Reconfigurable intelligent surface joint beamforming and interference nulling method

Abstract

The application discloses a reconfigurable intelligent metasurface joint beamforming and interference nulling method, and relates to the field of wireless communication and physical layer security. By using a narrowband pure power calibration technology based on a rotating electric field vector method, by switching the phase of each RIS unit in turn and recording the total power at the receiving end, only the secondary power measurement is needed to extract the relative amplitude and phase of each unit, so as to compensate for the initial inconsistency of the phases among the elements. On the basis of calibration, taking the continuous linear constraint minimum variance solution as the benchmark, the continuous phase optimization result is iteratively mapped to a 2-bit discrete phase codebook by a phase rotation algorithm, and the gain in the interference direction is minimized under the constraint that the gain loss in the expected direction is not more than 3dB, so that accurate beam pointing and deep interference nulling are realized. The application fully utilizes pure power measurement and discrete phase optimization, greatly reduces the system hardware complexity and measurement overhead, and verifies the effectiveness and engineering feasibility of the method on actual 2-bit passive RIS hardware.

Description

Technical Field

[0001] This invention relates to the fields of wireless communication and physical layer security technology, and in particular to a reconfigurable smart metasurface combined beamforming and interference nulling method. Background Technology

[0002] With the advancement of research into sixth-generation mobile communication systems, reconfigurable intelligent surfaces (RIS) have become a key enabling technology for improving wireless network coverage and spectral efficiency. RIS consists of a large number of low-cost, low-power passive subwavelength units. By dynamically controlling the reflection phase of each unit in a programmable manner, it can intelligently reshape the wireless propagation environment, creating equivalent reflection paths for users in signal dead zones, enhancing link quality, and expanding coverage.

[0003] In the field of Physical Layer Security (PLS), preventing unauthorized eavesdropping or suppressing the impact of strong interference sources on legitimate communication links is a core challenge. Traditional phased array-based beamforming techniques can suppress interference by precisely controlling amplitude and phase to focus signal energy onto the desired receiver and create nulls in other directions. However, this typically requires expensive high-resolution phase shifters and variable gain amplifiers, resulting in high cost and power consumption. While phase shaping is more power efficient, it still relies on multi-bit phase shifters, making it difficult to meet the requirements for low-cost and low-power deployment.

[0004] As a passive reflective array, RIS (Reflective Arrays) possess inherent advantages in low hardware cost and low power consumption, making them highly promising for physical layer security applications. However, practical RIS hardware typically employs 1-bit or 2-bit coarse phase resolution to further reduce system cost and complexity, and each reflective element exhibits unknown initial phase inconsistencies. Since interference nulling performance is highly sensitive to phase accuracy, traditional nulling algorithms designed based on ideal RIS models cannot achieve effective interference suppression on practical RIS hardware with coarse quantization precision. For example, the theoretically Continuous Phase Linear Constrained Minimum Variance (LCMV) algorithm can obtain extremely deep nulls under ideal conditions, but directly quantizing it to 2-bit phase results in a sharp deterioration in null depth, even becoming comparable to conventional beamforming performance, thus failing to achieve interference suppression.

[0005] Furthermore, in traditional RIS-assisted communication systems, beam management and channel estimation are hampered by the fact that the RIS lacks an RF link and cannot actively transmit or receive reference signals. This makes it difficult to obtain channel state information between the RIS and the transmitter and receiver using conventional methods. This poses a fundamental challenge to accurate beamforming and interference nulling of the RIS in practical deployment scenarios. Summary of the Invention

[0006] The technical problem this invention aims to solve is to overcome the shortcomings of existing technologies and provide a reconfigurable smart metasurface combined beamforming and interference nulling method. This method adapts to practical 2-bit passive RIS hardware constraints while simultaneously achieving precise beam pointing and deep interference nulling. This method can complete RIS unit calibration using only the total power information of the received signal, and achieves precise beam pointing and deep interference direction nulling under 2-bit discrete phase constraints through a specially designed phase rotation algorithm.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] A reconfigurable smart metasurface combined beamforming and interference nulling method proposed according to the present invention includes:

[0009] Step 1: For a reconfigurable smart metasurface with N RIS units, place the reference antenna in the known line-of-sight direction;

[0010] For each RIS unit, while keeping the phases of other RIS units unchanged, multiple phase change states are applied to it in sequence, and the total power value received by the reference antenna is recorded in each phase state.

[0011] Based on the total power received by the reference antenna in each phase state, the initial response vector of each RIS unit is extracted using the rotating electric field vector method.

[0012] The initial response vector of each extracted RIS unit is used to compensate and calibrate each RIS unit to obtain the calibrated effective channel;

[0013] Step 2: Based on the calibrated effective channel, solve the continuous linear constraint minimum variance beamforming problem to obtain a continuous phase weight vector that satisfies the requirements of maximizing the desired direction gain and nulling the interference direction.

[0014] The continuous phase weight vector is quantized into a 2-bit initial discrete phase codebook, and its phase values ​​are restricted to... ;

[0015] The phase rotation algorithm is used to iteratively optimize the process starting from a 2-bit initial discrete phase codebook: for each RIS unit, the four phase values ​​defined by the 2-bit discrete phase codebook are traversed. Under the condition that the loss of the desired directional gain relative to the continuous optimal solution does not exceed the preset gain loss threshold, the phase value that minimizes the interference directional gain is selected to update the phase of the RIS unit.

[0016] This process is repeated round by round. The iteration ends when the decrease in the interference direction gain corresponding to the discrete phase codebook in a complete round is lower than a preset value. This yields the optimal discrete phase codebook that satisfies the 2-bit phase constraint, thus achieving joint beamforming and interference nulling.

[0017] As a further optimization scheme for the reconfigurable intelligent metasurface combined beamforming and interference nulling method described in this invention, in step one, multiple phase change states are... , and The total number of measurements was Second-rate.

[0018] As a further optimization scheme for the reconfigurable smart metasurface joint beamforming and interference nulling method described in this invention, in step one, the initial response vector of each RIS unit is extracted. Specifically:

[0019] Step (1): Assume all phases of the RIS unit are... At that time, the amplitude and phase of the total field received by the reference antenna are respectively and , No. The initial complex response of each RIS unit is ;

[0020] Step (2): For the nth RIS element, 1≤n≤N, let the amplitude and phase of the field it generates at the reference antenna be respectively... and ,in for The absolute value, for The phase value;

[0021] Step (3): When the phase of the nth RIS unit is set to... At that time, the total power received by the reference antenna Represented as:

[0022] ;

[0023] make , , , , , All are intermediate variables. and They are the remaining units except for the nth RIS unit. The amplitude and phase of the total field formed by each RIS unit;

[0024] Step (4): Based on the phase measured at the reference antenna They are respectively , and Received power at time , , The values ​​of a, b, and c are obtained through algebraic solutions, and then the relative amplitude of the initial cascaded channel response of the nth RIS unit is extracted. and phase response ;

[0025] Step (5): Repeat steps (2) to (4) for all N RIS units to reconstruct the initial cascaded channel response vector from the transmitter to the RIS to the reference antenna. From the initial concatenated channel response vector Subtract the known transmitter-to-RIS channel from the middle and RIS to reference antenna channel The initial response vectors of each RIS unit are obtained. .

[0026] As a further optimization of the reconfigurable smart metasurface combined beamforming and interference nulling method described in this invention, the compensation calibration process is as follows: using Calculate the calibrated effective transmitter-to-RIS channel vector ,in This indicates element-wise division, thus eliminating the need for further beamforming calculations. The impact.

[0027] As a further optimization scheme for the reconfigurable intelligent metasurface joint beamforming and interference nulling method described in this invention, step two involves solving the continuous linear constraint minimum variance beamforming problem, specifically as follows:

[0028] definition These are the pitch and azimuth angles, respectively, in the desired far-field direction for the user. These represent the elevation and azimuth angles in the far-field direction of the interference source, respectively, and the desired user direction. and interference direction Corresponding channel vector , They are respectively:

[0029] ;

[0030] in, For the far-field steering vector in the desired user direction, The far-field steering vector is the direction of interference.

[0031] in The nth component for: , The nth component for: ,in The imaginary unit, For the operating wavelength, and These are the y-plane coordinates and z-plane coordinates of the nth RIS unit, respectively.

[0032] Constructing the constraint matrix and response vector The superscript T is the transpose operator;

[0033] Solve the linearly constrained minimum variance LCMV optimization problem:

[0034] , ;

[0035] in, For the optimal weights of unquantized RIS, It is the identity matrix, and the superscript H is the conjugate transpose;

[0036] Obtain the optimal weights of unquantized RIS The closed-form solution is:

[0037] ;

[0038] This leads to the unquantized RIS excitation vector. .

[0039] As a further optimization scheme for the reconfigurable smart metasurface joint beamforming and interference nulling method described in this invention, step two includes the phase rotation algorithm as follows:

[0040] Step 201, Initialization: The phase of each element is quantized to the closest 2-bit value. The initial discrete codebook is obtained. ;

[0041] Step 202: For each iteration k, randomly arrange all RIS unit indices;

[0042] Step 203: Arrange the four candidate phases sequentially according to the order obtained from the random arrangement in step 202. For the nth RIS unit, evaluate the four candidate phases in turn. ;

[0043] Construct temporary excitation vectors Except for the nth RISC unit, the phase is set to In addition, the remaining RIS units maintain the discrete phase codebook for the current iteration. Phase value in;

[0044] Calculate the temporary desired directional gain and temporary interference directional power They are respectively:

[0045] ;

[0046] ;

[0047] in This indicates that when the nth RIS unit takes a candidate phase Gain in the desired user direction This indicates that when the nth RIS unit takes a candidate phase Power in the direction of interference It is a column vector consisting entirely of 1s;

[0048] Step 204, With continuous optimal expected gain Comparison, among which For continuous optimal weights The gain generated in the desired direction, from satisfying Of all candidate phases, select The minimum phase value; if this phase value is different from the current phase value of the nth RIS unit in the discrete phase codebook, then the phase of the nth RIS unit in the discrete phase codebook is updated with this phase value.

[0049] Step 205: After all N RIS units have been traversed in one iteration, the interference direction power in step 203 is... The total decrease after this phase update is less than The phase rotation algorithm terminates when the maximum number of iterations is reached.

[0050] Step 206: Output the optimal 2-bit RIS phase excitation vector in It is a natural constant. The imaginary unit, The phase value of the nth RIS unit after optimization by the phase rotation algorithm. This represents the complex excitation coefficient corresponding to this unit.

[0051] As a further optimization of the reconfigurable smart metasurface combined beamforming and interference nulling method described in this invention, the reconfigurable smart metasurface is a 2-bit passive array, operating in the following frequency band: The phase state is Phase accuracy is .

[0052] As a further optimization of the reconfigurable smart metasurface joint beamforming and interference nulling method described in this invention, in step one, the reference antenna is located in the near-field region of the RIS, and the physical distance between the reference antenna and the RIS is less than the far-field conditional distance. Where D is the maximum aperture size of RIS, This is the operating wavelength.

[0053] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0054] (1) This invention proposes a joint beamforming and interference nulling framework for practical 2-bit passive RIS hardware. This framework compensates for the initial phase inconsistency between each reflecting unit through a fast calibration step using only power measurement, thereby reducing the main lobe pointing error from The above decreased to Within this range, beamforming accuracy is restored; then, a deep interference null is achieved under a 2-bit discrete phase constraint using a phase rotation algorithm. Experiments show that the null depth can reach 21.3dB and 35.7dB, while the expected directional gain loss is only 1.8dB and 2.4dB, respectively, which is within the 3dB design constraint.

[0055] (2) The calibration steps of the present invention only require The initial channel response of all RIS units can be extracted by the next power measurement (for The RIS prototype with each unit requires only 243 measurements, eliminating the need for complex vector network analyzers or reference signal transceivers, significantly reducing system hardware complexity and measurement overhead.

[0056] (3) The phase rotation algorithm of the present invention is designed directly for 2-bit phase constraints, which solves the problem of the rapid deterioration of null performance after direct quantization of traditional continuous phase algorithms. Simulation results show that the null depth of the continuous LCMV solution after direct quantization is only about 11dB, which is comparable to conventional beamforming; while the algorithm proposed in this invention can achieve a null depth of more than 20dB under the same conditions, which is an improvement of more than 10dB.

[0057] (4) The method of the present invention is applicable to real-world applications. A complete experimental verification was conducted on a 121-unit 2-bit passive RIS prototype and in a 5.8GHz anechoic chamber measurement environment, proving the feasibility and effectiveness of the method in practical engineering. It is applicable to strong active interference scenarios such as full-duplex self-interference suppression, millimeter-wave co-channel interference cancellation, and cross-system interference management in terminal-to-dual (D2D) cellular networks. Attached Figure Description

[0058] Figure 1 This is a schematic diagram of a RIS-assisted secure communication system;

[0059] Figure 2 This is a block diagram of the measurement system;

[0060] Figure 3 This is a schematic diagram of the experimental geometry configuration;

[0061] Figure 4a These are the initial amplitude response measurement results of each unit in the RIS;

[0062] Figure 4b These are the initial phase response measurement results of each unit in the RIS;

[0063] Figure 5 It is a comparison of the far-field radiation patterns of the RIS before and after calibration;

[0064] Figure 6 This is a comparison of the far-field radiation pattern measurement results for scenario A (desired direction). Interference direction (a) is reference beamforming, (b) is LCMV direct quantization, and (c) is the phase rotation method proposed in this invention.

[0065] Figure 7 This is a comparison of the far-field radiation pattern measurement results for scene B (desired direction). Interference direction (a) is reference beamforming, (b) is LCMV direct quantization, and (c) is the phase rotation method proposed in this invention.

[0066] Figure 8 This is a flowchart of the method of the present invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0068] The present invention proposes a joint beamforming and interference nulling method based on a 2-bit reconfigurable smart metasurface. The core idea is as follows: First, the REV calibration method based on pure power measurement is used to quickly obtain the initial channel response of each RIS unit to compensate for hardware inconsistencies. Then, based on the continuous LCMV beamforming solution, the codebook is iteratively optimized under 2-bit discrete phase constraints through a phase rotation algorithm to achieve joint beamforming and interference nulling.

[0069] Figure 1 This illustrates a typical scenario for a RIS-assisted secure communication system. In this scenario, the direct path between the transmitter (Alice) and the legitimate user (Bob) is blocked by an obstacle, requiring the use of RIS reflection to form a cascaded path for communication. Simultaneously, a potential eavesdropper (Eve) is located in the direction of interference, necessitating the use of RIS's spatial nulling capability to suppress signal leakage in that direction. A system with... The 2-bit passive RISC unit is placed in In the plane, the center is located at the origin. The transmitter is located at a known near-field position. The legitimate user is located in the far field direction. The eavesdropper is located in the far field. .

[0070] The apparatus system of the method of the present invention includes: a 2-bit passive reconfigurable smart metasurface, a transmitting antenna (feed), a reference antenna (probe), a vector network analyzer, a low-noise amplifier, and a control computer. Figure 2 As shown, the transmitting antenna radiates a signal to the RIS via a horizontally polarized feed. The signal reflected by the RIS is received by a vertically polarized waveguide probe, amplified by a low-noise amplifier, and then measured by a vector network analyzer. The parameters control computer is responsible for generating RIS phase control commands, instrument trigger control, data storage, and subsequent data processing.

[0071] Figure 3 The experimental geometry configuration is shown. In this embodiment, the RIS uses an 11×11 2-bit passive array with 121 cells, and operates at a frequency band of [frequency band missing]. The element spacing is approximately 0.5 times the wavelength, and the phase state is... Phase accuracy is The feed antenna is placed at... At this point, the reference probe is located in the line of sight direction. At this location, the probe is mounted on a planar near-field scanning frame, allowing it to operate within the yz plane. Precision movement.

[0072] Figure 8 The complete flow of the method of the present invention is shown. The following detailed explanation, in two steps, is based on specific embodiments.

[0073] Step 1: Rapid calibration based on pure power measurement.

[0074] First, phase position all RIS units as follows: Then, for the nth unit, while keeping the phase of all other units constant... In this case, the phase of the unit is switched sequentially to , and And a vector network analyzer was used to record the data received by the reference antenna in each phase state. Parameter amplitude. For operating frequency Only retain this frequency point The amplitude is used as the received power information.

[0075] Assume that all units of the RIS have phase. The amplitude and phase of the total field received by the reference antenna are respectively and The amplitude and phase of the field generated by the nth unit are respectively and Except for the nth unit, the rest The amplitude and phase of the total field formed by each unit are denoted as follows: and .

[0076] When the phase of the nth RISC unit is set to At that time, the total power received by the reference antenna can be expressed as the vector superposition of two electric fields:

[0077]

[0078] Define intermediate variables:

[0079]

[0080] The received power then simplifies to the standard cosine form:

[0081]

[0082] In the measurement , and Corresponding receiving power , , Then, the following system of equations was solved:

[0083]

[0084]

[0085]

[0086] Intermediate variables can be obtained:

[0087]

[0088] Then, the amplitude and phase corresponding to the nth RIS unit are obtained:

[0089]

[0090] After traversing all N = 121 units, the initial Tx-RIS-Rx cascaded channel response vector is obtained:

[0091]

[0092] in, Given a known spherical wavefront channel from the transmitter to the nth RIS unit, Let n be the Euclidean distance from the transmitter to the nth RIS unit. This refers to the position of the nth unit. For the known channel from the nth RIS element to the reference antenna, Let be the distance from the nth RIS element to the reference antenna. Since the probe distance to the RIS is 0.5m, it is much smaller than the distance under far-field conditions at 5.8GHz. (in This is the maximum aperture size of the RIS. (where the wavelength is 1), therefore the reference antenna is located in the near-field region of the RIS.

[0093] from Subtract the known from the middle and The channel contribution allows us to extract the initial complex response of each RIS unit. ,in The relative amplitude factor of the nth unit. This represents the initial phase deviation of the unit.

[0094] Figure 4a These are the initial amplitude response measurement results of each unit in the RIS; Figure 4b These are the initial phase response measurement results for each unit of the RIS. The measurement results show that the relative phase dispersion between units is approximately... This confirms the necessity of implementing calibration.

[0095] To eliminate The impact on beamforming accuracy is defined as follows: the calibrated effective transmitter-RIS channel vector is:

[0096]

[0097] in This indicates element-wise division. The effective channel will automatically compensate for the initial inconsistencies in the RIS units during subsequent beamforming calculations.

[0098] Figure 5 A comparison of the far-field radiation patterns of the RIS before and after calibration is shown. Plane pairs and One-dimensional far-field pattern scanning measurements were performed in two directions. Simulation results show that 2-bit phase quantization itself only causes approximately The main lobe gain is lost without affecting the main lobe direction, while the uncalibrated beam exhibits different characteristics at the four measurement angles. , , and The peak pointing error. After calibration, these pointing errors were reduced to... , , and This verified the effectiveness of the calibration procedure in restoring beamforming accuracy.

[0099] Step 2: Discrete phase optimization based on phase rotation.

[0100] Valid channel after calibration Based on this, firstly, targeting the intended user group. and interference direction Constructing LCMV beamforming problem.

[0101] For those located in the far field direction RIS-guided vector users The nth component is:

[0102]

[0103] in and These are the y-coordinate and z-coordinate of the nth RIS unit, respectively.

[0104] Define the calibrated channel vectors corresponding to the desired direction and the interference direction as follows:

[0105]

[0106]

[0107] Construct the LCMV optimization problem:

[0108]

[0109] in ( () is the numerically stabilized identity matrix. For the constraint matrix, Let be the response vector. This constraint means that the response in the desired direction is fixed at 1 (unity gain), and the response in the disturbance direction is forced to be 0 (complete null).

[0110] The above optimization problem has a closed-form solution:

[0111]

[0112] The solution This is the optimal weight vector for continuous phase and continuous amplitude. Its gain (linear scale) in the desired direction is:

[0113]

[0114] from After removing the influence of the calibrated transmitter-RIS channel, the desired continuous RIS excitation vector is obtained:

[0115]

[0116] While continuous LCMV solutions can theoretically achieve extremely deep zeros (>300dB), they assume that each RIS unit can provide continuous and precise amplitude and phase control, which is severely mismatched with the hardware capabilities of a practical 2-bit passive RIS. Directly addressing... The phase of each element is quantized using 2 bits (quantized to the closest value). The resulting null depth will deteriorate sharply to only about 11 dB, which is almost equivalent to conventional beamforming without null processing.

[0117] To address this, the present invention proposes a phase rotation algorithm that iteratively optimizes the phase of each unit based on the direct quantization results while maintaining a 2-bit phase constraint. The goal is to minimize the interference direction gain while keeping the expected direction gain loss within 3dB.

[0118] The specific steps of the phase rotation algorithm are detailed below:

[0119] Step (1): Initialization. [The following text appears to be a separate, unrelated section:] ... Phase of each element Quantize to the nearest 2-bit value, that is:

[0120]

[0121] Obtain the initial discrete phase codebook .

[0122] Step (2): Iterative optimization. Let the maximum number of iterations be set. For each iteration Perform the following sub-steps:

[0123] (2a) Randomly arrange the indices of the N RIS units (to shuffle the processing order and avoid getting trapped in local optima).

[0124] (2b) Perform the following steps on each unit n in the order they are arranged:

[0125] For four candidate phases For each of the elements, construct a temporary excitation vector. This vector, except for the nth element, is set as... In addition, the rest Each unit remains current. The phase value remains unchanged.

[0126] Calculate the temporary response:

[0127]

[0128] Calculate the desired directional gain (dB) and interference directional power (linear):

[0129]

[0130] in It is a column vector of all 1s. This represents the coherent summation of all channel signals.

[0131] Screening constraints: From the four candidate phases, first eliminate those that do not satisfy the following constraints:

[0132]

[0133] That is, the expected directional gain loss should not exceed 3dB.

[0134] Among the candidate phases that satisfy the constraints, select the one that makes... smallest This becomes the new phase value for that cell. If this value differs from the current one, then the nth cell is updated.

[0135] (2c) Convergence judgment. Calculate the linear interference power after traversing all units in this iteration. The total decrease. If this decrease is lower than a preset threshold. If the condition is met, it indicates that further iterations have negligible effect on interference suppression, and the algorithm converges and terminates; otherwise, it continues to the next iteration. The algorithm continues until the maximum number of iterations is reached. It can also be forcibly terminated at any time.

[0136] Step (3): Output. Obtain the optimal RIS excitation vector that satisfies the 2-bit phase constraint:

[0137]

[0138] By applying this codebook to RIS hardware, the joint design goal of high-gain beamforming in the desired direction and deep nulls in the interference direction can be achieved.

[0139] To verify the effectiveness of this method, experiments were conducted in a darkroom environment using a 121-cell 2-bit passive RIS prototype in two typical scenarios. Scenario A: Desired Direction Interference direction Scenario B: Desired Direction Interference direction For each scenario, three RIS codebooks were configured for comparison: (1) Base codebook: beamforming weights in the desired direction were directly quantized (without any processing of the interference direction); (2) LCMV direct quantization codebook: continuous LCMV solution was directly quantized by 2 bits; (3) Phase rotation algorithm codebook proposed in this invention.

[0140] During measurement, the waveguide probe performs a rectangular grid scan on the yz plane at a distance of 0.5m from RIS, with the azimuth angle... cover Pitch angle cover The step size is 2.5cm, which meets the Nyquist sampling criterion. A total of [number] samples were collected. A sampling point. The two-dimensional far-field radiation pattern is reconstructed by near-field to far-field transformation (NF2FF).

[0141] Figure 6 This is a comparison of the far-field radiation pattern measurement results for scenario A (desired direction). Interference direction ),in, Figure 6 (a) in the diagram represents the reference beamforming. Figure 6 (b) in the figure represents direct LCMV quantization. Figure 6 (c) in the figure represents the phase rotation method proposed in this invention. The null depth of the reference beamforming in the interference direction is only 10.4 dB; the null depth of LCMV direct quantization is slightly improved to 11.9 dB, verifying the limitations of the direct quantization method; while the null depth of the method proposed in this invention reaches 21.3 dB, an improvement of 10.9 dB relative to the reference, with the expected directional gain decreasing by only 1.8 dB, and the main lobe direction remaining unchanged. .

[0142] Figure 7 This is a comparison of the far-field radiation pattern measurement results for scene B (desired direction). Interference direction ),in, Figure 7 (a) in the diagram represents the reference beamforming. Figure 7 (b) in the figure represents direct LCMV quantization. Figure 7 (c) in the figure represents the phase rotation method proposed in this invention. The null depth of the reference beamforming in the interference direction is 10.5 dB; the direct quantization of LCMV is 11.3 dB; while the null depth of the method proposed in this invention reaches 35.7 dB, an improvement of 25.2 dB, with a desired directional gain decrease of only 2.4 dB, and the main lobe direction is... For this scenario, the half-power beamwidths of the method of the present invention in the azimuth and elevation directions are respectively... and .

[0143] The experimental results of the two scenarios consistently show that: (1) 2-bit phase coarse quantization severely restricts the performance of the LCMV null algorithm, and direct quantization cannot achieve meaningful interference suppression effect; (2) The phase rotation algorithm designed in this invention can effectively overcome the performance loss caused by coarse quantization and achieve a depth of more than 20dB under strict expected gain loss constraints; (3) After calibration compensation, the main lobe pointing accuracy is significantly restored.

[0144] In summary, the joint beamforming and interference nulling method based on a 2-bit reconfigurable smart metasurface proposed in this invention solves the core challenges of joint beamforming and interference nulling on practical low-precision RIS hardware through a two-step strategy of rapid calibration using only power measurement and specially designed phase rotation discretization optimization. This method offers advantages such as low measurement overhead, low hardware complexity, strong nulling capability, and high beam pointing accuracy, providing a feasible solution for practical RIS deployment in scenarios requiring physical layer security and strong interference suppression.

[0145] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for combining reconfigurable smart metasurfaces with beamforming and interference nulling, characterized in that, include: Step 1: For a reconfigurable smart metasurface with N RIS units, place the reference antenna in the known line-of-sight direction; For each RIS unit, while keeping the phases of other RIS units unchanged, multiple phase change states are applied to it in sequence, and the total power value received by the reference antenna is recorded in each phase state. Based on the total power received by the reference antenna in each phase state, the initial response vector of each RIS unit is extracted using the rotating electric field vector method. The initial response vector of each extracted RIS unit is used to compensate and calibrate each RIS unit to obtain the calibrated effective channel; Step 2: Based on the calibrated effective channel, solve the continuous linear constraint minimum variance beamforming problem to obtain a continuous phase weight vector that satisfies the requirements of maximizing the desired direction gain and nulling the interference direction. The continuous phase weight vector is quantized into a 2-bit initial discrete phase codebook, and its phase values ​​are restricted to... ; The phase rotation algorithm is used to iteratively optimize the process starting from a 2-bit initial discrete phase codebook: for each RIS unit, the four phase values ​​defined by the 2-bit discrete phase codebook are traversed. Under the condition that the loss of the desired directional gain relative to the continuous optimal solution does not exceed the preset gain loss threshold, the phase value that minimizes the interference directional gain is selected to update the phase of the RIS unit. This process is repeated round by round. When the decrease in the interference direction gain corresponding to the discrete phase codebook is lower than the preset value in a complete round of iteration, the iteration is terminated, and the optimal discrete phase codebook that satisfies the 2-bit phase constraint is obtained, thus achieving joint beamforming and interference nulling.

2. The method for reconfigurable smart metasurface combined with beamforming and interference nulling according to claim 1, characterized in that, In step one, multiple phase change states are: , and The total number of measurements was Second-rate.

3. The method for reconfigurable smart metasurface combined with beamforming and interference nulling according to claim 1, characterized in that, In step one, the initial response vector of each RIS unit is extracted. Specifically: Step (1): Assume all phases of the RIS unit are... At that time, the amplitude and phase of the total field received by the reference antenna are respectively and , No. The initial complex response of each RIS unit is ; Step (2): For the nth RIS element, 1≤n≤N, let the amplitude and phase of the field it generates at the reference antenna be respectively... and ,in for The absolute value, for The phase value; Step (3): When the phase of the nth RIS unit is set to... At that time, the total power received by the reference antenna Represented as: ； make , , , , , All are intermediate variables. and They are the remaining units except for the nth RIS unit. The amplitude and phase of the total field formed by each RIS unit; Step (4): Based on the phase measured at the reference antenna They are respectively , and Received power at time , , The values ​​of a, b, and c are obtained through algebraic solutions, and then the relative amplitude of the initial cascaded channel response of the nth RIS unit is extracted. and phase response ; Step (5): Repeat steps (2) to (4) for all N RIS units to reconstruct the initial cascaded channel response vector from the transmitter to the RIS to the reference antenna. From the initial concatenated channel response vector Subtract the known transmitter-to-RIS channel from the middle and RIS to reference antenna channel The initial response vectors of each RIS unit are obtained. .

4. The method for combined reconfigurable smart metasurface beamforming and interference nulling according to claim 3, characterized in that, The compensation calibration process is as follows: using Calculate the calibrated effective transmitter-to-RIS channel vector ,in This indicates element-wise division, thus eliminating the need for further beamforming calculations. The impact.

5. The method for combined reconfigurable smart metasurface beamforming and interference nulling according to claim 4, characterized in that, Step two involves solving the continuous linearly constrained minimum variance beamforming problem, specifically: definition These are the pitch and azimuth angles, respectively, in the desired far-field direction for the user. These represent the elevation and azimuth angles in the far-field direction of the interference source, respectively, and the desired user direction. and interference direction Corresponding channel vector , They are respectively: ； in, For the far-field steering vector in the desired user direction, The far-field steering vector is the direction of interference. in The nth component for: , The nth component for: ,in The imaginary unit, For the operating wavelength, and These are the y-plane coordinates and z-plane coordinates of the nth RIS unit, respectively. Constructing the constraint matrix and response vector The superscript T is the transpose operator; Solve the linearly constrained minimum variance LCMV optimization problem: ， ； in, For the optimal weights of unquantized RIS, It is the identity matrix, and the superscript H is the conjugate transpose; Obtain the optimal weights of unquantized RIS The closed-form solution is: ； This leads to the unquantized RIS excitation vector. .

6. The method for combined reconfigurable smart metasurface beamforming and interference nulling according to claim 5, characterized in that, In step two, the phase rotation algorithm includes: Step 201, Initialization: The phase of each element is quantized to the closest 2-bit value. The initial discrete codebook is obtained. ; Step 202: For each iteration k, randomly arrange all RIS unit indices; Step 203: Arrange the four candidate phases sequentially according to the order obtained from the random arrangement in step 202. For the nth RIS unit, evaluate the four candidate phases in turn. ; Construct temporary excitation vectors Except for the nth RISC unit, the phase is set to In addition, the remaining RIS units maintain the discrete phase codebook for the current iteration. Phase value in; Calculate the temporary desired directional gain and temporary interference directional power They are respectively: ； ； in This indicates that when the nth RIS unit takes a candidate phase Gain in the desired user direction, This indicates that when the nth RIS unit takes a candidate phase Power in the direction of interference It is a column vector consisting entirely of 1s; Step 204, With continuous optimal expected gain Comparison, among which For continuous optimal weights The gain generated in the desired direction, from satisfying Of all candidate phases, select The minimum phase value; if this phase value is different from the current phase value of the nth RIS unit in the discrete phase codebook, then the phase of the nth RIS unit in the discrete phase codebook is updated with this phase value. Step 205: After all N RIS units have been traversed in one iteration, the interference direction power in step 203 is... The total decrease after this phase update is less than The phase rotation algorithm terminates when the maximum number of iterations is reached. Step 206: Output the optimal 2-bit RIS phase excitation vector in It is a natural constant. The imaginary unit, The phase value of the nth RIS unit after optimization by the phase rotation algorithm. This represents the complex excitation coefficient corresponding to this unit.

7. A method for reconfigurable smart metasurface combined with beamforming and interference nulling according to any one of claims 1 to 6, characterized in that, The reconfigurable smart metasurface is a 2-bit passive array, operating in the following frequency band: The phase state is Phase accuracy is .

8. The method for reconfigurable smart metasurface combined with beamforming and interference nulling according to claim 1, characterized in that, In step one, the reference antenna is located in the near-field region of the RIS, and the physical distance between the reference antenna and the RIS is less than the far-field conditional distance. Where D is the maximum aperture size of RIS, This is the operating wavelength.

Images