Reconfigurable intelligent surface-assisted ocean communication method and system based on LR-BCD

Abstract

The invention relates to the technical field of communication, and particularly provides a reconfigurable intelligent surface-assisted ocean communication method and system based on LR-BCD, and the method comprises the steps: constructing an RIS-assisted ocean communication system model; constructing an ocean double-path channel model based on an ocean double-path propagation theory in combination with a sea surface reflection coefficient; singular value decomposition (SVD) is utilized to extract low-rank features of a channel, and high-dimensional beam forming and phase shift optimization problems are projected to a low-dimensional subspace; constructing a low-dimensional and rate maximization problem, and converting the problem into a convex subproblem through a weighted minimum mean square error (WMMSE) criterion; utilizing a block coordinate descent (BCD) algorithm to alternately iteratively update a low-dimensional precoding matrix and a phase shift vector; and finally, reconstructing the optimized low-dimensional variable back to a high-dimensional space and carrying out discretization quantization. According to the method, the sparsity of an ocean channel is utilized, the calculation complexity of large-scale RIS optimization is remarkably reduced through dimension reduction processing, and meanwhile the system and rate performance equivalent to that of a full-rank algorithm is kept.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, specifically relating to a reconfigurable smart surface-assisted marine communication method and system based on LR-BCD. Background Technology

[0002] To meet the demands of today's "smart ocean" and 6G integrated sea-air-space network construction, reconfigurable smart surface (RIS) technology has emerged, aiming to solve the problems of signal fading and obstruction under harsh sea conditions. However, simple RIS-assisted communication design often ignores the unique two-path propagation and sparsity characteristics of the marine environment, making it difficult to accurately match the actual channel environment. Therefore, passive beamforming of RIS to enhance the link is considered, but for large-scale RIS, the computational cost of traditional algorithms such as semi-definite relaxation (SDR) or genetic algorithm (GA) is too high and cannot meet real-time requirements. In addition, existing methods fail to utilize the low-rank characteristics of marine channels for dimensionality reduction, resulting in low optimization efficiency. Summary of the Invention

[0003] To address the aforementioned shortcomings of existing technologies, this invention provides a reconfigurable smart surface-assisted marine communication method and system based on LR-BCD, thereby solving the aforementioned technical problems.

[0004] In a first aspect, the present invention provides a reconfigurable smart surface-assisted marine communication method based on LR-BCD, comprising:

[0005] Construct a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users; Based on the ocean two-path propagation model and the aforementioned communication model, channel models are constructed for base station to reconfigurable smart surface, reconfigurable smart surface to user, and base station to user. The channel from the base station to the reconfigurable smart surface in the channel model is represented as a channel matrix, and singular value decomposition is performed on it. The main singular values ​​and the corresponding left and right singular vector matrices are retained according to the energy threshold. The high-dimensional beamforming matrix to be optimized is projected into a low-dimensional beamforming matrix using the right singular vector matrix, and the high-dimensional phase shift vector to be optimized is projected into a low-dimensional phase shift vector using the left singular vector matrix. Based on the low-dimensional beamforming matrix, the low-dimensional phase shift vector, and the channel model, a low-dimensional equivalent composite channel is constructed, and a power-constrained communication model and a rate maximization problem are constructed based on the low-dimensional equivalent composite channel. The problem is transformed using the weighted minimum mean square error criterion, and the low-rank block coordinate descent algorithm is used to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence. Using the right singular vector matrix and the left singular vector matrix, the converged low-dimensional beamforming matrix and low-dimensional phase shift vector are reconstructed into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication.

[0006] In one optional implementation, a communication model is constructed that includes a base station, a reconfigurable smart surface on the sea surface, and multiple maritime users, including: The base station is defined as a uniform linear array with M antennas, the reconfigurable smart surface is a uniform planar array with N reflective elements, and K maritime users are randomly distributed in the sea surface area. Define the base station transmit beamforming matrix as follows: It is limited by the total transmit power constraint, that is ,in This is the maximum transmission power; Define the phase shift matrix of a reconfigurable smart surface The reflection coefficient of the nth reflecting unit , For the b-bit quantized discrete set Discrete phase shift.

[0007] In an optional implementation, in the communication system model: The effective composite channel for the kth maritime user is formed by superimposing the direct link channel vector from the base station to the user and the cascaded link channel vector reflected by the reconfigurable smart surface. The user's received signal is the product of its effective composite channel and the superimposed signal vector transmitted by the base station, plus receiver noise, wherein the superimposed signal vector is a linear combination of data symbols transmitted by the base station to all users through the beamforming matrix; The signal-to-interference-plus-noise ratio (SIR) of the kth maritime user is defined as the ratio of its expected signal power to the sum of the interference signal power and noise power from other users, wherein the expected signal power and interference signal power are calculated based on its effective composite channel and the corresponding beamforming vector in the beamforming matrix. The achievable total rate of the maritime communication system is defined as the sum of the Shannon channel capacities calculated by all maritime users based on their respective self-defined interference-to-noise ratios.

[0008] In an optional implementation, based on the ocean two-path propagation model and the communication model, channel models are constructed for base station to reconfigurable smart surface, reconfigurable smart surface to user, and base station to user, including: For each communication path in the communication model, a two-path channel model is established, consisting of the superposition of the direct propagation path and the path reflected from the sea surface. The signal components of each communication path are characterized by path loss, array response, and phase difference caused by path length difference. For communication paths involving sea surface reflection, the reflection effect is modeled using the effective sea surface reflection coefficient, which combines the Fresnel reflection coefficient based on electromagnetic wave polarization and incident angle, and the roughness attenuation factor used to quantify the sea surface wave scattering effect.

[0009] In one optional implementation, the method for calculating the effective reflectance of the sea surface includes: Based on the polarization of the transmitted signal and the incident angle of the electromagnetic wave, calculate the Fresnel reflection coefficient of the sea surface under ideal smooth conditions; Based on the parameters characterizing sea surface roughness, the attenuation factor of the reflected signal caused by the scattering effect of sea surface waves is calculated, i.e., the roughness attenuation factor. The effective sea surface reflectance coefficient is obtained by multiplying the Fresnel reflectance coefficient by the roughness attenuation factor.

[0010] In an optional implementation, the channel from the base station to the reconfigurable smart surface in the channel model is represented as a channel matrix, and singular value decomposition is performed on it. Based on an energy threshold, the principal singular values ​​and their corresponding left and right singular vector matrices are retained, including: The channel model from the base station to the reconfigurable smart surface is represented as a channel matrix; Singular value decomposition is performed on the channel matrix to obtain a decomposition result consisting of a left singular vector matrix, a singular value matrix, and a right singular vector matrix; Based on a preset energy threshold, the top L largest singular values ​​are selected from the singular value matrix and retained, such that the sum of the energies of the retained singular values ​​accounts for a proportion of the total energy of all singular values ​​that is not less than the energy threshold. Simultaneously, the left singular vectors corresponding to the L retained singular values ​​are selected to form a truncated left singular vector matrix, and the right singular vectors corresponding to the L retained singular values ​​are selected to form a truncated right singular vector matrix.

[0011] In an optional implementation, the high-dimensional beamforming matrix to be optimized is projected into a low-dimensional beamforming matrix using the right singular vector matrix, and the high-dimensional phase shift vector to be optimized is projected into a low-dimensional phase shift vector using the left singular vector matrix, including: The high-dimensional beamforming matrix to be optimized on the base station side is operated on with the truncated right singular vector matrix, and mapped to the low-dimensional subspace spanned by the right singular vector matrix, thereby obtaining a low-dimensional beamforming matrix with significantly reduced dimensionality. The high-dimensional phase shift vector to be optimized on the reconfigurable smart surface is operated on with the truncated left singular vector matrix, and mapped to the low-dimensional subspace spanned by the left singular vector matrix to obtain the corresponding low-dimensional phase shift vector. The dimensions of the low-dimensional beamforming matrix and the low-dimensional phase shift vector are determined by the number of singular values ​​L retained, and the transmit power constraint satisfied by the high-dimensional beamforming matrix remains unchanged after the projection is mapped to the low-dimensional space.

[0012] In an optional implementation, the problem is transformed using a weighted minimum mean square error criterion, and a low-rank block coordinate descent algorithm is employed to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence, including: By introducing receiver gain variables and weighting factor variables corresponding to each maritime user as auxiliary variables, the system and rate maximization problem is transformed into an equivalent weighted minimum mean square error minimization problem. The variables to be optimized are divided into four independent variable blocks: the low-dimensional beamforming matrix, the low-dimensional phase shift vector, the set of receiver gain variables, and the set of weighting factor variables. In each iteration of the low-rank block coordinate descent algorithm, the following sub-steps are executed sequentially, and when updating any variable block, the values ​​of the other three variable blocks are fixed: Update receiver gain variables: For each user, calculate the optimal receiver gain in the sense of minimum mean square error based on the current low-dimensional beamforming matrix and low-dimensional phase shift vector; Update weight factor variables: For each user, calculate the corresponding optimal weight factor based on the updated receiver gain; Update the low-dimensional beamforming matrix: Under the conditions of fixed receiver gain, weighting factor and low-dimensional phase shift vector, solve a convex quadratic programming problem constrained by total power to obtain the updated low-dimensional beamforming matrix. Update the low-dimensional phase shift vector: Under the conditions of fixed receiver gain, weighting factor and updated low-dimensional beamforming matrix, solve an unconstrained quadratic programming problem about the vector to obtain the updated low-dimensional phase shift vector; The iteration is repeated until the change in system and rate calculated in two adjacent iterations is less than the preset convergence threshold, at which point the algorithm is considered to have converged.

[0013] In an optional implementation, the converged low-dimensional beamforming matrix and low-dimensional phase shift vector are reconstructed into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector using the right singular vector matrix and the left singular vector matrix, respectively. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication, including: The converged low-dimensional beamforming matrix is ​​then processed with the truncated right singular vector matrix to reconstruct the high-dimensional beamforming matrix on the base station side. The converged low-dimensional phase shift vector is operated on with the truncated left singular vector matrix to reconstruct the high-dimensional continuous phase shift vector of the reconstructable smart surface. For each element in the high-dimensional continuous phase shift vector, a discrete phase shift value is selected from a preset discrete phase shift set according to the hardware resolution of the reconfigurable smart surface, such that the Euclidean distance between the complex exponential unity modulus corresponding to the discrete phase shift value and the element value is minimized. Based on all selected discrete phase shift values, construct the final reconfigurable smart surface discrete phase shift matrix; The high-dimensional beamforming matrix and the discrete phase shift matrix are the final parameters used to control the base station and the reconfigurable smart surface to assist marine communication.

[0014] Secondly, the present invention provides a reconfigurable smart surface-assisted marine communication system based on LR-BCD, comprising: The first building module is used to build a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users. The second construction module is used to construct channel models from the base station to the reconfigurable smart surface, from the reconfigurable smart surface to the user, and from the base station to the user based on the ocean dual-path propagation model and the communication model. The channel decomposition module is used to represent the channel from the base station to the reconfigurable smart surface in the channel model as a channel matrix, perform singular value decomposition on it, and retain the main singular values ​​and the corresponding left and right singular vector matrices according to the energy threshold. The low-dimensional projection module is used to project the high-dimensional beamforming matrix to be optimized into a low-dimensional beamforming matrix using the right singular vector matrix, and to project the high-dimensional phase shift vector to be optimized into a low-dimensional phase shift vector using the left singular vector matrix. The problem construction module is used to construct a low-dimensional equivalent composite channel based on the low-dimensional beamforming matrix, the low-dimensional phase shift vector and the channel model, and to construct a power-constrained communication model and a rate maximization problem based on the low-dimensional equivalent composite channel. The problem transformation module is used to transform the problem using the weighted minimum mean square error criterion, and to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence using a low-rank block coordinate descent algorithm. The high-dimensional reconstruction module is used to reconstruct the converged low-dimensional beamforming matrix and low-dimensional phase shift vector into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively, using the right singular vector matrix and the left singular vector matrix. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication.

[0015] The beneficial effects of this invention are as follows: The reconfigurable intelligent surface-assisted marine communication method and system based on LR-BCD provided by this invention utilizes the inherent low-rank characteristics of marine two-path channels to reduce the dimensionality of large-scale RIS-assisted communication systems, significantly reducing the computational complexity of beamforming and phase shift joint optimization, making the algorithm more suitable for marine communication nodes with limited computational resources. While maintaining system and rate performance similar to full-rank optimization algorithms, it significantly reduces computation time and improves optimization efficiency. Furthermore, the scheme closely integrates marine environmental feature modeling and performs discretization and quantization processing on phase shifts, improving communication reliability and coverage while ensuring the engineering feasibility of the scheme. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic flowchart of a method according to an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of a RIS-assisted marine multi-user communication system provided in an embodiment of the present invention.

[0019] Figure 3 This is a schematic diagram of the marine two-path propagation model of the base station-RIS provided in an embodiment of the present invention.

[0020] Figure 4 The graph shows a comparison of the system and rate convergence performance of the LR-BCD algorithm provided in this embodiment of the invention with that of the benchmark algorithm.

[0021] Figure 5 The figure shows a comparison of the system performance and rate of the LR-BCD algorithm provided in this embodiment of the invention with the benchmark algorithm under different numbers of RIS units.

[0022] Figure 6 A comparison chart of CPU computation time between the LR-BCD algorithm provided in this embodiment of the invention and the benchmark algorithm under different numbers of RIS units.

[0023] Figure 7 The graph shows a comparison of the system and rate performance of the LR-BCD algorithm and the benchmark algorithm provided in this embodiment of the invention under different base station transmit powers.

[0024] Figure 8 The graph shows a comparison of the system and rate performance of the LR-BCD algorithm provided in this embodiment of the invention with that of the benchmark algorithm under different quantization bit precisions. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0027] Marine communication systems provide a foundation for achieving integrated sea, air, and space communication to meet the needs of "smart ocean" and 6G full coverage. In the complex marine communication environment, RIS (Radio Router Array) can assist coastal base stations in data transmission with users at sea. When the direct link between the base station and the user is blocked by waves or suffers severe transmission loss over long distances, RIS can intelligently reconstruct the wireless propagation environment and establish a virtual line-of-sight link, thereby significantly improving the coverage, system capacity, and speed of the wireless communication system.

[0028] This invention focuses on leveraging the low-rank characteristics of ocean channels to jointly optimize RIS phase shifting and base station beamforming, maximizing downlink speed and data rate while meeting power and hardware constraints. Since this problem is non-convex, and in large-scale RIS scenarios, traditional semi-definite relaxation (SDR) or genetic (GA) algorithms have excessively high computational complexity, making them unsuitable for real-time performance. Therefore, this invention utilizes the sparsity of ocean two-path channels and proposes a low-rank block coordinate descent (LR-BCD) algorithm based on singular value decomposition (SVD) to approximate the optimal solution. Results show that the proposed LR-BCD algorithm reduces the optimization dimension from the total number of RIS units to the effective rank of the channel. While maintaining system performance comparable to full-rank algorithms, it significantly reduces computational overhead and runtime, indicating that the LR-BCD algorithm is more suitable for ocean communication nodes with limited computational resources.

[0029] The method provided in this embodiment of the invention is executed by a computer device, and correspondingly, the reconfigurable smart surface-assisted marine communication system based on LR-BCD runs in the computer device.

[0030] Figure 1 This is a schematic flowchart illustrating a method according to an embodiment of the present invention. Wherein, Figure 1 The implementing entity can be a reconfigurable smart surface-assisted marine communication system based on LR-BCD. Depending on different requirements, the order of the steps in this flowchart can be changed, and some can be omitted.

[0031] like Figure 1 As shown, the method includes: S1. Construct a communication model: Construct a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users; S2. Constructing Channel Models: Based on the ocean two-path propagation model and the aforementioned communication model, construct channel models from the base station to the reconfigurable smart surface, from the reconfigurable smart surface to the user, and from the base station to the user; S3. Low-rank feature extraction and dimensionality reduction: The channel from the base station to the reconfigurable smart surface in the channel model is represented as a channel matrix. Singular value decomposition is performed on it, and the main singular values ​​and the corresponding left and right singular vector matrices are retained according to the energy threshold. The high-dimensional beamforming matrix to be optimized is projected into a low-dimensional beamforming matrix using the right singular vector matrix, and the high-dimensional phase shift vector to be optimized is projected into a low-dimensional phase shift vector using the left singular vector matrix. S4. Constructing the optimization problem: Based on the low-dimensional beamforming matrix, the low-dimensional phase shift vector, and the channel model, construct a low-dimensional equivalent composite channel, and based on the low-dimensional equivalent composite channel, construct a power-constrained communication model and a rate maximization problem; S5. Joint optimization solution: The problem is transformed using the weighted minimum mean square error criterion, and the low-rank block coordinate descent algorithm is used to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence. S6. Reconstruction and Quantization: Using the right singular vector matrix and the left singular vector matrix, the converged low-dimensional beamforming matrix and low-dimensional phase shift vector are reconstructed into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication.

[0032] Step S1: Construct a RIS-assisted marine multi-user communication system, the system including a base station, a RIS deployed on the sea surface, and... Marine user equipment.

[0033] In this embodiment of the invention, in step S1, a RIS-assisted marine multi-user communication system is constructed. The base station is located at a fixed position and is equipped with... A uniform linear array (ULA) of antennas; the RIS is deployed above the sea surface and equipped with Uniform planar array of passive reflective elements (UPA); Several offshore user equipment units are randomly distributed on the sea surface, among which ; Represent the channel matrix from the base station to the RIS, and the channel matrix from the RIS to the RIS, respectively. The channel vector of the i-th user, and the channel vector from the base station to the i-th user. Direct channel vectors for each maritime user.

[0034] The base station sent Superimposed signal vector of each user As shown:

[0035] in, For the first Transmit beamforming vectors for each user For the first Data symbols for each user; define the beamforming matrix as follows: .

[0036] The total transmit power of a base station is limited by its maximum power. That is, the constraints are satisfied:

[0037] The phase shift matrix of RIS is defined as a diagonal matrix. ,in Indicates the first The reflection coefficient of each reflecting unit. For phase shift; considering RIS hardware limitations, phase shift Values ​​are taken from Discrete sets with bit precision :

[0038] No. The effective composite channel for a maritime user can be represented as:

[0039] Accordingly, the first Received signal for each user Represented as:

[0040] in, Additive white Gaussian noise at the receiving end; Furthermore, the first The signal-to-interference-plus-noise ratio (SINR) of an individual maritime user can be expressed as:

[0041] Ultimately, the achievable total rate of the maritime communication system can be expressed as:

[0042] In embodiments of the present invention, such as Figure 2 As shown, a RIS-assisted marine multi-user communication system is illustrated. (The arrow...) To the arrow This represents the channel matrix from the base station to the maritime user via RIS reflection; arrows This represents the channel matrix from the base station directly to the user at sea.

[0043] Step S2: Based on the ocean two-path propagation theory, construct ocean two-path channel models from the base station to the RIS, from the RIS to the user, and from the base station to the user by combining the sea surface reflection coefficient.

[0044] In this embodiment of the invention, step S2 includes: Due to the sparse scattering characteristics of the marine environment, a two-path propagation model is used to construct the channels for each link. The channel from the base station to the RIS is used as an example. For example, it is represented as:

[0045] in, and These represent the path loss for the direct path and the reflected path, respectively. Indicates the sea surface reflectance; It is the phase difference between the two paths; Represents the array response vector of the base station, where The starting azimuth; Denotes the array response vector of the RIS, where and These correspond to the elevation angle and azimuth angle of arrival, respectively. make Indicates the horizontal distance between the transmitter and receiver. and These represent the antenna heights at the transmitting and receiving ends, respectively. Direct path length. and reflection path length It can be represented as:

[0046] Based on the free-space propagation model, the path loss of the direct path and the reflected path can be expressed as:

[0047] in, This represents the carrier wavelength. Additionally, the phase difference is caused by the path length difference. It can be represented as:

[0048] The effective reflectivity of the sea surface By Fresnel reflection coefficient and sea surface roughness attenuation factor Joint decision, that is ; Fresnel reflectance Assuming the transmitted signal is vertically polarized, it is defined as:

[0049] in, The reflection angle is the reflection path relative to the normal to the sea surface. The complex permittivity of seawater is related to seawater temperature T and salinity S, and the specific formula for its calculation is as follows:

[0050] in, The free space permittivity, The dielectric constant at infinite frequency, The static dielectric constant is It is the ionic conductivity. This is the relaxation time.

[0051] The sea surface roughness attenuation factor Used to quantify the scattering effect caused by ocean waves, defined as:

[0052] in, The glancing angle of the reflection path relative to the sea level. The covariance of wave height is used to characterize sea state roughness; RIS to User Channel and base station to user channel All models are constructed using the aforementioned two-path model, defined as follows:

[0053]

[0054] in, Corresponding to RIS The array response vector in the array, and The corresponding base station is The array response vector in the array.

[0055] In embodiments of the present invention, such as Figure 3 The diagram illustrates a two-path propagation model of a base station-RIS system at sea. The solid arrows represent the direct path, corresponding to their path loss. The dashed arrows represent the reflection path of the RIS after reflection from the sea surface, corresponding to its path loss. . This represents the angle of reflection relative to the normal to the sea surface. This represents the glancing angle of the reflection path relative to the sea level. The superposition of these two paths results in the unique two-path propagation effect of the ocean channel.

[0056] Step S3: Perform singular value decomposition (SVD) on the channel matrix from the base station to the RIS, extract low-rank features using the energy threshold, construct a low-dimensional subspace, and project the high-dimensional beamforming matrix and phase shift vector into low-dimensional variables to construct a low-dimensional equivalent composite channel.

[0057] In this embodiment of the invention, step S3 includes: Due to the ocean two-path propagation effect, the channel matrix It exhibits significant low-rank characteristics; for the channel matrix from the base station to the RIS Perform singular value decomposition (SVD), denoted as According to the preset energy threshold Before keeping The truncation of the channel is approximated by the maximal singular value:

[0058] in, and These are the truncated left and right singular vector matrices, respectively. For including the previous A diagonal matrix with singular values; The formula for determining the value of is: .

[0059] Using truncated right singular vector matrix High-dimensional beamforming matrix of the base station Projection as a low-dimensional beamforming matrix ,Right now:

[0060] because The total power constraint remains invariant in the low-dimensional space, that is:

[0061] Using the truncated left singular vector matrix The high-dimensional phase shift vector of RIS Projection is a low-dimensional phase shift vector ,Right now:

[0062] Also available .

[0063] To construct the objective function for the marine communication system, we substitute the low-dimensional representation into the first... In the equivalent channel expression for a maritime user, considering the sparse scattering characteristics of the marine two-path propagation environment, the approximate expression is... and Introducing the original effective composite channel , No. The equivalent received signal term for a maritime user can be expressed as:

[0064] use The above equation can be simplified to:

[0065] Therefore, the equivalent low-dimensional composite channel vector of maritime users

[0066]

[0067] At this point, the high-dimensional optimization problem has been transformed into... In a dimensional subspace.

[0068] Step S4: Based on the low-dimensional equivalent composite channel, determine the signal-to-interference-plus-noise ratio (SINR) and system rate of the maritime user, and construct a system rate maximization problem constrained by base station transmit power and RIS discrete phase shift.

[0069] In this embodiment of the invention, step S4 includes: Based on the low-dimensional equivalent composite channel, the first... Signal-to-interference-plus-noise ratio (SINR) for individual maritime users:

[0070] Constructing a power-constrained system and a rate maximization problem are represented as:

[0071]

[0072] Step S5: The system and rate maximization problem is transformed into a convex subproblem using the Weighted Minimum Mean Square Error (WMMSE) criterion, and the receiver gain, weighting factor, low-dimensional beamforming matrix, and low-dimensional phase shift vector are iteratively updated alternately using the LR-BCD algorithm.

[0073] In this embodiment of the invention, step S5 includes: To address the intractable nonconvexity problem in maritime communication systems and the rate maximization problem, this paper proposes a low-rank block coordinate descent (LR-BCD) algorithm based on weighted minimum mean square error (WMMSE). This method transforms the logarithmic objective function into an equivalent form of minimizing the weighted mean square error (MSE), making it easier to perform iterative optimization.

[0074] Introducing receiver gain and weighting factors These two auxiliary variables, and correspondingly the optimization variables, are divided into four independent blocks: the low-dimensional precoding matrix. Low-dimensional phase vector Receiver gain set and weight factor set .

[0075] For a given receive gain The mean square error (MSE) of the received signal for a single maritime user is defined as:

[0076] Accordingly, the original problem is restructured into the following problem of minimizing the augmented weighted mean squared error (WMSE):

[0077] To solve The block coordinate descent (BCD) method is employed. This iterative framework leverages the convexity of the objective function with respect to a single variable block (while fixing other variable blocks) to decompose the complex ocean optimization problem into a series of tractable convex subproblems, thereby efficiently updating variables in an alternating manner.

[0078] Update receive gain : Keeping other variables fixed, minimize the mean squared error , can obtain the first Optimal minimum mean square error (MMSE) receiver gain for individual maritime users :

[0079] Update weighting factors : Fix other variables and apply the objective function to Setting the first derivative to zero yields the optimal weighting factor. :

[0080] Update the low-dimensional precoding matrix By fixing other variables and ignoring constant terms, the problem is transformed into a convex quadratic programming problem under power constraints. Its optimal closed-form solution can then be obtained. :

[0081] in, It is to satisfy The Lagrange multipliers can be solved using the bisection method; yes The identity matrix, and matrix inversion only involves low-dimensional... .

[0082] Update the low-dimensional phase vector With other variables fixed, solve a quadratic programming problem involving . Recall the equivalent low-dimensional composite channel. .

[0083] To decouple the phase variables embedded in the equivalent composite channel, matrix identities are used. ,in Representative vector The complex conjugate, Represents Hadamaji. (Order) , , The signal term can be expanded as follows:

[0084] To simplify the expression, we define an intermediate constant vector. :

[0085] At this point, the original objective function can be rearranged as follows: Standard quadratic form:

[0086] Where the matrix sum vector They are defined as follows:

[0087]

[0088] Setting the gradient of the objective function to zero yields the unconstrained optimal closed-form solution:

[0089] Step S6: Reconstruct the converged low-dimensional beamforming matrix and low-dimensional phase shift vector back to the high-dimensional space using the low-rank subspace basis matrix, and discretize and quantize the phase shift according to the RIS hardware constraints to obtain the final base station beamforming scheme and RIS phase shift matrix.

[0090] In this embodiment of the invention, step S6 includes: Once the proposed LR-BCD algorithm converges in the low-dimensional subspace, the optimized low-dimensional continuous phase vector can be obtained. and precoding vector , by optimized low-dimensional beamforming vector An optimized low-dimensional beamforming matrix can be obtained. .

[0091] Using right strange subspace basis Reconstructing the high-dimensional transmission precoding matrix of the base station :

[0092] For RIS, consider its strict unity-mode constraint on phase shift (i.e. ), using left strange subspace basis Reconstructing the unquantized high-dimensional continuous phase vector :

[0093] Because the deployed RIS uses a discrete phase shifter as described in the system model, the reconstructed continuous phase vector It must be mapped to a feasible discrete set. To minimize the impact of quantization error on the performance of the ocean system, the minimum Euclidean distance criterion is used to map the ... The phase of each RIS unit is quantized:

[0094] The final discrete phase shift matrix is ​​constructed as follows:

[0095] To accurately assess the system performance under the actual constraints of the marine environment, it is necessary to reconstruct the high-dimensional beam matrix. With the quantized discrete RIS phase shift matrix Substitute back into the original system model. Accordingly, the first... The signal-to-interference-to-noise ratio for each marine user is:

[0096] in, The effective maritime channel, which includes discrete phase shifts, is expressed as:

[0097] in, Indicates the first Beamforming vectors.

[0098] The LR-BCD algorithm flow includes: Input: Ocean channel matrix , , Power constraints SVD energy threshold Iterative convergence tolerance Discrete phase shift Output: Optimal transmit beamforming matrix and phase shift matrix .

[0099] 1 pair Perform truncated SVD based on the threshold. Obtain subspace , .

[0100] 2. Initialize low-dimensional variables , And set the iteration index .

[0101] 3. Repeat the following steps: 4. According to the formula Update receiver gain ; 5. According to the formula Update weighting factors ; 6. According to the formula Update the low-dimensional beamforming matrix ; 7. According to the formula Update the low-dimensional phase shift vector ; 8. Update Iteration Count ; 9. Continue until the convergence formula is satisfied: ; 10. According to the formula and Reconstructing high-dimensional variables and ; 11. End: According to the formula Quantification and output and .

[0102] In summary, by utilizing SVD-based low-rank channel feature extraction and the LR-BCD algorithm for joint optimization of base station beamforming and RIS discrete phase shift, it is theoretically possible to optimize the data transmission rate between coastal base stations and offshore user equipment. Under the premise of satisfying the maximum transmit power constraint of the base station and the discrete phase shift constraint of the RIS hardware, the computational complexity of large-scale RIS can be significantly reduced and the system and rate in the optimization problem can be maximized.

[0103] The conclusions are analyzed based on the above methods: In this embodiment of the invention, the system simulation parameters are shown in Table 1: Table 1 System Simulation Parameters

[0104] This invention assumes that the base station is deployed at a high point along the coast to reduce wave obstruction, and its coordinates are set as follows: Meters; the RIS is deployed on the sea surface to ensure line-of-sight links, and its coordinates are set at... rice; The maritime users are randomly distributed in the following areas: Within the ocean area centered on meters.

[0105] Figure 4 This is a comparison chart of the system and rate convergence performance of the LR-BCD algorithm provided in this embodiment of the invention and the benchmark algorithm. Figure 4 As shown, the horizontal axis represents the number of iterations, and the vertical axis represents the system and rate. It can be seen that the LR-BCD algorithm proposed in this invention (red curve) has the characteristic of fast convergence, reaching a stable state after approximately 10 to 15 iterations. Compared with the computationally complex Full-Rank BCD algorithm (black curve), the LR-BCD algorithm not only converges faster, but its final convergence rate performance is also very close to that of the Full-Rank algorithm, while significantly outperforming the LR-SDR (pink curve) and LR-GA (blue curve) algorithms. This figure verifies that the LR-BCD algorithm has extremely high convergence efficiency while ensuring performance.

[0106] Figure 5 The following is a comparison chart of the system and rate of the LR-BCD algorithm and the benchmark algorithm provided in this embodiment of the invention under different numbers of RIS units, as shown in the figure. Figure 5 As shown, the horizontal axis represents the number of RIS reflection units. The vertical axis represents the system and the rate. With... With the increase of , the sum and rate of all schemes show an increasing trend, which is attributed to the higher beamforming gain. Especially when The performance improvement is most significant when increasing from 16 to 81; while when After exceeding 81, the growth curve tends to flatten out, indicating that the performance gain gradually saturates.

[0107] Figure 6 This is a comparison chart of the CPU computation time of the LR-BCD algorithm and the benchmark algorithm provided in this embodiment of the invention under different numbers of RIS units, as shown in the figure. Figure 6 As shown, the horizontal axis represents the number of RIS reflection units, and the vertical axis represents the average CPU computation time. With... As the rank of the channel increases, the computation time of the full-rank BCD algorithm grows exponentially, making it difficult to meet real-time requirements. In contrast, the LR-BCD algorithm proposed in this invention extracts the low-rank features of the channel using SVD and performs dimensionality reduction, resulting in a very slow increase in computation time.

[0108] Comprehensive comparison Figure 5 and Figure 6 It can be seen that determining the optimal number of RIS reflection units... A balance needs to be struck between system performance and computational complexity. Figure 5 It is evident that as N increases, the system's performance and speed gradually increase and then plateau; simultaneously, Figure 6 The CPU runtime of the display algorithm varies. The number of [something] is growing rapidly. Furthermore, from a hardware implementation perspective, in order to adapt to the binary addressing architecture of digital control systems and maximize hardware resource utilization, The value of should be an integer power of 2 (i.e. Although like While such values ​​are theoretically feasible, they cannot match standard binary address encoding, leading to wasted control address space and hardware design redundancy. Therefore, considering performance saturation point, computational overhead, and hardware encoding efficiency, the embodiment of this invention ultimately selects... (Right now As the number of reflection units, this effectively reduces computational and hardware costs while ensuring the acquisition of most beamforming gains.

[0109] Figure 7 The following is a comparison chart of the system and rate performance of the LR-BCD algorithm and the benchmark algorithm provided in this embodiment of the invention under different base station transmit powers, as shown in the figure. Figure 7 As shown, in a fixed In the case of [unspecified conditions], the system and data rate exhibit a monotonically increasing trend with the increase of base station transmit power. The performance curve of the LR-BCD algorithm proposed in this invention closely follows that of the full-rank BCD algorithm, indicating that it can achieve near-optimal performance gains. Furthermore, the performance of all RIS-assisted schemes is significantly higher than that without RIS assistance, fully demonstrating the necessity of deploying RIS and employing the algorithm of this invention for joint beamforming optimization in marine dual-path propagation environments with severe obstruction and fading.

[0110] Figure 8 The following is a comparison chart of the system and rate performance of the LR-BCD algorithm and the benchmark algorithm under different quantization bit precisions provided in this embodiment of the invention, as shown in the figure. Figure 8 As shown in the figure, the system and speed are compared for 1-bit, 2-bit, and 3-bit discrete phase shifts, as well as continuous phase shifts. Simulation results show that as the number of quantization bits increases... As the quantization increases, the system performance gradually approaches the ideal continuous phase shift condition. It is worth noting that 2-bit quantization can achieve performance very close to that of continuous phase shift, meaning that in practical engineering deployments, low-cost, low-power 2-bit or 3-bit discrete phase shift RIS can meet the high-performance communication requirements of the system of this invention.

[0111] This invention establishes a RIS-assisted multi-user marine communication system model suitable for dual-path propagation environments in the ocean. It proposes a joint optimization scheme for base station beamforming and RIS phase shifting based on singular value decomposition (SVD) and low-rank block coordinate descent (LR-BCD) algorithms to maximize the downlink sum and rate of the system while satisfying base station transmit power constraints and RIS discrete phase shift constraints. Results show that this invention utilizes the inherent sparsity and low-rank characteristics of marine channels for dimensionality reduction, effectively solving the problem of excessive computational complexity in large-scale RIS scenarios. Simulation results confirm that the proposed LR-BCD algorithm can significantly reduce the algorithm's running time while achieving system and rate performance comparable to the high-complexity full-rank BCD algorithm. Furthermore, through joint analysis of system performance and complexity, 64 RIS reflection units are determined to be the optimal configuration. The analysis also shows that even with low-precision (e.g., 2-bit) discrete phase shifting, the system can still maintain near-continuous phase shift performance. Simulation results verify the feasibility and effectiveness of this invention in marine communication nodes with limited computational resources.

[0112] The technical solution provided by this invention includes a method for constructing a RIS-assisted marine multi-user communication system, the system comprising a base station, a RIS deployed on the sea surface, and... Marine user equipment, of which Based on the ocean dual-path propagation theory, ocean dual-path channel models from the base station to the RIS, from the RIS to the user, and from the base station to the user are constructed using the sea surface reflection coefficient. Singular value decomposition (SVD) is performed on the channel matrix from the base station to the RIS, and low-rank features are extracted using an energy threshold to construct a low-dimensional subspace. The high-dimensional beamforming matrix and phase shift vector are projected as low-dimensional variables to construct a low-dimensional equivalent composite channel. Based on the low-dimensional equivalent composite channel, the signal-to-interference-plus-noise ratio (SINR) and system rate of the maritime user are determined, and a system rate maximization problem constrained by power and discrete phase shift is constructed. The system rate maximization problem is transformed into a convex subproblem using the weighted minimum mean square error (WMMSE) criterion, and the receiver gain, auxiliary weights, low-dimensional beamforming matrix, and low-dimensional phase shift vector are iteratively updated using the LR-BCD algorithm. The converged low-dimensional beamforming matrix and low-dimensional phase shift vector are reconstructed back into the high-dimensional space using the low-rank subspace basis matrix, and the phase shift is discretized and quantized according to the RIS hardware constraints to obtain the final base station beamforming scheme and the discrete phase shift matrix of the RIS. This method leverages the sparsity of ocean channels to reduce computational complexity while maintaining the high sum-rate performance of the wireless communication system.

[0113] In some embodiments, the system may include multiple functional modules composed of computer program segments. The computer programs for each program segment in the LR-BCD-based reconfigurable smart surface-assisted marine communication system may be stored in the memory of a computer device and executed by at least one processor to perform (see details). Figure 1 (Description) Functionality of reconfigurable smart surface-assisted marine communication based on LR-BCD.

[0114] In this embodiment, the LR-BCD-based reconfigurable smart surface-assisted marine communication system can be divided into multiple functional modules according to the functions it performs. A module, as referred to in this invention, is a series of computer program segments that can be executed by at least one processor and perform a fixed function, stored in memory. In this embodiment, the functions of each module will be detailed in subsequent embodiments.

[0115] The first building module is used to build a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users. The second construction module is used to construct channel models from the base station to the reconfigurable smart surface, from the reconfigurable smart surface to the user, and from the base station to the user based on the ocean dual-path propagation model and the communication model. The channel decomposition module is used to represent the channel from the base station to the reconfigurable smart surface in the channel model as a channel matrix, perform singular value decomposition on it, and retain the main singular values ​​and the corresponding left and right singular vector matrices according to the energy threshold. The low-dimensional projection module is used to project the high-dimensional beamforming matrix to be optimized into a low-dimensional beamforming matrix using the right singular vector matrix, and to project the high-dimensional phase shift vector to be optimized into a low-dimensional phase shift vector using the left singular vector matrix. The problem construction module is used to construct a low-dimensional equivalent composite channel based on the low-dimensional beamforming matrix, the low-dimensional phase shift vector and the channel model, and to construct a power-constrained communication model and a rate maximization problem based on the low-dimensional equivalent composite channel. The problem transformation module is used to transform the problem using the weighted minimum mean square error criterion, and to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence using a low-rank block coordinate descent algorithm. The high-dimensional reconstruction module is used to reconstruct the converged low-dimensional beamforming matrix and low-dimensional phase shift vector into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively, using the right singular vector matrix and the left singular vector matrix. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication.

[0116] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0117] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or modules may be electrical, mechanical, or other forms.

[0118] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0119] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0120] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the present invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the present invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should also be covered within the protection scope of the present invention.

Claims

1. A reconfigurable smart surface-assisted marine communication method based on LR-BCD, characterized in that, include: Construct a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users; Based on the ocean two-path propagation model and the aforementioned communication model, channel models are constructed for base station to reconfigurable smart surface, reconfigurable smart surface to user, and base station to user. The channel from the base station to the reconfigurable smart surface in the channel model is represented as a channel matrix, and singular value decomposition is performed on it. The main singular values ​​and the corresponding left and right singular vector matrices are retained according to the energy threshold. The high-dimensional beamforming matrix to be optimized is projected into a low-dimensional beamforming matrix using the right singular vector matrix, and the high-dimensional phase shift vector to be optimized is projected into a low-dimensional phase shift vector using the left singular vector matrix. Based on the low-dimensional beamforming matrix, the low-dimensional phase shift vector, and the channel model, a low-dimensional equivalent composite channel is constructed, and a power-constrained communication model and a rate maximization problem are constructed based on the low-dimensional equivalent composite channel. The problem is transformed using the weighted minimum mean square error criterion, and the low-rank block coordinate descent algorithm is used to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence. Using the right singular vector matrix and the left singular vector matrix, the converged low-dimensional beamforming matrix and low-dimensional phase shift vector are reconstructed into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication.

2. The method according to claim 1, characterized in that, Construct a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users, including: The base station is defined as a uniform linear array with M antennas, the reconfigurable smart surface is a uniform planar array with N reflective elements, and K maritime users are randomly distributed in the sea surface area. Define the base station transmit beamforming matrix as follows: It is limited by the total transmit power constraint, that is ,in This is the maximum transmission power; Define the phase shift matrix of a reconfigurable smart surface The reflection coefficient of the nth reflecting unit , For the b-bit quantized discrete set Discrete phase shift.

3. The method according to claim 2, characterized in that, In the communication system model: The effective composite channel for the kth maritime user is formed by superimposing the direct link channel vector from the base station to the user and the cascaded link channel vector reflected by the reconfigurable smart surface. The user's received signal is the product of its effective composite channel and the superimposed signal vector transmitted by the base station, plus receiver noise, wherein the superimposed signal vector is a linear combination of data symbols transmitted by the base station to all users through the beamforming matrix; The signal-to-interference-plus-noise ratio (SIR) of the kth maritime user is defined as the ratio of its expected signal power to the sum of the interference signal power and noise power from other users, wherein the expected signal power and interference signal power are calculated based on its effective composite channel and the corresponding beamforming vector in the beamforming matrix. The achievable total rate of the maritime communication system is defined as the sum of the Shannon channel capacities calculated by all maritime users based on their respective self-defined interference-to-noise ratios.

4. The method according to claim 1, characterized in that, Based on the ocean two-path propagation model and the aforementioned communication model, channel models are constructed for base station to reconfigurable smart surface, reconfigurable smart surface to user, and base station to user, including: For each communication path in the communication model, a two-path channel model is established, consisting of the superposition of the direct propagation path and the path reflected from the sea surface. The signal components of each communication path are characterized by path loss, array response, and phase difference caused by path length difference. For communication paths involving sea surface reflection, the reflection effect is modeled using the effective sea surface reflection coefficient, which combines the Fresnel reflection coefficient based on electromagnetic wave polarization and incident angle, and the roughness attenuation factor used to quantify the sea surface wave scattering effect.

5. The method according to claim 1, characterized in that, Methods for calculating the effective reflectance of the sea surface include: Based on the polarization of the transmitted signal and the incident angle of the electromagnetic wave, calculate the Fresnel reflection coefficient of the sea surface under ideal smooth conditions; Based on the parameters characterizing sea surface roughness, the attenuation factor of the reflected signal caused by the scattering effect of sea surface waves is calculated, i.e., the roughness attenuation factor. The effective sea surface reflectance coefficient is obtained by multiplying the Fresnel reflectance coefficient by the roughness attenuation factor.

6. The method according to claim 1, characterized in that, The channel from the base station to the reconfigurable smart surface in the channel model is represented as a channel matrix. Singular value decomposition is performed on this matrix, and the principal singular values ​​and their corresponding left and right singular vector matrices are retained based on an energy threshold. This includes: The channel model from the base station to the reconfigurable smart surface is represented as a channel matrix; Singular value decomposition is performed on the channel matrix to obtain a decomposition result consisting of a left singular vector matrix, a singular value matrix, and a right singular vector matrix; Based on a preset energy threshold, the top L largest singular values ​​are selected from the singular value matrix and retained, such that the sum of the energies of the retained singular values ​​accounts for a proportion of the total energy of all singular values ​​that is not less than the energy threshold. Simultaneously, the left singular vectors corresponding to the L retained singular values ​​are selected to form a truncated left singular vector matrix, and the right singular vectors corresponding to the L retained singular values ​​are selected to form a truncated right singular vector matrix.

7. The method according to claim 1, characterized in that, The process involves projecting the high-dimensional beamforming matrix to be optimized into a low-dimensional beamforming matrix using the right singular vector matrix, and projecting the high-dimensional phase shift vector to be optimized into a low-dimensional phase shift vector using the left singular vector matrix, including: The high-dimensional beamforming matrix to be optimized on the base station side is operated on with the truncated right singular vector matrix, and mapped to the low-dimensional subspace spanned by the right singular vector matrix, thereby obtaining a low-dimensional beamforming matrix with significantly reduced dimensionality. The high-dimensional phase shift vector to be optimized on the reconfigurable smart surface is operated on with the truncated left singular vector matrix, and mapped to the low-dimensional subspace spanned by the left singular vector matrix to obtain the corresponding low-dimensional phase shift vector. The dimensions of the low-dimensional beamforming matrix and the low-dimensional phase shift vector are determined by the number of singular values ​​L retained, and the transmit power constraint satisfied by the high-dimensional beamforming matrix remains unchanged after the projection is mapped to the low-dimensional space.

8. The method according to claim 1, characterized in that, The problem is transformed using the weighted least mean square error criterion, and a low-rank block coordinate descent algorithm is employed to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence, including: By introducing receiver gain variables and weighting factor variables corresponding to each maritime user as auxiliary variables, the system and rate maximization problem is transformed into an equivalent weighted minimum mean square error minimization problem. The variables to be optimized are divided into four independent variable blocks: the low-dimensional beamforming matrix, the low-dimensional phase shift vector, the set of receiver gain variables, and the set of weighting factor variables. In each iteration of the low-rank block coordinate descent algorithm, the following sub-steps are executed sequentially, and when updating any variable block, the values ​​of the other three variable blocks are fixed: Update receiver gain variables: For each user, calculate the optimal receiver gain in the sense of minimum mean square error based on the current low-dimensional beamforming matrix and low-dimensional phase shift vector; Update weight factor variables: For each user, calculate the corresponding optimal weight factor based on the updated receiver gain; Update the low-dimensional beamforming matrix: Under the conditions of fixed receiver gain, weighting factor and low-dimensional phase shift vector, solve a convex quadratic programming problem constrained by total power to obtain the updated low-dimensional beamforming matrix. Update the low-dimensional phase shift vector: Under the conditions of fixed receiver gain, weighting factor and updated low-dimensional beamforming matrix, solve an unconstrained quadratic programming problem about the vector to obtain the updated low-dimensional phase shift vector; The iteration is repeated until the change in system and rate calculated in two adjacent iterations is less than the preset convergence threshold, at which point the algorithm is considered to have converged.

9. The method according to claim 1, characterized in that, Using the right singular vector matrix and the left singular vector matrix, the converged low-dimensional beamforming matrix and low-dimensional phase shift vector are reconstructed into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication, including: The converged low-dimensional beamforming matrix is ​​then processed with the truncated right singular vector matrix to reconstruct the high-dimensional beamforming matrix on the base station side. The converged low-dimensional phase shift vector is operated on with the truncated left singular vector matrix to reconstruct the high-dimensional continuous phase shift vector of the reconstructable smart surface. For each element in the high-dimensional continuous phase shift vector, a discrete phase shift value is selected from a preset discrete phase shift set according to the hardware resolution of the reconfigurable smart surface, such that the Euclidean distance between the complex exponential unity modulus corresponding to the discrete phase shift value and the element value is minimized. Based on all selected discrete phase shift values, construct the final reconfigurable smart surface discrete phase shift matrix; The high-dimensional beamforming matrix and the discrete phase shift matrix are the final parameters used to control the base station and the reconfigurable smart surface to assist marine communication.

10. A reconfigurable smart surface-assisted marine communication system based on LR-BCD, characterized in that, include: The first building module is used to build a communication model that includes base stations, a reconfigurable smart surface on the sea surface, and multiple maritime users. The second construction module is used to construct channel models from the base station to the reconfigurable smart surface, from the reconfigurable smart surface to the user, and from the base station to the user based on the ocean dual-path propagation model and the communication model. The channel decomposition module is used to represent the channel from the base station to the reconfigurable smart surface in the channel model as a channel matrix, perform singular value decomposition on it, and retain the main singular values ​​and the corresponding left and right singular vector matrices according to the energy threshold. The low-dimensional projection module is used to project the high-dimensional beamforming matrix to be optimized into a low-dimensional beamforming matrix using the right singular vector matrix, and to project the high-dimensional phase shift vector to be optimized into a low-dimensional phase shift vector using the left singular vector matrix. The problem construction module is used to construct a low-dimensional equivalent composite channel based on the low-dimensional beamforming matrix, the low-dimensional phase shift vector and the channel model, and to construct a power-constrained communication model and a rate maximization problem based on the low-dimensional equivalent composite channel. The problem transformation module is used to transform the problem using the weighted minimum mean square error criterion, and to alternately update the low-dimensional beamforming matrix and the low-dimensional phase shift vector until convergence using a low-rank block coordinate descent algorithm. The high-dimensional reconstruction module is used to reconstruct the converged low-dimensional beamforming matrix and low-dimensional phase shift vector into a high-dimensional beamforming matrix and a high-dimensional continuous phase shift vector, respectively, using the right singular vector matrix and the left singular vector matrix. The high-dimensional continuous phase shift vector is then discretized and quantized to obtain the final base station beamforming matrix and reconfigurable smart surface phase shift matrix used for auxiliary communication.

Images