An improved DML-based high-resolution ranging method, system and storage medium

Abstract

The application discloses a high-resolution ranging method and system based on improved DML and a storage medium, and belongs to the field of radar signal processing. The application innovatively uses the DML algorithm, redefines the steering vector and the amplitude and phase vector of the DML algorithm, and uses the same for measuring the distance of a target to distinguish multiple targets with similar distances. Specifically, the position of the target and the target distance neighborhood range to be distinguished are initially determined by using a traditional FFT method, and then the DML algorithm is used to finely search the target distance neighborhood range, so that the range of the DML joint search is reduced, and the complexity of the algorithm operation is reduced. Meanwhile, the position of the target is further accurately determined by combining the phase method, so that two or more adjacent targets can be identified, and the accurate measurement of multiple target positions can be realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of radar signal processing, and more particularly, relates to a high-resolution ranging method based on improved DML, a system and a storage medium. BACKGROUND

[0002] Frequency Modulated Continuous Wave (FMCW) radar is increasingly widely used in close-range and high-resolution scenarios due to its small size, light weight, simple structure and high resolution. In the FMCW radar system, the transmitted waveform mostly adopts a multi-period FMCW signal, and the signal processing adopts two-dimensional FFT to obtain the distance and velocity spectrum of the target. In order to improve the measurement accuracy, some improved algorithms based on FFT such as zero padding method and CZT transformation are also proposed. Although these algorithms based on FFT can improve the estimation accuracy, due to the limitation of the available bandwidth of the radar system, they cannot achieve fine resolution of the target. This means that in the case where the distance or velocity of multiple targets is similar, they may be estimated as a single target although they are multiple targets.

[0003] In order to finely resolve each target, the existing processing method generally also needs to use a high-resolution angle measurement algorithm to estimate the target angle. Commonly used high-resolution angle measurement algorithms include MUSIC, ESPRIT and DML. Among them, DML has been widely used in high-resolution angle measurement and is used to distinguish targets with close angles.

[0004] Considering that ranging and angle measurement are essentially based on the idea of multi-frequency estimation, theoretically, the DML algorithm can be applied to high-resolution ranging. However, the existing DML algorithm needs to jointly search the distance of each target, and in the case of a large number of targets, the computational complexity is extremely high, which cannot be practically used. SUMMARY

[0005] In view of the defects and improvement needs of the prior art, the present application provides a high-resolution ranging method based on improved DML, a system and a storage medium, aiming to solve the technical problem that the existing ranging method cannot finely distinguish targets.

[0006] To achieve the above-mentioned purpose, according to the first aspect of the present application, a high-resolution ranging method based on improved DML is provided, comprising the following steps:

[0007] S1, calculating a preliminary target distance by using an FFT algorithm, and determining a target distance neighborhood range to be distinguished in combination with the preliminary target distance and the resolution of the radar;

[0008] S2, defining a steering vector matrix and an amplitude and phase vector to obtain a matrix form of the difference frequency signal;

[0009] S3, obtaining a spectral function of the beat signal matrix according to the DML algorithm;

[0010] S4, performing a multi-dimensional search on the spectral function in the target distance neighborhood range to be resolved, and a multi-dimensional variable when the spectral function reaches a maximum value being a coarse estimated distance of all targets in the target distance neighborhood range to be resolved;

[0011] S5, obtaining a clear steering vector matrix according to the coarse estimated distances of all targets, and obtaining an amplitude and phase vector according to a relationship between the steering vector matrix and the amplitude and phase vector to obtain phase information of the beat signal corresponding to all targets;

[0012] S6, for each target to be resolved, obtaining a phase ambiguity multiple of the corresponding beat signal by using the coarse estimated distance of the target, and obtaining an accurate distance of the target by combining the phase information of the beat signal corresponding to the target.

[0013] Further, in the S1, the target distance neighborhood range to be resolved is

[0014] wherein r' is a preliminary target distance, c is a light speed, and B is a bandwidth of the frequency modulation signal transmitted by the FMCW radar.

[0015] Further, in the S3, the spectral function of the beat signal matrix is:

[0016] lnω(X) = -Tr[(I-P)R(I-P) H ] = -Tr[(I-P)R] = Tr(PR) - Tr(R)

[0017] wherein ω(X) is a probability density function of the beat signal matrix X, Tr(·) represents a trace of a matrix, I represents a unit matrix, P = A(A H A) -1 A H , A is a steering vector matrix, R = XX H , and H represents a conjugate transpose of a matrix.

[0018] Further, in the S6, a phase ambiguity multiple of the beat signal corresponding to the kth target is

[0019]

[0020] wherein round(·) represents rounding to the nearest integer, λ represents a wavelength corresponding to a starting frequency f of the transmitted frequency modulation signal, λ = c / f, represents the coarse estimated distance of the kth target, represents the phase information of the beat signal corresponding to the kth target. ​

[0021] Further, in the S6, the precise distance of the kth target is

[0022]

[0023] To achieve the above object, according to a second aspect of the present application, there is provided a high-resolution ranging system based on improved DML, comprising: a computer readable storage medium and a processor; the computer readable storage medium is used to store executable instructions; the processor is used to read the executable instructions stored in the computer readable storage medium, and execute the high-resolution ranging method based on improved DML as described in the first aspect.

[0024] To achieve the above object, according to a third aspect of the present application, there is provided a computer readable storage medium, which stores a computer program, the program is executed by a processor to implement the high-resolution ranging method based on improved DML as described in the first aspect.

[0025] In general, through the above technical solutions conceived by the present application, the following beneficial effects can be achieved:

[0026] The present application innovatively utilizes the DML algorithm, redefines the steering vector and amplitude phase vector of the DML algorithm, and uses it to measure the distance of the target to distinguish multiple targets with similar distances. Specifically, first, the traditional FFT method is used to preliminarily determine the target position and the target distance neighborhood range to be distinguished, and then the DML algorithm is used to finely search the target distance neighborhood range, thereby reducing the range of DML joint search and reducing the complexity of algorithm operation; at the same time, the phase method is combined to further accurately determine the target position, which not only can identify two or more adjacent targets, but also can realize accurate measurement of multiple target positions. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a flowchart of a high-resolution ranging method based on improved DML provided by the present application;

[0028] Figure 2 is a spectrum diagram of the traditional FFT frequency method;

[0029] Figure 3 is an estimated distance diagram of the traditional FFT frequency method and the frequency phase combination method;

[0030] Figure 4 is a target rough estimated distance diagram obtained by searching the maximum value of the spectrum function in the present application;

[0031] Figure 5 is a target precise estimated distance diagram obtained by combining the phase method in the present application. DETAILED DESCRIPTION​

[0032] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application. In addition, the technical features involved in the various embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0033] In the present application, the terms "first", "second", etc. (if any) in the present application and the accompanying drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0034] Referring to Figure 1 , in combination Figures 2 to 5 , the present application provides a high-resolution ranging method based on improved DML, comprising the following steps:

[0035] S1, using the FFT algorithm to calculate the preliminary target distance, and combining the preliminary target distance and the resolution of the radar to determine the target distance neighborhood range to be resolved.

[0036] In this embodiment, the FMCW radar transmits a sawtooth wave signal, and the difference frequency signal model is as follows:

[0037]

[0038]

[0039]

[0040] where f k , and s k represent the frequency, phase and amplitude of the difference frequency signal corresponding to the kth target, N represents the number of sampling points, B is the bandwidth of the frequency modulation signal transmitted by the FMCW radar, f is the starting frequency of the transmitted frequency modulation signal, K represents K targets with similar distances, r k represents the distance of the kth target, and the maximum distance of the K targets does not exceed the distance resolution of the radar, and w(n) represents a Gaussian white noise sequence.

[0041] This embodiment takes the case of two targets with similar distances as an example, and the distances of the two targets are r0 and r1 respectively, and the received difference frequency signal is:

[0042]

[0043] The target distance calculated by the traditional FFT algorithm in the embodiment is denoted as r', which is coupled by the two target distances close to each other, because the available bandwidth limits the resolution of the radar system, and the fine resolution of the target cannot be achieved. The resolution of the radar is known as:

[0044]

[0045] Then, r0 and r1 satisfy:

[0046]

[0047] In the embodiment, r0 and r1 are taken as the target distance neighborhood range to be resolved.

[0048] S2, define the steering vector matrix and the amplitude and phase vector to obtain the matrix form of the beat signal.

[0049] Specifically, the kth steering vector is defined as:

[0050] a(r k )=[1 exp(j2πf k / N) … exp(j2(N-1)πf k / N)] T

[0051] Wherein, T represents the transpose of the matrix, and then the steering vector matrix can be obtained:

[0052] A=[a(r0) a(r1) … a(r K-1 )]

[0053] Define the amplitude and phase vector as:

[0054]

[0055] According to the steering vector matrix and the amplitude and phase vector, the beat signal can be written in the form of the matrix:

[0056] X=AS+W

[0057] Wherein

[0058] X=[x0 x1 … x N-1 ] T

[0059] W=[w(0) w(1) … w(N-1)] T

[0060] In the embodiment, the steering vector matrix of the two targets is:

[0061] ​A(r0, r1) = [a(r0) a(r1)]

[0062] The amplitude-phase vectors of the two targets are:

[0063]

[0064] S3, obtaining the spectral function of the difference frequency signal matrix according to the DML algorithm.

[0065] In this embodiment, the probability density function of the difference frequency signal matrix X is known as:

[0066]

[0067] wherein Tr(·) represents the trace of the matrix, H represents the conjugate transpose of the matrix, denotes the variance of the Gaussian white noise, and it is considered that:

[0068]

[0069] Substituting into the probability density function can obtain:

[0070]

[0071] wherein P = A(A H A) -1 A H , R = XX H , and I represents the unit matrix.

[0072] Taking the logarithm of the probability density function can obtain:

[0073] lnω(X) = -Tr[(I-P)R(I-P) H ] = -Tr[(I-P)R] = Tr(PR) - Tr(R)

[0074] Therefore, according to the DML algorithm, the maximization of lnω(X) is equivalent to the maximization of Tr(PR).

[0075] S4, performing a multi-dimensional search on the spectral function in the target distance neighborhood range to be distinguished, and the multi-dimensional variable when the spectral function reaches the maximum value is the coarse estimation distance of all targets in the target distance neighborhood range to be distinguished.

[0076] Specifically, the coarse estimation distance of all targets in the target distance neighborhood range to be distinguished is:

[0077]

[0078] wherein argmaxTr(P(r)R) represents the argument value when Tr(P(r)R) reaches the maximum value.

[0079] The embodiment takes the case of two targets with similar distances as an example, and performs a two-dimensional search on the maximum value of the simplified spectral function Tr(P(r)R) in the target distance neighborhood to be distinguished, that is,

[0080] S5, according to the coarse estimated distance of all targets, the definite steering vector matrix is obtained, and the amplitude and phase vector is obtained according to the relationship between the steering vector matrix and the amplitude and phase vector, so as to obtain the phase information of the difference frequency signal corresponding to all targets.

[0081] Specifically, the definite steering vector matrix is obtained according to the coarse estimated distance of the target:

[0082]

[0083] The amplitude and phase vector is obtained according to the relationship between the steering vector matrix and the amplitude and phase vector:

[0084]

[0085] Then the phase information of the difference frequency signal corresponding to the target can be obtained:

[0086]

[0087] Wherein, arg(·) represents the phase value.

[0088] The embodiment takes the case of two targets with similar distances as an example, and then:

[0089]

[0090]

[0091]

[0092] S6, for each target distinguished, the phase ambiguity multiple of the corresponding difference frequency signal is obtained by using the coarse estimated distance of the target, and the accurate distance of the target is obtained by combining the phase information of the difference frequency signal corresponding to the target.

[0093] Specifically, the phase ambiguity multiple of the difference frequency signal corresponding to the kth target:

[0094]

[0095] Wherein, round(·) represents rounding, λ represents the wavelength corresponding to the starting frequency f of the transmitted frequency modulation signal, λ = c / f.

[0096] The accurate distance of the kth target is obtained according to the phase method:

[0097] ​

[0098] In this embodiment, the phase ambiguity multiples corresponding to the two targets are and The final target accurate distances are and

[0099] In this embodiment, in order to more intuitively verify the resolution ability of the method to similar targets, a matlab simulation experiment with two similar targets is designed. The distance r0 of target 1 is set to 0.275m, and the corresponding difference frequency signal amplitude s0 is set to 0.75. The distance r1 of target 2 is set to 0.3m, and the corresponding difference frequency signal amplitude s1 is set to 1. The chirp number of FMCW radar in one frame is 125, the chirp signal repetition period is 86.9us, the starting frequency f of the transmitted signal is 60GHz, the signal bandwidth B is 3GHz, the sampling frequency f is 5MHz, the sampling point number N is 256, the distance resolution AR is 0.05m, and the simulation signal-to-noise ratio is 20dB. In order to make the preliminary measurement result obtained by the traditional FFT algorithm more accurate, the zero padding method is used to improve the measurement accuracy, and the zero padding number M is 2000, that is, the FFT point number is MxN. s

[0100] As shown in Figure 2 , it is a frequency spectrum diagram obtained by using the traditional FFT method, and it can be seen that there is only one peak position in the figure, that is, the traditional FFT method cannot distinguish two targets. As shown in Figure 3 , it is a comparison diagram of the distance estimated by the traditional FFT method and the distance estimated by the phase method, and it can be seen that the frequency combined phase method is more stable and has higher estimation accuracy than the traditional FFT frequency method, but both cannot distinguish two similar targets, and the distance values obtained are the coupling values of the distances of the two similar targets. As shown in Figure 4 , it is the target rough estimation distance obtained by directly searching the maximum value of the spectrum function in the method of the application, and as shown in Figure 5 , it is the accurate distance of the target obtained by combining the phase method in the method of the application, and it can be seen that the distance estimation accuracy of the phase combined method is higher and more stable. At the same time, as shown in Figure 3 , it can be seen that the method of the application can accurately distinguish two similar targets, and the error between the estimated target distance and the true target distance is very small.

[0101] Those skilled in the art can easily understand that the above description is only a preferred embodiment of the application, and is not intended to limit the application. Any modification, equivalent replacement and improvement made within the spirit and principle of the application shall be included in the protection scope of the application.​

Claims

1. A high-resolution ranging method based on improved DML, characterized in that, Includes the following steps: S1. Calculate the preliminary target distance using the FFT algorithm, and then determine the target distance neighborhood range to be distinguished by combining the preliminary target distance with the radar resolution. S2. Define the steering vector matrix and amplitude phase vector to obtain the matrix form of the difference frequency signal; S3. Obtain the spectral function of the difference frequency signal matrix according to the DML algorithm; S4. Perform a multidimensional search on the spectral function within the neighborhood of the target to be distinguished. The multidimensional variable when the spectral function reaches its maximum value is the coarse estimate of the distance of all targets within the neighborhood of the target to be distinguished. S5. Obtain the explicit steering vector matrix based on the coarse estimated distance of all targets, and then obtain the amplitude phase vector based on the relationship between the steering vector matrix and the amplitude phase vector, so as to obtain the phase information of the difference frequency signal corresponding to all targets. S6. For each identified target, the phase ambiguity factor of the corresponding difference frequency signal is obtained by using the coarse estimated distance of the target, and then the precise distance of the target is obtained by combining the phase information of the difference frequency signal corresponding to the target.

2. The high-resolution ranging method based on improved DML according to claim 1, characterized in that, In S1, the target distance neighborhood range to be distinguished is: Where r′ is the initial target distance, c is the speed of light, and B is the bandwidth of the frequency-modulated signal transmitted by the FMCW radar.

3. The high-resolution ranging method based on improved DML according to claim 1, characterized in that, In S3, the spectral function of the difference frequency signal matrix is: lnω(X)=-Tr[(I-P)R(I-P) H ]=-Tr[(I-P)R]=Tr(PR)-Tr(R) Where ω(X) is the probability density function of the difference frequency signal matrix X, Tr(·) represents the trace of the matrix, I represents the identity matrix, and P = A(A H A) -1 A H A is the guiding vector matrix, and R = XX H H represents the conjugate transpose of the matrix.

4. The high-resolution ranging method based on improved DML according to claim 1, characterized in that, In S6, the phase ambiguity factor of the difference frequency signal corresponding to the k-th target for: Where round(·) represents rounding to the nearest integer, and λ represents the wavelength corresponding to the starting frequency f of the transmitted FM signal, λ = c / f. This represents the coarse estimated distance to the k-th target. This represents the phase information of the difference frequency signal corresponding to the k-th target.

5. The high-resolution ranging method based on improved DML according to claim 4, characterized in that, In S6, the precise distance to the k-th target for:

6. A high-resolution ranging system based on improved DML, characterized in that, include: Computer-readable storage media and processors; The computer-readable storage medium is used to store executable instructions; The processor is configured to read executable instructions stored in the computer-readable storage medium and execute the high-resolution ranging method based on improved DML as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the high-resolution ranging method based on improved DML as described in any one of claims 1-5.

Images