the average received power; - compensating said amplitude according to said deviation.