Pre-coding method of multi-input multi-output constant envelope orthogonal frequency division multiplexing system
Through analog precoding methods and perfect channel estimation, a fully connected analog precoding method is designed to solve the problems of PAPR and nonlinear distortion in orthogonal frequency division multiplexing system, and efficient signal transmission and accurate data detection are achieved.
Patent Information
- Application Number
- CN202411084997.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-12-06
AI Technical Summary
In orthogonal frequency division multiplexing systems, precoding technology may cause PAPR to increase, resulting in nonlinear distortion and power efficiency decreases, especially under high-frequency amplifier performance, and digital precoding destroys the constant envelope characteristics.
Using the analog precoding method, the frequency domain response information is obtained through perfect channel estimation, and a fully connected analog precoding method is designed. Based on the minimized sum of the mean square errors of the subcarrier signal and the sent signal, the analog precoding matrix is iteratively calculated to realize the The signal is simulated precoding and demodulated using an NML receiver at the receiver.
It achieves the performance gains exceeding digital precoding, improves the efficiency and accuracy of signal transmission, and avoids the occurrence of PAPR problems and nonlinear distortion.
Smart Images

Figure CN119093980A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of communication technology, and in particular relates to a precoding method for a multiple-input multiple-output constant envelope orthogonal frequency division multiplexing system. Background Art
[0002] The orthogonal frequency division multiplexing system has high spectrum utilization but has a serious peak-to-average problem. Precoding technology uses channel state information at the transmitter to adjust the phase and amplitude of the signal to adjust the antenna radiation direction and shape in wireless transmission, thereby enhancing the signal transmission effect. However, precoding may introduce an increase in PAPR (Peak to Average Power Ratio). When the amplifier processes the signal, it may cause nonlinear distortion and power saturation, resulting in a decrease in power efficiency. In particular, the performance of high-frequency amplifiers will be seriously affected.
[0003] In order to solve the PAPR problem caused by the OFDM system, people have proposed a constant envelope OFDM waveform using nonlinear phase modulation, which can avoid the problem of large PAPR caused by multi-carrier waveforms, but digital precoding and hybrid precoding will destroy the constant envelope characteristics. Summary of the invention
[0004] In view of the above problems, the present invention provides an analog precoding method that can obtain performance gains that exceed digital precoding. The method of the present invention first obtains all frequency domain response information of the system through perfect channel estimation, and then designs a fully connected analog precoding method based on minimizing the sum of the mean square errors of all subcarrier received signals and transmitted signals. When the receiving end uses an NML receiver, its BER performance is better than the digital precoding of the MIMO system.
[0005] The technical solution of the present invention is:
[0006] A precoding method for a MIMO constant envelope orthogonal frequency division multiplexing system is defined in a single-user CE-OFDM system, where the number of transmitting antennas, receiving antennas, and data streams are N and t 、N r 、N s , the modulation symbol adopts M-order modulation, the subcarrier is N, and the number of effective subcarriers is Oversampling multiple Q, total number of carriers N FFT =N×Q, the number of RF chains at the transmitting end in full connection mode is N t RF =N s ; Characterized in that the method comprises the following steps:
[0007] Sending side:
[0008] A1. Generate information bit stream: Calculate N sThe total information bits of the stream data under M-order modulation are N s ×N QAM ×log2M, in order to obtain an OFDM signal that is all real numbers, let the data before IFFT be Where X ns for:
[0009]
[0010] Where, ns = 1, 2…N s , X ns (k), k=1,2,...,N QAM Represents the constellation data transmitted on the corresponding subcarrier after modulation, N zp The number of zero padding depends on the oversampling factor M. Two zeros are necessary to make all the signals real numbers after IFFT.
[0011] A2, phase modulation: X to N FFT Point IFFT transform:
[0012]
[0013] Then the obtained MIMO-OFDM real signal x t (n) Perform phase modulation to obtain the MIMO-CE-OFDM time domain signal ρ 2 Represents the transmission power of each stream data, and ρ is A is the carrier signal amplitude, Expressed as 2πh represents the phase modulation index, and C is the power normalization factor, which is σ I 2 It represents the variance when the M-QAM modulation symbols obey independent and identical distribution, and its value is
[0014] The time domain signal s t [n] Perform FFT transformation to obtain frequency domain signal
[0015] A3. Set the total number of iterations iter, initialize the number of iterations i = 1, and initialize the simulated precoding matrix Frequency domain response estimated from perfect channel Iterate the calculation of the simulated precoding matrix and obtain the iter-th simulated precoding result as A iter , the calculation process of the i-th iteration is:
[0016] A31, according to the simulated precoding matrix A i Get the equivalent channel Heq :
[0017]
[0018] The equivalent channel H of the kth subcarrier eq,k Perform MMSE equalization to obtain the matrix W k ', take the first N s The balanced matrix
[0019]
[0020] Where I represents the identity matrix, σ n 2 is the noise variance, is the channel transmission matrix H k The conjugate transpose of , and then find the equivalent matrix of the receiving end Among them, H k represents the original channel of the kth subcarrier;
[0021] A32. Calculate the precoded frequency domain signal S i :
[0022]
[0023] Calculate the power allocation p for subcarrier k k :
[0024]
[0025] in Represents S i The element in the mth row and kth column, and then normalize the power to get the power allocation factor P k :
[0026]
[0027] A33. Calculate the intermediate matrix
[0028] A34. Find the N before Temp s The eigenvectors corresponding to the eigenvalues are calculated, and the isometric matrix containing these eigenvectors is calculated, which is the unnormalized analog precoding matrix A. Then, the normalized analog precoding matrix A of the i-th cycle is obtained by energy normalization. i ,Right now:
[0029]
[0030] A4. Calculate the time domain signal A after analog precoding iter s t , sent through the antenna;
[0031] Receiver:
[0032] The frequency domain signal of the kth subcarrier at the receiving end is:
[0033] Y k =H k A iter S k +Z k
[0034]
[0035] in is the frequency domain received signal of the kth subcarrier and the rth receiving antenna, A represents the channel response matrix between the kth subcarrier receiving antenna r and the transmitting antenna t, iter is the simulated precoding matrix after iter iterations, S k is the frequency domain transmission signal of the kth subcarrier after phase modulation, The mean is 0 and the variance is Gaussian white noise vector; the demodulation process of the receiving end using the NML receiver includes:
[0036] B1. Perform FFT operation on the time domain signal received by the base station, and obtain preliminary detection results after equalization and phase resolution based on the equivalent channel of the receiving end. in Represents the preliminary detection result of the i-th subcarrier, i = 1, 2...N FFT :
[0037]
[0038] B2. All N s The possible constellation point combination of the stream data replaces the data of the first subcarrier carrying information, and the data of the remaining subcarriers remain unchanged, and we get Where Z is N s The set of all possible constellation data of stream data is kind; will To the time domain, perform phase modulation, and then FFT to the frequency domain to obtain the frequency domain signal
[0039] The transmission data on the current subcarrier is determined by the minimum Euclidean distance:
[0040]
[0041] Where H' is the equivalent channel frequency domain impulse response after analog precoding H' = HA iterThe first N / 2 data, Y′ is the first N / 2 data of Y after the received time domain symbol y is FFT;
[0042] B3, replace the data of the next subcarrier carrying information with all possible constellation point combinations, and let the data of the remaining subcarriers be the results obtained in B2, and we have:
[0043]
[0044] Then the result on the current subcarrier is obtained by the same operation as in B2.
[0045] B4, repeat B3 operation until all subcarriers containing data are re-detected, and finally the detected constellation data is The final detection result output after demodulation is the detection result of approximate maximum likelihood detection under multi-stream conditions.
[0046] The beneficial effect of the present invention is that it can obtain a performance gain exceeding digital precoding. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] Figure 1 The present invention is a flowchart of iteratively calculating a simulated precoding matrix in the method of the present invention. DETAILED DESCRIPTION
[0048] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and embodiments:
[0049] Example:
[0050] This example is a single-user CE-OFDM system. The number of transmitting antennas, receiving antennas, and data streams are 4, 4, and 2 respectively. The modulation symbol adopts QPSK modulation. The number of subcarriers is 8, and the number of effective subcarriers is N. QAM =3, oversampling factor 1, total number of carriers N FFT =8, the number of RF chains at the transmitting end in full connection mode is N t RF =N s =2.
[0051] Step 1. Generate information bit stream: Calculate the total information bit of the two-stream data under QPSK modulation to be 12 bits. Assume that the transmitted information is [0 0 0 1 0 1 0 0 0 0 1 0], and the data after constellation modulation is
[0052]
[0053] Step 2: Phase modulation: Take A = 1, 2πh = 0.7, and get the time domain signal s t [n].
[0054] Step 3: Estimating the frequency domain response based on the perfect channel Perform the calculation of the simulated precoding matrix for the i-th time (i=1, 2, 3, 4, 5). Initialize A 0 =I 4×4 ,like Figure 1 As shown, including:
[0055] Step 31: simulate the precoding matrix A i Get the equivalent channel H eq , calculated as follows:
[0056]
[0057] The equivalent channel H of the kth subcarrier eq,k Perform MMSE equalization to obtain the matrix W k ', take the first N s The balanced matrix
[0058]
[0059] Then find the equivalent matrix of the receiving end Among them, H k Represents the original channel of the kth subcarrier.
[0060] Step 32: Calculate the subcarrier power allocation ratio and calculate the precoded frequency domain signal S i
[0061]
[0062] Calculate the power allocation p for subcarrier k k
[0063]
[0064] in Represents S i The element in the mth row and kth column, and then normalize the power to get the power allocation factor P k :
[0065]
[0066] Step 33: Calculate the intermediate matrix
[0067] Step 34, find the eigenvectors corresponding to the first two eigenvalues of Temp, calculate the isometric matrix containing these eigenvectors, which is the unnormalized analog precoding matrix A, and then obtain the normalized analog precoding matrix A of the i-th cycle through energy normalization. i ,Right now
[0068]
[0069] i=i+1, repeat steps A31-A34 until i=5, and obtain the fifth simulation precoding result A iter ;
[0070] Step 4: Calculate the time domain signal A after analog precoding iter s t , sent through the antenna;
[0071] Step 5: Perform FFT operation on the time domain signal received by the base station, and obtain preliminary detection results after equalization and phase solution according to the equivalent channel of the receiving end;
[0072]
[0073] Step 6: Extract the data on the subcarrier that carries the information
[0074]
[0075] The phase detection result after demodulation is [0 0 0 0 0 1 0 1 0 0 0 0 1 0], and the corresponding constellation data is
[0076] Then replace all the constellation points of 2-stream QPSK (there are 16 combinations) with the first column of data in the detection result to obtain Where A represents all constellation points, and the Euclidean distances in formula (1) are calculated, which are [86 51 113 80 114 71 136 93 100 64 119 86 122 77 137 93], so the information on the subcarrier is judged as [0 0 0 1], and then the detection result is updated to [0 0 0 1 0 1 0 1 0 0 0 0 1 0], and then the remaining data are detected similarly to obtain the final detection result of [0 0 0 1 0 1 0 0 0 0 1 0].
Claims
1. A precoding method for a MIMO constant envelope orthogonal frequency division multiplexing system is defined in a single-user CE-OFDM system, where the number of transmitting antennas, receiving antennas, and data streams are N respectively. t 、N r 、N s , the modulation symbol adopts M-order modulation, the subcarrier is N, and the number of effective subcarriers is Oversampling multiple Q, total number of carriers N FFT =N×Q, the number of RF chains at the transmitting end in full connection mode is It is characterized in that The method comprises the following steps: Sending side: A1. Calculate N s The total information bits of the stream data under M-order modulation are N s ×N QAM ×log 2 M, in order to obtain an OFDM signal that is all real numbers, let the data before IFFT be Where X ns for: Where, ns = 1, 2…N s , X ns (k), k=1,2,...,N QAM Represents the constellation data transmitted on the corresponding subcarrier after modulation, N zp The number of zero padding depends on the oversampling factor M. Two zeros are necessary to make all the signals real numbers after IFFT. A2. Perform N on X FFT Point IFFT transform: Then the real signal x is obtained t (n) Perform phase modulation to obtain a time domain signal ρ 2 Represents the transmission power of each stream data, and ρ is A is the carrier signal amplitude, Expressed as 2πh represents the phase modulation index, and C is the power normalization factor, which is σ I 2 It represents the variance when the M-QAM modulation symbols obey independent and identical distribution, and its value is The time domain signal s t [n] Perform FFT transformation to obtain frequency domain signal A3. Set the total number of iterations iter, initialize the number of iterations i = 1, and initialize the simulated precoding matrix Frequency domain response based on perfect channel estimation Iterate the calculation of the simulated precoding matrix and obtain the iter-th simulated precoding result as A iter , the calculation process of the i-th iteration is: A31, according to the simulated precoding matrix A i Get the equivalent channel H eq : The equivalent channel H of the kth subcarrier eq,k Perform MMSE equalization to obtain the matrix W k ', take the first N s The balanced matrix Where I represents the identity matrix, σ n 2 is the noise variance, is the channel transmission matrix H k The conjugate transpose of , and then find the equivalent matrix of the receiving end Among them, H k represents the original channel of the kth subcarrier; A32. Calculate the precoded frequency domain signal S i : Calculate the power allocation p for subcarrier k k : in Represents S i The element in the mth row and kth column, and then normalize the power to get the power allocation factor P k : A33. Calculate the intermediate matrix A34. Find the N before Temp s The eigenvectors corresponding to the eigenvalues are calculated, and the isometric matrix containing these eigenvectors is calculated, which is the unnormalized analog precoding matrix A. Then, the normalized analog precoding matrix A of the i-th cycle is obtained by energy normalization. i ,Right now: A4. Calculate the time domain signal A after analog precoding iter s t , sent through the antenna; Receiver: The frequency domain signal of the kth subcarrier at the receiving end is: Y k =H k A iter S k +Z k in is the frequency domain received signal of the kth subcarrier and the rth receiving antenna, A represents the channel response matrix between the kth subcarrier receiving antenna r and the transmitting antenna t, iter is the simulated precoding matrix after iter iterations, S k is the frequency domain transmission signal of the kth subcarrier after phase modulation, The mean is 0 and the variance is Gaussian white noise vector; the demodulation process of the receiving end using the NML receiver includes: B1. Perform FFT operation on the time domain signal received by the base station, and obtain preliminary detection results after equalization and phase resolution based on the equivalent channel of the receiving end. in Represents the preliminary detection result of the i-th subcarrier, i = 1, 2...N FFT : B2. All N s The possible constellation point combination of the stream data replaces the data of the first subcarrier carrying information, and the data of the remaining subcarriers remain unchanged, and we get Where Z is N s The set of all possible constellation data of stream data is kind; will IFFT to time domain, perform phase modulation, and then FFT to frequency domain to obtain frequency domain signal The transmission data on the current subcarrier is determined by the minimum Euclidean distance: Where H' is the equivalent channel frequency domain impulse response after analog precoding H' = HA iter The first N / 2 data, Y′ is the first N / 2 data of Y after the received time domain symbol y is FFT; B3, replace the data of the next subcarrier carrying information with all possible constellation point combinations, and let the data of the remaining subcarriers be the results obtained in B2, and we have: Then the result on the current subcarrier is obtained by the same operation as in B2. B4, repeat B3 operation until all subcarriers containing data are re-detected, and finally the detected constellation data is The final detection result is output after demodulation.
