A method for underwater acoustic multicarrier communication based on two-dimensional rotational precoding

The underwater acoustic multicarrier communication method using two-dimensional rotational precoding solves the ICI problem caused by Doppler variation in the underwater acoustic channel in traditional OFDM and OTFS, achieving efficient information transmission in the underwater acoustic channel and enhancing the signal's anti-interference capability.

CN123122691APending Publication Date: 2026-11-17JIMEI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611401548.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-10
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In shallow sea and nearshore underwater acoustic channels, the Doppler variation of traditional OFDM is not negligible, which leads to a decrease in subcarrier orthogonality and the generation of ICI. Existing OTFS methods have limited degrees of freedom to adapt to different motion states and channel spread patterns.

Method used

A two-dimensional rotational precoding-based underwater acoustic multicarrier communication method is adopted. The time and frequency dimensions of rotation matrices are constructed using the chirped phase matrix and the unitary DFT matrix. The data matrix is ​​then subjected to two-dimensional rotational precoding, which diffuses a single information symbol to multiple effective subcarriers and multiple consecutive data moments. Synchronization, time scale correction, and MMSE equalization are performed at the receiving end to recover the source data.

Benefits of technology

While maintaining the existing pilot, cyclic prefix, and frequency domain equalization links, an adjustable two-dimensional rotating orthogonal layer was added to reduce the impact of local deep fading or ICI, thereby improving the reliability and anti-interference capability of signal transmission.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122691A_ABST
    Figure CN123122691A_ABST
Patent Text Reader

Abstract

This application discloses an underwater acoustic multicarrier communication method based on two-dimensional rotation precoding. The transmitting end encodes and modulates source data to obtain a data matrix; it constructs time- and frequency-dimensional rotation matrices using a chirped phase matrix and a unitary DFT matrix; it performs two-dimensional rotation precoding on the data matrix based on the time- and frequency-dimensional rotation matrices to diffuse a single information symbol to multiple effective subcarriers and multiple consecutive data moments according to a rotational orthogonal basis; it loads the precoding result into a multicarrier frequency domain frame and inserts pilot signals to obtain transmitted data; the receiving end performs synchronization, time-scale correction, pilot channel estimation, and MMSE equalization to obtain an equalized two-dimensional matrix, and recovers the source data based on the equalized two-dimensional matrix; it adds an adjustable two-dimensional rotational orthogonal layer while maintaining the existing frequency domain equalization link essentially unchanged; and it uses a rotational basis to diffuse a single information symbol to multiple time-frequency resources, which can reduce the concentrated impact of local deep fading or local ICI on a single symbol.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of communications, and more particularly to an underwater acoustic multicarrier communication method based on two-dimensional rotational precoding. Background Technology

[0002] Shallow and nearshore underwater acoustic channels typically exhibit long multipath propagation, narrowband Doppler shift, and wideband time-scale scaling simultaneously. Traditional OFDM relies on cyclic prefixes to suppress delay spread, but information symbols remain concentrated in local time-frequency resources; when Doppler variations become non-negligible within a frame, subcarrier orthogonality decreases, leading to inter-carrier interference (ICI). OTFS-type methods diffuse symbol energy through two-dimensional transforms, but the fixed Fourier map has limited degrees of freedom for adapting to different motion states and channel spread patterns. Summary of the Invention

[0003] The following is an overview of the topics described in detail in this article.

[0004] The purpose of this application is to at least partially solve one of the technical problems existing in the related technologies. The embodiments of this application provide an underwater acoustic multicarrier communication method based on two-dimensional rotation precoding.

[0005] An embodiment of this application provides an underwater acoustic multicarrier communication method based on two-dimensional rotational precoding, comprising:

[0006] The transmitting end encodes and modulates the source data to obtain a data matrix;

[0007] The time-dimensional rotation matrix and the frequency-dimensional rotation matrix are constructed using the chirped phase matrix and the unitary DFT matrix, respectively.

[0008] The data matrix is ​​precoded in two dimensions by rotation matrices in the time dimension and frequency dimension, so as to spread a single information symbol to multiple effective subcarriers and multiple consecutive data moments according to the rotation orthogonal basis.

[0009] The precoding result is loaded into a multicarrier frequency domain frame and pilot signals are inserted to obtain the transmission data.

[0010] The sending end sends the transmitted data, and the receiving end receives the transmitted data;

[0011] The receiving end performs synchronization, time scale correction, pilot channel estimation, and MMSE equalization on the transmitted data to obtain an equalized two-dimensional matrix, and then recovers the source data based on the equalized two-dimensional matrix.

[0012] According to certain embodiments of this application, source data is encoded and modulated to obtain a data matrix, including:

[0013] The source data is sequentially subjected to optional channel coding, scrambling, block interleaving, and quadrature phase shift keying modulation / quadrature amplitude modulation mapping, and then loaded into a data matrix by column.

[0014] In this data matrix, rows correspond to effective subcarriers, and columns correspond to the orthogonal frequency division multiplexing (OFDM) times of continuous data.

[0015] According to certain embodiments of this application, the rotation matrix in the frequency dimension is represented as:

[0016] ;

[0017] in, Represented as: ;

[0018] Represented as: ;

[0019] In the formula, Let be a rotation matrix in the frequency dimension. Let be the frequency-dimensional chirped phase matrix. Let be a frequency-dimensional unitary DFT matrix. Let be the element in the k-th row and n-th column of the unitary DFT matrix in the frequency dimension. For the frequency dimension of the data, For rotation parameters, diag{·} denotes a diagonal matrix formed by the sequence within the brackets.

[0020] According to certain embodiments of this application, the rotation matrix in the time dimension is represented as:

[0021] ;

[0022] in, Represented as: ;

[0023] Represented as: ;

[0024] In the formula, Let the rotation matrix be in the time dimension. Let be the chirped phase matrix in the time dimension. Let be the time-dimensional unitary DFT matrix. Let be the element in the k-th row and n-th column of the unitary DFT matrix in the time dimension. The time dimension of the data. For rotation parameters, diag{·} denotes a diagonal matrix formed by the sequence within the brackets.

[0025] According to certain embodiments of this application, the precoding result is represented as follows: ;

[0026] In the formula, For the precoding result, For the conjugate transpose of the frequency-dimensional rotation matrix, For data matrix, It is a rotation matrix in the time dimension;

[0027] The two-dimensional rotation precoding only applies to data blocks and does not change the pilot definition.

[0028] According to certain embodiments of this application, parameters The parameters are obtained based on the platform speed, center frequency, and Doppler extension configuration. The same parameters are used to control the tilt angle of the time-frequency basis function, and are applied to both the transmitting and receiving ends. .

[0029] According to certain embodiments of this application, loading the precoding result into a multi-carrier frequency domain frame and inserting pilot signals to obtain transmission data includes:

[0030] The columns of the precoding results are loaded into the effective subcarriers, and known pilots are inserted at fixed intervals;

[0031] Perform Inverse Fast Fourier Transform (IFFT), add a cyclic prefix, upsample, and perform real bandpass modulation on each frequency domain vector;

[0032] A linear frequency modulation (LFM) synchronization segment is inserted before each frame. Multi-frame playback uses an upper frequency sweep template / lower frequency sweep template to obtain the transmission data.

[0033] According to certain embodiments of this application, the LFM complete chain detection of the transmitted signal by the receiving end includes:

[0034] The transmitted signal is normalized and matched with the upper-sweep LFM template and the lower-sweep LFM template respectively to obtain two types of synchronization correlation sequences and their corresponding candidate peak positions;

[0035] Based on the pre-set alternation pattern of up-scanning and down-scanning frequencies at the transmitting end, candidate peaks are combined in a time sequence and filtered group by group.

[0036] When the following conditions are met: the peak amplitude reaches the detection threshold, the relative difference of peaks within the group is within the allowable range, the interval between adjacent peaks conforms to the frame structure constraints, and the number of detected peaks is consistent with the number of frames contained in a complete playback, the peak group is determined as a valid LFM frame chain, and the initial boundary of each frame in this playback is determined.

[0037] Isolated peaks, out-of-order peaks, and incomplete peak groups that do not meet the aforementioned joint conditions are removed.

[0038] According to certain embodiments of this application, the time scale correction includes:

[0039] For the complete LFM chain of the transmitted signal, a coarse estimate of the time scale is obtained based on the observation frame interval, and a set of candidate scales is constructed based on the coarse estimate of the time scale.

[0040] The LFM template or transmitted signal segment is scaled according to the candidate scale and the matching correlation peak of each frame is calculated to form the scale evaluation quantity;

[0041] Select the time scale with the largest scale evaluation quantity;

[0042] When the deviation of the time scale with the largest scale evaluation value relative to 1 exceeds the preset tolerance, the entire received waveform of the transmitted signal is corrected by inverse linear interpolation.

[0043] According to certain embodiments of this application, the MMSE equalization of the effective subcarrier is represented as follows:

[0044] ;

[0045] In the formula, The MMSE equalization output of the k-th effective subcarrier at the m-th data time. For the corresponding received frequency domain symbol, The channel frequency response is obtained by pilot estimation and interpolation. express The complex conjugate, Let be the regularization parameter for the m-th data time.

[0046] The above scheme has at least the following beneficial effects: it adds an adjustable two-dimensional rotating orthogonal layer while keeping the existing pilot, CP, FFT / IFFT and frequency domain equalization links basically unchanged; it uses the rotating basis to diffuse a single information symbol to multiple time-frequency resources, which can reduce the concentrated impact of local deep fading or local ICI on a single symbol; it increases the degree of freedom of the rotation parameters; the strict unitary nature of the matrix avoids linear energy amplification; it uses the data matrix without introducing additional linear energy amplification; and it maintains the existing pilot, CP, FFT / IFFT and frequency domain equalization links, which is convenient to implement through software upgrades on existing underwater acoustic modulation and demodulation platforms. Attached Figure Description

[0047] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0048] Figure 1 This is a flowchart illustrating the steps of an underwater acoustic multicarrier communication method based on two-dimensional rotational precoding.

[0049] Figure 2 This is a schematic diagram of the processing flow at the sending end;

[0050] Figure 3 This is a schematic diagram of the processing flow at the receiving end. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, or the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0053] An embodiment of this application provides a method for underwater acoustic multicarrier communication based on two-dimensional rotational precoding.

[0054] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0055] Reference Figure 1 A method for underwater acoustic multicarrier communication based on two-dimensional rotation precoding includes the following steps:

[0056] The transmitting end encodes and modulates the source data to obtain a data matrix; it constructs a time-dimensional rotation matrix and a frequency-dimensional rotation matrix using the chirped phase matrix and the unitary DFT matrix, respectively; it performs two-dimensional rotation precoding on the data matrix based on the time-dimensional and frequency-dimensional rotation matrices to spread a single information symbol to multiple effective subcarriers and multiple consecutive data moments according to the rotation orthogonal basis; it loads the precoding result into a multi-carrier frequency domain frame and inserts pilot signals to obtain the transmitted data.

[0057] The sending end sends the transmitted data;

[0058] The receiving end receives the transmitted data;

[0059] The receiving end performs synchronization, time-scale correction, pilot channel estimation, and MMSE equalization on the transmitted data to obtain an equalized two-dimensional matrix, and then recovers the source data based on the equalized two-dimensional matrix.

[0060] The implementation process of the underwater acoustic multicarrier communication method based on two-dimensional rotational precoding is explained step by step below.

[0061] The sending end executes steps S1 to S6. (Refer to...) Figure 2In general, the processing flow at the transmitting end is as follows: input bit / image data → channel coding and interleaving → QPSK mapping → symbol matrix rearrangement → two-dimensional inverse rotation mapping → pilot insertion → OFDM modulation / IFFT → cyclic prefix addition → up-conversion → power amplification and transmit transducer.

[0062] Step S1: Parameter reading and send / receive interface verification.

[0063] Read the received sequence, operating frequency band, transmit / receive sampling rate, FFT points, cyclic prefix, effective subcarrier index, number of data symbols per frame, LFM up and down scan templates, pilot sequence, and rotation parameters. And the configuration of the rotation matrix at the transceiver end.

[0064] Check the uniqueness of variables in the data file and parameter file.

[0065] examine , Check if the dimensions match the data matrix and calculate the error:

[0066] ;

[0067] In the formula, and These are the unitary errors of the rotation matrices in the frequency and time dimensions, respectively. and These are rotation matrices for the frequency dimension and the time dimension, respectively. express The conjugate transpose of . for The conjugate transpose of . This represents the identity matrix with the same dimension as the corresponding matrix. Describing the Frobenius norm, and These are the data dimensions for the frequency dimension and the time dimension, respectively.

[0068] Decoding only begins when the error is below the allowable threshold to prevent frame mismatch caused by inconsistent parameter versions or rotation angles. If any error exceeds the allowable threshold, the parameters and data are determined to be mismatched, and decoding stops.

[0069] Step S2: Data encoding, modulation, and two-dimensional grid formation.

[0070] The source data is encoded and modulated to obtain a data matrix, including the following steps:

[0071] The source data is sequentially subjected to optional channel coding, scrambling, block interleaving, and quadrature phase shift keying modulation / quadrature amplitude modulation mapping, and then loaded into a data matrix by column.

[0072] The data matrix is ​​represented as follows: The rows of the data matrix correspond to the effective subcarriers, and the columns correspond to the orthogonal frequency division multiplexing (OFDM) times of continuous data, thus providing a unified interface for two-sided rotation transformation.

[0073] Step S3, construct the unitary DFT basis.

[0074] Constructing a unitary DFT matrix of arbitrary length N, we have:

[0075] ;

[0076] satisfy .

[0077] according to To generate a unitary DFT matrix in the frequency dimension, we have:

[0078] .

[0079] according to To generate a unitary DFT matrix in the time dimension, we have:

[0080] .

[0081] Step S4: Construction of the quadratic phase matrix and the chirped-Fourier rotation matrix.

[0082] The chirped phase matrix is ​​defined as:

[0083] .

[0084] according to To generate the frequency-dimensional chirped phase matrix, we have:

[0085] .

[0086] according to To generate the chirped phase matrix in the time dimension, we have:

[0087] .

[0088] Using the chirped phase matrix and the unitary DFT matrix respectively, we can construct the rotation matrix:

[0089] ;

[0090] For the elements of the rotation matrix, we have:

[0091] ;

[0092] because and All are unitary matrices, therefore .

[0093] The rotation matrix in the frequency dimension is represented as: ;

[0094] The time-dimensional rotation matrix is ​​represented as: .

[0095] In the formula, Let be a rotation matrix in the frequency dimension. Let be the frequency-dimensional chirped phase matrix. Let be a frequency-dimensional unitary DFT matrix. Let be the element in the k-th row and n-th column of the unitary DFT matrix in the frequency dimension. The data dimension is the frequency dimension (i.e., the number of effective subcarriers participating in two-dimensional rotation precoding). For rotation parameters, diag{·} denotes a diagonal matrix formed by the sequence within the brackets.

[0096] Let be a rotation matrix in the frequency dimension. Let be the frequency-dimensional chirped phase matrix. Let be a frequency-dimensional unitary DFT matrix. Let be the element in the k-th row and n-th column of the unitary DFT matrix in the frequency dimension. The data dimension is the frequency dimension (i.e., the number of consecutive data moments involved in two-dimensional rotational precoding).

[0097] The second phase is compensated by the first chirped phase matrix to form a tilted wavefront; orthogonal mixing is achieved in the frequency domain using Fourier transform; and the second chirped phase matrix is ​​used to recover the second phase, resulting in the rotated orthogonal basis. The rotation matrix represents a global rotation in the time-frequency plane, suitable for dimensional rotation processing in systems such as OTFS.

[0098] In a preferred embodiment, .parameter The parameters are obtained based on the platform speed, center frequency, and Doppler extension configuration. The same parameters are used to control the tilt angle of the time-frequency basis function, and are applied to both the transmitting and receiving ends. Rotation parameters As a shared configuration quantity between the sender and receiver, the phase curvature of the two-dimensional basis function is adjusted while ensuring unitaryity and invertibility.

[0099] Step S5, two-dimensional rotation precoding.

[0100] Based on the implementation structure, let the frequency-dimensional left transform be... The time dimension right transformation is ,implement:

[0101]

[0102] This operation allows a single information symbol to be diffused across multiple effective subcarriers and multiple consecutive data moments according to a defined rotational orthogonal basis.

[0103] Since both left and right matrices are unitary matrices, the precoding maintains the Frobenius energy and does not require additional power normalization.

[0104] Step S6: Pilot, multi-carrier frame and acoustic waveform generation.

[0105] The precoding result is loaded into a multi-carrier frequency domain frame and pilot signals are inserted to obtain the transmitted data, including the following steps:

[0106] The columns of the precoding results are loaded into the effective subcarriers, and known pilots are inserted at fixed intervals;

[0107] Perform Inverse Fast Fourier Transform (IFFT), add a cyclic prefix, upsample, and perform real bandpass modulation on each frequency domain vector;

[0108] A linear frequency modulation (LFM) synchronization segment is inserted before each frame. Multi-frame playback uses an upper frequency sweep template / lower frequency sweep template to obtain the transmission data.

[0109] Among them, rotational precoding only applies to data blocks and does not change the pilot definition. Therefore, traditional OFDM pilot channel estimation and single-tap frequency domain equalization can still be used subsequently.

[0110] The transmitting end amplifies the data and transmits it to the transmitting transducer to send the data, while the receiving end receives the data.

[0111] The receiving end executes steps S7 to S12. The transmitted data is synchronized, time-scale corrected, pilot channel estimated, and MMSE equalized to obtain an equalized two-dimensional matrix. The source data is then recovered based on the equalized two-dimensional matrix.

[0112] Reference Figure 3 In general, the processing flow at the receiver is as follows: received signal → bandpass filtering and preprocessing → LFM synchronization detection → frame segmentation and start correction → time scale / Doppler estimation → time scale correction → cyclic prefix removal → FFT transform → pilot-aided channel estimation → phase-aware channel interpolation → MMSE equalization → constructing the equalized two-dimensional matrix → two-dimensional inverse rotation recovery → QPSK demapping → deinterleaving → channel decoding → image / bitstream recovery.

[0113] Pilot-assisted channel estimation and phase-aware channel interpolation are pilot-based phase compensation. The rotating time-frequency representation is recovered through a two-dimensional inverse rotation recovery step.

[0114] Step S7: Receive front-end resampling, amplitude normalization, and bandpass limiting.

[0115] The receiver reads the discrete received sequence from the hydrophone or simulation output. .

[0116] First, the discrete received sequence The DC component is removed, and the signal is normalized using a robust amplitude scale to avoid individual pulse peaks dominating subsequent related calculations.

[0117] If the received sampling rate Sampling rate at the sending end If there is a discrepancy, resampling is performed using a rational ratio or equivalent interpolation method to ensure that the synchronization template and the received data are on the same sampling time axis. After resampling, the result is... .

[0118] Subsequently, coverage of the operating frequency band was adopted. A finite impulse response bandpass filter suppresses out-of-band noise and low-frequency transducer drift, resulting in a front-end processing sequence. .

[0119] for For cases that are difficult to represent as small integer ratios, block interpolation or continuous time axis interpolation can be used to avoid the storage and computational overhead caused by large integer resampling factors.

[0120] Step S8: Alternate LFM complete chain detection and frame group confirmation.

[0121] The receiver performs LFM complete chain detection on the transmitted signal, including the following steps:

[0122] The transmitted signal is normalized and matched with the upper-sweep LFM template and the lower-sweep LFM template respectively to obtain two types of synchronization correlation sequences and their corresponding candidate peak positions;

[0123] Based on the pre-set alternation pattern of up-scanning and down-scanning frequencies at the transmitting end, candidate peaks are combined in a time sequence and filtered group by group.

[0124] When the following conditions are met: the peak amplitude reaches the detection threshold, the relative difference of peaks within the group is within the allowable range, the interval between adjacent peaks conforms to the frame structure constraints, and the number of detected peaks is consistent with the number of frames contained in a complete playback, the peak group is determined as a valid LFM frame chain, and the initial boundary of each frame in this playback is determined.

[0125] Isolated peaks, out-of-order peaks, and incomplete peak groups that do not meet the joint conditions are removed.

[0126] Specifically, using known upsweep frequency templates respectively and downscan frequency template right Perform normalized matching correlation.

[0127] For template Candidate delay The normalized correlation coefficient at point is defined as:

[0128] ;

[0129] In the formula, template In candidate delay Normalized matching correlation coefficient at the location, This is the received sequence after front-end preprocessing. For the first LFM-like template, ∈{u,d} correspond to the upper and lower frequency sweep templates, respectively. For candidate delay, This is the index of the discrete sampling points, and * indicates complex conjugate.

[0130] in, For LFM template length, To prevent small positive numbers with a denominator of zero.

[0131] right and Local peaks are detected separately. Noise peaks and repetitive multipath peaks near the same path are removed by meeting the following joint conditions: the peak amplitude reaches the detection threshold, the relative difference between peaks within the group is within the allowable range, the interval between adjacent peaks conforms to the frame structure constraints, and the number of detected peaks is consistent with the number of frames contained in a complete playback. Only a group of peaks that meets the preset UP / DOWN alternation order, the interval between adjacent peaks is close to the nominal frame interval, and the template type purity of each peak meets the requirements is considered a complete playback chain.

[0132] The purpose of this step is not only to find the maximum correlation peak, but also to verify the peak intensity, peak type, inter-peak time interval and frame integrity at the same time, so as to prevent strong multipath subpeaks, repeated playback segments or isolated noise peaks from being misjudged as valid frame start points.

[0133] Step S9, Time Scale / Broadband Doppler Estimation and Correction.

[0134] Time-scale correction includes the following steps:

[0135] For the complete LFM chain of the transmitted signal, a coarse estimate of the time scale is obtained based on the observation frame interval, and a set of candidate scales is constructed based on the coarse estimate of the time scale.

[0136] The LFM template or transmitted signal segment is scaled according to the candidate scale and the matching correlation peak of each frame is calculated to form the scale evaluation quantity;

[0137] Select the time scale with the largest scale evaluation quantity;

[0138] When the deviation of the time scale with the largest scale evaluation value relative to 1 exceeds the preset tolerance, the entire received waveform of the transmitted signal is corrected by inverse linear interpolation.

[0139] Specifically, for each complete LFM chain, a coarse estimate of the time scale is first obtained based on the observation frame interval, and then a candidate scale set is constructed around the coarse estimate. .

[0140] For candidate scale The LFM template or received segment is tentatively scaled according to this scale, and the matching correlation peak of each frame is recalculated to form a scale evaluation quantity.

[0141] The scale evaluation quantity is expressed as:

[0142] ;

[0143] in, The frame number, For the first The local search interval near the frame prediction start point.

[0144] Select the time scale with the largest scale evaluation value. As a time-scale estimate.

[0145] Understandably, in practice, a coarse search can be performed first with a larger step size, and then a fine search can be performed near the optimal candidate with a smaller step size, in order to balance computational cost and estimation accuracy.

[0146] when When the preset tolerance is exceeded, reverse linear interpolation correction is performed on the received waveform:

[0147] ;

[0148] in This represents the interpolation operator. (Corrected) Used for subsequent intra-frame precision timing, this suppresses the overall waveform stretching caused by broadband Doppler before entering the FFT.

[0149] Step S10: Cyclic prefix fine timing, pilot consistency screening, and FFT demodulation.

[0150] Centered on the nominal OFDM start position obtained by LFM mapping, in Search for timed offset within range .

[0151] For each candidate offset, the cyclic prefix relevance is calculated as follows:

[0152]

[0153] in The number of FFT points. Candidate timing offset The corresponding normalized cyclic prefix correlation, This is the timing offset relative to the nominal OFDM start position. The length of the cyclic prefix. This represents the received discrete sequence to be timed, where * denotes complex conjugate. To prevent small positive numbers with a denominator of zero.

[0154] First, a sparse coarse search is performed to retain several candidates with high relevance. Then, a fine search is performed on a sample-by-sample basis near the candidates. For the OFDM row where the pilot is located, the coherence and amplitude-phase continuity of the pilot channel response are examined at the same time. Finally, the frame starting point with the highest joint score is selected.

[0155] After determining the starting point, remove the cycle prefix from each OFDM symbol and execute. Point FFT is used to extract the pre-configured effective subcarrier indexes to obtain the received frequency domain data matrix. .

[0156] Step S11: Pilot-assisted channel estimation, phase-aware interpolation, and MMSE equalization.

[0157] For the At each pilot time, the known pilot vector is... Then the initial channel estimate is:

[0158] ;

[0159] In the formula, For the first Pilot time, the first pilot moment The initial channel estimate of each effective subcarrier, Yp[k] is the received frequency domain symbol at this pilot moment, and P[k] is the known pilot symbol at the transmitting end. For pilot time index, A small positive number introduced to avoid the denominator from approaching zero.

[0160] For adjacent pilots and Data time between ,make

[0161]

[0162] Linear interpolation can be directly performed on complex channels;

[0163] In the formula, The linear interpolation weights between two adjacent pilot timestamps are the values ​​for the current data time. Index for the current data moment, and They are located at The index of adjacent pilot times before and after, and < < .

[0164] Preferably, the amplitude and the expanded phase are interpolated separately to reduce phase skipping. Error jump during the time.

[0165] If full-frame pilot shape estimation is used, the local interpolation results can be weighted and fused with the full-frame common phase model to improve stability under fast time-varying conditions.

[0166] Regularized MMSE equalization is applied to the data subcarriers:

[0167]

[0168] In the formula, The MMSE equalization output of the k-th effective subcarrier at the m-th data time. For the corresponding received frequency domain symbol, The channel frequency response is obtained by pilot estimation and interpolation. express The complex conjugate, Let be the regularization parameter for the m-th data time.

[0169] in It can be adaptively obtained from the estimation of invalid subcarrier noise power, or upper and lower limit constraints can be set according to the channel average power.

[0170] all A balanced two-dimensional matrix is ​​formed according to the data time. This serves as the input to the corresponding orthogonal inverse transform module.

[0171] Step S12: Two-dimensional inverse rotation, phase correction, and information recovery.

[0172] Let the balanced two-dimensional matrix be denoted as... Calculated according to the reverse relationship of the sending end:

[0173]

[0174] This formula is related to the sending end. Strict matching.

[0175] The recovered complex symbols undergo common phase correction, decision-guided gain normalization, QPSK / QAM demapping, deinterleaving, descrambling, and channel decoding in sequence to finally reconstruct the original image or data.

[0176] If repeated playback occurs, a quality score can be generated based on LFM peak value, CP correlation, pilot consistency, and EVM, and the optimal playback can be selected or multiple soft symbol merging operations can be performed.

[0177] In one specific embodiment, the transmitting end sampling rate is 96 kHz, the underwater acoustic operating frequency band is 8-16 kHz, and the center frequency is 12 kHz; the baseband sampling rate is 16 kHz, the FFT number is 1024, the cyclic prefix length is 512 points, and the number of effective subcarriers is... Number of data symbols per frame QPSK modulation is used, and pilot signals with all effective subcarriers are inserted at fixed intervals. Rotation parameters are taken as follows: .

[0178] First generate and and set

[0179] Modulation matrix for each frame After performing bilateral precoding, the columns of the matrix are sequentially loaded into the effective subcarrier positions. Since the rotation matrix only affects the data blocks at the transmitter without changing the pilot definition, the traditional OFDM pilot channel estimation, cyclic prefix timing, and frequency domain single-tap equalization structure can be continued.

[0180] The receiving end obtains the balanced data matrix Then, calculate

[0181]

[0182] Then, phase correction, constellation normalization, QPSK / QAM demapping, deinterleaving, descrambling, and channel decoding are performed sequentially to recover the original bit or image information.

[0183] It should be noted that this method can also... As a system configuration parameter. Different parameters can be selected from a pre-defined set for different platform speeds, center frequencies, or Doppler extensions. As long as the transmitter and receiver use the same rotation matrix, the data grid size and subsequent OFDM interface remain unchanged.

[0184] In this embodiment, the underwater acoustic multicarrier communication method based on two-dimensional rotation precoding implemented in this application adds an adjustable two-dimensional rotation orthogonal layer while keeping the existing pilot, CP, FFT / IFFT, and frequency domain equalization links basically unchanged; by using the rotation basis to diffuse a single information symbol to multiple time-frequency resources, the concentrated influence of local deep fading or local ICI on a single symbol can be reduced; the rotation parameter degrees of freedom are increased; the strict unitary nature of the matrix avoids linear energy amplification; the data matrix does not introduce additional linear energy amplification; and the existing pilot, CP, FFT / IFFT, and frequency domain equalization links are maintained, making it easy to implement on existing underwater acoustic modulation and demodulation platforms through software upgrades.

[0185] The above is a detailed description of the preferred embodiments of this application, but this application is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for underwater acoustic multicarrier communication based on two-dimensional rotational precoding, characterized in that, include: The transmitting end encodes and modulates the source data to obtain a data matrix; The time-dimensional rotation matrix and the frequency-dimensional rotation matrix are constructed using the chirped phase matrix and the unitary DFT matrix, respectively. The data matrix is ​​precoded in two dimensions by rotation matrices in the time dimension and frequency dimension, so as to spread a single information symbol to multiple effective subcarriers and multiple consecutive data moments according to the rotation orthogonal basis. The precoding result is loaded into a multicarrier frequency domain frame and pilot signals are inserted to obtain the transmission data. The sending end sends the transmitted data, and the receiving end receives the transmitted data; The receiving end performs synchronization, time scale correction, pilot channel estimation, and MMSE equalization on the transmitted data to obtain an equalized two-dimensional matrix, and then recovers the source data based on the equalized two-dimensional matrix.

2. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The source data is encoded and modulated to obtain a data matrix, including: The source data is sequentially subjected to optional channel coding, scrambling, block interleaving, and quadrature phase shift keying modulation / quadrature amplitude modulation mapping, and then loaded into a data matrix by column. In this data matrix, rows correspond to effective subcarriers, and columns correspond to the orthogonal frequency division multiplexing (OFDM) times of continuous data.

3. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The rotation matrix in the frequency dimension is represented as: ; in, Represented as: ; Represented as: ; In the formula, Let be a rotation matrix in the frequency dimension. Let be the frequency-dimensional chirped phase matrix. Let be the frequency-dimensional unitary DFT matrix. Let be the element in the k-th row and n-th column of the unitary DFT matrix in the frequency dimension. For the frequency dimension of the data, For rotation parameters, diag{·} denotes a diagonal matrix formed by the sequence within the brackets.

4. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The time-dimensional rotation matrix is ​​represented as: ; in, Represented as: ; Represented as: ; In the formula, Let the rotation matrix be in the time dimension. Let be the chirped phase matrix in the time dimension. Let be the time-dimensional unitary DFT matrix. Let be the element in the k-th row and n-th column of the unitary DFT matrix in the time dimension. The time dimension is the data dimension. For rotation parameters, diag{·} denotes a diagonal matrix formed by the sequence within the brackets.

5. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The precoding result is represented as: ; In the formula, For the precoding result, For the conjugate transpose of the frequency-dimensional rotation matrix, For data matrix, It is a rotation matrix in the time dimension; The two-dimensional rotation precoding only applies to data blocks and does not change the pilot definition.

6. The underwater acoustic multi-carrier communication method based on two-dimensional rotation precoding according to claim 1, characterized in that, parameter The parameters are obtained based on the platform speed, center frequency, and Doppler extension configuration. The same parameters are used to control the tilt angle of the time-frequency basis function, and are applied to both the transmitting and receiving ends. .

7. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The precoding result is loaded into a multi-carrier frequency domain frame and pilot signals are inserted to obtain the transmitted data, including: The columns of the precoding results are loaded into the effective subcarriers, and known pilots are inserted at fixed intervals; Perform Inverse Fast Fourier Transform (IFFT), add a cyclic prefix, upsample, and perform real bandpass modulation on each frequency domain vector; A linear frequency modulation (LFM) synchronization segment is inserted before each frame. Multi-frame playback uses an upper frequency sweep template / lower frequency sweep template to obtain the transmitted data.

8. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The receiver performs LFM complete chain detection on the transmitted signal, including: The transmitted signal is normalized and matched with the upper-sweep LFM template and the lower-sweep LFM template respectively to obtain two types of synchronization correlation sequences and their corresponding candidate peak positions; Based on the pre-set alternation pattern of up-scan and down-scan frequencies at the transmitting end, candidate peaks are combined in time sequence and filtered group by group. When the following conditions are met: the peak amplitude reaches the detection threshold, the relative difference of peaks within the group is within the allowable range, the interval between adjacent peaks conforms to the frame structure constraints, and the number of detected peaks is consistent with the number of frames contained in a complete playback, the peak group is determined as a valid LFM frame chain, and the initial boundary of each frame in this playback is determined. Isolated peaks, out-of-order peaks, and incomplete peak groups that do not meet the aforementioned joint conditions are removed.

9. The underwater acoustic multi-carrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The time-scale correction includes: For the complete LFM chain of the transmitted signal, a coarse estimate of the time scale is obtained based on the observation frame interval, and a set of candidate scales is constructed based on the coarse estimate of the time scale. The LFM template or transmitted signal segment is scaled according to the candidate scale and the matching correlation peak of each frame is calculated to form the scale evaluation quantity; Select the time scale with the largest scale evaluation quantity; When the deviation of the time scale with the largest scale evaluation value relative to 1 exceeds the preset tolerance, the entire received waveform of the transmitted signal is corrected by inverse linear interpolation.

10. The underwater acoustic multicarrier communication method based on two-dimensional rotational precoding according to claim 1, characterized in that, The MMSE equalization for effective subcarriers is expressed as: ; In the formula, The MMSE equalization output of the k-th effective subcarrier at the m-th data time. For the corresponding received frequency domain symbol, The channel frequency response is obtained by pilot estimation and interpolation. express The complex conjugate, Let be the regularization parameter for the m-th data time.