MIMO Radar System

Abstract

To provide a MIMO radar system which can determine with high accuracy and a large range of uniqueness.SOLUTION: A system includes: a transmission array having a plurality of transmission antennas spaced apart from one another in an angular resolution direction: and a reception array having a plurality of reception antennas, where a transmission signal 32 according to time and a frequency division multiplexing pattern is transmitted through the transmission array, time and frequency space is divided into non-overlapping time slots 36 and frequency sub-bands 38, where only a single transmit antenna is active in each time slot, to transmit only in a single frequency sub-band. In a first evaluation stage, based on the signals received in a measurement cycle, priori estimation of distance and Doppler is performed on the basis of signals from the respective transmit antennas, and in a second evaluation stage, common distance, the Doppler, and an angle are estimated based on phases of the signals transmitted from the different transmit antennas by multi-dimensional estimation algorithm.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a MIMO radar system, in particular to a MIMO radar system for vehicles. [Background technology]

[0002] Radar systems are being used in vehicles to detect the traffic environment over a wider range and provide information on the distance, relative speed and direction angle of the vehicle or of located objects such as obstacles to one or more safety or comfort functions that relieve the driver from the burden of steering the vehicle or completely or partially replace the human driver. Here, multiple input-multiple output (MIMO) systems, in which multiple transmit and receive antennas are used, are increasingly being used.

[0003] WO 2018 / 076005 mentions different kinds of MIMO radar systems, where the transmitters and / or receivers can be located at different locations and can use Time Division Multiple Access (TDMA) and / or Frequency Division Multiple Access (FDMA).

[0004] From DE 10 2014 212 284 A1 a MIMO radar measurement method is known in which a transmission signal is frequency modulated in a ramp manner with a modulation model in which ramp trains are assigned to different transmission switching states, which differ in terms of the selection of antenna elements used for transmission, and which overlap each other in time. A transmission switching state is in turn assigned to a number of trains which overlap each other in time. Based on the peak positions of the two-dimensional spectrum of the signal obtained for the train, a value of the relative velocity of the radar target, which is periodic with a predefined velocity period, is determined. The phase relationship of the spectral values ​​of the spectrum of the train of transmission switching states is compared with the phase relationship expected for each periodic value of the relative velocity, and based on the result of the comparison an estimate of the relative velocity is selected.

[0005] The applicant has already proposed a MIMO radar system in which the repetition rate of the multiplexed array is below the Nyquist limit for an unambiguous Doppler measurement, so that the measurement of the relative velocity is ambiguous, albeit with high resolution. To resolve this ambiguity, the proposed method exploits the fact that demultiplexing methods, in which signals that can be uniquely assigned to the individual transmitting antennas of the transmitting array are reconstructed from the received signals, only give high quality results if the phase offset due to the Doppler effect is appropriately corrected. However, this requires knowledge of the relative velocity of the object in question. The ambiguity is resolved by testing various ambiguity hypotheses and selecting the one that provides the highest quality standard signal during demultiplexing. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] International Publication No. 2018 / 076005 [Patent Document 2] DE 102014212284 Summary of the Invention [Problem to be solved by the invention]

[0007] An object of the present invention is to provide a MIMO radar system that can determine the range, relative velocity and positioning angle of a radar target with high accuracy and within a higher range of unambiguity through efficient signal processing. [Means for solving the problem]

[0008] This problem is solved according to the present invention by: a transmitting array having a plurality of transmitting antennas spaced apart from one another in an angularly resolved direction; a receiving array having a plurality of receiving antennas spaced apart from one another in an angularly resolved direction; - Control evaluation device and This problem is solved by a MIMO radar system equipped with A control evaluation device - transmitting, during each of a number of repeated measurement cycles, a transmit signal in a time and frequency multiplexed pattern via a transmit array, the time and frequency space being divided into non-overlapping time slots and frequency sub-bands, and only a single transmit antenna being active in each time slot and transmitting in only a single frequency sub-band; - in a first evaluation phase, a priori estimate of range and Doppler is generated based on signals received in a measurement cycle, - in a second evaluation stage, a common range, Doppler and angle estimation is performed based on the phases of the signals transmitted from the different transmitting antennas by a multidimensional estimation algorithm, so that the results of the first evaluation stage are qualified by improving their accuracy and / or by eliminating ambiguities. It is configured for:

[0009] In the radar system according to the invention, in the first evaluation stage, so-called undersampling can be carried out at least in the measurement dimensions of the range and relative velocity (Doppler shift) and optionally in the angle dimension (azimuth and / or elevation), whereby all available information is evaluated in a compressed form, whereby initially some information loss is accepted. In the case of range or Doppler estimation, undersampling can be such that, for example, for the individual transmitting antennas, the measurement time is very short and only a limited resolution is achieved. In the case of Doppler estimation, undersampling can be such that, since the sampling frequency for the individual transmitting antennas is very low, longer measurement times and thus higher resolution are achieved, but some ambiguity of the measurement results must be accepted. In the case of angle estimation, undersampling can be such that the transmitting and / or receiving arrays have a small aperture, so that the angular resolution is limited, or the transmitting and / or receiving arrays have an antenna spacing larger than the Nyquist limit (and therefore a small spatial frequency), so that ambiguity occurs. However, now the transmit and receive arrays are collectively configured such that virtual arrays generated by different combinations of transmit and receive antennas have large apertures and antenna spacings that do not exceed the Nyquist limit.

[0010] In a second evaluation stage, the signals acquired from the different transmit antennas are combined together to exploit the complete information content. This exploits the fact that the signals transmitted from the different transmit antennas have time and frequency offsets and thus phases that are Doppler and range dependent, and that these phases contain angular information that is dependent on the aperture of the transmit array. Since the range, velocity and angle dependencies of the phase are coupled to each other, in this stage a multi-dimensional estimation algorithm is used to perform a common parameter estimation in all three measurement dimensions. In the three-dimensional parameter space, each point represents a virtual object whose range, relative velocity and positioning angle are given by the coordinates of that point. For each of these points, it is necessary to check whether the phases of the signals received on the various transmit and receive channels correspond to the values ​​expected for a real object for a given combination of range, relative velocity and angle.

[0011] This parameter estimation can be carried out effectively and resource-savingly, even with a large number of transmit and receive antennas, despite the multidimensionality of this parameter space, since the search area in which the object to be located must be located in the parameter space is already significantly limited by the results obtained in the first evaluation stage. Only those points in the parameter space that are compatible with the results of the first evaluation stage need to be searched. Compatibility in this case means that if the parameters in the first evaluation stage (distance, relative speed, angle) were measured only with limited resolution, the difference between the values ​​obtained in the first and second stages must not exceed the resolution limit of the first evaluation stage. The search can therefore be limited to the immediate surroundings of the point found in the first stage. On the other hand, if a parameter was measured in the first stage with high resolution, but the result was ambiguous, compatibility means that one of the ambiguous values ​​(ambiguity hypotheses) must differ from the value obtained in the second stage by at most the resolution limit of the second stage. The search can therefore be limited to the surroundings of the ambiguity hypothesis in question.

[0012] The disclosure of this application also includes a measurement and evaluation method implemented by the above system.

[0013] Advantageous embodiments and variants of the invention emerge from the dependent claims.

[0014] In an advantageous embodiment, in a first evaluation step, an angle estimate is also performed based on the phase of signals received by multiple receive antennas but transmitted from the same transmit antenna. Depending on the aperture of the receive array and the antenna spacing, these angle data have low resolution or are ambiguous. In a second evaluation step, the resolution is improved or the ambiguity is eliminated by comparing the phases of the signals from the different transmit antennas.

[0015] In one embodiment, the receive array has a large aperture that is not completely occluded so that the angle estimate based on the receive array is high resolution but ambiguous, while the transmit array has a small aperture that is completely occluded, thus allowing an unambiguous angle measurement at lower resolution, however, in other embodiments the transmit array may have a large aperture and the receive array may have a smaller aperture.

[0016] In one embodiment, the radar system is an FMCW or chirp train radar, in which the frequency of the transmitted signal is modulated according to a train of steep frequency ramps, so-called chirps, whose slope is so large that the Doppler effect on the ramps is negligible, and thus a pure time-of-flight measurement, i.e. a distance measurement, is performed. The relative velocity measurement is then performed by evaluating the ramp-to-ramp phase shift due to the Doppler effect.

[0017] According to the principle of frequency division multiplexing, the signals transmitted from different transmit antennas have different center frequencies. However, the frequency sub-bands swept by the frequency ramps may overlap each other. In one embodiment, the frequency ramps transmitted from different transmit antennas have the same duration and the same slope, so that they differ only with respect to the frequency sub-bands.

[0018] In one embodiment, the time division multiplexing is arranged such that each transmit antenna is associated with an equal length sequence of equidistant time slots, which represents the so-called Slow-Time Dimension, in which velocity estimation is performed, and within each time slot, a distance estimation is performed (in Fast-Time).

[0019] The distribution of the transmit antennas among the time slots and frequency subbands is preferably uniform in the sense that all transmit antennas are active the same number of times within a measurement cycle and all frequency subbands are used the same number of times, although the model of the distribution can be regular or irregular.

[0020] In one embodiment, the time slots associated with different transmit antennas are connected to each other such that the frequency ramp of a first transmit antenna follows several frequency ramps transmitted from other transmit antennas before the next frequency ramp in the sequence is transmitted from the first transmit antenna.

[0021] Depending on the time offset between the rows and between the frequency ramps within each row, in principle so-called migration effects can occur, i.e. signal distortions due to the distance and velocity of the object being located changing over the considered time interval.

[0022] However, preferably, the time interval of the gradients and the trains is selected so that this migration effect is negligible. Time division multiplexing of short time intervals is mitigated by the fact that the individual gradients have only a short duration due to frequency division multiplexing, in particular since the entire frequency subband is swept with a given gradient in an already short time. Range and velocity estimation can then be performed in a common stage by a two-dimensional Fourier transformation.

[0023] In another embodiment, the time slots associated with the different transmit antennas are arranged block by block such that in a first block a first transmit antenna transmits the entire train of ramps and in the next block another transmit antenna transmits the entire train, in which case migration effects arising due to time offsets between the blocks and between the ramps within a single block are preferably compensated for by appropriate signal correction.

[0024] The embodiments will be described in detail below with reference to the drawings. [Brief description of the drawings]

[0025] [Figure 1] Schematic diagram of the analog portion of a MIMO radar system with independent range and velocity determination. [Diagram 2] FIG. 2 illustrates the frequencies of an FMCW transmit signal in a frequency and time division multiplexing pattern. [Diagram 3] Diagram of the antenna array of the radar system. [Figure 4] Illustrates the angular positioning range of a radar system. [Diagram 5] 1 is a block diagram of a digital signal evaluation device according to an embodiment of the present invention. [Figure 6] 4 shows a multiplexing pattern according to another embodiment. [Figure 7] FIG. 13 is a block diagram of a digital signal evaluation device according to another embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] An embodiment of a high-speed chirp MIMO radar system will be described as an example of an FMCW-MIMO radar system with reference to Figures 1 to 5. Figure 1 shows a schematic simplified configuration of an analog section of the radar system.

[0027] The frequency modulation device 10 controls high frequency oscillators 12 which generate a train of signals in the form of frequency ramps for a number of transmitting antennas 14. Each of a number of transmitting channels is provided with an amplifier 16 which either blocks the signal or amplifies and forwards it to an associated antenna. The oscillators 12 and amplifiers 16 are driven by a multiplexer 18 according to a time and frequency multiplexing pattern 20 (FIG. 2) such that each of the transmitting antennas 14 transmits a frequency modulated signal in a particular frequency sub-band within a particular time slot.

[0028] The transmitted signal reflected by the object 24 is received by a number of receive antennas 26, and in each receive channel, the signal is mixed with a component from the high frequency oscillator 12 by a mixer 28 to bring it into the low frequency range, and A / D conversion is then performed by an A / D converter 30 in the usual manner.

[0029] In figure 2 a multiplexing pattern 20 is shown. The figure shows a complete measurement cycle of duration T.

[0030] A "fast chirp" frequency modulation pattern is used having a train of relatively "fast" frequency ramps 32, so that range and velocity estimates can be made substantially independent of each other, for example using a two-dimensional Fourier transform. In particular, Doppler shifts in the ramps can therefore be ignored or corrected in the second dimension based on the Doppler estimate.

[0031] The duration T of a measurement cycle is divided into a number ns of equal-length windows 34 of length ts, numbered consecutively from left to right at the bottom of the figure. Each window is divided into a number k=nTDM of non-overlapping time slots 36 of equal length, numbered consecutively at the top for window number 3.

[0032] The frequency f exhibited by oscillator 12 forms a frequency band that is divided into a number b of frequency sub-bands 38. The sub-bands may in principle overlap one another, but in the illustrated example are adjacent with no gaps or overlaps. At the left edge of the figure, the sub-bands are numbered consecutively from bottom to top.

[0033] In each timeslot 36, the frequency ramp 32 is transmitted from exactly one of the transmit antennas 14 and fills exactly one frequency sub-band 38 of the frequency sub-bands 38. The transmit antennas are numbered TX1, TX2, ..TXn and the assignment of the frequency ramps to the transmit antennas is indicated by different line widths and dashed-dot lines.

[0034] In the illustrated example, all frequency ramps 32 have the same slope and the same duration tTDM. The distribution of the frequency ramps among the time slots 36 and frequency subbands 38 is selected so that all time slots and subbands are approximately equally occupied. During the entire window 34, each transmit antenna is active at least once and each subband is also used at least once. However, it is not excluded that a transmit antenna is activated multiple times within a window, either in different subbands or in the same subband. For example, transmit antenna TX1 is active twice in subband 1 (in timeslots 1, 5) and once in subbands 3 and 6 (in timeslots 8, 10).

[0035] In the illustrated example, the division of frequency ramps between time slots 36 and frequency sub-bands 38 is repeated in each of the windows 34. Each frequency ramp in window number 1 thus marks the start of a train 40 of equidistant ramps with ramp spacing ts. All the ramps in a train are in the same frequency sub-band.

[0036] In other embodiments, the distribution may vary from window to window, forming rows with the same gradient spacing, but the gradients do not necessarily have to be in the same subband. For example, the center frequency of the gradient may vary linearly across the row.

[0037] 3, the transmit antennas 14 form a transmit array 42 and the receive antennas 26 form a receive array 44. In the example shown, both arrays are two-dimensional, allowing MIMO angle measurements in both azimuth and elevation.

[0038] In the receive array 44, the receive antennas 26 are equally spaced in the angular resolution direction y, e.g., azimuth. The distance between the individual receive antennas is so large that a large aperture and a correspondingly high angular resolution can be achieved with just a few antennas. However, the Nyquist uniqueness criterion is not met because the antenna-to-antenna distance is now greater than half the wavelength of the radar radiation.

[0039] 4 shows the field of view of the radar sensor including angles from -θ to +θ with respect to an axis X perpendicular to the plane of the receiving array 44. The result of the angle measurement is unambiguous only if the positioning angle is within a significantly small interval from -θa to +θa. If larger positioning angles cannot be excluded, the measurement is not unambiguous since there are several angles for which the phase relationship between the signals at the receiving antennas is the same.

[0040] In the example shown in FIG. 3, the receiving antennas 26 are also arranged at equal intervals in the elevation angle direction (angle resolution direction z), and since the antenna intervals are large in this direction as well, non-unique undersampling is performed.

[0041] The transmitting antennas 14 of the transmit array 42 are unevenly spaced in azimuth, with the spacing selected to provide unambiguous angle measurements, but with a significantly smaller aperture than that of the receive array 44, resulting in poorer angular resolution. Also, in elevation, the transmit array 42 is configured for unambiguous angle measurements with a small aperture.

[0042] 3 further shows a virtual array 46 that is obtained when combining each receive antenna 26 with each transmit antenna 14 such that the propagation time differences of the signal from the transmit antenna to the object and from the object to the receive antenna are additive. Ultimately, it is the aperture of this virtual array 46 that determines the resolution of the radar sensor. However, the signal components emanating from the various transmit antennas 14 need to be separated from each other in the received signal so that the receive array can be disambiguated.

[0043] Also, in the example shown in Fig. 3, for each y position of the receive antenna, all z positions are also occupied by the receive antenna, so the two angular resolved directions y and z in the receive array 44 are separated from each other. In contrast, the transmit array 42 is an example of a non-separated array where for some y positions (the two right positions in Fig. 3), all z positions are not occupied. In general, a separated array makes data evaluation easier, and a non-separated array requires fewer antenna elements. The decision between separated and non-separated arrays can be made differently for the transmit and receive sides depending on their respective requirements.

[0044] Also, an equidistant arrangement of the antenna elements (in azimuth and / or elevation) facilitates the evaluation of the data, since it allows, for example, the use of a Fast Fourier Transform (FFT), whereas a non-equidistant arrangement of the antennas has the advantage that a unique angular range (FIG. 4) can be optimized for a given aperture, as here is the case for the transmitting antenna 14.

[0045] Overall, all combinations of equidistant and non-equidistant or separated and non-separated geometries are contemplated in the radar systems described herein. Similarly, embodiments are possible in which the transmit array is configured for ambiguous high resolution angle measurements while the receive array is configured for unambiguous angle measurements with lower angular resolution.

[0046] Due to the multiplexing pattern 20 and the configuration of the antenna array, the radar signal from the transmitting antennas has the following properties: - The total bandwidth used is not fully occupied at each time point. Instead, only one sub-band is occupied at each time point, so that the entire bandwidth is used over the measurement time by the occupancy of all sub-bands. In one configuration, the entire bandwidth is occupied by each transmitting antenna over the measurement time. In an alternative configuration, the transmitting antennas occupy the bandwidth as a whole, but not by each transmitting antenna individually. - The different rows of transmitting antennas typically have different center frequencies (occupying different sub-bands) and are offset in time.

[0047] Thus, the phase of the transmitted signal at the receiver is affected by the following quantities: - Distance of the object (due to the frequency offset between the transmitting antennas) - Relative velocity (due to the time offset between the rows of transmitting antennas) - The positioning angle of the object (due to the spatial offset of the transmitting antenna).

[0048] Because no time-overlapping frequency ramps are used and only one subband is occupied per time slot, a single receive path of the radar system (Figure 1) is sufficient to detect radar signals of all transmit channels.

[0049] Next, an example of an evaluation device for evaluating a received signal obtained using the antenna array shown in FIG. 3 and the multiplexing pattern shown in FIG. 2 will be described with reference to FIG.

[0050] The digital data provided by the A / D converter 30 is sampled over each complete measurement cycle. The total number of (complex) signal values ​​recorded within a measurement cycle is given by the product of the number of receive antennas 26 of the receive array 44, nRX, the number of time slots 36 per window 34, nTDM, the number of windows 34 per measurement cycle, ns, and the number of scan points on each frequency ramp 32, nf.

[0051] In a first processing stage 48, the data sampled for each time slot are subjected to a four-dimensional Fourier transform (4D-FFT). This results in a four-dimensional spectrum with the dimensions "Distance 1", "Doppler 1", "Azimuth 1" and "Elevation 1". The "Azimuth 1" dimension shows the distribution of complex amplitudes over the azimuth measurement angle range based on data from the receiving antennas 26 arranged in a row in the azimuth direction. Correspondingly, the "Elevation 1" dimension shows the distribution over the elevation angle range based on data from the receiving antennas 26 arranged in a row in the elevation direction. The "Doppler 1" dimension shows the Doppler spectrum obtained by a "slow" scan at interval ts. The "Distance" dimension shows the distance spectrum based on a "fast" scan on the individual frequency gradients 32. A separate spectrum is obtained for each transmitted gradient. The result in this dimension is unambiguous, but has low resolution, since only a narrow sub-band is used in each time slot. The results for all the remaining three dimensions are ambiguous due to the respective undersampling (large temporal distance between successive ramps in the train or large spatial distance between the receiving antennas).

[0052] The 4D spectrum is integrated non-coherently (addition of the absolute values ​​of the complex amplitudes). This results in an amplitude distribution in a 4D parameter space 50. To each point of this 4D space is assigned a specific value of the amplitude sum, in which each localized object is represented in the form of a peak (local maximum) at a specific distance, a specific Doppler shift, a specific azimuth angle and a specific elevation angle, but the last three values ​​are each ambiguous, so that only several hypotheses regarding the relative velocity and only several hypotheses regarding the azimuth and elevation angles can be associated with the object. Then, in this parameter space 50, the 4D coordinates of the detected peaks are searched, each representing a detection result. For each of these points there are nTDM complex amplitudes, which (before the non-coherent integration) form a vector with nTDM components and are further evaluated to resolve the remaining ambiguities and improve the range resolution.

[0053] For this purpose, the (complex) amplitudes obtained in the different time slots 36 are (coherently) added for each detected object.

[0054] After detection, in a second processing stage 52, the phases of the signals of the different transmission channels are used to counteract the limitations of each estimation dimension (range, velocity, angle) in the common parameter estimation. This is possible because the sequences transmitted from the different transmission antennas have time and frequency offsets and thus Doppler and range dependent phases. Since the transmission array 42 occupies a specific aperture, the phase of the transmitted signal also contains angular information. The phase information of the transmission channels is a conjoint formula, so the estimation of range, velocity and angle is performed jointly. Multidimensional estimation algorithms can be used to expand the parameter range of each estimation parameter and calculate the combination of parameters that best represents the phase progression across the transmission channels. For example, the estimation can be performed using a maximum likelihood algorithm.

[0055] Depending on the configuration, the algorithm can be designed to be computationally efficient by searching the search dimensions sequentially, i.e., by not searching the entire parameter space. Alternatively, depending on the configuration, the algorithm can be accelerated by using a faster implementation such as an FFT in one or more measurement dimensions. Depending on the configuration, the algorithm can be designed to be versatile.

[0056] The result is higher resolution range values ​​"Range2" for each object, and still high resolution yet unambiguous velocity and angle estimates "Doppler2", "Azimuth2", "Elevation2".

[0057] An alternative multiplexing pattern 20' is shown in Fig. 6. According to this pattern, the measurement cycle is divided into a number nTDM measurement blocks 54 of length tTDM, each measurement block comprising an equal number ns of time slots 36 of length ts. In each measurement block 54, a train 40' of frequency ramps 32 in a single frequency subband is transmitted from exactly one transmit antenna, where the train 40' may be repeated periodically after a certain number of measurement blocks.

[0058] In both FIG. 2 and FIG. 6, tTDM denotes the period duration after the frequency sub-band and possibly the transmit antenna are changed, and ts denotes the ramp spacing within a row 40 or 40'.

[0059] In comparison with Fig. 2, the sequence 40' has a shorter measurement period, which does not correspond to the full length of the measurement cycle, but only to the length tTDM of the measurement block 54. This reduces the resolution in the Doppler dimension. However, in Fig. 6 the gradient spacing within the sequence has been reduced to the length ts of the time slot 36. Therefore, there is no undersampling and the result of the Doppler measurement is unambiguous.

[0060] Figure 7 shows an example of a signal evaluation device adapted to a multiplexing pattern 20' according to Figure 6. For simplicity, the elevation angle estimation is omitted here. In a first processing stage 48, the sampled data in each measurement block undergo a three-dimensional Fourier transformation (3D-FFT) of the dimensions "Distance 1", "Doppler 1" and "Azimuth 1". The results in the dimensions "Distance 1" and "Doppler 1" are unambiguous but of low resolution, only the results in the angle dimension are ambiguous.

[0061] In the three-dimensional spectrum, each localized radar target is shown as a peak at a specific coordinate position. When all radar targets are stationary relative to the radar sensor, the power spectra obtained for each successive measurement block are identical to each other. However, if one of the radar targets has a non-vanishing relative velocity, a significant change in the distance of this object can occur over the duration of the measurement cycle. A so-called migration effect then occurs, whereby the position of the distance coordinate of the peak in the spectrum moves slightly from one measurement block to the next, the extent of which depends on the relative velocity. This effect can be compensated for with a relatively small amount of computation by shifting the distance coordinate in the distance-velocity subspace so that the migration of the radar targets in the spectrum is reversed. This is done in the correction block 56 of FIG. 7. This correction results in a collection of spectra in which all peaks occupy the same position.

[0062] Then, non-coherent integration and parameter detection are performed in the three-dimensional parameter space 50'. In this way, the low-resolution but unambiguous results "Distance 1" and "Doppler 1" and the high-resolution but ambiguous result "Azimuth 1" are obtained.

[0063] A second processing stage 52 then improves the resolution in the "range2" and "doppler2" dimensions and disambiguates the angle estimates using the aperture of the transmit array 42.

Claims

1. a transmitting array (42) having a number of transmitting antennas (14) arranged at a distance from one another in an angularly resolved direction; a receiving array (44) having a number of receiving antennas (26) arranged at a distance from one another in an angularly resolved direction; - Control evaluation device and A MIMO radar system comprising: The control evaluation device, - transmitting, in each of a number of repeated measurement cycles, a transmit signal (32) in a time and frequency multiplexed pattern via said transmit array, the time and frequency space being divided into non-overlapping time slots (36) and frequency sub-bands (38), only a single transmit antenna being active in each time slot and transmitting in only a single frequency sub-band; in a first evaluation phase, a priori estimate of the range and the Doppler is made based on the signals received in the measurement cycle, in a second evaluation stage, a common range, Doppler and angle estimation is performed based on the phases of the signals transmitted by the different transmit antennas by means of a multidimensional estimation algorithm, the results of said first evaluation stage being qualified by improving their accuracy and / or by eliminating ambiguities; 2. A radar system configured for:

2. 2. The radar system of claim 1, wherein the transmitted signal comprises a train of frequency ramps (32) whose slopes are configured for distance measurement according to the FMCW principle, and the Doppler measurement is based on the relative phase of the signals obtained in the train of frequency ramps (40, 40').

3. 3. A radar system as claimed in claim 1 or 2, wherein each transmitting antenna (14) is associated with a sequence of equidistant time slots (36) of the same length.

4. 4. A radar system according to claim 2 or 3, wherein the time slots (36) associated with the different transmit antennas (14) are connected to each other such that, before the next frequency ramp of the series (40, 40') is transmitted from the first transmit antenna in the same sub-band, the frequency ramp (32) of the first transmit antenna is followed by several frequency ramps transmitted from other transmit antennas and / or in other sub-bands.

5. 3. A radar system according to claim 1 or 2, wherein the time slots (36) associated with the different transmit antennas (14) are arranged in blocks such that in a first block a first transmit antenna transmits an entire train of ramp waves (40') and in the next block another transmit antenna transmits the entire train.

6. 6. The radar system according to claim 1, wherein the control and evaluation device is further configured to perform a range- and velocity-dependent coordinate transformation on the phases of the signals transmitted and received at different measurement times in order to compensate for relative movements of objects in the time that elapses between the measurement times.

7. 7. The radar system according to claim 1, wherein in the first evaluation stage an angle estimation is also performed based on the phase of the signals received by the multiple receive antennas (26) but transmitted from the same transmit antenna (14).

8. 8. The radar system of claim 1, wherein the transmit array is configured for unambiguous angle measurements and the receive array is configured for high-resolution ambiguous angle measurements.

9. 8. The radar system of claim 1, wherein the receive array is configured for unambiguous angle measurements and the transmit array is configured for high-resolution ambiguous angle measurements.

10. 10. The radar system according to claim 1, wherein the receiving antennas (26) and / or the transmitting antennas (14) are arranged equidistant in the angularly resolved directions (y,z).

11. 11. The radar system of claim 10, wherein the receiving antennas (26) are equidistantly positioned in an angularly resolved direction (y,z) and the angle estimation is performed by a Fast Fourier Transform (FFT) based on the receiving array (44).

Images