A time-frequency synchronization method for a non-cooperative OFDM communication system

Abstract

The application discloses a time-frequency synchronization method of a non-cooperative OFDM communication system, a blind synchronization algorithm of cyclic prefix autocorrelation is used to perform initial frame synchronization on a received signal to realize frame header positioning, a sliding correlation window is used to position a time domain position of a synchronization sequence of the received signal, clustering decision is performed on the synchronization sequence in continuous multiple frames of signals, and an initial synchronization sequence is extracted through autocorrelation peak value statistics of multiple frame decision results. According to autocorrelation characteristics and cycle characteristics of the synchronization sequence, the application provides a fast symbol synchronization and optimal sampling point synchronization method, and frequency offset estimation is realized based on a Moose algorithm. After the received signal is subjected to time-frequency synchronization, the synchronized sequence is fed back to the clustering decision step to perform iterative updating on a local synchronization sequence, and the overall system performance is further improved.

Description

Technical Field

[0001] This 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 Technology

[0002] Non-cooperative communication refers to a communication method in which a third party, acting without authorization, accesses a cooperative communication system to conduct radio awareness and reconnaissance. It is currently widely used in both military and civilian fields. In civilian communications, to ensure the rational use of spectrum and a healthy communication environment, relevant regulatory authorities must monitor and control free-space signals. This requires third parties to blindly identify communication signals without prior information. In military communications, non-cooperative parties need to correctly identify and demodulate unknown signals to achieve monitoring and reconnaissance of communication signals.

[0003] Orthogonal Frequency Division Multiplexing (OFDM) technology has become a mainstay of modern communication systems due to its advantages such as strong resistance to multipath and interference, high spectrum utilization, and high-speed data transmission. Currently, OFDM technology is widely used in 4G and 5G mobile communications as well as satellite communications; therefore, research on non-cooperative blind reception of OFDM is of great significance.

[0004] Time-frequency synchronization is a key technology on the receiving side of a communication system. Correct time-frequency synchronization is the foundation and prerequisite for subsequent channel estimation, channel equalization, and proper demodulation of the received signal. For non-cooperative OFDM systems, existing typical blind time-frequency synchronization methods include: maximum likelihood algorithms based on cyclic prefixes (CP), algorithms based on signal cyclostationarity, and algorithms based on autocorrelation matrices. However, existing methods are often highly complex and their performance is limited when facing complex channel environments. Summary of the Invention

[0005] To address the limitations and defects of existing technologies, this invention provides a time-frequency synchronization method for a non-cooperative OFDM communication system, comprising:

[0006] Step S1: Use the blind synchronization algorithm of cyclic prefix autocorrelation to perform initial frame synchronization of the received signal, thereby realizing the frame header positioning of the received signal;

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

[0008] Step S3: Extract the synchronization sequence from the multi-frame signal based on the positioning results;

[0009] Step S4: Use constellation diagram clustering method to perform clustering decision on the synchronization sequence in the multi-frame signal, compare the autocorrelation peaks of the multi-frame decision results, and statistically extract the initial synchronization sequence.

[0010] Step S5: Use the initial synchronization sequence to perform time-frequency synchronization on the subsequently received signals;

[0011] After completing step S5, the time-frequency synchronized sequence 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.

[0015] Optionally, the steps preceding step S1 include:

[0016] Obtain signal parameters, including the orthogonal frequency division multiplexing system bandwidth, subcarrier spacing, number of fast Fourier transform points, and cyclic prefix length.

[0017] Optionally, step S2 includes:

[0018] Step S21: Set the number of Fast Fourier Transform points of the Orthogonal Frequency Division Multiplexing (OFDM) signal to nfft, the cyclic prefix length to cp_len, and each frame to contain frame_len OFDM symbols.

[0019] Step S22: Set a sliding correlation window with a window length of nfft+cp_len, starting from the frame header;

[0020] Step S23: Calculate the autocorrelation peak value of the elements within the sliding correlation window;

[0021] Step S24: Slide the sliding related window backward by nfft+cp_len, and repeat step S23;

[0022] Step S25: Compare the frame_len autocorrelation peaks and determine the position of the largest autocorrelation peak as the symbol containing the synchronization sequence.

[0023] Optionally, the step of using the constellation graph clustering method to perform clustering decisions on the synchronization sequences in the multi-frame signal includes:

[0024] Step S41: Determine the value of K based on the modulation order of the synchronization sequence, where K corresponds to the number of points in the ideal constellation;

[0025] Step S42: Randomly select K data points from the synchronization sequence as the initial centroids;

[0026] Step S43: Calculate the Euclidean distance between each constellation point and each centroid in the synchronization sequence, and assign each constellation point to the cluster represented by the centroid with the nearest Euclidean distance;

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

[0028] Repeat steps S43 and S44 until a stopping condition is met, wherein the change in the centroid during continuous iterations is less than a preset threshold, or the maximum number of iterations is reached.

[0029] Step S45: The synchronization sequence constellation points have been classified into class K. The synchronization sequence constellation points are then judged according to the modulation order of the synchronization sequence.

[0030] Optionally, the step of comparing the autocorrelation peaks of multi-frame decision results and statistically analyzing and extracting the initial synchronization sequence includes:

[0031] Obtain the normalized decision result of the N-frame signal synchronization sequence;

[0032] Autocorrelation calculations were performed on N sets of synchronization sequences to obtain N autocorrelation peaks.

[0033] Compare N autocorrelation peaks, select the synchronization sequence 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 from N sets of autocorrelation peak values. peak The maximum autocorrelation peak value R peak Used for iterative updates of the subsequent local synchronization sequence.

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

[0036] Set the sampling rate of the non-cooperative communication party's receiver to f. s The signal bandwidth is B, the number of subcarriers in the orthogonal frequency division multiplexing system is nfft, and the synchronization sequence length is nfft.

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

[0038] Based on 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, with the position of the correlation peak as the symbol timing start point.

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

[0041] Use the symbol timing start point to locate the rest. The sign timing start point of the data after group downsampling;

[0042] Get The first n sampled points of the downsampled data are respectively synchronized with the first n points of the local synchronization sequence;

[0043] Compare The group of data with the largest autocorrelation peak value corresponds to the optimal sample data.

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

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

[0046] Perform hard decision on the extracted synchronization sequence;

[0047] Normalize the synchronization sequence following the hard decision, and calculate the autocorrelation peak R′ based on the modulo operation of the normalized decision result. peak

[0048] Compare the autocorrelation peak value R′ peak With the maximum autocorrelation peak value R peak ;

[0049] If R p ′ eak >R peak Update R peak And the local synchronization sequence, ending this iteration; if R′ peak ≤R peak This concludes the current iteration.

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

[0051] This invention provides a time-frequency synchronization method for a non-cooperative OFDM communication system. It employs a blind synchronization algorithm based on cyclic prefix autocorrelation to perform initial frame synchronization of the received signal, achieving frame header localization. A sliding correlation window is used to locate the time-domain position of the synchronization sequence in the received signal. Clustering decisions are made on the synchronization sequences across multiple consecutive frames. The initial synchronization sequence is extracted by statistically analyzing the autocorrelation peaks of the multi-frame decision results. Based on the autocorrelation and cyclic characteristics of the synchronization sequence, this invention provides a fast symbol synchronization and optimal sampling point synchronization method, and frequency offset estimation is achieved based on the Moose algorithm. After time-frequency synchronization of the received signal, the synchronized sequence is fed back to the clustering decision step for iterative updates of the local synchronization sequence, further improving the overall system performance. Attached Figure Description

[0052] Figure 1 The above is a flowchart of the 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 This is a schematic diagram of the frame structure of an orthogonal frequency division multiplexing system provided in Embodiment 1 of the present invention.

[0054] Figure 3 The flowchart shows the low-complexity timing synchronization method provided in Embodiment 1 of the present invention.

[0055] Figure 4 This is a schematic diagram of the timing accuracy versus iteration count curves under different signal-to-noise ratio conditions provided in Embodiment 1 of the present invention.

[0056] Figure 5 This is a schematic diagram illustrating the timing accuracy of two types of methods under three frequency offsets provided in Embodiment 1 of the present invention.

[0057] Figure 6 This is a schematic diagram of the mean square error of frequency offset estimation for two types of methods under three frequency offset conditions provided in Embodiment 1 of the present invention.

[0058] Figure 7 This diagram illustrates the time complexity of two methods under three different cyclic prefix lengths provided in Embodiment 1 of the present invention. Detailed Implementation

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

[0060] Example 1

[0061] 1. Overview

[0062] This embodiment proposes a time-frequency synchronization method combining synchronization sequence estimation for non-cooperative OFDM systems. The overall process is as follows: Figure 1 As shown in the figure, firstly, a blind synchronization algorithm based on CP autocorrelation is used to perform initial frame synchronization of the signal to achieve frame header localization; secondly, a sliding correlation window is used to locate the time-domain position of the synchronization sequence of the received signal; then, clustering decisions are made on the synchronization sequences in multiple consecutive frames, and a more accurate synchronization sequence is determined by statistical analysis of the autocorrelation peaks of the multi-frame decision results; subsequently, based on the good autocorrelation and cyclic characteristics of the synchronization sequence, this embodiment designs a fast symbol synchronization and optimal sampling point synchronization method, and implements frequency offset estimation based on the Moose algorithm. After time-frequency synchronization of the received signal, the synchronized sequence is fed back to the clustering module for iterative updating of the local synchronization sequence, thereby further improving the overall system performance. In this figure, the CP autocorrelation blind synchronization and sliding window localization in the synchronization sequence extraction module within the dashed box only need to be executed at the beginning of blind reception. Once the synchronization sequence is initially determined, the received signal skips this part and directly enters the time-frequency synchronization module.

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

[0064] 2. Synchronization sequence extraction

[0065] To achieve rapid time-frequency synchronization of OFDM signals, this module, after correctly obtaining parameters such as OFDM system bandwidth, subcarrier spacing, FFT points, and CP length, first performs CP-related blind synchronization on the signal. Then, it uses a sliding correlation window to locate the synchronization sequence in the signal and determine its time-domain position. Finally, it uses the location results to extract the synchronization sequence from multiple frames of the signal, uses clustering to make decisions, and compares the autocorrelation peaks of the decision results across multiple frames to achieve the extraction of the initial synchronization sequence. This provides a foundation for low-complexity time-frequency synchronization and iterative updates of the synchronization sequence in subsequent signals.

[0066] 2.1 Synchronization Sequence Position Identification

[0067] Existing OFDM communication systems, such as LTE and 5G, typically use time-domain correlation for timing synchronization at the physical layer. Specifically, when transmitting signals, the sender inserts a synchronization sequence known to both the sender and receiver at a specific time-frequency position. During signal reception, the receiver uses its local synchronization sequence to perform correlation calculations on the received sequence and selects the relevant peak point as the synchronization start point.

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

[0069] The horizontal axis represents the time-domain OFDM symbol, and the vertical axis represents the frequency-domain subcarrier. The PSS is the synchronization sequence described above. Synchronization sequences generally possess good autocorrelation characteristics, and some communication systems also exhibit cyclic characteristics. The position of the synchronization sequence in the time domain symbol and the number of subcarriers it occupies may differ in different systems. This section utilizes the good autocorrelation characteristics of synchronization sequences to design a synchronization sequence position identification method based on a sliding correlation window.

[0070] The specific method and process for synchronous sequence position identification are as follows:

[0071] (1) Blind CP autocorrelation synchronization

[0072] Based on the accurate 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) Positioning the synchronization sequence by sliding the correlation window

[0074] Assuming the OFDM signal has nfft points in its Fast Fourier Transform (FFT), a CP length of cp_len, and each frame contains frame_len OFDM symbols, after locating the signal frame header using the CP in the previous step, based on the strong autocorrelation characteristic of the 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 length of nfft+cp_len, starting from the frame header;

[0076] b) Calculate the autocorrelation peak of the elements within the sliding window;

[0077] c) Slide the window backward by nfft+cp_len, and repeat operation b;

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

[0079] 2.2 Clustering Decision of Synchronous Sequences

[0080] After locating the synchronization sequence in the previous step, the receiver can extract the synchronization sequence of the signal. In practical OFDM communication systems, to ensure that the receiver can correctly synchronize the signal in time and frequency, the synchronization sequence often uses low-order modulation methods such as BPSK or QPSK. 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 a constellation diagram clustering method of multi-frame signals to implement synchronization sequence decision. The specific process is as follows:

[0081] a) Determine the value of K based on the modulation order of the synchronization sequence, where K corresponds to the number of points in the ideal constellation;

[0082] b) Randomly select K data points from the synchronization sequence and use their constellation coordinates as the initial centroids;

[0083] c) Assign constellation points to clusters. Calculate the Euclidean distance between each constellation point in the synchronization sequence and each centroid, 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 that mean as the new centroid;

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

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

[0087] 2.3 Initial Synchronization Sequence Determination

[0088] After locating and extracting the synchronization sequence of consecutive multi-frame signals and performing clustering decisions, the optimal result needs to be selected from multiple decisions 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) Obtain the normalized decision result of the synchronization sequence of N frames of signals using clustering;

[0090] b) Perform autocorrelation calculations on the N sets of synchronization sequences respectively to obtain N autocorrelation peaks;

[0091] c) Compare the N autocorrelation peaks, select the group with the largest peak as the initially determined local synchronization sequence, and use it for subsequent synchronization;

[0092] d) Save the maximum autocorrelation peak value R from the N sets of autocorrelation peak values. peak This is used for subsequent iterative updates of the local synchronization sequence.

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

[0094] 3.1 Scheduled 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 As shown.

[0096] Assume the sampling rate of the receiver of the non-cooperative communication party 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 timed synchronization process shown can be described as follows:

[0097] a) For f s The input data is downsampled to rate B.

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

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

[0100] d) For f s The input data is downsampled to obtain the remaining data. Group downsampling data;

[0101] e) Using the symbol timing start point obtained in c), quickly locate the remaining points in step d). The sign timing start point of the data after group downsampling;

[0102] f) Take The first n sampled points of the downsampled data are correlated with the first n points of the local PSS.

[0103] g) Comparison The peak values ​​of each group of data are obtained, and the group with the highest peak value corresponds to the best sample data, thereby completing symbol synchronization and optimal sampling synchronization.

[0104] The above timing synchronization method takes into account downsampling. Since the data sets share the same symbol start position, the symbol timing start point is obtained through a single correlation synchronization based on a certain set of downsampled data, and then applied to the remaining data sets. The optimal selection of sample points for the dataset is crucial. Furthermore, this method utilizes the signal-to-noise ratio (SNR) estimation results of the previous frame signal to propose an adaptive correlation length configuration. When the SNR of the previous frame signal is high, a low correlation length can be selected as the number of correlation points for synchronization in the current frame, thereby significantly reducing the algorithm complexity.

[0105] 3.2 Frequency Synchronization

[0106] After proper frame synchronization and optimal sample selection, the non-cooperative communication party can accurately locate the PSS sequence in the received signal. Subsequent frequency offset estimation will be performed using the Moose algorithm based on the training sequence, where the PSS sequence extracted from this frame will be used as the training sequence input for this algorithm. The detailed algorithm will not be elaborated here.

[0107] 3.3 Synchronous Sequence Iterative Update

[0108] Although Section 2.3 provided an initial estimate of the synchronization sequence, this estimate may be inaccurate in real-world channel environments. To further improve the accuracy of the synchronization sequence estimation, this section proposes an iterative update method. After correctly timing and estimating the frequency offset of the current frame signal in Sections 3.1 and 3.2, this method extracts and makes decisions from the time-frequency synchronized PSS sequence, and then inputs the autocorrelation peak value back into the clustering decision multi-frame statistics module. By comparing the correlation peak values, iterative updates to the local synchronization sequence are 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) Perform hard decision on the extracted PSS;

[0111] c) Perform modulo 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 iteration;

[0114] f) If R p ′ eak ≤R peak This concludes the current iteration.

[0115] 4. Simulation Results

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

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

[0118] 2. Simulation comparison of timing accuracy and frequency offset estimation MSE between the traditional maximum likelihood time-frequency synchronization method based on CP and the overall method of this embodiment under signal-to-noise ratio of 1-20dB, CP length of 48 and different frequency offset conditions;

[0119] 3. Simulation comparison of the time complexity of the traditional maximum likelihood time-frequency synchronization method based on CP and the method of this embodiment for time-frequency synchronization of 1-100 frames of signal under different CP lengths.

[0120] 4.1 Simulation Results 1

[0121] The synchronization sequence iterative update algorithm proposed in this embodiment shows the synchronization sequence accuracy-iteration number curves under seven conditions: 10kHz frequency offset and signal-to-noise ratios of 1, 2, 3, 4, 5, 7, and 9 dB, as shown below. Figure 4 As shown, under the same number of iterations, the synchronization sequence accuracy increases with the increase of the signal-to-noise ratio (SNR). Furthermore, under the same SNR, although the estimation accuracy of adjacent iterations may remain unchanged, the local synchronization sequence accuracy shows an overall upward trend with the increase of the number of iterations. This simulation result successfully verifies the effectiveness of the synchronization sequence iterative update algorithm proposed in this embodiment.

[0122] 4.2 Simulation Results 2

[0123] Under frequency offset settings of 10kHz, 15kHz, and 20kHz, the timing synchronization accuracy-signal-noise ratio curves of the time-frequency synchronization method proposed in this embodiment and the traditional CP-based maximum likelihood time-frequency synchronization algorithm are shown below. Figure 5 As shown. By 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. Its timing performance is better than that of the traditional maximum likelihood time-frequency synchronization algorithm based on CP.

[0124] Figure 6 When the signal-to-noise ratio changes, the final frequency offset estimation result obtained by the overall method in this embodiment and the frequency offset estimation result based on the traditional CP method are presented. 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 error of both methods decreases with the increase of signal-to-noise ratio. However, for the blind CP method, high frequency offset has a greater impact on the accuracy of 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 magnitude of the frequency offset.

[0125] 4.3 Simulation Results 3

[0126] At CP lengths of 16, 32, and 48, Figure 7The time consumption of the proposed method in this embodiment and the traditional CP-based method for processing 1-100 frames of data are compared. Clearly, the proposed method in this embodiment has superior time complexity performance. Furthermore, as the CP length increases, the complexity of the traditional method increases significantly, while the performance of the proposed method is almost unaffected by the CP length.

[0127] 5. Conclusion

[0128] This embodiment proposes a low-complexity time-frequency synchronization method for non-cooperative OFDM communication systems based on synchronization sequence detection and extraction. Initial frame synchronization is achieved using a blind synchronization algorithm based on CP autocorrelation. Then, a sliding correlation window is used to locate the synchronization sequence position of the received signal. A relatively accurate local synchronization sequence is obtained through clustering decision 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 the good autocorrelation and cyclic properties of the synchronization sequence, and uses the Moose algorithm based on the synchronization sequence to achieve frequency offset estimation. After correctly synchronizing the signal in time and frequency, this method re-determines the received synchronization sequence and designs a corresponding algorithm to iteratively update the local synchronization sequence to improve its accuracy and apply it to the time-frequency synchronization of the next frame. Finally, this embodiment designs three sets of comparative experiments to verify the effectiveness and good time complexity performance of the proposed time-frequency synchronization method.

[0129] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled 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, The method comprises the following steps: Step S1, performing initial frame synchronization on a received signal by using a blind synchronization algorithm based on cyclic prefix autocorrelation, so as to locate a frame header of the received signal; Step S2, locating a time domain position of a synchronization sequence of the received signal by using a sliding correlation window; Step S3, extracting a synchronization sequence from multiple frames of signals according to a locating result; Step S4, performing clustering decision on the synchronization sequence in the multiple frames of signals by using a constellation clustering method, performing autocorrelation peak comparison of multiple frame decision results, and counting and extracting an initial synchronization sequence; Step S5, performing time-frequency synchronization on a subsequent received signal by using the initial synchronization sequence; After step S5 is executed, the sequence after time-frequency synchronization is fed back to step S4, and step S4 is executed to update the local synchronization sequence iteratively; The step of performing clustering decision on the synchronization sequence in the multiple frames of signals by using the constellation clustering method comprises the following steps: Step S41, determining a K value according to a modulation order of the synchronization sequence, wherein K corresponds to a 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 an Euclidean distance between each constellation point in the synchronization sequence and each centroid, and assigning each constellation point to a cluster represented by a centroid with the nearest Euclidean distance; Step S44, calculating a mean value of all constellation points for each cluster, and setting the mean value as a new centroid; Steps S43 and S44 are repeated until a stop condition is met, the stop condition being that a change in 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 are classified into K categories, and the synchronization sequence constellation points are decided according to a modulation order of the synchronization sequence; The step of performing autocorrelation peak comparison of multiple frame decision results, counting and extracting an initial synchronization sequence comprises the following steps: acquiring N a normalized decision result of the frame signal synchronization sequence; respectively, and the self-correlation calculation is performed on the group synchronization sequence to obtain N N self-correlation peak values.​ Comparison N correlation peak, and selecting a group of synchronization sequences with the largest autocorrelation peak as the initially determined local synchronization sequences, and using the initially determined local synchronization sequences for subsequent synchronization. saving N the maximum autocorrelation peak in the set of autocorrelation peaks the maximum autocorrelation peak iterative updates for subsequent local synchronization sequences.

2. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 1, characterized in that, The step S5 comprises the following steps: Performing symbol synchronization and best sampling point synchronization; Performing frequency offset estimation on the sequence after symbol synchronization and best sampling point synchronization by using a Moose algorithm.

3. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 2, characterized in that, The step S1 comprises the following steps: Obtaining signal parameters, the signal parameters comprising 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 the following steps: Step S21, setting the fast Fourier transform point number of the orthogonal frequency division multiplexing signal as , the cyclic prefix length as , and each frame containing orthogonal frequency division multiplexing symbols; Step S22, setting the window length as the sliding correlation window with the frame header as the starting point; Step S23, calculating an autocorrelation peak value of elements in the sliding correlation window; Step S24, sliding the sliding correlation window backwards Step S23 is repeated. Step S25, comparison The position of the autocorrelation peak with the largest value is determined as the position of 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 performing symbol synchronization and best sampling point synchronization comprises the following steps: The non-cooperative communication receiver sampling rate is set as , the signal bandwidth is , the number of subcarriers of the OFDM system is , and the length of the synchronization sequence is ; to a sampling rate input data downsampled to a rate ; According to the signal-to-noise ratio estimation result of the previous frame, the correlation length of the current synchronization is adaptively configured ​​ Correlate the down-sampled data with the local synchronization sequence n The position of the auto-correlation peak is the starting point of symbol timing. to the sampling rate downsample the input data to obtain the remaining group downsampled data; positioning the remaining symbol timing starting point of the group downsampled data; acquiring the first n sampling points of the group down-sampled data are respectively correlated with the first n sampling points of the local synchronization sequence Comparison The respective autocorrelation peaks of the groups of data are obtained, and the group having the largest autocorrelation peak corresponds to the best sample data.

6. The time-frequency synchronization method of a non-cooperative OFDM communication system according to claim 5, characterized in that, The step of updating the local synchronization sequence iteratively comprises the following steps: Extracting a synchronization sequence of the signal after time-frequency synchronization; Performing hard decision on the extracted synchronization sequence; The normalized decision is made on the hard decision synchronization sequence, and a modulo calculation is made according to the normalized decision result to obtain an autocorrelation peak value ; comparing the autocorrelation peak values to the maximum autocorrelation peak value ; If , update and local synchronization sequence, end this round of iteration; if , end this round of iteration.

Images