A channel estimation method based on active amplification type reconfigurable intelligent surface

Abstract

The present application belongs to the technical field of wireless communication, and particularly relates to a channel estimation method based on an active amplification type reconfigurable intelligent surface. The method of the present application can separate the required estimated CSI at the base station by continuously switching the reflection pattern of the intelligent surface, and perform accurate estimation. Meanwhile, by using the signal power amplification function of the active amplification type intelligent surface, the reflection signal power can be improved, thereby improving the channel estimation performance. Compared with the channel estimation method suitable for the traditional passive type intelligent surface, the method of the present application can further reduce the channel estimation error, thereby significantly improving the auxiliary effect of the active amplification type reconfigurable intelligent surface on the wireless communication system. Through simulation and experimental verification, the channel estimation method of the present application can realize accurate estimation of the wireless channel environment, and at the same time promote the application of the active amplification type reconfigurable intelligent surface in the field of wireless communication, and has important application value and development potential.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to a channel estimation method based on an active amplified reconfigurable smart surface. Background Technology

[0002] Reconfigurable Intelligent Surfaces (RIS), as a technology with high spectral efficiency, high energy efficiency, and low cost, have attracted widespread attention from academia and industry in recent years. Traditional passive intelligent surfaces consist of a large number of passive, low-cost reflective elements. By artificially controlling the reflection coefficient of each reflective element, a phase shift can be applied to the incident signal, resulting in reflection and beamforming. Compared to traditional repeater technologies, intelligent surface technology naturally achieves "full-duplex" repeater operation by utilizing the principle of electromagnetic signal reflection. By rationally designing the reflection coefficient of each reflective element, the wireless propagation environment can be intelligently reconfigured to achieve purposes such as useful signal enhancement, interference suppression, and secure transmission.

[0003] Traditional passive smart surfaces cannot amplify the power of incident signals, thus limiting their effectiveness in assisting wireless communication systems. On the other hand, active amplified reconfigurable smart surfaces (ARIS), a novel smart surface technology, integrate power amplification devices in each reflective unit. Therefore, ARIS can effectively amplify the power of the incident signal while altering its phase state. Compared to passive smart surfaces, ARIS offers higher energy efficiency and can provide a stronger reflective communication link with the same power consumption, thus better supporting existing wireless communication systems. Summary of the Invention

[0004] The main content of this invention is to propose a channel estimation method based on an active amplified reconfigurable smart surface. The basic working principle of channel estimation based on an active amplified reconfigurable smart surface is as follows: Considering that the system adopts Time Division Duplexing (TDD), since the uplink and downlink channels are distinct, the uplink channel state information (CSI) estimated at the base station can be used for downlink data transmission. Simultaneously, since the positions of the base station and the smart surface are fixed, and their deployment ensures a direct link between the base station and the smart surface, the CSI of this link can be determined by prior known position information. Based on this, the specific channel estimation process is as follows: During the training phase, the user continuously sends multiple pilot signals to the base station; simultaneously, the smart surface switches its reflection pattern once for each transmitted pilot signal (this reflection pattern is referred to as the training reflection pattern below). Therefore, the base station receives the pilot signals from the user and those passing through the smart surface, and uses the known pilot information and training reflection pattern information to estimate the channels from the user to the base station and from the user to the smart surface respectively.

[0005] The technical solution of this invention is as follows:

[0006] A channel estimation method based on an active amplified reconfigurable smart surface, the system comprising a single-antenna user, a base station with K antennas, and an active amplified smart surface with M reflection elements; characterized in that the channel estimation method comprises:

[0007] During the training phase, the user continuously sends multiple pilot signals to the base station. Simultaneously, the smart surface switches its training reflection pattern for each transmitted pilot signal. The base station receives the pilot signals from the user and those passing through the smart surface, and uses the known pilot information and training reflection pattern information to estimate the channels from the user to the base station and from the user to the smart surface. Specifically, the signal received at the k-th antenna of the base station in the nth training time slot is defined as follows:

[0008]

[0009] in, It is the training reflection pattern of the smart surface in the nth time slot. h is a diagonal matrix formed by the channels from the smart surface to the k-th antenna of the base station. d,k It is the channel from the user to the k-th antenna of the base station. It is the channel from the user to the smart surface, P T This represents the transmission power, and s(n) is the pilot signal transmitted in the nth time slot. It is thermal noise at the smart surface. Let I be the variance of the thermal noise.M It is an M-dimensional identity matrix. It is the received noise at the k-th antenna of the base station. To receive the noise variance, pilot signals received in N time slots during the training phase are collected. The received signal at the k-th antenna of the base station is:

[0010]

[0011] Among them, y k =[y k (0),…,y k (N-1)] T It is the pilot signal collected at the k-th antenna, h k =[h d,k ,b T ] T S is the channel vector that needs to be estimated at the k-th antenna, and S = diag([s(0),…,s(N-1)]) is a diagonal matrix composed of the transmitted pilot signals. It is equivalent noise. z k =[z k (0),…,z k (N-1)] T ,

[0012]

[0013] When the number of transmitted pilots N≥M+1, h is obtained. k The estimated value is:

[0014]

[0015] Among them, I N This is an N-dimensional identity matrix. By estimating the pilot signals received at each antenna of the base station, the estimated value of the channel from the user to the base station can be obtained. for The first element.

[0016] The estimated user-to-smart surface channel at each antenna of the base station ( The average of the vector consisting of the second to the last element yields the estimate of the channel from the user to the smart surface:

[0017]

[0018] Furthermore, the training reflection pattern of the smart surface is defined to satisfy the following equation:

[0019]

[0020] in, G1 is a matrix composed of the elements from the 2nd to the (M+1)th columns of the N-point Fourier transform matrix, and G1 = diag(g1) is a diagonal matrix formed by the channels from the smart surface to the first antenna of the base station. The maximum amplitude a that can be achieved by the reflection coefficient of channel G1 and the smart surface is determined by the channel G1. max The determined scaling factor β shall satisfy the following condition:

[0021]

[0022] Where, ρ g For large-scale path fading from smart surfaces to base stations, a simplified h is obtained. k The estimate is:

[0023]

[0024] The matrix inversion can be directly obtained as follows:

[0025] The beneficial effects of this invention are as follows: This invention proposes a channel estimation method based on an actively amplified reconfigurable smart surface, including a channel estimation method and the design of a training reflection pattern for the smart surface. By continuously switching the reflection pattern of the smart surface, the required estimated CSI can be separated at the base station and accurately estimated. Simultaneously, utilizing the signal power amplification function of the actively amplified smart surface, the reflected signal power can be increased, thereby improving channel estimation performance. Furthermore, based on a minimum variance unbiased estimator and its covariance matrix, this invention designs the training reflection pattern and amplification coefficient of the smart surface, achieving optimal overall channel estimation performance. Compared with channel estimation methods applicable to traditional passive smart surfaces, the method proposed in this invention can further reduce channel estimation errors, thereby significantly improving the auxiliary effect of the actively amplified reconfigurable smart surface on wireless communication systems. Through simulation and experimental verification, the channel estimation method proposed in this invention can achieve accurate estimation of the wireless channel environment, while promoting the application of actively amplified reconfigurable smart surfaces in the field of wireless communication, demonstrating significant application value and development potential. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the system composition of the present invention;

[0027] Figure 2 The curve showing the relationship between the mean square error of the channel estimation and the scaling factor;

[0028] Figure 3 The curves showing the relationship between the mean square error of channel estimation and transmit power under different quantization bits of reflection phase shift;

[0029] Figure 4 The curves showing the relationship between the system's achievable rate and transmit power under different channel estimation schemes;

[0030] Figure 5 This is the curve showing the relationship between the mean square error of channel estimation and the maximum amplitude of the smart surface reflectivity. Detailed Implementation

[0031] The invention will now be described in detail with reference to the accompanying drawings and simulation examples.

[0032] This invention proposes a channel estimation method based on an active amplified reconfigurable smart surface. The system composition structure is as follows: Figure 1 As shown, it consists of a single-antenna user equipment (UE), a base station (BS) with K antennas, and an active amplified smart surface with M reflective elements.

[0033] The channel estimation principle based on an actively amplified reconfigurable smart surface (hereinafter referred to as the smart surface) is as follows: Considering that the system adopts Time Division Duplexing (TDD), since the uplink and downlink channels are distinct, the uplink channel state information (CSI) estimated at the base station can be used for downlink data transmission. Simultaneously, since the positions of the base station and the smart surface are fixed, and their deployment ensures a direct link between them, the CSI of this link can be determined by prior known position information. Based on this, the specific channel estimation process is as follows: During the training phase, the user continuously sends multiple pilot signals to the base station; simultaneously, the smart surface switches its reflection pattern once for each transmitted pilot signal (this reflection pattern is referred to as the training reflection pattern). Therefore, the base station receives the pilot signals from the user and those passing through the smart surface, and uses the known pilot information and training reflection pattern information to estimate the channels from the user to the base station and from the user to the smart surface respectively.

[0034] Based on the above description, the signal received at the k-th antenna of the base station during the nth training time slot is:

[0035]

[0036] in, It is the training reflection pattern of the smart surface in the nth time slot. h is a diagonal matrix formed by the channels from the smart surface to the k-th antenna of the base station. d,k It is the channel from the user to the k-th antenna of the base station. It is the channel from the user to the smart surface, P TThis represents the transmission power, and s(n) is the pilot signal transmitted in the nth time slot. It is thermal noise at the smart surface. This is the received noise at the k-th antenna of the base station. The pilot signals received in N time slots during the training phase are collected, and the received signal at the k-th antenna of the base station is...

[0037]

[0038] Among them, y k =[y k (0),…,y k (N-1)] T It is the pilot signal collected at the k-th antenna, h k =[h d,k ,b T ] T S is the channel vector that needs to be estimated at the k-th antenna, and S = diag([s(0),…,s(N-1)]) is a diagonal matrix composed of the transmitted pilot signals. It is equivalent noise. z k =[z k (0),…,z k (N-1)] T ,

[0039]

[0040] When the number of transmitted pilots N ≥ M + 1, estimate h k The minimum variance unbiased (MVU) estimator is the same as the least squares (LS) estimator. Therefore, h k It can be estimated by the following formula.

[0041]

[0042] The covariance matrix corresponding to the estimation result is

[0043]

[0044] After estimating the pilot signals received at each antenna of the base station as shown in formula (3), the estimated value of the channel from the user to the base station can be obtained. Meanwhile, the estimated value of the user-to-smart surface channel By averaging, we can obtain the estimate of the channel from the user to the smart surface.

[0045]

[0046] Based on the above estimation process, we can obtain and The sum of the mean square deviations is

[0047]

[0048] Since the sum of the mean square errors of the channel estimations mentioned above is affected by the training reflection patterns of the smart surface, it can be optimized to further reduce the channel estimation error. Therefore, in order to minimize the sum of the mean square errors of the channel estimations in formula (6), the N training reflection patterns {φ} of the smart surface are optimized. n} N Optimizing the design leads to the following optimization problem:

[0049]

[0050] Among them, a max This represents the maximum amplitude achievable by the reflection coefficient of the smart surface. The aforementioned optimization problem is a non-convex optimization problem, which is difficult to solve. Borrowing relevant properties of the Fisher matrix and the Fourier transform matrix from estimation theory, the training reflection pattern of the smart surface designed in this invention satisfies the following equation:

[0051]

[0052] in, G1 is a matrix composed of the elements from the 2nd to the (M+1)th columns of the N-point Fourier transform matrix, and G1 = diag(g1) is a diagonal matrix formed by the channels from the smart surface to the first antenna of the base station. It is composed of channel G1 and a max The scaling factor is determined. Furthermore, the sum of the mean square errors of the above channel estimates reaches its lower bound when the scaling factor β satisfies the following condition:

[0053]

[0054] Where, ρ g This represents large-scale path fading from the smart surface to the base station. Under this reflection pattern design and corresponding channel estimation method, h k The estimate can be simplified as follows:

[0055]

[0056] The matrix inversion can be directly obtained as follows: Simultaneously, the mean square errors of the user-to-base station and user-to-smart surface channel estimates can be obtained as follows:

[0057]

[0058]

[0059] The beneficial effects of this invention will be verified through simulation. The simulation parameters are set as follows: the antenna at the base station and the reflector elements at the smart surface are both deployed in a uniform planar array, M = K = 16. The channel modeling from the smart surface to the base station is as follows: Among them, the large-scale fading factor is modeled as ρ g =10 -3 d -2 The distance from the smart surface to the base station is d = 50m. The guiding vector is determined by the direction angle, and the incident direction is... and launch direction Randomly generated. The noise power at the smart surface and the base station are respectively... and The DFT channel estimation scheme designed for traditional passive smart surfaces (Zheng B, Zhang R. Intelligent reflecting surface-enhanced OFDM: Channel estimation and reflection optimization[J]. IEEE Wireless Communications Letters, 2019, 9(4): 518-522.) and the switch-based channel estimation scheme (Yang Y, Zheng B, Zhang S, Zhang R. Intelligent reflecting surface meets OFDM: Protocol design and rate maximization[J]. IEEE Transactions on Communications, 2020, 68(7): 4522-4535.) are used as reference schemes and their performance is compared with the scheme proposed in this patent.

[0060] like Figure 2 As shown, the theoretical and simulation performance of the method of this invention are in good agreement. Furthermore, within a given range of scaling factors, the estimation performance achieved by the method of this invention is superior to that of traditional DFT channel estimation schemes. In particular, the optimal scaling factor selected in the method of this invention achieves the best overall channel estimation performance.

[0061] like Figure 3 As shown, under different quantization bit numbers for the reflection phase shift, the training reflection pattern and corresponding channel estimation method proposed in this invention outperform traditional DFT and switching channel estimation schemes. However, due to the disruption of the orthogonality between the original training reflection patterns caused by the discrete quantization of the reflection phase shift, the mean square error of the channel estimation in this invention will not continue to decrease with the increase of the signal-to-noise ratio under high signal-to-noise ratio conditions.

[0062] like Figure 4 As shown, the reflection coefficient matrix of the smart surface can be solved using channel state information obtained from different channel estimation schemes, thereby obtaining the achievable rate of the corresponding system. A comparison reveals that the active amplified smart surface achieves better achievable rate performance than the traditional passive smart surface. Furthermore, because the channel estimation method proposed in this invention provides more accurate estimation results, it can achieve even better achievable rate performance for wireless communication systems assisted by the active amplified smart surface.

[0063] like Figure 5 As shown, when designing and selecting parameters for active amplified smart surfaces, their impact on system performance should also be considered. When a certain threshold is exceeded, increasing its value can no longer improve the channel estimation accuracy.

Claims

1. A channel estimation method based on an active amplified reconfigurable smart surface, the system comprising a single-antenna user, a base station with K antennas, and an active amplified smart surface with M reflective elements; characterized in that, The channel estimation method includes: During the training phase, the user continuously sends multiple pilot signals to the base station. Simultaneously, the smart surface switches its training reflection pattern for each transmitted pilot signal. The base station receives the pilot signals from the user and those passing through the smart surface, and uses the known pilot information and training reflection pattern information to estimate the channels from the user to the base station and from the user to the smart surface. Specifically, the signal received at the k-th antenna of the base station in the nth training time slot is defined as follows: in, It is the training reflection pattern of the smart surface in the nth time slot. h is a diagonal matrix formed by the channels from the smart surface to the k-th antenna of the base station. d,k It is the channel from the user to the k-th antenna of the base station. It is the channel from the user to the smart surface, P T This represents the transmission power, and s(n) is the pilot signal transmitted in the nth time slot. It is thermal noise at the smart surface. Let I be the variance of the thermal noise. M It is an M-dimensional identity matrix. It is the received noise at the k-th antenna of the base station. To receive the noise variance, pilot signals received in N time slots during the training phase are collected. The received signal at the k-th antenna of the base station is: Among them, y k =[y k (0),…,y k (N-1)] T It is the pilot signal collected at the k-th antenna, h k =[h d,k ,b T ] T S is the channel vector that needs to be estimated at the k-th antenna, and S = diag([s(0),…,s(N-1)]) is a diagonal matrix composed of the transmitted pilot signals. It is equivalent noise. z k =[z k (0),…,z k (N-1)] T , When the number of transmitted pilots N≥M+1, h is obtained. k The estimated value is: Among them, I N Given an N-dimensional identity matrix; by estimating the pilot signals received at each antenna of the base station, the estimated value of the user-to-base station channel can be obtained. for The first element; Depend on The vector consisting of the second to the last element represents the estimated user-to-smart surface channel at each antenna of the base station. right By averaging, we can obtain the estimate of the channel from the user to the smart surface as follows:

2. The channel estimation method based on an active amplified reconfigurable smart surface according to claim 1, characterized in that, The training reflection pattern of a smart surface is defined to satisfy the following equation: in, G1 is a matrix composed of the elements from the 2nd to the (M+1)th columns of the N-point Fourier transform matrix, and G1 = diag(g1) is a diagonal matrix formed by the channels from the smart surface to the first antenna of the base station. The maximum amplitude a that can be achieved by the reflection coefficient of channel G1 and the smart surface is determined by the channel G1. max The determined scaling factor β shall satisfy the following condition: Where, ρ g For large-scale path fading from smart surfaces to base stations, a simplified h is obtained. k The estimate is: The matrix inversion can be directly obtained as follows:

Images