Time-frequency synchronization method of non-cooperative OFDM communication system

Abstract

The invention discloses a time-frequency synchronization method of a non-cooperative OFDM (Orthogonal Frequency Division Multiplexing) communication system, which comprises the following steps of: performing primary frame synchronization on a received signal by using a cyclic prefix self-correlation blind synchronization algorithm to realize frame header positioning, positioning a time domain position of a synchronization sequence of the received signal by using a sliding correlation window, performing clustering judgment on the synchronization sequence in continuous multi-frame signals, and performing clustering judgment on the synchronization sequence in the continuous multi-frame signals. Counting and extracting an initial synchronization sequence through a multi-frame judgment result self-correlation peak value; according to the self-correlation characteristic and the cycle characteristic of a synchronization sequence, the invention provides a rapid symbol synchronization and optimal sampling point synchronization method, and frequency offset estimation is realized based on a Moose algorithm. According to the method, after time-frequency synchronization is carried out on the received signal, the synchronized sequence is fed back to the clustering judgment step to carry out iterative updating on the local synchronization sequence, and the overall system performance is further improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a time-frequency synchronization method for a non-cooperative OFDM communication system. Background Art

[0002] Non-cooperative communication refers to a communication method in which a third party, as a communicator, accesses a cooperative communication system and conducts radio cognition and reconnaissance in an unauthorized mode. It has been widely used in military and civilian fields. In the field of civilian communications, in order to ensure the rational use of spectrum space and a healthy and good communication environment, relevant regulatory authorities must monitor and control free space signals, which requires third parties to blindly identify communication signals without any prior information; in the field of military communications, non-cooperative parties need to correctly identify and demodulate unknown signals in order to monitor and reconnaissance communication signals.

[0003] Orthogonal Frequency Division Multiplexing (OFDM) technology has become the mainstay of contemporary communication systems due to its advantages of strong anti-multipath and anti-interference capabilities, high spectrum utilization, and high-speed data transmission. Currently, OFDM technology has been widely used in 4G and 5G mobile communications and satellite communications, so it is of great significance to study the non-cooperative blind reception of OFDM.

[0004] Time-frequency synchronization technology is a key technology on the receiving side of the communication system. Correct time-frequency synchronization is the basis and prerequisite for subsequent channel estimation, channel equalization and correct demodulation of the received signal. For non-cooperative OFDM systems, the existing typical methods of blind time-frequency synchronization include: maximum likelihood algorithm based on cyclic prefix (CP), algorithm based on signal cyclostationarity and algorithm based on autocorrelation matrix. However, the existing methods are often complex and have limited performance in complex channel environments. Summary of the invention

[0005] In order to solve the limitations and defects of the prior art, the present invention provides a time-frequency synchronization method for a non-cooperative OFDM communication system, comprising:

[0006] Step S1, using a blind synchronization algorithm based on cyclic prefix autocorrelation to perform initial frame synchronization on a received signal to achieve frame header positioning of the received signal;

[0007] Step S2, using a sliding correlation window to locate the time domain position of the synchronization sequence of the received signal;

[0008] Step S3, extracting synchronization sequences in the multi-frame signals according to the positioning results;

[0009] Step S4, using a constellation diagram clustering method to perform clustering judgment on the synchronization sequence in the multi-frame signal, performing autocorrelation peak comparison of the multi-frame judgment results, and counting and extracting the initial synchronization sequence;

[0010] Step S5, using the initial synchronization sequence to perform time-frequency synchronization on subsequent received signals;

[0011] After step S5 is executed, the sequence after time-frequency synchronization is fed back to step S4, and step S4 is executed to iteratively update the local synchronization sequence.

[0012] Optionally, step S5 includes:

[0013] Perform symbol synchronization and optimal sampling point synchronization;

[0014] The Moose algorithm is used to estimate the frequency offset of the sequence after symbol synchronization and optimal sampling point synchronization are achieved.

[0015] Optionally, before step S1, the following steps are included:

[0016] Signal parameters are obtained, where the signal parameters include an orthogonal frequency division multiplexing system bandwidth, a subcarrier spacing, a fast Fourier transform point number, and a cyclic prefix length.

[0017] Optionally, step S2 includes:

[0018] Step S21, setting the number of fast Fourier transform points of the OFDM signal to nfft, the cyclic prefix length to cp_len, and each frame to contain frame_len OFDM symbols;

[0019] Step S22, setting a sliding correlation window with a window length of nfft+cp_len with the frame header as the starting point;

[0020] Step S23, calculating the autocorrelation peak value of the elements in the sliding correlation window;

[0021] Step S24, slide the sliding correlation window backward by nfft+cp_len, and repeat step S23;

[0022] Step S25: compare frame_len autocorrelation peaks, and determine that the position with the largest autocorrelation peak is the symbol where the synchronization sequence is located.

[0023] Optionally, the step of performing clustering decision on the synchronization sequences in the multi-frame signal using a constellation diagram clustering method includes:

[0024] Step S41, determining a K value according to a synchronization sequence modulation order, wherein K corresponds to the number of ideal constellation points;

[0025] Step S42, randomly selecting constellation coordinates of K data points from the synchronization sequence as initial centroids;

[0026] Step S43, calculating the Euclidean distance between each constellation point in the synchronization sequence and each centroid, and assigning each constellation point to the cluster represented by the centroid with the shortest Euclidean distance;

[0027] Step S44: For each cluster, calculate the mean of all constellation points, and set the mean as a new centroid;

[0028] Repeat step S43 and step S44 until a stop condition is met, wherein the stop condition is that the change of the centroid in consecutive iterations is less than a preset threshold, or a maximum number of iterations is reached;

[0029] Step S45: The synchronization sequence constellation points have been classified into K categories, and the synchronization sequence constellation points are determined according to the synchronization sequence modulation order.

[0030] Optionally, the step of comparing the autocorrelation peak values ​​of the multi-frame decision results, counting and extracting the initial synchronization sequence includes:

[0031] Obtaining a normalized decision result of a synchronization sequence of N frames of signals;

[0032] Perform autocorrelation calculations on N groups of synchronization sequences respectively to obtain N autocorrelation peaks;

[0033] Compare N autocorrelation peaks, select a group of synchronization sequences with the largest autocorrelation peak as the initially determined local synchronization sequence, and use the initially determined local synchronization sequence for subsequent synchronization;

[0034] Save the maximum autocorrelation peak value R among N groups of autocorrelation peak values peak , the maximum autocorrelation peak R peak Used for subsequent iterative updates of local synchronization sequences.

[0035] Optionally, the steps of performing symbol synchronization and optimal sampling point synchronization include:

[0036] Set the sampling rate of the non-cooperative communication party receiver to f s , the signal bandwidth is B, the number of subcarriers in the OFDM system is nfft, and the synchronization sequence length is nfft;

[0037] For sampling rate f s The input data is downsampled to rate B;

[0038] According to the signal-to-noise ratio estimation result of the previous frame signal, the correlation length n of this synchronization is adaptively configured, where n≤nfft;

[0039] The downsampled data is correlated and synchronized with the first n points of the local synchronization sequence, and the position of the autocorrelation peak is the symbol timing starting point;

[0040] For sampling rate f s The input data is downsampled to obtain the remaining Group downsampled data;

[0041] Use the symbol timing starting point to locate the remaining The symbol timing starting point of the group down-sampled data;

[0042] Get The first n sampling points of the group down-sampled data are correlated and synchronized with the first n points of the local synchronization sequence respectively;

[0043] Compare The autocorrelation peak values ​​obtained for each group of data, the group with the largest autocorrelation peak value corresponds to the best sample value data.

[0044] Optionally, the step of iteratively updating the local synchronization sequence includes:

[0045] Extracting the synchronization sequence of the signal after time-frequency synchronization;

[0046] Making hard decisions on the extracted synchronization sequence;

[0047] Perform normalization on the synchronization sequence after hard decision, perform modulus calculation based on the normalization decision result, and obtain the autocorrelation peak value R′ peak

[0048] Compare the autocorrelation peak R′ peak The maximum autocorrelation peak R peak ;

[0049] If R p ' eak >R peak , update R peak and local synchronization sequence, ending this round of iteration; if R′ peak ≤R peak , ending this round of iteration.

[0050] The present invention has the following beneficial effects:

[0051] The present invention provides a time-frequency synchronization method for a non-cooperative OFDM communication system, which uses a blind synchronization algorithm based on cyclic prefix autocorrelation to perform initial frame synchronization on a received signal to achieve frame header positioning, uses a sliding correlation window to locate the time domain position of a synchronization sequence of the received signal, performs clustering judgment on the synchronization sequence in a continuous multi-frame signal, and extracts the initial synchronization sequence through statistics of the autocorrelation peak values ​​of the multi-frame judgment results. According to the autocorrelation characteristics and cyclic characteristics of the synchronization sequence, the present invention provides a fast symbol synchronization and optimal sampling point synchronization method, and realizes frequency offset estimation based on the Moose algorithm. After the present invention performs time-frequency synchronization on the received signal, the synchronized sequence is fed back to the clustering judgment step to perform iterative update of the local synchronization sequence, thereby further improving the overall system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is an overall flow chart of a time-frequency synchronization method for a non-cooperative orthogonal frequency division multiplexing communication system provided in Embodiment 1 of the present invention.

[0053] Figure 2 A schematic diagram of an orthogonal frequency division multiplexing system frame structure provided in Embodiment 1 of the present invention.

[0054] Figure 3 This is a flow chart of a low-complexity timing synchronization method provided in Embodiment 1 of the present invention.

[0055] Figure 4 This is a schematic diagram of a timing accuracy and iteration number curve under different signal-to-noise ratio conditions provided in the first embodiment of the present invention.

[0056] Figure 5 A schematic diagram of the timing accuracy of two methods under three frequency deviations provided in the first embodiment of the present invention.

[0057] Figure 6 A schematic diagram of mean square error of frequency offset estimation of two methods under three frequency offsets provided in embodiment 1 of the present invention.

[0058] Figure 7 A schematic diagram of the algorithm time complexity of two types of methods under three cyclic prefix lengths provided in Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0059] In order to enable those skilled in the art to better understand the technical solution of the present invention, the time-frequency synchronization method for a non-cooperative OFDM communication system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0060] Embodiment 1

[0061] 1. Overview

[0062] This embodiment proposes a time-frequency synchronization method combined with synchronization sequence estimation for a non-cooperative OFDM system. The overall process is as follows: Figure 1 As shown. First, the CP autocorrelation blind synchronization algorithm is used to perform the initial frame synchronization on the signal to realize frame header positioning; secondly, the sliding correlation window is used to locate the time domain position of the synchronization sequence of the received signal; then, the synchronization sequence in the continuous multi-frame signal is clustered and determined, and a more accurate synchronization sequence is determined by the autocorrelation peak statistics of the multi-frame decision results; then, based on the good autocorrelation characteristics and cyclic characteristics of the synchronization sequence, this embodiment designs a fast symbol synchronization and optimal sampling point synchronization method, and realizes frequency offset estimation based on the Moose algorithm. After the received signal is synchronized in time and frequency, the synchronized sequence is fed back to the clustering module and the local synchronization sequence is iteratively updated to further improve the overall system performance. In the figure, the CP autocorrelation blind synchronization and sliding window positioning in the synchronization sequence extraction module in the dotted box need to be executed only at the beginning of blind reception. Once the synchronization sequence is initially determined, the received signal skips this part and directly enters the time and frequency synchronization module.

[0063] The key modules of the proposed method are described below, including: synchronization sequence extraction module and time-frequency synchronization module.

[0064] 2. Synchronous sequence extraction

[0065] In order to achieve subsequent fast time-frequency synchronization of OFDM signals, this module first performs CP-related blind synchronization on the signal on the premise that multiple parameters such as OFDM system bandwidth, subcarrier spacing, FFT points and CP length have been correctly obtained; then the synchronization sequence existing in the signal is located by sliding correlation window to determine its time domain position; finally, the synchronization sequence in multi-frame signals is extracted using the positioning results, and clustering is used to make decisions and compare the autocorrelation peaks of multi-frame decision results to achieve the extraction of the initial synchronization sequence, providing a basis for low-complexity time-frequency synchronization of subsequent signals and iterative updating of the synchronization sequence.

[0066] 2.1 Synchronous sequence position identification

[0067] In existing OFDM communication systems, such as LTE and 5G, the physical layer usually uses time domain correlation for timing synchronization. Specifically, when sending a signal, the sender inserts a synchronization sequence known to both the sender and the receiver at a specific time-frequency position. When receiving the signal, the receiver uses the local synchronization sequence to perform correlation calculations on the received sequence and selects the correlation peak point as the synchronization starting point.

[0068] Usually, the frame structure design of OFDM communication system is as follows: Figure 2 As shown:

[0069] The horizontal axis is the time domain OFDM symbol, the vertical axis is the frequency domain subcarrier, and PSS is the synchronization sequence described above. The synchronization sequence generally has good autocorrelation characteristics, and some communication systems also have cyclic characteristics. In different systems, the time domain symbol position of the synchronization sequence and the number of subcarriers occupied may be different. This section uses the good autocorrelation characteristics of the synchronization sequence to design a synchronization sequence position identification method based on a sliding correlation window.

[0070] The specific method flow of synchronization sequence position identification is as follows:

[0071] (1) Blind CP autocorrelation synchronization

[0072] Based on the correct estimation of multiple parameters such as OFDM system bandwidth, subcarrier spacing, FFT points and CP length, the signal frame header is first located using blind CP autocorrelation.

[0073] (2) Sliding correlation window to locate the synchronization sequence position

[0074] Assume that the number of Fast Fourier Transform (FFT) points of OFDM signal is nfft, the length of CP is cp_len, and each frame contains frame_len OFDM symbols. After locating the signal frame head using CP in the previous step, according to the strong autocorrelation of synchronization sequence, the following method can be used to locate the synchronization sequence in a frame of OFDM signal:

[0075] a) Set a sliding window with a window length of nfft+cp_len starting from the frame header;

[0076] b) calculating the autocorrelation peak of the elements in the sliding window;

[0077] c) The window slides backward by nfft+cp_len, and operation b is repeated;

[0078] d) Compare frame_len correlation peaks, where the position with the largest correlation peak is the symbol where the synchronization sequence is located.

[0079] 2.2 Synchronous sequence clustering decision

[0080] After locating the synchronization sequence in the previous step, the receiving end can extract the synchronization sequence of the signal. In the actual OFDM communication system, in order to ensure that the receiving side can correctly synchronize the signal in time and frequency, the synchronization sequence often uses low-order modulation methods such as BPSK or QPSK. In order to reduce the impact of hard decision misjudgment caused by the distortion of the synchronization sequence constellation diagram due to factors such as channel noise, multipath effects, receiver sampling errors, and timing errors, this section will use the constellation diagram clustering method of multi-frame signals to realize synchronization sequence judgment. The specific process is as follows:

[0081] a) determining a K value according to a modulation order of the synchronization sequence, wherein K corresponds to the number of ideal constellation points;

[0082] b) randomly selecting the constellation coordinates of K data points from the synchronization sequence as the initial centroid;

[0083] c) Assign constellation points to clusters. Calculate the Euclidean distance between each constellation point in the synchronization sequence and each centroid, and then assign each point to the cluster represented by the nearest centroid;

[0084] d) Update the centroid. For each cluster, calculate the mean of all constellation points and set the mean as the new centroid;

[0085] e) Iteration process. Repeat steps c and d until a stopping condition is met, such as the change in the centroid in consecutive iterations is less than a certain threshold, or the maximum number of iterations is reached.

[0086] After the above operations, the synchronization sequence constellation points have been classified into K categories, and the synchronization sequence constellation points can be determined according to the synchronization sequence modulation order.

[0087] 2.3 Initial synchronization sequence determination

[0088] After locating and extracting the synchronization sequence of continuous multi-frame signals and performing clustering judgment, it is necessary to select the best result from multiple judgment results for subsequent time-frequency synchronization. This embodiment proposes a method for determining the initial synchronization sequence based on autocorrelation peak comparison. The specific process is as follows:

[0089] a) using clustering to obtain the normalized decision result of N-frame signal synchronization sequence;

[0090] b) performing autocorrelation calculations on N groups of synchronization sequences respectively to obtain N autocorrelation peaks;

[0091] c) comparing N autocorrelation peaks, selecting a group with the largest peak as the initially determined local synchronization sequence, and using it for subsequent synchronization;

[0092] d) Save the maximum autocorrelation peak value R among N groups of autocorrelation peak values peak , used for subsequent iterative updates of the local synchronization sequence.

[0093] 3. Low-complexity time-frequency synchronization based on iterative update of synchronization sequences

[0094] 3.1 Timing Synchronization

[0095] After obtaining the OFDM system synchronization sequence, this embodiment designs a fast timing synchronization method based on correlation synchronization, such as Figure 3 shown.

[0096] Assume that the sampling rate of the non-cooperative communication party receiver is f s , the signal bandwidth is B, the number of subcarriers in the OFDM system is nfft, and the length of the synchronization sequence PSS is nfft. Figure 3 The timing synchronization process shown can be described as follows:

[0097] a) for f s The sampling rate input data is downsampled to rate B;

[0098] b) Adaptively configure the correlation length n of this synchronization according to the signal-to-noise ratio estimation result of the previous frame signal (n≤nfft);

[0099] c) Correlate and synchronize the downsampled data with the first n points of the local PSS sequence, and the correlation peak position is the symbol timing starting point;

[0100] d) for f s The sampling rate input data is downsampled to obtain the remaining Group downsampled data;

[0101] e) Using the symbol timing starting point obtained in c), quickly locate the remaining The symbol timing starting point of the group down-sampled data;

[0102] f) Take The first n sampling points of the group downsampled data are correlated with the first n points of the local PSS respectively;

[0103] g) Comparison The peak-to-peak value of the correlation obtained for each group of data, the group with the highest peak value corresponds to the best sample value data, thereby achieving symbol synchronization and optimal sampling synchronization.

[0104] The above timing synchronization method takes into account the downsampling The group data has the same symbol starting position. Based on a group of downsampled data, the symbol timing starting point is obtained through a correlation synchronization and applied to the remaining subsequent In addition, this method uses the signal-to-noise ratio estimation result of the previous frame signal to propose an adaptive correlation length configuration. When the signal-to-noise ratio of the previous frame signal is high, a low correlation length can be selected as the correlation point number for the current frame synchronization, thereby greatly reducing the algorithm complexity.

[0105] 3.2 Frequency Synchronization

[0106] After correct frame synchronization and optimal sample point selection, the non-cooperative communication party can accurately locate the PSS sequence in the received signal. The subsequent frequency offset estimation will be performed using the Moose algorithm based on the training sequence, in which the PSS sequence extracted from this frame will be used as the training sequence input of the algorithm. The details of the algorithm are not described here.

[0107] 3.3 Synchronous Sequence Iteration Update

[0108] Although Section 2.3 has made an initial estimate of the synchronization sequence, the estimation result may be wrong in the actual channel environment. In order to further improve the estimation accuracy of the synchronization sequence, this section proposes an iterative update method for the synchronization sequence. After the correct timing synchronization and frequency offset estimation of the current frame signal in Sections 3.1 and 3.2, this method will extract and judge the PSS sequence after time-frequency synchronization, and input the autocorrelation peak value into the clustering decision multi-frame statistics module again. By comparing the correlation peak value, the iterative update of the local synchronization sequence is completed and applied to the time-frequency synchronization of the next frame signal. The specific method flow is as follows:

[0109] a) extract the synchronization sequence PSS of the signal after time-frequency synchronization of this frame;

[0110] b) making a hard decision on the extracted PSS;

[0111] c) Perform modulus calculation on the PSS normalized decision result in b) to obtain the correlation peak value R p ' eak ;

[0112] d) Compare R p ' eak The previous frame R saved by the clustering decision multi-frame statistics module peak ;

[0113] e) If R p ' eak >R peak , update R peak And the local synchronization sequence, ending this round of iteration;

[0114] f) If R p ' eak ≤R peak , ending this round of iteration.

[0115] 4. Simulation Results

[0116] The OFDM system symbol rate is set to 60Msps, the receiver sampling rate is 480MHz, the number of subcarriers is 1024, the CP length is 16 / 32 / 48, the PSS sequence is a 1024-point time domain M sequence and QPSK modulation is used. To verify the rationality and effectiveness of the method of this embodiment, this embodiment conducts the following three sets of simulation experiments:

[0117] 1. Simulation of the number of synchronization sequence iteration updates and the accuracy of the local synchronization sequence under seven conditions: frequency deviation of 10kHz and signal-to-noise ratio of 1, 2, 3, 4, 5, 7, and 9dB;

[0118] 2. Under the conditions of signal-to-noise ratio 1-20dB, CP length 48 and different frequency offsets, the timing accuracy and frequency offset estimation MSE of the traditional CP-based maximum likelihood time-frequency synchronization method and the overall method of this embodiment are compared and simulated;

[0119] 3. Under different CP lengths, a comparative simulation of the time complexity of time-frequency synchronization of 1-100 frames of signal between the traditional CP-based maximum likelihood time-frequency synchronization method and the method of this embodiment is performed.

[0120] 4.1 Simulation Results 1

[0121] The synchronization sequence accuracy-iteration number curves of the synchronization sequence iterative update algorithm proposed in this embodiment under the conditions of 10kHz frequency deviation and signal-to-noise ratio of 1, 2, 3, 4, 5, 7, and 9dB are as follows: Figure 4 As shown in the figure. Under the condition of the same number of iterations, the accuracy of the synchronization sequence increases with the increase of the signal-to-noise ratio. In addition, under the same signal-to-noise ratio, although the estimation accuracy of adjacent iterations may remain unchanged, as the number of iterations increases, the accuracy of the local synchronization sequence shows an overall upward trend. The simulation results successfully verify the effectiveness of the synchronization sequence iterative update algorithm proposed in this embodiment.

[0122] 4.2 Simulation Results 2

[0123] Under the frequency deviation setting conditions of 10kHz, 15kHz and 20kHz, the timing synchronization accuracy-SNR curves of the time-frequency synchronization method proposed in this embodiment and the traditional CP-based maximum likelihood time-frequency synchronization algorithm are as follows: Figure 5 As shown. Figure 5 It can be seen that the time-frequency synchronization method proposed in this embodiment can accurately reconstruct the optimal sample point position under three frequency offset conditions and a signal-to-noise ratio range of 1-20dB, and the timing performance is better than the traditional CP-based maximum likelihood time-frequency synchronization algorithm.

[0124] Figure 6 When the signal-to-noise ratio changes, the frequency offset estimation result finally obtained by the overall method of this embodiment and the frequency offset estimation result based on the traditional CP method are given. Figure 6 It can be seen that the method proposed in this embodiment has better frequency offset estimation performance than the blind CP method; the frequency offset estimation MSE errors of the two methods decrease with the increase of the signal-to-noise ratio. However, for the blind CP method, the high frequency offset has a greater impact on the accuracy of the frequency offset estimation under low signal-to-noise ratio conditions, while the frequency offset estimation performance of the method proposed in this embodiment is almost unaffected by the size of the frequency offset.

[0125] 4.3 Simulation Results 3

[0126] Under the three CP lengths of 16, 32, and 48, Figure 7The time taken by the method proposed in this embodiment and the traditional CP-based method to process 1-100 frames of data is compared. Obviously, the method proposed in this embodiment has better time complexity performance, and as the CP length increases, the complexity of the traditional method increases significantly, while the performance of the method proposed in this embodiment is almost unaffected by the CP length.

[0127] 5. Conclusion

[0128] This embodiment proposes a low-complexity time-frequency synchronization method for a non-cooperative communication OFDM system based on synchronization sequence detection and extraction. The signal is initially frame synchronized by a blind synchronization algorithm based on CP autocorrelation, and then the synchronization sequence position of the received signal is located by a sliding correlation window, and then a more accurate local synchronization sequence is obtained by clustering judgment and multi-frame correlation peak statistics. After obtaining the synchronization sequence, this embodiment designs an adaptive correlation length fast symbol synchronization and optimal sampling point synchronization method based on correlation synchronization based on the good autocorrelation characteristics and cyclic characteristics of the synchronization sequence, and uses the Moose algorithm based on the synchronization sequence to realize frequency deviation estimation. After the signal is correctly synchronized in time and frequency, this method re-judges the received synchronization sequence and designs a corresponding algorithm to iteratively update the local synchronization sequence to improve the accuracy of the local synchronization sequence and apply it to the next frame time and frequency synchronization. Finally, this embodiment designs three sets of comparative experiments to verify the effectiveness and good time complexity performance of the time and frequency synchronization method proposed in this embodiment.

[0129] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A time-frequency synchronization method for a non-cooperative OFDM communication system, characterized in that: include: Step S1, using a blind synchronization algorithm based on cyclic prefix autocorrelation to perform initial frame synchronization on a received signal to achieve frame header positioning of the received signal; Step S2, using a sliding correlation window to locate the time domain position of the synchronization sequence of the received signal; Step S3, extracting synchronization sequences in the multi-frame signals according to the positioning results; Step S4, using a constellation diagram clustering method to perform clustering judgment on the synchronization sequence in the multi-frame signal, performing autocorrelation peak comparison of the multi-frame judgment results, and counting and extracting the initial synchronization sequence; Step S5, using the initial synchronization sequence to perform time-frequency synchronization on subsequent received signals; After step S5 is executed, the sequence after time-frequency synchronization is fed back to step S4, and step S4 is executed to iteratively update the local synchronization sequence.

2. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 1, characterized in that: The step S5 comprises: Perform symbol synchronization and optimal sampling point synchronization; The Moose algorithm is used to estimate the frequency offset of the sequence after symbol synchronization and optimal sampling point synchronization are achieved.

3. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 2, characterized in that: The step S1 includes: Signal parameters are obtained, where the signal parameters include an orthogonal frequency division multiplexing system bandwidth, a subcarrier spacing, a fast Fourier transform point number, and a cyclic prefix length.

4. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 3, characterized in that: The step S2 comprises: Step S21, setting the number of fast Fourier transform points of the OFDM signal to nfft, the cyclic prefix length to cp_len, and each frame to contain frame_len OFDM symbols; Step S22, setting a sliding correlation window with a window length of nfft+cp_len with the frame header as the starting point; Step S23, calculating the autocorrelation peak value of the elements in the sliding correlation window; Step S24, slide the sliding correlation window backward by nfft+cp_len, and repeat step S23; Step S25: compare frame_len autocorrelation peaks, and determine that the position with the largest autocorrelation peak is the symbol where the synchronization sequence is located.

5. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 4, characterized in that: The step of using the constellation diagram clustering method to perform clustering decision on the synchronization sequence in the multi-frame signal comprises: Step S41, determining a K value according to a synchronization sequence modulation order, wherein K corresponds to the number of ideal constellation points; Step S42, randomly selecting constellation coordinates of K data points from the synchronization sequence as initial centroids; Step S43, calculating the Euclidean distance between each constellation point in the synchronization sequence and each centroid, and assigning each constellation point to the cluster represented by the centroid with the shortest Euclidean distance; Step S44: For each cluster, calculate the mean of all constellation points, and set the mean as a new centroid; Repeat step S43 and step S44 until a stop condition is met, wherein the stop condition is that the change of the centroid in consecutive iterations is less than a preset threshold, or a maximum number of iterations is reached; Step S45: The synchronization sequence constellation points have been classified into K categories, and the synchronization sequence constellation points are determined according to the synchronization sequence modulation order.

6. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 5, characterized in that: The steps of comparing the autocorrelation peak values ​​of the multi-frame decision results, counting and extracting the initial synchronization sequence include: Obtaining a normalized decision result of a synchronization sequence of N frames of signals; Perform autocorrelation calculations on N groups of synchronization sequences respectively to obtain N autocorrelation peaks; Compare N autocorrelation peaks, select a group of synchronization sequences with the largest autocorrelation peak as the initially determined local synchronization sequence, and use the initially determined local synchronization sequence for subsequent synchronization; Save the maximum autocorrelation peak value R among N groups of autocorrelation peak values peak , the maximum autocorrelation peak R peak Used for subsequent iterative updates of local synchronization sequences.

7. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 6, characterized in that: The steps of performing symbol synchronization and optimal sampling point synchronization include: Set the sampling rate of the non-cooperative communication party receiver to f s , the signal bandwidth is B, the number of subcarriers in the OFDM system is nfft, and the synchronization sequence length is nfft; For sampling rate f s The input data is downsampled to rate B; According to the signal-to-noise ratio estimation result of the previous frame signal, the correlation length n of this synchronization is adaptively configured, where n≤nfft; The downsampled data is correlated and synchronized with the first n points of the local synchronization sequence, and the position of the autocorrelation peak is the symbol timing starting point; For sampling rate f s The input data is downsampled to obtain the remaining Group downsampled data; Use the symbol timing starting point to locate the remaining The symbol timing starting point of the group down-sampled data; Get The first n sampling points of the group down-sampled data are correlated and synchronized with the first n points of the local synchronization sequence respectively; Compare The autocorrelation peak values ​​obtained for each group of data, the group with the largest autocorrelation peak value corresponds to the best sample value data.

8. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 7, characterized in that: The step of iteratively updating the local synchronization sequence comprises: Extracting the synchronization sequence of the signal after time-frequency synchronization; Making hard decisions on the extracted synchronization sequence; Perform normalization on the synchronization sequence after hard decision, perform modulus calculation based on the normalization decision result, and obtain the autocorrelation peak value R′ peak Compare the autocorrelation peak R′ peak The maximum autocorrelation peak R peak ; If R p ' eak >R peak , update R peak and local synchronization sequence, ending this round of iteration; if R′ peak ≤R peak , ending this round of iteration.

Images