MIMO radar system

Abstract

The invention relates to a MIMO radar system having a transmitting array (36) with a plurality of transmitting antennas (14), a receiving array (38) with a plurality of receiving antennas (26), an antenna spacing in one of the transmitting array and the receiving array being below a Nyquist limit for angle measurements of a single value, while the antenna spacing in their combination is above the Nyquist limit, a control and evaluation device for transmitting a sequence of transmitting signals in each measurement cycle, the measurement cycle being divided into a plurality of measurement blocks, a uniform multiplexing scheme being used within each measurement block, the multiplexing scheme differing from measurement block to measurement block, performing a Doppler estimation and an angle estimation on the basis of the transmitting array (38) on the received signals, performing a Doppler correction on the basis of the Doppler estimation on the received signals followed by a demultiplexing, taking into account the phases of the signals transmitted by the different transmitting antennas, refining the Doppler estimation and the angle estimation on the basis of the signals received in different measurement blocks of the measurement cycle.

Description

Technical Field

[0001] This invention relates to a MIMO radar system, and more particularly to a MIMO radar system for motor vehicles. Background Technology

[0002] Radar systems are increasingly being used in motor vehicles to detect traffic conditions and provide information on the distance, relative speed, and heading angle of located objects—such as vehicles or obstacles—to one or more safety or comfort functions. These functions reduce the driver's workload or completely or partially replace the human driver. MIMO (multiple input-multiple output) systems are increasingly being used, employing multiple transmitting and receiving antennas to achieve high angular resolution.

[0003] WO 2018 / 076005 A1 mentions different types of MIMO radar systems: transmitters and / or receivers can be arranged in different locations. Virtual channels can be generated by using mutually orthogonal codes. Channels can be separated using time division multiplexing (TDMA) or frequency division multiplexing (FDMA).

[0004] A MIMO radar measurement method is known from DE 10 2014 212 284 A1, in which the transmitted signal is frequency-modulated in a ramp-like manner using a modulation mode in which a sequence of ramps is assigned to different transmit switch states and is interleaved in time, the transmit switch states being distinguished by the selection of the antenna element used for transmission. Each transmit switch state is further assigned to multiple sequences, which are interleaved in time. A value for the relative velocity of the radar target is determined based on the peak position in the two-dimensional spectrum of the signal obtained for one sequence, the value being periodic with a predetermined velocity period. The phase relationship of the spectral values ​​in the spectrum of the sequence used for the transmit switch states is compared with the expected phase relationship for the corresponding periodic value of the relative velocity, and an estimated value of the relative velocity is selected based on the comparison result.

[0005] US 2017 / 0160380 A1 describes a MIMO radar system in which multiple transmit antennas transmit simultaneously. Pseudo-Random Phase Modulation (PRPM) is used to randomly change the phase of the signals guided to the respective transmit antennas in order to achieve a degree of orthogonality between the simultaneously transmitted and received signals.

[0006] Digital modulation methods with multiple carrier frequencies are known as OFDM (orthogonal frequency division multiplexing). The use of OFDM methods in radar systems is increasingly being studied. In OFDM, the frequency band is divided into multiple orthogonal sub-bands or subcarriers (FDM, frequency division multiplexing), and OFDM symbols are transmitted sequentially. The transmitted signal of an OFDM symbol consists of mutually orthogonal subcarrier signals modulated according to the symbol's modulation scheme, which are transmitted simultaneously within the OFDM symbol period.

[0007] The applicant has proposed a MIMO radar system in which the repetition rate of the multiplexing sequence is lower than the Nyquist extremum of a single-valued Doppler measurement, resulting in a relative velocity measurement that, while having high resolution, is multi-valued. To resolve this multi-valuedness, the proposed method fully utilizes the fact that the demultiplexing method provides high-quality results only when the phase deviation caused by the Doppler effect is properly corrected, allowing the reconstructing of signals that can be single-valuedly assigned to the individual transmit antennas of the transmit array from the received signals. However, this requires knowledge of the relative velocities of the relevant objects. The multi-valuedness is resolved by testing different multi-valued hypotheses and then selecting the one that provides the signal with the highest quality metric during demultiplexing. Summary of the Invention

[0008] The objective of this invention is to provide a MIMO radar system with a simply constructed antenna array, wherein the relative velocity and positioning angle of a radar object can be determined in a short measurement time with high accuracy and a large range of single-valuedness.

[0009] According to the present invention, this task is solved by a MIMO radar system, which has the following characteristics:

[0010] - A transmitting array having a plurality of transmitting antennas arranged at a distance from each other in an angularly resolved direction.

[0011] - A receiving array having a plurality of receiving antennas arranged at a distance from each other in the angular resolution direction.

[0012] - Wherein, the antenna spacing in one of the transmitting and receiving arrays is higher than the Nyquist extremum for single-value angle measurement (spatial frequency below the Nyquist extremum), however, the antenna spacing in the combination of the transmitting and receiving arrays is lower than the Nyquist extremum (spatial frequency above the Nyquist extremum), and has

[0013] - A control and analysis processing device, the control and analysis processing device being configured to:

[0014] - In each of a plurality of repeatedly performed measurement cycles, a sequence of transmitted signals is transmitted via the transmitting array. The measurement cycle is divided into a plurality of measurement blocks, wherein a uniform multiplexing scheme is used within each measurement block, and the multiplexing scheme varies from measurement block to measurement block.

[0015] - Based on the transmitting array, Doppler estimation and angle estimation are performed on the signals received in each measurement block.

[0016] - Perform Doppler correction on the received signal based on the Doppler estimation.

[0017] - Demultiplex the Doppler-corrected received signal, and

[0018] - Taking into account the phase of signals transmitted by different transmitting antennas, the Doppler estimation and angle estimation are improved based on the signals received in different measurement blocks of the measurement period.

[0019] The present invention modifies the previously proposed method by implementing undersampling, i.e., sampling below the Nyquist extreme, instead of during Doppler measurements, during angle measurements. To this end, the aperture of the transmitting array or the receiving array is chosen such that high angular separation capability is achieved, but multi-valued results are obtained. Conversely, the apertures of other corresponding arrays are chosen such that, in a virtual array representing different combinations of transmitting and receiving antennas, the gaps between the antennas are filled to such an extent that single-valued, but high-resolution results are obtained.

[0020] However, angle estimation based on virtual arrays requires demultiplexing the received signal, that is, reconstructing the signal that can be individually assigned to each transmitting antenna, which requires correction of the phase step caused by the Doppler effect.

[0021] The system proposed here is unique in that the repeated cycle duration of the multiplexing scheme extends over the entire duration of the measurement cycle, thus encompassing multiple measurement blocks. To achieve high-resolution and single-valued Doppler measurements, signals obtained from different measurement blocks can be combined to maximize the measurement time for a given total duration of the measurement cycle, thus enabling the determination of the relative velocity of the located object with greater accuracy. Since these values ​​form the basis of Doppler correction, higher signal quality is achieved during the reconstruction (demultiplexing) of the signals transmitted by the individual transmitting antennas, thereby improving the overall accuracy of the measurement results during signal analysis processing, given the complexity of the antenna array and the computational overhead.

[0022] The preferred configurations and extensions of the present invention are given below.

[0023] Given the extended multiplexing period, the objective is to correct for migration effects during signal analysis and processing. These migration effects arise from changes in the spacing of the located objects during the extended measurement duration. To this end, for signals obtained in different measurement blocks, the coordinates representing relative velocity and spacing are transformed in the multidimensional probe space such that the signal peaks representing estimates of the spacing and relative velocity of the located objects appear at the same location in all measurement blocks. This compensates for the migration effect, ensuring that it does not adversely affect the analysis and processing of signals obtained from different measurement blocks.

[0024] In one implementation, the radar system is an FMCW sequence radar or a linear frequency modulated sequence radar, in which the frequency of the transmitted signal is modulated by a sequence corresponding to a steep frequency ramp (so-called linear frequency modulation), the slope of which is so large that the Doppler effect on the ramp can be ignored, thus performing a pure propagation time measurement, i.e., a distance measurement. The relative velocity is then measured by analyzing the phase deviation caused by the Doppler effect, which varies with the ramp.

[0025] Based on the MIMO principle, suitable multiplexing and demultiplexing methods are needed to separate signals transmitted by different antennas. Code division multiplexing (CDM) and time division multiplexing (TDM) methods are particularly suitable for this purpose.

[0026] In code division multiplexing, signals simultaneously transmitted by transmitting antennas are encoded using orthogonal or quasi-orthogonal codes, and the codes are determined in a code matrix. The allocation of codes to different transmitting antennas remains constant within each measurement block, but varies from measurement block to measurement block.

[0027] If a complete measurement cycle comprises, for example, nM measurement blocks, the phases of the signals received in different measurement blocks form a vector with nM components, each measurement block having a pre-defined code-to-transmitting antenna partition. In the case of a quadratisch code matrix, nM equals the number of antennas transmitting simultaneously in the transmitting array. The received signal can then be decoded by multiplying the vector by the inverse of the code matrix. However, this requires that the orthogonality of the encoded signal be preserved or at least recovered when reflected from a radar target. When the relative velocity of the radar target is not zero, the Doppler effect causes some interference with orthogonality, resulting in the decoded signal for a given transmitting antenna containing signal portions from different transmitting antennas.

[0028] Therefore, in order to obtain measurement results with high quality, the received signal must be corrected according to the Doppler effect.

[0029] In another implementation, time-division multiplexing can be used instead of code-division multiplexing. In this case, only one transmit antenna of the transmit array is active in each measurement block, and the transmission is switched between different transmit antennas on a per-mesh-block basis according to a fixed scheme. The order in which the antennas are activated is typically different from the order in which they are spatially arranged in the transmit array. Therefore, the relative motion of the radar target results in a characteristic phase shift between the signals obtained from the different transmit antennas due to the time offset (with which the signal is transmitted), and this phase shift can be distinguished from the phase shift caused by an angle-dependent propagation time difference—which occurs when the signal is transmitted at a defined angle relative to the normal of the transmit array. In this case, phase errors caused by relative motion can also be corrected based on high-resolution Doppler measurements.

[0030] It is also conceivable that the present invention can be implemented using a combination of code division multiplexing and time division multiplexing.

[0031] Similarly, it is also conceivable to implement a non-square code matrix. If the number of columns (code instances) of the code matrix is ​​less than the number of transmitting antennas, an underdetermined set of equations is obtained during decoding; however, this can be resolved with reasonable additional assumptions (e.g., regarding the number of simultaneously located radar targets). Conversely, if there are more code instances than transmitting antennas, an overdetermined set of equations can be achieved. In this case, greater robustness is achieved for multivalued resolution, for example, relative to signal noise or other interference effects.

[0032] In one implementation, the receiving array has a large, partially filled aperture, making angle estimation based on the receiving array high-resolution but multi-valued, while the transmitting array has a fully filled but smaller aperture, thus enabling single-valued angle measurements with lower resolution. However, in another implementation, the transmitting array can also have a large aperture, while the receiving array has a smaller aperture. Attached Figure Description

[0033] The embodiments are described in more detail below with reference to the drawings. They show:

[0034] Figure 1 A schematic diagram of the simulation portion of a MIMO radar system with independent spacing and velocity determination is shown.

[0035] Figure 2 A diagram showing the frequency of the FMCW transmitted signal and the modulation scheme of the transmitted signal;

[0036] Figure 3 A diagram showing the antenna array of a radar system;

[0037] Figure 4 A diagram showing the positioning angle range of the radar system;

[0038] Figure 5 A block diagram of a digital signal analysis and processing apparatus according to an embodiment of the present invention is shown; and

[0039] Figure 6 and 7 A block diagram of a digital signal analysis and processing apparatus according to another embodiment of the present invention is shown. Detailed Implementation

[0040] according to Figures 1 to 4 An embodiment of a fast linear frequency modulated MIMO radar system will be illustrated as an example of an FMCW-MIMO radar system, in which the encoding of the transmitted signal is performed using phase modulation. Figure 1 The structure of the analog section of the radar system is shown schematically and in a simplified manner.

[0041] Frequency modulation device 10 controls HF oscillator 12, which generates a sequence of identical signals in the form of frequency ramps for multiple transmit antennas 14. In each of the multiple transmit channels, a corresponding phase modulator 16 connected in front of amplifier 18 modulates the phase of the signal according to a corresponding code 20 generated by code generator 22. The phase-modulated signal is transmitted through one of the transmit antennas 14. The transmitted signal, reflected on object 24, is received by multiple receive antennas 26 and, in each receive channel, is mixed with a portion of the unmodulated signal from HF oscillator 12 by mixer 28 and brought to a low-frequency region. Then, A / D conversion is performed in a conventional manner by A / D converter 30.

[0042] exist Figure 2 The diagram schematically illustrates the frequency variation curve of the signal transmitted during the measurement period. The measurement period is divided into multiple measurement blocks 32. A "fast linear frequency modulation" frequency modulation scheme with a sequence of relatively "fast" frequency ramps 34 is used, so that the analysis and processing of spacing and velocity can be performed substantially independently of each other, for example, by means of a two-dimensional Fourier transform. In particular, the Doppler shift within the ramps can be neglected. Each frequency ramp 34 has a duration T. R .

[0043] according to Figure 2 The code matrix M determines the code 20 for each transmit antenna 14. The code matrix assigns a code value A, B, C, ... of the associated code 20 to each signal for each transmit antenna. A single code value defines a phase, with which the phase modulator 18 modulates the signal. Each column of the code matrix (also called code instance I) defines a code multiplexing scheme that assigns a specific code value to each transmit antenna and thereby determines a specific phase modulation. The number of code instances corresponds to the number of transmit antennas transmitting simultaneously.

[0044] Within each measurement block 32, the code instance—and thus the multiplexing scheme—remains unchanged, and switches to the next code instance when transitioning to the next measurement block. The number of code instances is the same as the number of measurement blocks 32 within the measurement cycle.

[0045] In the sequence of measurement block 32, the multiplexing scheme traverses all code instances I = 1, ..., m of the code matrix M. Figure 2 In the example shown, the measurement blocks 32 are sequentially arranged without gaps and have a uniform block duration T. M The duration of this block corresponds to the clock rate 1 / T. M . Figure 2The indices C# = 1, ..., m in the code matrix are used to count the measurement blocks, while the indices TX = 1, ..., n (n = m) are used to number the transmitting antennas. The codes 20 in the code matrix are orthogonal to each other (preferably completely orthogonal or alternatively quasi-orthogonal, i.e., the cross-correlation between codes is small). Therefore, the signals of each transmitting antenna are encoded by the codes; the transmitted signals are orthogonal to each other so that signal separation can be achieved in the receiving channel.

[0046] In another embodiment, the measuring blocks can also be separated from each other by intervals of the same or different lengths.

[0047] like Figure 3 As shown, transmitting antenna 14 forms transmitting array 36, and receiving antenna 26 forms receiving array 38. In the example shown, both arrays are two-dimensional, thus enabling MIMO angle measurements not only in azimuth but also in elevation.

[0048] In the receiver array 38, the receiver antennas 26 are arranged at a uniform spacing in the angular resolution direction y, for example, in the azimuth direction. Here, the spacing between the individual receiver antennas is so large that a large aperture and correspondingly high angular resolution can be achieved with a small number of antennas. However, the spacing between the antennas is greater than half the wavelength of the radar radiation, thus failing to satisfy the Nyquist single-valuedness criterion.

[0049] exist Figure 4 The image shows the field of view of the radar sensor, which includes angles from -θ to +θ about the axis x, which is perpendicular to the plane of the receiving array 38. Only when the positioning angle is within -θ... a to +θ a When the angle measurement is within a significantly smaller range, the result is single-valued. If larger positioning angles cannot be excluded, multiple angles exist, for which the same phase relationship between the signals on the receiving antenna is obtained, thus the measurement is no longer single-valued.

[0050] exist Figure 3 In the example shown, the receiving antennas 26 are also arranged at a uniform spacing in terms of elevation (in the angular resolution direction z), and the antenna spacing is so large in this direction that non-single-value undersampling occurs.

[0051] The transmitting antennas 14 of the transmitting array 36 are arranged at a non-uniform spacing in terms of azimuth; however, this spacing is chosen to enable single-value angle measurements. However, for this purpose, the aperture is significantly smaller than that in the receiving array 38, resulting in lower angular resolution. Similarly, the transmitting array 36 is designed with a small aperture in terms of elevation for single-value angle measurements.

[0052] exist Figure 3 Additionally, a composite array 40 is shown, which is obtained by combining each receiving antenna in the receiving antenna 26 with each transmitting antenna in the transmitting antenna 14, such that the propagation time difference of the signal from the transmitting antenna to the object is added to the propagation time difference from the object to the receiving antenna. Ultimately, it is the aperture of this virtual array 40 that determines the resolving power of the radar sensor. However, it is necessary to separate the signal portions from the different transmitting antennas 14 in the received signal in order to resolve the multivalued nature of the receiving array.

[0053] In addition, Figure 3 In the example shown, the two angularly resolved directions y and z in the receiver array 38 are decoupled from each other because for every y-position of the receiver antenna, all z-positions are occupied by the receiver antenna. Conversely, the transmitter array 36 is an example of an undecoupled array, in which for some y-positions ( Figure 3 Regarding the two rightmost positions in the array, not all z-positions are occupied. Generally, decoupled arrays simplify data analysis and processing, while uncoiled arrays require fewer antenna elements. Both the transmitting and receiving sides can make different decisions between decoupled and uncoiled arrays based on their respective requirements.

[0054] The equidistant arrangement of antenna elements (in azimuth and / or elevation) simplifies data analysis and processing because, for example, a Fast Fourier Transform (FFT) can be used. On the other hand, a non-equidistant arrangement of antennas (as in transmitting antenna 14 here) has the advantage of optimizing the single-valued angular range when the aperture is given. Figure 4 ).

[0055] Generally, the radar system described herein can take into account all combinations of equidistant and non-equidistant arrangements, as well as decoupled and undecoupled arrangements. The following embodiments are also possible, in which the transmitting array is designed for multi-valued, high-resolution angle measurements, while the receiving array is designed for single-valued, lower-angular-resolution angle measurements.

[0056] Now we should base it on Figure 5 To illustrate an example of an analytical processing apparatus, which is used for analyzing and processing data using... Figure 3 Antenna array and by means of Figure 2 The received signal is obtained by the multiplexing scheme.

[0057] The digital data provided by the A / D converter 30 is sampled within a complete measurement cycle. The total number of (complex) signal values ​​received during the measurement cycle is given by the following product: the number of receiving antennas 26 of the receiving array 38, nRX; the number of measurement blocks, nM; the number of repetitions of the frequency ramp 34 within each measurement block, ns; and the number of sampling points on a single frequency ramp 34, nf. In the processing stage 42, a four-dimensional Fourier transform (4D-FFT) is performed on the data sampled during the measurement cycle. The result is a four-dimensional spectrum with the following dimensions: "azimuth 1", "elevation 1", "Doppler 1", and "spacing". The dimension "azimuth 1" describes the distribution of complex amplitudes within the azimuth angle range based on the data of the receiving antennas 26 arranged in the same row in the azimuth direction. Correspondingly, the dimension "elevation 1" describes the distribution within the elevation angle range based on the data of the receiving antennas 26 arranged in the same column in the elevation direction. The dimension "Doppler 1" describes the distribution of complex amplitudes within the elevation angle range via the clock rate ns / T of the frequency ramp. M (slope duration T) R The Doppler spectrum is obtained through "slow" sampling (the reciprocal of the original value). It should be noted that the results in all dimensions "Azimuth 1" and "Elevation 1" are multi-valued due to the corresponding undersampling. The results in dimension "Doppler 1" are single-valued due to the chosen clock rate, but have only low resolution due to the relatively short duration of individual measurement blocks. The dimension "Spacing" describes the spacing spectrum based on "fast" sampling on each frequency ramp 34. The results in this dimension are also single-valued. A separate spectrum is obtained for each transmitted ramp.

[0058] In a four-dimensional spectrum, each located radar target is represented as a peak at a defined coordinate position. If all radar targets are stationary relative to the radar sensor, the power spectrum obtained for each measurement block in a series of measurement blocks is identical. However, if one of the radar targets has an undisappeared relative velocity, the spacing of that object can change significantly over the duration of the measurement period. This then produces a so-called migration effect, which causes the position of the spacing coordinates of the peaks in the spectrum to shift slightly with each measurement block, the degree of which depends on the relative velocity. This effect can be compensated for with relatively little computational overhead by shifting the spacing coordinates into the spacing-velocity subspace such that the migration of the radar target in the spectrum is reversed. The result of this correction is a set of spectra in which all peaks occupy the same position.

[0059] These corrected four-dimensional spectra can then be integrated incoherently (the absolute values ​​of the complex amplitudes are added) without blurring the peaks. The result is an amplitude distribution in a four-dimensional detection space 44. Each point in this four-dimensional space is assigned a definite value of the sum of amplitudes, and given a definite spacing, a definite Doppler offset, a definite azimuth, and a definite elevation angle, each located object in this space is represented by a significant maximum value, where the azimuth and elevation parameters are multivalued, so that only one of the multiple hypotheses about the azimuth and elevation angles can be assigned to the object. The four-dimensional coordinates of the found peaks are then searched in this detection space 44, each peak representing a detection result. For each of these points, (before the incoherent integration) there are nM complex amplitudes, which form a vector with nM components and are now further analyzed to refine the measurement of relative velocity and simultaneously resolve the remaining multivaluedness during angle measurement.

[0060] Therefore, in the fusion phase 46, for each object being detected, the amplitudes obtained in different measurement blocks are added together (coherently) to obtain a higher resolution Doppler spectrum due to the measurement duration extended over the entire measurement cycle, and thus to obtain a more accurate result for relative velocity, "Doppler 2".

[0061] Now, these more accurate values ​​for relative velocity can be used to correct the phase appearing in the signal vector corresponding to the relative velocity. Thus, the orthogonality of the codes in the recovery code instance allows for correct decoding and the acquisition of a decoded signal vector, whose components describe the phase of a signal originating from one of the transmitting antennas 14. Then, based on the components of this vector belonging to the transmitting antenna 14 arranged in the azimuth direction, a single-valued (but low-resolution) value for the azimuth angle can be determined, and correspondingly, a single-valued value for the elevation angle can be determined based on the components belonging to the transmitting antenna 14 arranged in the elevation direction.

[0062] Now, based on these criteria, multivaluedness can be distinguished in the second detection phase 48. For this purpose, the phase-corrected and decoded signal vectors obtained in the fusion phase 46 are summarized, for example, coherently added, into a three-dimensional spectrum. This spectrum has dimensions “Doppler 2”, “azimuth 2”, and “elevation 2”, and the clearest (and highest) peak in this spectrum indicates the true and single-valued values ​​for the relative velocity, azimuth, and elevation of the object. Multiple targets in this 3D space can also be distinguished in the same way.

[0063] The results obtained in the first detection phase 14, "azimuth 1" and "elevation 1", are high-resolution but multi-valued, while the results obtained in the second detection phase 48 are low-resolution but single-valued. These results can now be fused by selecting the value among the multiple high-resolution values ​​for "azimuth 1" that best matches "azimuth 2" as the final, single-valued azimuth value. The values ​​for elevation are also fused in a corresponding manner.

[0064] When searching for "Doppler 2", "azimuth 2", and "elevation 2" in the spectrum, the following can be taken advantage of: since the results for "Doppler 1", "azimuth 1", and "elevation 1" already exist, only the relatively small area of ​​the probe space in which the value being sought must be located can be considered, thereby limiting the search to these areas.

[0065] Since each located object can have a different relative velocity, for each detection result obtained in the detection space 44, the calculations in the fusion phase 46 and the second detection phase 48 are performed separately.

[0066] A particular advantage of common velocity and angle estimation arises from the fact that the multidimensionality of the detection space leads to the entropy of individual peaks. For example, if two radar targets have nearly identical relative velocities, it is difficult to separate the peaks when observing only the "relative velocity" dimension. Conversely, in a two-dimensional or three-dimensional detection space with dimensions of "relative velocity," "azimuth," and "azimuth," the peaks will typically have significantly greater spacing, making it easier to separate them from each other and allowing for the identification and determination of even minor differences in relative velocity.

[0067] radar system in Figure 1 The structure shown also allows for alternative operating modes, where the transmitted signal is transmitted in time-division multiplexing instead of code-division multiplexing. In this case, the code generator 22 manipulates the respective amplifiers 18 such that only one transmit antenna is active at any given time, and switches between the transmit antennas in a defined order. These switches then separate the respective measurement blocks from each other. Encoding of the transmitted signal is unnecessary in this case because the transmitted signals are already separated from each other due to the time-shifted transmission. However, phase correction of the received signal is still required in time-division multiplexing to compensate for the time shift between the signals transmitted sequentially by the transmit antennas 14. Similar to the phase correction used to restore orthogonality in code-division multiplexing, the phase correction in this case is also based on a result that can be determined with high accuracy for the relative velocity.

[0068] Similarly, it is also conceivable to combine code division multiplexing with time division multiplexing. In this case, the transmitting antenna 14 is divided into multiple groups, which transmit simultaneously and encode the signals of each group using a correspondingly smaller code matrix.

[0069] For in Figure 5 The signal analysis and processing types shown in the diagram have different alternatives.

[0070] Figure 6 An example is shown where, in processing phase 42, a two-dimensional Fourier transform of only the dimensions “Doppler 1” and “spacing” is first performed on the digital data sampled within a measurement cycle. Thus, for each object detected, the first detection phase provides the value “spacing” and a low-resolution value “Doppler 1” for the relative velocity.

[0071] In fusion phase 6, the relative velocity measurement is refined based on the signals from all measurement blocks, then phase correction is performed on the corresponding relative velocity, and the phase-corrected signal vector is decoded. In this way, a high-resolution value for the relative velocity, "Doppler 2," and a set of signals TX, associated with the simultaneously activated transmitting antenna 14, are obtained. Then, based on these signals, in angle estimation phase 52, angle estimation is performed not only based on the transmitting array 36 but also on the receiving array 38, thereby obtaining a single-valued and high-resolution value for the azimuth angle, "azimuth angle." Therefore, in the second detection phase 48, for each detection result of the first phase, a two-dimensional spectrum of dimensions "Doppler 2" and "azimuth angle" is obtained.

[0072] exist Figure 7 Another variation of the analysis and processing method is illustrated. In this method, a three-dimensional Fourier transform of dimensions “Doppler 1”, “spacing”, and “azimuth 1” is performed in processing stage 42, resulting in single-valued spacing values, single-valued values ​​for relative velocity “Doppler 1”, and multi-valued values ​​for “azimuth 1” in the detection space 44. Then, in fusion stage 46, phase correction and decoding are performed for each detection result. Then, in angle estimation stage 52, angle estimation is performed on the decoded signal based on the transmitting array 36 to obtain single-valued values ​​for azimuth “azimuth 2”. Furthermore, this stage provides higher resolution values ​​for “Doppler 2”. Then, in the second detection stage 48, for each detection result, peaks are searched in the two-dimensional spectrum of dimensions “Doppler 2” and “azimuth 2”, and then the detection results obtained in detection stages 44 and 48 are fused in the manner already described.

[0073] In a kind of Figure 3In different radar systems, the transmitting array has a large, unfilled aperture while the receiving array has a smaller aperture. The described analytical processing method can still be used in this radar system, but the roles of the transmitting and receiving arrays are reversed. Then, during angle estimation, single-valued but low-resolution values ​​are obtained in the first step, and in the second step, only the environment of these values ​​needs to be studied to improve accuracy.

Claims

1. A MIMO radar system, comprising: A transmitting array (36) having a plurality of transmitting antennas (14) arranged at a distance from each other in angularly resolved directions (y, z). A receiving array (38) having a plurality of receiving antennas (26) arranged at a distance from each other in the angular resolution direction. in, The antenna spacing in one of the transmitting and receiving arrays is lower than the Nyquist extreme value used for single-value angle measurements; however, the antenna spacing in the combination of the transmitting and receiving arrays is higher than the Nyquist extreme value. And has Control and analysis processing apparatus, the control and analysis processing apparatus being configured to: In each of a plurality of repeated measurement cycles, a sequence of transmitted signals is transmitted via the transmitting array. The measurement cycle is divided into multiple measurement blocks, wherein a uniform multiplexing scheme is used within each measurement block, and the multiplexing scheme varies from measurement block to measurement block. Based on the receiving array (38), Doppler estimation and angle estimation are performed on the received signals received in each measurement block. Doppler correction of the received signal is performed based on the Doppler estimation. Demultiplexing the Doppler-corrected received signal, and Taking into account the phase of the transmitted signals transmitted by different transmitting antennas, the Doppler estimation and the angle estimation are improved based on the received signals received in different measurement blocks of the measurement period.

2. The radar system according to claim 1, wherein the control and analysis processing device is further configured to: In each measurement cycle, for the object being located, independent spacing measurements are performed according to the measurement blocks, and Before performing the Doppler correction, the spacing data obtained from the different measurement blocks are subjected to a velocity-dependent coordinate transformation in order to compensate for the relative motion of the object during the elapsed time between the measurement blocks.

3. The radar system according to claim 1 or 2, wherein the multiplexing scheme comprises code division multiplexing for at least some of the transmitting antennas in the transmitting antenna (14).

4. The radar system according to claim 1 or 2, wherein the multiplexing scheme comprises time-division multiplexing for at least some of the transmitting antennas in the transmitting antennas (14).

5. The radar system according to claim 1 or 2, wherein the transmitted signal includes a sequence of frequency ramps (34) within each measurement block, the slope of which is designed for spacing measurement according to the FMCW principle.

6. The radar system according to claim 1 or 2, wherein the transmitting array (36) is designed for single-value angle measurement in at least one angular dimension, and the receiving array (38) is designed for multi-value angle measurement with higher resolution.

7. The radar system according to claim 1 or 2, wherein the receiving array (38) is designed for single-value angle measurement in at least one angular dimension, and the transmitting array (36) is designed for multi-value angle measurement with higher resolution.

8. The radar system according to claim 1 or 2, wherein the receiving antenna (26) and / or the transmitting antenna (14) are arranged equidistantly in the angular resolution directions (y, z).

9. The radar system according to claim 1 or 2, wherein angle estimation based on the receiving array (38) is performed separately from angle estimation based on the transmitting array (36).

10. The radar system according to claim 9, wherein the receiving antenna (26) is arranged equidistantly in the angle-resolved directions (y, z), and the angle estimation based on the receiving array (38) is performed by fast Fourier transform (FFT).

Images