A pilot optimization method based on smart reflector OFDM sparse channel

Abstract

The application discloses a pilot optimization method for an OFDM sparse channel based on an intelligent reflecting surface, and mainly solves the problem that channel estimation accuracy is limited by wideband channel complexity, and the implementation steps are as follows: (1) constructing an OFDM system model based on the intelligent reflecting surface, and modeling the sparse channel; (2) from the perspective of minimizing the cross-correlation value of a measurement matrix, constructing an optimization problem; (3) designing a coding and decoding auxiliary whale optimization algorithm to solve the optimization problem. The application designs a coding variable to represent discrete variables and continuous variables at the same time, designs a decoding process to decode the coding variable into pilot positions and pilot powers, and jointly solves the optimal pilot positions and pilot powers through the whale optimization algorithm. The pilot optimization method for the OFDM sparse channel based on the intelligent reflecting surface provided by the application uses a lower calculation complexity, and improves the accuracy of channel estimation.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to a pilot optimization method for OFDM sparse channels based on intelligent reflectors. Background Technology

[0002] Reflecting Intelligent Surface (RIS) is a technology that dynamically alters wireless channels to improve communication performance. It utilizes numerous low-cost passive reflective elements to intelligently adjust the amplitude and phase of signal reflection. Compared to fifth-generation (5G) wireless communication networks, 6G wireless communication networks based on RIS offer advantages such as flexible deployment, low cost, and wide coverage. RIS-based wireless networks are more complex, and the ability to obtain accurate channel state information and correctly demodulate the transmitted signal at the receiver is a crucial indicator of RIS communication performance. Currently, channel estimation techniques for RIS-assisted wireless communication are primarily based on narrowband channels, with limited research on broadband channels, which are more suitable for practical applications.

[0003] Orthogonal Frequency Division Multiplexing (OFDM) can fully utilize spectrum resources, reduce frequency-selective fading, and improve communication performance. However, the large multipath delay spread on both sides of the smart reflector leads to sparse and complex channels, resulting in large pilot overhead and numerous estimation parameters for channel estimation. Efficiently and accurately estimating channel state information has become a challenge.

[0004] Compressed Sensing (CS) theory, as a novel signal processing approach, samples compressible high-dimensional sparse signals at a rate far lower than the Nyquist sampling rate and accurately reconstructs the sparse signal. In channel estimation for OFDM systems based on smart reflectors, comb pilots are inserted at the transmitter, and the inherent sparsity of the channel is used to extract the pilot signal at the receiver. Channel state information is then recovered using a CS-based signal reconstruction algorithm.

[0005] Currently, in broadband channel estimation based on smart reflectors, pilot sequences are typically distributed at equal intervals, and pilot powers are equally divided. However, the channel estimation performance of OFDM systems based on smart reflectors using pilot sequences with equal intervals and equal power distribution is not optimal. Therefore, it is necessary to design a channel pilot sequence optimization method to simultaneously optimize pilot positions and pilot powers, further improving the accuracy of channel estimation for OFDM systems based on smart reflectors without sacrificing computational complexity. Summary of the Invention

[0006] The purpose of this invention is to provide a pilot optimization method for OFDM sparse channels based on intelligent reflectors. The optimization problem is constructed by minimizing the cross-correlation value of the measurement matrix, and the whale optimization algorithm is assisted by encoding and decoding methods to simultaneously optimize the pilot position and pilot power, thereby improving the channel estimation accuracy of OFDM communication systems based on intelligent reflectors.

[0007] The objective of this invention is achieved through the following technical solution: a pilot optimization method for OFDM sparse channels based on intelligent reflectors, the method comprising the following steps:

[0008] The first step is to construct an OFDM system model and channel model based on a smart reflector:

[0009] (1) The OFDM uplink single-input single-output system based on a smart reflector consists of a base station, a user terminal, and a smart reflector. Both the base station and the user terminal are equipped with a single antenna. The smart reflector is composed of... It consists of a reflective unit, and is set to... The reflection amplitude of each reflection unit is 1, and the reflection phase makes the reflected signal point towards the base station.

[0010] (2) Model the OFDM sparse channel based on the smart reflector and construct the impulse response of the direct sparse channel without passing through the smart reflector. Impulse response of cascaded sparse channel through intelligent reflector ,in and Represents the maximum delay for two different channels, with channel sparsity of [missing information]. and ;

[0011] (3) The OFDM system based on the intelligent reflector divides the bandwidth at the user end. Parallel subcarriers, The set of subcarrier positions is represented as It adopts a comb-shaped pilot structure. Among the subcarriers One pilot is used to assist in channel estimation. The set of pilot positions corresponding to each pilot symbol is represented as follows: ;

[0012] The second step is to construct the pilot optimization problem:

[0013] (1) Set the pilot signal to be transmitted as The received pilot signal is represented as:

[0014] ,

[0015] in, This indicates that the base station receives signals consisting of comb-shaped pilot signals. dimensional vector, It is a diagonal matrix. It is a Fourier leaf matrix. Let be a sparse vector with a small number of non-zero elements, the number of non-zero elements being . , It is Gaussian white noise;

[0016] (2) Based on the sparsity of vector h, a measurement matrix is ​​constructed according to compressed sensing theory. Minimize the cross-correlation value of the measurement matrix From this perspective, we construct the following optimization problem:

[0017] ,

[0018] in, , Pilot position The corresponding pilot power, the set of pilot powers is represented as , , and These represent the total pilot power, the upper bound of the pilot power, and the lower bound of the pilot power, respectively.

[0019] The third step is to design an algorithm to solve the pilot optimization problem:

[0020] Using the concept of encoding and decoding, the encoded variables simultaneously represent discrete and continuous variables, and the whale optimization algorithm is used to assist the encoding and decoding method in solving the above optimization problem:

[0021] (1) The construction length is Encoding vector ,Right now ,in Each encoded variable Simultaneously represents a discrete variable and a continuous variable binary variables Indicates the first Whether a subcarrier is selected as a pilot. Indicates the first The weights of each subcarrier are calculated as follows:

[0022] ;

[0023] (2) Encoding vector The decoding process is as follows: select the encoding vector. The largest of them The position corresponding to each value is the pilot position; Discrete variables corresponding to each pilot position The value is 1, and the remainder is... The corresponding positions If the value is 0, calculate the vector. Weight of each element , The pilot power corresponding to each pilot position is expressed as:

[0024] ;

[0025] (3) The optimization problem is solved iteratively by using the whale optimization algorithm to assist the encoding and decoding algorithm. The search agent dimension is: Each search agent represents an encoded vector, and the fitness function is the cross-correlation value of the measurement matrix. When calculating the fitness function, the search agent is decoded into a set of pilot positions and a set of pilot powers, and the fitness value is calculated by substituting these into the cross-correlation value of the measurement matrix. The specific process for solving the optimization problem is as follows:

[0026] (3-1) Initialize parameters and population location, population size is The maximum number of iterations is Current iteration number ;

[0027] (3-2) Each search agent in the population Decoded into pilot location set and pilot power set ;

[0028] (3-3) Calculate the cross-correlation value of the measurement matrix for each search agent;

[0029] (3-4) Find the current optimal search agent;

[0030] (3-5) Update the population position according to the whale optimization algorithm;

[0031] (3-6) Each search agent in the population Decoded into pilot location set and pilot power set ;

[0032] (3-7) Calculate the cross-correlation value of the measurement matrix for each search agent;

[0033] (3-8) Find the current optimal search agent;

[0034] (3-9) Number of update iterations Determine if the maximum number of iterations has been reached. ;

[0035] (3-10) If Return to step (3-5) to update the population position based on the current optimal search agent; if Output the optimal search agent and decode the optimal search agent into a set of optimal pilot positions and pilot powers.

[0036] Compared with the prior art, the present invention has the following advantages:

[0037] (1) This invention considers the combined influence of pilot position and pilot power on channel estimation accuracy. Compared with the equal-spacing pilot allocation method, the constructed measurement matrix is ​​more helpful in recovering channel information and improves the accuracy of channel estimation.

[0038] (2) The present invention uses encoding and decoding to use a single variable to simultaneously represent the pilot position and pilot power, thereby reducing computational complexity while ensuring the accuracy of channel estimation. Attached Figure Description

[0039] Figure 1 This is a flowchart of the pilot optimization method described in this invention;

[0040] Figure 2 This is a schematic diagram of OFDM system communication based on intelligent reflective surfaces;

[0041] Figure 3 This is a flowchart of the encoding and decoding process in the pilot optimization method described in this invention;

[0042] Figure 4 This is a flowchart of the pilot optimization process for the codec-assisted whale optimization algorithm;

[0043] Figure 5 This is a comparative diagram showing the normalized mean square error of channel impulse response estimation as a function of channel signal-to-noise ratio for the algorithm described in this invention and the uniform pilot allocation method.

[0044] Figure 6 This is a schematic diagram comparing the bit error rate of the algorithm described in this invention and the uniform pilot allocation method as a function of the channel signal-to-noise ratio. Detailed Implementation

[0045] The technical solution and effects of the present invention will be further described in detail below with reference to the accompanying drawings.

[0046] For pilot optimization applications in practical systems, we often hope to use simpler pilot allocation algorithms to estimate more accurate channel state information. However, due to limitations such as the joint optimization complexity of pilot position and pilot power, and the need for intelligent reflector-assisted wireless channel structures, existing methods cannot simultaneously optimize these two aspects, often resulting in a trade-off. To improve the accuracy of channel estimation, this invention provides a pilot optimization method for OFDM sparse channels based on intelligent reflectors, referring to... Figure 1 The implementation steps of this invention are as follows:

[0047] The first step is to construct an OFDM system model based on a smart reflector and model the sparse channel. (Refer to...) Figure 2 Considering a single-input single-output (SSO) uplink system, with both the base station and user end using a single antenna, the number of intelligent reflector elements is... The reflection amplitude of each reflector element is set to 1 using a Smart Reflector Controller (RIS), and the reflection phase is set so that the reflected signal is directed towards the base station. This constructs the user-base station direct connection channel impulse response. User-Smart Reflector-Base Station Cascaded Channel Impulse Response , and Let represent the maximum delay of the direct channel and the cascaded channel, respectively. The sparsity of the two channels are respectively... and Cascaded channel impulse response It can be represented as:

[0048] ,

[0049] Where * represents the convolution symbol, For the first The reflection coefficient of each reflecting unit. and Representing user - number The first reflective unit and the first Channel impulse response of a single reflector-base station and Representing user - number The first reflective unit and the first Maximum channel delay per reflection unit-base station Indicates user - number Channel impulse response of a single reflector unit-base station.

[0050] OFDM systems based on smart reflectors divide the bandwidth at the user end. Parallel subcarriers, The set of subcarrier positions is represented as It also employs a comb-shaped pilot structure. Among the subcarriers One pilot is used to assist in channel estimation. The set of pilot positions corresponding to each pilot symbol is represented as follows: .

[0051] The second step is to construct an optimization problem from the perspective of minimizing the cross-correlation value of the measurement matrix. Let the transmitted pilot signal be... Received pilot signal It can be represented as:

[0052] ,

[0053] in, It is a diagonal matrix composed of pilot signals. and Let the pilot signals be the Fourier matrices for the cascaded and direct-connected channels, respectively. ,but and The former The columns are completely identical. Let be a sparse vector with sparsity of . , The noise is Gaussian white noise; according to compressed sensing theory, the measurement matrix for receiving the pilot signal is... .

[0054] Based on the fact that the cross-correlation value of a matrix is ​​the maximum value of the inner product of any two normalized columns, the cross-correlation value of the measurement matrix D is constructed. :

[0055] ,

[0056] in, Represents the L2 norm, Indicates the inner product. and These are any two columns of matrix D. It is the number of unique column vectors in D. Pilot position The corresponding pilot power, the set of pilot powers is represented as The variables for the cross-correlation values ​​of the above measurement matrices include the pilot position set. and pilot power set ,set up , and Let the total pilot power, upper bound of pilot power, and lower bound of pilot power be denoted as , respectively. The optimization problem based on minimizing the cross-correlation value of the measurement matrix can be expressed as:

[0057] .

[0058] The optimization problem described above is a mixed-integer nonlinear programming problem that includes pilot position variables and pilot power variables, and it is a nonconvex optimization problem.

[0059] The third step is to design a method for solving the optimization problem using an encoding / decoding-assisted whale optimization algorithm.

[0060] (1) Design the encoding vector to represent both discrete and continuous vectors: such as Figure 3 As shown, the encoding vector The length is ,Right now ,in Each encoded element Simultaneously represents a discrete element and a consecutive element binary elements Indicates the first Whether a subcarrier is selected as a pilot. Indicates the first The power weights of each subcarrier are calculated as follows:

[0061] ;

[0062] (2) Design the decoding process to decode the encoded vector into pilot positions and pilot power: Select the encoded vector The largest of them The position corresponding to each value is the pilot position; set this... Discrete elements corresponding to each pilot position The value is 1, and the remainder is... The corresponding positions If the value is 0, calculate the vector. Weight of each element , The pilot power corresponding to each pilot position is expressed as:

[0063] ;

[0064] (3) The optimization problem is solved iteratively by using encoding and decoding to assist the whale optimization algorithm, such as Figure 4 As shown, the search agent dimension is Each search agent represents an encoded vector, and the fitness function is the cross-correlation value of the measurement matrix. When calculating the fitness function, the search agent is decoded into a set of pilot positions and a set of pilot powers, and the fitness value is calculated by substituting these into the cross-correlation value of the measurement matrix. The specific process for solving the optimization problem is as follows:

[0065] (3-1) Initialize parameters and population location, population size is The maximum number of iterations is Current iteration number ;

[0066] (3-2) Each search agent in the population Decoded into pilot location set and pilot power set ;

[0067] (3-3) Calculate the cross-correlation value of the measurement matrix for each search agent;

[0068] (3-4) Find the current optimal search agent;

[0069] (3-5) Update the population position according to the whale optimization algorithm;

[0070] (3-6) Each search agent in the population Decoded into pilot location set and pilot power set ;

[0071] (3-7) Calculate the cross-correlation value of the measurement matrix for each search agent;

[0072] (3-8) Find the current optimal search agent;

[0073] (3-9) Number of update iterations Determine if the maximum number of iterations has been reached. ;

[0074] (3-10) If Return to step (3-5) to update the population position based on the current optimal search agent; if Output the optimal search agent and decode the optimal search agent into a set of optimal pilot positions and pilot powers.

[0075] On the one hand, this invention considers the impact of pilot position and pilot power on channel estimation performance simultaneously, and further reduces the cross-correlation value of the measurement matrix by compressed sensing, thereby improving the accuracy of channel estimation. On the other hand, it takes advantage of the mutual coupling between pilot position and pilot power, uses a single coded variable to represent both variables, and combines the whale optimization algorithm to solve the non-convex optimization problem, thereby controlling the complexity of pilot allocation.

[0076] The effects of the present invention will be further described below with reference to simulation examples.

[0077] Simulation conditions: Number of reflective elements in the intelligent reflective surface , the number of subcarriers Direct channel length sparsity Cascaded channel length sparsity The channel signal-to-noise ratio ranges from 0dB to 20dB, and the threshold for the number of iterations in the iterative optimization is set to... The number of search proxies is set to That is, the population size is .

[0078] Simulation Example 1: A comparative diagram showing the normalized mean square error (NMSE) of channel impulse response estimation between the algorithm described in this invention and the uniform pilot allocation method as a function of channel signal-to-noise ratio (SNR) is shown below. Figure 5 As shown in the figure, it can be seen that as the signal-to-noise ratio increases, the mean square error of channel estimation gradually decreases; with the increase of the number of pilots... =24 and When the frequency is 32, the channel recovery error using the pilot allocation method described in this invention is significantly smaller than the channel recovery error using the uniform pilot allocation method. This demonstrates that the algorithm described in this invention improves the accuracy of channel estimation. Furthermore, the pilot allocation algorithm described in this invention... =24 and uniform pilot distribution method The mean square error curves at 32 almost overlap, indicating that the pilot allocation scheme proposed in this invention improves the efficiency of pilot usage.

[0079] Simulation Example 2: A comparative diagram showing the bit error rate (BER) of the algorithm described in this invention and the uniform pilot allocation method as a function of channel signal-to-noise ratio (SNR) is shown below. Figure 6 As shown in the figure, the bit error rate gradually decreases as the signal-to-noise ratio increases; with the number of pilots... =24 and When the pilot frequency is 32, the bit error rate of restoring the transmitted signal using the pilot allocation method described in this invention is significantly lower than that of restoring the transmitted signal using the uniform pilot allocation method. This shows that the algorithm described in this invention improves the accuracy of transmitted signal restoration. Furthermore, when the channel signal-to-noise ratio is between 15dB and 20dB, the bit error rate of restoring the transmitted signal using the algorithm described in this invention with 24 pilots is lower than that using the uniform pilot allocation method with 32 pilots, demonstrating the advantage of pilot frequency utilization efficiency.

[0080] In summary, this invention primarily addresses the problem that existing pilot allocation techniques cannot simultaneously optimize the signal estimation performance of pilot position and pilot power. It considers the advantages of the coupling between pilot position and pilot power, using a single coded variable to represent both variables, and employs a whale optimization algorithm to iteratively optimize the cross-correlation value of the measurement matrix, thus controlling the complexity of pilot allocation. Simulation results demonstrate that the proposed method effectively improves channel estimation accuracy and further enhances the accuracy of reconstructing the transmitted signal, laying a stable foundation for achieving efficient and reliable communication.

Claims

1. A pilot optimization method for OFDM sparse channels based on intelligent reflectors, characterized in that, Includes the following steps: The first step is to construct an OFDM system model and channel model based on a smart reflector: (1) The OFDM uplink single-input single-output system based on a smart reflector consists of a base station, a user terminal, and a smart reflector. Both the base station and the user terminal are equipped with a single antenna. The smart reflector is composed of... It consists of a reflective unit, and is set to... The reflection amplitude of each reflection unit is 1, and the reflection phase makes the reflected signal point towards the base station. (2) Model the OFDM sparse channel based on the smart reflector and construct the impulse response of the direct sparse channel without passing through the smart reflector. Impulse response of cascaded sparse channel through intelligent reflector ,in and Represents the maximum delay for two different channels, with channel sparsity of [missing information]. and The cascaded sparse channel impulse response is obtained by convolving and superimposing the channel impulse responses from the user terminal to each reflection unit of the smart reflector, the reflection coefficients of each reflection unit, and the channel impulse responses from each reflection unit to the base station. Where * represents the convolution symbol, For the first The reflection coefficient of each reflecting unit. and Representing user - number The first reflective unit and the first Channel impulse response of a single reflector-base station Indicates user - number Channel impulse response of a single reflector unit-base station; (3) The OFDM system based on the intelligent reflector divides the bandwidth at the user end. Parallel subcarriers, The set of subcarrier positions is represented as It adopts a comb-shaped pilot structure. Among the subcarriers One pilot is used to assist in channel estimation. The set of pilot positions corresponding to each pilot symbol is represented as follows: ; The second step is to construct the pilot optimization problem: (1) Set the pilot signal to be transmitted as The received pilot signal is represented as: ， in, This indicates that the base station receives signals consisting of comb-shaped pilot signals. dimensional vector, It is a diagonal matrix. It is a Fourier leaf matrix. Let be a sparse vector with a small number of non-zero elements, the number of non-zero elements being . , It is Gaussian white noise; (2) Based on the sparsity of vector h, a measurement matrix is ​​constructed according to compressed sensing theory. Minimize the cross-correlation value of the measurement matrix From this perspective, we construct the following optimization problem: ， in, , Pilot position The corresponding pilot power, the set of pilot powers is represented as , , and 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 is to design an algorithm to solve the pilot optimization problem: Using pilot position and pilot power as variables to be optimized, and employing the concept of encoding and decoding, the coded variables simultaneously represent discrete and continuous variables. The whale optimization algorithm is then used to assist the encoding and decoding method in solving the above optimization problem. (1) The construction length is Encoding vector ,Right now ,in Each encoded variable Simultaneously represents a discrete variable and a continuous variable binary variables Indicates the first Whether a subcarrier is selected as a pilot. Indicates the first The weights of each subcarrier are calculated as follows: ； (2) Encoding vector The decoding process is as follows: select the encoding vector. The largest of them The position corresponding to each value is the pilot position; Discrete variables corresponding to each pilot position The value is 1, and the remainder is... The corresponding positions If the value is 0, calculate the vector. Weight of each element , The pilot power corresponding to each pilot position is expressed as: ； (3) The optimization problem is solved iteratively by using encoding and decoding to assist the whale optimization algorithm, with the search agent dimension being... Each search agent represents an encoded vector, and the fitness function is the cross-correlation value of the measurement matrix. When calculating the fitness function, the search agent is decoded into a set of pilot positions and a set of pilot powers, and the fitness value is calculated by substituting these into the cross-correlation value of the measurement matrix. The specific process for solving the optimization problem is as follows: (3-1) Initialize parameters and population location, population size is The maximum number of iterations is Current iteration number ; (3-2) Each search agent in the population Decoded into pilot location set and pilot power set ; (3-3) Calculate the cross-correlation value of the measurement matrix for each search agent; (3-4) Find the current optimal search agent; (3-5) Update the population position according to the whale optimization algorithm; (3-6) Each search agent in the population Decoded into pilot location set and pilot power set ; (3-7) Calculate the cross-correlation value of the measurement matrix for each search agent; (3-8) Find the current optimal search agent; (3-9) Number of update iterations Determine if the maximum number of iterations has been reached. ; (3-10) If Return to step (3-5) to update the population position based on the current optimal search agent; if Output the optimal search agent and decode the optimal search agent into a set of optimal pilot positions and pilot powers.

Images