A Method and System for Isolating Cross-Domain Interference in a DSSS Receiver Based on Multidimensional Adaptive Filtering

Abstract

This invention discloses a method and system for mitigating cross-address interference in a DSSS receiver based on multidimensional adaptive filtering. The method includes matched filtering and noncoherent accumulation steps, a detection and decision step, an open-loop adaptive filtering step, and interference reconstruction and elimination steps. The system includes a matched filtering and noncoherent accumulation module, a detection and decision module, an open-loop adaptive filtering module, and an interference reconstruction and elimination module. The open-loop adaptive filtering module is composed of a delay-frequency two-dimensional correlation function calculation unit, a filter coefficient calculation unit, and a filtering unit. This invention uses matched filtering for coarse parameter estimation and detection of strong interference signals. Based on the coarse estimation results, a fine parameter estimation range is defined. Local multidimensional adaptive filtering is used to obtain fine estimates of the signal complex amplitude at different channels, delays, and Doppler frequencies within the observation time, enabling the reconstruction and elimination of interference signals. This reduces algorithm complexity while suppressing cross-correlation caused by cross-address interference.

Description

Technical Field

[0001] This invention relates to the field of aerospace technology, specifically to a method and system for resisting cross-address interference in a DSSS receiver based on multidimensional adaptive filtering. Background Technology

[0002] Direct sequence spread spectrum (DSSS) signals possess advantages such as high concealment, strong resistance to interception and interference, and are widely used in satellite navigation and positioning systems, aerospace telemetry and communication systems, and civilian communications. DSSS receivers utilize the correlation characteristics of the modulation pseudocode in DSSS signals to achieve detection and synchronization of the received signal. However, in DSSS-based multi-user systems, because the pseudocode sequences allocated to different users are not completely orthogonal and their cross-correlation functions are not zero, cross-address interference exists between users. When the cross-address interference power is high, the performance of the DSSS receiver degrades sharply.

[0003] Anti-interference algorithms based on adaptive filtering have good anti-interference capabilities and have therefore attracted widespread attention.

[0004] In the 2006 IEEE Transactions on Aerospace and Electronic Systems, Vol. 42, No. 3, pp. 891-903, Shannon D. Blunt et al. proposed a multistatic adaptive pulse compression (MAPC) algorithm based on iterative minimum mean square error (RMMSE). The RMMSE-based MAPC algorithm retains the good anti-interference capability of adaptive filtering algorithms. It employs an open-loop structure, exhibiting higher convergence robustness and requiring fewer iterations compared to the LMS algorithm. However, the filter order of this algorithm is limited to the number of baseband transmit waveform sampling points in one cycle, resulting in a large RMMSE filter order and high algorithm complexity. In the 2015 IEEE Transactions on Aerospace and Electronic Systems, Volume 51, Issue 1, pp. 548-564, Zhengzheng Li et al. proposed a fast adaptive pulse compression algorithm that applies RMMSE filtering to the pulse compression results. By significantly reducing the filter order, the computational complexity is greatly reduced with minimal performance loss. In the 2016 IEEE Transactions on Aerospace and Electronic Systems, Volume 52, Issue 4, pp. 2044-2053, Yuyao Shen et al. extended the RMMSE filtering algorithm based on pulse compression results to suppress multiple access interference for direct-sequence spread spectrum (DSSS) signals.

[0005] However, this algorithm only focuses on delay-dimensional interference suppression and has not yet considered the effects of delay, non-zero Doppler frequency, and symbol transition in the signal model.

[0006] Therefore, to improve the robustness of the interference suppression dimension of the algorithm, the effects of delay, non-zero Doppler frequency and symbol transition should be considered in the signal model, the algorithm should be modified, the algorithm dimension should be expanded, and a low-complexity anti-interference algorithm based on multi-dimensional RMMSE filtering should be designed. However, such improvements to the algorithm have not been reported at home and abroad. Summary of the Invention

[0007] In view of this, the present invention provides a method and system for resisting cross-address interference in a DSSS receiver based on multidimensional adaptive filtering. It combines the basic ideas of serial interference cancellation and adaptive filtering, performs coarse parameter estimation and strong interference signal detection through matched filtering, defines the fine parameter estimation range based on the coarse estimation results, and obtains fine estimation of signal complex amplitude at different channels, delays and Doppler frequencies during the observation time through local multidimensional adaptive filtering, thereby reconstructing and eliminating interference signals. This reduces the algorithm complexity while suppressing cross-correlation caused by cross-address interference.

[0008] To achieve the above objectives, the technical solution of this invention is as follows: a multi-dimensional adaptive filtering-based anti-multiple access interference acquisition method, which first performs multi-dimensional adaptive filtering in each receiving channel of the DSSS receiver according to the delay search interval Δ T and frequency search interval Δ F The range of delay and frequency uncertainty is divided into a two-dimensional search grid, where each grid is denoted as a search cell; the following steps are performed:

[0009] Step 1: In the search grid of each receiving channel where no signal is detected, zero-padding is used to perform zero-padding matched filtering and non-coherent accumulation on the received signal to resist data jumps, and the energy accumulation value in each search cell is obtained.

[0010] Step 2: Use the peak value of the energy accumulation value in the search unit of each channel as the detection decision value. When the detection decision value is higher than the set detection threshold, it is considered that the channel has detected a signal.

[0011] After completing one round of signal detection, the signal detection count is incremented by 1, and the set of channels N for which no signals were detected and the set of channels M for which signals were detected are updated according to the detection results; initially, N is set to the entire set and M is set to the empty set.

[0012] If the target signal has been detected, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has been successfully acquired; if the target signal is not detected when the signal detection count reaches the preset upper limit, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has failed to be acquired.

[0013] If the signal detection count has not reached the preset upper limit and no target signal has been detected, the signal acquisition has not ended, and step 3 is executed.

[0014] Step 3: Estimate the complex amplitude vector of the signal using adaptive filtering in the search unit of the channel where the target signal is detected.

[0015] Step 4: Reconstruct all detected interference signals based on the complex amplitude vector estimate, remove all reconstructed signals from the received signal, and return to Step 1.

[0016] Further, step 3: In the search unit for the detected signal of the m-th channel, the complex amplitude vector of the signal is estimated using adaptive filtering, specifically as follows:

[0017] Start adaptive filtering, set the upper limit of iteration, set the initial iteration value to 1, and perform the following steps:

[0018] Step 301: Calculate and output the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal;

[0019] Step 302: Adaptive calculation of the filter coefficients for the m-th channel in the current iteration;

[0020] Step 303: Filter the delay-frequency two-dimensional correlation results in each channel of the detected signal to obtain the estimated value of the complex amplitude vector output in the current iteration;

[0021] Step 304: Increment the iteration count by 1. If the iteration converges or reaches the preset maximum number of iterations, return to step 301. Otherwise, use the output of the last iteration as the estimated value of the complex amplitude vector.

[0022] Furthermore, in step 1, the energy accumulation value in each search grid is...

[0023]

[0024] Where q = nL(2K+1) + l(2K+1) + K + k is a simplified representation of the three-dimensional search unit index (n, l, k) for the receiving channel, delay, and frequency. In the search unit index (n, l, k), n is the receiving channel index, l is the index of the delay dimension search unit, k is the index of the frequency dimension search unit, L is the number of delay dimension search units, and 2K+1 is the number of frequency dimension search units; the superscript H indicates the conjugate transpose; P is the number of non-coherent accumulations, and y p and y p+1 The received signal sampling sequences are respectively the p-th and p+1-th noncoherent accumulation time intervals, y p Represented as

[0025]

[0026] in, The superscript T in L indicates transpose. sτ represents the number of sampling points within the coherent accumulation time period, where the coherent accumulation time is equal to the code period, i represents the transmitter index, p is the p-th coherent accumulation time period, and τ i f is the delay of the i-th spread spectrum signal. i Where is the Doppler frequency, N represents the number of transmitters, and s +,i,p (τ i ,f i ) and s -,i,p (τ i ,f i S represents the baseband characteristic waveforms when there is an unsigned bit transition and when there is a sign bit transition, respectively. i,p (τ i ,f i )=[s +,i,p (τ i ,f i )s -,i,p (τ i ,f i [)] represents the three-dimensional characteristic waveform matrix of delay, frequency, and data; v p x is the combined Gaussian white noise of the orthogonal and in-phase branches. +,i,p and x -,i,p Represent the complex amplitudes for unsigned transitions and signed transitions, respectively, with the following expressions:

[0027]

[0028] Where A i d represents the product of the amplitude and phase of the i-th spread spectrum signal. i (·)∈{-1,1} represents the data symbol of the i-th spread spectrum signal, T coh τ is the coherent accumulation time, p is the p-th coherent accumulation time interval, and τ is the time interval of the coherent accumulation. i x is the delay of the i-th spread spectrum signal. i,p =[x +,i,p x -,i,p ] T Represents a complex magnitude vector;

[0029] w MF,q,p The matched filter coefficients in the search unit (n,l,k) are expressed as follows:

[0030]

[0031] in To delay segmented search indexing, Indicates rounding down. and They represent 1×(l-aL) s A vector consisting of all zeros in a column.

[0032] Furthermore, adaptive filtering is initiated, and the delay-frequency two-dimensional correlation result is calculated and output based on the characteristic waveform matrix and the input received signal, specifically as follows:

[0033] The calculated delay-frequency two-dimensional correlation result is z Acor,m,p ;

[0034]

[0035] In the formula, the subscript Acor indicates the relevant operation, the subscript m is the m-th channel of the detected signal, and the subscript p represents the p-th signal segment; W Acor,m The characteristic waveform matrix used for complex amplitude estimation is expressed as follows:

[0036]

[0037]

[0038] in For L T L is the number of one-sided correlators in the delay dimension. F The number of frequency-dimensional single-sided correlators. This is a delayed segmented search index.

[0039] Furthermore, adaptive calculation is performed to compute the filter coefficients of the m-th channel in each iteration of filtering, specifically:

[0040] Calculate the filter coefficients for the m-th channel in the j-th iteration of filtering.

[0041]

[0042] in The correlation coefficient matrix between the m-th spread spectrum signal and the i-th baseband characteristic waveform of the signal, i∈M, is input by the delay-frequency two-dimensional correlation function calculation module; For noise variance, L is the matrix related to the second moment of the complex amplitude. s This refers to the number of sampling points within the coherent accumulation time period; specifically...

[0043]

[0044] Based on normalized energy accumulation value set up Used as the initial value for the complex amplitude vector and the filter coefficients are calculated. Subsequently, the complex magnitude vector estimate is obtained from the output of the j-th iteration. Calculate the filter coefficients for Take the conjugate, where the superscript j is the iteration number, the subscript m is the m-th spread spectrum signal, and the subscript p is the p-th coherent accumulation time period; The expression is

[0045] Furthermore, the delay-frequency two-dimensional correlation results in each detected signal channel are filtered to obtain the estimated value of the complex amplitude vector output in the current iteration, specifically:

[0046] The delay-frequency two-dimensional correlation result calculated for each channel of the detected signal, i.e., the filter coefficients of the m-th channel in the j-th iteration of filtering. For z Acor,m,p After filtering, the estimated value of the complex amplitude vector of the signal output in the j-th iteration is obtained.

[0047] Further, step 4: Reconstruct all detected interference signals based on the complex amplitude vector estimate, and remove all reconstructed signals from the received signal. The specific calculation formula is as follows:

[0048]

[0049] in This indicates the received signal after removing all reconstructed signals. Feedback is sent to step 1 and used in subsequent execution processes. Replace y in equation (1) p Repeat steps 1 through 4 until the capture ends; This is an estimate of the delay time; This is an estimate of the Doppler frequency; This is an estimate of the complex magnitude vector; It is a three-dimensional feature waveform matrix. These are the delay estimate and Doppler estimate of the m-th spread spectrum signal, respectively.

[0050] A multi-dimensional adaptive filtering-based anti-multiple access interference acquisition system, wherein each receiving channel of the DSSS receiver is configured according to the delay search interval Δ T and frequency search interval Δ F The range of delay and frequency uncertainty is divided into a two-dimensional search grid, where each grid is denoted as a search unit.

[0051] This anti-multiple access interference acquisition system includes a matched filtering and non-coherent accumulation module, a detection and decision module, an open-loop adaptive filtering module, and an interference reconstruction and cancellation module.

[0052] The matched filtering and non-coherent accumulation module uses zero-padding to perform zero-padding matched filtering and non-coherent accumulation on the received signal in the search grid of each receiving channel where no signal is detected, in order to obtain the energy accumulation value in each search unit.

[0053] The detection decision module uses the peak value of the energy accumulation value in the search unit of each channel as the detection decision value. When the detection decision value is higher than the set detection threshold, it is considered that the channel has detected a signal.

[0054] After completing one round of signal detection, the signal detection count is incremented by 1, and the set of channels N for which no signals were detected and the set of channels M for which signals were detected are updated according to the detection results; initially, N is set to the entire set and M is set to the empty set.

[0055] If the target signal has been detected, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has been successfully acquired; if the target signal is not detected when the signal detection count reaches the preset upper limit, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has failed to be acquired.

[0056] If the signal detection count has not reached the preset upper limit and the target signal has not been detected, the signal acquisition has not ended, and the open-loop adaptive filtering module is started.

[0057] The open-loop adaptive filtering module estimates the complex amplitude vector of the detected signal in the search unit of the m-th channel using adaptive filtering.

[0058] Interference reconstruction and cancellation module: Reconstructs all detected interference signals based on the complex amplitude vector estimate, removes all reconstructed signals from the received signal, and feeds them back to the matched filtering and non-coherent accumulation module.

[0059] Furthermore, the open-loop adaptive filtering module includes a delay-frequency two-dimensional correlation function calculation unit, a filtering unit, and a filter coefficient calculation unit.

[0060] The delay-frequency two-dimensional correlation function calculation unit calculates the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal, and outputs it to the filtering unit.

[0061] The filter coefficient calculation unit calculates the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal, and outputs it to the filter unit.

[0062] The filtering unit filters the delay-frequency two-dimensional correlation results in each channel of the detected signal to obtain the estimated value of the complex amplitude vector output in the current iteration.

[0063] One iteration is defined as one execution of the delay-frequency two-dimensional correlation function calculation unit, the filtering unit, and the filter coefficient calculation unit. An upper limit for the iteration is set, and the initial value of the iteration is set to 1. The number of iterations is incremented by 1. When the iteration converges or the preset maximum number of iterations is reached, the process returns to the delay-frequency two-dimensional correlation function calculation unit. Otherwise, the output of the last iteration is used as the estimated value of the complex amplitude vector for output.

[0064] Beneficial effects:

[0065] This method combines the basic principles of serial interference cancellation and RMMSE filtering, alternately performing interference detection, cancellation, and RMMSE filtering parameter estimation on the received baseband digital signal, thereby suppressing the cross-correlation sidelobes of cross-address interference. Compared to existing single-dimensional delay RMMSE filtering algorithms, this method extends the algorithm dimension to three dimensions: delay, non-zero Doppler frequency, and symbol transition, improving its robustness. By replacing the tracking loop in interference cancellation with RMMSE filtering, the acquisition time is shortened compared to interference cancellation methods. The interference cancellation results are used to narrow the parameter estimation range, transforming global RMMSE filtering into local RMMSE filtering, effectively reducing the algorithm's complexity. This method employs adaptive filtering technology, exhibiting strong resistance to cross-address interference. For the commonly used 1023-bit Gold code, under a fixed target detection probability, the interference tolerance of the algorithm is effectively improved. Attached Figure Description

[0066] Figure 1 This is a block diagram of the anti-interference method for DSSS receivers based on multidimensional adaptive filtering structure according to the present invention;

[0067] Figure 2 This is a structural block diagram of the adaptive filtering module of the present invention. Detailed Implementation

[0068] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0069] Example 1:

[0070] This invention provides a multi-dimensional adaptive filtering-based method for acquiring anti-multiple access interference. First, in each receiving channel of the DSSS receiver, based on the delay search interval Δ... T and frequency search interval Δ F The range of delay and frequency uncertainty is divided into a two-dimensional search grid, where each grid is denoted as a search cell; the following steps are performed:

[0071] Step 1: In the search grid of each receiving channel where no signal is detected, zero-padding is used to perform zero-padding matched filtering and non-coherent accumulation on the received signal to resist data jumps, obtaining the energy accumulation value in each search cell; the energy accumulation value in each search cell is...

[0072]

[0073] Where q = nL(2K+1) + l(2K+1) + K + k is a simplified representation of the three-dimensional search unit index (n, l, k) for the receiving channel, delay, and frequency. In the search unit index (n, l, k), n is the receiving channel index, l is the index of the delay dimension search unit, k is the index of the frequency dimension search unit, L is the number of delay dimension search units, and 2K+1 is the number of frequency dimension search units; the superscript H indicates the conjugate transpose; P is the number of non-coherent accumulations, and y p and y p+1 The received signal sampling sequences are respectively the p-th and p+1-th noncoherent accumulation time intervals, y p Represented as

[0074]

[0075] in, The superscript T in L indicates transpose. s τ represents the number of sampling points within the coherent accumulation time period, where the coherent accumulation time is equal to the code period, i represents the transmitter index, p is the p-th coherent accumulation time period, and τ i f is the delay of the i-th spread spectrum signal. i Where is the Doppler frequency, N represents the number of transmitters, and s +,i,p (τ i ,f i ) and s -,i,p (τ i ,f i S represents the baseband characteristic waveforms when there is an unsigned bit transition and when there is a sign bit transition, respectively. i,p (τ i ,f i )=[s +,i,p (τ i ,f i )s -,i,p (τ i ,f i [)] represents the three-dimensional characteristic waveform matrix of delay, frequency, and data; v p x is the combined Gaussian white noise of the orthogonal and in-phase branches. +,i,p and x -,i,p Represent the complex amplitudes for unsigned transitions and signed transitions, respectively, with the following expressions:

[0076]

[0077] Where A i d represents the product of the amplitude and phase of the i-th spread spectrum signal. i (·)∈{-1,1} represents the data symbol of the i-th spread spectrum signal, T coh τ is the coherent accumulation time, p is the p-th coherent accumulation time interval, and τ is the time interval of the coherent accumulation. i x is the delay of the i-th spread spectrum signal. i,p =[x +,i,p x -,i,p ] T Represents a complex magnitude vector;

[0078] w MF,q,p The matched filter coefficients in the search unit (n,l,k) are expressed as follows:

[0079]

[0080] in To delay segmented search indexing, Indicates rounding down. and They represent 1×(l-aL) s A vector consisting of all zeros in a column.

[0081] Step 2: Use the peak value of the energy accumulation value in the search unit of each channel as the detection decision value. When the detection decision value is higher than the set detection threshold, it is considered that the channel has detected a signal.

[0082] After completing one round of signal detection, the signal detection count is incremented by 1, and the set of channels N for which no signals were detected and the set of channels M for which signals were detected are updated according to the detection results; initially, N is set to the entire set and M is set to the empty set.

[0083] If the target signal has been detected, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has been successfully acquired; if the target signal is not detected when the signal detection count reaches the preset upper limit, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has failed to be acquired.

[0084] If the signal detection count has not reached the preset upper limit and no target signal has been detected, the signal acquisition has not ended, and step 3 is executed.

[0085] Step 3: Estimate the complex amplitude vector of the signal using adaptive filtering in the search unit of the channel where the target signal is detected; this step includes the following specific steps:

[0086] Start adaptive filtering, set the upper limit of iteration, set the initial iteration value to 1, and perform the following steps:

[0087] Step 301: Calculate and output the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal; specifically:

[0088] The calculated delay-frequency two-dimensional correlation result is z Acor,m,p ;

[0089]

[0090] In the formula, the subscript Acor indicates the relevant operation, the subscript m is the m-th channel of the detected signal, and the subscript p represents the p-th signal segment; W Acor,m The characteristic waveform matrix used for complex amplitude estimation is expressed as follows:

[0091]

[0092]

[0093] in For L T L is the number of one-sided correlators in the delay dimension. F The number of frequency-dimensional single-sided correlators. This is a delayed segmented search index.

[0094] Step 302: Adaptively calculate the filter coefficients of the m-th channel in the current iteration of filtering; specifically:

[0095] Calculate the filter coefficients for the m-th channel in the j-th iteration of filtering.

[0096]

[0097] in The correlation coefficient matrix between the m-th spread spectrum signal and the i-th baseband characteristic waveform of the signal, i∈M, is input by the delay-frequency two-dimensional correlation function calculation module; For noise variance, L is the matrix related to the second moment of the complex amplitude. s This refers to the number of sampling points within the coherent accumulation time period; specifically...

[0098]

[0099] Based on normalized energy accumulation value set up Used as the initial value for the complex amplitude vector and the filter coefficients are calculated. Subsequently, the complex magnitude vector estimate is obtained from the output of the j-th iteration.

[0100] Calculate the filter coefficients for Take the conjugate, where the superscript j is the iteration number, the subscript m is the m-th spread spectrum signal, and the subscript p is the p-th coherent accumulation time period; The expression is

[0101] Step 303: Filter the delay-frequency two-dimensional correlation results in each channel of the detected signal to obtain the estimated value of the complex amplitude vector output in the current iteration; specifically:

[0102] The delay-frequency two-dimensional correlation result calculated for each channel of the detected signal, i.e., the filter coefficients of the m-th channel in the j-th iteration of filtering. For z Acor,m,p After filtering, the estimated value of the complex amplitude vector of the signal output in the j-th iteration is obtained.

[0103] Step 304: Increment the iteration count by 1. If the iteration converges or reaches the preset maximum number of iterations, return to step 301. Otherwise, use the output of the last iteration as the estimated value of the complex amplitude vector.

[0104] Step 4: Reconstruct all detected interference signals based on the complex amplitude vector estimate, remove all reconstructed signals from the received signal, and return to Step 1.

[0105] The specific calculation formula is as follows:

[0106]

[0107] in This indicates the received signal after removing all reconstructed signals. Feedback is sent to step 1 and used in subsequent execution processes. Replace y in equation (1) p Repeat steps 1 through 4 until the capture ends; This is an estimate of the delay time; This is an estimate of the Doppler frequency; This is an estimate of the complex magnitude vector; It is a three-dimensional feature waveform matrix. These are the delay estimate and Doppler estimate of the m-th spread spectrum signal, respectively;

[0108] The purpose of this invention is to provide a low-complexity DSSS receiver anti-interference method based on multidimensional adaptive filtering. The proposed anti-interference method combines the basic ideas of serial interference cancellation and adaptive filtering. It uses matched filtering for coarse parameter estimation and detection of strong interference signals. Based on the coarse estimation results, it defines the fine parameter estimation range. Then, through local multidimensional adaptive filtering, it obtains fine estimates of the signal complex amplitude at different channels, delays, and Doppler frequencies within the observation time, enabling the reconstruction and elimination of interference signals. This reduces algorithm complexity while suppressing cross-correlation caused by inter-interference.

[0109] Example 2:

[0110] The technical solution for implementing the present invention is as follows: The system adopted by the low-complexity DSSS receiver anti-interference method based on multidimensional adaptive filtering includes a matched filtering and non-coherent accumulation module, a detection and decision module, an open-loop adaptive filtering module, and an interference reconstruction and cancellation module; wherein the open-loop adaptive filtering module is composed of a delay-frequency two-dimensional correlation function calculation unit, a filter coefficient calculation unit, and a filtering unit.

[0111] The system performs the following steps (where bold letters indicate sets or matrices):

[0112] (1) Initialize the signal detection state, set the upper limit of the number of iterations (give a specific example 3-50), let N represent the set of undetected signals and set it to the whole set (containing all channels to be detected), and M represent the set of detected signals and set it to the empty set.

[0113] (2) In each receiving channel, based on the delay search interval Δ T =T s (T s (Sampling interval of the received signal) and frequency search interval Δ F The range of delay and frequency uncertainty is divided into a two-dimensional search grid.

[0114] (3) The matched filtering and non-coherent accumulation modules perform zero-padding matched filtering and non-coherent accumulation of the received signal in the search grid of each receiving channel where no signal is detected, to resist data jumps. The energy accumulation value in each search grid is...

[0115]

[0116] Where q = nL(2K+1) + l(2K+1) + K + k is a simplified representation of the three-dimensional search unit index (n, l, k) for the receiving channel, delay, and frequency dimensions; n is the receiving channel index, l is the index of the delay dimension search unit, L is the number of delay dimension search units, k is the index of the frequency dimension search unit, and 2K+1 is the number of frequency dimension search units. The superscript H indicates the conjugate transpose. P is the number of non-coherent accumulations, y p The received signal sampling sequence during the p-th coherent accumulation time period can be represented as:

[0117]

[0118] Where the superscript T indicates transpose, L s The number of sampling points within the coherent accumulation time period is given by the code period. i represents the transmitter index, N represents the number of transmitters, and s represents the number of sampling points within the coherent accumulation time period. +,i,p (τ i ,f i ) and s -,i,p (τ i ,f i S represents the baseband characteristic waveforms when there is an unsigned bit transition and when there is a sign bit transition, respectively. i,p (τ i ,f i )=[s +,i,p (τ i ,f i )s -,i,p (τ i ,f i [)] is a three-dimensional characteristic waveform matrix of delay, frequency, and data. v p x is the combined Gaussian white noise of the orthogonal and in-phase branches. +,i,p and x -,i,p Represent the complex amplitudes for unsigned transitions and signed transitions, respectively, with the following expressions:

[0119]

[0120] Where A i d represents the product of the amplitude and phase of the i-th spread spectrum signal. i (·)∈{-1,1} represents the data symbol of the i-th spread spectrum signal, T coh τ is the coherent accumulation time, p is the p-th coherent accumulation time interval, and τ is the time interval of the coherent accumulation. i x represents the delay of the i-th spread spectrum signal. i,p =[x +,i,p x -,i,p ] T This represents a complex magnitude vector.

[0121] w MF,q,pThe matched filter coefficients in the search unit (n,l,k) are expressed as follows:

[0122]

[0123] in To delay segmented search indexing, This indicates rounding down to the nearest integer.

[0124] (4) The detection and decision module uses the energy accumulation value h within the search unit of each channel. n,l,k The peak value is the detection decision value. When the peak value is higher than the detection threshold, it is considered that the nth signal has been detected (n is the receiving channel index).

[0125] After completing one round of signal detection, the signal detection count is incremented by 1, and the set N of channels where no signal was detected and the set M of channels where signals were detected are updated according to the detection results. If the target signal has been detected, the acquisition ends and the result of successful target signal acquisition is output; if the signal detection count reaches the upper limit (preset value) and no target signal is detected, the acquisition ends and the result of target signal acquisition failure is output; if the signal detection count has not reached the upper limit and no target signal has been detected, the signal acquisition is considered not to have ended, and execution continues to the search unit where the signal is detected in the m-th (m∈M) channel. The complex amplitude vector x of the signal is estimated using adaptive filtering. m,p The number of segments in segment p here.

[0126] (5) Input signal y p The detection and decision module only controls the initiation of adaptive filtering. After the open-loop adaptive filtering module is started, the delay-frequency two-dimensional correlation function calculation unit calculates the correlation output based on the characteristic waveform matrix and the input received signal, and outputs the result to the filtering module. The calculation formula is as follows:

[0127]

[0128] Among them, z Acor,m,p For Acor, perform correlation, where m is the m-th channel and p-th signal segment of the detected signal, and z is the correlation output; W Acor,m The characteristic waveform matrix used for complex amplitude estimation is expressed as follows:

[0129]

[0130]

[0131] Where L T L is the number of one-sided correlators in the delay dimension. F The number of frequency-dimensional single-sided correlators. This is a delayed segmented search index.

[0132] (6) The filter coefficient calculation unit calculates the filter coefficients of the m-th channel in the j-th (j=0,1,……) iteration. And output it to the filtering unit.

[0133]

[0134] in This is the correlation coefficient matrix between the m-th spread spectrum signal and the baseband characteristic waveform of the i-th (i∈M) signal. The superscript H indicates transpose, and it is input from the delay-frequency two-dimensional correlation function calculation module. For noise variance, L is the matrix related to the second moment of the complex amplitude. s The number of sampling points within the coherent accumulation time period; specifically

[0135]

[0136] Based on normalized energy accumulation value set up Used as the initial value for the complex amplitude vector and the filter coefficients are calculated. Subsequent iterations based on the filter module output

[0137] Calculate the filter coefficients j is the iteration number; The expression is for Take conjugate.

[0138] The filtering module calculates the correlation result z of the delay-frequency two-dimensional correlation function input to the module for each channel of the detected signal. Acor,m,p Filtering is performed to obtain the complex amplitude estimate of the output of the j-th iteration.

[0139]

[0140] If the iteration does not converge, then the iterative output will be... The output is sent to the filter coefficient calculation module. The iteration counter j is incremented by 1. Each execution of the delay-frequency two-dimensional correlation function calculation unit-filter unit-filter coefficient calculation unit constitutes one iteration process, until the iteration converges. The output of the last iteration is used as the basis for the iteration. As an estimate of the complex magnitude vector

[0141] (8) After the RMMSE filtering iteration converges in each channel of the detected signal, the interference reconstruction and cancellation module reconstructs all detected interference signals based on the complex amplitude vector estimate obtained by the adaptive filtering module, and removes all reconstructed signals from the received signal. The calculation formula is as follows:

[0142]

[0143] in This indicates the received signal after removing all reconstructed signals. Feedback is sent to the matched filtering and non-coherent accumulation modules, and used in subsequent iterations. Replace y p Return to the matched filtering and non-coherent accumulation module, that is, repeat steps (3) to (8) until the capture ends. This is an estimate of the delay time; This is an estimate of the Doppler frequency; This is an estimate of the complex magnitude vector; It is a three-dimensional feature waveform matrix;

[0144] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for acquiring anti-multiple access interference based on multidimensional adaptive filtering, characterized in that, First, in each receiving channel of the DSSS receiver, based on the delay search interval... and frequency search interval The range of delay and frequency uncertainty is divided into a two-dimensional search grid, where each grid is denoted as a search cell; the following steps are performed: Step 1: In the search grid of each receiving channel where no signal is detected, zero-padding is used to perform zero-padding matched filtering and non-coherent accumulation on the received signal to resist data jumps, and the energy accumulation value in each search unit is obtained. Step 2: Use the peak value of the energy accumulation value in the search unit of each channel as the detection decision value. When the detection decision value is higher than the set detection threshold, it is considered that the channel has detected a signal. After completing one round of signal detection, the signal detection count is incremented by 1, and the set of channels N for which no signals were detected and the set of channels M for which signals were detected are updated according to the detection results; initially, N is set to the entire set and M is set to the empty set. If the target signal has been detected, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has been successfully acquired; if the target signal is not detected when the signal detection count reaches the preset upper limit, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has failed to be acquired. If the signal detection count has not reached the preset upper limit and no target signal has been detected, the signal acquisition has not ended, and step 3 is executed; Step 3: In the search unit of the channel where the target signal is detected, the complex amplitude vector of the signal is estimated using adaptive filtering; Step 4: Reconstruct all detected interference signals based on the complex amplitude vector estimate, remove all reconstructed signals from the received signal, and return to Step 1.

2. The anti-multiple access interference acquisition method based on multidimensional adaptive filtering as described in claim 1, characterized in that, Step 3: In the search unit of the detected signal of the m-th channel, the complex amplitude vector of the signal is estimated using adaptive filtering, m∈M, specifically as follows: Start adaptive filtering, set the upper limit of iteration, set the initial iteration value to 1, and perform the following steps: Step 301: Calculate and output the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal; Step 302: Adaptive computation of the first... m The filter coefficients for each channel in the current iteration of filtering; ； Step 303: Filter the delay-frequency two-dimensional correlation results in each channel of the detected signal to obtain the estimated value of the complex amplitude vector output in the current iteration; Step 304: Increment the iteration count by 1. If the iteration converges or reaches the preset maximum number of iterations, return to step 301. Otherwise, use the output of the last iteration as the estimated value of the complex amplitude vector.

3. The anti-multiple access interference acquisition method based on multidimensional adaptive filtering as described in claim 2, characterized in that, In step 1, the energy accumulation value in each search grid is... in, Index for the three-dimensional search unit of the receiving channel, delay, and frequency. Simplified representation, search unit index middle, n For the receive channel index, l For the index of the delayed dimension search unit, k For the index of the frequency dimension search unit, L The number of delayed-dimensional search units, The number of search units in the frequency dimension; the superscript H indicates conjugate transpose; P For non-coherent accumulation counts, and The first p The and the first p +1 non-coherent accumulation time period of received signal sampling sequence, Represented as ； in, The superscript T in the text indicates transpose. This refers to the number of sampling points within the coherent accumulation time period, where the coherent accumulation time is equal to the code period. i This represents the transmitter index, where p is the p-th coherent accumulation time period. The delay of the i-th spread spectrum signal, For Doppler frequency, N Indicates the number of transmitters. and These represent the baseband characteristic waveforms when there is an unsigned bit transition and when there is a sign bit transition, respectively. The waveform matrix represents the three-dimensional characteristics of the delay, frequency, and data. The combined Gaussian white noise of orthogonal and in-phase branches. and Represent the complex amplitudes for unsigned transitions and signed transitions, respectively, with the following expressions: ； in Indicates the first i The product of the amplitude and phase of the spread spectrum signal. Indicates the first i Data symbols of the spread spectrum signal, Accumulate time for coherence, p For the first p A coherent cumulative time period, The delay of the i-th spread spectrum signal, Represents a complex magnitude vector; Indicates the search unit The matched filter coefficients in the formula are expressed as follows: in To delay segmented search indexing, Indicates rounding down. and They represent A vector whose columns are all zeros.

4. The anti-multiple access interference acquisition method based on multidimensional adaptive filtering as described in claim 3, characterized in that, The adaptive filtering is initiated by calculating and outputting the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal, specifically as follows: The calculated delay-frequency two-dimensional correlation results are as follows: ; In the formula, the subscript Acor indicates the relevant operation, the subscript m is the m-th channel of the detected signal, and the subscript p indicates the p-th signal segment; The characteristic waveform matrix used for complex amplitude estimation is expressed as follows: in for The number of one-sided correlators in the delay dimension, The number of frequency-dimensional single-sided correlators. This is a delayed segmented search index.

5. The anti-multiple access interference acquisition method based on multidimensional adaptive filtering as described in claim 4, characterized in that, The adaptive calculation m The filter coefficients for each channel in each iteration of filtering, specifically: Calculate the first m The first channel j The filter coefficients of the next iteration of filtering in For the first m Road spread spectrum signal and the first i The correlation coefficient matrix of the baseband characteristic waveform of the signal. The input is from the delay-frequency two-dimensional correlation function calculation module; For noise variance, The matrix is ​​related to the second moment of the complex amplitude. L s This refers to the number of sampling points within the coherent accumulation time period; specifically... Based on normalized energy accumulation value set up Used as the initial value for the complex amplitude vector and the filter coefficients are calculated. Subsequently, the complex magnitude vector estimated by the output of the j-th iteration is used. Calculate the filter coefficients , for Take the conjugate, where the superscript j is the iteration number and the subscript j is the subscript number. m For the first m Road spread spectrum signal, subscript p For the first p A coherent cumulative time period; The expression is .

6. The anti-multiple access interference acquisition method based on multidimensional adaptive filtering as described in claim 5, characterized in that, The filtering of the delay-frequency two-dimensional correlation results in each channel of the detected signal to obtain the estimated value of the complex amplitude vector output in the current iteration is as follows: The delay-frequency two-dimensional correlation result calculated for each channel of the detected signal is the... m The first channel j The filter coefficients of the next iteration of filtering ,for Perform filtering to obtain the first... j The estimated value of the complex amplitude vector of the signal output in the next iteration is : .

7. The anti-multiple access interference acquisition method based on multidimensional adaptive filtering as described in claim 6, characterized in that, Step 4: Reconstruct all detected interference signals based on the complex amplitude vector estimate, and remove all reconstructed signals from the received signal. The specific calculation formula is as follows: in This indicates the received signal after removing all reconstructed signals. Feedback is sent to step 1 and used in subsequent execution processes. replace Repeat steps 1 through 4 until the capture ends; This is an estimate of the delay time; This is an estimate of the Doppler frequency; This is an estimate of the complex magnitude vector; It is a three-dimensional feature waveform matrix. These are the delay estimate and Doppler estimate of the m-th spread spectrum signal, respectively.

8. A multi-dimensional adaptive filtering-based anti-multiple access interference acquisition system, characterized in that, The delay search interval is used in each receive channel of the DSSS receiver. and frequency search interval The range of delay and frequency uncertainty is divided into a two-dimensional search grid, where each grid is denoted as a search unit; The anti-multiple access interference acquisition system includes a matched filtering and non-coherent accumulation module, a detection and decision module, an open-loop adaptive filtering module, and an interference reconstruction and cancellation module; The matched filtering and non-coherent accumulation module uses zero-padding to perform zero-padding matched filtering and non-coherent accumulation on the received signal in the search grid of each undetected receiving channel to obtain the energy accumulation value in each search unit. The detection decision module uses the peak value of the energy accumulation value in the search unit of each channel as the detection decision value. When the detection decision value is higher than the set detection threshold, it is considered that the channel has detected a signal. After completing one round of signal detection, the signal detection count is incremented by 1, and the set of channels N for which no signals were detected and the set of channels M for which signals were detected are updated according to the detection results; initially, N is set to the entire set and M is set to the empty set. If the target signal has been detected, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has been successfully acquired; if the target signal is not detected when the signal detection count reaches the preset upper limit, the DSSS receiver ends the acquisition and outputs a result indicating that the target signal has failed to be acquired. If the signal detection count has not reached the preset upper limit and the target signal has not been detected, the signal acquisition has not ended, and the open-loop adaptive filtering module is started. The open-loop adaptive filtering module estimates the complex amplitude vector of the signal using adaptive filtering in the search unit of the channel where the target signal is detected. The interference reconstruction and elimination module reconstructs all detected interference signals based on the complex amplitude vector estimate, removes all reconstructed signals from the received signal, and feeds them back to the matched filtering and non-coherent accumulation module.

9. The anti-multiple access interference acquisition system based on multidimensional adaptive filtering as described in claim 8, characterized in that, The open-loop adaptive filtering module includes a delay-frequency two-dimensional correlation function calculation unit, a filtering unit, and a filter coefficient calculation unit. The delay-frequency two-dimensional correlation function calculation unit calculates the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal, and outputs it to the filtering unit. The filter coefficient calculation unit calculates the delay-frequency two-dimensional correlation result based on the characteristic waveform matrix and the input received signal, and outputs it to the filtering unit. The filtering unit filters the delay-frequency two-dimensional correlation results in each channel of the detected signal to obtain the estimated value of the complex amplitude vector output in the current iteration. One iteration is defined as one execution of the delay-frequency two-dimensional correlation function calculation unit, the filtering unit, and the filter coefficient calculation unit. An upper limit for the iteration is set, and the initial value of the iteration is set to 1. The number of iterations is incremented by 1. When the iteration converges or the preset maximum number of iterations is reached, the process returns to the delay-frequency two-dimensional correlation function calculation unit. Otherwise, the output of the last iteration is used as the estimated value of the complex amplitude vector for output.

Images