Radar system

Abstract

The invention relates to a radar system having a transmitting and receiving device which is designed to transmit a transmission signal having a series of measurement periods (Tf) separated from one another by pauses (Tp), at least one sequence of frequency-modulated signals being transmitted in each measurement period, having a digital evaluation device, the digital evaluation device is configured to determine the distance and the radial relative speed of the positioned radar target (12, 14), characterized in that the measurement period (Tf) is divided into a plurality of modulation bursts (36) which are not equidistant.

Description

TECHNICAL FIELD

[0001] The present application relates to a radar system having a transmitting and receiving device configured for transmitting a transmission signal having a series of measurement periods separated from each other by pauses, wherein in each measurement period a sequence of at least one frequency-modulated signal is transmitted, and having a digital evaluation device configured to determine the distance and the radial relative velocity of a located radar target.

[0002] In particular, the present application relates to a radar system for use in a driver assistance system or in an autonomous driving system of a motor vehicle for detecting the traffic environment. BACKGROUND

[0003] Such known radar sensors have a cyclic modulation pattern. Each cycle comprises one measurement period and one pause. In so-called Chirp-Sequence radar systems, the transmission signal is modulated in each measurement period to form a sequence of equidistant frequency ramps. The transmitted and after reflection at a radar target again received signal is mixed with a part of the transmission signal at the reception instant, so that by the difference frequency a baseband signal is obtained, whose frequency is equal to the frequency difference between transmission signal and reception signal. Due to the ramp-shaped frequency modulation, this frequency difference is proportional to the steepness of the frequency ramp and to the propagation time of the signal from the radar sensor to the radar target and back. If a Fourier transformation is performed on the signal obtained on the frequency ramp, a spectrum is obtained, in which each located object appears in the form of a peak, the frequency position of which indicates the distance of the object. If the radial relative velocity of the object with respect to the radar sensor is not zero, an additional frequency shift occurs due to the Doppler effect, which, however, is negligible in the distance measurement as long as the ramp steepness is sufficiently large. The Doppler frequency shift, however, leads to a phase step of the reception signal from ramp to ramp, on the basis of which the radial velocity of the object can be determined. For this purpose, a Fourier transformation is performed on the reception signal received at mutually corresponding sampling time points on successive frequency ramps.

[0004] The steeper the frequency ramp and the greater the bandwidth of the transmission signal, the better the resolution of the radar sensor in the distance dimension (small Δd). Conversely, the longer the observation duration, i.e. the duration of the measurement period, the better the resolution in the relative velocity or Doppler dimension (small Δv). However, due to the periodicity of the radar signal, ambiguities occur in both the distance measurement and the velocity measurement if the sampling time points are too far apart. If d max is the effective range of the radar sensor, i.e. the maximum distance of a radar target that can just still be detected, then in the case of equidistant sampling, d must be distributed over the bandwidthmax Δd sample values. If the speed interval in which the object's relative speed is expected in practice is limited by a lower limit of v min (usually negative) and an upper limit of v max , then the v max -v min ) / Δv time sample values must be distributed over the measurement duration.

[0005] For signal analysis and object tracking, it is advantageous if the frequency ramps sent during the measurement period are equidistant and the duration of the measurement cycle is also constant.

[0006] Another known modulation method is the OFDM method (Orthogonal Frequency Division Multiplexing), which uses a plurality of orthogonal subcarriers. The time samples spanning the measurement duration are achieved by sending a plurality of so-called OFDM symbols. This method also allows distance and speed measurements, but with limited resolution and the corresponding explicitness criterion.

[0007] In order to achieve as high a resolution as possible in the speed dimension, the duration of the measurement cycle should be as large as possible. However, increasing the measurement duration is limited, because with increasing measurement duration, the latency time, i.e. the time elapsed between the generation of the localization result and its availability for the necessary reaction decisions of the driver assistance system, also increases. Another limitation of the measurement duration comes from the thermal capacity of the radar sensor.

[0008] In order to increase the resolution, it is also known to perform the analysis on the basis of signals obtained in a plurality of successive measurement cycles. However, in this case, the migration effects must be generally considered and compensated for in the analysis, in particular the change in the radar target speed over a long observation period. SUMMARY

[0009] It is the object of the present invention to create a radar system which enables high-resolution and explicit relative speed measurements.

[0010] This object is solved according to the invention by dividing the measurement cycle into a plurality of non-equidistant modulation bursts, respectively.

[0011] A modulation burst should be understood here as a sequence of frequency-modulated signals, the time interval of which is typically less than the minimum time interval between consecutive modulation bursts. Therefore, a modulation burst can be considered a “packet” of signals transmitted in a dense sequence, separated by slightly larger time gaps. Signals transmitted in a dense sequence can be, for example, frequency ramps or OFDM symbols. While the signals within a single burst can be equidistant (and should be equidistant for reasons of signal analyzability and processability), it is crucial that the modulation bursts are not equidistant from each other; that is, the time interval between different modulation bursts varies within the measurement period. Compared to conventional radar systems with consecutive frequency ramps having the same (minimum possible) spacing, according to the present invention, the signal sequence is to some extent “staggered”, thus creating larger gaps between individual bursts. Overall, this increases the effective observation duration (the time from the first signal of the first burst to the last signal of the last burst), thereby improving Doppler separation capability without increasing the total number of signals transmitted within the measurement period. Therefore, it does not lead to an increase in the thermal load on the radar sensor. Furthermore, by varying the intervals between modulation bursts, the symmetry of the Doppler spectrum is broken, thereby partially suppressing some periodically occurring peaks in the spectrum that cause blurring. This facilitates blur resolution.

[0012] Advantageous configurations and extensions of the invention are described below.

[0013] To further improve the resolution capability of the Doppler dimension, the analysis process can be performed across multiple cycles, as is known to be. Non-uniform modulation abruptly avoids blurring, although the measurement cycle considered during analysis—and therefore the effective measurement time point—remains within an equidistant grid, which facilitates object tracking in subsequent analysis stages.

[0014] Furthermore, modulation bursts within a measurement period can vary in duration. However, for signal processing purposes, it may be advantageous if the modulation bursts have a uniform duration.

[0015] Multiple analysis processes with different numbers of bursts can also be considered. For example, different numbers of bursts can be analyzed and processed together based on distance range. Similarly, a single burst from the second cycle can also be used, for example.

[0016] In cross-cycle analysis, the precision can be further improved by varying the timing of modulation bursts from measurement cycle to measurement cycle.

[0017] The total time gap duration between modulation bursts in the measurement period represents a degree of freedom, which can be used to optimize the ratio between the measurement period duration and the period time (measurement period duration + pause duration) to suppress sidelobes in the spectrum. This facilitates further resolution of ambiguities.

[0018] The cross-cycle analysis processing used to determine range and relative velocity can be performed in different ways. For example, the received signals from all modulation bursts occurring in a continuous measurement cycle can be analyzed together. However, in that case, migration effects caused by radar target motion or acceleration must be compensated for using known methods.

[0019] Another possibility is to first analyze and process the signals in each measurement cycle separately, thereby forming a two-dimensional spectrum (range-Doppler spectrum) for each measurement cycle based on the modulation bursts transmitted in that measurement cycle, and then fuse the spectra obtained from different measurement cycles within the scope of cross-cycle analysis and processing.

[0020] Another possibility is to first generate a separate two-dimensional spectrum for each individual modulation burst, and then fuse these spectra within the scope of burst and cross-cycle analysis processing. The burst and cross-cycle analysis processing can also be performed in multiple steps, for example, first generating the spectra of each burst, which can then be combined into a three-dimensional spectrum, where the third dimension represents the sequence of bursts in the measurement cycle, and then fusing the 3D spectra obtained for different measurement cycles. Attached Figure Description

[0021] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings. The drawings show:

[0022] Figure 1 A simplified block diagram of the radar system according to the present invention;

[0023] Figure 2 Frequency / time diagram of a transmitted signal modulated in a conventional manner;

[0024] Figure 3 (A) According to Figure 2 The modulation scheme of the transmitted signal;

[0025] Figure 3 (B) A modulation scheme with equidistant modulation bursts;

[0026] Figure 3 (C)-(E) Examples of three modulation schemes with non-equidistant modulation bursts according to the present invention;

[0027] Figure 4 against Figure 3 Doppler spectra of modulation schemes (B) and 3(C);

[0028] Figure 5 against Figure 3 (A) Doppler spectrum of the modulation scheme during analysis across four measurement cycles;

[0029] Figure 6 against Figure 3 (B) Doppler spectrum of the modulation scheme during analysis across four measurement cycles;

[0030] Figure 7 against Figure 3 (C) Doppler spectrum of the modulation scheme during analysis across four measurement cycles;

[0031] Figure 8 against Figure 3 (C) Doppler spectrum of the modulation scheme during analysis across two measurement cycles;

[0032] Figure 9 against Figure 3 (C) The Doppler spectrum of the modulation scheme during analysis across four measurement cycles; and

[0033] Figure 10 Examples of four different analysis and processing methods (A)-(D). Detailed Implementation

[0034] exist Figure 1 The diagram illustrates a simplified FMCW radar sensor 10, mounted, for example, at the front of a motor vehicle, for measuring the distance *d* and relative speed *v* of objects 12 and 14, such as a vehicle traveling ahead. The radar sensor 10 has a voltage-controlled oscillator 16 that provides a frequency-modulated transmission signal to a transmitting and receiving device 20 via a mixer 18, the signal being transmitted from the device toward the objects 12 and 14. Signals reflected from the objects are received by the transmitting and receiving device 20 and mixed with a portion of the transmission signal in the mixer 18. A baseband signal *b* is thus obtained, which is further analyzed and processed in a digital analysis and processing unit 22. The analysis and processing unit 22 includes a control section 24 that controls the function of the oscillator 16. In radar measurements, the frequency of the transmission signal provided by the oscillator is modulated using a rising or falling ramp sequence.

[0035] Figure 2 A diagram illustrates a traditional radar signal modulation scheme, showing the transmitted signal frequency f as a function of time t. The duration is T. z A single measurement cycle 26 includes a duration of T. f The measurement period is 28, followed by a duration of T. pPause 30. After pause 30, a new measurement cycle begins, which is identical to measurement cycle 26. Measurement cycle 28 is in... Figure 2 The modulation block 32 is also symbolically represented in the middle. In each measurement cycle 28, a sequence of equidistant linear frequency ramps 34 is transmitted.

[0036] Figure 3 The same modulation scheme is also shown in (A), but without a frequency / time diagram.

[0037] Figure 3 (B) shows a modulation scheme in which, instead of transmitting a modulation block 32 in each measurement cycle, a series of equally spaced modulation bursts 35 are transmitted. Each modulation burst 35 consists of a dense sequence of frequency ramps, similar to frequency ramps 34. These frequency ramps can have the same characteristics as... Figure 2 The intermediate frequency ramps 34 have the same steepness and the same spacing, but fewer in number because the duration of a single modulation burst 3 is shorter than the duration of modulation block 32. Figure 3 In (B), the modulation block 32 is divided to some extent into four separate, temporally successive modulation bursts 34. Therefore, the duration T of the measurement period 28 is... f Pause for 30 seconds T p Increased at a cost, while the duration T of cycle 26. z It remains unchanged. In another implementation, T p It can also be reduced to zero.

[0038] Figure 3 (C)-(E) show examples of embodiments of the invention, in which non-equidistant modulation bursts (36) are used.

[0039] exist Figure 3 In (C), the duration T of the measurement period f and the duration of the pause T p and Figure 3 (B) is the same, and the spacing pattern between these modulation bursts 36 is the same in each measurement cycle.

[0040] Figure 3 (D) shows an example where the time interval between modulation bursts 36 is less than [the time interval between the 36 modulation bursts is less than 1]. Figure 3 (C). Therefore, with Figure 3 (C) Compared to the measurement period 28, the duration t f It decreased, but is still greater than Figure 3 (A). The change in the interval between modulation bursts 36 also alters the shape of the Doppler spectrum formed based on the received signal. Therefore, changing the duration T of the measurement period 28... f(Without changing the duration of the modulation burst (36)) can be used to modify the obtained spectrum to suppress sidelobes as much as possible, thereby making it easier to resolve ambiguities. Suppressing sidelobes can be a standard even when bursts are not equidistantly arranged within the measurement period.

[0041] Figure 3 (E) shows an example where the duration T of period 28 is measured. f and the duration T of the entire cycle 26 z and Figure 3 (C) is the same, but in the second measurement cycle, the time interval separating the individual modulation bursts 36 is different from that in the first measurement cycle. These variations in intervals from one cycle to another also contribute to improved ambiguity resolution.

[0042] Measurement data contained in a single measurement cycle can be analyzed by performing a two-dimensional digital Fourier transform on the baseband signal b, such as a Fast Fourier Transform (FFT). The range spectrum is obtained by performing a Fourier transform on the sampling points within a single frequency ramp. Each located radar target appears as a peak in this spectrum, its frequency position depending on the object's range. The Doppler spectrum is obtained by performing a Fourier transform on the sampling time points located within a continuous ramp of a modulation burst. This spectrum illustrates the complex amplitude of the received signal as a function of the Doppler frequency. Since this Doppler frequency is proportional to the relative velocity of the object, different points on the frequency axis of this spectrum represent assumptions about the object's relative velocity v. The closer the velocity assumption matches the actual velocity of the object, the greater the received radiated power. For example, if a radar echo is received from a target with a relative velocity v = 0, a peak will appear at v = 0 in the spectrum. The sharper this peak, the greater the observation duration (T). f The longer. For example, according to Figure 3 (A) modulation mode, T f =20ms and T p =30ms, the peak width, i.e., the relative velocity resolution, is approximately 0.098m / s. On the other hand, if the observation duration T f The value is 35ms, therefore, under the same cycle time, T p Only 15ms, such as Figure 4 As shown in (B) and (C), the relative velocity resolution is improved to approximately 0.056 m / s. This resolution also determines the size of the Doppler-Bins interval in the two-dimensional spectrum.

[0043] However, due to the periodicity of the signal, the spectrum for a single radar target does not produce a single peak, but rather a regular sequence of peaks at different Doppler frequencies, and each associated velocity could be the target's true velocity. Therefore, velocity measurements are ambiguous. The smaller the interval between peaks, the smaller the time interval between consecutive frequency ramps 34 within a modulation block or burst. However, in the case of multiple modulation bursts 36 in each measurement cycle, they also depend on the time sequence of the bursts.

[0044] Figure 3 Showing targets Figure 3 The Doppler spectra of the modulation modes shown in (B) and (C) both assume the localization of a single target with zero relative velocity. Curve B, drawn with a thinner line, corresponds to the... Figure 3 Modulation mode (B) has four modulation bursts, each 5ms long, with equal start times of 0, 10, 20, and 30ms. It can be seen that the distance from peak to peak is approximately 0.2m / s. Curve C, drawn with thicker lines, is based on... Figure 5 (C) The spectrum of the modulation pattern, where the 5ms-long modulation burst has non-equidistant start times: 0ms, 5ms, 20ms, and 30ms. This sequence of modulation bursts increases the peak-to-peak distance to 0.4m / s, effectively doubling the velocity range for precise measurements. Simultaneously, this sequence of modulation bursts reduces the mutual interference between two radar targets within the same Doppler range.

[0045] The following section compares the Doppler spectra obtained when different modulation modes are analyzed over multiple measurement cycles (four measurement cycles in this example). Figure 3 Curve a in the figure shows the results based on Figure 6 (A) The spectrum of the modulation mode, Figure 3 Curve b in the figure shows the results based on Figure 7 (B) The spectrum of the modulation mode, Figure 3 Curve c in the figure shows the result according to Figure 5 (C) Spectrum of the modulation mode. It can be seen that in the standard modulation method with only one modulation block 32 per cycle ( Figure 6 The blur appears at intervals of only 0.04ms. And according to... Figure 7 In equidistant modulation bursts, blurring occurs at significantly larger intervals of 0.2 m / s; in modulation using non-equidistant bursts ( Figure 3 The interval between blurs is doubled again to 0.4 m / s.

[0046] Several methods are known to resolve these ambiguities, allowing for the definitive determination of relative velocities over a larger velocity range. Examples of such methods are described in DE102014212280A1, DE102014212284A1, DE102017200317A1, US10921436B2, US11614531B2, and US11774552B2. However, ambiguity resolution is typically only possible if no ambiguity is present within ±0.5 Doppler intervals. In the examples described here, only according to Figure 7 This can only be achieved with modulation mode (C). Furthermore, from... Figure 8 It can be seen that, when using modulation with non-equidistant bursts, sidelobes (targets with other relative velocities) can be suppressed by more than 5 dB.

[0047] Figure 3 and 9 Showing according to Figure 8 (D) The spectrum of the modulation mode, where, to further suppress sidelobes in the spectrum, the duration of the modulation burst is shortened to 3.8 ms, and non-equidistant start times are used: 0 ms, 3.8 ms, 15.2 ms, and 22.8 ms. Cycle time T z The time is 50ms, the same as other modes. Figure 9 The curve d2 in the figure involves analysis and processing across two measurement cycles, while Figure 9 The curve d4 in the figure involves analysis and processing across four measurement cycles. Figure 10 As can be seen, the side lobes were significantly suppressed by more than 5 dB.

[0048] Figure 10 (A)-(D) schematically represent different analysis and processing strategies.

[0049] exist Figure 10 In (A), modulation bursts 36 from multiple consecutive measurement cycles are analyzed and processed jointly (across bursts and across cycles), compensating for migration effects where necessary. For example, if the analysis and processing spans four measurement cycles, a common two-dimensional spectrum 38 (with "distance" and "velocity" dimensions) is obtained for all four measurement cycles.

[0050] exist Figure 3 In (B), each individual measurement cycle is first analyzed separately to obtain its own two-dimensional spectrum 40. Then, cross-cycle analysis involves performing a third Fast Fourier Transform (FFT) 42 across the cycles. The 2D spectrum 40 obtained for each cycle is then ready for object detection.

[0051] exist Figure 10In (C), the first step forms a 2D spectrum 44 for each individual modulation burst 36, and then the modulation bursts across multiple measurement cycles perform a third FFT 46.

[0052] exist Figure 10 In (D), the first analysis and processing step is... ​ (C) Same. However, in the next step, the third FFT is performed only on the spectrum of a single period, thus obtaining a three-dimensional spectrum with "distance", "velocity", and "burst" dimensions for each period. Then, the cross-period analysis process involves performing a fourth FFT50 across periods.

[0053] If necessary, migration effect compensation can be performed before the third FFT 42 or 46, or before the fourth FFT 50.

[0054] In cross-cycle analysis and processing, in various variants, multi-objective models with improved frequency separation capabilities can be used instead of single-objective models to model signals.

[0055] Although in the method described herein, the frequency ramps 34 within a single modulation burst have the same slope, the same spacing, and the same center frequency, variations in these parameters changing within a burst or from one burst to another are also conceivable. For example, a modulation burst could consist of a series of frequency ramps with linearly rising or falling center frequencies.

[0056] Furthermore, the modulation parameters can also vary from measurement cycle to measurement cycle. In that case, the parameter variations must be taken into account appropriately during analysis and processing.

Claims

1. A radar system having a transmitting and receiving device (20) which is configured for transmitting a transmission signal, the transmission signal having a series of measurement cycles (26) separated from one another by pauses (30), wherein, At least one sequence of frequency-modulated signals (34) is transmitted in each measurement cycle (28), the radar system having a digital evaluation device (22) which is configured for determining the range and the radial relative velocity of a located radar target (12, 14), characterized in that the measurement cycles (28) are each divided into a plurality of non-equidistant modulation bursts (36).

2. The radar system of claim 1, wherein, The evaluation for determining the relative velocity is performed across a plurality of measurement cycles (26).

3. The radar system of claim 2, wherein, The evaluation device (22) is configured for calculating a two-dimensional spectrum (38) based on the measurement signals for the modulation bursts (36) in the plurality of measurement cycles (26).

4. The radar system of claim 2, wherein, The evaluation device (22) is configured for first calculating a two-dimensional spectrum (40) for each individual measurement cycle (26) and then evaluating the spectra across cycles in a second step.

5. The radar system of claim 2, wherein, The evaluation device (22) is configured for first calculating a two-dimensional spectrum (44) for each individual one of the modulation bursts (36) in the plurality of measurement cycles (26) and then evaluating the spectra (44) across bursts and across cycles.

6. The radar system of claim 2, wherein, The evaluation device (22) is configured for first calculating a two-dimensional spectrum (44) for each individual one of the modulation bursts (36) in the plurality of measurement cycles (26) and then calculating a three-dimensional spectrum (48) from the spectra (44) obtained for each individual measurement cycle and then evaluating the three-dimensional spectra (48) obtained for the plurality of measurement cycles across cycles.

7. The radar system of any one of the preceding claims, wherein, The modulation bursts (36) have a uniform duration within a measurement cycle (28).

8. The radar system of any one of the preceding claims, wherein, The temporal arrangement of the modulation bursts (36) varies from measurement cycle to measurement cycle.

9. The radar system of any one of the preceding claims, wherein, To determine the relative velocity, a Doppler spectrum is calculated, wherein the ratio between the duration (T f ) of the measurement period (26) and the duration (T z ) of the entire measurement cycle (26) is chosen such that a maximum suppression of side lobes in the Doppler spectrum is achieved.

Images