Low-complexity receiver suitable for double selective fading channel and signal detection method

Abstract

The invention discloses a low-complexity receiver suitable for a dual-selective fading channel, and mainly solves the problems of inaccurate data detection and high operation complexity of a receiver of an existing wireless communication system in a dual-selective fading scene. The device comprises a channel estimator and a channel equalizer which are bidirectionally connected. The channel estimator carries out channel estimation and sends an obtained channel estimation value and an equivalent noise variance to the equalizer, the equalizer detects a received signal by using the channel estimation value and the equivalent noise variance, then splices an output posteriori mean value and a known pilot frequency into a new pilot frequency sequence, and inputs the new pilot frequency sequence to the channel estimator for channel estimation optimization; and the channel estimator sends the optimized channel estimation value and the equivalent noise variance into the equalizer, and the circulation interaction is carried out in this way, so that the accurate estimation of the channel state and the accurate detection of the data signal are realized. According to the method, the calculation complexity can be reduced, the receiving stability and the data detection accuracy are improved, and the method can be used for a communication system under a double-selective fading channel.

Description

Technical Field

[0001] The present invention belongs to the field of communication technology, and further provides a low-complexity receiver and signal detection method, which can be used in a communication system under a dual-selective fading channel to achieve efficient and accurate recovery of the transmitted data signal at the receiving end. Background Art

[0002] With the development of 5G / B5G and 6G communication technologies, the demand for highly efficient and reliable communications has increased significantly. However, the combined effects of multipath and Doppler shift can cause severe distortion of the received signal and reduce channel capacity. Therefore, achieving efficient and reliable data transmission in dual-selective fading channels has become an urgent task in wireless communications. To measure the reliability level of data transmission, the bit error rate (BER) is defined as the ratio of the number of bits received in error at the receiver to the total number of bits transmitted. Reducing the BER has become an urgent task in receiver design in wireless communication systems. In dual-selective fading channels, the OAMP equalization algorithm can effectively recover the data signal, but its computational complexity is too high. Therefore, reducing computational complexity and improving resource utilization while ensuring transmission reliability are important goals in wireless communication receiver design.

[0003] Patent document No. 202310956461.5 discloses a "single-carrier MIMO underwater acoustic communication method", in which the communication receiver operates in a block-by-block iterative manner: (1) an initialization model of the received signal is established based on the received data, and the LS algorithm is used for channel estimation. (2) interference elimination and soft equalization based on time-domain vector approximate message passing are performed, and the output result is demapped to obtain log-likelihood ratio information. (3) deinterleaving and channel decoding. This method has two shortcomings: first, due to the use of the VAMP equalization algorithm, high-dimensional vector state evolution is required in each iteration, which is highly complex and computationally expensive, and requires the storage of high-dimensional intermediate variables, resulting in high memory usage; second, due to the poor dynamic channel adaptability of the receiver, statistical parameters need to be frequently updated in fast time-varying channels, resulting in a decrease in the receiver's data detection speed and accuracy.

[0004] In their paper "An Iterative Receiver With OrthogonalAMP-Based Equalization for Doubly Selective Fading Channels" (2022 IEEE / CIC International Conference on Communications in China (ICCC), 2022, pp. 196-201), Yizhuo Wang et al. proposed an iterative receiver design that considers information exchange between different receiver modules. They first quantize the noise level from the output of the channel estimator to provide data detection for the equalizer. They then employ an orthogonal approximate message passing (OAMP) algorithm for equalization. Channel interpolation and decoding feedback strategies are then used to mitigate the effects of time-varying channels, while short pilot sequences are used to achieve efficient transmission. The OAMP equalization algorithm used in this method suffers from high computational complexity and high computational complexity due to the complex matrix inversion operations involved in its LMMSE linear detector module. This channel equalization scheme fails when the channel transmission matrix is ​​high in dimensionality, resulting in poor resource utilization. Summary of the Invention

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and propose a low-complexity receiver and signal detection method suitable for dual-selective fading channels, so as to improve the accuracy of data detection and resource utilization under dual-selective fading channels.

[0006] The technical approach to achieving the objectives of the present invention is to convert the complex matrix inversion operation in the LMMSE detector into a diagonal matrix inversion operation by using the same pilot sequence, and to reduce the complexity and resource utilization of the OAMP equalizer by utilizing frequency domain characteristics. Furthermore, the accuracy of channel estimation and the adaptability to time-varying channels are improved through joint iteration between the channel estimator and the equalizer, thereby increasing the accuracy of data detection in dual-selective fading channels.

[0007] According to the above ideas, the technical solution of the present invention includes:

[0008] 1. A low-complexity receiver for a dual-selective fading channel, comprising a channel estimator and a channel equalizer, characterized in that:

[0009] The channel equalizer includes an LMMSE linear detector, a decorrelation module and a nonlinear demodulator:

[0010] The LMMSE linear detector is used to calculate its own posterior mean and variance based on the received signal, and input it to the decorrelation module for decorrelation operation to obtain the prior mean and variance of the nonlinear demodulator;

[0011] The nonlinear demodulator demodulates the prior mean and variance of its input to obtain its own posterior mean, variance, and log-likelihood ratio information, performs decorrelation operation on these information, and returns the result to the LMMSE linear detector to complete OAMP iterative equalization.

[0012] The channel estimator is bidirectionally connected to the channel equalizer for iterative information interaction between the two. That is, the OAMP equalization result completed by the channel equalizer is spliced ​​with the known pilot to form a new pilot sequence, which is input into the channel estimator for channel estimation optimization. The optimized channel estimation error is then quantized into an equivalent noise variance and sent to the equalizer. This cyclic interaction achieves accurate estimation of the channel state and accurate detection of the data signal.

[0013] Furthermore, the LMMSE linear detector comprises:

[0014] A fast Fourier transform module is used to perform Fourier transform on the received signal to obtain a frequency domain received signal;

[0015] The linear detection module uses the characteristics of the received signal in the frequency domain to transform the complex matrix inversion process into a diagonal matrix inversion process.

[0016] 2. A method for detecting a received signal using the above receiver, comprising:

[0017] Perform least squares LS channel estimation on the received pilot signal and use linear interpolation to obtain the channel estimation value h of the unknown data segment between two known pilot segments. k , and quantify the error of channel estimation as the equivalent noise variance of the equalizer Transmit to the equalizer;

[0018] The equalizer uses the channel estimation value, equivalent noise variance, and the received signal of the unknown data segment to perform low-complexity frequency domain OAMP iterative equalization to obtain an accurate mean value. variance Log-likelihood ratio information LLR k ;

[0019] The average value of the frequency domain OAMP equalizer output As known data, and the pilot vector p k+1 Splice into a new pilot vector The signal is transmitted to the channel estimator for channel estimation optimization, and then the optimized channel estimation result is transmitted to the equalizer. This cycle of interaction is repeated to achieve accurate estimation of fast time-varying channels and accurate detection of data signals.

[0020] Furthermore, the channel estimator optimizes the channel estimation result, and its implementation includes:

[0021] The posterior mean of the frequency domain OAMP equalizer output Treat it as known data and compare it with the pilot vector p k+1 Splice into a new pilot vector Transmit to the channel estimator for a new channel estimation;

[0022] The kth unknown data segment receives the signal and the k+1th known pilot segment received signal Splice into a new received signal

[0023] According to the new received signal and the new pilot vector Get the new channel estimate h k , completing the optimization of channel estimation.

[0024] Compared with the prior art, the present invention has the following advantages:

[0025] First, the low-complexity OAMP frequency domain equalizer in the receiver of the present invention uses the same pilot sequence to convert the complex matrix inversion operation into a diagonal matrix inversion, which can greatly reduce the complexity of the calculation, overcome the problem that the existing technology is difficult to meet the low latency requirements of the real-time communication system, and enable the present invention to improve resource utilization.

[0026] Secondly, the present invention introduces iterative interaction between the channel estimator and the equalizer when using the receiver for data detection, and uses the mean value obtained by the equalizer as a new pilot to update the channel estimation result, thereby improving the accuracy of channel estimation; at the same time, since the estimation error is quantified and the parameters of the equalizer are adjusted to further optimize the performance of the receiver, it can overcome the problem that the existing technology cannot accurately track channel changes in fast time-varying channels, and significantly improves the accuracy of data detection under the same pilot resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a structural block diagram of a low-complexity receiver suitable for dual-selective fading channels provided by an embodiment of the present invention;

[0028] Figure 2 yes Figure 1 The OAMP equalizer structure diagram in the figure;

[0029] Figure 3 is a flowchart of an implementation of detecting a transmitted data signal using the receiver provided in an embodiment of the present invention;

[0030] Figure 4 This is a simulation diagram of the bit error rate of data detected using the signal detection method of the present invention and the existing signal detection method. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0032] Embodiment 1: Low-complexity receiver for dual-selective fading channels

[0033] Reference Figure 1 and Figure 2 , this example includes a channel estimator 1 and a channel equalizer 2, wherein:

[0034] The channel estimator 1 includes: a pilot extraction module 11, a least squares estimation module 12, and a channel interpolation module 13. The pilot extraction module 11 is used to separate a known pilot signal from a received signal; the least squares estimation module 12 is used to calculate the channel response at the pilot location based on the pilot signal separated by the pilot extraction module 11 and the known transmitted pilot signal; and the channel interpolation module 13 is used to estimate the channel response at a non-pilot location by linear interpolation based on the channel response at the pilot location obtained by the least squares estimation module 12.

[0035] The channel equalizer 2 includes a LMMSE linear detector 21, a decorrelation module 22 and a nonlinear demodulator 23:

[0036] The LMMSE linear detector 21 is used to calculate its own posterior mean and variance based on the received signal, and includes a fast Fourier transform module 211 and a linear detection module 212; the fast Fourier transform module 211 performs Fourier transform on the received signal to obtain a frequency domain received signal, and transmits it to the linear detection module 212; the linear detection module 212 uses the characteristics of the frequency domain received signal to transform the complex matrix inversion process into a diagonal matrix inversion process.

[0037] The decorrelation module 22 has an input end connected to the LMMSE linear detector 21 and an output end connected to the nonlinear demodulator 23. It decorrelates the posterior mean and variance output by the LMMSE linear detector using the Schmidt orthogonalization method to obtain the prior mean and variance, thereby ensuring the independence and Gaussianity of the estimation error.

[0038] The nonlinear demodulator 23 adopts a demodulation scheme based on the maximum likelihood criterion to demodulate the priori mean and variance output by the decorrelation module 22 to effectively suppress the inter-symbol interference caused by the dual-selective fading channel, obtain accurate posterior mean, variance and log-likelihood ratio information, and then transmit it to the decorrelation module 22 for decorrelation operation to obtain the priori mean and variance, and then return it to the LMMSE linear detector 21 to complete the iterative interaction.

[0039] The above-mentioned channel estimator 1 and channel equalizer 2 are bidirectionally connected to perform information iterative interaction, that is, channel estimator 1 performs channel estimation and sends the obtained channel estimation value and equivalent noise variance to equalizer 2, equalizer 2 splices the output posterior mean with the known pilot into a new pilot sequence, inputs it to channel estimator 1 for channel estimation optimization, and then quantizes the optimized channel estimation error into an equivalent noise variance and sends it to equalizer 2. This cyclic interaction realizes accurate estimation of the channel state and accurate detection of the data signal.

[0040] It should be noted that the above-mentioned functional modules can be implemented in whole or in part through software, hardware, firmware, or any combination thereof. When implemented using software, they can be implemented in whole or in part in the form of a program instruction product. The program instruction product includes one or a group of program instructions. When the program instructions are loaded and executed on a computer, the process or function described is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The program instructions can be stored in a computer-readable and writable storage medium, or transferred from a computer-readable and writable storage medium to another computer-readable and writable storage medium.

[0041] The direct coupling or communication connection between the modules shown or discussed in this embodiment can be achieved through indirect coupling or communication connection of some interfaces, devices or modules. The various functional modules and submodules in this embodiment can be dynamically located in a processing component, or each module can exist physically separately, or two or more modules can be dynamically located in a processing component. When the above-mentioned dynamic components are implemented in the form of software functional modules and sold or used as independent products, they can also be stored in a computer-readable and writable storage medium. The storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0042] Embodiment 2 is a method for detecting data using the above-mentioned low-complexity receiver.

[0043] This example uses the low-complexity receiver described above to efficiently and accurately detect the signal sent by the transmitter under a doubly selective fading channel. That is, the pilot signal and unknown data signal sent by the transmitter are processed, and the unknown data signal sent is recovered at the receiver.

[0044] Reference Figure 3 The implementation steps of this example include the following:

[0045] Step 1: Perform least squares (LS) channel estimation on the received pilot signal to obtain channel estimation values ​​for the left and right pilot segments.

[0046] 1.1) The cyclic pilot matrix P constructed using the k-th pilot sequence k , perform the least squares estimation on it to obtain the channel estimation value of the left pilot segment:

[0047] 1.2) The cyclic pilot matrix P constructed using k+1 pilot sequences k+1 , perform the least squares estimation on it to obtain the channel estimation value of the right pilot segment:

[0048] Among them C n represents the covariance matrix of the noise, represents the received signal of the k-th pilot segment, represents the received signal of the k+1th pilot segment, and H represents the conjugate transpose.

[0049] Step 2: Use linear interpolation to obtain the channel estimation value h of the unknown data segment k , and calculate the equivalent noise variance of the data segment

[0050] 2.1) Calculate the channel estimation value h of the middle unknown data segment based on the channel estimation values ​​of the two known pilot segments on the left and right k :

[0051] h k =(h left +h right ) / 2;

[0052] 2.2) According to the cyclic pilot matrix P k The noise variance introduced by channel estimation is calculated by adding channel noise w

[0053]

[0054] in, P represents the cyclic pilot matrix k Pseudo-inverse matrix, Tr represents trace operation.

[0055] 2.3) Noise variance introduced by channel estimation and channel noise variance Get the equivalent noise variance

[0056]

[0057] Step 3: The equalizer uses the channel estimation value h k , equivalent noise variance Unknown data segment reception signal Perform low-complexity frequency-domain OAMP iterative equalization.

[0058] 3.1) The prior mean of the LMMSE linear detector Set to 0, the prior variance Set to 1;

[0059] 3.2) According to the prior mean of the current LMMSE linear detector and the prior variance Calculate the posterior mean of the linear detector and the posterior variance

[0060]

[0061] Among them C x Indicates the use of the detector prior mean The constructed diagonal matrix, F is the normalized DFT matrix, Represents the received signal in the frequency domain, and D is composed of diagonal elements The diagonal matrix formed, h k It is H k The first column, H k The channel estimate h k Constructed cyclic channel response matrix, N = U n +K n Indicates the processing length, U n Indicates unknown data segment length, K n represents the known pilot segment length, and T represents transposition;

[0062] 3.3) The decorrelation module uses the Schmidt orthogonalization method to calculate the k-th posterior mean of the LMMSE linear detector output. and the posterior variance Perform decorrelation operation to obtain the prior mean of the demodulator and the prior variance

[0063]

[0064]

[0065] in, represents the demodulator prior mean The nth element in, n∈{1,2,…,U n}, Represents the posterior mean of the LMMSE linear detector The nth element in Represents the LMMSE linear detector prior mean The nth element in

[0066] 3.4) Nonlinear demodulator using prior mean and the prior variance The calculated k-th log-likelihood ratio information sequence:

[0067]

[0068] Among them, a k,m represents the mth bit of the symbol generated after modulation, Indicates that the received signal is Time a k,m =1, Indicates that the received signal is Time a k,m =0 conditional probability;

[0069] 3.5) Using the log-likelihood ratio sequence LLR(a k,m ) Calculate the posterior probability of the nth bit being 1

[0070]

[0071] Among them, p1 represents the prior probability that the nth bit is 1;

[0072] 3.6) According to Get the posterior mean of the nonlinear detector and the posterior variance

[0073]

[0074] in represents the posterior mean of the nonlinear detector The nth element in , c0 represents the constellation point mapped by bit 0, and c1 represents the constellation point mapped by bit 1;

[0075] 3.7) The decorrelation module uses the Schmidt orthogonalization method to calculate the kth segment posterior mean of the nonlinear demodulator output and the posterior variance Perform decorrelation operation to obtain the prior mean of LMMSE linear detector and the prior variance

[0076]

[0077] in represents the demodulator posterior mean The nth element in ;

[0078] 3.8) The prior mean of the LMMSE linear detector obtained in step 3.7) is and the prior variance Return to the linear detector, i.e. return to step 3.2);

[0079] 3.9) Repeat steps 3.2) to 3.8) to achieve OAMP iterative balancing.

[0080] The number of cycles in this example is set but not limited to 5 times.

[0081] Step 4: Use the frequency domain OAMP equalizer detection results to optimize channel estimation:

[0082] 4.1) The posterior mean of the frequency domain OAMP equalizer output after iteration Treat it as known data and compare it with the pilot vector p k+1 Splice into a new pilot vector Transmit to the channel estimator for a new channel estimation;

[0083] 4.2) The kth unknown data segment receives the signal and the k+1th known pilot segment received signal Splice into a new received signal

[0084] 4.3) According to the new received signal and the new pilot vector Calculate the new channel estimate h k :

[0085]

[0086] in, Indicates the use of new pilot vector a cyclic pilot matrix obtained by cyclic shift;

[0087] 4.4) According to the channel noise w and the cyclic pilot matrix Calculate the noise variance introduced by channel estimation and the new equivalent noise variance

[0088]

[0089] in, represents the channel noise variance, represents the cyclic pilot matrix The pseudo-inverse matrix of .

[0090] Step 5: The optimized channel estimation value h obtained in step 4 is k and the equivalent noise variance The signal is sent back to the equalizer, and the equalizer performs step 3 again based on the new channel estimation value and equivalent noise variance.

[0091] Repeat steps 3 to 4 to implement iterative interaction between the channel estimator and the equalizer, and finally obtain the information sequence of the transmitted data signal.

[0092] The number of cycles in this example is set but not limited to 5 times.

[0093] It should be noted that the flowchart representations or method representations of the above embodiments can be understood as representing a module, segment, or portion of code that includes one or a group of executable instructions configured to implement a specific logical function or process. The present invention is not limited to the disclosed preferred embodiments, and its implementation may not follow the order presented or discussed. In other words, the step numbers are only for the purpose of clarifying the embodiment of the present invention and facilitating understanding, and the order of the step numbers is not limiting.

[0094] The effects of the present invention can be further illustrated by the following simulation results:

[0095] 1. Simulation conditions

[0096] The communication system is simulated using the MATLAB2023a platform.

[0097] In the transmitting end framing, the pilot length is set to 20, the unknown data segment length is set to 128, a total of 10 packets are transmitted, a 1 / 2 code rate convolutional code is used for channel coding, the number of internal iterations of the OAMP equalizer is set to 5, the number of iterations of the channel estimator and equalizer is set to 5, and the simulation channel uses a double-selective fading channel.

[0098] 2. Simulation content and results

[0099] Under the above conditions, the bit error rates of the transmitted data signals are compared using the present invention and the existing method without iterating the channel estimator and the equalizer. The results are as follows: Figure 4 shown.

[0100] from Figure 4 It can be seen that under the condition of the same channel noise power, the bit error rate of the present invention is much lower than that of the prior art. When the channel noise power is 0 dB, the bit error rate of the present invention is 3×10 -4 , the bit error rate of the existing technology is 10 -2 ; When the bit error rate is equal to 10 -3When , the present invention can obtain a gain of 12dB, indicating that the receiving performance of the present invention is significantly better than that of the prior art.

Claims

1. A low-complexity receiver for a dual-selective fading channel, comprising a channel estimator and a channel equalizer, characterized in that: The channel equalizer includes an LMMSE linear detector, a decorrelation module and a nonlinear demodulator: The LMMSE linear detector is used to calculate its own posterior mean and variance based on the received signal, and input it to the decorrelation module for decorrelation operation to obtain the prior mean and variance of the nonlinear demodulator; The nonlinear demodulator demodulates the prior mean and variance of its input to obtain its own posterior mean, variance, and log-likelihood ratio information, performs decorrelation operation on these information, and returns the result to the LMMSE linear detector to complete OAMP iterative equalization. The channel estimator is bidirectionally connected to the channel equalizer for iterative information interaction between the two. That is, the OAMP equalization result completed by the channel equalizer is spliced ​​with the known pilot to form a new pilot sequence, which is input into the channel estimator for channel estimation optimization. The optimized channel estimation error is then quantized into an equivalent noise variance and sent to the equalizer. This cyclic interaction achieves accurate estimation of the channel state and accurate detection of the data signal.

2. The receiver according to claim 1, wherein The channel estimator comprises: A pilot extraction module is used to separate the known pilot signal from the received signal; A least squares estimation module for calculating the channel response based on the received pilot signal and the known transmitted pilot signal; The channel interpolation module is used to estimate the channel response of non-pilot positions by linear interpolation.

3. The receiver according to claim 1, wherein The LMMSE linear detector comprises: A fast Fourier transform module is used to perform Fourier transform on the received signal to obtain a frequency domain received signal; The linear detection module uses the characteristics of the received signal in the frequency domain to transform the complex matrix inversion process into a diagonal matrix inversion process.

4. The receiver according to claim 1, wherein The decorrelation module performs decorrelation operation on the posterior mean and variance output by the LMMSE linear detector and the nonlinear demodulator through the Schmidt orthogonalization method, thereby ensuring the independence and Gaussianity of the estimation error.

5. The receiver according to claim 1, wherein The nonlinear demodulator adopts a demodulation scheme based on the maximum likelihood criterion to demodulate the input priori mean and variance to effectively suppress the inter-symbol interference caused by the dual selective fading channel and output accurate posterior mean, variance and log-likelihood ratio information.

6. A method for detecting a received signal using the receiver of claim 1, characterized in that: include: Perform least squares LS channel estimation on the received pilot signal and use linear interpolation to obtain the channel estimation value h of the unknown data segment between two known pilot segments. k , and quantify the error of channel estimation as the equivalent noise variance of the equalizer Transmit to the equalizer; The equalizer uses the channel estimation value, equivalent noise variance, and the received signal of the unknown data segment to perform low-complexity frequency domain OAMP iterative equalization to obtain an accurate mean value. variance Log-likelihood ratio information LLR k ; The average value of the frequency domain OAMP equalizer output As known data, and the pilot vector p k+1 Splice into a new pilot vector The signal is transmitted to the channel estimator for channel estimation optimization, and then the optimized channel estimation result is transmitted to the equalizer. This cycle of interaction is repeated to achieve accurate estimation of fast time-varying channels and accurate detection of data signals.

7. The method according to claim 6, characterized in that: The least squares LS channel estimation is performed on the received pilot signal, and the channel estimation value h of the unknown data segment between two known pilot segments is obtained by linear interpolation. k , whose implementation includes: The cyclic pilot matrix P constructed using the k-th pilot sequence k , perform the least squares estimation on it to obtain the channel estimation value of the left pilot segment: The cyclic pilot matrix P constructed using k+1 pilot sequences k+1 , perform the least squares estimation on it to obtain the channel estimation value of the right pilot segment: The channel estimation value h of the middle unknown data segment is calculated based on the channel estimation values ​​of the two known pilot segments on the left and right. k : h k =(h left +h right ) / 2, Among them C n represents the covariance matrix of the noise, represents the received signal of the k-th pilot segment, represents the received signal of the k+1th pilot segment, and H represents the conjugate transpose.

8. The method according to claim 6, characterized in that The error of channel estimation is quantified as the equivalent noise variance of the equalizer Its implementation includes: According to the cyclic pilot matrix P k And the channel noise w is used to calculate the noise variance introduced by the channel estimation: According to the noise variance introduced by channel estimation and the variance of the channel noise w Calculate the equivalent noise variance in P represents the cyclic pilot matrix k Pseudo-inverse matrix, H represents conjugate transpose.

9. The method according to claim 6, characterized in that The equalizer performs low-complexity frequency-domain OAMP iterative equalization using channel estimation values, equivalent noise variance, and received signals of unknown data segments. The implementation includes: (9a) The prior mean of the LMMSE linear detector Set to 0, the prior variance Set to 1; (9b) According to the prior mean of the current LMMSE linear detector and the prior variance Calculate the posterior mean of the linear detector and the posterior variance Among them C x Indicates the use of the detector prior mean The constructed diagonal matrix, F is the normalized DFT matrix, represents the received signal in the frequency domain, Represents the received signal of the kth unknown data segment, D is composed of the diagonal elements The diagonal matrix formed, h k It is H k The first column, H k The channel estimate h k Constructed cyclic channel response matrix, N = U n +K n Indicates the processing length, U n Indicates unknown data segment length, K n represents the known pilot segment length, and T represents transposition; (9c) The decorrelation module uses the Schmidt orthogonalization method to calculate the k-th posterior mean of the LMMSE linear detector output. and the posterior variance Perform decorrelation operation to obtain the prior mean of the demodulator and the prior variance in, represents the demodulator prior mean The nth element in, n∈{1,2,…,U n }, Represents the posterior mean of the LMMSE linear detector The nth element in Represents the LMMSE linear detector prior mean The nth element in (9d) Nonlinear demodulator using a priori mean and the prior variance The calculated k-th log-likelihood ratio information sequence: Among them, a k,m represents the mth bit of the symbol generated after modulation, Indicates that the received signal is Time a k,m =1, Indicates that the received signal is Time a k,m =0 conditional probability; (9e) Using the log-likelihood ratio sequence LLR(a k,m ) Calculate the posterior probability of the nth bit being 1 Among them, p1 represents the prior probability that the nth bit is 1; (9f) According to Get the posterior mean of the nonlinear detector and the posterior variance in represents the posterior mean of the nonlinear detector The nth element in , c0 represents the constellation point mapped by bit 0, and c1 represents the constellation point mapped by bit 1; (9g) The decorrelation module uses the Schmidt orthogonalization method to calculate the kth posterior mean of the nonlinear demodulator output. and the posterior variance Perform decorrelation operation to obtain the prior mean of LMMSE linear detector and the prior variance in represents the demodulator posterior mean The nth element in ; (9h) Substitute the prior mean of the LMMSE linear detector obtained in (9f) and the prior variance The signal is sent back to the linear detector, i.e., it returns to (9b), and the cycle repeats to achieve iterative interaction.

10. The method according to claim 6, characterized in that The channel estimator optimizes the channel estimation results. Its implementation includes: (10a) The posterior mean of the frequency domain OAMP equalizer output Treat it as known data and compare it with the pilot vector p k+1 Splice into a new pilot vector Transmit to the channel estimator for a new channel estimation; (10b) The kth unknown data segment receiving signal and the k+1th known pilot segment received signal Splice into a new received signal (10c) Based on the new received signal and the new pilot vector Get the new channel estimate h k and the new equivalent noise variance Complete the optimization of channel estimation: in, Indicates the use of new pilot vector The constructed cyclic pilot matrix, represents the noise variance introduced by channel estimation, w represents the channel noise, represents the channel noise variance, represents the cyclic pilot matrix The pseudo-inverse matrix of .

Images