Large frequency offset estimation method based on time-frequency domain combination under OFDM (Orthogonal Frequency Division Multiplexing) system

Abstract

The invention relates to the technical field of wireless communication, and provides a large frequency offset estimation method based on time-frequency domain combination under an OFDM (Orthogonal Frequency Division Multiplexing) system, which comprises the following steps of: performing IFFT (Inverse Fast Fourier Transform) on an original frequency domain signal under an OFDM symbol where a DMRS (Demodulation Reference Signal) is located to a time domain to obtain a first time domain signal; performing inter-symbol signal conjugate correlation processing on the first time domain signal to obtain a correlation value; performing Alpha filtering on the correlation value to obtain a filtered correlation value; performing phase calculation and conversion processing on the filtered correlation value to obtain a coarse frequency offset value; and finally, performing frequency offset estimation under a traditional frequency domain through FFT transformation to the frequency domain to obtain a fine frequency offset value. The method has the beneficial effects that the frequency offset in a subcarrier range is estimated by using the phase difference of the time domain signal under the OFDM symbol where the DMRS is located between the same symbols, and the estimation range is obviously far larger than that of the existing frequency offset estimation scheme.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular, to a large frequency offset estimation method combining time and frequency domains based on an OFDM (Orthogonal Frequency Division Multiplexing) system. Background Art

[0002] With the gradual popularization of 5G, the application fields of wireless mobile communication are becoming more and more extensive. The 5G mobile communication system is based on OFDM technology. OFDM is a multi-carrier data transmission technology. Its basic principle is to convert the originally high-speed serial transmission data to multiple parallel low-speed channels for transmission. By using the orthogonality between carriers, not only can the spectrum utilization rate of the system be significantly improved, but also the influence of the frequency-selective fading channel on the signal can be reduced, and the anti-interference ability of the system can be improved. However, when there is a Doppler frequency offset in the system, it will cause a frequency deviation between carriers and generate inter-carrier interference. With the continuous enhancement of the country's comprehensive strength and the continuous improvement of social infrastructure construction, people have an urgent need for communication problems in high-speed mobile scenarios such as high-speed rails, civil airliners, and fighter jets. In the above high-speed mobile scenarios of the OFDM system, there will be a more serious Doppler effect between the transmitter and the receiver, generating a larger Doppler frequency offset, seriously damaging the normal communication ability of the system. The current maximum maneuvering speed of the high-speed rail mode supported by the 5G-based OFDM system is 500 km / h, and it cannot meet the communication requirements of civil airliners at 1000 km / h and fighter jets at Mach level and other higher speeds. Therefore, large frequency offset estimation and compensation technologies are technical problems that need to be solved when the current OFDM system is applied to ultra-high-speed mobile communication scenarios.

[0003] In a wireless communication system, the relative movement between the transmitter and the receiver will cause the Doppler effect. The calculation formula of the Doppler frequency offset f d is as follows,

[0004] f d = vf c ·cosθ / c

[0005] In the formula, v refers to the relative speed between the transmitter and the receiver in the communication system, with the unit of m / s. f c refers to the carrier frequency, with the unit of Hz. c refers to the speed of light, usually taking 3×10 8 m / s. θ refers to the angle between the signal arrival direction and the receiver movement direction.

[0006] Common frequency offset estimation schemes in OFDM systems include time-domain frequency offset estimation schemes based on CP (Cyclic Prefix) and frequency-domain frequency offset estimation schemes based on DMRS (Demodulation reference signal) pilots.

[0007] Among them, the frequency offset estimation scheme based on CP: CP refers to copying the tail data of the OFDM symbol to the starting position of the symbol to achieve cyclic extension of the OFDM symbol. Its main function is to eliminate inter-symbol interference and inter-carrier interference. The frequency offset estimation scheme based on CP is a frequency offset estimation scheme carried out in the time domain. It mainly calculates the conjugate multiplication of the two segments of data, namely the tail data and the head data copied to the starting position, so as to calculate the phase difference between adjacent points. Finally, it is converted into the estimated frequency offset value through parameters such as the number of sampling points and the FFT length. Taking the receiving-end processing of the 5G physical uplink shared channel (PUSCH) as an example, the specific calculation steps of the frequency offset estimation scheme are as follows:

[0008] (1) CP data conjugate correlation:

[0009] Assume that the received data in the time domain at the receiving end is x(n), then the formula for CP data conjugate correlation is:

[0010]

[0011] In the formula, represents the number of symbols included in a slot, and the default value is 14; represents the starting position of the data of symbol l in the time-domain signal, N CP,l represents the time-domain CP length of symbol l, N FFT represents the length of the FFT, () * represents the conjugate operation.

[0012] (2) Frequency offset calculation:

[0013]

[0014] In the formula, Δf represents the subcarrier spacing, and phase() represents the phase operation.

[0015] It can be seen from here that the range of phase calculation of phase(R cpCorr ) is [-π, π], so the estimation range of Fo est is That is, the maximum estimated frequency offset value is half of the subcarrier.

[0016] (3) Frequency offset compensation for time-domain signals

[0017]

[0018] In the formula, N CP represents the length of CP, and x foComp (n) is the time-domain signal after frequency offset compensation and can be output to subsequent processing such as CP removal for subsequent calculations.

[0019] The frequency offset estimation scheme based on DMRS pilots is a frequency offset estimation scheme carried out in the frequency domain. It mainly inserts DMRS pilot signals between normal data symbols, performs channel estimation based on the pilot signals, and uses the differential correlation between different symbols of the estimated channel to calculate the phase difference between adjacent points. Then, through parameters such as the number of sampling points and the length of FFT (Fast Fourier Transform), it is converted into the estimated frequency offset value. Taking the receiving-end processing of the 5G data sharing channel (PUSCH) as an example, the specific calculation steps of the frequency offset estimation scheme are as follows:

[0020] (1) LS (Least Squares Estimation) channel estimation

[0021] Assume that the frequency-domain signal at the receiving end is X l (n), and the formula for LS channel estimation is as follows:

[0022]

[0023] In the formula, represents the DMRS base sequence, represents the frequency-domain data of the symbol where DMRS is located, and l DMRS represents the OFDM symbol where DMRS is located, and N ScNum represents the number of frequency-domain subcarriers of DMRS. The result output by the formula is the channel of LS channel estimation

[0024] (2) Correlation between OFDM symbols where different DMRSs are located

[0025]

[0026] In the formula, R corr represents the sum of the conjugate correlation values of different DMRS symbols in the frequency domain. Taking the phase of it can obtain the phase difference represented by the frequency offset. L dmrsSet represents the set of OFDM symbols where DMRS is located, and N DMRS represents the number of OFDM symbols occupied by DMRS, and () * represents the conjugate operation..

[0027] (3) Alpha Filtering

[0028] R corrFilter = R corr · (1 - α filter ) + R corrHis · α filter

[0029] In the formula, α filrer represents the Alpha filtering factor, and the default value is selected as 0.7. R corrHis represents the historical value of the fine frequency offset estimation in the frequency domain of the previous TTI (Transmission Time Interval), and R corrFilter is the correlation value output after filtering.

[0030] (4) Frequency Offset Calculation

[0031]

[0032] In the formula, Δf represents the subcarrier spacing, and l Δt represents the difference in the time domain of different DMRS symbols. When there are more than 2 DMRS symbols, that is, there are multiple l Δt , the calculation of the variable l Δt can be placed in the calculation formula of R corr corresponding to multiple symbols. N FFT represents the length of the FFT, and N CP represents the length of the CP.

[0033] In the above formula, the value range of phase(R corrFilter ) is [-π, π], then the value range of Fo est is Since in the OFDM system, the length of N FFT is generally much larger than the length of N CP , so can be approximately considered equal to 1. Finally, the estimation range of Fo est is Its range is mainly limited by the subcarrier spacing and the difference l Δt in the time domain of the DMRS symbol. In the data sharing channel of 5G, the minimum value of l Δt is 3, then the maximum range of frequency offset estimation is It can be seen that the maximum frequency offset estimation value in this scheme is 1 / 3 of the frequency offset estimation scheme based on the CP.

[0034] (5) Frequency Offset Compensation

[0035] When interpolation is performed in channel estimation, inter-symbol frequency offset compensation is performed for different OFDM symbols.

[0036]

[0037] l minDmrs = argmin{|l - L dmrsset |}

[0038] l diff = l minDmrs - l

[0039] where y LsInterp (l, n) represents the channel estimate value of the interpolated data symbol, l minDmrs represents the position of the DMRS symbol with the smallest difference in the current OFDM symbol l from the DMRS symbol set, and l diff represents the difference between the current symbol l and the symbol where the DMRS is located. When l diff is 0, it means the current symbol is the symbol where the DMRS is located.

[0040] The CP-based frequency offset estimation scheme performs frequency offset estimation in the time domain and processes the time-domain data at the time-slot level at the receiving end. Therefore, in the scenario of multiple users, the Doppler frequency offset between different users and the transmitting end cannot be estimated; moreover, the maximum estimation range of this scheme is half of the subcarrier frequency value, that is and since the CP itself has the function of resisting multipath channels, in a multipath scenario, its frequency offset estimation effect will be severely damaged by the influence of delay.

[0041] The DMRS pilot-based frequency offset estimation scheme performs frequency offset estimation in the frequency domain and belongs to the user-level frequency offset estimation scheme. Its estimation performance is relatively accurate, but its frequency offset estimation range depends on the position difference of the DMRS pilot symbols in the time domain, and the maximum frequency offset estimation range is The estimation range is small and cannot meet the frequency offset estimation ability under high-speed movement. SUMMARY OF THE INVENTION

[0042] The present invention aims to provide a large frequency offset estimation method based on the time-frequency domain joint in an OFDM system to solve the problems that the traditional frequency offset estimation scheme has inaccurate user-level frequency offset estimation or relatively small estimation range in the ultra-high-speed movement scenario of the OFDM system and cannot cope with the Doppler effect of large frequency offsets.

[0043] A large frequency offset estimation method based on the time-frequency domain joint in an OFDM system provided by the present invention includes:

[0044] Performing an IFFT transform on the original frequency-domain signal under the OFDM symbol where the DMRS is located to the time domain to obtain a first time-domain signal;

[0045] Perform conjugate correlation processing on the inter-symbol signals of the first time-domain signal to obtain correlation values;

[0046] Perform Alpha filtering on the correlation values to obtain filtered correlation values;

[0047] Perform phase calculation and conversion processing on the filtered correlation values to obtain the coarse frequency offset value;

[0048] Compare the coarse frequency offset value with the estimation threshold to adaptively determine whether to perform coarse frequency offset compensation;

[0049] According to the adaptive decision result, perform fine frequency offset estimation on the frequency-domain signals under all OFDM symbols to obtain the fine frequency offset estimation value. In the present invention, performing IFFT transformation on the original frequency-domain signal under the OFDM symbol where the DMRS is located to the time domain to obtain the first time-domain signal includes:

[0050] x(l dmrs ,n) = ifft(X(l dmrs ,n))

[0051] where x(l dmrs ,n) is the first time-domain signal, l dmrs is the symbol position occupied by the DMRS, is the number of symbols included in one slot, is the number of frequency-domain subcarriers occupied by the PUSCH, ifft() is the IFFT transformation, and X(l dmrs ,n) is the original frequency-domain signal under the OFDM symbol where the DMRS is located.

[0052] In the present invention, performing conjugate correlation processing on the inter-symbol signals of the first time-domain signal to obtain correlation values includes:

[0053]

[0054] where R tdCorr is the correlation value, N dmrs is the number of OFDM symbols occupied by the DMRS, () * represents the conjugate operation.

[0055] In the present invention, performing Alpha filtering on the correlation values to obtain filtered correlation values includes:

[0056] R tdCorrFilter = R tdCorr ·(1 - α filter ) + R tdCorrHis ·α filter ;

[0057] where R tdCorrFilterIs the correlated value after filtering, R tdCorr Is the correlated value, α filter Is the Alpha filtering factor, R tdCorrHis Is the historical value of the fine frequency offset estimation in the frequency domain of the previous TTI.

[0058] In the present invention, phase calculation and conversion processing are performed on the correlated value after filtering to obtain a coarse frequency offset value, including:

[0059]

[0060] Among them, FO est Is the coarse frequency offset value, Δf is the subcarrier spacing, phase() is to calculate the phase, and the value range of phase(R tdCorrFilter ) is [-π, π], and the value range of FO est is [-Δf, Δf].

[0061] In the present invention, the method further includes:

[0062] If the coarse frequency offset value is greater than or equal to the estimation threshold, then perform IFFT transformation on all OFDM symbols according to the coarse frequency offset value to the time domain for coarse frequency offset compensation, and then obtain the compensated frequency domain signal through FFT;

[0063] According to DMRS, perform fine frequency offset estimation and frequency offset compensation on the compensated frequency domain signal;

[0064] If the coarse frequency offset value is less than the estimation threshold, then directly perform fine frequency offset estimation and fine frequency offset compensation based on the original frequency domain signal.

[0065] In the present invention, the estimation threshold is the maximum value of the fine frequency offset estimation.

[0066] In the present invention, performing coarse frequency offset compensation on all OFDM symbols according to the coarse frequency offset value to obtain the compensated frequency domain signal includes:

[0067] Perform IFFT transformation on all PUSCH frequency domain data symbols except DMRS to obtain a second time domain signal;

[0068] Perform frequency offset compensation on the second time domain signal and the first time domain signal to obtain the compensated time domain signal of all OFDM symbols;

[0069] Perform FFT transformation on the compensated time domain signal to obtain the compensated frequency domain signal.

[0070] In the present invention, performing IFFT transformation on all PUSCH frequency domain data symbols except DMRS to obtain a second time domain signal includes:

[0071] x(l pusch ,n) = ifft(X(lpusch , n))

[0072] wherein, x(l pusch , n) is the second time-domain signal, and l pusch is the symbol position occupied by PUSCH,

[0073] In the present invention, frequency offset compensation is performed on the second time-domain signal and the first time-domain signal to obtain a compensated time-domain signal, including:

[0074]

[0075] N DsCP = round(N CP / (N FFT / N DsFFT ))

[0076] wherein, x foComp (l, n) is the compensated time-domain signal, represents the starting position of each symbol in the time domain after downsampling due to domain transformation, l dmrs and l pusch together constitute the position of l in the time domain in the time domain, i.e., l = l dmrs + l pusch ,

[0077] The round() function is a rounding operation, N CP represents the length of the cyclic prefix CP, and N FFT represents the length of the system FFT transformation. j is the imaginary part symbol of a complex number. In the present invention, an FFT transformation is performed on the compensated time-domain signal to obtain a compensated frequency-domain signal, including:

[0078] X foComp (l, n) = fft(x foComp (l, n))

[0079] wherein, X foComp (l, n) represents the compensated frequency-domain signal.

[0080] In the present invention, after adaptively determining the coarse frequency offset value and the estimation threshold, fine frequency offset estimation and compensation based on DMRS pilots are performed according to the compensated frequency-domain signal or the frequency-domain signal.

[0081] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are:

[0082] 1. The present invention estimates the frequency offset within a sub - carrier range by using the phase difference of the time - domain signals in the OFDM symbol where the DMRS is located among the same symbols. The estimation range is significantly larger than that of the existing frequency - offset estimation schemes.

[0083] 2. The large - frequency - offset estimation scheme proposed by the present invention can perform adaptive processing based on the estimated value in the coarse frequency - offset estimation process. When the frequency - offset value obtained from the coarse frequency - offset estimation is greater than the estimation threshold, time - domain compensation of the frequency offset and domain transformation of the OFDM data are performed; when the frequency - offset value obtained from the coarse frequency - offset estimation is less than the estimation threshold, the frequency - offset compensation and domain - transformation processing of the coarse frequency - offset estimation are skipped, and the process directly enters the fine frequency - offset estimation process. Therefore, the computational complexity of the system can be greatly saved.

[0084] 3. The large - frequency - offset estimation scheme proposed by the present invention jointly estimates the frequency offset in the time - frequency domain. This time - frequency domain joint - estimation scheme can effectively solve the Doppler frequency - offset problem in the ultra - high - speed scenario under the OFDM system. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 It is a schematic diagram of the overall process provided by an embodiment of the present invention.

[0086] Figure 2 It is a schematic diagram of the detailed process provided by an embodiment of the present invention.

[0087] Figure 3a It is the simulation result provided by an embodiment of the present invention Figure 1 .

[0088] Figure 3b It is the simulation result provided by an embodiment of the present invention Figure 2 .

[0089] Figure 3c It is Figure 3 of the simulation result provided by an embodiment of the present invention.

[0090] Figure 3d It is the simulation result provided by an embodiment of the present invention Figure 4 .

[0091] Figure 4 It is the residual - frequency - offset curve graph provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0092] To make the objectives, 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 with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated herein can be arranged and designed in various different configurations.

[0093] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts fall within the scope of protection of the present invention.

[0094] First, the abbreviations in the present invention are shown in Table 1.

[0095] Table 1, Explanation of Abbreviations:

[0096] Abbreviation English Explanation Chinese Explanation OFDM Orthogonal Frequency Division Multiplexing Orthogonal Frequency Division Multiplexing CP Cyclic Prefix Cyclic Prefix DMRS Demodulation reference signal Demodulation reference signal FFT Fast Fourier Transform Fast Fourier Transform IFFT inverse fast Fourier transform inverse fast Fourier transform AWGN additive white Gaussian noise additive white Gaussian noise TDL Tapped Delay Line Tapped Delay Line MCS Modulation and Coding Scheme Modulation and Coding Scheme TTI Transmission Time Interval Transmission Time Interval PUSCH Physical Uplink Shared Channel Physical Uplink Shared Channel RB Resource Block Resource Block RE Resource Element Resource Element Least Squares Estimation Least Squares Estimation Least Squares Estimation

[0097] Traditional frequency offset estimation schemes have inaccurate user-level frequency offset estimation or relatively small estimation ranges in the ultra-high-speed mobile scenarios of OFDM systems, and cannot cope with the problem of the Doppler effect of large frequency offsets. Embodiments of the present invention propose a large frequency offset estimation method based on the joint time-frequency domain in an OFDM system. First, it can estimate the frequency offset at the user level; second, it can estimate large frequency offsets, and the estimation range is the frequency value of one subcarrier; then it can perform adaptive processing according to the estimated frequency offset value, reducing the processing flow and thus saving the system computing complexity; finally, it can ensure high estimation accuracy while estimating large frequency offsets, and the estimation results are accurate and reliable, ensuring the normal communication ability of the OFDM communication system in ultra-high-speed mobile scenarios.

[0098] Figure 1 It is a schematic diagram of the overall process provided for the embodiments of the present invention. The position of the frequency offset estimation scheme of the present invention in the PUSCH communication process is as Figure 1 shown. The frequency domain data processed at the front end is directly transmitted to the coarse frequency offset estimation module, the output data of the coarse frequency offset estimation module is transmitted to the channel estimation, and the frequency domain data after the channel estimation is transmitted to the fine frequency offset estimation, thus completing the entire frequency offset estimation process. The data after the fine frequency offset estimation is transmitted to subsequent processing modules such as channel equalization, demodulation, and decoding.

[0099] The present invention includes two steps: coarse frequency offset estimation in the time domain and fine frequency offset estimation in the frequency domain. First is the coarse frequency offset estimation in the time domain. The IFFT transform is performed on the user-level frequency domain data, the coarse frequency offset estimation and compensation of the time domain signal are carried out, and then it is transformed back to the frequency domain data through the FFT transform; then is the fine frequency offset estimation in the frequency domain, and the fine frequency offset estimation and frequency offset compensation based on the DMRS pilot signal are carried out.

[0100] In the above process, the coarse frequency offset estimation based on the time domain can estimate the frequency offset at the subcarrier level, which is the main technical means for the present invention to cope with the Doppler effect at ultra-high speeds. It specifically includes the following steps:

[0101] (1) Perform IFFT transformation on the original frequency-domain signal under the OFDM symbol where the DMRS is located to obtain the time-domain signal of the symbol where the DMRS is located;

[0102] (2) Use the time-domain signal of the symbol where the DMRS is located to perform conjugate correlation of the signals between symbols to obtain the correlation value;

[0103] (3) Perform Alpha filtering on the correlation value, considering the factor of the historical value, and perform smoothing filtering on the current instantaneous value;

[0104] (4) Calculate the phase based on the filtered correlation value and convert it into a frequency offset value;

[0105] (5) Threshold decision: If the frequency offset value is greater than the threshold, perform the subsequent steps; otherwise, directly skip steps (6), (7), and (8), and directly use the original frequency-domain signal to enter the fine frequency offset estimation;

[0106] (6) Perform IFFT transformation on the frequency-domain signals of all OFDM symbols except the DMRS symbol to the time domain;

[0107] (7) Perform frequency offset compensation on all time-domain OFDM symbols using the estimated frequency offset value;

[0108] (8) Perform FFT transformation on the time-domain OFDM symbols to the frequency domain to obtain the compensated frequency-domain signal after coarse frequency offset compensation.

[0109] (9) Perform fine frequency offset estimation using the original frequency-domain signal or the compensated frequency-domain signal.

[0110] The fine frequency offset estimation based on the frequency domain mainly processes according to the DMRS pilot signal, which is consistent with the traditional frequency offset estimation scheme based on the DMRS pilot signal. In the present invention, it is mainly to perform residual frequency offset estimation on the coarse frequency offset estimation and the compensated data in the time domain. The residual frequency offset range is relatively small, and exactly the DMRS pilot signal can be used for estimation, which plays a role in further improving the accuracy of frequency offset estimation. Through the combined frequency offset estimation scheme of two steps of coarse frequency offset estimation in the time domain and fine frequency offset estimation in the frequency domain, it is possible to estimate an extremely large frequency offset value at the subcarrier level, and at the same time ensure that the accuracy of frequency offset estimation is at a relatively high level.

[0111] Figure 2 This is the detailed flow diagram provided by the embodiment of the present invention. Taking the 5G PUSCH link as an example, referring to Figure 2 the processing flow diagram of the large frequency offset estimation scheme of the present invention at the PUSCH receiving end and the specific processing flow steps of the present invention, the detailed implementation manner of the present invention is as follows:

[0112] The first part, coarse frequency offset estimation

[0113] (1) Perform an IFFT transformation on the original frequency-domain signal in the OFDM symbol where the DMRS is located to the time domain to obtain the first time-domain signal:

[0114] Assume that the frequency-domain signal output by the front-end processing module at the receiving end is X(n). Performing an IFFT transformation on the original frequency-domain signal in the OFDM symbol where the DMRS is located is expressed as:

[0115] x(l dmrs ,n) = ifft(X(l dmrs ,n))

[0116] In the formula, x(l dmrs ,n) is the first time-domain signal, l dmrs represents the symbol position occupied by the DMRS, represents the number of symbols included in a slot, with a default value of 14; n ∈ [0, NSCPUSCH - 1], represents the number of frequency-domain subcarriers occupied by the PUSCH, ifft() is the IFFT transformation, and X(l dmrs ,n) is the original frequency-domain signal in the OFDM symbol where the DMRS is located.

[0117] (2) Perform conjugate correlation processing on the inter-symbol signals of the first time-domain signal:

[0118]

[0119] In the formula, R tdCorr is the correlation value, N dmrs represents the number of OFDM symbols occupied by the DMRS, () * represents the conjugate operation.

[0120] (3) Perform Alpha filtering on the correlation value:

[0121] R tdCorrFilter = R tdCorr ·(1 - α filter ) + R tdCorrHis ·α filter

[0122] In the formula, R tdCorrFilter is the filtered correlation value, α filter represents the Alpha filtering factor, with a default value selected as 0.7. R tdCorrHis represents the historical value of the frequency-domain coarse frequency offset estimation in the previous TTI.

[0123] (4) Perform phase calculation and conversion processing on the filtered correlation value:

[0124]

[0125] In the formula, FO est is the coarse frequency offset value, Δf represents the subcarrier spacing, phase() represents calculating the phase, indicating that only half of the data is used to estimate the frequency offset, and it needs to be multiplied by the ratio of the length of the complete frequency domain data to obtain the complete frequency offset value. The phase operation phase(R tdCorrFilter ) has a value range of [-π, π], so the value range of FO est is [-Δf, Δf], which is one subcarrier spacing.

[0126] (5) Threshold decision:

[0127] Assume that FO TH is the maximum value of the coarse frequency offset estimation. Taking FO TH as the estimation threshold, when the coarse frequency offset value FO est is less than this threshold, the frequency offset can be estimated through the subsequent fine frequency offset. Therefore, in the process of fine frequency offset estimation, there is no need to perform subsequent steps, and directly enter the processing flow of fine frequency offset estimation. At the same time, it also reduces the FFT transformation operation in the subsequent steps of coarse frequency offset estimation, saving the complexity of the calculation process; when the coarse frequency offset value FO est is greater than FO TH , it indicates that the frequency offset estimation value exceeds the range of fine frequency offset estimation, and it is necessary to continue to execute the subsequent frequency offset compensation process of the coarse frequency offset estimation, and perform frequency offset compensation on the time domain signal for subsequent signal processing.

[0128] (6) Frequency offset compensation for all OFDM symbols:

[0129] (a) Perform IFFT transformation on all PUSCH frequency domain data symbols except DMRS symbols:

[0130] Assume that the frequency domain data is X(n),

[0131] x(l pusch ,n) = ifft(X(l pusch ,n))

[0132] In the formula, x(l pusch ,n) is the second time domain signal, that is, the transformed PUSCH frequency domain data symbol, l pusch represents the symbol position occupied by PUSCH, represents the number of symbols included in one slot, and the default value is 14; represents the number of frequency domain subcarriers occupied by PUSCH.

[0133] (b) The corresponding lengths of FFT and CP after IFFT transformation:

[0134] Since the FFT length during IFFT transformation is the length of the PUSCH frequency-domain subcarriers, the sampling rate in the time domain will also decrease as the IFFT transformation length shortens. Compared with the FFT length of the normal system domain transformation, the sampling rate in the time domain decreases proportionally:

[0135] The FFT length is N DsFFT , expressed as:

[0136]

[0137] The length of the CP is N DsCP , expressed as:

[0138] N DsCP = round(N CP / (N FFT / N DsFFT ))

[0139] In the formula, the round() function represents the rounding operation.

[0140] (c) Frequency offset compensation of the time-domain signal:

[0141]

[0142]

[0143]

[0144] In the formula, represents the starting position of each symbol in the time domain after downsampling due to domain transformation, l dmrs and l pusch together form the position of l in the time domain, that is, l = l dmrs + l pusch ,

[0145] (7) FFT of all OFDM symbols:

[0146] Through the previous process, the frequency offset estimation and compensation of the time-domain signal are completed. Since the subsequent frequency offset estimation and equalization processing at the receiving end are signal processing in the frequency domain, it is necessary to transform the time-domain signal after frequency offset compensation to the frequency-domain signal through FFT.

[0147] X foComp (l,n) = fft(x foComp (l,n))

[0148] In the formula,

[0149] The second part, fine frequency offset estimation

[0150] The steps of fine frequency offset estimation are mainly based on the frequency offset estimation method of the DMRS pilot signal in the conventional LS channel estimation. After the previous fine frequency offset estimation and compensation, the residual frequency offset of the signal is small. The conventional frequency offset estimation method can be used to correct and compensate the residual frequency offset again to obtain a smaller comprehensive frequency offset estimation error, so as to achieve better anti-frequency offset ability.

[0151] (1) LS channel estimation:

[0152] Assume that the frequency-domain signal at the receiving end is X l (n), and the formula for Ls channel estimation is as follows:

[0153]

[0154] In the formula, represents the DMRS base sequence, l DMRS represents the OFDM symbol where the DMRS is located, B ScNum represents the number of frequency-domain subcarriers of the DMRS. The result output by the formula is the channel estimated by LS

[0155] (2) Correlation between OFDM symbols where different DMRSs are located:

[0156]

[0157] In the formula, R corr represents the conjugate correlation value of different DMRS symbols in the frequency domain. Taking the phase of it can obtain the phase difference represented by the frequency offset, H LS (L dmrsSet (l),n) represents the frequency-domain channel estimation value after LS estimation, L dmrsSet represents the set of OFDM symbols where the DMRS is located, N dMRS represents the number of OFDM symbols occupied by the DMRS, () * represents the conjugate operation.

[0158] (3) Alpha filtering:

[0159] R corrFilter = R corr ·(1 - α filter ) + R corrHis ·α filter

[0160] In the formula, α filter represents the Alpha filtering factor, and the default value is selected as 0.7; R corrHis represents the historical value of the fine frequency offset estimation in the frequency domain of the previous TTI.

[0161] (4) Frequency offset calculation:

[0162]

[0163] Wherein, Fo est is the fine frequency offset value, Δf represents the subcarrier spacing, and l Δt represents the difference in the time domain between different DMRS symbols. When there are more than 2 DMRS symbols, that is, there are multiple l Δt , the calculation of the variable l Δt can be placed in the calculation formula of R corr corresponding to multiple symbols.

[0164] (5) Frequency offset compensation:

[0165] When time domain interpolation is performed for channel estimation, inter-symbol frequency offset compensation is performed for different OFDM symbols:

[0166]

[0167] l minDmrs = argmin{|l - L dmrsSet |}

[0168] l diff = l minDmrs - l

[0169] Wherein, y FoComp (l, n) represents the frequency offset compensation result, and y LsInterp (l, n) represents the channel estimation value of the data symbol after time domain interpolation, l minDmrs represents the position of the DMRS symbol with the smallest difference in the DMRS symbol set from the current OFDM symbol l, and l diff represents the time domain difference between the current symbol l and the DMRS symbol. When l diff is 0, it indicates that the current symbol is the symbol where the DMRS is located.

[0170] The technical effects of the present invention are verified as follows:

[0171] For the frequency offset estimation design scheme of the present invention, the simulation demonstration of the design scheme effect is carried out taking the 5G uplink shared data channel (PUSCH) as an example.

[0172] (1) Simulation parameter configuration is shown in Table 1.

[0173] Table 1, simulation parameters:

[0174]

[0175]

[0176] (2) Configuration of simulation test cases, as shown in Table 2.

[0177] Table 2, simulation test cases:

[0178]

[0179] (3) Simulation results:

[0180] Figure 3a 、 Figure 3b 、 Figure 3c 、 Figure 3d are the schematic diagrams of the simulation results provided by the embodiments of the present invention. Figure 3a 、 Figure 3b 、 Figure 3c 、 Figure 3d In, FO = 25 kHz indicates the fixed frequency offset value of 25 kHz added at the transmitting end of the link. From Figure 3a 、 Figure 3b 、 Figure 3c 、 Figure 3d of the 4 test cases, it can be seen that for the same test case, the demodulation performance with different fixed frequency offsets is basically the same, indicating that the frequency offset estimation and compensation ability of the scheme is good, and different frequency offsets can be correctly estimated and compensated without affecting the demodulation performance of the link. Through the above analysis of the simulation results, it can be known that the frequency offset estimation and compensation ability of this scheme can resist large frequency offsets under 29 kHz, approaching the frequency offset estimation ability of one subcarrier. Using the frequency offset calculation formula f d = vf c ·cosθ / c, within sub6GHz specified by FR1, the movement speed at Mach level can be satisfied. For example, when calculating with a carrier frequency of 3 GHz, the highest supported movement speed is 10800 km / h, that is, 8.8 Mach; when the carrier frequency is lower, higher movement speeds can be supported.

[0181] Figure 4 is the residual frequency offset curve diagram provided by the embodiments of the present invention. From Figure 4 it can be seen that the curves of the two colors respectively represent the residual frequency offset errors of the AWGN channel at different MCS levels. When adding a fixed frequency offset of 25 kHz at low SNR, the residual frequency offset error is less than 300 Hz. Compared with the subcarrier spacing of 30 kHz, the normalized residual frequency offset is less than 0.01, having very good frequency offset estimation ability; and as the SNR increases, when at MCS13, adding a fixed frequency offset of 29 kHz, the residual frequency offset values are all less than 100 Hz, and the frequency offset estimation effect is better.

[0182] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A large frequency offset estimation method based on time-frequency domain joint in OFDM system, characterized in that: include: Performing IFFT transformation on the original frequency domain signal under the OFDM symbol where the DMRS is located to the time domain to obtain a first time domain signal; Performing conjugate correlation processing of inter-symbol signals on the first time domain signal to obtain a correlation value; Performing Alpha filtering on the correlation value to obtain a filtered correlation value; Performing phase calculation and conversion processing on the filtered correlation value to obtain a coarse frequency deviation value; Comparing the coarse frequency offset value with an estimated threshold to adaptively determine whether to perform coarse frequency offset compensation; According to the adaptive decision result, the frequency domain signals under all OFDM symbols are precisely estimated to obtain a precise frequency offset estimation value.

2. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 1 is characterized in that: The performing IFFT transformation on the original frequency domain signal under the OFDM symbol where the DMRS is located to the time domain to obtain the first time domain signal includes: x(l dmrs ,n)=ifft(X(l dmrs ,n)) Among them, x(l dmrs , n) is the first time domain signal, l dmrs is the symbol position occupied by DMRS, is the number of symbols contained in a slot, is the number of frequency domain subcarriers occupied by PUSCH, ifft() is IFFT transform, X(l dmrs ,n) is the original frequency domain signal under the OFDM symbol where the DMRS is located.

3. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 2 is characterized in that: The performing conjugate correlation processing of the inter-symbol signal on the first time domain signal to obtain a correlation value includes: Among them, R tdCorr is the correlation value, N dmrs is the number of OFDM symbols occupied by the DMRS, () * Represents the conjugate operation.

4. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 3 is characterized in that: The step of performing Alpha filtering on the correlation value to obtain a filtered correlation value includes: R tdCorrFilter =R tdCorr ·(1-a filter )+R tdCorrHis ·a filter Among them, R tdCorrFilter is the correlation value after filtering, α filter is the Alpha filter factor, R tdCorrHis It is the historical value of the frequency domain precise frequency offset estimation of the previous TTI.

5. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 4 is characterized in that: The performing phase calculation and conversion processing on the filtered correlation value to obtain a coarse frequency deviation value includes: Among them, FO est is the coarse frequency deviation value, Δf is the subcarrier spacing, phase() is the phase, phase(R tdCorrFilter ) has a value range of [-π,π], FO est The value range of is [-Δf,Δf].

6. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to any one of claims 1 to 5, characterized in that: The method of performing precise frequency offset estimation on the frequency domain signals under all OFDM symbols according to the adaptive decision result to obtain a precise frequency offset estimation value includes: If the coarse frequency offset value is greater than or equal to the estimation threshold, all OFDM symbols are transformed into the time domain by IFFT according to the coarse frequency offset value to perform coarse frequency offset compensation, and then the compensated frequency domain signal is obtained by FFT; According to the DMRS, performing precise frequency offset estimation and precise frequency offset compensation on the compensated frequency domain signal; If the coarse frequency offset value is smaller than the estimation threshold, fine frequency offset estimation and fine frequency offset compensation are performed according to the original frequency domain signal.

7. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 6 is characterized in that: The estimation threshold is the maximum value of the precise frequency offset estimation.

8. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 6 is characterized in that: The performing coarse frequency offset compensation on all OFDM symbols according to the coarse frequency offset value to obtain a compensated frequency domain signal includes: Performing IFFT transformation on all PUSCH frequency domain data symbols except the DMRS to obtain a second time domain signal; Performing frequency offset compensation on the second time domain signal and the first time domain signal to obtain compensated time domain signals of all OFDM symbols; Perform FFT transformation on the compensated time domain signal to obtain a compensated frequency domain signal.

9. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 8, characterized in that: Performing IFFT transformation on all PUSCH frequency domain data symbols except the DMRS to obtain a second time domain signal, including: x(l pusch ,n)=ifft(X(l pusch ,n)) Among them, x(l pusch , n) is the second time domain signal, l pusch is the symbol position occupied by PUSCH, 10. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 8, characterized in that: The performing frequency offset compensation on the second time domain signal and the first time domain signal to obtain compensated time domain signals of all OFDM symbols includes: N DsCP =round(N CP / (N FFT / N DsFFT )) Among them, x foComp (l,n) is the time domain signal after compensation, represents the starting position of each symbol in the time domain after downsampling due to domain transformation, l dmrs With l pusch Together they form the position of l in the time domain, i.e. l ​​= l dmrs +l pusch , The round() function is a rounding operation. CP Indicates the length of the cyclic prefix CP, N FFT It represents the length of the system FFT transformation, and j is the sign of the imaginary part of the complex number.

11. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 8, characterized in that: The performing FFT transformation on the compensated time domain signal to obtain the compensated frequency domain signal includes: X foComp (l,n)=fft(x foComp (l,n)) Among them, X foComp (l,n) represents the frequency domain signal after compensation.

12. The large frequency offset estimation method based on time-frequency domain joint in OFDM system according to claim 6, characterized in that: The method further comprises: After adaptively determining the coarse frequency offset value and the estimation threshold, fine frequency offset estimation and compensation based on the DMRS pilot are performed according to the compensated frequency domain signal or the original frequency domain signal.

Images