Radar distance spectrum generation method

Abstract

Disclosed in the present application is a radar distance spectrum generation method, comprising: on the basis of a preset sparse frequency point sequence, controlling a radar to transmit frequency points and receive signals; performing a fourth-order virtual operation on the received signals, so as to obtain a virtual frequency point sequence; performing merging and sorting according to elements having the same virtual numbers, so as to obtain a denoised virtual frequency point sequence; and selecting some continuous virtual numbers to form an observation vector, and performing windowed Fourier transform to obtain a radar distance spectrum. The present invention improves robustness and the signal-to-interference-plus-noise ratio of a virtual frequency point.

Description

A radar range spectrum generation method

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on October 24, 2023, with application number CN202311377188.7 and application name “A Method for Generating Radar Range Spectrum”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present invention relates to the technical field of radar signal processing, and in particular to a radar range spectrum generating method. Background Art

[0003] With the widespread application of radar for target perception in emerging fields such as autonomous driving and smart homes, the requirements for range resolution are constantly increasing. How to effectively improve range resolution is a key focus of radar target perception.

[0004] Based on radar target resolution characteristics, resolution can be improved by reducing the radar pulse duration or increasing the radar signal scanning bandwidth. Since reducing the pulse duration reduces target echo energy and reduces radar detection range, increasing the radar operating bandwidth is a more common and effective method for improving radar resolution. However, due to relevant constraints or environmental electromagnetic interference, the available radar operating spectrum is limited. For example, the available bandwidth in the 24 GHz unlicensed band does not exceed 250 MHz, the available bandwidth in the 5.8 GHz unlicensed band does not exceed 125 MHz, and the available bandwidth in the 2.4 GHz unlicensed band does not exceed 83.5 MHz. This restricted operating frequency band limits radar target resolution, which in turn affects the target perception experience in actual applications, severely restricting the application of radar in narrow bandwidth bands.

[0005] Existing approaches to improving radar range resolution primarily rely on compressed sensing technology for sparse restoration imaging. These methods require real-time solving of complex optimization problems, and the algorithms are susceptible to robustness issues due to observation perturbations and noise. Furthermore, virtual expansion methods, widely used in sparse array direction finding, can also be applied to improve radar range resolution. In this case, the target of virtual expansion is the radar bandwidth, not the array aperture. While the underlying principles are similar, there are significant differences in the specific implementation of direction finding and ranging. In direction finding, multiple array antennas are arranged at sparse intervals. Cost and physical size constraints limit the number of array elements that can be deployed. Therefore, when designing sparse array arrangements, the fundamental goal is to reduce redundancy after virtual expansion, aiming to maximize the spatial degrees of freedom after virtualization within the constraints of a limited number of array elements. These approaches, such as minimum redundancy array design or nested array design, are often employed. Sparse array virtualization schemes in direction finding solutions are characterized by low virtual redundancy and typically discard redundant virtual results, resulting in a low signal-to-interference-noise ratio (SINR) for a single virtual result.

[0006] Summary of the Invention

[0007] The present invention aims to solve the robustness problem of existing solutions for improving radar resolution and the problem of low signal-to-interference-noise ratio of virtual frequency points, and proposes a radar range spectrum generation method.

[0008] The technical solution adopted by the present invention to solve the above technical problems is:

[0009] A radar range spectrum generation method, the method comprising:

[0010] Step 1: Control the radar transmission frequency according to the preset sparse frequency sequence;

[0011] Step 2: Receive the echo signal reflected by the object in the environment to obtain the received signal;

[0012] Step 3: Perform a fourth-order virtual operation on the received signal to obtain a virtual frequency point sequence, where each element in the virtual frequency point sequence has a virtual number;

[0013] Step 4: determining the frequency of each virtual number in the virtual frequency point sequence, deleting the elements corresponding to the virtual numbers whose frequencies are lower than a frequency threshold, and obtaining an optimal virtual frequency point sequence;

[0014] Step 5: Merge the elements with the same virtual number in the preferred virtual frequency point sequence, and sort all the merged results according to the virtual number to obtain a noise reduction virtual frequency point sequence;

[0015] Step 6: For the noise reduction virtual frequency point sequence, select a continuous part of the virtual number to form an observation vector, and perform a windowed Fourier transform on the observation vector to obtain a radar range spectrum.

[0016] Furthermore, the preset sparse frequency sequence is S={z1,z2,…,z n ,…,z N}, the vector of the received signal is r={r(1),r(2),…,r(n),…,r(N)}, the sparse frequency sequence S is consistent with the nth frequency f actually transmitted by the radar n The relationship is:

[0017] f n =f0+z n f step ;

[0018] Among them, N is the number of frequency points, f0 is the starting frequency of the radar carrier, and f step is the unit frequency shift step value, z n is the nth frequency point in the sparse frequency point sequence S, r(n) is the nth element in the vector r, that is, the nth frequency point z in the sparse frequency point sequence Sn The corresponding received signal is n∈[1,2,…,N].

[0019] Furthermore, if the vector r of the received signal is a real number vector, before performing the fourth-order virtual operation on the received signal, the method further includes:

[0020] Performing Hilbert transform processing on each element in the vector r of the received signal in the slow time dimension to obtain a complex vector of the received signal, and performing a fourth-order virtual operation on the complex vector of the received signal.

[0021] Furthermore, the expression of the Hilbert transform processing is as follows:

[0022] Hilbert{r 1 (n),r 2 (n),…,r k (n),…,r K (n)};

[0023] Where Hilbert{*} represents the Hilbert transform processing operation, K represents the number of detections since the radar was turned on to the current moment, r k (n) represents the nth element in the vector r of the received signal obtained by the kth radar detection, k∈[1,2,…,K], n∈[1,2,…,N].

[0024] Furthermore, the step 3 specifically includes:

[0025] Four elements are selected from the vector r of the received signal in an orderly and repeatable manner to form a quaternary sequence, fourth-order virtual operation results of all permutable quaternary sequences and their corresponding virtual numbers are calculated, and all fourth-order virtual operation results are combined into a virtual frequency point sequence.

[0026] Furthermore, in step 3, the fourth-order virtual operation formula is as follows:

[0027] in, is the fourth-order virtual operation result with virtual number u obtained by using r(m), r(n), r(p) and r(q), u=z m +z n -z p -z q , z m ,z n ,z p ,z q∈S, r(m), r(n), r(p), r(q)∈r, r(m) is the m-th element in vector r, r(n) is the n-th element in r, r(p) is the p-th element in vector r, r(q) is the q-th element in vector r, m, n, p, q∈[1, 2, …, N], the superscript * indicates the conjugate operation.

[0028] Furthermore, the step 4 specifically includes:

[0029] selecting only one of the multiple elements in the virtual frequency point sequence that have the same virtual number and an identity relationship and retaining it in the virtual frequency point sequence to obtain a coarsely selected virtual frequency point sequence;

[0030] In the roughly selected virtual frequency point sequence, the number of elements corresponding to each virtual number is used as the frequency of the virtual number, and the elements corresponding to the virtual numbers whose frequencies are lower than the frequency threshold are deleted to obtain a preferred virtual frequency point sequence;

[0031] Among them, multiple elements with the same virtual number and identity relationship refer to: if one element in the virtual frequency point sequence is but and are elements with the same virtual number and an identity relation.

[0032] Furthermore, in step 5, the method for merging elements with virtual numbers u in the virtual frequency point sequence preferably includes:

[0033] Calculate the average value of multiple fourth-order virtual operation results corresponding to the virtual number u, and use the average value as the element merging result corresponding to the virtual number u.

[0034] Furthermore, in step 6, before performing the windowed Fourier transform on the observation vector, the following steps are further included:

[0035] Perform slow-time low-pass filtering on each element in the observation vector. The operation of the slow-time low-pass filtering is as follows:

[0036] Among them, Lowpass (*) represents the low-pass filter operation, K represents the number of detections since the radar was turned on to the current moment, It represents the merged result of the elements corresponding to the virtual number u obtained by the k-th radar detection calculation, k∈[1,2,…,K].

[0037] Furthermore, in step 6, the windowing sequence for performing windowed Fourier transform on the observation vector is a Hanning window, a Hamming window, a Blackman window, a Chebyshev window, a Kaiser window, a Gaussian window, a Bartlett window, a triangular window or a rectangular window.

[0038] The present invention has the following beneficial effects: The radar range spectrum generation method described herein applies fourth-order virtual expansion technology to radar ranging. It merges redundant frequency points after virtual expansion rather than discarding them directly, fully utilizing the redundant virtual frequency points, suppressing high-frequency interference distributed on the virtual frequency points after the fourth-order expansion operation, and improving the signal-to-interference-and-noise ratio of the virtual frequency points. Furthermore, the present invention does not require real-time solution of complex optimization problems and is highly robust to the effects of noise or disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] FIG1 is a schematic flow chart of a radar range spectrum generating method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0041] The present invention aims to suppress interference components after fourth-order virtual expansion operations and improve the signal-to-interference-noise ratio of virtual frequency points. A radar range spectrum generation method is proposed. The main technical solution includes: controlling radar transmission frequency points according to a preset sparse frequency point sequence; receiving echo signals reflected by objects in the environment to obtain a received signal; performing a fourth-order virtual operation on the received signal to obtain a virtual frequency point sequence, wherein each element in the virtual frequency point sequence has a virtual number; determining the frequency of each virtual number in the virtual frequency point sequence, deleting elements corresponding to virtual numbers with a frequency lower than a frequency threshold to obtain a preferred virtual frequency point sequence; merging elements with the same virtual number in the preferred virtual frequency point sequence, and sorting all the merged results according to the virtual number to obtain a de-noised virtual frequency point sequence; and selecting continuous parts of the virtual number in the de-noised virtual frequency point sequence to form an observation vector, and performing a windowed Fourier transform on the observation vector to obtain a radar range spectrum.

[0042] During radar ranging, sparsely arranged frequency points can be transmitted in a time-division manner through a single antenna. In the scenario of the intelligent Internet of Things, a relatively large number of sparse frequency points can be deployed and transmitted within the coherence time. The number of sparse frequencies is no longer a core constraint. When designing the sparse frequency point arrangement, low virtual redundancy can be used as a constraint, and even a certain amount of virtual frequency point redundancy can be introduced to improve the signal-to-interference-to-noise ratio of the virtual frequency points. Based on this, the present invention controls the radar transmitting frequency points by designing a sparse frequency point sequence, and applies the fourth-order virtual extension technology to radar ranging. A virtual frequency point sequence is obtained by performing a fourth-order virtual operation on the radar echo signal, and elements with frequencies lower than the frequency threshold are eliminated to obtain a preferred virtual frequency point sequence. Redundant merging is performed on elements with the same virtual number in the preferred virtual frequency point sequence, and redundant virtual frequencies are fully utilized, thereby suppressing the high-frequency interference distributed on the virtual post-frequency points after the fourth-order correlation operation, further improving the signal-to-interference-to-noise ratio of the virtual frequency points, and finally selecting the continuous part of the virtual number to form an observation vector and performing a Fourier transform on it after windowing to obtain the radar range spectrum.

[0043] Example

[0044] Referring to FIG1 , a radar range spectrum generation method according to an embodiment of the present invention includes the following steps:

[0045] Step 1: Control the radar transmission frequency points according to the preset sparse frequency point sequence S.

[0046] In this embodiment, the preset sparse frequency sequence is S={z1,z2,…,z n ,…,z N}, sparse frequency sequence S and radar actual transmitting frequency f n The relationship is:

[0047] f n =f0+z n f step

[0048] Among them, N is the number of frequency points, f0 is the starting frequency of the radar carrier, and f n is the nth frequency point actually transmitted by the radar, z n is the n elements in the sparse frequency sequence S, n∈[1,2,…,N], f step is the unit frequency shift step value.

[0049] Since this embodiment merges the redundant results mapped to the same virtual frequency point, when designing the sparse frequency point sequence S, it is first required that the number of effective frequencies (effective redundancy) of the virtual frequency points must be higher than the set virtual frequency point frequency lower limit N. min-redun , thereby ensuring the redundant merging effect. If the radar available bandwidth is B, in order to improve the range resolution and make full use of the limited bandwidth, it is also necessary to require (z N-z1)f step =B,z N and z1 are the maximum values ​​z in the sparse frequency sequence S respectively. N And the minimum value z1. Unit frequency shift step value f step The sum of the number of frequency points N needs to be systematically designed in combination with the hardware operating characteristics of the radar equipment and the distance to the target in the detection environment. This is the a priori input information when designing the sparse frequency point sequence S in this embodiment. Based on the above analysis, this embodiment designs the sparse frequency point sequence S through the following optimization problem:

[0050] Given a unit frequency shift step value f step The number of frequency points N and the radar available bandwidth B are used to design the sparse frequency sequence S = {z1, z2, ..., z n ,…,z N}, maximizing the continuous portion of the virtual numbering that is sufficiently redundant.

[0051] Among them, the continuous part of the virtual number that is fully redundant is: for all u∈J=[-N c ,-N c +1,…,-1,0,1,…,N c -1,N c ], there are at least N min-redun There are unequal value sequences {z m ,z n ,z p ,z q}, so that u=z m +z n -z p -z q , but there is no value sequence {z t ,z y ,z i ,z j} so that u=z t +z y -z i -z j =N c +1, or there is {z t ,z y ,z i ,z j} so that u=z t +z y -z i -z j =N c +1 holds true, but makes u=z t +z y -z i -z j =N c+1 holds true for the unequal value sequence {z t ,z y ,z i ,z j The number of} is less than N min-redun Then J is the continuous part of the virtual number generated by the sparse frequency sequence S with sufficient redundancy, z t ,z y ,z i ,z j ∈S,z m ,z n ,z p ,z q ∈S.

[0052] In this embodiment, for a given sequence {z m ,z n ,z p ,z q}, except {z n ,z m ,z p ,z q}、{z m ,z n ,z q ,z p} and {z n ,z m ,z q ,z p}, a 4-element sequence consisting of elements selected from the sparse frequency sequence S is combined with {z m ,z n ,z p ,z q} are not equivalent. According to the subsequent step 4, the value sequence {z m ,z n ,z p ,z q}、{z n ,z m ,z p ,z q}、{z m ,z n ,z q ,z p} and {z n ,z m ,z q ,z p}The fourth-order virtual operation result and the virtual number corresponding to it are always the same. In order to ensure the effective frequency number of the fourth-order virtual operation result under each virtual number, when designing the sparse frequency point sequence S, the number of value sequences is counted according to the number of unequal value sequences.

[0053] The above maximization problem is an integer combination optimization problem, which is difficult to solve analytically. It can be solved by using an enumeration optimal or heuristic suboptimal algorithm, which is not limited in this embodiment.

[0054] Step 2: Receive the echo signal reflected by the object in the environment to obtain the received signal.

[0055] In step 1 above, the radar transmits N frequency points and receives the echo signals reflected by objects in the environment, and obtains the corresponding received signal vector r = {r(1), r(2), ..., r(n), ..., r(N)}, where r is an N × 1 vector and the n elements r(n) in r are the frequency points z n The received signal, z n Represents the nth element in the sparse frequency sequence S, n∈[1,2,…,N].

[0056] Step 3: Perform a fourth-order virtual operation on the received signal to obtain a virtual frequency point sequence, where each element in the virtual frequency point sequence has a virtual number.

[0057] If the radar receiver is a single-channel (single I-channel receiving or single Q-channel receiving) receiver, the vector r of the received signal is a real vector. Directly performing a fourth-order virtual operation on the real vector will produce a large amount of intermodulation interference, which will deteriorate the signal-to-interference-noise ratio of the virtual frequency. If the vector r of the received signal is a real vector, in order to reduce the impact of intermodulation interference, this embodiment also includes performing Hilbert transform processing on each element of the vector r of the received signal in the slow time dimension to obtain the complex vector rc of the received signal, and performing a fourth-order virtual operation on the obtained rc. More specifically, after performing Hilbert transform processing on the nth element in r in the slow time dimension, the nth element rc(n) of the complex vector rc is obtained, wherein, for the nth element in r, the implementation method of the slow time dimension Hilbert transform processing is as follows:

[0058] Hilbert{r 1 (n),r 2 (n),…,r k (n),…,r K (n)};

[0059] Wherein, Hilbert{*} represents the Hilbert transform processing operation, which is an existing public technology. Specifically, the discrete Fourier transform method or the Hilbert (transform) filter method can be used. This embodiment will not be repeated. K represents the number of detections since the radar was turned on to the current moment, r k (n) represents the nth element in the vector r of the received signal obtained by the kth radar detection, k∈[1,2,…,K], n∈[1,2,…,N].

[0060] If the radar receiver uses an orthogonal reception scheme, the received signal vector r is a complex vector, and fourth-order virtual operations can be performed directly on the received signal vector r. The subsequent processing operations for the two complex vectors are identical, so in the following descriptions, the vector r is used uniformly to represent the complex vector of the received signal.

[0061] This embodiment sequentially and repeatedly selects four elements from the vector r to form a quaternary sequence, calculates the fourth-order virtual operation result and the corresponding virtual number of each permutable quaternary sequence, and forms all the fourth-order virtual operation results into a virtual frequency point sequence.

[0062] For example, for the four-element sequence {r(m), r(n), r(p), r(q)}, the virtual number u=z can be obtained. m +z n -z p -z q The fourth-order virtual operation result

[0063] Where r(m), r(n), r(p), r(q)∈r, r(m) is the m-th element in vector r, r(n) is the n-th element in r, r(p) is the p-th element in vector r, r(q) is the q-th element in vector r, m, n, p, q∈[1, 2, …, N], and the superscript * indicates the conjugate operation.

[0064] After the fourth-order virtual operation, multiple fourth-order virtual operation results with the same virtual number can be obtained, and for the fourth-order virtual operation results generated by the same virtual number and unequal value combinations, the noise and high-frequency interference components they contain are different.

[0065] All fourth-order virtual operation results are combined into a virtual frequency point sequence. Each element in the virtual frequency point sequence has a virtual number, and multiple elements can have the same virtual number.

[0066] Step 4: Determine the frequency of each virtual number in the virtual frequency point sequence, delete the elements corresponding to the virtual numbers whose frequencies are lower than the frequency threshold, and obtain an optimal virtual frequency point sequence.

[0067] This embodiment first performs frequency statistics pre-processing on the virtual frequency point sequence to remove repeated elements with the same virtual number and identity relationship in the virtual frequency point sequence. but and are elements with the same virtual number and an identity relation.

[0068] To prevent the same interference and noise from coherently accumulating and deteriorating the signal-to-interference-plus-noise ratio of the virtual frequency point in the subsequent step 5, in this embodiment, according to the relationship judgment rule, only one of the multiple elements in the virtual frequency point sequence that have the same virtual number and an identity relationship is selected and retained in the virtual frequency point sequence, thereby obtaining a coarsely selected virtual frequency point sequence.

[0069] In the coarse selection virtual frequency point sequence, it is assumed that there are N R (u) kind of value sequence {z m ,z n ,z p ,z q} satisfies u=z m +z n -z p -z q , then the frequency of virtual number u is N R (u). If N R (u) is lower than a preset threshold, the element with the virtual number u in the coarsely selected virtual frequency point sequence is discarded, and the preferred virtual frequency point sequence is obtained after deleting the low-frequency virtual numbers.

[0070] The preset threshold is usually the same as N in step 1. min-redun Corresponding.

[0071] Step 5: Merge the elements with the same virtual number in the preferred virtual frequency point sequence, and sort the merged results according to the virtual number to obtain a noise reduction virtual frequency point sequence.

[0072] In this embodiment, N of the preferred virtual frequency points are R The method for merging (u) elements with virtual number u includes: calculating N R The average value of the fourth-order virtual operation results with (u) virtual numbers u is used as the merging result of the elements with the virtual number u, that is, the estimated value of the virtual frequency point.

[0073] Step 6: For the noise reduction virtual frequency point sequence, select a continuous part of the virtual number to form an observation vector, and perform a windowed Fourier transform on the observation vector to obtain a radar range spectrum.

[0074] Specifically, based on the continuous part J of the virtual number determined in step 1, the elements with virtual number u∈J are selected in the denoised virtual frequency point sequence, and the observation vector is composed according to the increasing order of the virtual number u. The observation vector is subjected to a windowed Fourier transform to obtain the distance spectrum.

[0075] In order to improve the signal-to-interference-noise ratio of the observation vector, this embodiment further includes performing a slow-time low-pass filter on each element in the observation vector after obtaining the observation vector. The operation of the slow-time low-pass filter is as follows:

[0076] Among them, Lowpass (*) represents the low-pass filter operation, K represents the number of detections since the radar was turned on to the current moment, It represents the merged result of the elements corresponding to the virtual number u obtained by the k-th radar detection calculation, k∈[1,2,…,K].

[0077] The slow-time low-pass filtering operation can further filter out various intermodulation interferences, and performing slow-time low-pass filtering on each element in the observation vector can further improve the signal-to-interference-noise ratio of the observation vector.

[0078] After low-pass filtering the observation vector, the radar range spectrum can be obtained by performing a windowed Fourier transform on the observation vector. The windowing sequence can be one of the Hanning window, Hamming window, Blackman window, Chebyshev window, Kaiser window, Gaussian window, Bartlett window, triangular window and rectangular window.

[0079] In summary, the radar range spectrum generation method provided by this embodiment applies fourth-order virtual expansion technology to radar ranging, virtually expanding the radar detection bandwidth and improving the range resolution, achieving better results than the traditional second-order cumulant method. This embodiment designs the distribution of sparse frequencies under the constraints of a given number of frequency points, sweep width, and virtual frequency point frequency, merges the redundant frequency points after virtual expansion, fully utilizes the redundant virtual frequency points, suppresses the high-frequency interference distributed on the virtual frequency points after the fourth-order expansion operation, and improves the signal-to-interference-noise ratio of the virtual frequency points. In addition, this embodiment does not require real-time solution of complex optimization problems and is highly robust to the effects of noise or disturbances.

Claims

1. A radar range spectrum generation method, wherein: The method comprises: Step 1: Control the radar transmission frequency point according to the preset sparse frequency point sequence; Step 2, receiving the echo signal reflected by the object in the environment to obtain a received signal; Step 3: Perform a fourth-order virtual operation on the received signal to obtain a virtual frequency point sequence, wherein each element in the virtual frequency point sequence has a virtual number; Step 4: determining the frequency of each virtual number in the virtual frequency point sequence, deleting the elements corresponding to the virtual numbers whose frequencies are lower than the frequency threshold, and obtaining a preferred virtual frequency point sequence; Step 5: merge the elements with the same virtual number in the preferred virtual frequency point sequence, and sort all the merged results according to the virtual number to obtain a noise reduction virtual frequency point sequence; Step 6: For the noise reduction virtual frequency point sequence, select the continuous part of the virtual number to form an observation vector, and perform a windowed Fourier transform on the observation vector to obtain a radar range spectrum.

2. The radar range spectrum generating method according to claim 1, wherein: The preset sparse frequency point sequence is S={z1,z2,…,z n ,…,z N }, the vector of the received signal is r = {r(1), r(2), ..., r(n), ..., r(N)}, the sparse frequency sequence S is consistent with the nth frequency f actually transmitted by the radar n The relationship is: n =f0+z n f step ; Where N is the number of frequency points, f0 is the radar carrier starting frequency, and f step is the unit frequency shift step value, z n is the nth frequency point in the sparse frequency point sequence S, r(n) is the nth element in the vector r, that is, the nth frequency point z in the sparse frequency point sequence S n The corresponding received signal, n∈[1,2,…,N].

3. The radar range spectrum generating method according to claim 2, wherein: If the vector r of the received signal is a real number vector, before performing the fourth-order virtual operation on the received signal, the method further includes: Perform Hilbert transform processing on each element in the vector r of the received signal in the slow time dimension to obtain a complex vector of the received signal, and perform a fourth-order virtual operation on the complex vector of the received signal.

4. The radar range spectrum generating method according to claim 3, wherein: The expression of the Hilbert transform process is as follows: 1 (n),r 2 (n),…,r k (n),…,r K (n)}; Among them, Hilbert{*} represents the Hilbert transform processing operation, K represents the time since the radar was turned on. The number of detections at the previous moment, r k (n) represents the nth element in the vector r of the received signal obtained by the kth radar detection, k∈[1,2,…,K], n∈[1,2,…,N].

5. The radar range spectrum generating method according to claim 2, wherein: The step 3 specifically includes: Four elements are selected in an orderly and repeatable manner from the vector r of the received signal to form a quaternary sequence, fourth-order virtual operation results of all permutable quaternary sequences and their corresponding virtual numbers are calculated, and all fourth-order virtual operation results are combined into a virtual frequency point sequence.

6. The radar range spectrum generating method according to claim 5, wherein: In step 3, the fourth-order virtual operation formula is as follows: in, is the fourth-order virtual operation result with virtual number u obtained by using r(m), r(n), r(p) and r(q), u=z m +z n -z p -z q , z m ,z n ,z p ,z q ∈S, r(m), r(n), r(p), r(q)∈r, r(m) is the m-th element in vector r, r(n) is the n-th element in r, r(p) is the p-th element in vector r, r(q) is the q-th element in vector r, m, n, p, q∈[1, 2, …, N], the superscript * indicates the conjugate operation.

7. The radar range spectrum generating method according to claim 6, wherein: The step 4 specifically includes: Among multiple elements in the virtual frequency point sequence that have the same virtual number and an identity relationship, only one is selected and retained in the virtual frequency point sequence to obtain a roughly selected virtual frequency point sequence; In the roughly selected virtual frequency point sequence, the number of elements corresponding to each virtual number is used as the frequency of the virtual number, and the elements corresponding to the virtual number whose frequency is lower than the frequency threshold are deleted to obtain a preferred virtual frequency point sequence; Among them, multiple elements with the same virtual number and identity relationship refer to: if one element in the virtual frequency point sequence is but and are elements with the same virtual number and an identity relation.

8. The radar range spectrum generating method according to claim 6, wherein: In step 5, the method for merging elements with virtual number u in the virtual frequency point sequence preferably includes: The average value of multiple fourth-order virtual operation results corresponding to the virtual number u is calculated, and the average value is used as the element merging result corresponding to the virtual number u.

9. The radar range spectrum generating method according to claim 6, wherein: In the step 6, before performing the windowed Fourier transform on the observation vector, the step further includes: A slow-time low-pass filter is performed on each element in the observation vector, and the operation of the slow-time low-pass filter is as follows: Among them, Lowpass (*) represents the low-pass filter operation, K represents the number of detections since the radar was turned on to the current moment, It represents the merged result of the elements corresponding to the virtual number u obtained by the k-th radar detection calculation, k∈[1,2,…,K].

10. The radar range spectrum generating method according to claim 1, wherein: In step 6, the windowing sequence for performing windowed Fourier transform on the observation vector is a Hanning window, a Hamming window, a Blackman window, a Chebyshev window, a Kaiser window, a Gaussian window, a Bartlett window, a triangular window or a rectangular window.

Images