A joint optimization method of a rotating antenna system with polarization reconstruction function

Abstract

This invention belongs to the field of wireless communication technology, specifically relating to a joint optimization method for a rotating antenna system with polarization reconfiguration capability. This invention introduces two additional degrees of freedom—three-dimensional mechanical rotation and polarization reconfigurability—at the base station end, and endows the antenna with polarization reconfigurability at the user end, constructing a "space-polarization" joint optimization mechanism to achieve precise matching between beam pointing and transmit / receive polarization states. Under this mechanism, the system can effectively eliminate polarization mismatch losses caused by antenna rotation and channel scattering in complex multipath and multi-user interference environments, thereby significantly reducing base station transmit power and comprehensively improving the system's energy efficiency while ensuring communication quality for all users.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to a joint optimization method for a rotating antenna system with polarization reconfiguration function. Background Technology

[0002] As the demands for data rates and system capacity in 6G communication continue to increase, traditional multi-antenna technologies are limited by physical aperture bottlenecks, making it difficult to further improve performance. Reconfigurable antennas (RAs) have attracted much attention due to their ability to introduce additional degrees of freedom in the antenna domain. Among them, rotatable antennas are typically equipped with highly directional antenna elements, enabling high peak gain. To fully utilize this characteristic, existing research mainly focuses on spatial beamalignment, which involves mechanically rotating a narrow beam to precisely point it at a target user in three-dimensional (3D) space, thereby maximizing the gain in the desired direction. However, such designs generally overlook a crucial physical coupling characteristic: in three-dimensional space, the mechanical rotation of the antenna, while changing the beam direction, inevitably alters its radiation polarization. This inherent coupling between rotation and polarization can easily lead to severe polarization mismatch, making it impossible to effectively receive signals even with perfect spatial beamalignment.

[0003] Furthermore, in complex multipath and multi-user communication environments, factors such as channel scattering can induce significant channel depolarization effects, causing complex changes in the polarization state of the signal arriving at the receiver. In this situation, relying solely on mechanical rotation for spatial beam alignment is insufficient for efficient transmission; dynamic control of the polarization domain must be introduced to achieve precise matching of the transceiver polarization state. However, traditional rotating antennas generally lack the ability to dynamically reconstruct the polarization state, making it difficult to effectively overcome the channel depolarization effect. Consequently, the high directional gain gained through physical rotation is severely offset by potential polarization mismatch losses.

[0004] Therefore, how to break through the limitations of traditional rotating antennas that only focus on spatial beam alignment, and accurately compensate for polarization mismatch while obtaining high pointing gain, so as to achieve coordinated optimization of beam pointing and transmit / receive polarization state, has become a technical problem that needs to be solved in current multi-antenna communication systems. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing rotating antennas that only focus on spatial beam alignment and ignore polarization mismatch, providing a rotating antenna system with polarization reconfiguration capabilities and its joint optimization method. This method introduces two additional degrees of freedom—three-dimensional mechanical rotation and polarization reconfigurability—at the base station end, and endows the antenna with polarization reconfigurability at the user end, constructing a "space-polarization" joint optimization mechanism to achieve precise matching between beam pointing and transmit / receive polarization states. Under this mechanism, the system can effectively eliminate polarization mismatch losses caused by antenna rotation and channel scattering in complex multipath and multi-user interference environments, thereby significantly reducing base station transmit power and comprehensively improving the system's energy efficiency while ensuring communication quality for all users.

[0006] The technical solution of this invention is as follows:

[0007] A joint optimization method for a rotating antenna system with polarization reconfiguration capability is proposed. This method involves configuring a reconfigurable antenna array with three-dimensional mechanical rotation capability and dual polarization ports at the base station. The three-dimensional mechanical rotation capability refers to each antenna element possessing independent three-dimensional mechanical rotation capability, and the three-dimensional rotation matrix of each antenna element containing a special orthogonal group. The constraints are as follows: dual-polarized ports refer to two orthogonal linear polarization ports, defined as the horizontal H port and the vertical V port, respectively. The transmit polarization state vector is adjusted by adjusting the phase shifter and the power divider. Correspondingly, a dual-polarized receiving antenna is configured at the receiving end, and the receiving polarization state vector is adjusted by adjusting the phase shifter.

[0008] The base station has been defined. A reconfigurable rotating antenna, with a receiver having... The number of single-antenna users, the first The three-dimensional rotation matrix of the antenna is , ,in, Determines the physical beam direction of the antenna. and Determine the polarization direction of the antenna in three-dimensional space; [and] the base station end [of the antenna]. The transmit polarization state vector of each antenna The model is as follows:

[0009] ,

[0010] in, This represents the amplitude coefficient, with subscripts H and V corresponding to the horizontal H-port and vertical V-port, respectively, satisfying the power constraint. The phase introduced by the phase shifter is ;

[0011] Definition of the first The received polarization vector for each user is:

[0012] ,

[0013] in, The phase controlled by the phase shifter at the receiving end;

[0014] Define user The equivalent downlink channel vector is , No. The antenna to the first Channel coefficients for individual users Consisting of a Loss path and It is formed by the superposition of NLoS scattering paths:

[0015] ,

[0016] Among them, the The channel coefficients for the scattering paths are as follows:

[0017]

[0018] in, For carrier wavelength, For the first The propagation distance of each path. For the first The large-scale channel fading coefficient of the path. In particular, due to the base station employing a highly directional rotating antenna, the... Large-scale fading of the path Combining Friis transmission loss and directional antenna gain, the specific model is as follows:

[0019]

[0020] in, This represents the physical area of ​​the antenna. This represents the maximum directional gain of the antenna in the main lobe direction. The directivity factor characterizes the beamwidth. The physical beam pointing vector of the antenna (i.e., rotation matrix) The spatial deviation angle between the third column and the actual propagation direction. Clearly, this gain is related to the mechanical rotation matrix. Strong correlation. and These are the projection matrices of the polarization substrates at the transmitting and receiving ends onto the electromagnetic wave cross-section, respectively. For the LoS path, since the polarization state of the electromagnetic wave is preserved, its polarization coupling matrix degenerates into a second-order identity matrix, i.e. For NLoS scattering paths, the depolarization effect of the scatterer plays a significant role. Modeling as

[0021]

[0022] in, This represents the common polarization power ratio, which reflects the cross-polarization discrimination rate of the environment. Indicates the emission polarization along the scattering path. To receive polarization ( Independent random phase shift introduced by ).

[0023] At the base station end, define For users The digital beamforming vector. Simultaneously, let... , as well as Let represent the set of three-dimensional rotation matrices for the base station antenna, the set of base station transmit polarization state vectors, and the set of user receive polarization state vectors, respectively. With the objective of minimizing the total transmit power of the base station, the following joint optimization problem is established:

[0024] in, Ensure that it is assigned to the first achievable transmission rate for individual users It must be greater than or equal to the minimum rate threshold preset by the system. ; To receive constant mode constraints for polarization; The total power of the transmitter polarization state vector is kept normalized. and Together they form a special orthogonal group manifold constraints, where It is a third-order identity matrix; The Z-axis unit vector in the global coordinate system represents the initial pointing direction of the antenna. The spatial angle between the antenna's physical beam pointing and its initial pointing direction is limited to no more than the maximum deflection threshold allowed by the physical hardware. ;

[0025] The optimization problem is solved by breaking it down into two sub-problems and solving them alternately. The first sub-problem is to optimize the digital beamforming matrix of the base station under the condition of fixed antenna physical parameters, including the rotation matrix, transmit polarization state vector, and receive polarization state vector. The second sub-problem is to optimize the physical hardware parameters in sequence, under the condition of fixed beamforming matrix, by optimizing the rotation matrix, transmit polarization state vector, and receive polarization state vector.

[0026] Furthermore, the solution to the first type of subproblem is to define a positive semi-definite matrix. Furthermore, the rank-one constraint is equivalently transformed into the difference between the matrix trace and the spectral norm. Adding this as a penalty term to the objective function yields a standard convex optimization problem:

[0027] ,

[0028] in, , for The eigenvector corresponding to the largest eigenvalue; Represents the inner product of matrices; The penalty parameter is used; by solving this convex optimization problem and performing eigenvalue decomposition, the updated digital beamforming matrix can be recovered.

[0029] The solution to the second type of subproblem is to transform the optimization objective into maximizing the minimum user signal-to-interference-plus-noise ratio in the system. The received signal-to-interference-plus-noise ratio is expressed as:

[0030] ,

[0031] in, For additive noise variance, , , Introducing the Log-Sum-Exp function to construct a smoother alternative objective function:

[0032] ,

[0033] in, To control the smoothing parameter of the Log-Sum-Exp function's approximation accuracy, a Softplus function is introduced to address the constraints on discrete rates in the system and the maximum antenna rotation angle. Transform it into a smooth penalty term, where, This parameter controls the smoothness of the penalty term. After smoothing, the penalty term for exceeding the limit of the antenna rotation angle is applied. Penalties for User Speed ​​Violations They are represented as follows:

[0034] ,

[0035] ,

[0036] in, ;

[0037] Optimization of the three-dimensional rotation matrix of the antenna in a special orthogonal group manifold Above, we construct and minimize a joint cost function that includes a smoothing objective function and double penalties for angle and rate. ;

[0038] Optimize the base station transmit polarization state in a complex unit spherical manifold. Above, construct and minimize the cost function. ;

[0039] Optimization of user receiver polarization state in complex circular manifolds Above, construct and minimize the cost function. ;

[0040] in, All penalty parameters are dynamically updated during the iteration process. The RCG algorithm is used to iteratively solve the cost function sequentially, specifically: First, the regular Euclidean gradient of the smooth cost function with respect to the current optimization variable is calculated; then, the Euclidean gradient is mapped to the tangent space of the manifold where the current iteration point is located using the corresponding orthogonal projection operator to obtain the corresponding Riemann gradient; next, the current search direction is updated according to the conjugate gradient criterion, combined with the search direction and vector transfer mechanism of the previous round, and the optimal iteration step size on the manifold is determined using the Armijo backtracking search algorithm; finally, a shrinking mapping operation is performed to backtrack the updated point on the tangent space to the original physical manifold surface; finally, the optimized physical hardware parameters are obtained.

[0041] The beneficial effects of this invention are as follows:

[0042] 1) Breaking through the performance bottleneck of traditional rotating antennas, achieving precise matching in both spatial and polarization dimensions: Existing rotatable antenna technologies rely solely on mechanical rotation for spatial beam alignment, completely ignoring the polarization direction deflection caused by the rotation itself. This invention reveals and actively compensates for the physical coupling between three-dimensional rotation and polarization direction. Through the joint control of mechanical rotation and polarization reconstruction, this invention not only achieves high directional gain from spatial alignment but also mitigates polarization mismatch losses caused by misalignment or depolarization, fully utilizing the antenna's degrees of freedom in both the spatial and polarization domains.

[0043] 2) Significantly reduces total system transmit power consumption and effectively overcomes channel depolarization effects in complex multipath environments: In practical multipath and multi-user communication environments, channels exhibit significant depolarization effects. The joint optimization method proposed in this invention can achieve precise matching of polarization states at both the transmitting and receiving ends based on channel state information. Compared with traditional fixed antenna systems or systems that only optimize physical beam alignment, this invention greatly reduces the total transmit power required by the base station while ensuring the same quality of service for all communication users, thus significantly improving the system's energy efficiency.

[0044] 3) The algorithm exhibits strong convergence and is feasible for engineering deployment: it addresses the unique orthogonal groups specific to rotating antenna hardware. This invention combines convexity technology with the Riemann conjugate gradient algorithm to impose hardware constraints on manifold constraints and polarization states. While ensuring strict monotonic convergence of the algorithm, it maintains low computational complexity and has engineering value for efficient operation in actual wireless base station hardware. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the antenna hardware architecture with polarization reconfiguration function in the transceiver end (base station transmitter and user receiver) of the present invention.

[0046] Figure 2 This is a comparison chart of the total system transmit power performance of the proposed solution and various benchmark solutions under different target rate requirements;

[0047] Figure 3 This is a comparison chart of the total transmit power performance of the proposed scheme and various benchmark schemes under different maximum antenna rotation angle constraints. Detailed Implementation

[0048] The present invention will now be described in detail with reference to the accompanying drawings and simulation examples.

[0049] The method of the present invention can be logically summarized into the following steps:

[0050] Step 1: Construct a rotation- and polarization-coupled MU-MISO system and channel model; specifically including:

[0051] Configure the transceiver hardware architecture: such as Figure 1 As shown, a reconfigurable antenna array with three-dimensional mechanical rotation capability and dual polarization ports is configured at the base station end, and the transmit polarization state vector is adjusted by adjusting the phase shifter and power divider; a dual polarization receive antenna is configured at the receiver end, and the receive polarization state vector is adjusted by adjusting the phase shifter.

[0052] Constructing a three-dimensional rotation and polarization mapping relationship: defining a special orthogonal group for each antenna element of the base station. A constrained three-dimensional rotation matrix is ​​used to accurately characterize the geometric effects of mechanical rotation on the antenna beam pointing and radiation polarization direction.

[0053] Constructing an equivalent downlink channel model: Taking into account the channel depolarization effects of line-of-sight (LoS) and non-line-of-sight (NLoS) paths, based on the three-dimensional rotation matrix, transmit polarization state vector, and receive polarization state vector, an equivalent channel model including spatial propagation fading and polarization domain coupling relationship is derived and established.

[0054] Step 2: Construct a joint optimization problem with the objective of minimizing transmit power; construct a joint optimization problem with the objective of minimizing the total transmit power of the base station; the constraints of this optimization problem include: rate constraints for each user, and a special orthogonal group of the base station antenna rotation matrix. The optimization variables include the base station's digital beamforming matrix, the antenna's three-dimensional rotation matrix set, the transmit polarization state vector set, and the receive polarization state vector set.

[0055] Step 3: Decomposition of the optimization problem; Considering the highly coupled and non-convex characteristics of the joint optimization problem, an Alternating Optimization (AO) solution framework is designed; the original highly non-convex joint optimization problem is decoupled into four easily handled sub-problems, namely: solving the digital beamforming sub-problem with fixed antenna physical parameters, and solving the antenna three-dimensional rotation matrix sub-problem, the base station transmit polarization state sub-problem, and the user receive polarization state sub-problem under the condition of fixed beamforming.

[0056] Step 4: Digital beamforming optimization based on positive semidefinite programming and convexity algorithm; For a given antenna rotation matrix and transceiver polarization state vector, solve the digital beamforming subproblem; Introduce positive semidefinite programming (SDP) to linearize the non-convex quadratic terms containing the beamforming vector; For the non-convex rank-one constraint introduced after SDP transformation, use the difference-of-convex (DC) technique to introduce a penalty term, and use a first-order Taylor expansion to construct its convex upper bound, transforming the original non-convex subproblem into an iteratively solvable convex optimization problem, and then obtain the updated digital beamforming matrix through singular value decomposition.

[0057] Step 5: Joint optimization of rotation and polarization states based on Riemann conjugate gradient; For a given digital beamforming matrix, since the physical hardware parameters are independent of the transmit power objective function, the original optimization problem is transformed into maximizing the minimum signal-to-interference-plus-noise ratio (Max-min SINR) problem to expand the system rate constraint satisfaction margin; For the antenna rotation matrix, transmit polarization state vector and receive polarization state vector respectively constrained by special orthogonal group manifolds, complex unit spherical manifolds and complex circular manifolds, the Riemannian conjugate gradient (RCG) algorithm is used to perform unconstrained optimization in their respective manifold spaces, thereby efficiently solving the above three physical hardware sub-problems and obtaining the updated antenna rotation matrix and transceiver polarization state vector.

[0058] Example:

[0059] In this example, consider a downlink MU-MISO communication system, which includes a device equipped with A base station with a reconfigurable rotating antenna and Single-antenna user.

[0060] The optimization process for this example is described in detail below, combining the steps outlined above:

[0061] In step one, the base station adopts a uniform planar array (UPA), each antenna element has independent three-dimensional mechanical rotation capability, and two orthogonal linear polarization ports (horizontal H port and vertical V port) are co-located.

[0062] Definition of the first The three-dimensional rotation matrix of the antenna is This matrix is ​​constrained by a special orthogonal group manifold, namely This is to ensure the rigid rotational characteristics of the antenna. Among them, The physical beam pointing (Boresight) of the antenna is determined, and and This determines its polarization direction in three-dimensional space. (Base station number...) The transmit polarization state vector of each antenna This is achieved by adjusting the power divider and phase shifter, and can be modeled as follows:

[0063] ,

[0064] in, This represents the amplitude coefficient, satisfying the power constraint. The phase introduced by the phase shifter is .

[0065] The user end also uses a dual-polarized antenna, and to reduce hardware complexity, a pure phase-controlled polarization architecture is adopted. The received polarization vector for each user is defined as:

[0066] ,

[0067] in, The phase is controlled by the phase shifter at the receiving end.

[0068] Considering the multipath propagation and channel depolarization effects in real-world environments, the first The base station antenna to the first Overall channel coefficient of individual users Consisting of a Loss path and It is formed by the superposition of NLoS scattering paths:

[0069] ,

[0070] Combining the relationship between antenna three-dimensional rotation and spatial polarization projection, the channel coefficients of the LoS path are expressed as:

[0071] ,

[0072] in, For carrier wavelength, This represents the propagation distance of the Loss of Suppression (LoS) path. This represents the large-scale channel fading coefficient of the LoS path. Specifically, due to the use of a highly directional rotating antenna at the base station, the large-scale fading of the LoS path... Combining Friis transmission loss and directional antenna gain, the specific model is as follows:

[0073] ,

[0074] in, This represents the physical area of ​​the antenna. This represents the maximum directional gain of the antenna in the main lobe direction. The directivity factor characterizes the beamwidth. The physical beam pointing vector of the antenna (i.e., rotation matrix) The spatial deviation angle between the third column and the actual propagation direction. Clearly, this gain is related to the mechanical rotation matrix. Strong correlation. and These are the projection matrices of the polarization substrates at the transmitting and receiving ends onto the electromagnetic wave cross-section, respectively. For the LoS path, since the polarization state of the electromagnetic wave is preserved, its polarization coupling matrix degenerates into a second-order identity matrix, i.e. Similarly, the first The channel coefficients of each scattering path are expressed as:

[0075] ,

[0076] Among them, the polarization coupling matrix Characterizes the channel depolarization effect introduced by scatterers in the physical environment. Among them, This indicates the common polarization power ratio. Indicates the emission polarization along the scattering path. To receive polarization ( This introduces independent random phase shifts. Ultimately, the user... The equivalent downlink channel vector is represented as .

[0077] At the base station end, For users Digital beamforming vector, Normalized data symbols. Signals from other users are considered interference. The received signal-to-interference-plus-noise ratio (SINR) is expressed as:

[0078] ,

[0079] in, The variance is additive noise; the equivalent channel vector is explicitly parameterized in the global optimization variables, i.e., the set of antenna rotation matrices. Emission polarization state set and the set of received polarization states .

[0080] In step two, to achieve efficient transmission, this invention constructs a joint optimization problem with the objective of minimizing the total transmit power of the base station:

[0081] ,

[0082] in, Ensure that it is assigned to the first achievable transmission rate for individual users It must be greater than or equal to the minimum rate threshold preset by the system. This ensures the basic communication quality for every communication user; To receive the constant mode constraint of polarization, it is shown that the user end only adjusts the polarization phase without changing the amplitude through the phase shifter, so as to adapt to the pure phase control architecture with low hardware cost. The total power of the transmitter polarization state vector is kept normalized to ensure that the electronic polarization reconstruction process itself does not cause additional energy amplification; and Together they form a special orthogonal group manifold constraints, where It is a third-order identity array. This constraint ensures that the antenna does not deform during physical mechanical rotation and maintains the spatial orthogonality between its polarization port and the beam direction; furthermore, The Z-axis unit vector in the global coordinate system represents the initial pointing direction of the antenna. The spatial angle between the antenna's physical beam pointing and its initial pointing direction is limited to no more than the maximum deflection threshold allowed by the physical hardware. This constraint not only matches the physical rotation limit of the mechanical gimbal, but also effectively avoids physical collisions between adjacent antenna elements in the base station array due to large-scale rotation, and suppresses the electromagnetic mutual coupling effect between antennas.

[0083] Based on step three, the original joint optimization problem, which is not highly convex and involves multivariable coupling, is decomposed into two easily handled subproblems for alternating optimization: one subproblem is solved by fixing the antenna physical parameters (i.e., the rotation matrix). Emission polarization state vector and receive polarization state vector Optimize the digital beamforming matrix of the base station under the condition of ) The second type is a sub-problem in a fixed beamforming matrix. Under certain conditions, this is a sub-problem of optimizing the physical hardware parameters of the antenna's three-dimensional rotation matrix, transmit polarization state vector, and receive polarization state vector in sequence.

[0084] Step four addresses the digital beamforming subproblem. Since the rate constraints include non-convex quadratic terms in the beamforming vector, this invention introduces a positive semidefinite programming approach, defining a positive semidefinite matrix. Transforming it into a linear term introduces a non-convex rank-one constraint. To address this, our method further employs the DC technique, which transforms the rank-one constraint into the difference between the matrix trace and the spectral norm. This is then added as a penalty term to the objective function. Since the negative term of the spectral norm is a concave function, this invention performs a first-order Taylor expansion on it at the previous iteration point to construct a strictly convex upper bound. Substituting this linear approximation into the objective function, the original non-convex subproblem is successfully transformed into a standard convex optimization problem:

[0085] ,

[0086] in, , for The eigenvector corresponding to the largest eigenvalue; Represents the inner product of matrices; The penalty parameter is used; by solving this convex optimization problem and performing eigenvalue decomposition, the updated digital beamforming matrix can be recovered.

[0087] In step five, given a beamforming matrix Under these conditions, since rotation and polarization variables do not directly appear in the total power objective function, this invention transforms its optimization objective into maximizing the minimum user signal-to-interference-plus-noise ratio (Max-min SINR) in the system, thereby expanding the system rate constraint satisfaction margin and enabling further reduction of base station transmit power in the next round of alternating optimization. To address the unique physical and geometric constraints of this problem, this invention employs the RCG algorithm for efficient solution. To address the non-smoothness of the Max-min SINR optimization objective, this invention introduces the Log-Sum-Exp function to construct a smooth alternative objective function:

[0088] ,

[0089] in, This is a smoothing parameter used to control the approximate accuracy of the Log-Sum-Exp function.

[0090] Meanwhile, to address the constraints on the discrete rates in the system and the maximum antenna rotation angle, this invention introduces the Softplus function. Transform it into a smooth penalty term, where, This is a parameter used to control the smoothness of the penalty term. Specifically, it refers to the penalty term for exceeding the limit of the antenna rotation angle. Penalties for User Speed ​​Violations They are represented as follows:

[0091] ,

[0092] ,

[0093] in, .

[0094] Combining the aforementioned smooth substitution function, this invention transforms the original subproblems with complex physical and geometric constraints into smooth, unconstrained optimization problems on their respective manifold spaces:

[0095] Antenna 3D Rotation Matrix Optimization: In Special Orthogonal Group Manifolds Above, we construct and minimize a joint cost function that includes a smoothing objective function and double penalties for angle and rate. ;

[0096] Base station transmit polarization state optimization: in complex unit spherical manifold Above (i.e., satisfying) Construct and minimize the cost function ;

[0097] User receiver polarization state optimization: in complex circular manifolds Above (i.e., satisfying the constant mode constraint of pure phase control) Construct and minimize the cost function .

[0098] in, All of these are penalty parameters that are dynamically updated during the iteration process.

[0099] For the optimization problems on the three manifolds mentioned above, this invention employs the RCG algorithm for iterative solutions. The core solution idea and steps of the RCG algorithm are as follows: First, calculate the regular Euclidean gradient of the smoothing cost function with respect to the current optimization variable; then, use the corresponding orthogonal projection operator to map this Euclidean gradient onto the tangent space of the manifold where the current iteration point is located, and obtain the corresponding Riemannian gradient; next, combine the search direction and vector transport mechanism from the previous round, update the current search direction according to the conjugate gradient criterion, and use the Armijo backtracking search algorithm to determine the optimal iteration step size on the manifold; finally, perform a retraction operation (e.g., for...). The polar decomposition and contraction of the manifold, the normalization operation for the unit sphere, and the phase extraction operation for the complex circular manifold bring the update point on the tangent space back to the original physical manifold surface. This ensures that the algorithm is stable and monotonically convergent while strictly guaranteeing that all geometric constraints of the antenna hardware are constant.

[0100] Simulation example:

[0101] This embodiment utilizes MATLAB simulation software to verify the performance of the proposed rotating antenna system with polarization reconfiguration function. The simulation parameters are set as follows: system carrier frequency is 2.4 GHz; the base station is equipped with... (Right now A uniform planar array of antennas, with an antenna element spacing of half a wavelength; serving a number of users. The channel contains a Loss path and A single NLoS scattering path. For each scattering path, a common polarization power ratio is set. (Right now This simulates the channel depolarization effect in a real-world environment. Users are randomly distributed at distances from the base station. meters, yaw angle Within the range. The receiver noise power is fixed at [value]. dBm. The default minimum communication rate requirement for each user is [value missing]. bps / Hz, antenna pattern directivity factor set to .

[0102] To highlight the performance advantages of this invention, the following three benchmark schemes were introduced for comparison in the simulation:

[0103] Traditional fixed antenna scheme: The three-dimensional rotation and polarization state of the base station antenna are fixed, and only the digital beamforming matrix is ​​optimized. This scheme serves as the lower bound for performance.

[0104] Beam alignment only: The base station antenna only optimizes the direction of the physical beam (i.e., aligns with the user), but does not actively optimize the polarization direction deflection caused by rotation.

[0105] Three-dimensional rotation only: The base station antenna supports three-dimensional rigid body rotation, but the polarization state is fixed and it does not have the ability to reconstruct electronic polarization.

[0106] The present invention provides a solution in which the base station antenna simultaneously supports three-dimensional mechanical rotation and electronic polarization reconstruction, performing joint optimization of beamforming, antenna rotation, and polarization state.

[0107] like Figure 2 As shown, the present invention's solution and various benchmark solutions are presented under different target rate requirements. The graph compares the total transmit power performance of the system under various conditions. The horizontal axis represents the target rate requirement (bps / Hz), and the vertical axis represents the total transmit power required by the base station (dBm). As the target rate requirement increases, the transmit power required by all schemes shows a monotonically increasing trend. However, under any rate requirement, the transmit power of the scheme in this invention is consistently and significantly lower than all benchmark schemes. Specifically, at the target rate… At bps / Hz, compared to traditional fixed antenna solutions, this invention can save up to approximately 12 dB of transmit power by utilizing the additional degrees of freedom in spatial and polarization dimensions, demonstrating extremely excellent energy efficiency.

[0108] like Figure 3 As shown, the method of the present invention and various reference schemes are presented at different maximum rotation angles of the antenna. Comparison of transmit power under constraints and in-depth decomposition of triple physical gain. The horizontal axis in the figure represents the maximum deflection angle under constraints. The vertical axis represents the total transmit power of the system. Firstly, it can be observed that, due to the lack of mechanical rotation capability, the transmit power curve of a traditional fixed antenna scheme is a curve that does not follow the direction of rotation. A changing horizontal straight line. However, for the proposed solution and the three-dimensional rotation-only solution, the transmission power varies with the maximum permissible rotation angle. The relaxation of restrictions has led to a sustained and significant decline. When Reaching approximately At this point, the transmit power curve reaches saturation and no longer decreases. This saturation point perfectly matches the maximum yaw angle range of the user distribution in the simulation scenario. This fully demonstrates that introducing mechanical rotational degrees of freedom can effectively achieve spatial coverage for all service users. Furthermore, comparing the step differences between the four curves in the figure clearly shows that the performance benefits obtained by this invention can be decomposed into progressive physical gains, which is precisely the core technological advancement of this invention.

[0109] Spatial directional alignment gain mainly refers to the power difference resulting from the curve of a traditional fixed antenna scheme falling to that of a beam-aligned scheme only. This gain comes from changing the physical direction of the antenna (i.e., beam alignment) so that the narrow, high-gain main beam is precisely aimed at the user's direction, thereby effectively compensating for the huge directional gain attenuation caused by the angular deviation of the fixed antenna.

[0110] The polarization alignment gain mainly refers to the power difference generated when the curve of the beam alignment scheme is further reduced to the curve of the three-dimensional rotation scheme. This gain reveals a blind spot that is generally ignored in the prior art: based on beam alignment, this invention further aligns the spatial geometric direction of the physical polarization ports of the transmitting and receiving antennas in three-dimensional space by optimizing the roll angle of the antenna, thereby effectively avoiding polarization mismatch loss caused by mechanical rotation.

[0111] The polarization state matching gain mainly refers to the power difference generated when the curve of the rotating scheme finally falls to the curve of the scheme of this invention. This gain demonstrates the irreplaceable nature of polarization reconstruction. Since complex radio waves inevitably undergo depolarization or cross-polarization coupling in multipath propagation, simple physical mechanical rotation cannot change the polarization state of the transmitted signal; while this invention generates a polarization state that perfectly matches the depolarized channel through electronic polarization reconstruction, achieving superior system performance.

Claims

1. A joint optimization method for a rotating antenna system with polarization reconfiguration capability, characterized in that, A reconfigurable antenna array with three-dimensional mechanical rotation capability and dual-polarization ports is configured at the base station. The three-dimensional mechanical rotation capability means that each antenna element has independent three-dimensional mechanical rotation capability, and the three-dimensional rotation matrix of each antenna element contains a special orthogonal group. The constraints are as follows: dual-polarized ports refer to two orthogonal linear polarization ports, defined as the horizontal H port and the vertical V port, respectively. The transmit polarization state vector is adjusted by adjusting the phase shifter and the power divider. Correspondingly, a dual-polarized receiving antenna is configured at the receiving end, and the receiving polarization state vector is adjusted by adjusting the phase shifter. The base station has been defined. A reconfigurable rotating antenna, with a receiver having... The number of single-antenna users, the first The three-dimensional rotation matrix of the antenna is , ,in, Determines the physical beam direction of the antenna. and Determine the polarization direction of the antenna in three-dimensional space; [and] the base station end [of the antenna]. The transmit polarization state vector of each antenna The model is as follows: ， in, This represents the amplitude coefficient, with subscripts H and V corresponding to the horizontal H-port and vertical V-port, respectively, satisfying the power constraint. The phase introduced by the phase shifter is ; Definition of the first The received polarization vector for each user is: ， in, The phase controlled by the phase shifter at the receiving end; Define user The equivalent downlink channel vector is , No. The antenna to the first Channel coefficients for individual users Consisting of a Loss path and It is formed by the superposition of NLoS scattering paths: ， Among them, the The channel coefficients for the scattering paths are as follows: in, For carrier wavelength, For the first The propagation distance of each path, For the first The large-scale channel fading coefficient of the path, the first Large-scale fading of the path Combining Friis transmission loss and directional antenna gain, the model is as follows: ， in, This represents the physical area of ​​the antenna. This represents the maximum directional gain of the antenna in the main lobe direction. The directivity factor characterizing beamwidth, The physical beam pointing vector of the antenna Spatial deviation angle from the actual propagation direction, and These are the projection matrices of the polarization substrates at the transmitting and receiving ends onto the electromagnetic wave cross-section, respectively. For the polarization coupling matrix, for the LosS path, For NLoS scattering paths, The model is as follows: ， in, Indicates the common polarization power ratio. Indicates the emission polarization along the scattering path. To receive polarization Introduced independent random phase shift, ; At the base station end, define For users The digital beamforming vector, at the same time, makes , as well as Let the set of three-dimensional rotation matrices of the base station antenna, the set of base station transmit polarization state vectors, and the set of user receive polarization state vectors be represented respectively. A joint optimization problem is established with the objective of minimizing the total transmit power of the base station: , in, Ensure that it is assigned to the first achievable transmission rate for individual users It must be greater than or equal to the minimum rate threshold preset by the system. ; To receive constant mode constraints for polarization; The total power of the transmitter polarization state vector is kept normalized. and Together they form a special orthogonal group manifold constraints, where It is a third-order identity matrix; Let Z be the unit vector of the global coordinate system, representing the initial pointing direction of the antenna. The spatial angle between the antenna's physical beam pointing and its initial pointing direction is limited to no more than the maximum deflection threshold allowed by the physical hardware. ; The optimization problem is solved by breaking it down into two sub-problems and solving them alternately. The first sub-problem is to optimize the digital beamforming matrix of the base station under the condition of fixed antenna physical parameters, including the antenna three-dimensional rotation matrix, transmit polarization state vector, and receive polarization state vector. The second sub-problem is to optimize the physical hardware parameters of the antenna three-dimensional rotation matrix, transmit polarization state vector, and receive polarization state vector in sequence under the condition of fixed beamforming matrix.

2. The joint optimization method for a rotating antenna system with polarization reconfiguration function according to claim 1, characterized in that, The solution to the first type of subproblem is to define a positive semi-definite matrix. Furthermore, the rank-one constraint is equivalently transformed into the difference between the matrix trace and the spectral norm. Adding this as a penalty term to the objective function yields a standard convex optimization problem: , in, , for The eigenvector corresponding to the largest eigenvalue; Represents the inner product of matrices; The penalty parameter is used; by solving this convex optimization problem and performing eigenvalue decomposition, the updated digital beamforming matrix can be recovered. The solution to the second type of subproblem is to transform the optimization objective into maximizing the minimum user signal-to-interference-plus-noise ratio in the system. The received signal-to-interference-plus-noise ratio is expressed as: ， in, For additive noise variance, , , Introducing the Log-Sum-Exp function to construct a smoother alternative objective function: ， in, To control the smoothing parameter of the Log-Sum-Exp function's approximation accuracy, a Softplus function is introduced to address the constraints on discrete rates in the system and the maximum antenna rotation angle. Transform it into a smooth penalty term, where, The parameter controls the smoothness of the penalty term; after smoothing, the penalty term for exceeding the limit of antenna rotation angle. Penalties for User Speed ​​Violations They are represented as follows: ， ， in, ; Optimization of the three-dimensional rotation matrix of the antenna in a special orthogonal group manifold Above, we construct and minimize a joint cost function that includes a smoothing objective function and double penalties for angle and rate. ; Optimize the base station transmit polarization state in a complex unit spherical manifold. Above, construct and minimize the cost function. ; Optimization of user receiver polarization state in complex circular manifolds Above, construct and minimize the cost function. ; in, All penalty parameters are dynamically updated during the iteration process. The RCG algorithm is used to iteratively solve the cost function sequentially, specifically: First, the regular Euclidean gradient of the smooth cost function with respect to the current optimization variable is calculated; then, the Euclidean gradient is mapped to the tangent space of the manifold where the current iteration point is located using the corresponding orthogonal projection operator to obtain the corresponding Riemann gradient; next, the current search direction is updated according to the conjugate gradient criterion, combined with the search direction and vector transfer mechanism of the previous round, and the optimal iteration step size on the manifold is determined using the Armijo backtracking search algorithm; finally, a shrinking mapping operation is performed to backtrack the updated point on the tangent space to the original physical manifold surface; finally, the optimized physical hardware parameters are obtained.

Images