A timing synchronization and frequency offset estimation method based on multi-segment repeated preamble sequence

Abstract

The application discloses a timing synchronization and frequency offset estimation method based on a multi-segment repeated preamble sequence, which comprises the following steps: S1, receiving the transmission data with a multi-segment repeated preamble sequence transmitted by a transmission end by using a plurality of receiving antennas to obtain a receiving signal; S2, performing a plurality of rounds of sliding autocorrelation on the receiving signal, and performing weighted summation on the obtained autocorrelation values to obtain a timing measurement variable; S3, if the timing measurement variable is greater than a preset detection threshold η, it is considered that a timing synchronization point is reached, and S4 is performed, otherwise, S2 is returned; S4, on the basis of the timing synchronization point, continuing to slide backward to find the maximum value of the timing measurement variable, and obtaining the phases of a plurality of groups of autocorrelation values corresponding to the maximum value of the timing measurement variable; S5, iteratively solving the signal frequency offset according to the phases of the plurality of groups of autocorrelation values corresponding to the maximum value of the timing measurement variable to obtain a frequency offset estimation value. The timing synchronization point is found while the frequency offset is estimated, and high-precision time-frequency synchronization under a low signal-to-noise ratio can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to a timing synchronization and frequency offset estimation method of single carrier communication system, in particular to a timing synchronization and frequency offset estimation method based on multiple segment repeated preamble sequence. BACKGROUND

[0002] Timing synchronization and frequency offset estimation are very important in wireless digital communication system, which are the premise of the correct processing of the received signal at the receiving end. The role of timing synchronization is to determine whether the signal has arrived and when it has arrived. Frequency offset estimation is to calculate the deviation of the center frequency of the received signal from the carrier frequency of the receiving end.

[0003] In recent years, there have been many studies on timing synchronization technology. Cox and Schmidl proposed "Robust frequency and timing synchronization for OFDM" in IEEE Transactions on Communications, vol. 45, no. 12, pp. 1613-1621. This algorithm uses two repeated training sub-sequences for delay autocorrelation. When the frame synchronization metric function of timing synchronization reaches the maximum value, the corresponding position is the frame header. On this basis, Minn published "On timing offset estimation for OFDM systems" in IEEE Commun. Lett, vol. 4, pp. 242-244, which improved the above algorithm and eliminated the influence of the cyclic prefix. This algorithm is achieved by introducing different symbols in the two repeated sequences before and after, but the performance is limited in multipath channel conditions, and the complexity of hardware implementation is also large. In 2017, Mu Pengcheng and Chen Jiancheng et al. applied for a patent "A smoothing autocorrelation timing coarse synchronization method based on multiple segment repeated preamble sequence". This method uses the characteristics of multiple repeated sequences to smooth the signal at the receiving end to improve the equivalent signal-to-noise ratio, which can realize the detection of the synchronization signal under low signal-to-noise ratio (below 0 dB), but this method cannot work normally when there is a large frequency offset in the receiving signal.

[0004] After the timing synchronization is achieved, the system also needs to estimate and compensate the frequency offset of the received signal. Chen Dafu et al. proposed a fast Fourier transform carrier frequency offset estimation algorithm in Circuit and System Letters, 2006(02):128-132. The algorithm is based on maximum likelihood estimation, and selects the maximum value in the amplitude-frequency characteristics of the signal, and the corresponding frequency is the estimation of the frequency offset, but the frequency offset estimation accuracy of this method is limited by the number of FFT. Gao Yunfeng et al. published a new algorithm for blind estimation of carrier frequency offset in digital communication in Electronic Technology Application, 2005(08):53-54, which uses the phase difference between the front and rear sampling points to estimate the carrier frequency offset, and its capture range can theoretically reach the symbol rate, but it is only suitable for higher signal-to-noise ratio environment. SUMMARY

[0005] The present application aims at the shortcomings of the existing time-frequency synchronization algorithm, such as high signal-to-noise ratio required, low estimation accuracy, etc., and proposes a timing synchronization and frequency offset estimation method based on multi-segment repeated preamble sequence, which can realize high-precision time-frequency synchronization under low signal-to-noise ratio by finding the timing synchronization point and estimating the frequency offset at the same time.

[0006] The present application is realized by the following technical solutions:

[0007] A timing synchronization method based on multi-segment repeated preamble sequence, comprising:

[0008] S1, receiving the transmitted data with multi-segment repeated preamble sequence transmitted by the transmitting end using multiple receiving antennas to obtain a received signal;

[0009] S2, performing multi-round sliding autocorrelation on the received signal, and performing weighted summation on the obtained autocorrelation values to obtain a timing measurement variable;

[0010] S3, if the timing measurement variable is greater than a preset detection threshold η, it is considered that the timing synchronization point is reached, otherwise, return to S2; wherein the value of the preset detection threshold η satisfies 0<η<1.

[0011] Preferably, S2 is specifically: sampling the received signal to obtain a digital baseband signal at time d, denoted as x t (d), t=1,…,T, performing N-1 rounds of sliding autocorrelation on the digital baseband signal of each receiving antenna, the interval point number of the front and rear two data of the i-th round of sliding autocorrelation is iL, and the sampling point number of each data in the front and rear two data is (N-i)L, wherein i=1, 2,…, N-1, a group of autocorrelation values is obtained corresponding to each round of sliding autocorrelation, the autocorrelation values obtained by each round of sliding autocorrelation of each receiving antenna are normalized, and then weighted summation is performed to obtain the timing measurement variable at time d;

[0012] Wherein, T is the number of receiving antennas, N is the number of repeated segments of the preamble sequence, and L is the number of sampling points of each segment of the preamble sequence.

[0013] Further, the timing measurement variable λ(d) at time d is expressed as:

[0014]

[0015] Wherein:

[0016]

[0017]

[0018] Wherein, α>0 is a preset parameter, [·] * represents a conjugate complex number, Z i (d) represents the autocorrelation value at time d, P i (d) represents the normalized variable at time d, n is the index value when the weighted sum is taken, n=0, 1, 2, …, (N-i)L-1.

[0019] Preferably, S3 is specifically: if the timing measurement variable is greater than a preset detection threshold η for consecutive M points, it is considered that the timing synchronization point is reached, M≥1 and is an integer.

[0020] Further, the preset detection threshold η is obtained by using Monte Carlo simulation to obtain the timing synchronization false alarm probability under different detection threshold values, and taking the timing synchronization false alarm probability equal to the false alarm probability preset threshold ξ FA as the preset detection threshold η; wherein, 0<ξ FA <1, the timing synchronization false alarm probability refers to the probability that the timing measurement variable is higher than the detection threshold value for consecutive M points under the condition that the transmitted data does not contain the preamble sequence, M≥1 and is an integer.

[0021] A frequency offset estimation method based on a multi-segment repeated preamble sequence, based on the timing synchronization method based on the multi-segment repeated preamble sequence, when the timing synchronization point is reached, the following operations are performed:

[0022] S4, on the basis of the timing synchronization point, continue to slide backward to find the maximum value of the timing measurement variable, and find the phase of the multiple sets of autocorrelation values corresponding to the maximum value of the timing measurement variable;

[0023] S5, according to the phase of the multiple sets of autocorrelation values corresponding to the maximum value of the timing measurement variable, iteratively solve the signal frequency offset to obtain the frequency offset estimation value.

[0024] Preferably, S4 is specifically: continuing to slide K points backward on the basis of the timing synchronization point, finding the maximum value of the timing measurement variable in the K points, wherein K>=NL, and obtaining the phase of the N-1 groups of autocorrelation values corresponding to the maximum value of the timing measurement variable i=1,...,N-1

[0025] Wherein, N is the number of repeated segments of the preamble sequence, and L is the number of sampling points of each segment of the preamble sequence.

[0026] Further, S5 is specifically: on the basis of the phase of the N-1 groups of autocorrelation values obtained in S4 , the phase of the i-1 group of corrected autocorrelation values is recorded as i=2,3,...,N-1, Let i=2, and the frequency offset estimation value is calculated in the following way:

[0027] 1) Calculate the phase deviation of the phase of the i group of autocorrelation values Wherein

[0028] 2) According to the phase deviation, the phase of the i group of corrected autocorrelation values is calculated as Wherein β>0 is a preset parameter;

[0029] 3) Let i=i+1, if i=N-1, end the frequency offset estimation and obtain the frequency offset estimation value, otherwise return to step 1).

[0030] Further, the frequency offset estimation value is Wherein f s is the sampling frequency.

[0031] Compared with the prior art, the present application has the following beneficial effects:

[0032] The present application is a timing synchronization method based on pilot under wireless line-of-sight conditions, and the receiving end performs multiple autocorrelation calculations of different lengths on the received signal, and performs weighted summation on the obtained autocorrelation values. Compared with the existing time-frequency synchronization algorithm, the multi-segment preamble sequence is fully utilized, the equivalent signal-to-noise ratio is improved, the lowest signal-to-noise ratio of the signal can be detected by the present method, which is lower than that of the Cox & Schmidl method, and the false alarm probability and the miss probability are guaranteed to be very low. At the same time, since the present method does not need to smooth the received signal, it can still work normally when there is a large frequency offset in the received signal.

[0033] The application carries out frequency deviation estimation while searching timing synchronization point, uses the phase of the autocorrelation value of multiple sets of different length autocorrelation to iteratively correct the original phase, and calculates the frequency deviation according to the corrected phase, so that the phase ambiguity problem in the process of calculating the phase angle is effectively avoided, and the precision of the frequency deviation estimation is improved under the condition of ensuring the same frequency deviation estimation range as the classical method. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 The figure is a schematic diagram of the preamble sequence of the transmitting end and the autocorrelation detection of the receiving end of the application.

[0035] Figure 2 The timing synchronization false alarm probability of the application and the Cox & Schmidl method under different detection threshold values is compared.

[0036] Figure 3 The timing synchronization false alarm probability of the application and the Cox & Schmidl method under different signal-to-noise ratios is compared when the timing synchronization false alarm probability is 1e-4.

[0037] Figure 4 The frequency deviation estimation performance of the application and the Cox & Schmidl method under different signal-to-noise ratios is compared. DETAILED DESCRIPTION

[0038] In order to further understand the application, the application will be described below in conjunction with examples, which are only used to further explain the features and advantages of the application, and are not used to limit the claims of the application.

[0039] The application is based on the transmitting end adding a preamble sequence with N repeated segments before the data to be transmitted, and obtaining the transmitting data; wherein the sampling point number of each preamble sequence is L, and the total length of the N preamble sequences is NL, such as Figure 1 .

[0040] Based on the transmitting data with multiple repeated preamble sequences transmitted by the transmitting end, the application is a timing synchronization and frequency deviation estimation method based on multiple repeated preamble sequences, which includes seven steps, and the specific introduction is as follows.

[0041] S1, such as Figure 1 The receiving signal (i.e. the transmitting data in S1 transmitted by the transmitting end) is obtained by using T receiving antennas, the receiving signal is sampled, and the digital baseband signal at time d can be expressed as x t (d), t = 1, …, T.

[0042] S2, to construct the timing measurement variable, firstly, the digital baseband signal of each receiving antenna is subjected to N-1 rounds of sliding autocorrelation, and a set of autocorrelation values is obtained for each round of sliding autocorrelation. Specifically, the interval point number of the two pieces of data before and after the i-th round of sliding autocorrelation is iL, and the sampling point number of each piece of data in the two pieces of data is (N-i)L, where i = 1, 2, …, N-1; then, the autocorrelation values obtained by each round of sliding autocorrelation of each receiving antenna are normalized, and then weighted summation is performed to obtain the final timing measurement variable. At this time, the timing measurement variable λ(d) at time d can be represented as:

[0043]

[0044] wherein:

[0045]

[0046]

[0047] wherein α>0 is a preset parameter, [·] * denotes a conjugate complex, Z i (d) denotes the autocorrelation value at time d, P i (d) denotes the normalization variable at time d, and n is an index value during the weighted summation, n = 0, 1, 2, …, (N-i)L-1.

[0048] S3, the timing synchronization false alarm probability under different detection thresholds is obtained by using Monte Carlo simulation, and the detection threshold corresponding to the timing synchronization false alarm probability equal to the false alarm probability preset threshold ξ FA (0<ξ FA <1) is taken as the preset detection threshold η (0<η<1); the timing synchronization miss probability under different signal-to-noise ratios is obtained by using Monte Carlo simulation under the preset detection threshold η, and the signal-to-noise ratio corresponding to the timing synchronization miss probability equal to the miss probability preset threshold ξ MD (0<ξ MD <1) is taken as the minimum signal-to-noise ratio that can be detected by the method. In the present application, the timing synchronization false alarm probability is defined as the probability that the timing measurement variable is higher than the detection threshold value for M (M≥1 and is an integer) consecutive points under the condition that the transmitted data does not contain a preamble sequence; and the timing synchronization miss probability is defined as the probability that the timing measurement variable is lower than the detection threshold value under the condition that the preamble sequence is transmitted.

[0049] S4, the timing measurement variable λ(d) of the received signal is calculated according to the method described in the second step, and when the calculated value of the timing measurement variable is greater than the preset detection threshold η determined in the third step for M consecutive points, it is considered that the timing synchronization point is reached (i.e., the initial position of the preamble sequence is detected), and S5 is performed, otherwise it is considered that the synchronization is not reached (i.e., the preamble sequence is not detected in the received signal), and S4 is repeated.

[0050] S5, continue to slide K points backward on the basis of the timing synchronization point, find the maximum value of the timing measurement variable in the K points, wherein K≥NL, the phase of N-1 groups of autocorrelation values at the corresponding position can be obtained according to the position of the maximum value of the timing measurement variable i=1,...,N-1.

[0051] S6, on the basis of the phase of N-1 groups of autocorrelation values obtained in S5 i=1,...,N-1, record the phase of the i-1 group of corrected autocorrelation values as (especially, i=2, 3,...,N-1), let i=2, and the frequency offset estimation value is calculated in the following way:

[0052] 1) calculate the phase deviation of the phase of the i group of autocorrelation values Wherein

[0053] 2) calculate the phase of the i group of corrected autocorrelation values according to the phase deviation Wherein β>0 is a preset parameter;

[0054] 3) let i=i+1, if i=N-1, end the frequency offset estimation and obtain the frequency offset estimation value Wherein f s is the sampling frequency, otherwise return to step 1).

[0055] Embodiment

[0056] The specific parameter setting of the embodiment of the application is: the number of receiving antennas T=4, the number of repeated segments of the preamble sequence N=10, the number of sampling points of each segment of the preamble sequence L=8, the total length of the preamble sequence NL=80, the sampling frequency f s =16MHz, the true value of the frequency offset is set as Δf real =800KHz, the false alarm probability threshold ξ FA =1e-4, the false alarm probability threshold ξ MD =1e-4, the number of continuously detected timing measurement variables M=5, the remaining parameters are K=80, α=1, β=2, and the channel model is a Gaussian white noise channel. The signal-to-noise ratio is defined as Wherein is the received signal power, is the variance of the Gaussian white noise.

[0057] Figure 2The false alarm probability of timing synchronization under different detection thresholds using the method of the application and the method proposed by Cox and Schmidl in the paper of "Robust frequency and timing synchronization for OFDM" published in IEEE Transactions on Communications, vol. 45, no. 12, pp. 1613-1621 is compared.

[0058] Figure 3 The false alarm probability of timing synchronization under different detection thresholds using the method of the application and the method proposed by Cox and Schmidl in the paper of "Robust frequency and timing synchronization for OFDM" published in IEEE Transactions on Communications, vol. 45, no. 12, pp. 1613-1621 is compared.

[0059] Figure 4 The false alarm probability of timing synchronization under different detection thresholds using the method of the application and the method proposed by Cox and Schmidl in the paper of "Robust frequency and timing synchronization for OFDM" published in IEEE Transactions on Communications, vol. 45, no. 12, pp. 1613-1621 is compared. Wherein, is the estimated frequency offset value. As can be seen from the figure, when the MMSE is 1e3, the SNR required by the method of the application is 3dB lower than that of the Cox & Schmidl method, especially at low SNR (below 0dB), the frequency offset error estimated by the method of the application is significantly lower than that of the Cox & Schmidl method.

[0060] The method of the application can complete timing synchronization at a lower SNR by performing multiple autocorrelation calculations on the received signal and performing weighted summation on the obtained autocorrelation values. At the same time, the phases of multiple sets of autocorrelation values are used to iteratively solve the signal frequency offset, avoiding the phase ambiguity problem in the process of solving the phase angle, and improving the accuracy of frequency offset estimation while ensuring the estimable frequency offset range.

Claims

1. A timing synchronization method based on multiple repeated preamble sequences, characterized in that: include: S1, using multiple receiving antennas to receive the transmission data with multiple repeated preamble sequences transmitted by the transmitting end to obtain a received signal; S2, performing multiple rounds of sliding autocorrelation on the received signal, performing weighted summation on the obtained autocorrelation values, and obtaining a timing measurement variable; S3, if the timing measurement variable is greater than the preset detection threshold , it is considered that the timing synchronization point has been reached, otherwise, return to S2; Among them, the preset detection threshold The value of ; S2 is specifically: sampling the received signal to obtain The digital baseband signal at time , the digital baseband signal of each receiving antenna is The sliding autocorrelation of the wheel, The number of interval points between the two segments of data before and after the wheel slip autocorrelation is , the number of sampling points in each of the two segments of data is ,in , a set of autocorrelation values ​​is obtained for each round of sliding autocorrelation, and the autocorrelation values ​​obtained for each round of sliding autocorrelation of each receiving antenna are normalized and then weighted summed to obtain Timed measurement variables at the moment; in, is the number of receiving antennas, is the number of repeated segments of the leading sequence, is the number of sampling points of each leading sequence; Time measurement variables at the moment Expressed as: in: in, For the preset parameters, represents the conjugate complex number, represents the autocorrelation value at time d, represents the normalized variable at time d, n is the index value for weighted summation, n =0,1,2,…,( N - i ) L -1.

2. The timing synchronization method based on multiple repeated preamble sequences according to claim 1, characterized in that: S3 is specifically: If the timed measurement variable is continuous The point is greater than the preset detection threshold , it is considered that the timing synchronization point is reached, And it is an integer.

3. The timing synchronization method based on multiple repeated preamble sequences according to claim 2, characterized in that: The preset detection threshold The method for obtaining is: using Monte Carlo simulation to obtain the timing synchronization false alarm probability under different detection threshold values, taking the timing synchronization false alarm probability equal to the false alarm probability preset threshold The corresponding detection threshold value is used as the preset detection threshold ;in, The timing synchronization false alarm probability refers to the probability of the timing measurement variable being continuously The probability that the point is above the detection threshold, And it is an integer.

4. A frequency offset estimation method based on multiple repeated preamble sequences, characterized in that: The timing synchronization method based on multiple repeated preamble sequences according to any one of claims 1 to 3 is configured to perform the following operations after reaching a timing synchronization point: S4, continuing to slide backward based on the timing synchronization point, finding the maximum value of the timing measurement variable, and calculating the phases of multiple groups of autocorrelation values ​​corresponding to the maximum value of the timing measurement variable; S5, iteratively solving the signal frequency offset according to the phases of multiple groups of autocorrelation values ​​corresponding to the maximum values ​​of the timing measurement variables to obtain a frequency offset estimation value.

5. The frequency offset estimation method based on multiple repeated preamble sequences according to claim 4, characterized in that: S4 is specifically: continue to slide backward based on the timing synchronization point Point, here Find the maximum value of the timed measurement variable within the range of points, where , find the maximum value of the timed measurement variable corresponding to Phase of the group autocorrelation value ; in, is the number of repeated segments of the leading sequence, is the number of sampling points in each leading sequence.

6. The frequency offset estimation method based on multiple repeated preamble sequences according to claim 5, characterized in that: S5 is specifically: obtained in S4 Phase of the group autocorrelation value On the basis of The phase of the autocorrelation value after group correction is , , ,make , the frequency offset estimate is calculated as follows: 1) Calculate the Phase deviation of the group autocorrelation value phase ,in ; 2) Calculate the first The phase of the autocorrelation value after group correction is ,in are the preset parameters; 3) Order ,like , then the frequency offset estimation is completed and the frequency offset estimation value is obtained, otherwise return to step 1).

7. The frequency offset estimation method based on multiple repeated preamble sequences according to claim 6, characterized in that: The frequency offset estimate is ,in is the sampling frequency.

Images