Angle-resolved radar sensor

Abstract

An angle-resolved radar sensor is provided that reduces the calculation cost for signal analysis and enables unambiguous angle measurement with high resolution. [Solution] This angle-resolved radar sensor comprises an antenna array 10 including N antenna elements 12 offset from one another in the scanning direction y, a digital beamforming device 20, and an angle estimation device 22 configured to estimate the angle based on the signal from the beamforming device 20, wherein the aperture A of the antenna array 10 is greater than (N-1) / 2 in units of wavelength λ, and the center-to-center distances between adjacent antenna elements 12 are different from one another but do not deviate by more than a predetermined amount from the value A / (N-1).

Description

[Technical Field]

[0001] The present invention relates to an angularly resolved radar sensor comprising an antenna array with N antenna elements offset from one another in a scanning direction, a digital beamforming device and an angle estimation device configured to estimate an angle based on a signal of the beamforming device. [Background technology]

[0002] Radar sensors are used in automobiles, for example, to measure the distance, relative speed, and azimuth angle of a target vehicle or other object ahead of the vehicle. Here, multiple antenna elements are arranged, for example, horizontally at intervals so that different azimuth angles of the target object result in different propagation distances that radar signals must cover from the object to each antenna element. Such propagation distance differences result in corresponding phase differences in the signals received by the multiple antenna elements, which are then analyzed by corresponding analysis channels. By comparing the (complex) amplitudes received by the multiple different channels with the corresponding amplitudes on an antenna diagram, the angle of incidence of the radar signal and the azimuth angle of the target object can be determined.

[0003] The elevation angle of an object can be measured in a similar manner, where the scanning direction in which the antenna elements are offset from one another is vertical rather than horizontal.

[0004] To achieve high angular resolution, it is desirable for the antenna aperture to be as large as possible. However, if the spacing between adjacent antenna elements is too large, the same phase relationship will be obtained between multiple received signals for propagation distances that differ by an integer multiple of the wavelength λ, which can lead to ambiguity in the angle measurement. Unambiguous angle measurement can be achieved, for example, using a uniform linear array (ULA) structure in which multiple antenna elements are spaced λ / 2 apart.

[0005] A technique called "digital beamforming" allows the antenna array's main sensitivity direction to be changed, thereby directing the radar sensor's so-called main receive lobe in a specific direction. However, side lobes with varying sensitivity also appear on either side of the main receive lobe. In digital beamforming, the complex amplitude received by each antenna element is weighted by an angle-dependent complex phase factor, which corresponds to the difference in propagation distance of the radar beam at a given angle. When multiple beams with different azimuth angles are formed in this way, a gain function is obtained for each beam, which indicates the antenna gain for each angle when the target object is located in that angle. Ideally, for an object located at a particular angle, it is desirable to measure a signal strength at each beam that corresponds to the theoretical antenna gain for that beam. The measured amplitudes should have the correct relationship to each other only at exactly one angle—the angle at which the object is actually located.

[0006] However, in reality, the measured signals contain more or less noise, and the positioning angle can only be estimated by finding the angle at which the amplitudes measured in different beams best correlate with the theoretical value. This correlation can be expressed, for example, by a so-called DML function (Deterministic Maximum Likelihood Function), and the angle estimation consists in finding the maximum of the DML function.

[0007] If it is desired to enlarge the aperture of the antenna array to achieve higher angular resolution and still maintain unambiguous measurements, the number of antenna elements needs to be increased, however this also increases the number of analysis channels required, which increases the required computational power and therefore the hardware costs.

[0008] From WO 2013 / 056880 a radar sensor is known that operates with a thin antenna array, in which the spacing between at least some pairs of adjacent antenna elements is increased, thereby achieving a larger aperture for a given number of analysis channels. By alternately measuring with different combinations of antenna elements to fill the gaps in the antenna array, the unambiguity of the angle measurement is restored. Summary of the Invention [Problem to be solved by the invention]

[0009] An object of the present invention is to provide an angle-resolved radar sensor that reduces the calculation cost for signal analysis and enables unambiguous angle measurement with high resolution. [Means for solving the problem]

[0010] According to the present invention, the aperture A of the antenna array is greater than (N-1) / 2 in units of wavelength λ, and the center-to-center distances between adjacent antenna elements are different but do not deviate from the value A / (N-1) by more than a predetermined amount, thereby achieving the above objectives.

[0011] Since the aperture is larger than (N-1) / 2, ambiguities in the angle estimate may occur in principle. In the DML function for the ULA, this is manifested by the fact that the function has multiple maxima of the same magnitude at different angles. In accordance with the present invention, the antenna elements are not arranged with exactly equal spacing, but rather the spacing varies slightly between pairs, so that all maxima of the DML function are lower except for one, and the function again has a unique absolute maximum, allowing for unambiguous angle estimation. However, the deviation of the antenna spacing from the value A / (N-1), which corresponds to the spacing within the ULA, is limited in such a way that the same techniques as for the ULA for digital beamforming can still be used and the sidelobes are sufficiently attenuated. In particular, digital beamforming can be performed particularly efficiently using the Fast Fourier Transform (FFT).

[0012] Therefore, the radar sensor according to the invention has a quasi-regular array instead of a ULA, but the deviation from the perfect ULA is such that an unambiguous angle estimate is still possible, taking into account the expected noise level. The "predetermined amount" by which the antenna element spacing may deviate from the ULA value A / (N-1) is selected so that, on the one hand, sufficient robustness against signal noise is achieved, but, on the other hand, the angle-dependent gain function is not distorted too strongly compared to the ULA.

[0013] Advantageous configurations and refinements of the invention are set forth in the dependent claims.

[0014] In one embodiment, the deviation of the center-to-center distance between the antenna elements from the value A / (N-1) is less than 25%, i.e., less than A / 4(N-1) in absolute value, and preferably the deviation is less than 15%.

[0015] The difference in spacing between different pairs of adjacent antenna elements may vary regularly, for example according to a linear function, a quadratic function or according to a higher order polynomial, or for example according to a sinusoidal function. The number N of antenna elements in the array is preferably a power of 2, for example N=8 or N=16, which allows efficient digital beamforming by FFT.

[0016] Exemplary embodiments are explained in more detail below with reference to the drawings. [Brief explanation of the drawings]

[0017] [Figure 1] FIG. 1 is a block diagram illustrating a radar sensor according to the present invention having a quasi-regular antenna array. [Figure 2] FIG. 1 illustrates a conventional radar sensor regular antenna array (ULA). [Figure 3]3 is a diagram showing the gain distribution functions of a plurality of different beams formed by digital beamforming using the antenna array shown in FIG. 2. FIG. [Figure 4] FIG. 4 is a schematic diagram similar to FIG. 3, showing only the gain functions of two beams for clarity. [Figure 5] FIG. 3 shows a DML function for the antenna array shown in FIG. 2. [Figure 6] FIG. 1 illustrates an example of a ULA with an enlarged aperture. [Figure 7] 7 is a diagram similar to FIG. 4 but showing two gain functions for the array shown in FIG. 6. [Figure 8] FIG. 7 illustrates a DML function for the array of FIG. 6. [Figure 9] 8 is a diagram similar to FIGS. 4 and 7, but showing two gain functions of the array according to the invention shown in FIG. 1; [Figure 10] FIG. 2 illustrates a DML function for the array shown in FIG. DETAILED DESCRIPTION OF THE INVENTION

[0018] FIG. 1 shows a radar sensor having a quasi-regular antenna array 10 with eight antenna elements 12. Each antenna element 12 is formed by a row of eight series-fed antenna patches 14, ..., 14 extending in a longitudinal direction z. The row of antenna elements 12 is aligned in a scanning direction y, which corresponds to the horizontal direction of the vehicle in the case of an automotive radar sensor. The radar sensor has angular resolution in azimuth. The spacing between the antenna elements 12, 12 and the spacing between each antenna patch 14, 14 in each row are given in units of the wavelength λ of the radar beam. The spacing between adjacent pairs of antenna elements 12, 12 is also quantitatively shown in FIG. 1 and is approximately 2, with a deviation of less than 7% from the average value of 2.00.

[0019] The width of the antenna array 10 in the scan direction y is approximately 14λ, and the array has an aperture in azimuth A=14. In general, the average spacing between antenna elements 12, 12 (2.00 in this example) is equal to A / (N-1), where N is the number of antenna elements 12, ..., 12 in the array.

[0020] In the illustrated example, the spacing between pairs of antenna elements 12, . . . , 12 increases linearly from 1.87 to 2.13.

[0021] The eight antenna elements 12,...,12 are connected via respective signal lines 16,...,16 to an analysis circuit 18, where the received signals are analyzed in separate receive channels. For example, the radar sensor shown herein may be an FMCW (Frequency Modulated Continuous Wave) radar. In this case, each analysis channel includes a mixer that mixes the signal received by the antenna element with a portion of the transmitted radar signal, thereby obtaining an intermediate frequency signal having a frequency that depends on, on the one hand, the elapsed time of the radar signal traveling from the radar sensor to the object and returning, and on the other hand, on the relative velocity of the object. The intermediate frequency signals are digitized by the analysis circuit 18 and recorded over a predetermined sampling period. During this sampling period, the frequency of the transmitted signal is modulated in a ramp pattern. Based on the frequencies of the intermediate frequency signals obtained by the multiple modulation ramps with different ramp slopes, the distance and relative velocity of the object to be located can be determined in a known manner.

[0022] The signal lines 16, ..., 16 connecting the multiple antenna elements to the analysis circuit 18 are all configured to have the same length to prevent distortion of the phase relationship between the signals on their way to the analysis circuit 18. By comparing the amplitudes and phases (i.e., complex amplitudes) of the signals received on the eight receive channels, it is possible to determine, for each object to be positioned at a specific distance and a specific relative velocity, the angle (azimuth angle) indicating the direction from the radar sensor to that object. For this purpose, the signals received on the eight receive channels are subjected to digital beamforming, e.g., by a fast Fourier transform (FFT), in a beamforming unit 20. The beamforming results are transmitted to an angle estimation unit 22, which detects the azimuth angle φ of the object to be positioned using maximum likelihood estimation.

[0023] To explain the operation of the present invention, we first consider a fully aligned regular antenna array 24 (ULA) as shown in Figure 2, used in conventional radar sensors. The spacing between adjacent antenna elements 12 is uniformly λ / 2 in this array, thereby satisfying the condition of uniqueness. However, this array has only a single aperture A = 3.5 for eight receive channels, which severely limits angular resolution.

[0024] FIG. 3 shows the antenna gain of the beamforming device 20 as a function of the azimuth angle φ for the antenna array 24 shown in FIG. 2. Specifically, this figure shows graphs of normalized gain functions 28-40 for ten receive beams with sensitivity maxima at angles greater than 0°, ±15°, ±30°, ±45°, and ±60°. Each beam has a main lobe with maximum gain and multiple side lobes attenuated by approximately 13 dB. This figure is based on beamforming using an FFT with a rectangular window.

[0025] For clarity, the graphs of gain functions 30 and 36 are again shown separated in Figure 4. The graph of gain function 30 is shown in bold, while the graph of gain function 36 is shown in dashed. It can be seen that gain function 30, which has a maximum at 0°, has symmetric sidelobes, while the gain function for the beam with a sensitivity maximum at -45° is asymmetric.

[0026] In digital beamforming, a weighted sum is formed from the complex amplitudes of the signals received at the eight antenna elements 12,...,12, using complex weighting coefficients that reflect the propagation distance differences from antenna element to antenna element. Because these propagation distance differences depend on the azimuth angle φ, a different set of weighting coefficients is obtained for each beam (having a sensitivity maximum at a particular azimuth angle). If an object is located at a certain azimuth angle φ, the beam with a sensitivity maximum at 0° will produce a signal with an intensity proportional to gain function 28, and for the same object, the beam with a sensitivity maximum at -45° will produce a signal proportional to gain function 36. Similarly, each of the other beams will produce a value given by its corresponding gain function.

[0027] From the different amplitude values ​​obtained for the different beams after beamforming, the azimuth angle at which the object to be located is actually located can be estimated by finding the angle at which the measurement value best correlates with the value given by the gain function.

[0028] As an example, Figure 5 shows a DML function 42 (deterministic maximum likelihood function), also called the "angular spectrum," for a target located at an azimuth angle φ = 0°. The DML function 42 shows the correlation between the measured values ​​for several different beams and the gain function for each azimuth angle φ. The function has been normalized so that the maximum value has a value of 1. It can be seen that the DML function 42 has only a single, clearly significant maximum at angle φ = 0°, where the object to be positioned is located. Instead, if a signal is received from an object located at angle φ = 20°, the DML function shifts to have a maximum at 20°.

[0029] To improve the angular resolution, in the ULA 24 shown in Figure 2, the spacing between the antenna elements 12, ..., 12 is increased to 2λ without increasing the number of antenna elements, resulting in a ULA 44 with an aperture A = 14, as shown in Figure 6.

[0030] Figure 7 shows how this change affects gain functions 28-40. As with Figure 4, only graphs of gain functions 28 and 36 are shown in Figure 7. It can be seen that each of these gain functions has multiple maxima of approximately the same height. For example, gain function 28 has a maximum at 0° (as in Figure 4), but also has other maxima of the same height at + / - 30°. Between these maxima are numerous side lobes that are more strongly attenuated. The same is true for gain function 36 and each of the other gain functions (not shown in Figure 7).

[0031] The DML function 42 of the ULA shown in Figure 6 is shown in Figure 8. This function also has several main maxima with the same height, so that an unambiguous angle estimation is no longer possible.

[0032] In contrast, unambiguous angle estimation is again possible when using the quasi-regular antenna array 10 shown in Figure 1. The gain functions 28, 36 of this array are shown in Figure 9. As in Figure 7, each gain function has three prominent main maxima in the angular range from -50° to +50°, but the side lobes are now more strongly "shouldered" - i.e., they are pushed up and closer to the edges of the main maxima.

[0033] Figure 10 shows the associated DML function 42. Beamforming in the beamforming device 20 is performed using the same weighting coefficients as in the ULA shown in Figures 2 and 6, but the spacing between adjacent antenna elements 12, 12 is slightly non-uniform, so that only one main maximum 46 is present at full height at φ=0° (for a target at 0°), while the side maxima 48, 48 are more strongly suppressed at + / -30°. This means that a unique maximum can be found during angle estimation, and therefore a unique angle estimation is possible.

[0034] Misestimations can occur when the signal is noisy enough that the difference between maxima 46 and maxima 48 is unclear, leading to one of the maxima 48,...,48 being erroneously selected as the highest value to determine the azimuth angle. The more non-uniform the spacing between adjacent antenna elements 12,...,12, the more strongly the maxima 48,...,48 are suppressed, and the angle estimation becomes more robust to signal noise. However, as the non-uniformity of the array 20 increases, sidelobes become increasingly pronounced. However, with appropriate selection of the non-uniform spacing between the antenna elements 12,...,12, unambiguous angle estimation is generally possible even in the presence of signal noise, and higher angular resolution can be achieved based on the enlarged aperture without the need for additional analysis channels.

Claims

1. 1. An angularly resolved radar sensor, comprising: an antenna array (10) including N antenna elements (12) offset from one another in a scanning direction (y); A digital beamforming device (20); an angle estimation device (22) configured to estimate an angle from a maximum value of a DML function based on a signal from the digital beamforming device (20); an aperture A of the antenna array (10) in wavelength λ greater than (N-1) / 2, and the center-to-center distances between adjacent antenna elements (12) are different from one another and are greater than 75% of the value A / (N-1) and less than 125% of the value A / (N-1); An angle-resolved radar sensor characterized by:

2. 2. The radar sensor of claim 1, wherein the center-to-center distance of the adjacent antenna elements is greater than 85% of the value A / (N-1) and less than 115% of the value A / (N-1).

3. 3. A radar sensor according to claim 1 or 2, wherein the center-to-center distance between the antenna elements (10) varies according to a regular pattern.

4. 4. A radar sensor as claimed in claim 3, wherein the centre-to-centre distance between the antenna elements (12) varies according to a polynomial function.

5. 5. A radar sensor according to claim 4, wherein the center-to-center distance between the antenna elements (12) varies according to a linear function.

6. 6. A radar sensor according to any one of claims 1 to 5, wherein N is a power of two.

7. 7. A radar sensor according to any one of claims 1 to 6, wherein A is 1 or more, preferably A is 2 or more.

Images