Universal 3D codebook design method and device for super-large-scale MIMO

Abstract

The invention discloses a universal 3D codebook design method and device for super-large-scale MIMO (Multiple Input Multiple Output). The 3D codebook design method comprises angle design and distance design. According to the design method, a design scheme of an angle and a distance is obtained based on a target of minimizing a correlation function. Although the angle and distance design scheme is initially designed for the near-field area, the design principle and the implementation effect of the angle and distance design scheme are also suitable for the far-field area, and the angle and distance design scheme has good universality. Compared with the existing work, the 3D codebook design theory and method can realize higher discrimination of the wave beam in each sampling angle and sampling distance, and reveals the performance gain brought by the excavation of the degree of freedom of the vertical dimension under the ideal condition.

Description

Technical Field

[0001] The present invention relates to the field of codebook design, and particularly to a general 3D codebook design method and apparatus for ultra-large-scale MIMO. Background Art

[0002] In the field of wireless communication, the multiple-input multiple-output (MIMO) technology plays a crucial role. By deploying multiple antennas at the transmitter and receiver ends, the MIMO technology realizes advantages such as improved spectrum utilization, enhanced signal transmission ability, and increased system capacity, which are extremely crucial for meeting the growing data transmission demands and improving the user experience. Therefore, the MIMO technology has become one of the core technologies for enhancing the performance of 5G and future communication technologies.

[0003] However, although according to the Hannan limit theory, increasing the aperture area of the planar antenna array to increase the number of antennas can enhance the system performance, the physical size, construction cost, and complexity of the base station actually restrict the expansion of the aperture area of the antenna array. To achieve more efficient signal transmission and extended coverage range and break through the performance limit of two-dimensional (2D) antennas, a possible approach is to construct a three-dimensional (3D) antenna array.

[0004] A 3D antenna is an antenna array that fully explores and utilizes the degrees of freedom in the vertical dimension to optimize the antenna performance. Compared with the traditional 2D antenna array, the array elements in the 3D antenna array have spatial distribution differences both in the transverse plane and the longitudinal plane. The introduction of this spatial height difference enables the projection area of the 3D antenna array in the transverse plane to be expanded, thereby increasing the degrees of freedom of the array. By utilizing the additional space in the vertical dimension, the 3D antenna array can achieve more excellent performance.

[0005] Although 3D antennas show great potential in enhancing the performance of communication systems, they still face many challenges in practical applications. In particular, in the core field of codebook design, there are still many problems to be solved. A codebook is a finite set of predefined quantization phases or amplitudes and quantization phases, which is used at the transmitter end to optimize the signal transmission effect. It plays an important role in reducing inter-user interference and improving spectrum efficiency, etc. The goals of codebook design usually include minimizing quantization error, improving coding quality, reducing signal distortion, and achieving effective channel coding under limited feedback conditions, etc., and these goals play a key role in the application of antennas. Given the significant impact of codebook design on antenna performance, it has become a hot issue in antenna technology research.

[0006] In addition, with the continuous development of 6G technology, the extremely high frequencies and significantly increased antenna array sizes applied have expanded the range of the near-field region. In the future development of communication technology, the exploration and research of the near-field region will gradually become the focus. In the near-field region, the spherical wave propagation characteristics make distance information non-negligible. Therefore, designing a codebook suitable for the near-field region and transferring beamforming technology from simple beam steering to beam focusing is a key research topic.

[0007] Currently, although there are already various general codebooks (such as Grassmannian codebooks, DFT codebooks, etc.), the design of near-field codebooks for 3D antennas has not been fully studied. The introduction of layer spacing in 3D codebook design can optimize the beam performance obtained from codebook design, effectively reduce inter-user interference, and improve the overall system performance. Therefore, 3D codebook design is a research field with practical significance and application value, which needs to be adapted to the structure of 3D antennas and can introduce additional gains in the vertical direction based on existing codebooks.

[0008] In summary, 3D codebook design is an important topic in the development of wireless communication technology, and how to design the codebook for 3D antennas in the near-field region has become a challenge. Summary of the Invention

[0009] The purpose of the present invention is to propose a general 3D codebook design method and device for ultra-large-scale MIMO aiming at the deficiencies of the existing technology.

[0010] The purpose of the present invention is achieved through the following technical solutions: A general 3D codebook design method for ultra-large-scale MIMO, the method includes:

[0011] S1. Establish a two-layer antenna array model, including a two-layer ULA antenna model or a two-layer UPA antenna model;

[0012] S2. Perform Taylor expansion on the distance between the transmitting antenna and the receiving end, obtain the approximate formula of the distance and substitute it into the response vector, so as to obtain the array response vector generated by the entire antenna array as the codebook;

[0013] S3. Decompose the phase part in the correlation function of any two antenna response vectors into two parts, one containing only angles and the other containing both angles and distances. Based on the goal of minimizing the correlation function, find the optimal sampling angles and sampling distances

[0014] S4. Construct codewords according to the optimal sampling angles and sampling distances to obtain the optimal codebook.

[0015] Further, the two-layer antenna array model specifically includes:

[0016] Dual-layer ULA antenna: It consists of 2N antenna elements, with each layer containing N antenna elements; the second-layer antenna is offset by a distance of half an antenna spacing in the arrangement direction compared to the first-layer antenna.

[0017] The above-mentioned dual-layer UPA antenna consists of 2 × N 1 × N 2 antenna elements. Each layer is a UPA array and contains N 1 N 2 antenna elements. N 1 is the number of antenna elements arranged in the vertical direction, and N 2 is the number of antenna elements arranged in the horizontal direction; compared with the first-layer antenna, the second-layer antenna has an offset of half an antenna spacing in the vertical direction and an offset of half an antenna spacing in the horizontal direction.

[0018] Furthermore, the specific process of performing Taylor expansion on the distance between the transmitting antenna and the receiving end includes: performing Taylor expansion on the distance between the transmitting antenna and the user in the dual-layer ULA antenna model to obtain:

[0019]

[0020] r l represents the distance between the reference antenna and the l-th user, and θ l represents the elevation angle of the l-th user, and the value range is p is a variable indicating the layer where the antenna element is located, d is the spacing between adjacent antenna elements in the same layer of the antenna; n is the index of each antenna element in the antenna of this layer, and h is the inter-layer spacing between the two layers of the antenna;

[0021] Performing Taylor expansion on the distance in the dual-layer UPA antenna to obtain:

[0022]

[0023] φ l represents the horizontal azimuth angle of the l-th user, d 1 is the spacing between adjacent antenna elements in the vertical direction in the same layer of the antenna, d 2 is the spacing between adjacent antenna elements in the horizontal direction in the same layer of the antenna; n 1 is the index of each antenna element in the vertical direction of the antenna of this layer, and n 2 is the index of each antenna element in the horizontal direction of the antenna of this layer.

[0024] Furthermore, the correlation function of any two antenna response vectors specifically includes:

[0025] Based on the correlation function of the response vectors between any two users in the dual-layer ULA antenna model:

[0026]

[0027] Correlation function of any two user response vectors in the double-layer UPA antenna model:

[0028]

[0029] where θ a , θ b , φ a , φ b , r a , r b are respectively the elevation angles, horizontal azimuth angles, and reference distances of users a and b.

[0030] Furthermore, the phase part decomposition in S3 is specifically:

[0031] The decomposition in the double-layer ULA antenna array is:

[0032] Phase containing only angle information:

[0033]

[0034] where j is the imaginary symbol, is the wave number; p is the variable indicating the layer where the antenna element is located, n is the index of each antenna element in the antenna of that layer, h is the inter-layer spacing between the two layers of antennas, and d is the spacing between adjacent antenna elements in the same layer of antennas;

[0035] Phase containing angle information and distance information:

[0036]

[0037] The decomposition in the double-layer UPA antenna array is:

[0038] Phase containing only angle information:

[0039]

[0040] Phase containing angle information and distance information:

[0041]

[0042] θ a , θ b , φ a , φ b , r a , r b are respectively the elevation angles, horizontal azimuth angles, and reference distances of users a and b, and d 1 is the spacing between adjacent antenna elements in the vertical direction in the same layer of antennas, and d 2n is the spacing between adjacent antenna elements in the horizontal direction in the same layer of antennas 1 n is the index of each antenna element in the vertical direction of the antennas in this layer 2 n is the index of each antenna element in the horizontal direction of the antennas in this layer

[0043] Furthermore, the objectives of minimizing the correlation function in different decomposition cases include: solving the angle sampling scheme, making the part of the phase that contains distance information zero, and converting the summation term for the number of layers in the objective function, resulting in the angle condition when the objective is satisfied as N is the number of antenna elements included in each layer

[0044] Setting the angle information to zero in the distance sampling scheme and transforming the problem of minimizing the correlation function into a univariate minimization problem with only one variable; obtaining the optimal solution by plotting the curve of this correlation function with respect to the variable at any angle and taking the interval

[0045] Furthermore, in the angle sampling scheme, there are the following specific operation steps according to different antenna arrays

[0046] In the ULA antenna array, the angle solving scheme is

[0047] Making the part that contains distance information At this time, the problem of minimizing the correlation function is transformed into

[0048]

[0049] Since the value of p has only two categories, 0 and 1, the summation operation for p is rewritten as adding a summation term in the summation expression and converting it into a geometric summation with respect to n, that is

[0050]

[0051] The minimum value of the expression is 0, and the angle θ that makes the equation 0 should satisfy

[0052]

[0053] where i is a positive integer, which is the optimal angle sampling scheme

[0054] In the double-layer UPA antenna array, the angle solving scheme is

[0055] Making the part that contains distance information zero, and at this time the problem of minimizing the correlation function is also transformed into

[0056]

[0057] Since the value of p has only two categories, 0 and 1, the summation operation over p is rewritten as adding a summation term in the summation expression, together with sinθ b -sinθ a The only index related to sinθ is n 1 , and the only index related to cosθ b sinφ b -cosθ a -sinφ a is also only n 2 . Therefore, the double summation problem is converted into the product of two summation problems, that is:

[0058]

[0059] N 1 is the number of antenna elements arranged in the vertical direction, and N 2 is the number of antenna elements arranged in the horizontal direction. Using the conclusion obtained in the ULA, the angles θ and φ that make the equation equal to 0 should satisfy:

[0060]

[0061]

[0062] where both i and j are positive integers, which is the optimal angle sampling scheme.

[0063] Furthermore, in the distance sampling scheme, there are the following specific operation steps according to different antenna arrays:

[0064] In the double-layer ULA antenna array, let θ a = θ b = θ, and the problem of minimizing the correlation function to be solved is transformed into:

[0065]

[0066] For this expression, solve for r a , r b ; among them, there is only one variable, which is transformed into the solution of a single-variable minimization problem:

[0067]

[0068] Plot the curve of this correlation function with respect to at any angle. As long as the minimum correlation function obtained meets a certain threshold within the set range, the optimal solution can be obtained;

[0069] In the double-layer UPA antenna array,

[0070] Let the part that only contains angles θ a = θ b = θ, φ a = φ b = φ, the problem of minimizing the correlation function is transformed into:

[0071]

[0072] For this equation, solve for r that minimizes it a , r b ; There is only one variable, and this equation is transformed into the solution of a single-variable minimization problem:

[0073]

[0074] Plot the curve of this correlation function at any angle with respect to variation, set the variable range to obtain the optimal solution within the range. When solving this minimization problem, for N 1 = N 2 = N case, UPA sets a smaller range.

[0075] According to another aspect of the specification, a general 3D codebook design system for ultra-large-scale MIMO is provided, including an antenna parameter acquisition unit, a sampling angle selection unit, and a codeword construction unit;

[0076] The antenna parameter acquisition unit is used to acquire the antenna parameters of the antenna array model;

[0077] The sampling angle selection unit is used to decompose the phase part in the correlation function of any two antenna response vectors into two parts, one that only contains angles and one that contains both angles and distances, and based on the goal of minimizing the correlation function, find the optimal sampling angle and sampling distance;

[0078] The codeword construction unit constructs a codeword according to the optimal sampling angle and sampling distance to obtain an optimal codebook.

[0079] According to another aspect of the specification, a general 3D codebook design device for ultra-large-scale MIMO is provided, including a memory and one or more processors. The memory stores executable code, and when the processor executes the executable code, it implements the general 3D codebook design method for ultra-large-scale MIMO described above.

[0080] Advantages of the present invention: By incorporating the structure of the double-layer antenna array into the modeling of the communication system, the design method uses the derived near-field response vector of the double-layer 3D antenna and proposes a novel 3D codebook design method based on the criterion of minimizing the correlation between different response vectors. Compared with previous codebook design schemes, the method takes into account the layer spacing between different-layer antenna arrays, fully utilizes the degrees of freedom brought by the extra space in the vertical dimension, and thus significantly improves the performance of the codebook design.

[0081] In the communication scenario of 3D antenna applications, the codebook design can achieve beam alignment with high directivity and high discrimination for users at different angles and different distances, effectively reducing interference between users, improving communication quality, and achieving precise focusing of the beam. By exploring the degrees of freedom in the vertical dimension, the codebook design scheme makes more effective use of space resources, enables the beam to better adapt to complex environments, and provides a wider signal coverage.

[0082] In the present invention, although the original intention of the codebook design is for the near-field communication scenario, after a brief analysis of the far-field communication environment, we find that the angular part of the near-field codebook design scheme matches the requirements in far-field communication. Therefore, this codebook design scheme is not only applicable to near-field communication but also to far-field communication, showing good universality.

[0083] Aiming at the gap in 3D antenna codebook design, we consider from several aspects of the correlation between the antenna array structure and the response vector, weigh and compromise the computational complexity and the intensity of interference between users, and obtain a 3D codebook design method with actual performance gain, ensuring computational efficiency while significantly reducing interference between users.

[0084] In addition, in the design process of the present invention, the propagation characteristics of electromagnetic waves are fully considered, and the spherical wave characteristics in the near-field region are approximated to design the scheme, which is easy to combine with electromagnetic simulation technology in the subsequent optimization process, so as to obtain a codebook design scheme closer to the actual application requirements. Brief Description of the Drawings

[0085] Figure 1 is a schematic diagram of near-field communication of the double-layer ULA model of the present invention;

[0086] Figure 2 is a schematic diagram of distinguishable beams in the angular domain under the double-layer ULA model of the present invention;

[0087] Figure 3 is a schematic diagram of distinguishable beams in the distance domain under the double-layer ULA model of the present invention;

[0088] Figure 4 is a schematic diagram of the change of the correlation function with the distance interval under the double-layer ULA model of the present invention;

[0089] Figure 5 It is the near - field communication schematic diagram of the double - layer UPA model of the present invention;

[0090] Figure 6 It is the schematic diagram of the angle - domain distinguishable beam under the double - layer UPA model of the present invention;

[0091] Figure 7 It is the schematic diagram of the distance - domain distinguishable beam under the double - layer UPA model of the present invention;

[0092] Figure 8 It is the schematic diagram of the correlation function varying with the distance interval and angle under the double - layer UPA model of the present invention;

[0093] Figure 9 It is the schematic diagram of the change of the angle - domain beam with the layer spacing under the double - layer UPA model of the present invention;

[0094] Figure 10 It is the schematic diagram of the change of the distance - domain beam with the layer spacing under the double - layer UPA model of the present invention;

[0095] Figure 11 It is the far - field communication schematic diagram of the double - layer ULA model of the present invention;

[0096] Figure 12 It is the schematic diagram of generating a beam for a far - field user through near - field and far - field codebooks under the double - layer UPA model of the present invention;

[0097] Figure 13 It is the structural diagram of the 3D codebook design system of the present invention;

[0098] Figure 14 It is the schematic diagram of the 3D codebook design device of the present invention. Specific Embodiments

[0099] The following further elaborates on the specific embodiments of the present invention with reference to the accompanying drawings.

[0100] The present invention describes an example of a general 3D codebook design method for ultra - large - scale MIMO. The disclosed general 3D codebook design method for ultra - large - scale MIMO includes two parts: angle design and distance design. The applied models of the disclosed 3D codebook design method are two types: double - layer ULA model and double - layer UPA model. The design method is based on the goal of minimizing correlation, thereby obtaining the sampling expressions of angle and distance, effectively restricting the interference between users, and improving the system performance.

[0101] In the embodiments of the present invention, the array design of the double - layer antenna includes: the array includes two layers of ULA or UPA models with multiple radiation units.

[0102] In any of the above aspects / embodiments, the multiple radiation units are evenly arranged on the same layer of antenna.

[0103] In any of the above aspects / embodiments, the multiple radiation units are offset by half an antenna spacing (d / 2) along the arrangement direction on the second layer of antenna.

[0104] In any of the above aspects / embodiments, the electromagnetic coupling effect is not considered among the multiple radiation units.

[0105] The 3D codebook design method includes the calculation of the correlation between different array response vectors. Among them, the array response vector is determined by the relative displacement of each antenna unit in the array. Among them, the relative displacement is determined by the difference between the distance vector from each antenna unit to a certain user and the distance vector from the reference antenna unit to the user.

[0106] In the communication model of the embodiments of the present invention, there are multiple objects in the communication environment, and the antenna array response vector is obtained from the distances between these objects and the antenna units. Since the forms of all antenna array responses are the same, only the unknown angle and distance parameters are different, the codebook design method uses a unified form of correlation function to perform minimization to obtain the angle and distance intervals of precoding. Among them, in the double-layer ULA model, first, the angle part is processed by geometric summation to obtain the optimal angle sampling scheme; then, for the part containing both angle and distance, the angle value is fixed, and minimization is performed to obtain the best distance sampling scheme at the same angle. Among them, in the double-layer UPA model, an extension is made in the horizontal dimension, and the conclusion is generalized according to the processing method in the double-layer ULA. The angle part in the phase is transformed into two independent geometric summations to obtain the best sampling intervals of the azimuth angle and the elevation angle; then, the angle value is fixed, and a similar univariate minimization is performed on the correlation function part containing the distance to obtain the best distance sampling scheme at the same angle.

[0107] In any of the above aspects / embodiments, the 3D antenna array can have an arbitrary antenna spacing;

[0108] In any of the above aspects / embodiments, the 3D antenna array can have an arbitrary placement angle;

[0109] In any of the above aspects / embodiments, the 3D codebook design algorithm can be applicable to the 3D antenna structure composed of the double-layer ULA model;

[0110] In any of the above aspects / embodiments, the 3D codebook design algorithm can be applicable to the 3D antenna structure composed of the double-layer UPA model;

[0111] In any of the above aspects / embodiments, the 3D codebook design algorithm can be applicable to a communication scenario where there are multiple users located at different angles in space.

[0112] In any of the above aspects / embodiments, the 3D codebook design algorithm can be applicable to a communication scenario where there are multiple users located at different distances in space.

[0113] In an embodiment of the present invention, the base station includes a double-layer 3D antenna array for transmitting and receiving in wireless communication. The base station performs precoding using the proposed 3D codebook design scheme to optimize the signal spatial distribution, effectively reduce the interference between users, and improve the communication quality.

[0114] Specifically, the specific implementation steps of the present invention in the double-layer ULA model are as follows:

[0115] S1. Double-layer ULA antenna array modeling: As Figure 1 shown, the present invention studies the codebook design scheme for the double-layer antenna model. The double-layer ULA antenna consists of 2N antenna elements, with each layer containing N antenna elements. The second-layer antenna is offset by in the arrangement direction compared to the first-layer antenna. Taking the central antenna of the first-layer antenna as the reference antenna, in the illustrated model, the coordinates of each antenna element are represented by . Where h is the inter-layer spacing between the two layers of antennas; p is a variable indicating the layer where the antenna element is located, p = 0 indicates that the antenna element is located in the first-layer antenna array, and p = 1 indicates that the antenna element is located in the second-layer antenna array. d is the spacing between adjacent antenna elements in the same layer of the antenna; n is the index of each antenna element in the antenna of this layer. If the index of the reference antenna is specified as 0, then the value range of n is

[0116] S2. Codebook design scheme: Existing codebook design schemes are all based on the 2D antenna model, and the introduction of the vertical dimension of the double-layer 3D antenna increases the design difficulty. The 3D codebook design scheme proposed by the present invention is adapted to the double-layer 3D antenna model. The introduction of the inter-layer spacing in the 3D model optimizes the beam characteristics obtained according to this codebook design scheme and improves the system performance. At the same time, this scheme focuses on the near-field communication range, and the beam generated by it can more accurately focus on the user.

[0117] S2.1. Antenna array response derivation: According to the Figure 1 shown model, let r l represent the distance between the reference antenna and the l-th user, and θ l represent the elevation angle of the l-th user, with the value range of . The distance between the antenna element with index n on the (p + 1)-th layer of the antenna and this user is:

[0118]

[0119] Taylor expansion based on near - field characteristics: For r l (n,p) After simplification, we can get:

[0120]

[0121] Thus, for this antenna element, the phase shift can be expressed as jk c (r l (n,p) - r l ), where j is the imaginary symbol, and k is the wave number. The resulting response vector is:

[0122]

[0123] The array response vector generated by the entire antenna array is:

[0124]

[0125] That is, in the form of a codeword.

[0126] S2.2, Derivation of the correlation function: Based on the antenna array response derived in S2.1, the correlation function of the response vectors between the ath user and the bth user can be obtained:

[0127]

[0128] where θ a , θ b , r a , r b are the elevation angles and reference distances of users a and b respectively. A reasonable codebook design criterion is to minimize the antenna response vectors, that is, the correlation function between codewords, namely:

[0129]

[0130] S2.3, Angle sampling scheme: Observing the exponential part form of the exponential function, it can be decomposed into two parts, one containing only angles and the other containing both angles and distance information

[0131]

[0132] To solve the angle sampling scheme, let the part containing distance information At this time, the problem of minimizing the correlation function is transformed into:

[0133]

[0134] Since the value of p has only two categories, 0 and 1, the summation operation on p can be rewritten by adding a summation term to the summation expression. Thus, the problem is transformed into a geometric summation for n, that is:

[0135]

[0136] Obviously, for this formula, its minimum value is 0, and the angle θ that makes the equation equal to 0 should satisfy:

[0137]

[0138] where i is a positive integer, and this is the optimal angle sampling scheme.

[0139] S2.4. Distance sampling scheme: Similar to the angle sampling process, we let For this equation, when θ a = θ b it is obviously true. We may directly assume that θ a = θ b = θ. At this time, the problem of minimizing the relevant function to be solved is transformed into:

[0140]

[0141] For this formula, solve for r a , r b . By observing it, it is found that there is only one variable term. Therefore, it is transformed into the solution of a single-variable minimization problem:

[0142]

[0143] We may plot the curve of this relevant function varying with at any angle, as shown in Figure 4 (the angle is 40°). The change of the relevant function is fluctuating. Therefore, when solving this minimization problem, the range of needs to be set carefully. An overly small range will cause the value of the relevant function not to meet the expectation, while an overly large range will bring unnecessary computational overhead. In fact, considering the fluctuating characteristics of the relevant function, the range of the sampling interval does not need to be too large, as long as the minimum relevant function obtained within the set range is lower than a certain threshold.

[0144] The specific implementation steps of the present invention in the double-layer UPA model are as follows:

[0145] S1. Double-layer UPA antenna array modeling: AsFigure 5 As shown. The dual-layer UPA antenna consists of 2×N 1 ×N 2 antenna array elements. Each layer is a UPA array, containing N 1 N 2 antenna array elements. N 1 is the number of antenna array elements arranged along the vertical direction (z-axis in the figure), and N 2 is the number of antenna array elements arranged along the horizontal direction (x-axis in the figure). Compared with the first-layer antenna, the second-layer antenna has a offset in the vertical direction and a offset in the horizontal direction. Taking the central antenna of the first-layer antenna as the reference antenna, in the illustrated model, the coordinates of each antenna array element are represented by . Where h is the inter-layer spacing between the two layers of antennas; p is a variable indicating the layer where the antenna array element is located. p = 0 indicates that the antenna array element is located in the first-layer antenna array, and p = 1 indicates that the antenna array element is located in the second-layer antenna array. d 1 is the spacing between adjacent antenna array elements in the vertical direction in the same layer of antennas, and d 2 is the spacing between adjacent antenna array elements in the horizontal direction in the same layer of antennas; n 1 is the index of each antenna array element in the vertical direction of the layer of antennas. If the index of the reference antenna is specified as 0, then n 1 takes values of Similarly, n 2 is the index of each antenna array element in the horizontal direction of the layer of antennas. If the index of the reference antenna is specified as 0, then n 2 takes values of

[0146] S2. Codebook design scheme: Compared with the codebook design of the dual-layer ULA model, the dual-layer UPA model introduces an additional horizontal dimension, and the model establishment and scheme derivation will be more complex.

[0147] S2.1. Antenna array response derivation: According to the Figure 5 model shown, let r l represent the distance between the reference antenna and the l-th user, θ l represent the elevation angle of the l-th user, and φ l represent the horizontal azimuth angle of the l-th user. The value ranges of both angles are The distance between the antenna array element with index (n 1 , n 2 ) located on the (p + 1)-th layer of the antenna and the user is:

[0148]

[0149] Taylor expansion based on near-field characteristics: For Simplifying, we get:

[0150]

[0151] Thus, for this antenna element, the phase shift can be expressed as j is the imaginary symbol, is the wave number. The array response vector generated by the entire antenna array is:

[0152]

[0153] That is, in the form of a codeword.

[0154] S2.2, Derivation of the correlation function: Based on the antenna array response derived in S2.1, the correlation function of the response vectors between the ath user and the bth user can be obtained:

[0155]

[0156] Where θ a , θ b , φ a , φ b , r a , r b are the elevation angles, horizontal azimuth angles, and reference distances of users a and b respectively. A reasonable codebook design criterion is to minimize the antenna response vector, that is, the correlation function between codewords, namely:

[0157]

[0158] S2.3, Angle sampling scheme: Observing the form of the exponential part of the exponential function, it can also be decomposed into two parts that only contain angles and that contain both angles and distance information

[0159]

[0160]

[0161] To solve the angle sampling scheme, let the part containing distance information be 0. At this time, the problem of minimizing the correlation function is transformed into:

[0162]

[0163] Since the value of p has only two categories, 0 and 1, the summation operation on p can be rewritten as adding a summation term in the summation expression. In addition, note that the index related to sinθ b -sinθ a is only n1 , and the index associated with cosθ b sinφ b -cosθ a sinφ a is only n, so this double summation problem can be transformed into the product of two summation problems, that is: 2 , so this double summation problem can be transformed into the product of two summation problems, namely:

[0164]

[0165] Using the conclusion obtained in ULA, it is not difficult for us to obtain that the angles θ and φ that make the equation equal to 0 should satisfy:

[0166]

[0167] where both i and j are positive integers, and this is the optimal angle sampling scheme

[0168] S2.4. Distance sampling scheme: Similar to the angle sampling process, we let the part that only contains angles

[0169] For this equation, when θ a = θ b , φ a = φ b it is obviously true. We may as well directly assume that θ a = θ b = θ, φ a = φ b = φ. At this time, the problem of minimizing the correlation function to be solved is transformed into:

[0170]

[0171] For this formula, solve for r that minimizes it a , r b . Although the formula becomes very complex, there is still only one variable, so it can still be transformed into the solution of a single-variable minimization problem:

[0172]

[0173] We may as well plot the curve of this correlation function with respect to changing at any angle, as shown in Figure 9 (the angle is θ = 40°, φ = 30°). The change of the correlation function is fluctuating. In addition, compared with the correlation function obtained by the double-layer ULA, the value of this correlation function decreases more rapidly with the decrease of the sampling interval. Therefore, when solving this minimization problem, for N 1 = N2 For the case of = N, the UPA can set a smaller range compared to the two-layer ULA model. Similarly, in actual operation, the sampling interval range does not need to be too large, as long as the correlation function obtained within the set range is lower than a certain threshold.

[0174] The codebook designed by the present invention aims to achieve high discrimination in both the angular and distance dimensions. To separately demonstrate these two advantages, we conducted targeted simulation experiments. Given the key role of the layer spacing parameter in the 3D antenna array, we specifically analyzed the specific impact of the layer spacing on the codebook design and performance. In addition, regarding the universality of the near-field codebook design, we also conducted corresponding simulation experiments to prove it. The functions and effects of the present invention are further illustrated and demonstrated through the following simulation experiments:

[0175] (1) Two-layer ULA model multi-user angular domain near-field scenario

[0176] (1.1) Simulation conditions

[0177] The operating frequency of the antenna array is 1.5 GHz, the number of antenna layers is 2, each layer of the antenna consists of N = 31 antenna elements, the distance between adjacent antenna elements in the same layer of the antenna array is d = 0.1 m, the reference layer spacing h between the two layers of the antenna is 0.1 m, and the range of the elevation angle θ is To test the performance of the codebook in the angular domain, we fixed the user distance and set the distance r of all users to the critical near-field Rayleigh distance of 384.4 m. Considering the codebook design in the near-field range in the present invention, to ensure the effectiveness of the codebook, the user distance needs to be within the effective communication area defined by the Rayleigh distance. For simplicity of processing, let the user distance be equal to Rui Li distance.

[0178] (1.2) Simulation results

[0179] Figure 2 Plotted the discrimination effect of the angular domain beam when the user distance is fixed. From the simulation results, it can be seen that within the angular range, the codebook design scheme proposed in the embodiment of the present invention can ensure that the beam has good discrimination at different angles, which means that in an actual communication system, it can allocate beams with extremely low interference to users at different angles. From Figure 2 , it can be observed that the main lobe of the beam shows a wider distribution at the edge angles, and this phenomenon can be attributed to the reduction of the effective projection area of the antenna aperture at these edge angles.

[0180] (2) Two-layer ULA model multi-user distance domain near-field scenario

[0181] (2.1) Simulation conditions

[0182] The antenna array operates at a frequency of 1.5 GHz, with 2 antenna layers. Each layer of antennas consists of N = 125 antenna elements. The distance between adjacent antenna elements in the same layer of antenna arrays is d = 0.1 m, the reference layer spacing between two layers of antennas is h = 0.1 m, and the Rayleigh distance is 1562.5 m. The sampling distance should not exceed this range. To test the performance of the codebook in the distance domain, we fixed the user angle and set the elevation angle θ of all users to 0°.

[0183] The present invention sets a consistent sampling distance interval during the ULA example simulation, so that The sampling interval is based on the codebook design proposed in the present invention. Under this sampling distance interval, the correlation is kept at a low level of about 0.4. This setting is intended to demonstrate the precise range focusing achieved in the range domain and the good discrimination ability of the beam in the range domain.

[0184] (2.2) Simulation results

[0185] Figure 3 The discrimination effect of the distance domain beam when the user angle is fixed is plotted. The simulation results show that the beam obtained by adopting the codebook design scheme proposed in the embodiment of the present invention also has good discrimination ability in the distance domain, which is of great significance in the actual communication system, which means that it can allocate beams with minimal interference to users at the same angle but different distances, and control the interference between users at a low level.

[0186] In addition, according to Figure 3 From the results shown, we observe that the range domain beam achieves maximum gain at the target distance, while the gain decreases rapidly on both sides of the target distance, indicating that the codebook design successfully achieves near-field beam focusing. However, it is noted that at a distance of 156.25m, or 1 / 10 of the Rayleigh distance, the beam gain drop at long distances slows down significantly, which may lead to higher interference at the remaining distances. We believe that the beam focusing effect at this time is close to failure. In fact, range focusing is not a universal phenomenon in the near-field region. In order to keep inter-beam interference at a low level, we found that effective beam focusing can only be achieved within about 1 / 10 of the near-field range.

[0187] (3) Double-layer UPA model for multi-user near-field scenarios

[0188] (3.1) Simulation conditions

[0189] The antenna array operates at a frequency of 1.5 GHz, with 2 antenna layers and N horizontal and vertical antenna elements. 1 =N 2= 7, and each layer of antennas consists of N 1 N 2 = 49 antenna elements. To test the performance of the codebook in the angular domain, we fixed the user distance and set all user distances r to the critical near-field Rayleigh distance of 9.8 m calculated by the empirical formula. At the same time, the layer spacing h between the two layers of antennas was fixed at The present invention considers the codebook design in the near-field range. To ensure the effectiveness of the codebook, the user distance needs to be within the effective communication area defined by the Rayleigh distance. For simplicity of processing, the user distance is set equal to the Rayleigh distance.

[0190] (3.2) Simulation results

[0191] Figure 6 The discrimination effect of the angular domain beam is plotted when the user distance is fixed. From the simulation results, it can be seen that within the angle range, the codebook design scheme proposed in the embodiment of the present invention can ensure that the beam has good discrimination at different angles, which means that in an actual communication system, beams with extremely low interference can be allocated to users at different angles. From Figure 6 , it can be observed that at different values of the elevation angle θ, the values and numbers of the azimuth angle φ are different, and this result corresponds to the angular sampling scheme proposed by the present invention. This formula shows that for different values of θ, the value of cosθ is different, which in turn leads to different values of the azimuth angle φ.

[0192] (4) Near-field scenario in the multi-user distance domain of the double-layer UPA model

[0193] (4.1) Simulation conditions

[0194] The operating frequency of the antenna array is 1.5 GHz, the number of antenna layers is 2, and the number of horizontal and vertical antenna elements is N 1 = N 2 = 25, and each layer of antennas consists of N 1 N 2 = 625 antenna elements. The distance between adjacent antenna elements in the same layer of the antenna array is d = 0.1 m, the reference layer spacing h between the two layers of antennas is 0.1 m, and the Rayleigh distance calculated by the empirical formula is about 250 m. To test the performance of the codebook in the distance domain, we fixed the user angle and set all user elevation angles θ to 0° and the azimuth angle to 0°.

[0195] During the simulation, the present invention example set the same sampling distance interval to 0.05, and this sampling interval is obtained according to the codebook design scheme proposed by the present invention. At this sampling distance interval, the correlation is at a relatively low level. Such a setting aims to demonstrate the precise distance focusing achieved in the distance domain and the good discrimination ability of the beam in the distance domain.

[0196] (4.2) Simulation Results

[0197] Figure 7 The discrimination effect of the distance-domain beam is plotted at a fixed user angle. The simulation results show that the beam obtained by using the codebook design scheme proposed in the embodiments of the present invention also has good discrimination ability in the distance domain, which is of great significance in practical communication systems. This means that it can allocate beams with extremely low interference to users at the same angle but different distances, achieving beam focusing. In the comparative analysis Figure 7 of the beams in the double-layer UPA model shown in Figure 3 and the beam performance in the double-layer ULA model in

[0198] Figure 7

[0199]

[0200] (5) Influence of the layer spacing in the double-layer UPA model on the sampling scheme

[0201] (5.1) Simulation Conditions

[0202] The operating frequency of the antenna array is 1.5 GHz, the number of antenna layers is 2, the number of horizontal and vertical antenna elements is N 1 = N 2 = 15, each layer of the antenna consists of N 1 N 2 = 225 antenna elements. The distance between adjacent antenna elements in the same layer of the antenna array is d = 0.1 m, and the reference layer spacing h between the two layers of antennas is 0.5λ = 0.1 m.

[0202] During simulation, the example of the present invention simultaneously compared the results of the correlation functions with different layer spacings h with those obtained using the reference layer spacing. Additionally, since the sampling points in the angle sampling part were shown to be independent of h in the previous derivation results, the correlation functions here are only the correlation function values related to distance.

[0203] (5.2) Simulation Results

[0204] Figure 8 It shows the variation of the correlation function sampling interval when both the pitch angle θ and the azimuth angle φ are between 0° and 90°, and at the same time, the results for different layer spacings are compared. It can be observed that when the user is at the central angle, that is, θ = φ = 0°, the influence of the layer spacing on the correlation function is difficult to clearly distinguish in the figure, indicating that the layer spacing h has little impact on the entire distance sampling scheme at this angle. However, when the user is at the edge angle, here it is the limit of 90°, Figure 8 The second figure shows the variation of the correlation function with the sampling interval for three layer spacings of h = 0, 0.5, 1λ. At this time, there are obvious distinctions among the three cases. This shows that at the edge angle, the layer spacing has a more significant impact on the correlation function, and thus has a more obvious impact on our sampling scheme. However, due to the volatility of the correlation function itself, the influence of the layer spacing on it is not monotonic, and this influence is closely related to specific angles, the number of antennas, and the layer difference, etc.

[0205] (6) Influence of Layer Spacing on Beam Performance in the Double-Layer UPA Model

[0206] (6.1) Simulation Conditions

[0207] The operating frequency of the antenna array is 1.5 GHz, the number of antenna layers is 2, and the number of horizontal and vertical antenna elements are N 1 = N 2 = 9, and each layer of the antenna consists of N 1 N 2 = 81 antenna elements. The distance between adjacent antenna elements in the same layer of the antenna array is d = 0.1 m, and the reference layer spacing h between the two layers of antennas is 0.5λ = 0.1 m.

[0208] During simulation, the example of the present invention simultaneously compared the results with different layer spacings h with the codebook obtained using the reference layer spacing. In the simulation, the angles were fixed at θ = φ = 60°, and the distance was set to r = 3 m. Such a setting was because r = 3 m is in the near-field region and the beam can be focused at this distance, which meets the requirements of the scenario we discussed.

[0209] (6.2) Simulation Results

[0210] Figure 9It shows the comparison of the angle-domain beam and the range-domain beam at different layer spacings. Figure 9 The first three figures in Figure 9 are the angle-domain beams generated by using the codebook proposed in the present invention under simulation conditions. It can be clearly seen that as the layer spacing increases, the beam width gradually becomes narrower, and the directivity of the beam is improved. The specific numerical results of the three figures show that under this condition, the beam width is reduced by about 27%, and the directivity is improved by about 14%, achieving performance improvement. In addition, it should be noted that there are two main lobes in the figure when h = 0λ. The main reason is that the feasible region of the antenna array is generated by the overlapping part of two cones. In the 2D case, there are symmetric cones on both sides, so two main lobes will be generated. It is normal to have this phenomenon without adding a reflector.

[0211] Figure 10 The second figure in Figure 9 shows the range-domain beam generated by using the codebook proposed in the present invention under simulation conditions. It is observed that as h increases, the beam change speed accelerates, and the interference to distances outside the target distance is reduced more correspondingly. According to Figure 10 the numerical results of the first figure, the width of the range-domain beam is reduced by about 10.7%, also showing performance improvement.

[0212] The emergence of the above results is attributed to the introduction of the layer spacing h, which provides additional degrees of freedom for the antenna array in the vertical direction. Using these degrees of freedom, we can achieve a relative performance improvement while keeping the aperture area unchanged, and the obtained simulation results are also consistent with the theoretical analysis under ideal conditions.

[0213] (7) Far-field scenario of the double-layer UPA model

[0214] (7.1) Simulation conditions

[0215] The operating frequency of the antenna array is 1.5 GHz, the number of antenna layers is 2, and the number of horizontal and vertical antenna elements is N 1 = N 2 = 5. Each layer of the antenna consists of N 1 N 2 = 25 antenna elements. The distance between adjacent antenna elements in the same layer of the antenna array is d = 0.1 m, and the layer spacing h between the two layers of the antenna is 0.1 m. To test the performance of the near-field and far-field codebooks in the far-field environment, we fixed the user angle, set the pitch angle θ of all users to 40° and the azimuth angle to 40°. For the near-field codebook, we set the distance to the Rayleigh distance that distinguishes the radiative near field from the far field. The Rayleigh distance calculated according to the empirical formula is approximately 5 m.

[0216] (7.2) Simulation results

[0217] Figure 12The angular domain beams based on the near-field codebook design and obtained using the far-field response vector are respectively shown. Since distance focusing cannot be achieved in the far-field case, no comparison of the range domain images is made. Figure 12 The first figure of Figure 12 is the beam pattern obtained when the distance is set to the Rayleigh distance according to the derived near-field codebook design theory, while the second figure is the beam pattern generated based on the derived far-field response vector. By comparing the beam images, we find that they are consistent in terms of beam width and angle pointing, etc. The numerical results show that for the beam obtained using the near-field codebook design, its directivity is 15.0338 dBi, the azimuth angle domain bandwidth is 27.34°, and the elevation angle domain bandwidth is 21°; while for the beam obtained from the far-field codebook, the directivity is 15.0526 dBi, the azimuth angle domain bandwidth is 27.33°, and the elevation angle domain bandwidth is 20.93°. In terms of the two key indicators of directivity and bandwidth, both show extremely high similarity.

[0218] Therefore, we can draw the conclusion that by appropriately setting the distance in the near-field codebook to be greater than or equal to the Rayleigh distance, the near-field codebook can provide beams with similar performance to the far-field codebook, thus effectively serving far-field users.

[0219] The above analysis results show that the present invention can not only provide beams with high angular and range resolution, but also improve the beam performance to a certain extent by introducing an additional vertical dimension degree of freedom. In addition, the present invention can be applied to both near-field and far-field environments, having a certain universality.

[0220] The present invention proposes a 3D codebook design theory and method applicable to large-scale MIMO systems in near-field communication scenarios. The disclosed 3D codebook design theory and method are adapted to the 3D antenna structure, and by minimizing the correlation function, the mutual correlation in the beamforming process is reduced, thereby designing beams with high directivity and high discrimination. These beams can significantly reduce the interference between users. In addition, the present invention further optimizes the beam characteristics, enhances the directivity of the beam, improves the performance of the communication system, and can serve far-field users by setting appropriate parameters by utilizing the vertical dimension degree of freedom.

[0221] On the other hand, the present invention also provides a system, as Figure 13 shown, which is a schematic structural diagram of a system for the general 3D codebook design of the present invention for ultra-large-scale MIMO. The implementation processes of the functions and roles of each unit in this device are specifically detailed in the implementation processes of the corresponding steps in the above method, and will not be elaborated here.

[0222] Corresponding to the embodiment of the foregoing general 3D codebook design method for ultra-large-scale MIMO, the present invention also provides an embodiment of a general 3D codebook design device for ultra-large-scale MIMO.

[0223] Referring to Figure 14 , a general 3D codebook design device for ultra-large-scale MIMO provided by an embodiment of the present invention includes a memory and one or more processors. Executable code is stored in the memory. When the processor executes the executable code, it is used to implement the general 3D codebook design method for ultra-large-scale MIMO in the foregoing embodiment.

[0224] The embodiment of the general 3D codebook design device for ultra-large-scale MIMO provided by the present invention can be applied to any device with data processing capabilities. Such a device with data processing capabilities can be a device or apparatus such as a computer. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a logically meaningful device, it is formed by a processor of any device with data processing capabilities reading the corresponding computer program instructions in the non-volatile memory into the memory for operation. From a hardware perspective, as Figure 14 shown, it is a hardware structure diagram of any device with data processing capabilities where the general 3D codebook design device for ultra-large-scale MIMO provided by the present invention is located. In addition to Figure 14 the shown processor, memory, network interface, and non-volatile memory, any device with data processing capabilities where the device in the embodiment is located usually also includes other hardware according to the actual functions of the device with data processing capabilities, which will not be elaborated herein.

[0225] The implementation processes of the functions and roles of each unit in the above device are specifically described in detail in the implementation processes of the corresponding steps in the above method, which will not be elaborated herein.

[0226] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. Each component in the device can be either an independent physical entity or a logical unit. They can be concentrated in one place or distributed on different network nodes. According to actual needs, some or all modules of the present invention can be selected for implementation. For those skilled in the art, these contents are easy to understand and can be executed without additional creative work.

[0227] The present disclosure can be embodied in other specific forms without departing from the subject matter of the claims. The described exemplary embodiments are illustrative only and not restrictive in all respects. Selected features from one or more of the above embodiments can be combined to create alternative embodiments not explicitly described, and features suitable for such combinations are understood to fall within the scope of the present invention. For example, although the calculation order of the disclosed 3D codebook design theory and method is shown, other representation methods and solution orders can be used.

[0228] In addition, although the systems, algorithms, and processes disclosed and shown herein may include a specific number of elements / components and solution methods, these systems, devices, and components can be modified to include more or fewer such elements / components and solution methods with the same key steps. For example, although any disclosed 3D codebook design algorithm may be the calculation result under specific conditions, the embodiments disclosed herein can be modified to include more or fewer such conditions and the calculation results of replacement steps. The present invention is intended to cover all reasonable technical variations and adaptations.

[0229] An embodiment of the present invention also provides a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, it implements a general 3D codebook design method for ultra-large-scale MIMO in the above embodiments.

[0230] The computer-readable storage medium can be an internal storage unit of any device with data processing capabilities described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device of any device with data processing capabilities, such as a plug-in hard disk, a Smart Media Card (SMC), an SD card, a Flash Card, etc. equipped on the device. Further, the computer-readable storage medium can also include both an internal storage unit and an external storage device of any device with data processing capabilities. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capabilities, and can also be used to temporarily store data that has been output or will be output.

[0231] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the general 3D codebook design method for ultra-large-scale MIMO described above.

[0232] Other embodiments of the present application will be readily contemplated by those skilled in the art upon considering the specification and practicing the disclosure herein. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include known common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and examples are only illustrative, and the true scope and spirit of the present application are pointed out by the claims.

[0233] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. The present application is not limited to the exact structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A general 3D codebook design method for ultra-large-scale MIMO, characterized in that: The method includes: S1. Establish a double-layer antenna array model, including a double-layer ULA antenna model or a double-layer UPA antenna model; S2. Perform Taylor expansion on the distance between the transmitting antenna and the receiving end, obtain an approximate formula of the distance and bring it into the response vector, thereby obtaining the array response vector generated by the entire antenna array as a codebook; S3. Decompose the phase part of the correlation function of any two antenna response vectors into two parts: one containing only the angle and the other containing both the angle and the distance. Based on the goal of minimizing the correlation function, find the optimal sampling angle and sampling distance. S4. Construct codewords according to the best sampling angle and sampling distance to obtain an optimal codebook.

2. The method for designing a universal 3D codebook for ultra-large-scale MIMO according to claim 1, characterized in that: The double-layer antenna array model specifically includes: Double-layer ULA antenna: It consists of 2N antenna elements, with each layer containing N antenna elements. The second layer of antennas is offset by half the antenna spacing in the arrangement direction compared to the first layer of antennas. The double-layer UPA antenna is composed of 2×N1×N2 antenna array elements, each layer is a UPA array, including N1N2 antenna array elements, N1 is the number of antenna array elements arranged in the vertical direction, and N2 is the number of antenna array elements arranged in the horizontal direction; compared with the first-layer antenna, the second-layer antenna has an offset of half the antenna spacing in the vertical direction and an offset of half the antenna spacing in the horizontal direction.

3. The method for designing a universal 3D codebook for ultra-large-scale MIMO according to claim 1, characterized in that: The Taylor expansion of the distance between the transmitting antenna and the receiving end specifically includes: Taylor expansion of the distance between the transmitting antenna and the user in the double-layer ULA antenna model to obtain: r l represents the distance between the reference antenna and the lth user, θ l Represents the pitch angle of the lth user, and its value range is p is a variable indicating the layer where the antenna array element is located, d is the spacing between adjacent antenna array elements in the same layer of antennas; n is the index of each antenna array element in the layer of antennas, and h is the inter-layer spacing between two layers of antennas; Taylor expansion of the distance in the double-layer UPA antenna yields: φ l represents the horizontal azimuth angle of the lth user, d1 is the spacing between adjacent antenna elements in the vertical direction in the same layer of antennas, d2 is the spacing between adjacent antenna elements in the horizontal direction in the same layer of antennas; n1 is the index of each antenna element in the vertical direction of the antenna layer, and n2 is the index of each antenna element in the horizontal direction of the antenna layer.

4. The method for designing a universal 3D codebook for ultra-large-scale MIMO according to claim 3, characterized in that: The correlation function of any two antenna response vectors specifically includes: Based on the correlation function of the response vector between any two users in the double-layer ULA antenna model: Correlation function based on any two user response vectors in the double-layer UPA antenna model: Among them, θ a ,θ b ,φ a ,φ b ,r a ,r b They are the pitch angle, horizontal azimuth angle and reference distance of users a and b respectively.

5. The method for designing a universal 3D codebook for ultra-large-scale MIMO according to claim 1, characterized in that: The phase part in S3 is specifically decomposed as follows: The double-layer ULA antenna array is decomposed into: Phase containing angle information only: Where j is the imaginary number symbol, is the wave number; p is a variable indicating the layer where the antenna element is located, n is the index of each antenna element in the antenna layer, h is the inter-layer spacing between two layers of antennas, and d is the spacing between adjacent antenna elements in the same layer of antennas; Phase containing both angle information and distance information: The double-layer UPA antenna array is decomposed into: Phase containing angle information only: Phase containing both angle information and distance information: θ a ,θ b ,φ a ,φ b ,r a ,r b are the elevation angle, horizontal azimuth angle and reference distance of users a and b respectively; d1 is the spacing between adjacent antenna elements in the vertical direction in the same layer of antennas; d2 is the spacing between adjacent antenna elements in the horizontal direction in the same layer of antennas; n1 is the index of each antenna element in the vertical direction of the antenna in this layer; n2 is the index of each antenna element in the horizontal direction of the antenna in this layer.

6. A general 3D codebook design method for ultra-large-scale MIMO according to claim 5, characterized in that: The goal of minimizing the correlation function under different decomposition conditions includes: solving the angle sampling scheme, setting the part of the phase containing distance information to zero, and converting the sum of the number of layers into the sum term in the objective function, and obtaining the angle condition when the goal is met: N is the number of antenna elements contained in each layer; In the distance sampling scheme, the angle information is set to zero, which transforms the problem of minimizing the correlation function into only A single variable minimization problem of one variable; by plotting the curve of this correlation function as the variable changes at any angle, the optimal solution is obtained by taking the interval.

7. The method for designing a universal 3D codebook for ultra-large-scale MIMO according to claim 6, characterized in that: The angle sampling scheme has the following specific operation steps according to different antenna arrays: In the ULA antenna array, the angle solution is: Let the part containing distance information At this time, the problem of minimizing the correlation function is transformed into: Since p can only take two values, 0 and 1, the summation operation on p is rewritten as adding a summation term to the summation expression, converting it into a geometric summation of n, that is: The minimum value of the equation is 0, and the angle θ that makes the equation 0 should satisfy: Where i is a positive integer, which is the optimal angle sampling scheme; In a double-layer UPA antenna array, the angle solution is: Let the part containing distance information be 0, then the problem of minimizing the correlation function is transformed into: Since p can only take two values, 0 and 1, the summation operation on p is rewritten as adding a summation term to the summation expression, which is the same as sinθ b -sinθ a The only relevant index is n1, and cosθ b sinφ b -cosθ a -sinφ a The only associated index is n2, so the double summation problem is converted into the product of two summation problems, namely: N1 is the number of antenna elements arranged in the vertical direction, and N2 is the number of antenna elements arranged in the horizontal direction. Using the conclusions drawn in ULA, the angles θ and φ that make the equation equal to 0 should satisfy: Where i and j are both positive integers, which is the optimal angle sampling scheme.

8. The method for designing a universal 3D codebook for ultra-large-scale MIMO according to claim 6, characterized in that: The distance sampling scheme has the following specific operation steps according to different antenna arrays: In the double-layer ULA antenna array, let θ a =θ b =θ, the problem of minimizing the correlation function to be solved is transformed into: For this formula, find the r that minimizes it. a ,r b ; only One variable is transformed into the solution of a single variable minimization problem: Make this correlation function at any angle For a changing curve, the optimal solution can be obtained as long as the minimum correlation function obtained within the set range is lower than a certain threshold; In the double-layer UPA antenna array, Let only the part containing the angle θ a =θ b =θ,φ a =φ b =φ, the problem of minimizing the correlation function is transformed into: For this formula, find the r that minimizes it. a ,r b ; Only One variable, transform the formula into a solution to a single variable minimization problem: Make this correlation function at any angle The variable range is set to obtain the optimal solution within the range. When solving the minimization problem, for the case of N1=N2=N, UPA sets a smaller range.

9. A universal 3D codebook design system for ultra-large-scale MIMO for implementing the method according to any one of claims 1 to 8, characterized in that: It includes an antenna parameter acquisition unit, a sampling angle selection unit and a codeword construction unit; The antenna parameter acquisition unit is used to acquire antenna parameters of the antenna array model; The sampling angle selection unit is used to decompose the phase part in the correlation function of any two antenna response vectors into two parts, one containing only the angle and the other containing both the angle and the distance, and find the best sampling angle and sampling distance based on the goal of minimizing the correlation function; The codeword construction unit constructs codewords according to the best sampling angle and sampling distance to obtain an optimal codebook.

10. A general 3D codebook design device for ultra-large-scale MIMO, comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that: When the processor executes the executable code, a general 3D codebook design method for ultra-large-scale MIMO is implemented as described in any one of claims 1-8.

Images