Method and device for determining the angle of a target object

By using multiple radar antennas in the host vehicle for signal conversion and correlation processing, and combining the signal model to calculate the angle of the target object, the problems of high angle estimation cost and low resolution in the radar system are solved, achieving high accuracy and multi-objective Angle estimation of resolution.

CN114942408BActive Publication Date: 2024-11-15APTIV TECHNOLOGIES AG
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210137764.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2021-02-17
Filing Date
2022-02-15
Publication Date
2024-11-15
Estimated Expiration
2042-02-15

AI Technical Summary

Technical Problem

In the prior art, target object angle estimation based on radar systems has problems such as high cost, large hardware size, difficulty in installation and limited angle resolution, especially when expensive lidar systems are not used.

Method used

By using multiple radar transmit and receive antennas in the host vehicle, signal transformation and correlation processing are performed, long beam vectors and reference vectors are generated, and the angle of the target object is calculated in combination with the signal model, side lobes and peak widths are reduced, and angle estimation is improved. Spend.

Benefits of technology

It is realized that without increasing the calculation workload and hardware cost, the accuracy and resolution of the angle estimation of the target object are improved, the ambiguity of the angle estimation is reduced, and the angles of multiple target objects can be accurately determined.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114942408B_ABST
    Figure CN114942408B_ABST
Patent Text Reader

Abstract

The present disclosure relates to a method and apparatus for determining the angle of a target object. A method for estimating the angle of an object relative to a vehicle is provided. A transformation of a reflected radar signal is calculated, wherein the result of the transformation depends on the distance relative to the vehicle. A long beam vector is generated for a corresponding distance quantization unit provided by the transformation by rearranging the result of the transformation so that the corresponding long beam vector includes elements of the transformation from all receiver elements of the radar for each distance quantization unit. A reference vector is calculated for each distance quantization unit based on a signal model, the reference vector depending on the motion of the target object relative to the vehicle and being parameterized with respect to the angle of the target object. The long beam vector is correlated with the reference vector, and the angle of the target object is determined based on the correlation result.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to methods and apparatus for estimating an angle of a target object relative to a host vehicle by using a radar system of the host vehicle. Background Art

[0002] Driver assistance systems and autonomous driving rely on reliable perception of the vehicle environment. If objects in the vehicle environment are detected, the angles of these objects relative to the host vehicle (i.e. azimuth and elevation relative to the longitudinal axis of the vehicle) are often important parameters for the proper performance of the assistance systems. For autonomous driving, reliable and precise angle determination is even essential.

[0003] To determine the azimuth and / or elevation of objects within the vehicle environment, a lidar system may be used. Lidar systems provide high angular resolution and accuracy because they utilize a sharp beam to scan the system's instrument field of view. However, lidar systems are expensive, for example, compared to radar systems. Therefore, it is desirable to make angular determinations of objects in the host vehicle environment based solely on a radar system, without the need for an expensive lidar system.

[0004] However, the angle estimation based on radar systems is a very important process that determines the performance of the system. Existing methods for angle estimation based on radar systems are, for example, Fourier transform techniques or iterative adaptive methods (IAA). However, in order to increase the angular resolution, the cost of the required radar system may increase significantly, since a greater number of antennas and, therefore, a larger package size is required. Furthermore, the increased package size of the radar system may prevent the installation of such a system in certain vehicles, for example, in the bumper area. Furthermore, methods like IAA require a high computational effort.

[0005] Artefacts of the antenna configuration used, vehicle integration and signal processing result in a broadening of the response of the "true" angle in the angular response of the signal reflected by the vehicle environment, as well as additional unwanted side lobes. That is, when using known techniques, there may be some ambiguity regarding the correct determination of the angle and limited resolution for separation of multiple target objects. Similar problems may arise in Doppler processing involving broadened responses and side lobe issues. However, these effects are different from those occurring during antenna processing, so they may compensate for each other.

[0006] Therefore, there is a need to provide a method and apparatus that can accurately determine the angle of a target object relative to a host vehicle and resolve multiple target objects by using an inexpensive radar system. Summary of the invention

[0007] The present disclosure provides a computer-implemented method, a computer system, and a non-transitory computer-readable medium.

[0008] In one aspect, the present disclosure relates to a computer-implemented method for estimating an angle of a target object relative to a host vehicle by using a radar system of a host vehicle. The radar system includes at least one radar transmitting element adapted to transmit a radar signal toward the target object and a plurality of antenna receiver elements, or a plurality of radar transmitting elements and at least one antenna receiver element, each antenna receiver element being adapted to receive a radar signal reflected by the target object. According to the method, a transformation of the reflected signal is calculated via a processing unit of the host vehicle, wherein the result of the transformation depends on the distance relative to the host vehicle. Via the processing unit, a long beam vector for each range quantization bin in a set of range quantization bins provided by the transformation is generated by rearranging the result of the transformation so that the corresponding long beam vector includes elements of the transformation from all receiver elements for the corresponding range quantization bin (range bin). Via the processing unit, a reference vector for each range quantization bin in the set of range quantization bins is calculated based on a signal model, the reference vector depending on the motion of the target object relative to the radar system and being parameterized with respect to the angle of the target object. Via the processing unit, the long beam vector is correlated with the reference vector for a predefined range of angles. Finally, via the processing unit, the angle of the target object is determined based on the result of the correlation.

[0009] The angle relative to the host vehicle can be defined relative to the longitudinal axis of the host vehicle (i.e., relative to the direction of travel of the host vehicle in the case of zero steering angle). Furthermore, the angle can be an azimuth angle defined in a plane parallel to the ground on which the vehicle is currently traveling, wherein the origin of the coordinate system can be located at the antenna receiver element of the radar system in order to define the azimuth angle. A vehicle coordinate system can also be used. Additionally or alternatively, an elevation angle can be considered, which is defined in a plane perpendicular to the plane defining the azimuth angle and relative to the same coordinate system origin. For the estimation of the azimuth and / or elevation angle, the transmitting and receiving antennas can form a virtual array that extends in the azimuth and / or elevation plane and whose elements are not at the same azimuth and / or elevation angle.

[0010] Information about relative motion may include the rate of change of distance of the target object, i.e., the change in the distance of the target object from the host vehicle over time. The antennas used to transmit and receive radar signals may include multiple physically separated antenna devices and / or virtual multiple antennas, which may be provided by a single physical device and may be virtually separated mechanically and / or electronically. The number of transmitting antennas and the number of receiving antennas must be selected so that the product of these numbers is at least two.

[0011] The transformation of the reflected signal with respect to the distance may be a one-dimensional Fourier transform that provides a transformation result as a function of the distance quantization unit (i.e., the discretized distance step). Typically, when a Fourier transform is performed numerically on a time-dependent signal, the transform provides a complex value as a function of the discretized frequency (i.e., the frequency quantization unit). However, alternatives to the Fourier transform may be used for the transformation of the reflected signal with respect to the distance, such as other related processes.

[0012] In order to detect the environment of the host vehicle, the radar signal transmitted by the radar transmitting element may have a frequency modulated continuous wave (FMCW) waveform, commonly referred to as a chirp. That is, the transmitted radar signal is frequency modulated on a fast time scale t within the so-called chirp period T. Therefore, the transformation of the reflected signal with respect to the distance can be performed on this fast time scale, thereby providing a transformation value that depends on the discretized distance step size or distance quantization unit.

[0013] Typically, a further Fourier transformation of the reflected chirped radar signal with respect to the Doppler frequency is performed on a slower time scale, which typically results in a so-called range-Doppler spectrum. However, the method according to the present disclosure differs from this standard way of processing reflected chirped radar signals in that only a first transformation with respect to the range is performed before further method steps. Instead of a second Fourier transformation, a long beam vector is generated as a rearrangement of the transformation results and includes information on all receiver elements and chirps for the corresponding range quantization unit.

[0014] Furthermore, a reference vector is calculated based on the signal model, wherein the reference vector comprises the same dimension or number of elements as the long beam vector. The correlation of the beam vector with the reference vector is performed by using a known correlation procedure, for example by calculating an inner product as in a common discrete Fourier transform, as described below. The angle (i.e., azimuth and / or elevation) of the target object relative to, for example, the longitudinal axis of the host vehicle is determined directly based on the correlation, since the signal model comprises the corresponding angles as parameters. For example, the angle can be varied with a predefined step size over a predefined angular range, and for each step size a corresponding result can be calculated for the correlation of the long beam vector with the reference vector.

[0015] Compared to angle estimation known from the background art, i.e. based on Fourier transform techniques or based on iterative adaptive methods (IAA), the method according to the present disclosure provides an angle estimation with suppressed side lobes and a reduced peak width at the angle to be determined. Thus, the method according to the present disclosure significantly reduces the ambiguity regarding the angle estimation and, overall, improves the accuracy of the angle estimation without significantly increasing the required computational effort or the cost of the required hardware. The improvement in the accuracy of the angle estimation also improves the ability to resolve multiple target objects.

[0016] The method according to the present disclosure is performed separately for each distance quantization unit, that is, one distance quantization unit after another. In order to reduce the required calculation time, the focus area of ​​the distance quantization unit can be predefined, that is, a subset of all available distance quantization units. Therefore, the method can be performed only for a subset of the distance quantization units or a subset of other parameters (such as azimuth or elevation).

[0017] The method may include one or more of the following features:

[0018] The transforming of the reflected signal may include a Fourier transform. The correlating of the long beam vector with the reference vector may include calculating an inner product of the long beam vector and the reference vector. The angle of the target object may be identified by finding at least one maximum of the correlation of the long beam vector with the reference vector.

[0019] A region of interest may be defined for an angle range and the long beam vector may be correlated with a reference vector only for the region of interest. An additional angle finding process may be combined with the correlation of the long beam vector with the reference vector in order to determine the angle. The additional angle finding process may be based on an iterative adaptive method.

[0020] The movement of the target object relative to the radar system may be measured by an additional detection unit installed in the host vehicle. Alternatively, information about the movement of the target object relative to the radar system may be derived from previous measurements that have been performed via the radar system.

[0021] According to one embodiment, the transformation of the reflected signal may include a Fourier transformation. In this case, the transformation may require a lower computational effort.

[0022] Furthermore, the correlation of the long beam vector with the reference vector may include calculating an inner product of the long beam vector with the reference vector. This correlation process is a direct means of performing the correlation and requires a low amount of mathematical and computational effort. For example, to calculate the inner product, each element of the long beam vector may be multiplied by the corresponding element of the reference vector, and the result of the correlation may simply be the sum of all these multiplications or products.

[0023] The angle of the target object can be identified by finding at least one maximum of the correlation of the long beam vector with the reference vector. Since it has been shown that the side lobes in the correlation results as a function of angle are strongly suppressed and the peak width within the correlation results is reduced, as described above, a straightforward maximum finding procedure may be sufficient to determine the appropriate angle of the target object from the correlation results. This can again reduce the computational requirements of the method.

[0024] Furthermore, a region of interest may be defined for a range of angles and the long beam vector may be correlated with the reference vector only for the region of interest. As described above, the reference vector is calculated based on an angle as a parameter, wherein the angle varies over a predefined range of angles, for example in steps of 1 degree. If a specific region of interest is known or defined for the angle to be estimated, the angle may be varied only over that region of interest when calculating the reference vector. Thus, the computation time required to calculate the reference vector for different angles may be reduced. The region of interest may be known from measurements provided by other systems of the vehicle or may be predefined based on the configuration of the vehicle.

[0025] According to another embodiment, an additional angle finding process and the correlation of the long beam vector with the reference vector may be combined to determine the angle. In detail, the additional angle finding process may be based on an iterative adaptive method (IAA). It turns out that the above method may be compatible to be combined with other frequency estimation concepts (e.g., iterative adaptive methods). It may also be compatible with subsequent estimation processes that require a complex-valued response (e.g., azimuth elevation estimation as a separate process). The additional angle finding process may further improve the resolution of the angle estimate by further suppressing side lobes and reducing peak width in the correlation results that depend on the angle as a parameter. The iterative adaptive method may focus on certain angle ranges and provide lower side lobes and lower peak widths.

[0026] The motion of the target object relative to the radar system can be measured by an additional detection unit installed in the host vehicle. Typically, the method requires independent determination of the relative motion of the target object. At least one parameter related to the relative motion of the target object is necessary for applying the signal model required to calculate the reference vector. If the relative motion of the target object is measured by an additional detection unit of a vehicle, for example, as part of another component of the driver assistance system, the reliability when calculating the reference vector can be improved because the information about the relative motion is provided by an independent measurement. On the other hand, if the information about the relative motion is obtained from a previous measurement of the radar system that is also used by the method, no additional hardware is required to provide the information about the relative motion. In addition, multiple relative motions of the parameter vector that can extend the signal vector can be considered. This can bring about an increase in the response dimension so that subsequent processing that can handle multiple hypotheses becomes possible.

[0027] In another aspect, the present disclosure relates to an apparatus for estimating an angle of a target object relative to a host vehicle. The apparatus comprises a radar system installed in the host vehicle and comprising at least one radar transmitting element and a plurality of antenna receiver elements adapted to transmit a radar signal toward the target object, or a plurality of radar transmitting elements and at least one antenna receiver element, wherein each antenna receiver element is adapted to receive a radar signal reflected by the target object. The apparatus further comprises a processing unit and the processing unit is configured to: calculate a transformation of the reflected signal, wherein the result of the transformation depends on the distance relative to the host vehicle; generate a long beam vector for each distance quantization unit in a set of distance quantization units provided by the transformation by rearranging the result of the transformation so that the corresponding long beam vector includes elements of the transformation from all receiver elements for the corresponding distance quantization unit; calculate a reference vector for each distance quantization unit in the set of distance quantization units based on a signal model, the reference vector depending on the motion of the target object relative to the radar system and parameterized with respect to the angle of the target object; correlate the long beam vector with the reference vector for a predefined angle range; and determine the angle of the target object based on the result of the correlation.

[0028] As used herein, the terms "processing module" and "processing unit" may refer to, be part of, or include an application specific integrated circuit (ASIC); an electronic circuit; a combinational logic circuit; a field programmable gate array (FPGA); a processor (shared, dedicated, or group) that executes code; other suitable components that provide the functionality; or a combination of some or all of the above, such as in a system on a chip. The term "module" or "unit" may include a memory (shared, dedicated, or group) that stores code executed by a processor.

[0029] In summary, the device according to the present disclosure comprises a radar system and a processing unit configured to perform the steps of the above-described corresponding method. Therefore, the benefits, advantages and disclosures of the above-described method are equally valid for the device according to the present disclosure. The processing unit can be installed in the host vehicle, but as an alternative, it can also be an external unit that may be associated with, for example, a cloud system.

[0030] The processing unit may also be configured to perform an additional angle finding process based on an iterative adaptive method. The apparatus may also include an additional detection unit configured to measure the motion of the target object relative to the radar system. Alternatively, the processing unit may also be configured to derive information about the motion of the target object relative to the radar system from previous measurements that have been performed via the radar system.

[0031] In another aspect, the present disclosure relates to a computer system configured to perform some or all of the steps of the computer-implemented method described herein.

[0032] The computer system may include a processing unit, at least one memory unit and at least one non-transitory data storage. The non-transitory data storage and / or the memory unit may include a computer program for instructing a computer to perform several or all steps or aspects of the computer-implemented method described herein.

[0033] In another aspect, the present disclosure relates to a non-transitory computer-readable medium comprising instructions for performing some or all steps or aspects of the computer-implemented methods described herein. The computer-readable medium can be configured as: an optical medium, such as a compact disk (CD) or a digital versatile disk (DVD); a magnetic medium, such as a hard disk drive (HDD); a solid-state drive (SSD); a read-only memory (ROM); a flash memory; etc. In addition, the computer-readable medium can be configured as a data storage unit accessible via a data connection such as an Internet connection. For example, the computer-readable medium can be an online data repository or cloud storage.

[0034] The present disclosure also relates to a computer program for instructing a computer to perform several or all steps or aspects of the computer-implemented method described herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Herein, exemplary embodiments and functions of the present disclosure are described in conjunction with the following drawings, which schematically illustrate:

[0036] Figure 1 depicts a diagram of a radar system used in accordance with the method of the present disclosure,

[0037] Figure 2 depicts the transmitted and reflected radar signals and their processing,

[0038] Figure 3A and Figure 3B The generated long beam vector is depicted,

[0039] Figure 4 depicting generation of reference vectors based on a signal model and relating these reference vectors to long beam vectors, and

[0040] FIG. 5A to FIG. 5C The results according to the method of the present disclosure are compared with the results according to the background art.

[0041] Reference numerals list

[0042] 10 Radar System

[0043] 12 Antenna Array

[0044] 14 Transmitter

[0045] 16 Antenna

[0046] 16A First Antenna

[0047] 16B Second Antenna

[0048] 18 Transmitting radar signals

[0049] 20 Reflecting radar signals

[0050] 24A First Object

[0051] 24B Second object

[0052] 26 Controller

[0053] 27 Processor

[0054] 28 Transmit Signal

[0055] 29 Receiver

[0056] 30 Detected signals

[0057] θ azimuth

[0058] 31 Delay of reflected signal or echo chirp

[0059] 41 Distance Quantization Unit

[0060] 43 Long beam vector

[0061] 45 Reference vector

[0062] 47 Related

[0063] 49 Distance change rate

[0064] 51 Angle Quantization Unit

[0065] 53 Curves found by Fourier transform angles

[0066] 55 Curves found through the perspective of iterative adaptive method (IAA)

[0067] 57 Side Lobe

[0068] 59 Max.

[0069] 61 Curves including additional iterative adaptive method (IAA) DETAILED DESCRIPTION

[0070] Figure 1A vehicle-based radar system 10 is depicted. The system 10 includes an antenna array 12 that includes at least one transmit element 14 and an array of at least two receive elements, also referred to as a plurality of antennas 16. One or more of the receive elements or antennas 16 may also be used to transmit radar signals 18, and they output detected signals 30 indicative of reflected radar signals 20 reflected by a first target object 24A or a second target object 24B in an instrument field of view of the system 10. The transmit element 14 and the plurality of antennas 16 are shown as distinct elements in this example simply to simplify explanation of the system 10.

[0071] The system 10 includes a controller or processing unit 26 configured to output a transmit signal 28 to the transmit element 14 and to collect receive signals 30 from each antenna 16, such as a first signal 30A from the first antenna 16A and a second signal 30B from the second antenna 16B. Each of the detected signals 30 corresponds to a reflected radar signal 20 detected by one of the plurality of antennas 16. The controller 26 includes a processor 27, such as a microprocessor, a digital signal processor, or other control / signal conditioning circuitry, such as analog and / or digital control circuitry including an application specific integrated circuit (ASIC) for processing data. The controller 26 also includes a memory (not shown) including a non-volatile memory, such as an electrically erasable programmable read-only memory (EEPROM) for storing one or more routines, thresholds, and captured data.

[0072] In addition, the controller 26 also includes a receiver 29 configured to receive antenna signals (e.g., first signal 30A and second signal 30B) from respective antennas (e.g., first antenna 16A and second antenna 16B), the antenna signals corresponding to the reflected radar signals 20 collected from respective antennas of the plurality of antennas 16. The controller 26 may include a mixer (not shown) and a local oscillator (not shown) to demodulate the detected signal 30. The mixer and the local oscillator may be part of the receiver 29.

[0073] The transmitting element 14 radiates or sends out a radar signal 18 toward a first target object 24A and / or a second target object 24B in the instrument field of view 22, and each of the plurality of antennas 16 detects a reflected radar signal reflected by the first target object 24A and / or the second target object 24B. The characteristics of the reflected radar signal 20 depend on the backscattering characteristics or radar cross section (RCS) of the first target object 24A or the second target object 24B. These characteristics also depend on the distance, direction, and relative motion of the first object 24A and / or the second object 24B relative to the antenna array 12, which affects the Doppler shift of the reflected radar signal 20.

[0074] The first target object 24A is located at a distance or distance r1 relative to the antenna array 12 of the system 10 and therefore relative to the host vehicle (not shown) on which the system 10 is installed, while the second target object 24B is located at a distance or distance r2 relative to the antenna array 12 of the system 10. The relative position difference between the first target object 24A and the second target object 24B is shown as Δrx and Δry. In addition, the second target object 24B is located at an azimuth angle θ relative to the longitudinal axis of the host vehicle or relative to the normal of the antenna array 12. Since the first object 24A is aligned with the normal, the azimuth angle of the first object 24A is almost zero and therefore is not in the normal. Figure 1 The method according to the present disclosure focuses on accurately estimating the azimuth angle θ of objects in the host vehicle's environment.

[0075] Figure 2 The frequency of a transmitted radar signal 18 and a reflected radar signal 20 over time and the processing of these signals according to the prior art are schematically depicted. Figure 2 The schematic diagram is taken from Sun et al.: MIMO Radar for Advanced Driver Assistance Systems and Autonomous Driving, IEEE Signal Processing Magazine, vol. 37, issue 4, pp. 98-117, July 29, 2020. The transmitted radar signal 18 and the reflected radar signal 20 have FMCW (frequency modulated continuous wave) waveforms, also known as chirps. That is, both signals include complex sine waves whose frequency increases linearly within the time interval T according to the following formula:

[0076]

[0077] Where B is the signal bandwidth, f c is the carrier frequency. For the carrier frequency, a typical frequency band of 76 to 81 GHz is used in automotive radar systems, such as Figure 1 The system 10 is shown.

[0078] The reflected radar signal 20 (also called the echo chirp) is a delayed and attenuated copy of the transmitted signal or chirp 18. For a target object 24A, 24B located at a distance R and moving at a radial velocity v, the delay time 31 (see Figure 2 ) is given by:

[0079] τ=(2(R+vt)) / c (2)

[0080] Wherein time t runs over a number of time periods T of the chirp signal 18, 20 and c is the speed of light.

[0081] Typically, the reflected radar signal 20 is mixed with the transmit signal 18 or transmit chirp, which results in a complex sine wave called a beat signal. The beat signal has a frequency f b =f R +f D , where f R =2RB / (Tc) is the so-called range frequency, f D =(2v / c)f c is the so-called Doppler frequency.

[0082] For the processing of the beat signal, the time during one period or chirp is usually referred to as fast time, and the time across multiple periods T or chirps is referred to as slow time. When sampling the beat signal, the samples of each chirp or antenna 16 are arranged into columns of a matrix, where the row index of the matrix corresponds to the fast time and the column index corresponds to the slow time, such as Figure 2 As shown. D Usually much smaller than f R , so f D can be considered as a constant within each chirp. Therefore, a Fourier transform is usually applied to the sampled beat signal along fast time for each antenna or channel 16, which allows identification of the respective distances of the target objects 24A, 24B relative to the antenna array 12 used as a reference point at the host vehicle. In detail, the distance or distance of the objects 24A, 24B is obtained as:

[0083] R = cf R T / (2B) (3).

[0084] To obtain the velocity of the objects 24A, 24B, a second Fourier transform is typically then performed along slow time. Application of the first and second Fourier transforms is equivalent to a two-dimensional Fourier transform of the beat signal in fast and slow time, and the result is typically referred to as a range-Doppler spectrum.

[0085] According to the background art, an angle estimation can also be performed based on the range-Doppler spectrum. For the angle estimation, the angle is defined relative to the longitudinal axis of the host vehicle and is an azimuth and / or elevation angle. Furthermore, the angle estimation according to the background art can be based on a further Fourier transformation or on an iterative adaptive method (IAA).

[0086] However, such angle estimation based on the range-Doppler spectrum mostly requires high computational and / or hardware costs. In addition, the angle estimation may be subject to some ambiguity due to the side lobes generated by the corresponding angle finding process. In order to overcome these drawbacks, the method according to the present disclosure is based on a different approach that does not rely on the range-Doppler spectrum. After performing the first Fourier transform as described above at a fast time related to the range determination, the result of the first Fourier transform is rearranged so that for each range quantization unit (i.e., the discretized distance generated by the first Fourier transform) a vector running over all chirps or antenna channels 16 is generated (see Figure 1 ). This vector is called the long beam vector and carries the antenna information as well as Doppler information about the target object's velocity.

[0087] exist Figure 3A and Figure 3B The generation of long beam vectors 43 for each range quantization unit 41 is schematically depicted in FIG. Figure 3A The generation of the range-Doppler spectrum is again illustrated, which is shown as depending on the beam vectors of the corresponding range quantization unit and Doppler quantization unit (according to Figure 2 ). For example, the beam vector is based on the result of a two-dimensional Fourier transform performed on the beat signal at fast time and slow time, as described above.

[0088] In contrast, Figure 3B The corresponding beam vectors after a first Fourier transformation in fast time are schematically depicted, which include the corresponding chirp or antenna channel 16 (see Figure 1 ) distance quantization unit. To generate the long beam vector 43, the beam vector elements of all chirps are rearranged one by one over all chirps to generate a vector called the long beam vector 43, which is Figure 3B The lower part is shown schematically.

[0089] For the angle estimation based on the long beam vector 43, a reference vector 45 is calculated based on the signal model (see Figure 4 ), which reference vector 45 takes into account the motion of the target object 24A, 24B relative to the radar system 10 (i.e., relative to the antenna 16 installed in the vehicle) and depends on the angle to be determined, i.e., the azimuth and / or elevation. The concept is explained in detail for a uniform linear array (ULA) and a linear relative motion of one of the target objects 24A, 24B. For more complex antenna arrangements and more complex object motions, the signal model can be extended as known in the background art.

[0090] The signal model is based on the following formula:

[0091]

[0092] Where T is the slow time represented by the slow time index relative to the first chirp. n is the slow time index relative to the first antenna 16A (see Figure 1 )’s antenna channel index. is the rate of change of the distance of the corresponding target object 24A, 24B in m / s. α is the electric angle defined for the interval from -1 to +1, where α=sinθ and θ is the mechanical angle to be determined in rad, where θ=0 defines the visual axis. a is the complex amplitude. The corresponding coefficient k is defined by the following formula:

[0093]

[0094]

[0095] Among them, T chirp is the Doppler sampling period or chirp period in seconds, T Space is the spatial sampling period in m, which corresponds to the spacing of the antennas 16. The factor 2 included in the formula for the coefficients is due to the bidirectional propagation of the radar signal, i.e., from the corresponding antenna 16 to one of the target objects 24A, 24B and back (see Figure 1 ).

[0096] Figure 4 The method steps of generating a reference vector depending on the angle θ and of associating the reference vector with the long beam vector 43 are schematically depicted in . The angle θ is used as a parameter and is defined for a predetermined angle range. For the present example, the angle range extends from -90° to +90° and the angle θ varies in steps of 1 degree. Ie, 181 reference vectors are generated based on the signal model in order to cover the entire angle range. In order to calculate the corresponding reference vector 45, the rate of change of the distance of the target objects 24A, 24B must be given from further measurements. The relative movement of the target objects can be determined via a further measuring device or detection unit installed in the host vehicle or can rely on previous measurements performed by the radar system, such as Figure 1 As shown, a further slow-time-based Fourier transform is performed on it in order to determine the velocity.

[0097] Based on a given distance change rate or speed 49 of the respective target object 24A, 24B and based on a preset angle θ as a parameter, elements of the respective reference vector 45 corresponding to respective elements of the long beam vector 43 are calculated.

[0098] For Figure 4 At each angular step size shown in , the corresponding reference vector 45 is correlated with the long beam vector 43. This correlation is given by Figure 447 in FIG. The correlation is performed by computing the inner product of the corresponding reference vector 45 and the long beam vector 43 (i.e., by multiplying corresponding pairs of elements and adding all products of the corresponding elements). The result of this correlation provides the value of the corresponding angle or angle quantization unit (i.e., for this example, the discretized angle step size running from -90° to +90°).

[0099] exist Figure 5B The overall result of correlating the long beam vector 43 with the reference vector 45 over all angular quantization units 51 is depicted in , wherein the angular quantization units are represented on the x-axis and the result of the correlation is represented by the y-axis. Figure 5B , note that the 256 angle quantization units cover the angle range from -90° to +90° to provide Figure 4 Schematic diagram of greater angular resolution ( Figure 4 180 steps of 1 degree have been assumed).

[0100] For each distance quantization unit 41 (see Figure 3A and Figure 3B ) respectively perform iteration of the angle θ over a predefined angle range. That is, a similar Figure 5B FIG. 4 is a diagram showing the correlation results of the long beam vector 43 and the reference vector 45. Figure 5B The angle θ to be determined for the respective distance quantization unit 41 is given by the maximum 59 of the correlation curve, which is located at the angle quantization unit 128 corresponding approximately to the viewing axis (ie 0 degrees).

[0101] In order to reduce the computation time required to perform the method according to the present disclosure, corresponding regions of interest may be predefined for the angle θ and for the distance or distance of the target object 24A, 24B. Such regions of interest may be based on the detection of target motion performed by another system of the host vehicle, which also provides, for example, the rate of change of distance or the speed 49.

[0102] exist Figure 5A , the results of two classical angle finding processes based on the range-Doppler spectrum as described above are shown. Curve 53 represents the result of the angle finding via a further Fourier transform, while curve 55 represents the result of an iterative adaptive method (IAA) as known in the background art. Although one maximum of both curves 53, 55 occurs at the correct angle (angle quantization unit 128) of about 0 degrees, both curves include many additional maxima and strong side lobes, indicated by 57. This makes the angle finding based on the above-mentioned curve 53 and curve 55 appear to be the same as the angle finding of the previous embodiment. Figure 5AThe results shown in the background art complicate the proper angle estimation, especially for angle estimation based on further Fourier transformation. Although the side lobes 57 are reduced for angle finding via IAA, the curve 55 also includes strong maxima at small angles. Therefore, Figure 5A It shows the ambiguity that may arise when applying the process of finding the background technology from a perspective.

[0103] In contrast, Figure 5B The results of the correlation of the long beam vector 43 with the reference vector 45 over all angular quantization units are depicted. Figure 5A In contrast, the sidelobes indicated at 57 are strongly suppressed. In addition, Figure 5B The maximum shown at 59 in FIG. 1 is dependent on a peak having a smaller width, especially when compared to the peaks shown in FIG. Figure 5A The results of angle finding via Fourier transform are shown in comparison.

[0104] For Figure 5C The curve 61 shown, except for the Figure 4 and Figure 5B In addition to the correlation of the long beam vector 53 with the reference vector, an additional angle finding process is applied. In detail, an additional iterative adaptive method (IAA) has been applied to the correlation result. Figure 5C As can be seen in , the side lobe 57 is additionally suppressed and the width of the peak indicating the correct maximum of the angle θ is also greatly reduced, as can be seen at 59. Therefore, in addition to further reducing the ambiguity due to the side lobes, the angular resolution of the method is enhanced by the additional angle finding process, such as Figure 5C shown.

Claims

1. A computer-implemented method for estimating an angle (θ) of a target object (24A, 24B) relative to the host vehicle by using a radar system (10) of the host vehicle, wherein: The radar system (10) comprises: a plurality of antenna receiver elements (16) and at least one radar transmitting element (14) adapted to transmit a radar signal (18) towards the target object (24A, 24B), or a plurality of transmitting elements (14) and at least one antenna receiver element (16), each antenna receiver element (16) being adapted to receive a radar signal (20) reflected by the target object (24A, 24B), The computer-implemented method comprises the following steps: calculating a transformation of the reflection signal (20) via a processing unit (26) of the host vehicle, wherein the result of the transformation depends on the distance relative to the host vehicle, generating, via the processing unit (26), a long beam vector (43) for each range quantization unit (41) in the set of range quantization units provided by the transform by rearranging the result of the transform so that the corresponding long beam vector (43) includes elements of the transform from all receiver elements (16) for the corresponding range quantization unit (41), calculating, via the processing unit (26), a reference vector (45) for each range quantization unit (41) in the set of range quantization units based on a signal model, the reference vector (45) depending on the motion of the target object (24A, 24B) relative to the radar system (10) and parameterized with respect to the angle (θ) of the target object (24A, 24B), correlating, via the processing unit (26), the long beam vector (43) with the reference vector (45) for a predefined range of angles (θ), and The angle (θ) of the target object (24A, 24B) is determined via the processing unit (26) based on a result of the correlation.

2. The computer-implemented method of claim 1 , wherein: The transformation of the reflection signal (20) comprises a Fourier transformation.

3. The computer-implemented method of claim 1 or 2, wherein: The correlating of the long beam vector (43) with the reference vector (45) includes calculating an inner product of the long beam vector (43) and the reference vector (45).

4. The computer-implemented method of claim 1 , wherein: The angle (θ) of the target object (24A, 24B) is identified by finding at least one maximum of the correlation of the long beam vector (43) with the reference vector (45).

5. The computer-implemented method of claim 1 , wherein: A region of interest is defined for said range of angles (θ), and The long beam vector (43) is correlated with the reference vector (45) for the region of interest only.

6. The computer-implemented method of claim 1 , wherein: An additional angle finding process is applied to the correlation of the long beam vector (43) with the reference vector (45) in order to determine the angle (θ).

7. The computer-implemented method of claim 6, wherein: The additional angle finding process is based on an iterative adaptive method.

8. The computer-implemented method of claim 1 , wherein: The movement of the target object (24A, 24B) relative to the radar system (10) is measured by an additional detection unit installed in the host vehicle.

9. The computer-implemented method of claim 1 , wherein: Information about the movement of the target object (24A, 24B) relative to the radar system (10) is obtained from previous measurements that have been performed via the radar system (10).

10. An apparatus for estimating an angle (θ) of a target object (24A, 24B) relative to a host vehicle, the apparatus comprising: A radar system (10) installed in the host vehicle and comprising at least one radar transmitting element (14) adapted to transmit a radar signal (18) towards the target object (24A, 24B), and a plurality of antenna receiver elements (16), each antenna receiver element (16) adapted to receive a radar signal (20) reflected by the target object (24A, 24B), A processing unit (26), wherein the processing unit (26) is configured to: calculating a transformation of the reflected signal (20), wherein the result of the transformation depends on the distance relative to the host vehicle, generating a long beam vector (43) for each range quantization unit (41) in the set of range quantization units provided by the transform by rearranging the result of the transform so that the corresponding long beam vector (43) includes elements of the transform from all receiver elements (16) for the corresponding range quantization unit (41), calculating a reference vector (45) for each range quantization unit (41) in the set of range quantization units based on a signal model, the reference vector (45) depending on the motion of the target object (24A, 24B) relative to the radar system (10) and parameterized with respect to the angle (θ) of the target object (24A, 24B), relating the long beam vector (43) to the reference vector (45) for a predefined range of angles (θ), and The angle (θ) of the target object (24A, 24B) is determined based on a result of the correlation.

11. The device according to claim 10, wherein: The processing unit (26) is further configured to apply an additional angle finding process to the correlation of the long beam vector (43) with the reference vector (45) in order to determine the angle (θ), wherein the additional angle finding process is based on an iterative adaptive method.

12. The device according to claim 10 or 11, wherein: The apparatus further comprises an additional detection unit configured to measure the movement of the target object (24A, 24B) relative to the radar system (10).

13. The device according to claim 10 or 11, wherein: The processing unit (26) is further configured to derive information about the movement of the target object (24A, 24B) relative to the radar system (10) from previous measurements that have been performed via the radar system (10).

14. A computer system configured to perform the computer-implemented method according to any one of claims 1 to 9.

15. A non-transitory computer-readable medium comprising instructions for executing the computer-implemented method according to any one of claims 1 to 9.