Active reconfigurable intelligent metasurface beam training method based on actual amplification model

Abstract

The invention discloses an active reconfigurable intelligent metasurface beam training method based on an actual amplification model, and belongs to the field of wireless communication. The method comprises the following steps: generating a final transmitting beam forming vector codebook based on an initial transmitting beam forming codebook and a transmitting power codebook, so that signal power incident on the active reconfigurable intelligent metasurface is in a power interval for triggering linear amplification of a reflection amplifier; a reflection beam forming codebook is generated based on an array steering vector of an active reconfigurable intelligent metasurface-assisted cascade channel arranged in a uniform plane array form; and comparing and judging the signal-to-noise ratio of the received signal to obtain the optimal transmitting power and the optimal reflected beam forming vector, and obtaining the maximum total rate of the communication system. According to the method, the power of the signal incident to the active reconfigurable intelligent metasurface can enable the amplifier to work in a linear amplification interval, effective amplification of the signal is achieved, and complete feedback of CSI is not needed. In this way, the feedback overhead is reduced while the beam forming effect is ensured.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to an active reconfigurable intelligent metasurface beam training method based on a practical amplification model. Background Art

[0002] In recent years, with the rapid development of advanced technologies, reconfigurable smart metasurfaces have attracted great interest from researchers as a revolutionary technology that can help meet the various needs of 6G communication systems. A reconfigurable smart metasurface is a plane composed of low-cost passive reflective units. Each reflective unit can independently cause controllable amplitude and / or phase changes in the incident signal, thereby realizing an intelligent and programmable wireless environment. However, in the cascaded channel consisting of the base station-reconfigurable smart metasurface and the reconfigurable smart metasurface-user link, a fatal "double fading" effect occurs, which causes the path loss of the signal reflected from the reconfigurable smart metasurface to be several orders of magnitude larger than that of the unobstructed direct link. Therefore, in actual deployment, when the reconfigurable smart metasurface is not close enough to the user or base station, the system can only achieve a small performance gain.

[0003] In order to effectively solve this problem, an active reconfigurable smart metasurface architecture equipped with a reflection amplifier is proposed. Compared with the traditional passive reconfigurable smart metasurface, the active reconfigurable smart metasurface can not only adjust the phase shift of the incident signal, but also amplify the weak incident signal propagated through the base station-reconfigurable smart metasurface channel at the electromagnetic level, thereby overcoming the influence of the "double fading" effect. Without the need for expensive high-power radio frequency (RF) chains, active reconfigurable smart metasurfaces can be implemented through hardware-efficient active components such as tunnel diodes (see reference: Amato F, W. Peterson C, P. Degnan B, et al. Tunneling RFID tags for long-range and low-power microwave applications [J]. IEEE Journal of Radio Frequency Identification, 2018, 2 (2): 93-103.).

[0004] Due to the low bias requirement and negative differential resistance characteristics, tunnel diodes are being used to make reflection amplifiers with high gain. The tunnel diode reflection amplifier is designed based on the concept of oscillators and implemented on microstrip lines. When properly biased, it will lock the frequency of the input signal; when the impedance of the reflection amplifier matches the impedance of the microstrip line, the reflection amplifier converts the DC power provided by the bias source into RF power in the form of injection locking, which is enhanced to the incident signal when the incident signal is reflected. The occurrence of locking and the impedance of the reflection amplifier are both related to the amplitude of the incident signal. Most active reconfigurable smart metasurface beamforming design studies are based on an ideal reflection amplification signal model without considering the characteristics of the amplifier hardware structure. Taking the tunnel diode-based reflection amplifier as an example, due to the hardware characteristics and limitations of the reflection amplifier circuit, the linear amplification mechanism of the amplifier can only be triggered when the power of the incident signal falls within a specific linear amplification interval, effectively amplifying the incident signal without distortion; when the incident signal power falls outside this specific interval, due to the characteristics of the amplifier circuit, the device can only perform nonlinear amplification with significant distortion, or only reflect the signal without amplification. Existing active reconfigurable smart metasurface beamforming designs do not consider the actual response of the amplification circuit to incident signals with different intensity levels, which will lead to performance degradation when inaccurate amplification modeling is used in active reconfigurable smart metasurface optimization.

[0005] For the passive reconfigurable smart metasurface assisted system, the design of the transmitter at the base station and the receiver at the user only requires the cascaded channel state information (CSI) of the transmitter-reconfigurable smart metasurface-receiver link. Due to the introduction of the amplifier, for the active reconfigurable smart metasurface assisted system, the transceiver design needs to consider the amplified reconfigurable smart metasurface transmit power and thermal noise, which requires separate CSI for the base station-reconfigurable smart metasurface link and the reconfigurable smart metasurface-user link. However, due to the lack of RF chain, the active reconfigurable smart metasurface cannot estimate two separate channels. Therefore, the existing work on active reconfigurable smart metasurfaces assumes that the base station-reconfigurable smart metasurface and the reconfigurable smart metasurface-user link have perfect CSI, which is difficult to obtain in practice. The introduction of the codebook allows the receiver to only select the best codeword in the predefined codebook for feedback without the need to fully feedback the CSI. This approach reduces the feedback overhead while ensuring the beamforming effect. Therefore, it is very important to design an active reconfigurable smart metasurface beam training method based on a practical amplification model. Summary of the invention

[0006] In order to solve the problems that most of the existing active reconfigurable intelligent metasurface assisted communication systems are based on ideal amplification models, ignoring the hardware limitations of actual amplifiers, resulting in significant signal distortion and additional performance loss when the incident signal power exceeds the linear amplification range of the amplifier; and the problem that obtaining accurate cascade link individual CSI for active reconfigurable intelligent metasurface reflection phase shift beamforming design is limited, the present invention provides an active reconfigurable intelligent metasurface beam training method based on an actual amplification model, which realizes the actual and effective linear amplification effect of the amplifier, ensures that the active reconfigurable intelligent metasurface can still achieve effective linear signal amplification under the condition of incomplete CSI feedback, improves beamforming performance, optimizes the overall rate performance of the communication system while reducing dependence on feedback information.

[0007] The objective of the present invention is achieved through the following technical solutions:

[0008] An active reconfigurable intelligent metasurface beam training method based on an actual amplification model is applicable to a communication system including a multi-antenna base station, an active reconfigurable intelligent metasurface and a single-antenna receiving end device, wherein the direct path between the base station and the receiving end is blocked; the base station is equipped with M transmitting antennas, which are arranged in the form of a uniform linear array; the active reconfigurable intelligent metasurface is equipped with N reflection units, which are arranged in the form of a uniform planar array, each reflection unit is connected to a reflection amplifier based on a tunnel diode, and phase shift adjustment and electromagnetic level power amplification are performed on the incident signal; the base station antenna spacing and the reflection unit spacing of the active reconfigurable intelligent metasurface are both set to half the signal wavelength; through codebook generation, power control based on the actual amplification model and exhaustive search, the main path beamforming between the base station and the active reconfigurable intelligent metasurface is realized, and the signal is controlled within the linear amplification range of the reflection amplifier based on the tunnel diode to achieve accurate beam steering and signal amplification; the method specifically includes the following steps:

[0009] Step 1: Design a beamforming codebook for a base station based on a uniform linear array arrangement, and generate an initial transmit beamforming vector codebook based on the channel characteristics from the base station to the active reconfigurable smart metasurface; and then further generate the final base station transmit beamforming vector codebook through the transmit power codebook;

[0010] Step 2, generating an active reconfigurable intelligent metasurface reflection beamforming codebook based on an array steering vector of a cascade channel assisted by an active reconfigurable intelligent metasurface arranged in a uniform planar array;

[0011] Step 3: Perform codebook training to determine the response of the tunnel diode-based reflection amplifier by calculating the incident signal power of the active reconfigurable intelligent metasurface under each transmit beamforming vector codeword one by one. Then, search the reflection beamforming vector codewords pre-calculated in the reflection beamforming codebook one by one to determine the codeword that can maximize the signal-to-noise ratio of the received signal at the receiving end, that is, the optimal transmit beamforming vector codeword and the optimal reflection beamforming vector codeword, so as to obtain the maximum total rate of the communication system.

[0012] Further, in step 1, the initial transmit beamforming vector codebook is generated as follows:

[0013] The channel G between the base station and the active reconfigurable smart metasurface is decomposed by SVD, and the obtained right singular vector is used as the candidate direction of the base station beamforming to optimize the signal gain in the main path direction of the channel; the formula is:

[0014] G=UΣV H (1)

[0015] in: represents the channel from the base station to the active reconfigurable smart metasurface; U is the left singular vector matrix, which describes the receiving direction space of the channel; Σ is a diagonal matrix, whose diagonal elements are arranged in descending order according to the non-negative singular values; V is the right singular vector matrix, which describes the transmitting direction space of the channel;

[0016] The initial transmit beamforming vector codebook generation method is expressed as:

[0017]

[0018] The right singular vector V(:,1) corresponding to the maximum singular value represents the optimal transmission direction at the transmitting end; V * (:, 1) represents the conjugate of V(:, 1), which enables the transmitted signal to obtain the best gain in the main path direction, thereby maximally concentrating the signal energy on the main path from the transmitter to the active reconfigurable smart metasurface, improving the transmission signal-to-noise ratio and system performance.

[0019] Furthermore, in step 1, in order to further optimize the incident signal power, the base station transmit power is adjusted by a transmit power codebook, which contains a series of candidate power values. By traversing the transmit power codebook, the optimal transmit power value that can make the signal power incident on the active reconfigurable smart metasurface meet the linear range of the amplifier is selected to avoid signal distortion and performance degradation due to abnormal amplifier amplification. Therefore, the codeword of the final base station transmit beamforming vector codebook is expressed as:

[0020]

[0021] Among them, P BS Represents the base station's transmit power codebook, which is generated at equal intervals according to the step size within the set power range; P BS (i) represents the base station transmission power codebook P BS The ith value of .

[0022] Furthermore, in step 2, by modeling the array steering characteristics of the cascaded channel between the base station and the active reconfigurable intelligent metasurface, a phase-shifted reflection codebook arranged in a uniform planar array is generated to ensure the phase adjustment and reflection direction optimization of the reflected signal; the active reconfigurable intelligent metasurface reflection beamforming codebook is expressed as:

[0023]

[0024] Wherein, a(α,β) represents the array response vector of the active reconfigurable smart metasurface related to the uniform planar array arrangement; α, β represent the spatial angles related to the active reconfigurable smart metasurface, n1 = [0, 1, ..., N1-1] T , n2=[0,1,...,N2-1] T , the number of active reconfigurable smart metasurface reflection units N = N1 × N2, where N1 and N2 are the number of columns and rows of the active reconfigurable smart metasurface, respectively.

[0025] Furthermore, in step 3, an exhaustive search method is used to perform codebook training, and the specific process is as follows:

[0026] For the i-th transmit beamforming vector codeword w(i), the incident signal power of the n-th reflector unit on the active reconfigurable smart metasurface is expressed as:

[0027] p in,n =|G(n,:)w(i) 2 (5)

[0028] Where G(n,:) represents the nth row of channel G, corresponding to the nth reflection unit of the active reconfigurable smart metasurface, n = 1,...,N;

[0029] In the present invention, the linear amplification mechanism of the tunnel diode reflection amplifier that triggers the active reconfigurable smart metasurface is p in,n The tunnel diode reflection amplifier linearly amplifies it at a constant multiple. in,n When p is less than the linear amplification requirement, the active reconfigurable smart metasurface only adjusts the phase and reflects the signal because the impedance of the tunnel diode does not match the microstrip line and the incident signal power is too weak to trigger the locking phenomenon. in,nWhen it is greater than the linear amplification requirement, the codebook training procedure stops to avoid abnormal amplification of the amplifier and cause additional uncontrollable losses.

[0030] The signal received by the user is expressed as:

[0031]

[0032] in: represents the channel from the active reconfigurable smart metasurface to the user; represents the transmission signal of the base station, x = w(i)s, s represents the transmission signal, and satisfies They represent the amplification factor matrix and reflection phase shift matrix of the active reconfigurable smart metasurface, The first codebook representing the active reconfigurable smart metasurface reflective beamforming Code words, represents the thermal noise introduced by the active reconfigurable smart metasurface, n represents the additive Gaussian white noise at the user, with a mean of 0 and a variance of σ 2 ;

[0033] The user receiving signal-to-noise ratio is expressed as:

[0034]

[0035] The total rate of the communication system is expressed as:

[0036] sum_rate=log(1+SNR) (8)

[0037] By searching the transmit beamforming power codewords pre-calculated in the transmit beamforming codebook and the reflection beamforming vector codewords pre-calculated in the reflection beamforming codebook one by one, the transmit beamforming vector codeword w that maximizes the signal-to-noise ratio of the received signal at the receiving end shown in formula (7) is found. best Active reconfigurable smart metasurface reflective beamforming vector codeword

[0038] Compared with the prior art, the present invention has the following advantages:

[0039] 1. Codebook-based beamforming optimization to reduce feedback overhead and improve system feasibility: The present invention proposes a beamforming method under non-perfect CSI conditions, using a predefined codebook and a beam training scheme based on exhaustive search, so that the system can still achieve optimal beamforming effects through codebook optimization in the absence of independent CSI feedback. The receiving end only needs to select the best codeword in the predefined codebook for feedback, without the need to feedback the complete CSI, thus avoiding the complex CSI feedback process. Through codebook design and one-by-one selection, the present invention makes it unnecessary for the system to re-estimate the complete CSI when each channel changes, thereby effectively reducing the burden on the feedback link, significantly improving the system response speed and operating efficiency, and greatly improving the real-time and feasibility of the system.

[0040] 2. Beam training based on actual amplification model ensures efficient operation of the amplifier: The present invention proposes a beam training method based on actual amplification model, introduces the actual amplification model of active reconfigurable smart metasurface for the first time in the beamforming design of the system, fully explores the hardware characteristics and working mechanism of the reflection amplifier based on tunnel diode, and comprehensively considers the triggering mechanism, nonlinear factors and impedance mismatch problems of the amplifier, so that the signal power incident on the active reconfigurable smart metasurface is always kept within the interval of linear amplification of the trigger amplifier, thereby avoiding signal distortion and additional power consumption caused by abnormal operation of the amplifier. Through exhaustive search, the optimal transmit beamforming vector codeword is selected at the base station end to ensure that the signal can be linearly amplified by the amplifier without distortion and achieve the maximum cascade channel gain, thereby optimizing the beamforming effect and amplification efficiency of the system, reducing unnecessary power consumption, and effectively avoiding the potential signal distortion and performance degradation caused by ideal modeling in traditional methods.

[0041] 3. The codebook design is combined with actual channel conditions to simplify system design requirements: By designing the reflection beamforming and transmission power codebooks, the present invention effectively reduces the dependence on the base station-reconfigurable intelligent metasurface and reconfigurable intelligent metasurface-receiver link CSI. The system searches for candidate beam vectors and power values ​​one by one to obtain the optimal configuration, ensuring that reliable beamforming can still be achieved under complex channel conditions. In addition, the present invention uses SVD decomposition to generate the beamforming vector of the base station to the active reconfigurable intelligent metasurface channel, and extracts the right singular matrix vector corresponding to the maximum eigenvalue of the channel as the candidate direction of the base station beamforming, which effectively improves the channel gain utilization. This design ensures the efficient operation of the system, enables the optimal configuration of the channel gain to be achieved, and reduces the feedback complexity and system design requirements while ensuring performance.

[0042] 4. Since the active reconfigurable intelligent metasurface has outstanding advantages such as low profile, light weight and low cost, it can be easily installed on / removed from the wall or ceiling of a building. In addition, the active reconfigurable intelligent metasurface beam training method based on the actual amplification model proposed in the present invention can greatly reduce the number of components of the reconfigurable intelligent metasurface, solve the problem of insufficient passive reconfigurable intelligent surface beamforming gain and limited service range in the existing transmission system, and further improve the communication performance of the receiving user in the system by making full use of the accurate amplification model based on reality.

[0043] In summary, the present invention reduces feedback overhead and improves system stability through beamforming and power control strategies based on actual amplification models, while ensuring that the amplifier of the active reconfigurable smart metasurface works efficiently and stably during signal transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of an active reconfigurable intelligent metasurface assisted communication system in the present invention;

[0045] Figure 2 It is a schematic diagram of the response characteristics of the reflection amplifier based on the tunnel diode in the present invention;

[0046] Figure 3 It is a simulation diagram comparing the total communication rate with different numbers of transmitting antennas in the present invention;

[0047] Figure 4 This is a simulation diagram of the number of active reconfigurable smart surface reflection units versus the total communication rate in the present invention. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, not all embodiments. The following description of the exemplary embodiment is actually only illustrative and is by no means intended to limit the present invention and its application or use. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0049] The present invention provides an active reconfigurable intelligent metasurface beam training method based on an actual amplification model, including active reconfigurable intelligent metasurface amplifier hardware characteristics exploration, transmitter beamforming codebook design, active reconfigurable intelligent metasurface reflection beamforming codebook design and codebook training;

[0050] The exploration of the hardware characteristics of the active reconfigurable intelligent metasurface amplifier includes the exploration of the amplification principle of the reflection amplifier based on the tunnel diode, the exploration of the amplification characteristics, and the exploration of the effective linear amplification range of the reflection amplifier;

[0051] The base station transmit beamforming codebook design includes an initial transmit beamforming codebook design and a transmit power codebook design, thereby obtaining a final transmit beamforming vector codebook and an effective transmit signal, so that the signal power incident on the active reconfigurable smart metasurface is within a power range that triggers linear amplification of the reflective amplifier;

[0052] The active reconfigurable intelligent metasurface reflection beamforming codebook is generated by array steering vectors of cascade channels of an active reconfigurable intelligent metasurface auxiliary system arranged in a uniform planar array form;

[0053] The codebook training method is an exhaustive search, which obtains the best codeword, that is, the optimal transmission power and the optimal reflection beamforming vector, by comparing and judging the signal-to-noise ratio of the signal received at the receiver, and obtains the maximum total rate of the communication system;

[0054] The tunnel diode-based reflection amplifier is based on the concept of an oscillator and consists of an inductor, a capacitor and a tunnel diode. A tunnel diode under proper bias naturally produces a negative differential resistance. Due to such low bias requirements and negative differential resistance characteristics, it is used to manufacture a reflection amplifier with high gain. The amplification principle of the reflection amplifier is that the properly biased tunnel diode reflection amplifier acts as a locked oscillator. The oscillator locks the frequency of the input signal and can amplify and reflect the impinging external signal. The occurrence of locking depends on multiple factors, among which the amplitude of the external signal (i.e., the power intensity of the incident signal) is an important factor. In addition, the impedance of the reflection amplifier is also related to the power intensity of the incident signal. For different incident signal powers, the impedance value of the reflection amplifier shows a changing trend. In summary, as Figure 2 As shown, for incident signal powers of different intensities, the reflection amplifier will exhibit the characteristics of linear amplification, nonlinear amplification, and reflection without amplification.

[0055] It is worth noting that experiments have shown that a DC power consumption of 45 μW can usually support a power gain of 40 dB. Therefore, the active reconfigurable intelligent metasurface based on the tunnel diode reflection amplifier adopted in the present invention has very low energy consumption, low hardware complexity and strong feasibility.

[0056] An active reconfigurable intelligent metasurface beam training method based on a practical amplification model, wherein the base station is equipped with M transmitting antennas to perform beamforming on the transmitting signal;

[0057] The effective transmission signal x is expressed as:

[0058] x=w(i)s (9)

[0059] Where: s represents the transmission signal of the transmitter, satisfying represents the transmitting end beamforming vector, through the channel from the base station to the active reconfigurable smart metasurface Perform SVD decomposition to generate an initial transmit beamforming vector codeword, and then further generate a final transmit beamforming vector codebook through a transmit power codebook;

[0060] Initial transmit beamforming vector codebook w ini The generation method is:

[0061] G=UΣV H (10)

[0062]

[0063] Among them: the left singular vector matrix U generated by SVD decomposition describes the receiving direction space of the channel, the diagonal elements of the diagonal matrix Σ are arranged in descending order according to the non-negative singular values, and the right singular vector matrix V describes the transmitting direction space of the channel; V * (:,1) represents the conjugate of V(:,1); the right singular vector V(:,1) corresponding to the maximum singular value represents the optimal transmission direction at the transmitting end; V * (:,1) The transmitted signal obtains the best gain in the main path direction, so that the signal energy is maximally concentrated on the main path from the base station to the active reconfigurable smart metasurface, thereby improving the transmission signal-to-noise ratio and system performance.

[0064] The codeword w(i) of the final transmit beamforming vector codebook is generated in detail as follows:

[0065]

[0066] Among them, P BS Represents the base station's transmit power codebook, which is generated at equal intervals according to the step size within the set power range; P BS (i) represents the base station transmission power codebook P BS The ith value of .

[0067] The signal power incident on the nth reflection unit on the active reconfigurable smart metasurface is expressed as:

[0068] p in,n =|G(n,:)w(i) 2 (13)

[0069] Where: G(n,:) represents the nth row of channel G, corresponding to the nth reflection unit of the active reconfigurable smart metasurface.

[0070] The active reconfigurable intelligent metasurface is equipped with N reflection units, each of which is connected to a reflection amplifier based on a tunnel diode to perform phase shift adjustment and electromagnetic level power amplification on the incident signal; the N reflection units are arranged in a uniform planar array; in the active reconfigurable intelligent metasurface structure, the number of reflection units affects the beamforming gain: the more reflection units, the greater the beamforming gain; the fewer reflection units, the smaller the beamforming gain.

[0071] Defining the reflection phase shift vector of active reconfigurable smart metasurface and the reflected phase shift matrix where θ n Represents the phase shift of the nth reflection unit; defines the amplification factor vector and the magnification matrix

[0072] The active reconfigurable intelligent metasurface reflection beamforming codebook design is based on array steering vector generation of cascade channels. By modeling the array steering characteristics of the active reconfigurable intelligent metasurface cascade channels arranged in a uniform planar array, the codewords of the reflection beamforming codebook are generated to ensure the phase adjustment and reflection direction optimization of the reflected signal. Specifically, the active reconfigurable intelligent metasurface reflection beamforming codebook is designed as follows:

[0073]

[0074] Wherein, a(α,β) represents the array response vector of the active reconfigurable smart metasurface related to the uniform planar array arrangement; α, β represent the space angles related to the active reconfigurable intelligent metasurface: α = Δdsin(θ)cos(ν) / λ, β = Δdsin(ν) / λ, where θ and ν represent the azimuth and elevation of the physical space, respectively, λ represents the wavelength of the electromagnetic wave of the transmitted signal, Δd represents the array spacing, and in the present invention, Δd = λ / 2; n1 = [0,1,…,N1-1] T , n2=[0,1,...,N2-1] T , N1 and N2 represent the number of columns and rows of the active reconfigurable smart metasurface, respectively; the number of active reconfigurable smart metasurface reflection units N = N1 × N2.

[0075] In the present invention, for the incident signal that triggers the linear amplification mechanism of the tunnel diode reflection amplifier of the active reconfigurable smart metasurface, the tunnel diode reflection amplifier linearly amplifies it at a constant multiple. in,nWhen the incident signal power is less than the linear amplification requirement, the active reconfigurable smart metasurface only adjusts the phase and reflects the signal because the impedance of the tunnel diode does not match the microstrip line and the incident signal power is too weak to trigger the locking phenomenon. in,n When it is greater than the linear amplification requirement, the codebook training procedure stops to avoid abnormal amplification of the amplifier and cause additional uncontrollable losses.

[0076] The codebook training method is an exhaustive search, which determines the codeword that can maximize the signal-to-noise ratio of the received signal at the receiving end by searching the transmit beamforming vector codebook and the reflection beamforming codebook one by one, thereby obtaining the maximum total rate of the communication system. By using the transmit beamforming vector codebook and the reflection beamforming codebook of the base station and the active reconfigurable intelligent metasurface, the system only needs to directly use the transmit beamforming power codebook and the reflection beamforming vector codebook pre-calculated in the codebook, test and select the appropriate transmit power value and reflection beamforming vector one by one to maximize the quality of the user's received signal.

[0077] Example 1

[0078] The following takes the design of a single-user multi-input single-output wireless communication system assisted by an active reconfigurable intelligent metasurface as an example to further illustrate the technical solution of the present invention. Figure 1 As shown in the figure, the base station as the transmitting end is equipped with M = 4 antennas, and communicates with a single-antenna receiving end user with the help of an active reconfigurable smart metasurface with N = N1 × N2 = 8 × 8 reflection units. Each reflection unit is connected to a reflection amplifier based on a tunnel diode to perform phase shift adjustment and electromagnetic level power amplification on the incident signal. Specifically, in the Cartesian coordinate system, the base station is located at (5m, 0m, -30m), the active reconfigurable smart metasurface is located at (0m, 250m, 15m), and the receiving end user is located at (-5m, 250m, 0m).

[0079] The present invention assumes that the direct channel between the transmitter and the receiver is blocked by an obstacle. By deploying an active reconfigurable intelligent metasurface, a virtual line-of-sight (LoS) path is established, and cascade reflection channels of the transmitter-active reconfigurable intelligent metasurface and the active reconfigurable intelligent metasurface-receiving end are constructed. There are deterministic LoS paths in both reflection channels.

[0080] The present invention adopts a commonly used geometric channel model, and the channel G from the base station to the active reconfigurable intelligent metasurface and the channel h from the active reconfigurable intelligent metasurface to the user are RU Respectively expressed as:

[0081]

[0082] in, represents the large-scale fading of the correlated channel and is expressed as:

[0083] PL=PL0+10μlogd (17)

[0084] Wherein, PL0 is the reference path loss, which is usually set to 30 dB; μ is the path loss index, and the path loss index of channel G in the present invention is set to 2.8, and the path loss index of channel h is set to 2.8. RU The path loss index is set to 2; d is the distance between the two communication devices;

[0085] a(α, β) and b(α, β) represent the array steering vectors associated with the active reconfigurable smart metasurface arranged in a uniform planar array and the base station antennas arranged in a uniform linear array, respectively, which are specifically expressed as:

[0086]

[0087] b(α)=[1,e -j2πα ,...,e -j2π(M-1)α ] T (19)

[0088] Where: n1=[0,1,…,N1-1] T , n2=[0,1,…,N2-1] T ; α, β represent the spatial angles related to the active reconfigurable smart metasurface: α = Δd sin(θ)cos(ν) / λ, β = Δdsin(ν) / λ, where θ and ν represent the azimuth and elevation angles of the physical space, respectively, λ represents the wavelength of the electromagnetic wave of the transmitted signal, and Δd represents the array spacing. In this embodiment, Δd = λ / 2;

[0089] The base station's transmit beamforming vector codebook w is designed based on SVD decomposition, specifically:

[0090] 1) Perform SVD decomposition on the channel matrix G from the base station to the active reconfigurable intelligent metasurface, obtain the right singular vector corresponding to the maximum singular value of G, and obtain the optimal transmission direction of the transmitting antenna at the transmitting end, so that the transmitted signal can obtain the best gain in the main path direction, so that the signal energy is concentrated to the maximum extent on the main path from the base station to the active reconfigurable intelligent metasurface, thereby improving the transmission signal-to-noise ratio and system performance. The formula is expressed as:

[0091] G=UΣV H (20)

[0092]

[0093] Where: V * (:,1) represents the conjugate of V(:,1).

[0094] 2) To ensure the maximization of channel gain, when selecting the transmit beamforming vector, the adjustment of the incident signal power needs to be considered to meet the amplifier linear range requirements of the active reconfigurable smart metasurface. To this end, the transmit power codebook P is defined. BS , and test all possible transmission power values ​​one by one. By calculating the power incident on each reflection unit n of the active reconfigurable smart metasurface, it is ensured that the optimal power codeword is selected under the condition that the amplifier linear amplification working condition is met, so that the system energy efficiency is optimized.

[0095] The signal power incident to the nth reflection unit is expressed as:

[0096] p in,n =|G(n,:)w(i) 2 (twenty two)

[0097] Where: G(n,:) represents the nth row of channel G, corresponding to the nth reflection unit of the active reconfigurable smart metasurface.

[0098] The method for generating the codeword w(i) of the final transmit beamforming vector codebook is described in detail as follows:

[0099]

[0100] Ensure that ||w(i)|| 2 =P BS (i). Where P BS (i) represents the base station transmission power codebook P BS The i-th value of BS In the set power range, the power is generated at equal intervals according to the step length. BS Generate in the interval [0dBm, 30dBm] with a step size of 1.

[0101] The reflection beamforming codebook of the active reconfigurable smart metasurface is generated based on the array steering vector of the cascaded reflection channel to ensure that the signal is amplified in the optimal direction; the specific steps include:

[0102] 1) Generate a reflection beamforming codebook based on the array steering vector of the cascaded channel to ensure that the reflection direction of the active reconfigurable smart metasurface can be aligned with the user receiving end to achieve the optimal reflection gain; the reflection beamforming codebook is expressed as:

[0103]

[0104] 2) If the incident signal power meets the linear range condition of the amplifier, the active reconfigurable intelligent metasurface amplifies the signal. This embodiment assumes a 100-fold constant amplification.

[0105] Beam training and optimal codeword selection are performed by exhaustively searching the base station transmit beamforming vector codebook w and the active reconfigurable smart metasurface reflection beamforming codebook. Calculate each candidate pair of transmit beamforming vector w(i) and reflected beamforming vector The received signal-to-noise ratio under , is expressed as:

[0106]

[0107] in: represents the effective transmission signal of the base station, x = w(i)s, and satisfies represent the amplification factor matrix and reflection phase shift matrix of the active reconfigurable smart metasurface respectively; The first codebook representing the active reconfigurable smart metasurface reflective beamforming Code words, represents the thermal noise introduced by the active reconfigurable smart metasurface, n represents the additive Gaussian white noise at the user, with a mean of 0 and a variance of σ 2 .

[0108] By selecting the beamforming configuration that can maximize the received signal-to-noise ratio, the base station transmits the beamforming vector codeword w best Active reconfigurable smart metasurface reflective beamforming vector codeword

[0109] The beam training process also includes power control and jump-out mechanism, which specifically includes the following two aspects:

[0110] 1) The incident signal power is too small: If the incident signal power p in,n If the value is less than the preset amplifier linear amplification start threshold, the active reconfigurable smart metasurface only reflects the signal without gain amplification;

[0111] 2) The incident signal power is too large: If the incident signal power p in,n If the upper limit of the linear operating range of the amplifier is exceeded, the exhaustive search loop is exited and the current best configuration is selected as the final beamforming solution to prevent the amplifier from entering the nonlinear operating region.

[0112] Figure 3 This is a simulation diagram comparing the total communication rate with different numbers of transmitting antennas; Figure 4The figure is a simulation diagram of the number of active reconfigurable smart surface elements versus the total communication rate in the present invention. It can be seen that the active reconfigurable smart metasurface beam training method based on the actual amplification model proposed in the present invention can achieve a higher total communication rate and a more significant performance gain. In particular, as the number of active reconfigurable smart metasurface reflective units increases, the method proposed in the present invention can achieve a more lasting performance improvement.

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. An active reconfigurable intelligent metasurface beam training method based on a practical amplification model, characterized in that: include: Step 1: Design a beamforming codebook for a base station based on a uniform linear array arrangement, generate an initial transmit beamforming vector codebook in combination with the channel characteristics from the base station to the active reconfigurable smart metasurface; and then generate a final base station transmit beamforming vector codebook in combination with the transmit power codebook; Step 2, generating an active reconfigurable intelligent metasurface reflection beamforming codebook based on an array steering vector of a cascade channel assisted by an active reconfigurable intelligent metasurface arranged in a uniform planar array; Step 3: Perform codebook training to determine the response of the tunnel diode-based reflection amplifier by calculating the incident signal power of the active reconfigurable intelligent metasurface under each transmit beamforming vector codeword one by one. Then, search the reflection beamforming vector codewords pre-calculated in the reflection beamforming codebook one by one to determine the codeword that can maximize the signal-to-noise ratio of the received signal at the receiving end, that is, the optimal transmit beamforming vector codeword and the optimal reflection beamforming vector codeword, so as to obtain the maximum total rate of the communication system.

2. The active reconfigurable intelligent metasurface beam training method based on the actual amplification model according to claim 1 is characterized in that: In step 1, the initial transmit beamforming vector codebook is generated as follows: The channel G between the base station and the active reconfigurable smart metasurface is decomposed by SVD, and the obtained right singular vector is used as the candidate direction of the base station beamforming to optimize the signal gain in the main path direction of the channel; the formula is: G=UΣV H (1) in: represents the channel from the base station to the active reconfigurable smart metasurface; U is the left singular vector matrix, which describes the receiving direction space of the channel; Σ is a diagonal matrix, whose diagonal elements are arranged in descending order according to the non-negative singular values; V is the right singular vector matrix, which describes the transmitting direction space of the channel; The initial transmit beamforming vector codebook generation method is expressed as: The right singular vector V(:,1) corresponding to the maximum singular value represents the optimal transmission direction at the transmitting end; V * (:,1) represents the conjugate of V(:,1).

3. The active reconfigurable intelligent metasurface beam training method based on the actual amplification model according to claim 2 is characterized in that: In step 1, the codeword of the final base station transmit beamforming vector codebook is expressed as: Among them, P BS Represents the base station's transmit power codebook, which is generated at equal intervals according to the step size within the set power range; P BS (i) represents the base station transmission power codebook P BS The ith value of .

4. The active reconfigurable intelligent metasurface beam training method based on the actual amplification model according to any one of claims 1 to 3, characterized in that: In step 2, the active reconfigurable intelligent metasurface reflection beamforming codebook is expressed as: Wherein, a(α,β) represents the array response vector of the active reconfigurable smart metasurface related to the uniform planar array arrangement; α, β represent the spatial angles related to the active reconfigurable smart metasurface, n1 = [0, 1, ..., N1-1] T , n2=[0,1,…,N2-1] T , the number of active reconfigurable intelligent metasurface reflection units N = N1 × N2, N1, N2 are the number of columns and rows of the active reconfigurable intelligent metasurface respectively.

5. The active reconfigurable intelligent metasurface beam training method based on the actual amplification model according to claim 3 is characterized in that: In step 3, for the i-th transmit beamforming vector codeword w(i), the incident signal power of the n-th reflector unit on the active reconfigurable smart metasurface is expressed as: p in,n =|G(n,:)w(i) 2 (5) Where G(n,:) represents the nth row of channel G, corresponding to the nth reflection unit of the active reconfigurable smart metasurface, n = 1,...,N; For the linear amplification mechanism of the tunnel diode reflectoamplifier that triggers the active reconfigurable smart metasurface in,n , the tunnel diode reflection amplifier linearly amplifies it at a constant multiple; when p in,n When p is less than the linear amplification requirement, the active reconfigurable smart metasurface only adjusts the phase and reflects the signal; when p in,n When it is greater than the linear amplification requirement, the codebook training procedure stops.

6. The active reconfigurable intelligent metasurface beam training method based on the actual amplification model according to claim 4 is characterized in that: In step 3, the user receiving signal-to-noise ratio is expressed as: in, represents the channel from the active reconfigurable smart metasurface to the user; represents the transmission signal of the base station, x = w(i)s, s represents the transmission signal, and satisfies They represent the amplification factor matrix and reflection phase shift matrix of the active reconfigurable smart metasurface, The first codebook representing the active reconfigurable smart metasurface reflective beamforming Code words, represents the thermal noise introduced by the active reconfigurable smart metasurface, σ 2 represents the additive white Gaussian noise variance at the user.

7. The active reconfigurable intelligent metasurface beam training method based on the actual amplification model according to claim 6 is characterized in that: In step 3, the total rate of the communication system is expressed as: sum_rate=log(1+SNR) (7).

Images