Low-precision quantization MIMO-OTFS signal detection method based on serial interference cancellation

Abstract

The invention discloses a low-precision quantization MIMO-OTFS signal detection method based on serial interference cancellation, and relates to the technical field of wireless communication, and the method comprises the following steps: S1, processing original data to obtain a transmission signal, and transmitting the transmission signal; s2, sequentially performing low-precision quantization and demodulation on the received signal to obtain low-precision quantized received data; s3, calculating a signal to interference plus noise ratio by using a channel matrix, and inputting one path of low-precision quantized receiving data with the maximum signal to interference plus noise ratio into a dequantized Gaussian approximation message passing detector for detection; s4, eliminating interference by using a serial interference elimination method according to a detection result, and updating the channel matrix; s5, repeating the steps S3-S4, and detecting each path of low-precision quantized received data; and S6, outputting the detected data symbols and carrying out symbol demodulation. According to the method, the problems of difficulty in signal detection and high bit error rate in a low-precision quantization MIMO-OTFS system are solved, and better bit error rate performance can be realized.

Description

Technical Field

[0001] The invention relates to the technical field of wireless communications, and in particular to a low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation. Background Art

[0002] The next generation of wireless communication systems requires faster rates and higher reliability. OFDM is seriously affected by fast time-varying channels. It is difficult to maintain strict orthogonality of subcarriers under Doppler frequency shift expansion, resulting in inter-carrier interference (ICI), which will significantly reduce system performance and reliability. It performs poorly in scenarios such as satellite communications, airborne communications, vehicle communications, and high-speed rail. Orthogonal Time Frequency Space (OTFS) modulation is a new modulation technology. Compared with OFDM, which processes data in the time-frequency domain, OTFS maps valid data to the Delay-Doppler (DD) domain for transmission through a set of two-dimensional transformations. The dual-selective channel in the time-frequency domain is approximately a time-invariant channel in the DD domain, which can provide good transmission performance and reliability in scenarios such as satellite communications, airborne communications, vehicle communications, and high-speed rail.

[0003] Multiple-Input Multiple-Output (MIMO) is a technology that can improve spectrum efficiency and solve problems such as signal attenuation and multipath effects in wireless channels. Combining MIMO with OTFS can further improve transmission reliability and increase the channel capacity of the system, which also puts higher requirements on the signal detection algorithm. Linear algorithms such as linear minimum mean-square error (LMMSE) and zero forcing (ZF) need to invert the channel matrix. As the grid size and the number of antennas increase, the algorithm complexity will also increase, which is not practical in OTFS. The nonlinear algorithm message passing algorithm (MP) is considered to be a method to achieve the best performance of signal detection in the OTFS system, but it still has problems such as high complexity. When there is hardware damage at the receiving end or a low-precision ADC is used to reduce cost and power, the above algorithms have poor performance. Therefore, an algorithm is urgently needed to solve the signal detection problem in the above scenarios. Summary of the invention

[0004] In order to solve the above problems, an object of the present invention is to provide a low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation.

[0005] In order to achieve the above object, the present invention adopts the following technical solution.

[0006] S1. Process the original data to obtain a transmission signal and transmit it;

[0007] S2, performing low-precision quantization and demodulation on the received signal in sequence to obtain low-precision quantized received data;

[0008] S3, calculating the minimum mean square error weight matrix according to the delay-Doppler domain channel matrix H; calculating the signal to interference plus noise ratio SINR of each signal according to the minimum mean square error weight matrix, and inputting the low-precision quantized received data with the largest signal to interference plus noise ratio into the dequantized Gaussian approximate message passing detector for detection;

[0009] At the receiving end, the minimum mean square error weight matrix W is calculated through the delay-Doppler domain channel matrix H MMSE , the calculation formula is:

[0010]

[0011] In the formula, is the noise variance, (·) H is the conjugate transpose operation, and I is the identity matrix with the same dimension as the channel matrix H.

[0012] Furthermore, the signal-to-interference-to-noise ratio (SINR) of each signal i The calculation formula is:

[0013]

[0014] in,

[0015] In the formula, w i W MMSE The i-th column, h i is the i-th column of the channel matrix, E x is the signal mean, |·| is the absolute value operation, and ||·|| is the binary norm operation.

[0016] S4, using a serial interference cancellation method to eliminate interference according to the detection result, and updating the channel matrix;

[0017] S5, repeat S3-S4, and detect the low-precision quantized received data of each channel in a traversal manner;

[0018] S6. Output the detected data symbols and perform symbol demodulation.

[0019] As a specific implementation of the present invention, the processing of the original data in step S1 includes: at the transmitting end, placing the symbols mapped into the bit stream in the delay-Doppler domain grid to obtain the delay-Doppler data x{k, 1}, and then modulating the delay-Doppler data x{k, 1} by OTFS to obtain the time domain transmission signal s(t), and constructing the channel matrix through the known channel state information.

[0020] As a specific implementation of the present invention, step S2 includes: at the receiving end, the received time domain receiving signal r(t) is quantized by low precision to obtain a low precision quantized receiving signal, and then the low precision quantized receiving signal is demodulated by OTFS to obtain the delay-Doppler domain to obtain the delay-Doppler domain receiving data Y{k, 1}.

[0021] As a specific implementation of the present invention, step S3 includes:

[0022] In the present invention, the low-precision quantization process is expressed as:

[0023] q (r) =Q(r (r) )

[0024] Q(·) is a low-precision quantizer. When the quantization accuracy is B-bit, it maps the real value and imaginary value of the received signal into 2 B One of the discrete output values, assuming the quantization step is Δ, when the real value of the input falls within the interval When within the range, the output is r (r) represents the time domain received signal of the rth receiving antenna; q (r) Represents the quantized result of the time domain received signal of the rth receiving antenna.

[0025] Since the quantization process is a nonlinear process, the analysis is difficult, which will lead to high algorithm complexity. Therefore, as a specific implementation of the present invention, the low-precision ADC quantization process is modeled by an additive quantization noise model (AQNM, Additive Quantization Noise Model), the impact of the quantization error is linearly analyzed, and the quantization error is modeled in the form of covariance to eliminate the quantization error in the dequantization message transmission, thereby improving the detection performance.

[0026] The beneficial effects of the present invention are as follows: the present invention overcomes the problems of difficult signal detection and high bit error rate in a low-precision quantized MIMO-OTFS system, and can achieve better bit error performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flowchart of a low-precision quantized MIMO-OTFS signal detection method based on interference elimination in an exemplary embodiment of the present invention.

[0028] Figure 2 It is a bit error rate curve diagram of a low-precision quantized MIMO-OTFS signal detection method based on interference elimination in an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0030] Example 1

[0031] The MIMO-OTFS system provided by the present invention mainly consists of N t single-antenna users and N r The number of users is N. t =4, the number of receiving antennas N r =4, the carrier frequency is 1.5GHz, the subcarrier spacing is 15KHz, the maximum user speed is 500km / h, the number of subcarriers in each OTFS data frame is M=32, the number of time slots is N=32, the number of multipaths between a single transmitting antenna and a single receiving antenna is 4, the symbol modulation method is QPSK, the channel model is Rayleigh channel, it is assumed that the channel information is known, and the quantization accuracy of the analog-to-digital converter is 3 bits.

[0032] A low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation, the main steps of which are as follows: Figure 1 As shown, the following steps are included:

[0033] S1. Process the original data to obtain a transmission signal and transmit it;

[0034] S2, performing low-precision quantization and demodulation on the received signal in sequence to obtain low-precision quantized received data;

[0035] S3, using the channel matrix to calculate the signal to noise ratio, and inputting the low-precision quantized received data with the largest signal to noise ratio into the dequantized Gaussian approximate message passing detector for detection;

[0036] S4, using the serial interference cancellation method to eliminate interference according to the detection result, and updating the channel matrix;

[0037] S5, repeat S3-S4, and detect the low-precision quantized received data of each channel in a traversal manner;

[0038] S6. Output the detected data symbols and perform symbol demodulation.

[0039] The processing of the original data in step S1 includes: at the transmitting end, placing the symbols mapped into the bit stream in the delay-Doppler domain grid to obtain the delay-Doppler data x{k,l}, and then modulating the delay-Doppler data x{k,l} by OTFS to obtain the time domain transmission signal s(t). In the specific modulation, the OTFS modulation at the transmitting end includes: transforming the delay-Doppler data x{k,l} to the time-frequency domain by symplectic finite-inverse Fourier transform to obtain the transmitting end time-frequency domain data X[n,m], and then using Heisenberg Transform (HT) to transform the time-frequency domain signal to the time domain to obtain the time domain transmission signal s(t).

[0040] According to the symplectic finite-inverse Fourier transform, we have

[0041]

[0042] Step S2 includes: at the receiving end, the received time domain received signal r(t) is quantized by a low-precision ADC to obtain a low-precision quantized received signal, and then the low-precision quantized received signal is demodulated by OTFS to obtain a delay-Doppler domain to obtain delay-Doppler domain received data Y{k,l}, and the delay-Doppler domain received data Y{k,l} is obtained according to the delay-Doppler domain received data Y{k,l}.

[0043] Step S3 includes: calculating the minimum mean square error weight matrix according to the delay-Doppler domain channel matrix H; calculating the signal to interference noise ratio SINR of each signal according to the minimum mean square error weight matrix; sorting the signals according to the signal to interference noise ratio SINR value from large to small, performing dequantized Gaussian approximate message passing (DQ-GAMP) detection from large to small, and eliminating interference according to the detected signal after each detection. The OTFS demodulation at the receiving end includes: demodulating the quantized signal by OTFS to obtain the delay-Doppler domain to obtain the delay-Doppler domain received data Y{k,l}, including restoring the received time domain received signal r(t) to the time-frequency domain through Wigner Transform (WT) to obtain the time-frequency domain received data Y[m,n], and then restoring it to the delay-Doppler domain through Symplectic Finite Fourier Transform (SFFT) to obtain the delay-Doppler domain received data Y{k,l}.

[0044] In this embodiment, the quantization process of the low-precision ADC is to convert the complex nonlinear quantization process into a linear model that is easy to analyze through an additive quantization noise model (AQNM). The model formula is as follows:

[0045] q=αGP 1 / 2 r+αz+e=A+αz+e

[0046] α is the distortion factor related to the number of quantization bits, r is the time domain received signal, z is the time domain additive white Gaussian noise, G is the time domain channel, P is the row and column interleaving matrix, e is the quantization error, and q is the quantized output signal.

[0047] At the receiving end, the minimum mean square error weight matrix is ​​calculated through the delay-Doppler domain channel matrix H. The calculation formula is:

[0048]

[0049] In the formula, is the noise variance, (·) H is the conjugate transpose operation, where I is the identity matrix of the same dimension as H.

[0050] SINR of each signal i The calculation formula is:

[0051]

[0052] in,

[0053] In the formula, w i W MMSE The i-th column, h i is the i-th column of the channel matrix, E x is the signal mean, |·| is the absolute value operation, and ||·|| is the binary norm operation.

[0054] In this embodiment, after each detection is completed, interference is eliminated according to the detected signal, that is, in, is the estimated value of the received data in the delay-Doppler domain of the i-th channel output by DQ-GAMP detection, y is obtained by OTFS demodulation of the received time-domain quantized signal q; H i is the channel matrix between the i-th receiving antenna and the i-th transmitting antenna.

[0055] The dequantized Gaussian approximation message passing detection includes:

[0056] S1, input parameters: delay-Doppler domain received signal y, time domain channel matrix G, algorithm iteration number T, damping factor β, noise variance σ 2 ;

[0057] S2. Parameter initialization:

[0058] Initialize estimates for the data, Initialize the variance for the data, is the initial transfer factor, is the prior mean of A, is the prior variance of A;

[0059] λ=diag(Λ)

[0060] diag(·) is the operation of taking diagonal elements

[0061]

[0062] H is the delay-Doppler domain channel, and x is the delay-Doppler domain signal;

[0063] S3. Calculate the posterior mean of A The R superscript and the I superscript indicate the real part and the imaginary part;

[0064] S4. Calculate the posterior variance of A The mean and variance are calculated using the following formula. The real and imaginary parts are calculated in the same way. The real part is used as an example below, where the R superscript represents the real part and the I superscript represents the imaginary part.

[0065]

[0066]

[0067] in and The upper and lower quantization thresholds for the quantizer.

[0068] S5. Calculate the average posterior variance

[0069] S6. Calculate the external variance of B

[0070] S7, start iteration: for t=1:T

[0071] Update the factor node variance

[0072] Calculate the first intermediate variable

[0073] Calculate the second intermediate variable

[0074] For w i Make an update

[0075] Update the second intermediate variable

[0076] Update the first intermediate variable

[0077] Calculate data estimates

[0078] Update the variance of the variable node

[0079] in

[0080]

[0081] When the transmitted signal is K-order QAM modulation, there are:

[0082]

[0083] The iteration ends when the maximum number of iterations T is reached.

[0084] The above are only preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the embodiments of the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation, characterized in that: The steps include: S1. Process the original data to obtain a transmission signal and transmit it; S2, performing low-precision quantization and demodulation on the received signal in sequence to obtain low-precision quantized received data; S3, calculating the minimum mean square error weight matrix according to the delay-Doppler domain channel matrix; calculating the signal to interference plus noise ratio SINR of each signal according to the minimum mean square error weight matrix, and inputting the low-precision quantized received data with the largest signal to interference plus noise ratio into the dequantized Gaussian approximate message passing detector for detection; The calculation formula of the minimum mean square error weight matrix is ​​as follows: W MMSE =(H H H+δ(Z) 2 I) -1 H H Where W MMSE is the minimum mean square error weight matrix; δ(Z) 2 is the noise variance, H is the channel matrix; (·) H is the conjugate transpose operation, I is the identity matrix with the same dimension as the channel matrix H; SINR of each signal i The calculation formula is: SINR i =E x |w i,MMSE h i |÷(E x ∑ l≠i |w i,MMSE h i |+δ(Z) 2 ∥w i,MMSE h i ∥ 2 ) Among them, w i,MMSE = argmax(|wh i | 2 E x ÷(E x ∑ j≠i |wh j | 2 +δ(Z) 2 ∥w∥ 2 )),j∈(1,2,…,N t ),w∈(w1,w2,…,w Nt ) where w i W MMSE The i-th column, h i is the i-th column of the channel matrix, E x is the signal mean, |·| is the absolute value operation, ∥·∥ is the binary norm operation; S4, using the serial interference cancellation method to eliminate interference according to the detection result, and updating the channel matrix; S5, repeat S3-S4, and detect the low-precision quantized received data of each channel in a traversal manner; S6. Output the detected data symbols and perform symbol demodulation.

2. A low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation according to claim 1, characterized in that: The processing of the original data in step S1 includes: at the transmitting end, placing the symbols mapped into the bit stream in the delay-Doppler domain grid to obtain delay-Doppler data, then modulating the delay-Doppler data by OTFS to obtain a time domain transmission signal, and constructing a channel matrix through known channel state information.

3. The low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation according to claim 1, characterized in that: Step S2 includes: at the receiving end, the received time domain reception signal is quantized by low precision to obtain a low precision quantized reception signal, and then the low precision quantized reception signal is demodulated by OTFS to obtain a delay-Doppler domain to obtain delay-Doppler domain reception data.

4. A low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation according to claim 1, characterized in that: In step S2, the low-precision quantization process is modeled by an additive quantization noise model, the influence of the quantization error is linearly analyzed, and the quantization error is modeled in the form of covariance to eliminate the quantization error in the dequantization message transmission and improve the detection performance.

5. A low-precision quantized MIMO-OTFS signal detection method based on serial interference cancellation according to claim 1, characterized in that: The detection of the quantized Gaussian approximate message passing detector includes: S31, input parameters: delay-Doppler domain received signal y, time domain channel matrix G, algorithm iteration number T, damping factor β, noise variance σ 2 ; S32, parameter initialization: S33, obtaining indicative quantization error information and external variance by calculating the posterior mean and posterior variance of the quantized signal and performing operations with the prior information; S34, treating the external variance as noise and adding it into the Gaussian approximate message passing process to eliminate the interference of the quantization error; S35, performing iterative detection on the quantized signal, transmitting the quantization error, noise and other symbol interference in the factor graph, stopping the iteration after the maximum number of iterations is met, and outputting the detection result.

Images