Array signal high-precision direction finding method based on two-dimensional improved co-prime array
By introducing compression factors and offset factors into the mutually qualitative array, the array structure is optimized, the virtual array element holes are eliminated and the aperture is expanded, the problem of low direction finding accuracy of the mutually qualitative array is solved, and signal processing with higher accuracy is achieved.
Patent Information
- Application Number
- CN202410963635.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-08
AI Technical Summary
In the prior art, when the mutual array expands the array aperture and improves the direction finding accuracy, there is a problem of waste of resources caused by duplication of virtual array elements, and the direction finding accuracy is relatively low.
By introducing compression factors and offset factors, the array element spacing and position of the subarray are optimized, the holes of virtual array elements are eliminated, the physical aperture of the array is expanded, and the rank loss MUSIC algorithm is applied in two-dimensional space for final DOA estimation.
Under the condition that the number of array elements remains unchanged, the holes of virtual array elements are eliminated, the accuracy and aperture of array direction finding are improved, and the performance of signal processing is significantly improved.
Smart Images

Figure CN118914966A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of array signal processing, and in particular relates to a high-precision direction finding method for array signals based on a two-dimensional improved coprime array. Background Art
[0002] Array signal processing, also known as spatial signal processing, is a very important branch of modern signal processing. It has been widely used and developed rapidly in radar, wireless communication, seismic exploration, biomedicine and other fields. At present, many classic direction-finding algorithms are proposed for uniform arrays. In order to avoid direction-finding angle ambiguity, the spacing between adjacent array elements in the array is often limited to within half the wavelength of the signal. Since the direction-finding accuracy of the array is positively correlated with the array aperture, the traditional uniform array can generally only expand the array aperture by stacking the number of array elements, thereby improving the array performance, but doing so will increase the system cost and complexity. In addition, the tight arrangement of array elements makes the influence of mutual coupling effect non-negligible, resulting in a decrease in array performance.
[0003] The introduction of sparse arrays provides a new idea for solving the above problems. It breaks through the limitation that the array element spacing is no more than half a wavelength. Compared with the traditional full-rank array, it has the advantages of higher direction finding accuracy and smaller mutual coupling effect. However, ordinary sparse arrays cannot resolve ambiguity well. Mutually prime arrays are a main type of sparse arrays. A typical mutually prime array is composed of two sparse uniform sub-arrays arranged in an interlaced manner. The number of array elements and the array element spacing of the two sub-arrays are mutually prime. The mutual prime characteristics of the array structure can be used to eliminate direction finding ambiguity. Under the condition of the same number of array elements, compared with uniform arrays, mutually prime arrays can expand the array aperture, increase the spatial degree of freedom, and improve the accuracy and resolution of spatial spectrum estimation. It is becoming a hot research direction in sparse arrays. However, the virtual array elements of ordinary mutually prime arrays will be repeated, resulting in a waste of resources. Therefore, arranging a set of arrays with a low repetition rate of virtual array elements will be very meaningful for improving the direction finding accuracy of the system. Summary of the invention
[0004] Purpose of the invention: In order to overcome the deficiencies in the prior art, a high-precision direction finding method for array signals based on a two-dimensional improved coprime array is provided. By introducing compression factors and offset factors, the direction finding accuracy is further improved, which is very meaningful for the study of array signal processing.
[0005] Technical solution: To achieve the above object, the present invention provides a high-precision direction finding method for array signals based on a two-dimensional improved coprime array, comprising the following steps:
[0006] S1: Set two mutually prime numbers M1 and M2 as the number of array elements of sub-array 1 and sub-array 2, where M1 = M2-1;
[0007] S2: Introducing a compression factor to compress the element spacing of subarray 1 to eliminate holes in virtual elements;
[0008] S3: Introduce an offset factor to offset the initial position of subarray 2, and delete the first array element of subarray 2, keeping the total number of array elements unchanged;
[0009] S4: Expand the one-dimensional coprime array to two-dimensional space to obtain a two-dimensional coprime array;
[0010] S5: Use the two-dimensional mutually prime array in step S4 as a receiving array, place an orthogonal antenna at each array element position, and use the rank-deficient MUSIC algorithm to complete the final DOA estimation.
[0011] Furthermore, in step S1, the array element spacing of subarray 1 is M2λ / 2, and the array element spacing of subarray 2 is M1λ / 2, where λ represents wavelength.
[0012] Furthermore, the set of all array element positions of subarray 1 and subarray 2 in step S1 is:
[0013] {S1∪S2=mM2λ / 2∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-1} (1)
[0014] Substituting M2=M1+1 into equation (1), we can obtain:
[0015] {S1∪S2=m(M1+1)λ / 2∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (2)
[0016] The set of virtual element positions of subarray 1 and subarray 2 through self-difference and mutual difference is:
[0017] {ld1∪ld2=m(M1+1)λ / 2∪nM1λ / 2∪±(M1(nm)-m)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1}(3)
[0018] Furthermore, in step S2, based on the set of all array element positions of subarray 1 and subarray 2, the virtual array element at position ±(M1 2 -2)λ / 2 will inevitably have holes, which can be proved by contradiction:
[0019] Assume that at position +(M1 2 -2)λ / 2 will not have holes, so the array element position set virtualized by self-difference and mutual difference must have a point +(M1 2 -2)λ / 2;
[0020] Assuming that the position of occurrence is in the self-difference set of sub-matrix 1, then:
[0021] (M1 2 -2)λ / 2=m(M1+1)λ / 2,m=0,1,2...M1-1 (4) Solve for m:
[0022] m=M1-1-1 / (M1+1) (5)
[0023] Since M1 is a natural number greater than 2, the obtained m must be a decimal, which is inconsistent with the value range of m, and the assumption is not valid;
[0024] Assuming that the position of occurrence is in the self-difference set of sub-matrix 2, then:
[0025] (M1 2 -2)λ / 2=nM1λ / 2,n=0,1,2...M1 (6) Solve for n:
[0026] n=M1-2 / M1 (7)
[0027] Since M1 is a natural number greater than 2, the obtained n must be a decimal, which is inconsistent with the value range of n, and the assumption is not valid;
[0028] Assuming that the position of occurrence is in the mutual difference set, then:
[0029] (M1 2 -2)λ / 2=±(M1(nm)-m)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1 (8) After simplification, we can get:
[0030] M1 2 -2=±(M1n-(M1+1)m) (9)
[0031] When the sign on the right side of the equation is positive, M1 2 -2=M1n-(M1+1)m, when n=M1, m=0, the right side of the equation takes the maximum value, and the equation does not hold; when n=M1-1, m=0 or n=M1, m=1, the right side of the equation is smaller than the left side and the right side of the equation takes the second largest value. Therefore, when the sign of the right side of the equation is positive, any m and n cannot satisfy the equation;
[0032] When the right side of the equation is negative, M1 2 -2=(M1+1)m-M1n, when m=M1-1,n=0, the right side of the equation takes the maximum value, and the equation does not hold; when m=M1-2,n=0 or m=M1-1,n=1, the right side of the equation is smaller than the left side and the right side of the equation takes the second largest value. Therefore, when the sign of the right side of the equation is negative, any m and n cannot satisfy the equation;
[0033] In summary, the virtual array element is at position ±(M1 2 -2)d / 2, the assumption that there will be no holes is not true, which proves that the virtual array element is at the position ±(M1 2 -2)λ / 2 will inevitably have holes.
[0034] Furthermore, in step S2, the compression factor is defined as ρ, which is an integer divisible by the natural number M2. By introducing the compression factor ρ, the array element spacing of the subarray 1 is compressed. The compressed array element spacing is M2λ / 2 / ρ. At this time, the set of all array element positions of the subarray 1 and the subarray 2 is:
[0035] {S1∪S2=mM2λ / 2 / ρ∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-1} (10)
[0036] Substituting M2=M1+1 into equation (10), we can obtain:
[0037] {S1∪S2=m(M1+1)λ / 2 / ρ∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (11)
[0038] At this time, the virtual set of array element positions obtained by self-difference and mutual difference is:
[0039] {ld1∪ld2=m(M1+1)λ / 2 / ρ∪nM1λ / 2∪±(nM1-m(M1+1) / ρ)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (12)
[0040] It can be seen from formula (12) that if the compression factor is exactly equal to M2, the virtual array elements have no holes within the aperture range. Therefore, the problem of holes in the original coprime array can be eliminated by introducing the compression factor ρ.
[0041] Furthermore, in step S3, the offset factor is defined as a natural number k, and the sub-array 2 is shifted rightward as a whole by k(M2 / ρ)λ / 2 using the offset factor. Since the sub-array is offset, the total number of array elements increases, and the first array element of the sub-array 2 can be removed to keep the total number of array elements unchanged.
[0042] Define the offset factor as a natural number k, and use the offset factor to shift subarray 2 to the right by kλ / 2. Since subarray 2 is offset, there may be no shared array elements at this time, so the total number of array elements increases. The first array element of subarray 2 can be removed to keep the total number of array elements unchanged. At this time, since the first array element of subarray 2 is deleted, the maximum continuous part of the array elements virtualized by self-difference and mutual difference in the offset array will be reduced. Combined with formula (12), in order to ensure that the maximum continuous part of the virtual array elements does not decrease compared to before the offset, the offset distance of subarray 2 should be limited to an integer multiple of the spacing between subarray elements, that is, subarray 2 should be shifted to the right by k(M2 / ρ)λ / 2.
[0043] Furthermore, the set of all array element positions of subarray 1 and subarray 2 in step S3 is:
[0044] {S1∪S2=mM2λ / 2 / ρ∪(nM1+kM2 / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-2} (13)
[0045] Substituting M2=M1+1 into equation (7), we can obtain:
[0046] {S1∪S2=m(M1+1)λ / 2 / ρ∪(nM1+k(M1+1) / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1-1} (14)
[0047] At this time, the virtual set of array element positions obtained by self-difference and mutual difference is:
[0048] {ld1∪ld2=m(M1+1)λ / 2 / ρ∪(nM1+k(M1+1) / p)λ / 2∪±(nM1+(km)(M1+1) / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1-1} (15)
[0049] Furthermore, the maximum value of the offset factor k in step S3 does not exceed M2-1. To ensure that the virtual array elements have no holes, the maximum value of k cannot exceed M2-1. By introducing the offset factor, the aperture of the array can be expanded while keeping the total number of array elements unchanged.
[0050] Furthermore, in step S5, the array received data is modeled based on the array element position to obtain the signal source covariance matrix, and the noise subspace E is obtained by performing eigenvalue eigenvector decomposition on the covariance matrix. N Then use the spectrum peak coarse search to estimate the signal DOA and then use the spectrum peak fine search to complete the final direction finding:
[0051]
[0052] in, It represents the part of the signal's steering vector that does not contain polarization parameters.
[0053] Beneficial effect: Compared with the prior art, the present invention eliminates the holes of virtual array elements by introducing compression factors and offset factors while ensuring that the number of array elements remains unchanged, and at the same time expands the physical aperture of the array, which can ultimately effectively improve the accuracy of array direction finding. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 It is the overall flow chart of the method of the present invention;
[0055] Figure 2 Schematic diagram of a standard coprime array and its virtual array elements;
[0056] Figure 3 Schematic diagram of the coprime array and its virtual array elements after the compression factor is introduced;
[0057] Figure 4 A schematic diagram of a coprime array and its virtual array elements after the introduction of compression factor and offset factor;
[0058] Figure 5 Schematic diagram of the position of two-dimensional coprime array elements;
[0059] Figure 6 Schematic diagram of signal incident;
[0060] Figure 7 This is the azimuth error traversal result diagram at different azimuth angles at a pitch angle of 60°;
[0061] Figure 8 This is the traversal result diagram of the pitch angle error at different azimuth angles at a pitch angle of 60°. DETAILED DESCRIPTION
[0062] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.
[0063] like Figure 1 As shown, the present invention provides a high-precision direction finding method for array signals based on a two-dimensional improved coprime array, comprising the following steps:
[0064] S1: Set array parameters:
[0065] Usually, the study of coprime arrays is carried out under one-dimensional conditions. Coprime arrays are composed of two sparse arrays, namely subarray 1 and subarray 2, and the array element spacing of these two sparse arrays is different. Among them, subarray 1 has M1 array elements, and the array element spacing is M2λ / 2, and subarray 2 has M2 array elements, and the array element spacing is M1λ / 2, where M1 and M2 are prime numbers to each other, and λ represents the wavelength;
[0066] Set two coprime numbers M1 and M2 as the number of array elements of subarray 1 and subarray 2, where M1 = M2-1, and both numbers are natural numbers greater than 2. Since the greatest common divisor of two consecutive natural numbers is 1, the natural numbers M1 and M2 must be coprime;
[0067] The set of all array element positions of subarray 1 and subarray 2 is:
[0068] {S1∪S2=mM2λ / 2∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-1} (1)
[0069] Substituting M2=M1+1 into equation (1), we can obtain:
[0070] {S1∪S2=m(M1+1)λ / 2∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (2)
[0071] The set of virtual element positions of subarray 1 and subarray 2 through self-difference and mutual difference is:
[0072] {ld1∪ld2=m(M1+1)λ / 2∪nM1λ / 2∪±(M1(nm)-m)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (3) From equation (3), we can see that the physical aperture of the array is 2M1 2 In addition, at positions ±(M1 2 -2)λ / 2 will inevitably have a hole, which can be proved by contradiction (here we prove that +(M1 2 -2) There is a hole at λ / 2, and the other point is similar), specifically:
[0073] Assume that at position +(M1 2 -2)λ / 2 will not have holes, so the array element position set virtualized by self-difference and mutual difference must have a point +(M1 2 -2)λ / 2;
[0074] Assuming that the position of occurrence is in the self-difference set of sub-matrix 1, then:
[0075] (M1 2-2)λ / 2=m(M1+1)λ / 2,m=0,1,2...M1-1 (4) Solve for m:
[0076] m=M1-1-1 / (M1+1) (5)
[0077] Since M1 is a natural number greater than 2, the obtained m must be a decimal, which is inconsistent with the value range of m, and the assumption is not valid;
[0078] Assuming that the position of occurrence is in the self-difference set of sub-matrix 2, then:
[0079] (M1 2 -2)λ / 2=nM1λ / 2,n=0,1,2...M1 (6) Solve for n:
[0080] n=M1-2 / M1 (7)
[0081] Since M1 is a natural number greater than 2, the obtained n must be a decimal, which is inconsistent with the value range of n, and the assumption is not valid;
[0082] Assuming that the position of occurrence is in the mutual difference set, then:
[0083] (M1 2 -2)λ / 2=±(M1(nm)-m)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1 (8) After simplification, we can get:
[0084] M1 2 -2=±(M1n-(M1+1)m) (9)
[0085] When the sign on the right side of the equation is positive, M1 2 -2=M1n-(M1+1)m, when n=M1, m=0, the right side of the equation takes the maximum value, and the equation does not hold; when n=M1-1, m=0 or n=M1, m=1, the right side of the equation is smaller than the left side and the right side of the equation takes the second largest value. Therefore, when the sign of the right side of the equation is positive, any m and n cannot satisfy the equation;
[0086] When the right side of the equation is negative, M1 2 -2=(M1+1)m-M1n, when m=M1-1,n=0, the right side of the equation takes the maximum value, and the equation does not hold; when m=M1-2,n=0 or m=M1-1,n=1, the right side of the equation is smaller than the left side and the right side of the equation takes the second largest value. Therefore, when the sign of the right side of the equation is negative, any m and n cannot satisfy the equation;
[0087] In summary, the virtual array element is at position ±(M1 2-2)d / 2, the assumption that there will be no holes is not true, which proves that the virtual array element is at the position ±(M1 2 -2)λ / 2 will inevitably have holes; therefore, existing coprime arrays will have holes, resulting in low direction finding accuracy.
[0088] In order to show it more intuitively in this embodiment, M1=3, M2=4, and its array schematic diagram and virtual array element distribution diagram are as follows: Figure 2 As shown in the figure, the array elements at positions 0, 4, and 8 form subarray 1, and the array elements at positions 0, 3, 6, and 9 form subarray 2 (the unit length is half a wavelength, the same below). Through the self-difference and mutual difference of the array elements, its virtual array element schematic diagram can be obtained. There are 6 array elements in the two subarrays, and through mutual difference, a virtual array element There are only 17 non-repeated virtual elements in the array, and the effective mutual error rate is 0.5667. Figure 2 It can be seen that there are holes at the ±7 positions, so its maximum continuous part is -6:6 and the physical aperture is 18 times the half wavelength.
[0089] S2: Introduce a compression factor to compress the element spacing of subarray 1 to eliminate holes in virtual elements:
[0090] The compression factor is defined as ρ, which is an integer divisible by the natural number M2. By introducing the compression factor ρ, the element spacing of subarray 1 is compressed. The compressed element spacing is M2λ / 2 / ρ. At this time, the set of all element positions of subarray 1 and subarray 2 is:
[0091] {S1∪S2=mM2λ / 2 / ρ∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-1} (10)
[0092] Substituting M2=M1+1 into equation (10), we can obtain:
[0093] {S1∪S2=m(M1+1)λ / 2 / ρ∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (11)
[0094] At this time, the virtual set of array element positions of sub-array 1 and sub-array 2 through self-difference and mutual difference is:
[0095] {ld1∪ld2=m(M1+1)λ / 2 / ρ∪nM1λ / 2∪±(nM1-m(M1+1) / ρ)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (12)
[0096] It can be seen from formula (12) that if the compression factor is exactly equal to M2, the virtual array elements have no holes within the aperture range. Therefore, the problem of holes in the original coprime array can be eliminated by introducing the compression factor ρ.
[0097] In this embodiment, ρ=M2=4 is obtained, and the spacing of subarray 1 is compressed to one-fourth of the original spacing. The schematic diagram of the array after compression and the distribution diagram of virtual array elements are shown in FIG. Figure 3 As shown. At this time, the array elements at positions 0, 1, and 2 form subarray 1. The virtual array elements have no holes, and 19 non-repeating continuous array elements can be virtualized, with an effective mutual error rate of 0.6333. Its aperture remains unchanged, still only 18 times the half wavelength.
[0098] S3: Introduce an offset factor to offset the initial position of subarray 2, and delete the first array element of subarray 2, keeping the total number of array elements unchanged:
[0099] Define the offset factor as a natural number k, and use the offset factor to shift subarray 2 to the right by kλ / 2. Since subarray 2 is offset, there may be no shared array elements at this time, so the total number of array elements increases. The first array element of subarray 2 can be removed to keep the total number of array elements unchanged. At this time, since the first array element of subarray 2 is deleted, the maximum continuous part of the array elements virtualized by self-difference and mutual difference in the offset array will be reduced. Combined with formula (12), in order to ensure that the maximum continuous part of the virtual array elements does not decrease compared to before the offset, the offset distance of subarray 2 should be limited to an integer multiple of the spacing between subarray elements, that is, subarray 2 should be shifted to the right by k(M2 / ρ)λ / 2;
[0100] At this time, the set of all array element positions of subarray 1 and subarray 2 is:
[0101] {S1∪S2=mM2λ / 2 / ρ∪(nM1+kM2 / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-2} (13)
[0102] Substituting M2=M1+1 into equation (7), we can obtain:
[0103] {S1∪S2=m(M1+1)λ / 2 / ρ∪(nM1+k(M1+1) / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1-1} (14)
[0104] At this time, the virtual set of array element positions obtained by self-difference and mutual difference is:
[0105] {ld1∪ld2=m(M1+1)λ / 2 / ρ∪(nM1+k(M1+1) / p)λ / 2∪±(nM1+(km)(M1+1) / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1-1} (15)
[0106] The maximum value of the offset factor k does not exceed M2-1. To ensure that the virtual array elements have no holes, the maximum value of k cannot exceed M2-1. By introducing the offset factor, the aperture of the array can be expanded while keeping the total number of array elements unchanged, thereby improving the direction finding accuracy.
[0107] In this embodiment, k=M2-1=3. At this time, the array diagram and the virtual array element distribution diagram are as follows: Figure 4 As shown. The array elements at positions 0, 1, and 2 form subarray 1, and the array elements at positions 6, 9, and 12 form subarray 2. At this time, 25 continuous non-repeating array elements can be virtualized, and the effective mutual error rate is 0.8333. Its aperture is increased to 24 times the half wavelength, and the aperture is greatly improved.
[0108] S4: Expand the one-dimensional coprime array to two-dimensional space to obtain a two-dimensional coprime array:
[0109] In actual direction finding, not only the azimuth of the signal needs to be measured, but also its elevation angle needs to be measured. Therefore, it is necessary to expand the one-dimensional array to two dimensions. The schematic diagram of the expanded two-dimensional coprime array in this embodiment is as follows: Figure 5 As shown, it can virtualize all array elements whose absolute values of the horizontal and vertical coordinates on the coordinate axis do not exceed 12.
[0110] S5: Use the two-dimensional mutually prime array in step S4 as the receiving array, place an orthogonal antenna at each array element position, and use the rank-deficient MUSIC algorithm to complete the final DOA estimation:
[0111] In order to obtain a signal model that is closer to the actual scene, this embodiment uses orthogonal polarization antennas to receive signals. Figure 5 An orthogonal antenna is placed at each array element position, and the signal incident diagram is shown in Figure 6 As shown;
[0112] The array receiving data is modeled based on the array element position to obtain the signal source covariance matrix. The noise subspace E is obtained by performing eigenvalue and eigenvector decomposition on the covariance matrix. N Then use the spectrum peak coarse search to estimate the signal DOA and then use the spectrum peak fine search to complete the final direction finding:
[0113]
[0114] in, It represents the part of the signal's steering vector that does not contain polarization parameters.
[0115] Based on the above content, in order to verify the effectiveness and practical effect of the method of the present invention, in this embodiment, the array before and after the improvement is simulated, and 300 Monte Carlo experiments are performed at each angle. The simulation results are as follows: Figure 7 and Figure 8 As shown, Figure 7 This is the azimuth error traversal result diagram at different azimuth angles at a pitch angle of 60°. Figure 8 1 and 1 are the elevation angle error traversal result diagrams at different azimuth angles at an elevation angle of 60°. It can be seen from the two figures that the improved array provided by the present invention can significantly improve the direction finding accuracy.
Claims
1. A high-precision direction finding method for array signals based on a two-dimensional improved coprime array, characterized in that: The steps include: S1: Set two mutually prime numbers M1 and M2 as the number of array elements of sub-array 1 and sub-array 2, where M1 = M2-1; S2: Introducing a compression factor to compress the element spacing of subarray 1 to eliminate holes in virtual elements; S3: Introduce an offset factor to offset the initial position of subarray 2, and delete the first array element of subarray 2, keeping the total number of array elements unchanged; S4: Expand the one-dimensional coprime array to two-dimensional space to obtain a two-dimensional coprime array; S5: Use the two-dimensional mutually prime array in step S4 as a receiving array, place an orthogonal antenna at each array element position, and use the rank-deficient MUSIC algorithm to complete the final DOA estimation.
2. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 1, characterized in that: In step S1, the array element spacing of subarray 1 is M2λ / 2, and the array element spacing of subarray 2 is M1λ / 2, where λ represents the wavelength.
3. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 2, characterized in that: The set of all array element positions of subarray 1 and subarray 2 in step S1 is: {S1∪S2=mM2λ / 2∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-1} (1) Substituting M2=M1+1 into equation (1), we can obtain: {S1∪S2=m(M1+1)λ / 2∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (2) The set of virtual array element positions through self-difference and mutual difference is: {ld1∪ld2=m(M1+1)λ / 2∪nM1λ / 2∪±(M1(nm)-m)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1}(3).
4. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 3, characterized in that: In step S2, based on the set of all array element positions of subarray 1 and subarray 2, the virtual array element at position ±(M1 2 -2)λ / 2 will inevitably have holes, which can be proved by contradiction: Assume that at position +(M1 2 -2)λ / 2 will not have holes, so the array element position set virtualized by self-difference and mutual difference must have a point +(M1 2 -2)λ / 2; Assuming that the position of occurrence is in the self-difference set of sub-matrix 1, then: (M1 2 -2)λ / 2=m(M1+1)λ / 2,m=0,1,2...M1-1 (4) Solve for m: m=M1-1-1 / (M1+1) (5) Since M1 is a natural number greater than 2, the obtained m must be a decimal, which is inconsistent with the value range of m, and the assumption is not valid; Assuming that the position of occurrence is in the self-difference set of sub-matrix 2, then: (M1 2 -2)λ / 2=nM1λ / 2,n=0,1,2...M1 (6) Solve for n: n=M1-2 / M1 (7) Since M1 is a natural number greater than 2, the obtained n must be a decimal, which is inconsistent with the value range of n, and the assumption is not valid; Assuming that the position of occurrence is in the mutual difference set, then: (M1 2 -2)λ / 2=±(M1(nm)-m)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1 (8) After simplification, we can get: M1 2 -2=±(M1n-(M1+1)m) (9) When the sign on the right side of the equation is positive, M1 2 -2=M1n-(M1+1)m, when n=M1, m=0, the right side of the equation takes the maximum value, and the equation does not hold; when n=M1-1, m=0 or n=M1, m=1, the right side of the equation is smaller than the left side and the right side of the equation takes the second largest value. Therefore, when the sign of the right side of the equation is positive, any m and n cannot satisfy the equation; When the right side of the equation is negative, M1 2 -2=(M1+1)m-M1n, when m=M1-1,n=0, the right side of the equation takes the maximum value, and the equation does not hold; when m=M1-2,n=0 or m=M1-1,n=1, the right side of the equation is smaller than the left side and the right side of the equation takes the second largest value. Therefore, when the sign of the right side of the equation is negative, any m and n cannot satisfy the equation; In summary, the virtual array element is at position ±(M1 2 -2)d / 2, the assumption that there will be no holes is not valid, which proves that the virtual array element is at the position ±(M1 2 -2)λ / 2 will inevitably have holes.
5. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 3, characterized in that: In step S2, the compression factor is defined as ρ, which is an integer divisible by the natural number M2. The element spacing of subarray 1 is compressed by introducing the compression factor ρ. The compressed element spacing is M2λ / 2 / ρ. At this time, the set of all element positions of subarray 1 and subarray 2 is: {S1∪S2=mM2λ / 2 / ρ∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-1} (10) Substituting M2=M1+1 into equation (10), we can obtain: {S1∪S2=m(M1+1)λ / 2 / ρ∪nM1λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (11) At this time, the set of virtual array element positions obtained by self-difference and mutual difference is: {ld1∪ld2=m(M1+1)λ / 2 / ρ∪nM1λ / 2∪±(nM1-m(M1+1) / ρ)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1} (12) It can be seen from formula (12) that if the compression factor is exactly equal to M2, the virtual array elements of the array have no holes within the aperture range. Therefore, by introducing the compression factor ρ, the problem of holes in the original coprime array can be eliminated.
6. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 5, characterized in that: In step S3, the shift factor is defined as a natural number k, and the second subarray is shifted rightward as a whole by the shift factor k(M2 / ρ)λ / 2. Since the subarray is shifted, the total number of array elements increases, and the first array element of the second subarray can be removed to keep the total number of array elements unchanged.
7. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 6, characterized in that: The set of all array element positions of subarray 1 and subarray 2 in step S3 is: {S1∪S2=mM2λ / 2 / ρ∪(nM1+kM2 / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M2-2} (13) Substituting M2=M1+1 into equation (7), we can obtain: {S1∪S2=m(M1+1)λ / 2 / ρ∪(nM1+k(M1+1) / p)λ / 2,m=0,1,2...M1-1,n=0,1,2...M1-1}(14) At this time, the virtual set of array element positions obtained by self-difference and mutual difference is: {ld1∪ld2=m(M1+1)λ / 2 / ρ∪(nM1+k(M1+1) / p)λ / 2∪±(nM1+(km)(M1+1) / p)λ / 2, m=0,1,2...M1-1, n=0,1,2...M1-1} (15).
8. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 7, characterized in that: The maximum value of the offset factor k in step S3 does not exceed M2-1.
9. The high-precision direction finding method for array signals based on a two-dimensional improved coprime array according to claim 7, characterized in that: In step S5, the array received data is modeled based on the array element position to obtain the signal source covariance matrix, and the noise subspace E is obtained by performing eigenvalue eigenvector decomposition on the covariance matrix. N Then use the spectrum peak coarse search to estimate the signal DOA and then use the spectrum peak fine search to complete the final direction finding: in, It represents the part of the signal's steering vector that does not contain polarization parameters.
