DE102023204002A1 - antenna array for a radar sensor - Google Patents

Abstract

Antenna array for a radar sensor with angle resolution in azimuth and elevation, with two linear subarrays (Tx, Rx) of antennas (10, 30) which are arranged on two non-parallel straight lines (Y, G), characterized in that the two straight lines (Y, G) run obliquely to one another and that the aperture of a first (TX) of the two subarrays in the direction of the straight line (Y) on which the antennas of this first subarray lie is smaller than the aperture of the second subarray (RX) in the direction of the same straight line (Y).

Description

The invention relates to an antenna array for a radar sensor with angle resolution in azimuth and elevation, with two linear subarrays of antennas arranged on two non-parallel straight lines.

In particular, the invention relates to an antenna array for radar sensors which are intended for installation in motor vehicles and are used in conjunction with a driver assistance system for measuring distances, relative speeds and angles of objects (such as other vehicles and obstacles).

State of the art

MIMO systems (MIMO: multiple input, multiple output), which use multiple transmitting and receiving antennas, are increasingly being used in driver assistance systems. The MIMO principle enables particularly precise angle measurements to be made, whereby the antenna aperture (antenna area) that is important for angle measurement is virtually enlarged. Several transmitting antennas send out their signals without being influenced by one another, and these signals are separated in the receiving channels. The virtual enlargement of the aperture is achieved by the fact that the distance between the transmitting antennas and the receiving antennas is different, meaning that calculations can be made as if there were only one transmitting antenna, but the number of receiving antennas is multiplied, thus creating a virtual larger antenna aperture.

In order to operate multiple transmit antennas in a MIMO configuration without interference, the transmit signals from the transmit antennas must be clearly separable at the receiver. This requires multiplexing the transmit signals, i.e. separating the signals in the frequency or time domain, or using orthogonal code sequences. Separation in the frequency domain is typically achieved by having different transmit antennas occupy different frequency ranges at the same time. The disadvantage of this method is the reduced available bandwidth per transmit channel. The distance separation capability of a radar system is directly proportional to its bandwidth, so the distance separation capability decreases with conventional frequency multiplexing.

The separation of the transmitting antennas is therefore often done in the time domain, i.e. the antennas transmit one after the other in time multiplex. The disadvantage here is that the measurement time increases due to the sequential measurement, and objects may have moved significantly during the increased measurement time, which reduces the accuracy. In addition, with such a time multiplex, the time interval between two sequential measurements of the respective transmitting antenna is increased, which can lead to a reduction in the maximum clearly measurable speed range.

The use of orthogonal code sequences allows the simultaneous operation of the transmit antennas, whereby they can use the same bandwidth. The orthogonality (or cross-correlation) of the code sequences is influenced by time and frequency shifts of the radar signal, so that with time and frequency shifted signals the separation between transmit channels can be impaired or can fail completely.

Out of DE 10 2018 221 085 A1 A chirp sequence method with multiplexing of the transmitting antennas is known, which results in a reduction of the clearly measurable speed range. The signals from the transmitting antennas are designed in such a way (e.g. by a correspondingly selected sequence in time-division multiplexing) that the phases of the transmitting antennas contain a component caused by the Doppler effect. The ambiguities in the speed estimate that arise from time-division or code-division multiplexing can then be resolved using the phase information from the transmitting antennas. However, this resolution of the speed ambiguity must take place together with the angle estimate, since the phase relationships depend on both the speed and the location angle of the object.

Angular resolution in both azimuth and elevation can be achieved if the antenna array has at least two linear subarrays, one of which is oriented horizontally and the other vertically in the installation position of the radar sensor. In order to achieve the highest possible angular resolution, the subarrays should have the largest possible aperture. The aperture can be increased by increasing the number of antennas, but this leads to a corresponding increase in the number of transmit and/or receive channels and thus to a higher computational effort in signal evaluation. If the aperture is increased without increasing the number of antennas, the distance between the antennas must be increased. However, as soon as this distance becomes greater than half the wavelength of the radar radiation, the relationship between the phases of the received signals and the detection angle is no longer clear for a large detection angle range, i.e. the angle range in which a clear angle estimate is still possible is reduced. In a MIMO radar, the antenna array is often designed by combining an array, a transmit or receive array, that has a large aperture and large antenna spacing in each dimension (azimuth and elevation) with a receive or transmit array that has smaller antenna spacing and a smaller aperture. The array with the large aperture then enables an ambiguous angle estimate with high resolution, while the other array enables an unambiguous angle estimate with low resolution. By combining both estimates, the ambiguities in the high-resolution angle spectrum can be resolved.

However, this approach also requires a relatively high computational effort, especially if the angle estimation is to be accompanied by a resolution of the velocity ambiguities, since a complete two-dimensional angle estimation is then necessary for each velocity hypothesis.

disclosure of the invention

The object of the invention is to create an antenna array that enables high-resolution two-dimensional angle estimation with low computational effort.

This object is achieved according to the invention in that the two straight lines run obliquely to one another and in that the aperture of a first of the two subarrays in the direction of the straight line on which the antennas of this first subarray lie is smaller than the aperture of the second subarray in the direction of the same straight line.
The antenna array according to the invention therefore has only two linear subarrays, which are not oriented orthogonally to one another, but rather run at an angle to one another. Accordingly, the planes in which a one-dimensional angle estimation takes place are also not orthogonal to one another, with the result that the result of the angle estimation in one of these planes also contains angular information about the location angle of the object in the other plane.

The radar sensor can be installed in a vehicle in such a way that both subarrays lie in a vertical plane and the straight line on which the first subarray lies forms a smaller angle with the horizontal than the other straight line. The second subarray is then primarily sensitive to the elevation angle because of the steeper course of the associated straight line. Since the objects that must be detected by a radar sensor in a vehicle scatter over a much smaller angular range in elevation than in azimuth, the angular range in which clear angle estimates are possible does not need to be as large for this subarray as for the other array, and antenna elements and computing effort can therefore be saved by thinning out this subarray. Due to its large aperture and its inclination, this second subarray also provides high-resolution angular information in the direction of the first subarray. Although this high-resolution angular information is ambiguous, this ambiguity can be resolved using the first subarray, which is primarily sensitive to the azimuth angle and, due to its smaller aperture, has a low resolution but a larger unambiguous angular range.

In this way, it is possible to carry out unambiguous angle estimates with high angular resolution in both dimensions using only two linear subarrays, both of which only need to have a limited number of antenna elements. Since only the signals from these two subarrays need to be evaluated, the computational effort is considerably reduced.

Advantageous embodiments and further developments of the invention emerge from the subclaims.

The two subarrays do not need to be equidistant. In an advantageous embodiment, however, the antennas are each located on a grid whose grid spacing corresponds to the antenna spacing, which enables clear angle estimates in the required angle range. In this case, the angle estimation can be carried out computationally efficiently using fast Fourier transformation (FFT).

In one embodiment, the subarray located on the steeper straight line is a transmit array formed by transmit antennas, while the other subarray is a receive array.

It is advantageous if the receiving array is a ULA (Uniform Linear Array), since then several targets can already be separated by evaluating the angle for which this subarray is sensitive.

In a special embodiment, the receiving antennas are located on a horizontal line, so that the receiving array is directly sensitive to the azimuth angle. The transmitting array is then sensitive to a mixed angle due to its inclination, which contains both a component in the azimuth direction and a component in the elevation direction.

The angle between the lines on which the two subarrays lie is preferably 45° or less.

The invention also relates to a radar sensor which has the antenna array according to the invention and an associated control and evaluation device.

A chirp sequence method can be implemented in the control and evaluation device. The transmitting antennas can be operated in time multiplex or code multiplex. The resulting speed ambiguity can be resolved according to the method described in DE 10 2018 221 085 A1 Since this method requires an angle estimate to be carried out for each of the several velocity hypotheses, the antenna array according to the invention and the resulting reduction in the computational effort in the angle estimation enables a considerable increase in efficiency.

In a particularly advantageous embodiment of the method, the angle estimates based on the transmit array and the receive array are carried out in separate steps. Since the phase differences that arise in the receive array are independent of the speed of the object, the angle estimate only needs to be carried out once in this step. Only the angle estimate based on the transmit array needs to be carried out separately for each speed hypothesis. However, the computational effort can be further reduced by carrying out object detection after the first angle estimate based on the receive array, so that objects can be separated from one another based on the first angle estimate and thus fewer multi-target scenarios arise when estimating the angle based on the transmit array.

• 1 ein Blockdiagramm eines Chirp Sequence MIMO-Radarsensors;
• 2 Ein Flussdiagramm für ein in dem Radarsensor nach 1 ausführbares Verfahren nach dem Stand der Technik.
• 3 ein erfindungsgemäßes Antennenarray in einer Frontansicht;
• 4 das Antennenarray nach 3 und ein vom Radarsensor geortetes Objekt in einer Ansicht aus Richtung der Pfeile III in 2;
• 5 ein Diagramm zur Erläuterung einer Umrechnung der mit dem Antennenarray nach 2 und 3 erhaltenen Winkelinformation in Azimut- und Elevationswinkel und
• 6 Ein Flussdiagramm für ein in dem erfindungsgemäßen Radarsensor auszuführendes Verfahren.

In the following, an example of implementation is explained in more detail using the drawing. They show:

• 1 a block diagram of a Chirp Sequence MIMO radar sensor;
• 2 A flow chart for a radar sensor according to 1 executable method according to the state of the art.
• 3 an antenna array according to the invention in a front view;
• 4 the antenna array after 3 and an object located by the radar sensor in a view from the direction of arrows III in 2 ;
• 5 a diagram to explain a conversion of the antenna array to 2 and 3 obtained angle information in azimuth and elevation angles and
• 6 A flow chart for a method to be carried out in the radar sensor according to the invention.

Based on 1 and 2 First, a state-of-the-art fast-chirp MIMO radar system is explained, in which transmission signals are encoded using phase modulation. 1 shows a schematic of the structure of the radar system and its control and evaluation device, whereby, to simplify the illustration, only one receiving channel with one receiving antenna is shown instead of several receiving channels. 2 shows corresponding procedural steps.

In step S10, transmitted signals are encoded. In this process, sequences of identical signals in the form of frequency ramps are generated for all transmitting antennas 10 by an RF oscillator 14, which is controlled by a frequency modulation device 16. In each of the several transmitting channels, a respective phase modulator 18, which is connected upstream of an amplifier 19, modulates the phases of the signals 12 according to a respective code, which is generated by a code generator 22. The phase-modulated signal is emitted via a transmitting antenna 10 of the transmitting channel.

A "fast chirp" frequency modulation scheme with a sequence of relatively "fast" frequency ramps is used so that the evaluation of distance d and velocity v can be carried out essentially independently of each other, for example by means of two-dimensional Fourier transformation. In particular, the Doppler shift within a ramp can be neglected.

1 shows a summary of the codes of the individual transmitting antennas 10 in a code block 26. The code block 26 assigns a code value A, B, C, ... of the relevant code to each individual signal of a transmitting antenna 10. The individual code value defines a phase with which the phase modulator 18 modulates the signal. For each position within the code, the code block 26 defines a relevant code value for each of the transmitting antennas 10. The number of codes in a code block 26 corresponds to the number of transmitting antennas that are transmitting simultaneously. In a sequence of code instances, the phase modulation runs through the code values of the relevant code for each transmitting antenna 10. The code blocks are repeated identically. The codes in a code block 26 are orthogonal to one another. The signals of the individual transmitting antennas 10 are thus encoded by the codes; the transmitted signals are orthogonal to each other to enable signal separation in the receiving channel.

After the transmitted signals 12 are reflected by a radar object 28, the signal 32 received by the relevant receiving antenna 30 is mixed in each receiving channel with the non-phase-modulated signal of the RF oscillator 14 in a mixer 34 and brought into a low-frequency range. An A/D conversion is then carried out in the usual way by an A/D converter 36.

The received signal contains temporally shifted, and in the case of a relatively moving radar object 28 also frequency-shifted, superimposed reflections of the transmitted phase-modulated signals 12. The received signals 32 corresponding to the different code instances are separated in a demultiplexer 38. For each code instance, a 2D FFT 40 is then calculated (step S20) by Fourier transformation in a first dimension over the course of the individual ramp-shaped signal and in a second dimension over the sequence of successive code blocks 26.

The resulting 2D spectrum corresponds to a distance-velocity space in which detected radar objects appear as a complex amplitude of the spectrum. Due to the time interval between consecutive identical code instances, i.e. between identical positions in consecutive code blocks, the speed is only determined within a uniqueness range. The uniqueness range, i.e. the width of the uniqueness range, is determined by the repetition rate of the code blocks.

In a complex 2D spectrum obtained, an object detection device 42 detects radar objects 28 (step S21) based on peaks of the spectrum, i.e. based on the position of maxima at the corresponding distance and speed positions. The detection can be carried out, for example, using a non-coherent integration, such as a summation of the absolute values, of the complex 2D spectra of the individual code instances. This non-coherent integration can combine information from the partial measurements corresponding to the code instances, so that the detection is improved.

For a detected radar object 28, an estimated value for the distance d of the radar object, corresponding to a distance estimate (step S22), and an estimated value for the periodically ambiguous speed v_amb, corresponding to an estimate of the Doppler shift of the signals (step S24), are calculated from the position of the relevant peak in the (integrated) two-dimensional spectrum by a speed estimator 43. The radar system is designed for a speed measurement range that exceeds the width v_u of the unambiguous range and can, for example, be a multiple of the width v_u of the unambiguous range. The true speed v of the radar object within the speed measurement range for which the radar system is designed can be equal to v_amb or differ from v_amb by an integer multiple of v_u, for example corresponding to one of the values v_amb - vu, v_amb, v_amb + vu, v_amb + 2 vu.

Based on the ambiguous velocity estimate v_amb, the signals are then further processed in different processing branches 44 for different ambiguity hypotheses, whereby a Doppler shift is assumed in each case corresponding to the “true” velocity of the radar object 28 that results for the respective ambiguity hypothesis.

Based on the respective assumed ambiguity hypothesis for the speed estimate v_amb, i.e. for the respective hypothetical "true" speed of the radar object 28, a Doppler compensation device 46 calculates Doppler shift-compensated signals for the respective signals of the code instances (step S30). For this purpose, a phase compensation is carried out for the complex amplitude of the peak corresponding to the radar object 28 in the respective 2D spectrum, which corresponds to a compensation of the Doppler shift of the phase expected for the respective code instance. Depending on the position of the code instance in the transmitted signal sequence, a corresponding Doppler shift of the phase expected for the hypothetical speed is compensated.

The Doppler shift-compensated signals of the code instances are decoded by a decoder 48 (step S32) by multiplying the signal vector by a decoding matrix. The result of the decoding is a vector of the signal components assigned to the different transmitting antennas 10.

The signal components are fed to an angle estimator 50, which, by comparison with signal values expected for the respective angles, provides a result of an angle estimation (both in azimuth and in elevation) as well as a quality of the angle estimation (step S34). For example, an angle spectrum can be obtained in which the height of a peak at an angle of the quality of the Estimation of this angle as the correct heading angle of the radar object.

The results of the respective angle estimates and the associated qualities of the angle estimates are thus obtained in the processing branches 44. The steps S30 to S34 are thus carried out for the respective ambiguity hypotheses.

A selection device 52 selects the angle estimate with the highest quality from the results of the angle estimates (step S36). The corresponding ambiguity hypothesis is then determined to be correct, and a unique estimate is determined for the relative speed v of the radar object 28 within the speed measurement range for which the radar system is designed (step S38).

Thus, a distance estimate, an angle estimate and a unique speed estimate are obtained for the radar object 28. Steps S22 to S38 are carried out for each detected radar object 28, since a separate Doppler estimate must be made for each radar object.

The method described above can be considerably simplified if the transmitting antennas 10 and the receiving antennas 30 form a planar antenna array 54, which, for example, has the 3 shown. In this example, four transmitting antennas 10 form a linear subarray, which will be referred to below as the transmitting array Tx. All transmitting antennas 10 lie on a straight line G. Four receiving antennas 30 form a further subarray, which will be referred to below as the receiving array Rx. The receiving array Rx is also linear, ie the receiving antennas 30 lie on a straight line Y.

The entire antenna array 54 is arranged on the surface of a circuit board 56 which, when the radar sensor is installed in a motor vehicle, is oriented such that its side edges run vertically and its upper and lower edges run horizontally. The straight line Y runs parallel to the upper and lower edges of the circuit board 56 and thus also runs horizontally in the installed position, while the straight line G runs obliquely and forms an angle of slightly less than 45° with the straight line Y.

The receiving array Rx is a ULA, so that clear angle estimates in azimuth can be made over the entire detection angle range of, for example, -80° to +80°, but only with relatively low resolution, since the aperture of the receiving array Rx is relatively small. The angle spectra for these angle estimates can be carried out very efficiently using FFT.

The transmitting array Tx has a much larger aperture and extends over the entire diagonal of the rectangular board 56. The distances between the phase centers of the transmitting antennas 10 are uneven, but lie on a regular grid, the grid points of which can, for example, have the same distances as the distances between the receiving antennas 30. One of the receiving antennas 30 is also located on the grid of the transmitting array Tx.

The phase relationships between the signals transmitted by different transmitting antennas 10 and received with the same receiving antenna 30 depend on the location angle of the respective located object in the direction in which the straight line G runs.

4 shows the board 56 in a side view from the direction of arrows III in 3 . The viewing direction is perpendicular to the line G, which in 4 is also shown. A straight line X does not pass through the intersection of the straight lines G and Y and is perpendicular to the plane of the board. This straight line X defines the optical axis of the radar sensor. Furthermore, in 4 a located object 28 is shown. The line of sight from the coordinate origin (intersection of the lines G, Y and X) to the object 28 forms an angle δ with the line X, which can be measured using the transmitting array Tx. This means that the transmitting array is sensitive to angles in the plane spanned by the lines G and X. If the radar measurement for the object 28 results in the distance d, the distance of this object from the line X is equal to d sin(δ).

In 5 the field of view of the radar sensor for a given object distance d cos(δ) (in the direction X) is given in a coordinate system with coordinate axes y and z. The axis y corresponds to the straight line Y in 3 and the axis z is the vertical at the origin. The line g in 5 is the projection of the line G into 3 on the field of view. The line g is provided with a length scale that indicates the distance of the points on this line from the x-axis.

It is assumed that the width of the field of view in azimuth ranges from -80° to +80° and the height of the field of view in elevation ranges from -20° to + 40°. The y-coordinates of the left and right edges of the field of view are then d sin(-80°) and d sin(+80°), and correspondingly the z-coordinates of the lower and upper edges of the field of view are d sin(-20°) and d sin(+40°). The object 28 is positioned in the plane spanned by the lines G and X. plane at the angle δ. If α is the azimuth angle of the object and ε is the elevation angle, the Cartesian coordinates of the object 28 in 5 given by y = d sin(α) and z = d sin(ε). Using these relationships, the angle measured by the transmit array Tx in the GX plane can be converted into the azimuth angle α and the elevation angle ε.

The distances between the transmitting antennas 10 in the transmitting array Tx determine the angular range in the direction of the straight line G in which the measurements of the angle δ are unambiguous. These distances are chosen so that the unambiguous range (along the straight line g in 5 ) covers the entire height of the field of view. The measurements of the elevation angle ε are therefore unambiguous. The measurements of the azimuth angle α, on the other hand, are ambiguous because the unambiguous range on the straight line G does not cover the entire width of the field of view. The measurement of the azimuth angle α based on the transmit array Tx is therefore ambiguous but high-resolution. In addition, the azimuth angle α can also be measured with the receive array Rx. These measurements have a lower resolution due to the smaller apparatus, but they are unambiguous and therefore make it possible to select the correct azimuth angle from the high-resolution ambiguous values. In this way, high-resolution and unambiguous angle data is ultimately obtained in both azimuth and elevation.

The based on 2 The procedure described for resolving the velocity ambiguity and for two-dimensional angle estimation can now be replaced by a less complex procedure that is 6 is shown as a flow chart. The steps S10 to S22 in 6 are identical to the corresponding steps in 2 . However, step S22 is now followed by a step S22a of angle estimation based on the receiving array Rx. Since the signals transmitted by the active transmitting antenna 10 and then received by the receiving antennas 30 are simultaneously reflected by the object, the phase progression from receiving antenna to receiving antenna is only dependent on the azimuth angle and not on the speed of the object. Knowledge of the relative speed is therefore not yet required at this point. The angle estimation in step S22a requires only a single FFT per object.

In step S22b, object detection is then carried out again in order to separate objects that differ in azimuth angle, so that the number of multi-target scenarios that require more complex handling is reduced.

The following steps S24 to S32 are again identical to the corresponding steps in 2 , except that the number of targets may have increased due to object separation.

In step S34a, an angle estimate is then made for each speed hypothesis on the basis of the transmit array Tx. In contrast to step S34 in 2 However, this angle estimation requires only a single FFT, which requires little computational effort. The larger the number of velocity hypotheses, the greater the saving in computational effort.

In steps S36 and S38, the angle estimate with the highest quality is then selected and a unique value for the relative velocity is determined.

QUOTES INCLUDED IN THE DESCRIPTION

This list of documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions.

Cited patent literature

• DE 102018221085 A1 [0007, 0021]

Claims (14)

Antenna array for a radar sensor with angle resolution in azimuth and elevation, with two linear subarrays (Tx, Rx) of antennas (10, 30) which are arranged on two non-parallel straight lines (Y, G), characterized in that the two straight lines (Y, G) run obliquely to one another and that the aperture of a first (TX) of the two subarrays in the direction of the straight line (Y) on which the antennas of this first subarray lie is smaller than the aperture of the second subarray (RX) in the direction of the same straight line (Y). antenna array according to claim 1 , in which the subarray (Tx) on the steeper straight line (G) in the installation position of the radar sensor has a larger aperture than the other subarray (Rx). antenna array according to claim 1 or 2 , in which the antennas (10, 30) of at least one of the two subarrays (Tx, Rx) are located on a regular grid. antenna array according to claim 3 , in which at least one of the two subarrays (Tx, Rx) is a ULA. Antenna array according to one of the preceding claims, in which the subarray (Tx) lying on the steeper straight line (G) is a transmitting array formed by transmitting antennas (30), while the other subarray (Rx) is a receiving array. Antenna array according to one of the preceding claims, in which the straight line (Y) on which the subarray (Rx) with the smaller aperture lies runs horizontally in the installed position of the radar sensor. Antenna array according to one of the preceding claims, wherein the angle (γ) between the two straight lines (Y, G) is 45° or less. Radar sensor, in particular for motor vehicles, with an antenna array (54) according to one of the preceding claims. radar sensor after claim 8 , with a control and evaluation device designed for a chirp sequence radar measurement method. radar sensor after claim 9 , in which the control and evaluation device is designed for a method in which the transmitting antennas (10) are operated in multiplex and ambiguities resulting therefrom in the measurement of the relative speed are resolved in the course of an angle estimation. Radar sensor according to one of the Claims 8 until 10 , in which the control and evaluation device is designed for a method in which an unambiguous estimate of a first angle with a small aperture from the first subarray and an estimate of a mixed angle from the second subarray are made, wherein the components in the mixed angle from the first angle are detected ambiguously and the two angles are offset against one another in such a way that the result is an unambiguous and high-resolution estimate of the first angle and an unambiguous estimate of the second angle. radar sensor after claim 10 or 11 in which the subarray (Rx), which is less inclined to the horizontal in the installation position of the radar sensor, is a receiving array and in which an angle estimate is made on the basis of the receiving array independently of the ambiguous speed determination. radar sensor after claim 12 , in which object detection and object separation takes place after the angle estimation based on the receiving array (Rx) and before the ambiguous velocity determination. Motor vehicle with a radar sensor according to one of the Claims 8 until 13 .

Images