An estimation method of a vehicle-mounted millimeter wave radar direction of arrival and a vehicle-mounted millimeter wave radar

Abstract

The application provides an estimation method for a vehicle-mounted millimeter wave radar direction of arrival, all information contained in a sparse virtual array is expressed in the form of a uniform virtual array by using a virtual element interpolation skill, the definition of an atomic norm is introduced, and an optimization problem based on atomic norm minimization is constructed, a covariance matrix satisfying Toeplitz is obtained after the minimization problem is solved, and finally the direction of arrival is estimated by using a MUSIC algorithm; compared with the prior art, all information in the non-uniform virtual array is fully utilized, the problem of lack of rank of the covariance matrix caused by coherent sources is effectively solved, the influence of noise is effectively reduced, and the resolution and estimation accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and more particularly to a method for estimating the direction of arrival of a vehicle-mounted millimeter-wave radar and a vehicle-mounted millimeter-wave radar. Background Technology

[0002] Frequency Modulated Continuous Wave (FMCW) radar, with its advantages of simplicity, high accuracy, and good stability, has been widely used as a type of millimeter-wave radar. During vehicle operation, onboard radar acquires target position information in real time by determining the target's angle of arrival (Angle of Arrival). Direction of Arrival (DoA) estimation, an important branch of array signal processing, refers to using an array antenna to receive spatial signals and then processing the received signals through statistical signal processing and optimization methods to recover the angular information contained in the transmitted signal.

[0003] Traditional angle estimation algorithms suffer from resolution limitations due to the Rayleigh limit of antenna aperture, resulting in less than ideal angle measurement accuracy. When using receiver arrays for signal processing to estimate the angles of multiple targets, the angle resolution is directly related to the array aperture; a larger aperture provides higher angle resolution. However, automotive radar installation locations are limited, and increasing the number of receiver elements increases costs. Current receiver arrays typically consist of only four elements, leading to a sparse array configuration to maximize aperture. However, in virtual domain signal processing, excessive sparsity results in a discontinuous virtual array, causing signal model mismatch. To avoid information loss and improve angle measurement accuracy, we introduce a covariance matrix reconstruction method based on an interpolated virtual array.

[0004] Furthermore, signal sources encountered in real-world environments are often not independent but rather correlated. This can lead to performance degradation or even failure of subspace-based methods widely used in DoA estimation. Therefore, DoA estimation of coherent signal sources is also an important topic in this field. Summary of the Invention

[0005] To overcome the above-mentioned technical defects, the technical solution in this invention uses virtual array element interpolation techniques to represent all the information contained in the sparse virtual array in the form of a uniform virtual array, introduces the definition of atomic norm and constructs an optimization problem based on minimizing atomic norm, solves the minimization problem to obtain the covariance matrix that satisfies Toeplitz, and finally estimates the direction of arrival using the MUSIC algorithm.

[0006] Specifically, the purpose of this invention is to provide a method for estimating the direction of arrival of vehicle-mounted millimeter-wave radar, comprising: A virtual planar array for a vehicle-mounted millimeter-wave radar is established, wherein both the transmitting and receiving arrays of the vehicle-mounted millimeter-wave radar are sparse arrays. Based on the original virtual planar array, the first received signal acquired by the receiving array of the vehicle-mounted millimeter-wave radar is... Perform modeling; Perform a two-dimensional FFT on the first received signal to calculate the target's distance and velocity information; Perform constant false alarm target detection; Select the range-Doppler cell where the target is located, and after detecting the target, process the target's range and velocity information; Based on interpolation techniques and the extracted angular dimension data vector, a first interpolated virtual array is interpolated into the original virtual planar array to obtain an interpolated uniform linear array ULA, and the second received signal Y of the interpolated uniform linear array ULA is obtained. U (f θ ), where the element positions of the interpolated uniform linear array ULA are U={0≤l≤max(S)}; Calculate the sampling covariance matrix of the interpolated uniform linear array ULA; Based on the definition of atomic norm, the covariance matrix T(z) of the interpolated uniform linear array ULA is reconstructed by constructing an atomic norm minimization problem. Calculate the optimal solution Based on the optimal solution The covariance matrix is ​​obtained by the Toeplitz property of the covariance matrix T(z) of the interpolated uniform linear array ULA. ; The angle parameters of the coherent source are obtained through a multiple signal classification method. The spatial spectrum is represented as: in, For the noise subspace of the covariance matrix, The steering vector for the virtual signal; The peak values ​​of the spatial spectrum are obtained through peak search, and the peak values ​​are arranged from largest to smallest. The top peak values ​​are then selected. K The angle value corresponding to each peak is the estimated result of the direction of arrival.

[0007] Preferably, the transmitting array and receiving array of the vehicle-mounted millimeter-wave radar are respectively , The azimuth antenna element spacing is half the wavelength of the incident narrowband signal. The pitch-angle antenna element spacing is half the wavelength of the incident narrowband signal; The vehicle-mounted millimeter-wave radar has one antenna element at the transmitting end and [number] receiver elements. The virtual array is equivalent to this, thus establishing the original virtual planar array of the vehicle-mounted millimeter-wave radar.

[0008] Preferably, the received signal acquired by the receiving end... Modeling includes: have K from θ=[θ1,θ2,...,θ K ] T A far-field narrowband coherent signal with consistent direction, distance, and velocity is incident on the original virtual planar array of the vehicle-mounted millimeter-wave radar, where θ k For the first k The angle between the direction of the target and the axis of the original virtual planar array, the distance of the target is R, the velocity is V, and the transmitter uses a frequency sweep slope. μ The FMCW signal with frequency diversity transmission can then be used for the first... n Each sweep cycle corresponds to a time t Received signal Modeling: in, The attenuation coefficient is... The center frequency of the carrier. T For the frequency sweep period, This represents independent and identically distributed zero-mean additive white Gaussian noise. Indicates noise power. For identity matrix, array manifold The first in k Column corresponds to the first k The steering vectors of the incident signals are as follows: ,in, Represents the first in the original virtual plane array l The location of each array element The imaginary unit is ; the coherent signal waveform vector is ,in for .

[0009] Preferably, the step of performing a two-dimensional FFT on the first received signal to obtain the target's distance and velocity information includes: Performing a two-dimensional FFT on the first received signal in both fast and slow time dimensions to transform the time-domain signal into the range-Doppler domain yields: Here, the slow time dimension refers to both the distance dimension and the velocity dimension. M The number of frequency sweep cycles is given. The sawtooth wave signal of each frequency sweep cycle is stored in a column after performing a one-dimensional FFT. The received signals obtained from different frequency sweep cycles are then stored in different columns after performing another-dimensional FFT, thus forming a two-dimensional data plane composed of different distance-Doppler units.

[0010] Preferably, the constant false alarm rate (CFAR) target detection includes: The measured value of the discrete sampling unit of the first received signal is compared with a threshold value. If the measured value exceeds the threshold value, it is determined that the first received signal contains the signal of the target. Specifically, for the range-Doppler domain, guard intervals and reference windows are set before and after the range and velocity dimensions of the unit to be detected, respectively, to obtain the threshold values ​​for the range and velocity dimensions. and ,in accordance with and Determine the threshold value for two-dimensional detection .

[0011] Preferably, after detecting the target, processing the target's distance and speed information includes: Read the velocity dimension frequency corresponding to each peak in the two-dimensional data plane. and distance dimension frequency The distance and velocity information of the target are calculated according to the following formula. in, The carrier center frequency of the FMCW signal; The extracted angular dimension data vector includes: Based on the target's distance and velocity information, extract the angular dimension data vector corresponding to the range-Doppler unit. .

[0012] Preferably, the second received signal of the interpolated uniform linear array ULA is obtained by interpolating the first virtual array element at the hole position of the original virtual planar array, treating the first virtual array element as an antenna element that is not in operation, and the corresponding received signal is zero. This allows us to obtain the position of the interpolated uniform linear array ULA within the... ld The second received signal of the array element at the location: .

[0013] Preferably, calculating the sampling covariance matrix of the interpolated uniform linear array ULA includes: the expression is as follows. , The sampling covariance matrix Theoretical covariance matrix The maximum likelihood estimate.

[0014] Preferably, the covariance matrix of the interpolated uniform array ULA, reconstructed by constructing an atomic norm minimization problem based on the definition of atomic norm, is... include: Based on the Hermitian Toeplitz property satisfied by the covariance matrix of the interpolated uniform array ULA, the vector The atomic norm is expressed as: in, Represents the atomic norm. Denotes the infimum for the vector. The covariance matrix of the reconstructed interpolated uniform array ULA, when decomposed with the minimum number of atoms, has the highest good fit to the ideal matrix; The optimization problem of reconstructing the covariance matrix of the interpolated uniform array ULA based on minimizing the atomic norm: subject to , , The atomic norm minimization problem is equivalent to the following optimization problem: subject to in, The regularization coefficient is . Describing the Frobenius norm, The binary matrix, the binary vector Used to distinguish between the original virtual planar array and the first virtual array in the interpolated uniform linear array ULA, each element of the binary vector corresponds one-to-one with each array element in the interpolated uniform linear array ULA. When an array element belongs to the original virtual planar array, the binary vector... The element at the corresponding position in the binary vector is 1; if the array element belongs to the first virtual array, the binary vector... The element at the corresponding position in the middle is 0.

[0015] In another aspect, the present invention provides an on-vehicle millimeter-wave radar that uses the method for estimating the direction of arrival of on-vehicle millimeter-wave radar as described above.

[0016] The sparse array vehicle-mounted millimeter-wave radar direction-of-arrival estimation method for coherent sources in this invention, compared with the prior art, makes full use of all the information in the non-uniform virtual array, effectively solves the problem of rank deficiency in the covariance matrix caused by coherent sources, and effectively reduces the influence of noise, improving resolution and estimation accuracy. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the direction-of-arrival estimation method for vehicle-mounted millimeter-wave radar as described in an embodiment of the present invention. Figure 2 The positions of the array elements and holes in the original virtual array used in this embodiment of the invention; Figure 3a This invention provides the normalized spatial spectrum for direction-of-arrival estimation of two coherent sources using the MUSIC algorithm when the signal-to-noise ratio is 20dB and the number of snapshots is 512. Figure 3b The normalized spatial spectrum of the direction-of-arrival estimation of two coherent sources using the algorithm proposed in this invention is shown in an embodiment of the present invention when the signal-to-noise ratio is 20dB and the number of snapshots is 512. Figure 4 This is a comparison of the direction-of-arrival estimation performance of embodiments of the present invention under different signal-to-noise ratios. Detailed Implementation Plan The advantages of the present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments.

[0018] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0019] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0020] In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the convenience of the description of the invention and have no specific meaning in themselves. Therefore, "module" and "part" can be used interchangeably.

[0021] One embodiment of the present invention provides a method for estimating the direction of arrival (DOA) of a sparse array vehicle-mounted millimeter-wave radar oriented towards coherent sources, such as... Figure 1 As shown, the specific steps include the following: S1: Establish the original virtual planar array of the vehicle-mounted millimeter-wave radar, wherein both the transmitting and receiving arrays of the vehicle-mounted millimeter-wave radar are sparse arrays.

[0022] In this embodiment, the transmitting array and receiving array of the vehicle-mounted millimeter-wave radar are respectively , The azimuth antenna element spacing is half the wavelength of the incident narrowband signal. The pitch-angle antenna element spacing is half the wavelength of the incident narrowband signal; the antenna of the vehicle-mounted millimeter-wave radar has one element at the transmitting end and the number of elements at the receiving end. The virtual array is equivalent to this, thus establishing the original virtual planar array of the vehicle-mounted millimeter-wave radar.

[0023] S2: Based on the original virtual planar array, the first received signal collected by the receiver of the vehicle-mounted millimeter-wave radar is... Modeling is performed.

[0024] Specifically, the modeling process includes: There are K elements from θ = [θ1, θ2, ..., θ] K ] T A far-field narrowband coherent signal with consistent direction, distance, and velocity is incident on the original virtual planar array of the vehicle-mounted millimeter-wave radar, where θ k For the first k The angle between the direction of the target and the axis of the original virtual planar array, the distance of the target is R, the velocity is V, and the transmitter uses a frequency sweep slope. μ For FMCW signals with frequency diversity transmission, the received signal at time t corresponding to the nth sweep cycle can be... Modeling: in, The attenuation coefficient is... The center frequency of the carrier. T For the frequency sweep period, This represents independent and identically distributed zero-mean additive white Gaussian noise. Indicates noise power. For identity matrix, array manifold The first in k Column corresponds to the first k The steering vectors of the incident signals are as follows: ,in, Represents the first in the original virtual plane array l The location of each array element The imaginary unit is ; the coherent signal waveform vector is ,in for .

[0025] It should be noted that Kr is the attenuation coefficient of the received signal. Currently, the received signal is directly simulated at the receiving end during the simulation, so the method of this invention does not involve the setting of this parameter.

[0026] S3: Perform a two-dimensional FFT on the first received signal to calculate the target's distance and velocity information.

[0027] Specifically, in this embodiment, a two-dimensional FFT with fast and slow time dimensions is performed on the first received signal to transform the time-domain signal into the range-Doppler domain, resulting in: Here, the slow time dimension refers to both the distance dimension and the velocity dimension. M The number of frequency sweep cycles is given. The sawtooth wave signal of each frequency sweep cycle is stored in a column after performing a one-dimensional FFT. The received signals obtained from different frequency sweep cycles are then stored in different columns after performing another-dimensional FFT, thus forming a two-dimensional data plane composed of different distance-Doppler units.

[0028] S4: Perform constant false alarm rate (CFAR) target detection, select the range-Doppler unit where the target is located, and process the target's range and velocity information after the target is detected.

[0029] The constant false alarm rate (CFAR) target detection includes: comparing the measured value of the discrete sampling unit of the first received signal with a threshold value; if the measured value exceeds the threshold value, it is determined that the first received signal contains the signal of the target. Specifically, for the range-Doppler domain, guard intervals and reference windows are set before and after the range and velocity dimensions of the unit to be detected, respectively, to obtain the threshold values ​​for the range and velocity dimensions. and ,in accordance with and Determine the threshold value for two-dimensional detection .

[0030] Subsequently, the velocity-dimensional frequency corresponding to each peak in the two-dimensional data plane is read. and distance dimension frequency The distance and velocity information of the target are calculated according to the following formula. in, The carrier center frequency of the FMCW signal; The extracted angular dimension data vector includes: extracting the angular dimension data vector corresponding to the distance-Doppler unit based on the target's distance and velocity information. .

[0031] S5: Based on the interpolation technique and the extracted angular dimension data vector, interpolate the first interpolated virtual array into the original virtual planar array to obtain the interpolated uniform linear array ULA, and obtain the second received signal of the interpolated uniform linear array ULA. The element positions of the interpolated uniform linear array ULA are U={0≤l≤max(S)}. Figure 2 This refers to the positions of the array elements and holes in the original virtual array used in this embodiment.

[0032] The second received signal of the interpolated uniform linear array ULA is obtained by interpolating the first virtual array element at the hole position of the original virtual planar array, treating the first virtual array element as an antenna element that is not in operation, and the corresponding received signal is zero. Therefore, the location of the interpolated uniform linear array ULA can be determined. ld The second received signal of the array element at the position: .

[0033] S6: Calculate the sampling covariance matrix of the interpolated uniform linear array ULA, as shown in the following expression: , in, This represents the conjugate transpose operator. Expressing expectations, L The number of snapshots; the sampling covariance matrix Theoretical covariance matrix Maximum likelihood estimation, when the number of sample snapshots L As it approaches infinity, infinitely close to .

[0034] S7: Based on the definition of atomic norm, construct an atomic norm minimization problem to reconstruct the covariance matrix T(z) of the interpolated uniform linear array ULA. The atomic norm minimization problem is modeled as follows: subject to in, The regularization coefficient is . Describing the Frobenius norm, The binary matrix, the binary vector Used to distinguish between the original virtual planar array and the first virtual array in the interpolated uniform linear array ULA, each element of the binary vector corresponds one-to-one with each array element in the interpolated uniform linear array ULA. When an array element belongs to the original virtual planar array, the binary vector... The element at the corresponding position in the binary vector is 1; if the array element belongs to the first virtual array, the binary vector... The element at the corresponding position in the middle is 0.

[0035] Based on the Hermitian Toeplitz property satisfied by the covariance matrix of the interpolated uniform array ULA, the vector The atomic norm is expressed as: in, Represents the atomic norm. Denotes the infimum for the vector. The covariance matrix of the reconstructed interpolated uniform array ULA, when decomposed with the minimum number of atoms, has the highest good fit to the ideal matrix; Specifically, the optimization problem for reconstructing the covariance matrix of the interpolated uniform array ULA based on minimizing the atomic norm is as follows: subject to , , The atomic norm minimization problem is equivalent to the following optimization problem: subject to in, is the regularization coefficient.

[0036] S8: Calculate the optimal solution According to the covariance matrix of the interpolated uniform linear array ULA The Toeplitz property yields the covariance matrix. In this embodiment, the optimization problem obtained in the previous step is solved using the CVX tool to obtain the optimal solution. .

[0037] S9: Obtain the angle parameters of the coherent source through a multiple signal classification method. The spatial spectrum can be represented as: in, Describe the noise subspace of the covariance matrix. The steering vector representing the virtual signal.

[0038] S10: Obtain the peak values ​​of the spatial spectrum through peak search, arrange the peak values ​​from largest to smallest, and take the first one. K The angle value corresponding to each peak is the estimated result of the direction of arrival, and the K value is the number of targets.

[0039] Based on the above embodiments, in practical applications, it is assumed that the following is used: A sparse array is used for both transmission and reception. It is assumed that the number of incident far-field narrowband coherent signals is 2, and the target's range and velocity are set to... The angles of the two coherent targets are set to The sawtooth wave pulse count is 512, the distance dimension sampling count is 512, and the appendix is... Figure 3a , 3b This refers to the comparison of the normalized spatial spectra of two coherent sources using the traditional MUSIC (Multiple Signal Classification) algorithm and the algorithm proposed in this invention, respectively, under a signal-to-noise ratio of 20dB.

[0040] Figure 4 To compare the statistical performance of the traditional MUSIC algorithm and the algorithm proposed in this invention under different signal-to-noise ratios, the criterion for successful detection is that the estimation error of any target does not exceed 1 degree. (See attached...) Figure 3a As can be seen, in the case of extremely sparse arrays and the presence of coherence, the traditional MUSIC algorithm completely fails and cannot successfully distinguish between two targets. However, the algorithm proposed in this invention... Figure 3b Two relatively sharp spectral peaks were obtained, indicating that this algorithm can achieve super-resolution direction-of-arrival estimation in the presence of coherence. As demonstrated in this embodiment, the method proposed in this invention can effectively detect coherent signals, and the estimation accuracy improves with increasing signal-to-noise ratio. Furthermore, the traditional MUSIC algorithm cannot effectively distinguish coherent signals under all signal-to-noise ratio settings.

[0041] In another embodiment of the present invention, a vehicle-mounted millimeter-wave radar is provided, which uses the above-described estimation method. Its features are the same as those of the above-described estimation method, and will not be repeated here.

[0042] In summary, the proposed method for estimating the direction of arrival (DOA) of a sparse array vehicle-mounted millimeter-wave radar for coherent sources can accurately estimate the DOA in scenarios where the source is coherent. This invention obtains a uniform virtual linear array using virtual element interpolation techniques and leverages the correspondence between sparse and uniform arrays. It introduces the definition of the atomic norm and constructs an optimization problem based on minimizing the atomic norm. Solving this minimization problem yields a covariance matrix satisfying the Toeplitz property. Finally, the DOA is estimated using the MUSIC algorithm. This invention fully utilizes all information in the non-uniform virtual array, effectively solving the problem of rank-deficient covariance matrix caused by coherent sources, thus improving resolution and estimation accuracy.

[0043] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for estimating the direction of arrival of a vehicle-mounted millimeter-wave radar, characterized in that, include: A virtual planar array for a vehicle-mounted millimeter-wave radar is established, wherein both the transmitting and receiving arrays of the vehicle-mounted millimeter-wave radar are sparse arrays. Based on the original virtual planar array, the first received signal acquired by the receiving array of the vehicle-mounted millimeter-wave radar is... Perform modeling; Perform a two-dimensional FFT on the first received signal to calculate the target's distance and velocity information; Perform constant false alarm target detection; Select the range-Doppler cell where the target is located, and after detecting the target, process the target's range and velocity information; Based on interpolation techniques and the extracted angular dimension data vector, a first interpolated virtual array is interpolated into the original virtual planar array to obtain an interpolated uniform linear array ULA, and the second received signal Y of the interpolated uniform linear array ULA is obtained. U (f θ ), where the element positions of the interpolated uniform linear array ULA are U={0≤l≤max(S)}; Calculate the sampling covariance matrix of the interpolated uniform linear array ULA; Based on the definition of atomic norm, the covariance matrix T(z) of the interpolated uniform linear array ULA is reconstructed by constructing an atomic norm minimization problem. Calculate the optimal solution Based on the optimal solution The covariance matrix is ​​obtained by the Toeplitz property of the covariance matrix T(z) of the interpolated uniform linear array ULA. ; The angle parameters of the coherent source are obtained through a multiple signal classification method. The spatial spectrum is represented as: in, For the noise subspace of the covariance matrix, The steering vector for the virtual signal; The peak values ​​of the spatial spectrum are obtained through peak search, and the peak values ​​are arranged from largest to smallest. The top peak values ​​are then selected. K The angle value corresponding to each peak is the estimated result of the direction of arrival.

2. The estimation method as described in claim 1, characterized in that, The transmitting array and receiving array of the vehicle-mounted millimeter-wave radar are respectively , The azimuth antenna element spacing is half the wavelength of the incident narrowband signal. The pitch-angle antenna element spacing is half the wavelength of the incident narrowband signal; The vehicle-mounted millimeter-wave radar has one antenna element at the transmitting end and [number] receiver elements. The virtual array is equivalent to this, thus establishing the original virtual planar array of the vehicle-mounted millimeter-wave radar.

3. The estimation method as described in claim 2, characterized in that, The received signal collected by the receiving end Modeling includes: There are K elements from θ = [θ1, θ2, ..., θ] K ] T A far-field narrowband coherent signal with consistent direction, distance, and velocity is incident on the original virtual planar array of the vehicle-mounted millimeter-wave radar, where θ k For the first k The angle between the direction of the target and the axis of the original virtual planar array, the distance of the target is R, the velocity is V, and the transmitter uses a frequency sweep slope. μ For FMCW signals with frequency diversity transmission, the received signal at time t corresponding to the nth sweep cycle can be... Modeling: in, The attenuation coefficient is... The center frequency of the carrier. T For the frequency sweep period, This represents independent and identically distributed zero-mean additive white Gaussian noise. Indicates noise power. For identity matrix, array manifold The first in k Column corresponds to the first k The steering vectors of the incident signals are as follows: ,in, Represents the first in the original virtual plane array l The location of each array element The imaginary unit is ; the coherent signal waveform vector is ,in for .

4. The estimation method as described in claim 3, characterized in that, The step of performing a two-dimensional FFT on the first received signal to obtain the target's distance and velocity information includes: Performing a two-dimensional FFT on the first received signal in both fast and slow time dimensions to transform the time-domain signal into the range-Doppler domain yields: Here, the slow time dimension refers to both the distance dimension and the velocity dimension. M The number of frequency sweep cycles is given. The sawtooth wave signal of each frequency sweep cycle is stored in a column after performing a one-dimensional FFT. The received signals obtained from different frequency sweep cycles are then stored in different columns after performing another-dimensional FFT, thus forming a two-dimensional data plane composed of different distance-Doppler units.

5. The estimation method as described in claim 4, characterized in that, The constant false alarm rate (CFAR) target detection includes: The measured value of the discrete sampling unit of the first received signal is compared with a threshold value. If the measured value exceeds the threshold value, it is determined that the first received signal contains the signal of the target. Specifically, for the range-Doppler domain, guard intervals and reference windows are set before and after the range and velocity dimensions of the unit to be detected, respectively, to obtain the threshold values ​​for the range and velocity dimensions. and ,in accordance with and Determine the threshold value for two-dimensional detection .

6. The estimation method as described in claim 5, characterized in that, After detecting the target, the process of processing the target's distance and speed information includes: Read the velocity dimension frequency corresponding to each peak in the two-dimensional data plane. and distance dimension frequency The distance and velocity information of the target are calculated according to the following formula. in, The carrier center frequency of the FMCW signal; The extracted angular dimension data vector includes: Based on the target's distance and velocity information, the angular dimension data vector corresponding to the range-Doppler unit is extracted. .

7. The estimation method as described in claim 6, characterized in that, The second received signal of the interpolated uniform linear array ULA is obtained by interpolating the first virtual array element at the hole position of the original virtual planar array, treating the first virtual array element as an antenna element that is not in operation, and the corresponding received signal is zero. Therefore, the location of the interpolated uniform linear array ULA can be determined. ld The second received signal of the array element at the location: 。 8. The estimation method as described in claim 7, characterized in that, The calculation of the sampling covariance matrix of the interpolated uniform linear array ULA includes: the expression is as follows. ， The sampling covariance matrix Theoretical covariance matrix The maximum likelihood estimate.

9. The estimation method as described in claim 8, characterized in that, The covariance matrix of the interpolated uniform array ULA, reconstructed by the atomic norm minimization problem based on the definition of atomic norm, is then constructed. include: Based on the Hermitian Toeplitz property satisfied by the covariance matrix of the interpolated uniform array ULA, the vector The atomic norm is expressed as: in, Represents the atomic norm. Denotes the infimum for the vector. The covariance matrix of the reconstructed interpolated uniform array ULA, when decomposed with the minimum number of atoms, has the highest good fit to the ideal matrix; The optimization problem of reconstructing the covariance matrix of the interpolated uniform array ULA based on minimizing the atomic norm: subject to , ， The atomic norm minimization problem is equivalent to the following optimization problem: subject to in, The regularization coefficient is . Describing the Frobenius norm, The binary matrix, the binary vector Used to distinguish between the original virtual planar array and the first virtual array in the interpolated uniform linear array ULA, each element of the binary vector corresponds one-to-one with each array element in the interpolated uniform linear array ULA. When an array element belongs to the original virtual planar array, the binary vector... The element at the corresponding position in the binary vector is 1; if the array element belongs to the first virtual array, the binary vector... The element at the corresponding position in the middle is 0.

10. A vehicle-mounted millimeter-wave radar, characterized in that, The method for estimating the direction of arrival of vehicle-mounted millimeter-wave radar as described in any one of claims 1-9 is used.

Images