Low signal-to-noise ratio direct spread signal detection method based on noise cancellation

Abstract

The invention discloses a low signal-to-noise ratio direct spread signal detection method based on noise cancellation, and belongs to the field of communication spectrum sensing. The self-adaptive noise cancellation method based on the minimum mean square error is adopted, non-stationary noise interference can be tracked and eliminated in real time, filtering parameters are automatically optimized in an unknown channel environment, and a signal detection system can adapt to different background noise conditions. And stable signal detection is realized in a low signal-to-noise ratio environment by utilizing a cyclic spectrum analysis method and extracting the cyclic stability characteristic of the direct spread signal. The existence of the signal is judged through the characteristic spectrum peak on the non-zero cyclic frequency, and the carrier frequency and the pseudo code rate are further estimated, so that more accurate signal identification and parameter extraction are realized, the influence of noise uncertainty on the detection performance is avoided, and the detection robustness and reliability are improved. Welch smoothing processing and short-time Fourier transform are combined, the variance of spectrum estimation is reduced when the cyclic spectrum is calculated, and the detection robustness is improved.

Description

Technical Field

[0001] The present invention belongs to the field of communication spectrum sensing and relates to a signal detection method for non-cooperative conditions, and in particular to a method for detecting direct-spread spectral signals and estimating parameters under extremely low signal-to-noise ratio conditions. Background Art

[0002] Signal detection technology, a method for determining channel occupancy in non-cooperative communications, is a key technology in spectrum sensing. Direct-sequence spread spectrum (DSS) technology has been widely used in both civilian and military communications due to its strong anti-interference and anti-multipath capabilities, signal concealment, low interception rate, and ease of implementing code division multiple access. Therefore, spread spectrum technology is crucial for electronic countermeasures and civilian radio resource management. DSS signal detection is crucial for maintaining control of the electromagnetic spectrum on future battlefields.

[0003] Currently, several methods exist for detecting and estimating parameters of direct sequence spread spectrum signals. While these methods offer good results for a given parameter, their performance deteriorates at low signal-to-noise ratios. Therefore, effectively detecting and estimating direct sequence spread spectrum signals without prior knowledge is a field worthy of further research, with theoretical implications for communication countermeasures, military reconnaissance, electronic spectrum management, and interference identification.

[0004] Different methods can effectively estimate different DSSS signal parameters. Some methods are very effective at estimating specific DSSS signal parameters, but they can be computationally complex and computationally intensive. Furthermore, these methods face challenges such as achieving a lower signal-to-noise ratio (SNR) to accommodate the estimation of signal parameters buried in noise, which warrants further in-depth research.

[0005] Due to the signal stealth of DSSS technology, its non-cooperative detection faces enormous challenges, especially in low signal-to-noise ratio environments where the signal is drowned out by background noise, making detection and identification more difficult.

[0006] In recent years, with the rapid development of computer computing power, DSSS signal detection and parameter estimation technologies have advanced rapidly. Energy detection is the simplest method, but its model does not incorporate the exploitable characteristics of DSSS signals, making it relatively crude and susceptible to noise and interference. Correlation methods exploit the correlated delays of DSSS signals for detection, but are susceptible to noise, converge slowly under low signal-to-noise ratio conditions, and require the assistance of prior conditions. While higher-order spectra can reveal more information about the signal, improvements are needed to achieve lower signal-to-noise ratio tolerances. Cyclic spectra offer better detection performance but require a long period of accumulation.

[0007] Due to the uncertainty of random signals and the fact that DS signals are buried in noise, traditional detection methods require a long period of accumulation of DS signals to obtain the statistical characteristics of the signal under extremely low signal-to-noise ratios. 6 , under the condition of -20dB signal-to-noise ratio, it takes about 10 7 The lower the signal-to-noise ratio (SNR), the more sampling points are needed. Therefore, low SNR and limited sampling sequences are two major challenges for cyclic spectrum detection. Therefore, detecting finite-length DSSS signal sampling sequences at extremely low SNRs is a key issue that needs to be addressed urgently. Summary of the Invention

[0008] In view of the problems existing in existing energy detection methods, such as poor detection effect and long time consumption for extremely low signal-to-noise ratio direct-spread spectral signals, the purpose of the present invention is to provide a low signal-to-noise ratio direct-spread spectral signal detection method based on noise cancellation, which uses an LMS adaptive filter to cancel noise, improve the signal-to-noise ratio, and make the characteristics of the direct-spread spectral signal clearer in the frequency domain; due to the difference in cyclostationary characteristics between the direct-spread spectral signal and the noise, the signal and noise can be effectively distinguished in a low signal-to-noise ratio environment, and efficient detection can be completed with shorter sampling time and lower computational complexity, thereby improving the real-time performance and noise resistance of the direct-spread spectral signal detection system.

[0009] The purpose of the present invention is achieved through the following technical solutions.

[0010] The present invention discloses a low signal-to-noise ratio direct-spread signal detection method based on noise cancellation, which performs noise cancellation through an LMS adaptive filter, and uses iterative learning to automatically adjust weights to match interference noise. While improving the signal-to-noise ratio of the direct-spread signal, it makes it clearly present an outline in the frequency domain and forms a significant spectrum peak, thereby achieving efficient noise suppression. Furthermore, the cyclostationary characteristics of the direct-spread signal are utilized to distinguish the signal from the noise by calculating the cyclic spectrum, and to extract the target signal characteristics to improve the detection accuracy in a low signal-to-noise ratio environment. On this basis, based on the specific frequency characteristics of the cyclic spectrum, the existence of the direct-spread signal is determined, and the parameters of the carrier frequency and the pseudo-code rate are estimated. Compared with traditional energy detection methods, the present invention has stronger anti-interference ability and higher detection stability in a low signal-to-noise ratio environment, and can improve the detection accuracy and reliability of the direct-spread signal.

[0011] The present invention discloses a low signal-to-noise ratio direct-spread spectral (DSS) signal detection method based on noise cancellation, comprising the following steps:

[0012] Step 1: Construct a noise cancellation method based on minimum mean square error adaptive filtering. By constructing an adaptive filter, the noise is estimated in real time, and the filter parameters are adaptively adjusted using the error minimization criterion. Through noise cancellation based on the adaptive filter LMS, the weights are automatically adjusted through iterative learning to match the interference noise, thereby enhancing the characteristics of the direct-spread signal, improving the signal-to-noise ratio of the direct-spread signal, clearly presenting the direct-spread signal profile in the frequency domain and forming a significant spectral peak, thereby achieving efficient noise suppression.

[0013] In low signal-to-noise ratio environments, DSSS signals are easily masked by background noise, resulting in reduced detection accuracy. This paper proposes a noise cancellation method based on minimum mean square error (MMSE) adaptive filtering. By constructing an adaptive filter to estimate noise in real time and adaptively adjusting the filter parameters using an error minimization criterion, the output signal is minimized from noise interference, thereby enhancing the characteristics of the DSSS signal and improving the accuracy of subsequent cyclic spectrum detection. This method can dynamically adjust the filter parameters through recursive calculations even when signal characteristics are unknown, achieving adaptive noise suppression and improving the signal-to-noise ratio of the sampling sequence.

[0014] The adaptive noise canceller has two input channels: a main input channel and a reference input channel. The main input channel contains the useful signal s(n) and the mutually uncorrelated interference noise n0(n), while the reference input channel contains only the noise n1(n) that is correlated with n0(n) but uncorrelated with s(n). The DSSS signal detection system adjusts the tap weight ω through the adaptive filter. i (n), so that the filter output y(n) approaches n0(n), and the optimized signal estimation value is obtained by error calculation:

[0015] e(n)=d(n)-y(n)=s(n)+n0(n)-y(n) (1)

[0016] The mean square error E[e 2 ] to minimize:

[0017] E[e 2 ]=E[(s+n0-y) 2 ]=E[s 2 ]+E[(n0-y) 2 ] (2)

[0018] When the filter output y(n) optimally approximates n0(n), E[e 2 ] is minimal, and the output signal e(n) of the DSSS signal detection system is mainly composed of s(n), achieving effective noise reduction. The specific implementation method of adaptive noise cancellation is:

[0019] Step 101: Initialize filter parameters, with initial tap weight ω(0)=0.

[0020] Step 102: The filter tap weights at time n are ω(n)=[ω0(n),ω1(n),...,ω M-1 (n)] T , the input vector is represented as u=[u(n),u(n-1),...,u(n-M+1)] T , calculate the output of the LMS adaptive filter

[0021] y(n)=ω H (n)u(n) (3)

[0022] Step 103, calculate the error between the main channel signal d(n) and the filter output:

[0023] e(n)=d(n)-y(n)=d(n)-ω H (n)u(n) (4)

[0024] The error signal e(n) reflects the noise components that have not been eliminated by the filter.

[0025] Step 104: Update the tap weights. When the prior information of the unknown signal is known, the steepest descent algorithm is used to adjust the filter weight vector along the steepest descent direction of the performance surface, and the minimum point of the performance surface is searched. The iterative formula for calculating the weight vector is:

[0026]

[0027] Where μ is the step size factor, is the descent gradient, and its calculation formula is:

[0028]

[0029] Step 105, error correction and convergence judgment. If the mean square error E[e 2 ]=E[(s+n0-y) 2 ] is less than the set threshold, stop the iteration; otherwise, continue to adjust the filter weights.

[0030] Step 106, repeating steps 102-105 until the system output signal e(n) is mainly composed of s(n), that is, the noise suppression reaches the optimal state.

[0031] This step dynamically adjusts the weights through an iterative learning mechanism, so that the filter output optimally approximates the interference noise, achieving real-time adaptive suppression of noise interference, thereby effectively improving the signal-to-noise ratio of the DS-SS signal, making the DS-SS signal clearly outlined in the frequency domain and forming a significant spectral peak at the carrier frequency, ultimately achieving efficient and accurate noise suppression and enhancing signal detection capabilities in low signal-to-noise ratio environments.

[0032] Step 2 leverages the cyclostationary properties of the DSSS signal to distinguish between signal and noise and extract the target signal's characteristics by calculating the cyclic spectrum. The adaptive noise cancellation in step 1 already improves the signal-to-noise ratio. This step further extracts the cyclic characteristics of the DSSS signal in the frequency domain, improving detection accuracy in low signal-to-noise ratio environments.

[0033] Since the DSSS signal is modulated by pseudo-random noise (PN) sequence, the cyclic frequency corresponding to the carrier frequency is α The energy distribution at the cyclic frequency f is stable, while the traditional Gaussian white noise has no significant component at the non-zero cyclic frequency f. Therefore, cyclic spectrum analysis can effectively highlight the DSSS signal while suppressing the noise, making the signal detection more stable and reliable.

[0034] The present invention converts the signal with improved signal-to-noise ratio in step 1 into the cyclic spectrum domain, and then uses the cyclic frequency dimension to further suppress noise, thereby significantly improving the detection accuracy of the direct-spread signal in a low signal-to-noise ratio environment.

[0035] Due to the instability of random signals, the variance is large. This paper uses the Welch algorithm to divide a long sequence into multiple fixed-length random sequences, calculate the cyclic spectrum of each, and finally average it. This segmented averaging method gives the resulting cyclic spectrum certain statistical characteristics, reduces the variance, and achieves a smoothing effect in the time domain.

[0036] The discrete implementation of the spectral correlation function is obtained by replacing the corresponding quantities in the continuous quantity formula of the cyclic spectral correlation function with discrete values:

[0037]

[0038] The specific implementation steps of step 2 are as follows:

[0039] Step 201: Sampling sequence preprocessing.

[0040] To improve the stability of frequency domain estimation, the input sampling sequence is segmented. When the total number of sampling points is N, it is divided into N1 groups, each with N2 data points:

[0041]

[0042] This segmentation strategy uses the Welch smoothing method for preprocessing, which effectively reduces the variance of spectrum estimation and improves signal stability.

[0043] Step 202: Short-time Fourier transform (STFT).

[0044] Perform a Fast Fourier Transform (FFT) on each set of data to convert the time domain signal to the frequency domain:

[0045]

[0046] The transformation is used to obtain the energy distribution of the signal at different frequencies, providing a basis for subsequent cyclic spectrum calculation.

[0047] Step 203: Calculate the spectral correlation function.

[0048] In cyclic spectrum calculation, the periodic characteristics of the signal can be extracted through the correlation between different frequency components. Calculate the cross-correlation of different frequency components:

[0049]

[0050] At this time, if α≠0, the spectral correlation of white noise tends to 0, while the DS signal will be 0 at α=2f due to its periodicity. c There is a significant spectrum peak at the carrier frequency.

[0051] Step 204: Calculate the cyclic spectrum and output it.

[0052] The spectral correlation functions calculated for all segments are averaged to obtain the final cyclic spectrum estimate:

[0053]

[0054] The statistical stability of the cyclic spectrum is ensured by obtaining the final cyclic spectrum estimation value, making the signal detection more robust.

[0055] This step uses cyclic spectrum analysis to clearly reveal the characteristics of the DSSS signal at non-zero cyclic frequencies. Since noise lacks periodicity, its energy distribution at these frequencies tends to zero. This characteristic enables cyclic spectrum analysis to accurately extract key DSSS signal parameters, including carrier frequency and spreading code rate, even in low signal-to-noise ratio environments, effectively improving detection performance.

[0056] Step 3: Determine the presence of a DSSS signal based on the specific frequency characteristics of the cyclic spectrum and perform parameter estimation. The parameter estimation result is the DSSS signal detection result at low signal-to-noise ratio (SNR). This means that DSSS signal detection at low SNR is achieved based on noise cancellation. The parameters include carrier frequency and pseudo-code rate.

[0057] In the preceding steps, adaptive noise cancellation improves the signal-to-noise ratio (SNR), and the cyclic spectrum enhances the characteristics of the DSSS signal. Next, based on the cyclic spectrum characteristics, the presence of the DSSS signal is determined, and parameters such as the carrier frequency and pseudo-code rate are extracted. Traditional energy detection methods are susceptible to noise interference at low SNRs, while the cyclic spectrum-based method effectively distinguishes signal from noise, making detection more stable and reliable.

[0058] In the cyclic spectrum, the DSSS signal, due to its periodicity, forms a distinct peak at the cyclic frequency α = 2f0, twice the carrier frequency, while Gaussian white noise is concentrated at the zero cyclic frequency. Therefore, by analyzing the distribution of the cyclic spectrum at different cyclic frequency dimensions, we can determine whether the signal exists.

[0059] Step 301: Obtain a two-dimensional slice of the cyclic spectrum and perform feature extraction.

[0060] Since the energy of the DSSS signal is concentrated at the cyclic frequency α=2f0, and Gaussian white noise only exists at the zero cyclic frequency, the distribution characteristics of the cyclic spectrum in different frequency dimensions can be used to detect the DSSS signal.

[0061] From the calculated three-dimensional cyclic spectrum, extract the two-dimensional slice at frequency f = 0, that is:

[0062] |S x (α,0)| (12)

[0063] In this slice, the DS signal shows a significant spectrum peak at the non-zero cycle frequency, while the noise part is concentrated at the zero cycle frequency.

[0064] Step 302: Signal detection and decision making.

[0065] Set the detection threshold and search for the maximum value S on the cycle frequency axis max , and calculate the average energy S of the cyclic spectrum avg .

[0066] In order to improve the accuracy of the decision, the peak-to-average ratio is calculated:

[0067]

[0068] And with the set decision threshold P th If the value exceeds the preset threshold, it is determined that there is a DS signal in the sampling sequence; if it does not reach the threshold, it is determined that the detection has failed and the current signal detection process is terminated.

[0069] Compared to directly comparing the maximum spectral peaks, using peak-to-average ratio for judgment can more effectively avoid the influence of background noise and reduce the possibility of false detection. This method is particularly suitable for signal detection in low signal-to-noise ratio environments. This method utilizes the periodic characteristics of the signal to avoid the problem that traditional energy detection methods are susceptible to noise uncertainty in low signal-to-noise ratio environments.

[0070] Step 303: Estimate the carrier frequency of the DS signal.

[0071] The maximum value S obtained in step 302 max The corresponding cycle frequency is 2f0:

[0072]

[0073] where α max is the maximum value of the cyclic spectrum S max The corresponding cycle frequency.

[0074] Based on the peak characteristics of the cyclic spectrum, the carrier frequency is estimated blindly without prior knowledge.

[0075] Step 304: Estimate the pseudo code rate.

[0076] According to the signal carrier frequency f0 estimated in step 303, another two-dimensional slice in the cyclic spectrum is selected, that is, the cyclic spectrum slice at frequency f=f0:

[0077] |S x (α,f0)| (15)

[0078] In this slice, the pseudo code rate R of the DS signal c This is reflected in the peak cyclic frequency of the slice maximum value:

[0079] R c =α max (16)

[0080] This step utilizes a cyclic spectrum-based signal detection method to overcome the limitations of traditional energy detection methods at low signal-to-noise ratios, making DSSS signal detection more accurate. Extracting the carrier frequency and pseudo-code rate requires no prior information, making it suitable for signal analysis in complex environments. Combined with the aforementioned steps, the complete signal detection method can accurately isolate DSSS signals in strong noise environments and estimate key parameters, providing reliable data for subsequent communication or signal processing.

[0081] Beneficial effects:

[0082] 1. This invention discloses a noise cancellation-based method for detecting low-SNR DS-S signals. This method employs an adaptive noise cancellation method based on minimum mean square error (MMSE). This method tracks and eliminates nonstationary noise interference in real time, automatically optimizing filtering parameters in unknown channel environments and enabling the signal detection system to adapt to varying background noise conditions. Compared to traditional fixed filtering methods, this adaptive filtering strategy minimizes the impact of noise without losing target signal characteristics, improving detection sensitivity and enabling the system to detect even weaker DS-S signals.

[0083] 2. This invention discloses a low-SNR DS-type signal detection method based on noise cancellation. This method utilizes cyclic spectrum analysis to extract the cyclostationary characteristics of DS-type signals, enabling robust signal detection in low SNR environments. Unlike traditional energy detection methods that rely on signal strength, this method determines the presence of a signal using characteristic spectrum peaks at non-zero cyclic frequencies, and further estimates the carrier frequency and pseudo-code rate. This allows for more accurate signal identification and parameter extraction, mitigates the impact of noise uncertainty on detection performance, and improves detection robustness and reliability.

[0084] 3. The present invention discloses a low-SNR DS-Signal Spread Spectrum (DSS) signal detection method based on noise cancellation. This method, combined with Welch smoothing and short-time Fourier transform (SFT), reduces the variance of spectrum estimation when calculating cyclic spectra, thereby improving detection robustness. Even in extremely low SNR environments, the system maintains good detection performance, avoiding the signal overwhelm caused by noise dominance in traditional methods and improving the success rate of signal detection. Furthermore, the peak-to-average ratio (PAR) decision method optimizes the signal detection strategy, effectively distinguishing DS-Signal Spread Spectrum (DSS) signals from random noise interference, reducing false detection rates compared to the traditional maximum peak threshold method. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 This is a general flow chart of the DSSS signal detection method under low signal-to-noise ratio based on noise cancellation disclosed in the present invention;

[0086] Figure 2 is a structural diagram of a system for noise cancellation in the present invention;

[0087] Figure 3 This is a block diagram of the LMS adaptive filter principle in the present invention;

[0088] Figure 4 This is a diagram showing the effect of the noise canceller of the present invention on a direct-spread signal with a signal-to-noise ratio of -15dB. Figure 4 (a) is the original signal, Figure 4 (b) is the power spectrum density of the DSSS signal after noise cancellation;

[0089] Figure 5 The present invention is Figure 5 The 3D image of cyclic spectral density obtained by processing the noise-reduced signal;

[0090] Figure 6 It is a two-dimensional image obtained by performing specific frequency slicing on the three-dimensional image of the cyclic spectral density in the present invention, Figure 6 (a) The figure is a two-dimensional slice of f = 0, Figure 6 (b) The figure shows f = f c Two-dimensional slices. DETAILED DESCRIPTION

[0091] In order to better illustrate the purpose and advantages of the present invention, the invention is further described below with reference to the accompanying drawings and examples.

[0092] Example 1:

[0093] The technical problem to be solved by the noise cancellation-based cyclic spectrum detection method for DSSS signals under low SNR conditions in this embodiment is to detect a finite-length DSSS signal sampling sequence under extremely low SNR conditions.

[0094] The implementation parameters are as follows: the number of information code bits is 200, the length of the pseudo-random sequence is 31, the pseudo-code rate is 46.5kHz, the signal code rate is 1kHz, the carrier frequency is 93kHz, the sampling rate is 744kHz, the signal-to-noise ratio is -15dB, and the power spectrum of the original sampling sequence is as follows: Figure 4 As shown in (a), the DSSS signal will be submerged in the noise when the signal-to-noise ratio is -15dB, and the spectrum peak of the DSSS signal at the carrier frequency cannot be observed.

[0095] like Figure 1 As shown, this embodiment discloses a low signal-to-noise ratio DS-SNR signal detection method based on noise cancellation, and the specific implementation steps are as follows:

[0096] Step 1: Construct a noise cancellation method based on minimum mean square error adaptive filtering. By constructing an adaptive filter, the noise is estimated in real time, and the filter parameters are adaptively adjusted using the error minimization criterion. Through noise cancellation based on the adaptive filter LMS, the weights are automatically adjusted through iterative learning to match the interference noise, thereby enhancing the characteristics of the direct-spread signal, improving the signal-to-noise ratio of the direct-spread signal, clearly presenting the direct-spread signal profile in the frequency domain and forming a significant spectral peak, thereby achieving efficient noise suppression.

[0097] The adaptive noise canceller contains two input channels: the main input channel and the reference input channel. Figure 2 As shown in Figure 2, the main input channel contains the useful signal s(n) and the uncorrelated interference noise n0(n), while the reference input channel only contains the noise n1(n) that is correlated with n0(n) but uncorrelated with s(n). The DSSS signal detection system adjusts the tap weight ω through the adaptive filter. i (n), so that the filter output y(n) approaches n0(n), and the optimized signal estimation value is obtained by error calculation:

[0098] e(n)=d(n)-y(n)=s(n)+n0(n)-y(n) (17)

[0099] Since s(n) is independent of n0(n) and y(n), we can calculate the mean square error of the above formula and get

[0100] E[e 2 ]=E[(s+n0-y)2 ]=E[s 2 ]+E[(n0-y) 2 ] (18)

[0101] To minimize the mean square error, it is necessary to adjust the tap weight ω of the LMS adaptive filter. i (n), which minimizes the mean square error, is

[0102] minE[e 2 ]=E[s 2 ]+minE[(n0-y) 2 ] (19)

[0103] The above equation shows that y(n) achieves the best estimate of the noise n0(n) in the main channel d, with its mean square error approaching zero. The error e(n) = s(n) + n0(n) - y(n) is statistically close to s(n). As a result, the system's output signal e(n) approaches the desired signal s(n), significantly reducing the noise in the system's output signal.

[0104] The adaptive filter structure is as follows Figure 3 As shown, the adaptive filter order is set to 12. The filter tap weights at time n The input vector is represented as d(n) is the main channel input signal, y(n) is the expected output response of the filter, e(n) is the error output signal, and M is the length of the filter. It is impossible to achieve optimal filter weighting when the prior knowledge of the signal is unknown. The instantaneous gradient value will be used instead of the expectation factor in the steepest descent method.

[0105] Step 1: The specific implementation method of adaptive noise cancellation is:

[0106] Step 101: Initialize the tap weight ω(0)=0.

[0107] Step 102: At time n, calculate the output of the LMS adaptive filter

[0108] y(n)=ω H (n)u(n) (20)

[0109] Step 103: Estimate the error at the current time n

[0110] e(n)=d(n)-y(n)=d(n)-ω H (n)u(n) (21)

[0111] The error signal e(n) reflects the noise components that have not been eliminated by the filter.

[0112] Step 104: Update tap weights

[0113]

[0114] Among them, μ is the step size factor, which is used to control the convergence rate and stability. In the implementation stage, the step size factor μ is set to 0.0008. is the descent gradient, and its calculation formula is:

[0115]

[0116] Step 105, error correction and convergence judgment. If the mean square error E[e 2 ]=E[(s+n0-y) 2 ] is less than the set threshold, stop the iteration; otherwise, continue to adjust the filter weights.

[0117] Step 106, repeating steps 102-105 until the system output signal e(n) is mainly composed of s(n), that is, the noise suppression reaches the optimal state.

[0118] This step dynamically adjusts the weights through an iterative learning mechanism, so that the filter output is optimally close to the interference noise, and realizes real-time adaptive suppression of noise interference, thereby effectively improving the signal-to-noise ratio of the DSSS signal, making the DSSS signal clearly present its outline in the frequency domain, and forming a significant spectrum peak at the carrier frequency, ultimately achieving efficient and accurate noise suppression and enhancing the signal detection capability in low signal-to-noise ratio environments. Figure 4 As shown, Figure 4 (a) is the original spectrum. Figure 4 (b) is the spectrum after processing in step 1. The signal-to-noise ratio has been significantly improved from the initial setting of -15dB to about 0dB.

[0119] Step 2 leverages the cyclostationary properties of the DSSS signal to distinguish between signal and noise and extract the target signal's characteristics by calculating the cyclic spectrum. The adaptive noise cancellation in step 1 already improves the signal-to-noise ratio. This step further extracts the cyclic characteristics of the DSSS signal in the frequency domain, improving detection accuracy in low signal-to-noise ratio environments.

[0120] The frequencies in the cyclic spectral density function extend from the coordinate domain of traditional power spectrum analysis to a dual-frequency coordinate system of spectral frequency and periodic frequency. In non-stationary signals, some spectral lines overlap on the power spectrum, and the spectral envelope obscures these characteristic lines. This prevents the detection of line spectrum peaks when using the power spectrum density function, and therefore prevents the eigenvalues ​​from being obtained. However, when using the cyclic spectral density function for analysis, where α ≠ 0, some of the signal's characteristics are re-expressed as discretely distributed periodic frequency lines. This allows the extraction of periodic information, such as the symbol rate, that is not available using power spectrum analysis.

[0121] When data is limited, noise instability often significantly impacts the cyclic spectrum estimation results, and the result can only be obtained through a single estimation. This results in large variance. Using the Welch algorithm, we divide the long sequence into multiple random segments of fixed length, calculate the cyclic spectrum for each, and then average it. This segmented averaging method can impart statistical properties to the resulting cyclic spectrum, reduce variance, and achieve a smoothing effect in the time domain.

[0122] By replacing the corresponding quantities in the continuous quantity formula of the cyclic spectral correlation function with discrete values, we can obtain the discrete implementation of the spectral correlation function:

[0123]

[0124] According to the formula, the general calculation steps for calculating the cyclic spectrum of the sampling sequence in step 2 are:

[0125] Step 201: Split the sequence after noise elimination in step 1. The total number of sampling points (99,200) is divided into 96 groups, each with 1,024 sequence points, and the redundant 896 points are discarded.

[0126] x 1024 (1,f+α / 2),x 1024 (2,f+α / 2),...,x 1024 (96,f+α / 2) (25)

[0127] Step 202: For each set of sampling points, perform a fixed-point FFT with a length of 1024.

[0128]

[0129] Step 203: Cross-multiply each frequency component to calculate a spectral correlation function value.

[0130]

[0131] Step 204 : To reduce the random influence, the spectrum correlation function results obtained from all groups are averaged to obtain a more accurate spectrum correlation function estimation value as the final output.

[0132]

[0133] The cyclic spectral density function obtained by the above steps is as follows Figure 5Peaks appear on both the cyclic frequency axis and the spectral frequency axis. The Gaussian stationary white noise superimposed on the signal has very little interference with the characteristic peaks of the signal at non-zero cyclic frequencies, as the cyclic spectral density of the stationary noise is mainly concentrated at the zero cyclic frequency. Especially in the presence of strong background noise, the signal spectrum at the zero cyclic frequency is often submerged in the noise. Therefore, it is effective to use the cyclic spectral density function to detect DS signals in noise, thus getting rid of the drawbacks of conventional signal energy detection methods.

[0134] Step 3: Determine the presence of a DSSS signal based on the specific frequency characteristics of the cyclic spectrum and perform parameter estimation. The parameter estimation result is the DSSS signal detection result at low signal-to-noise ratio (SNR). This means that DSSS signal detection at low SNR is achieved based on noise cancellation. The parameters include carrier frequency and pseudo-code rate.

[0135] In order to make the observation more intuitive and reduce the amount of computational simulation, it is necessary to slice the three-dimensional graph of the cyclic spectrum and perform a one-dimensional search on the non-zero cyclic frequency axis to effectively estimate the carrier frequency and pseudo-code rate.

[0136] Make a two-dimensional slice when f = 0, and we get

[0137]

[0138] Among them, T c is the spread spectrum symbol period of the direct spread signal, f0 is the carrier frequency of the signal, and Q(f) is the spectrum of the signal.

[0139] In the above formula, the zero-frequency slice is symmetrically distributed, so we only need to consider the case of α>0. When α=2f0, the maximum value appears. Therefore, the carrier frequency and pseudo code rate of the DSSS signal can be estimated by searching for the maximum and second maximum values ​​on the non-zero cyclic frequency axis at f = 0.

[0140] Make a two-dimensional slice when f=f0, and we get

[0141]

[0142] In the above formula, the carrier frequency slices are also symmetrically distributed, so we only need to consider the case where α>0. Therefore, the pseudo code rate of the DSSS signal can be estimated by searching for the maximum value on the non-zero cyclic frequency axis at f = f0.

[0143] According to the formula, the general calculation steps for step 3 to determine whether there is a DS signal and perform parameter estimation are:

[0144] Step 301: intercept the center plane of frequency f in the three-dimensional cyclic spectrum, which is the two-dimensional slice at f=0. Figure 6 As shown in (a).

[0145] Step 302: Set the signal-to-noise ratio index according to the current experimental environment and the test, set the threshold as the decision threshold for whether the signal exists, search for the maximum value on the cyclic frequency axis, and calculate the peak-to-average ratio.

[0146] By observing the two-dimensional slice at f = 0, if the peak-to-average ratio is greater than the set threshold, the presence of a DS-type signal in the sampling sequence is determined. Clearly symmetrical peaks are present along the cyclic frequency axis. This confirms the presence of a DS-type signal in the sampling sequence, consistent with the experimental settings.

[0147] In step 303, the cyclic frequency corresponding to the maximum value obtained in step 302 is twice the carrier frequency, i.e., 2f0. In this experiment, 2f0 = 186 kHz. Dividing the frequency corresponding to the peak by 2 yields the signal's carrier frequency, f0 = 93 kHz.

[0148] Step 304: intercept the two-dimensional slice of the three-dimensional cyclic spectrum at frequency f=93kHz, such as Figure 6 (b) is shown. Search for the maximum value on the cyclic frequency axis. The cyclic frequency value corresponding to the maximum value is the pseudo code rate R c Therefore, the pseudo code rate R is obtained c =46.5kHz, which is consistent with the experimental design.

[0149] So far, the above example uses a low SNR DSSS signal detection method based on noise cancellation to complete DSSS signal detection and parameter estimation for a limited short sequence at extremely low SNR. The flow chart is as follows: Figure 1 As shown in the figure, accurate detection was achieved for a signal-to-noise ratio of -15dB, a short sequence of 200 information code bits, and a pseudo-random sequence length of 31. This example demonstrates that the inventive method has excellent detection results for short sequence signals at extremely low signal-to-noise ratios, and the capability demonstrated in this example is far beyond the detection limit.

[0150] The above specific description further illustrates the purpose and technical solution of the invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is 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 in the scope of protection of the present invention.

Claims

1. A method for detecting low signal-to-noise ratio (SNR) DS-SS signals based on noise cancellation, characterized by: The following steps are included: Step 1: Construct a noise cancellation method based on minimum mean square error adaptive filtering. By constructing an adaptive filter to estimate the noise in real time, and using the error minimization criterion to adaptively adjust the filter parameters, through noise cancellation based on the adaptive filter LMS, the weights are automatically adjusted through iterative learning to match the interference noise, thereby enhancing the characteristics of the DS-S signal, improving the DS-S signal noise ratio, clearly presenting the DS-S signal profile in the frequency domain and forming a significant spectral peak, achieving efficient noise suppression; Step 2: Using the cyclostationary characteristics of the DSSS signal, the cyclic spectrum is calculated to distinguish the signal from the noise and extract the target signal characteristics; Step 3: Determine the presence of a DS-SS signal based on the specific frequency characteristics of the cyclic spectrum and perform parameter estimation. The parameter estimation result is the DS-SS signal detection result under low signal-to-noise ratio, that is, DS-SS signal detection under low signal-to-noise ratio is achieved based on noise cancellation. The parameters include carrier frequency and pseudo-code rate. Step 301: Obtain a two-dimensional slice of the cyclic spectrum |S x (α,0)| and perform feature extraction; Step 302: Detect and judge the signal using the peak-to-average ratio. Step 303: estimating the carrier frequency of the DS signal; Step 304: Estimate the pseudo code rate.

2. The method for detecting a low signal-to-noise ratio (SNR) DS-SS signal based on noise cancellation according to claim 1, wherein: The adaptive noise canceller contains two input channels: a main input channel and a reference input channel; The main input channel contains the useful signal s(n) and the uncorrelated interference noise n0(n), while the reference input channel only contains the noise n1(n) that is correlated with n0(n) but uncorrelated with s(n). The DSSS signal detection system adjusts the tap weight ω through an adaptive filter. i (n), so that the filter output y(n) approaches n0(n), and the optimized signal estimation value is obtained by error calculation: e(n)=d(n)-y(n)=s(n)+n0(n)-y(n) (1) The mean square error E[e 2 ] to minimize: E[e 2 ]=E[(s+n0-y) 2 ]=E[s 2 ]+E[(n0-y) 2 ] (2) When the filter output y(n) optimally approximates n0(n), E[e 2 ] is minimum, and the output signal e(n) of the direct-spread signal detection system is mainly composed of s(n), achieving effective noise reduction.

3. The method for detecting low signal-to-noise ratio DS-SNR signals based on noise cancellation according to claim 2, wherein: The specific implementation method of the noise cancellation method of the minimum mean square error adaptive filtering in step 1 is: Step 101, initializing filter parameters, initial tap weight ω(0)=0; Step 102: The filter tap weights at time n are ω(n)=[ω0(n),ω1(n),...,ω M-1 (n)] T , the input vector is represented as u=[u(n),u(n-1),...,u(n-M+1)] T , calculate the output of the LMS adaptive filter y(n)=ω H (n)u(n) (3) Step 103, calculate the error between the main channel signal d(n) and the filter output: e(n)=d(n)-y(n)=d(n)-ω H (n)u(n) (4) The error signal e(n) reflects the noise components that have not been eliminated by the filter; Step 104, tap weight update; when the prior information of the unknown signal is known, the steepest descent algorithm is used to adjust the filter weight vector along the direction of steepest descent of the performance surface, and the minimum point of the performance surface is searched. The iterative formula for calculating the weight vector is: Where μ is the step size factor, is the descent gradient, and its calculation formula is: Step 105, error correction and convergence judgment; if the mean square error E[e 2 ]=E[(s+n0-y) 2 ] is less than the set threshold, stop the iteration; otherwise, continue to adjust the filter weights; Step 106, repeating steps 102-105 until the system output signal e(n) is mainly composed of s(n), that is, the noise suppression reaches the optimal state.

4. The method for detecting a low signal-to-noise ratio (SNR) DS-SS signal based on noise cancellation according to claim 3, wherein: In step 2, Using the Welch algorithm, the long sequence is divided into multiple random sequences of fixed length, and the cyclic spectrum is calculated for each of them, and finally the average is calculated. Through the segmented averaging method, the final cyclic spectrum has statistical characteristics, the variance is reduced, and the effect of time domain smoothing is achieved. The discrete implementation of the spectral correlation function is obtained by replacing the corresponding quantities in the continuous quantity formula of the cyclic spectral correlation function with discrete values:

5. The method for detecting a low signal-to-noise ratio DS-SNR signal based on noise cancellation according to claim 4, wherein: The specific implementation steps of step 2 are as follows: Step 201: sampling sequence preprocessing; In order to improve the stability of frequency domain estimation, the input sampling sequence is segmented; when the total number of sampling points is N, it is divided into N1 groups, each with N2 data points: This segmentation strategy uses the Welch smoothing method for preprocessing to reduce the variance of spectrum estimation and improve signal stability; Step 202: Short-time Fourier transform; Perform a fast Fourier transform on each set of data to convert the time domain signal to the frequency domain: The transformation is used to obtain the energy distribution of the signal at different frequencies; Step 203: Calculate the spectral correlation function; In cyclic spectrum calculation, the periodic characteristics of the signal can be extracted through the correlation between different frequency components; Compute the cross-correlations of different frequency components: At this time, if α≠0, the spectral correlation of white noise tends to 0, while the DS signal will be 0 at α=2f due to its periodicity. c There is a significant spectrum peak at the carrier frequency; Step 204: Calculate the cyclic spectrum and output it; The spectral correlation functions calculated for all segments are averaged to obtain the final cyclic spectrum estimate: The statistical stability of the cyclic spectrum is ensured by obtaining the final cyclic spectrum estimation value, making the signal detection more robust.

6. The method for detecting a low signal-to-noise ratio (SNR) DS-SS signal based on noise cancellation according to claim 1, wherein: Step 301: Obtaining a two-dimensional slice of the cyclic spectrum |S x (α,0)| and the method for feature extraction is as follows, Since the energy of the DS signal is at the cycle frequency α=2f c The cyclic spectrum is concentrated at , while the Gaussian white noise only exists at the zero cyclic frequency. Therefore, the distribution characteristics of the cyclic spectrum in different frequency dimensions are used to detect the direct spread signal. From the calculated three-dimensional cyclic spectrum, extract the two-dimensional slice at frequency f = 0, that is: |S x (α,0)| (12) In this slice, the DS signal shows a significant spectrum peak at the non-zero cycle frequency, while the noise part is concentrated at the zero cycle frequency.

7. The method for detecting low signal-to-noise ratio DS-SNR signals based on noise cancellation according to claim 1, wherein: The method for signal detection and judgment using peak-to-average ratio in step 302 is as follows: Set the detection threshold and search for the maximum value S on the cycle frequency axis max , and calculate the average energy S of the cyclic spectrum avg ; In order to improve the accuracy of the decision, the peak-to-average ratio is calculated: And with the set decision threshold P th Make comparisons; If the threshold is exceeded, it is determined that there is a DS signal in the sampling sequence; if the threshold is not reached, it is determined that the detection has failed and the current signal detection process is terminated.

8. The method for detecting low signal-to-noise ratio DS-SNR signals based on noise cancellation according to claim 1, wherein: The method for estimating the carrier frequency of the DSSS signal in step 303 is as follows: The maximum value S obtained in step 302 max The corresponding cycle frequency is 2f c : where α max is the maximum value of the cyclic spectrum S max The corresponding cycle frequency; Based on the peak characteristics of the cyclic spectrum, the carrier frequency is estimated blindly without prior knowledge.

9. The method for detecting low signal-to-noise ratio DS-SNR signals based on noise cancellation according to claim 1, wherein: The method for estimating the pseudo code rate in step 304 is as follows: The signal carrier frequency f estimated in step 303 c , select another two-dimensional slice in the cyclic spectrum, that is, frequency f = f c Cyclic spectrum slice at : |S x (a,f c )| (15) In this slice, the pseudo code rate R of the DS signal c This is reflected in the peak cyclic frequency of the slice maximum value: R c ＝α max (16)

Images