A pilot design method for channel estimation of an intelligent reflecting surface assisted OFDM system

Abstract

The application provides a pilot design method for channel estimation of an intelligent reflecting surface assisted OFDM system, and belongs to the technical field of wireless communication, and the implementation steps are as follows: (1) constructing an intelligent reflecting surface assisted OFDM system model and a received pilot signal model; (2) constructing an optimization problem based on pilot positions and pilot powers; and (3) designing an alternating optimization algorithm to sequentially optimize the pilot positions and the pilot powers. The application provides a design method for jointly optimizing the pilot positions and the pilot powers based on minimizing the cross-correlation value of a measurement matrix, and the method can generate an optimal pilot pattern according to a system model to improve the channel estimation accuracy of the system and reduce the pilot overhead.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology and relates to a pilot design method that combines a reconfigurable intelligent surface (RIS) with orthogonal frequency division multiplexing (OFDM) technology to improve channel estimation accuracy. By jointly designing pilot positions and pilot power, the method achieves the minimization of cross-correlation values ​​based on compressed sensing, thereby accurately estimating channel state information. Background Technology

[0002] Compared to fifth-generation (5G) wireless communication network technology, sixth-generation (6G) wireless transmission solutions based on smart reflectors can improve data transmission rates and energy efficiency without incurring excessive costs. Smart reflectors can flexibly configure wireless channels to achieve ideal distribution by controlling the amplitude and phase of the incident signal. In terms of deployment, smart reflectors require no radio frequency (RF) link; the reflecting unit passively reflects the signal, thus generating no power consumption during signal processing. This results in low cost, flexible application scenarios, and easy widespread deployment.

[0003] Channel estimation in narrowband communication assisted by smart reflectors is relatively simple and has been extensively studied. OFDM technology, as a promising broadband communication technology, not only fully utilizes spectrum resources to improve communication performance, but its complexity also makes it more suitable for practical applications. However, channel estimation in OFDM systems based on smart reflectors is quite complex, with the following problems: 1. The cascaded channel convolution caused by multipath delay spread and time-domain impulse response on both sides of the smart reflector increases the difficulty of channel estimation; 2. Channel estimation based on OFDM systems can be obtained by inserting comb pilots, but the number of pilots consumed is extremely high. Therefore, in OFDM systems assisted by smart reflectors, how to design appropriate pilot positions and pilot powers to improve channel estimation accuracy has become a challenge. Summary of the Invention

[0004] This invention addresses the problem of obtaining accurate channel state information in intelligent reflector-assisted OFDM systems by providing a joint pilot position and pilot power allocation method based on an alternating optimization algorithm, namely the Whale Optimization Algorithm-Second Order Cone Programming (WOA-SOCP) algorithm. This method can obtain accurate channel information while reducing pilot overhead, providing a stable foundation for the deployment of intelligent reflectors.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0006] A pilot design method for channel estimation in an OFDM system with intelligent reflector-assisted design includes the following steps:

[0007] The first step is to construct a smart reflector-assisted OFDM system model and a pilot signal receiving model:

[0008] (1) The OFDM uplink single-input single-output system assisted by intelligent reflector consists of a base station, a user terminal and an intelligent reflector. Both the base station and the user terminal are equipped with a single antenna. The intelligent reflector consists of M reflector elements. The reflection amplitude of the M reflector elements is set to 1, and the reflection phase makes the reflected signal direction point to the base station.

[0009] (2) Set the pilot signal to be transmitted as The model for receiving pilot signals is as follows:

[0010]

[0011] in, This indicates that the N signal received by the base station consists of comb-shaped pilot signals. P A dimensional vector, where the user's bandwidth is equally divided into N parallel subcarriers, and the set of N subcarrier positions is represented as... Select N subcarriers from N subcarriers. P N is used as a pilot symbol. P The set of pilot positions corresponding to each pilot symbol is represented as follows: F is a diagonal matrix P It is a Fourier leaf matrix. Cascaded channel consisting of user-smart reflector-base station Direct connection channel between user and base station A sparse vector consisting of a small number of non-zero elements;

[0012] The second step involves minimizing the measurement matrix D = X based on the received pilot signal model derived in the first step, according to compressed sensing theory. P F P Based on the principle of cross-correlation value μ, an optimization problem is constructed based on pilot position and pilot power:

[0013]

[0014] This optimization problem is a nonconvex optimization problem, where L = max{L r ,L d Let} be the number of unique column vectors in matrix D, and v(i) = |x p(i) | 2 ,1≤i≤N PLet p(i) be the pilot power corresponding to the pilot position. The set of pilot powers is represented as follows: V T V H and V L These represent the total pilot power, the upper bound of the pilot power, and the lower bound of the pilot power, respectively.

[0015] The third step involves designing an alternating optimization algorithm to sequentially optimize the pilot position and pilot power. The specific steps are as follows:

[0016] (1) First, solve for the optimal set of pilot locations, and then distribute the pilot power evenly, i.e. At this point, the optimization problem based on pilot position and pilot power is transformed into an optimization problem based on pilot position:

[0017]

[0018] This optimization problem is nonconvex, and a discrete whale optimization algorithm is used. The fitness function is set as the cross-correlation value of the measurement matrix D. The optimal set of pilot positions is obtained through iterative solutions.

[0019] (2) Next, solve for the optimal pilot power set, and obtain the optimal pilot position set. Substituting these factors into an optimization problem based on pilot position and pilot power, the optimization problem is now transformed into an optimization problem based on pilot power:

[0020]

[0021] This optimization problem is a non-convex optimization problem involving complex values. It can be transformed into a second-order cone programming problem:

[0022]

[0023] in, ||·||2 is the L2 norm, v=[v(1),v(2),...,v(N) P )] T It is a vector consisting of a set of pilot power signals. i = 1, 2, ..., L-1, q i Let i be the i-th row of matrix F. Matrix F is specifically:

[0024]

[0025] This second-order cone programming problem was solved using MATLAB's CVX toolkit, and the optimal pilot power vector was calculated. Obtain the optimal pilot power set

[0026] The beneficial effect of this invention is that, in an OFDM system assisted by a smart reflector, the channel estimation accuracy is maximized by minimizing the cross-correlation value of the measurement matrix as the objective function and through joint optimization design of pilot position and pilot power. Attached Figure Description

[0027] Figure 1 A schematic diagram of the uplink of an OFDM system assisted by a smart reflector.

[0028] Figure 2 The flowchart shows the pilot design process based on the WOA-SOCP algorithm.

[0029] Figure 3 Flowchart of signal processing for an OFDM wireless communication system assisted by a smart reflector;

[0030] Figure 4 The normalized mean square error (NMSE) of channel impulse response estimation varies with signal-to-noise ratio (SNR) under pilot allocation schemes for the WOA-SOCP algorithm and the traditional (Uniform Scheme, US) algorithm.

[0031] Figure 5 The bit error rate (BER) varies with the signal-to-noise ratio (SNR) under the pilot allocation schemes of the WOA-SOCP and US algorithms. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0033] See Figure 1 A schematic diagram of the uplink of an OFDM system assisted by a smart reflector. The system consists of a base station, a user terminal, and a smart reflector. The number of base stations and users is 1. Both the base station and user terminal are equipped with a single antenna. The smart reflector has M passive reflector elements, and the reflection coefficient matrix is ​​denoted as... The amplitude and phase ranges of each reflecting unit are β. m ∈[0,1],θ m ∈(0,2π],m=1,2,...,M. In addition, the amplitude and phase of the intelligent reflector can be sent from the base station to the intelligent reflector controller (RIS controller link). At the same time, in order to maximize the reflection power and simplify the hardware installation, the reflection amplitude of all reflection units is set to 1, and the reflection phase makes the direction of the reflected signal point to the base station.

[0034] See Figure 2 The pilot design flowchart based on the WOA-SOCP algorithm is shown below, with specific steps as follows:

[0035] The first step is to construct the model of the intelligent reflector-assisted OFDM system and the model of the received pilot signal: For the channel model of the intelligent reflector-assisted OFDM system, this invention considers a quasi-static frequency-selective channel, in which the channel state remains approximately unchanged within a specific coherence time. The channel model includes a direct-connection channel that does not pass through the intelligent reflector. and cascaded channels via intelligent reflectors The channel impulse responses for user-base station, user-mth reflecting unit, and mth reflecting unit-base station are respectively... and The channel sparsity is k d The maximum channel delays for k1 and k2 are L, respectively. d L1 and L2, satisfying L d >>k d L1 >> k1 and L2 >> k2. Cascaded channel h of user-smart reflector-base station. r It can be represented as:

[0036]

[0037] Where * represents the convolution symbol, φ m Let L be the reflection coefficient of the m-th reflecting unit. r =L1+L2-1 is the maximum delay of the cascaded channel, h c,m This represents the cascaded channel of user-mth reflecting unit-base station, where k is set. r The sparsity of the cascaded channel.

[0038] A smart reflector-assisted OFDM system divides the bandwidth into N parallel subcarriers, and the set of N subcarrier positions is represented as... Select N subcarriers from N subcarriers. P N is used as a pilot symbol. P The set of pilot positions corresponding to each pilot symbol is represented as follows: The pilot signal to be transmitted is set as Received pilot signal for:

[0039]

[0040] in, F is a diagonal matrix composed of pilot signals. P1 and F P2 They are respectively composed of N P line L r Columns and N P line L d Fourier leaf matrix composed of columns. Sparse vectors. The sparsity is k r+k d , It is Gaussian white noise.

[0041] The second step, based on the received pilot signal model obtained in the first step, is to construct an optimization problem based on the pilot position and pilot power: From the received pilot signal model, the measurement matrix is ​​D = X. P F P The cross-correlation value μ(D) of the measurement matrix D is the maximum value of the inner product of any two normalized columns of D:

[0042]

[0043] Where ||·||2 represents the L2 norm, and <·> represents the inner product, D m and D n Let L be any two columns of matrix D, and L = max{L r ,L d} is the number of unique column vectors in D, v(i) = |x p(i) | 2 ,1≤i≤N P Let p(i) be the pilot power corresponding to the pilot position. The set of pilot powers is represented as follows:

[0044] Since vector h is a sparse vector, based on the principle of minimizing the cross-correlation value of the measurement matrix in compressed sensing theory, an optimization problem based on pilot position and pilot power is constructed:

[0045]

[0046] This optimization problem is a non-convex optimization problem, where V T V H and V L These represent the total pilot power, the upper bound of the pilot power, and the lower bound of the pilot power, respectively.

[0047] The third step involves designing an alternating optimization algorithm, namely the WOA-SOCP algorithm, to sequentially optimize pilot position and pilot power. The optimization problem based on pilot position and pilot power is a problem containing a set of discrete integer variables. and continuous variable set For nonconvex optimization problems, an alternating optimization algorithm is used to optimize sequentially. and

[0048] (1) First, solve for the set of pilot locations that are easier to enumerate, assuming that all pilot powers are equal, i.e. At this point, the optimization problem based on pilot position and pilot power is transformed into an optimization problem based on pilot position:

[0049]

[0050] This optimization problem is nonconvex, and a discrete whale optimization algorithm is used. The fitness function is set as the cross-correlation value of the measurement matrix D. The optimal set of pilot positions is obtained through iterative solutions.

[0051] (2) Next, solve for the optimal pilot power set, and obtain the optimal pilot position set. Substituting these factors into an optimization problem based on pilot position and pilot power, the optimization problem is now transformed into an optimization problem based on pilot power:

[0052]

[0053] To simplify this optimization problem, the following settings are made:

[0054]

[0055] At this point, the cross-correlation value of the measurement matrix can be expressed as:

[0056]

[0057] Among them, ||·|| ∞ For the infinite norm, v = [v(1), v(2), ..., v(N)]. P )] T Let be a vector consisting of a set of pilot powers. After simplification, the optimization problem can be expressed as:

[0058]

[0059] Divide matrix F into real and imaginary parts:

[0060]

[0061] Where Re(·) represents the real part, Im(·) represents the imaginary part, and q i Let F be the i-th row of matrix F. The optimization problem can then be expressed as:

[0062]

[0063] Where z = μ(v), ​​the optimization problem here is a second-order cone programming problem with second-order cone constraints and linear constraints. The optimal pilot power vector can be calculated using MATLAB's CVX toolkit to solve the above second-order cone programming problem. Obtain the optimal pilot power set

[0064] See Figure 3The signal processing flowchart of the intelligent reflector-assisted OFDM wireless communication system is as follows: The user's input signal undergoes Quadrature Amplitude Modulation (QAM) modulation, converting the serial signal to a parallel signal. Then, a pilot signal is inserted into the signal using a pilot allocation scheme based on an alternating optimization algorithm proposed in this invention. The signal undergoes an Inverse Fast Fourier Transform (IFFT), inserting a cyclic prefix and converting the parallel signal back to a serial signal. The received signal is then transmitted through cascaded and direct channels to the base station. To estimate channel state information and further estimate the transmitted signal at the base station, the serial signal is converted back to a parallel signal, the cyclic prefix is ​​removed, and an IFFT is performed to extract the received pilot signal. The channel impulse response is estimated using a traditional orthogonal matched pursuit algorithm. The parallel signal is then converted back to a serial signal, and QAM demodulation is performed. The transmitted signal is then reconstructed based on the estimated channel and the demodulated received signal.

[0065] To demonstrate the effectiveness of pilot allocation, simulation results were compared between this scheme and a scheme that distributes pilot positions and pilot power on an average basis.

[0066] See Figure 4 The normalized mean square error (NMSE) of channel impulse response estimation varies with signal-to-noise ratio (SNR) under pilot allocation schemes using the WOA-SOCP algorithm and the traditional (Uniform Scheme, US) algorithm. With SNR set from 0 dB to 20 dB, the figure shows that the NMSE of channel estimation gradually decreases as the SNR increases. To illustrate the impact of the number of pilots, we set N... P =24, 32. The results show that as the number of pilots increases, the mean square error decreases, and the mean square error of WOA-SOCP is significantly smaller than that of the US scheme. Furthermore, the WOA-SOCP algorithm performs better in N... P =24 hours and US algorithm in N P The mean square error curves at a value of 32 almost overlap. Therefore, the WOA-SOCP pilot allocation scheme proposed in this invention improves pilot utilization efficiency.

[0067] See Figure 5The bit error rate (BER) varies with signal-to-noise ratio (SNR) under pilot allocation schemes using the WOA-SOCP and US algorithms. The results show that the BER gradually decreases as the SNR increases, with the WOA-SOCP algorithm showing a more significant BER reduction, indicating that this method can more accurately estimate the original transmitted signal. At high SNR, the difference in BER between the WOA-SOCP and US algorithms is greater than at low SNR. Furthermore, the BER with 32 pilots is lower than the BER with 24 pilots, suggesting that a higher number of pilots leads to more accurate recovery of the original transmitted signal.

[0068] Combination Figure 4 and Figure 5 It can be observed that in the scenario of OFDM system assisted by intelligent reflector, the pilot allocation scheme proposed in this invention can achieve a high accuracy of channel estimation, thereby more accurately restoring the transmitted signal and improving pilot efficiency.

Claims

1. A pilot design method for channel estimation in an OFDM system assisted by a smart reflector, characterized in that, Includes the following steps, The first step is to construct a smart reflector-assisted OFDM system model and a pilot signal receiving model: (1) The OFDM uplink single-input single-output system assisted by intelligent reflector consists of a base station, a user terminal and an intelligent reflector. Both the base station and the user terminal are equipped with a single antenna. The intelligent reflector consists of M reflector elements. The reflection amplitude of the M reflector elements is set to 1, and the reflection phase makes the reflected signal direction point to the base station. (2) Set the pilot signal to be transmitted as The model for receiving pilot signals is as follows: in, This indicates that the N signal received by the base station consists of comb-shaped pilot signals. P A dimensional vector, where the user's bandwidth is equally divided into N parallel subcarriers, and the set of N subcarrier positions is represented as... Select N subcarriers from N subcarriers. P N is used as a pilot symbol. P The set of pilot positions corresponding to each pilot symbol is represented as follows: F is a diagonal matrix P It is a Fourier leaf matrix. Cascaded channel consisting of user-smart reflector-base station Direct connection channel between user and base station A sparse vector consisting of a small number of non-zero elements; The second step involves minimizing the measurement matrix D = X based on the received pilot signal model derived in the first step, according to compressed sensing theory. P F P Based on the principle of cross-correlation value μ, an optimization problem is constructed based on pilot position and pilot power: This optimization problem is a nonconvex optimization problem, where L = max{L r ,L d Let} be the number of unique column vectors in matrix D, and v(i) = |x p(i) | 2 ,1≤i≤N P Let p(i) be the pilot power corresponding to the pilot position. The set of pilot powers is represented as follows: V T V H and V L These represent the total pilot power, the upper bound of the pilot power, and the lower bound of the pilot power, respectively. The third step involves designing an alternating optimization algorithm to sequentially optimize the pilot position and pilot power. The specific steps are as follows: (1) First, solve for the optimal set of pilot locations, and then distribute the pilot power evenly, i.e. At this point, the optimization problem based on pilot position and pilot power is transformed into an optimization problem based on pilot position: This optimization problem is nonconvex, and a discrete whale optimization algorithm is used. The fitness function is set as the cross-correlation value of the measurement matrix D. The optimal set of pilot positions is obtained through iterative solutions. (2) Next, solve for the optimal pilot power set, and obtain the optimal pilot position set. Substituting these into the optimization problem based on pilot position and pilot power, the optimization problem is now transformed into an optimization problem based on pilot power: This optimization problem is a non-convex optimization problem involving complex values. It can be transformed into a second-order cone programming problem: min z in, ||·||2 is the L2 norm, v=[v(1),v(2),...,v(N) P )] T It is a vector consisting of a set of pilot power signals. q i Let i be the i-th row of matrix F. Matrix F is specifically: This second-order cone programming problem was solved using MATLAB's CVX toolkit, and the optimal pilot power vector was calculated. Obtain the optimal pilot power set

Images