Narrowband interference elimination method based on spread spectrum communication

Abstract

The invention relates to the technical field of satellite transmission service or radio transmission service of a space station or an airborne station, in particular to a narrow-band interference elimination method based on spread spectrum communication applied to satellite transmission service, which comprises the following steps: S1, performing FFT (Fast Fourier Transform) operation on a digital baseband signal, converting the digital baseband signal into a frequency domain signal, determining a judgment threshold according to the background noise power, and judging a component greater than the threshold as an interference signal; s2, based on a sequential return-to-zero algorithm or a sequential pruning algorithm, frequency points with interference signal components in each frequency domain under the AWGN condition are cut off; s3, using windowed FFT operation, and popularizing the application under the AWGN condition in the S2 to the application under the narrowband interference condition; and S4, reducing the frequency domain signal after the interference signal component frequency points are removed into a time domain signal by adopting IFFT (Inverse Fast Fourier Transform) operation. The method has a good application prospect in the fields of manufacturing of space-based satellite systems and other spacecrafts, manufacturing of satellite antennas and other communication system equipment, satellite transmission services and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite transmission services or radio transmission services for space stations or airborne stations, and more particularly to a narrowband interference cancellation method based on spread spectrum communication applied to satellite transmission services. Background Art

[0002] With the continuous development of satellite communication technology, anti-jamming technology has gradually become one of the important research directions in the field of satellite communication. Interference signals will seriously affect the stability and reliability of satellite communication systems. Therefore, it is very necessary to improve the anti-jamming ability of satellite communication. In recent years, anti-jamming technology has developed rapidly and many achievements have been made. Generally speaking, there are mainly nulling antenna technology, waveform design with anti-jamming ability, hard and soft limiting technology in radio frequency or intermediate frequency, reducing information rate, using high-gain channel coding, baseband signal processing anti-jamming technology, etc.

[0003] There are mainly two determining factors for interference cancellation algorithms: interference detection and notch filtering. Interference detection is to determine which frequency points have interference. Notch filtering is to map the amplitude at the interference frequency points into new values based on interference detection to remove the interference energy. Different interference detection algorithms and notch filtering strategies determine different interference cancellation algorithms. However, the existing interference cancellation algorithms have problems such as poor real-time performance and accuracy, and large damage to useful signals in these two determining factors respectively. Summary of the Invention

[0004] The present invention provides a narrowband interference cancellation method based on spread spectrum communication with less computational effort, higher accuracy, and less damage to useful signals, which can solve at least one of the above technical problems.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0006] A narrowband interference cancellation method based on spread spectrum communication includes the following steps:

[0007] S1. Perform FFT operation on the digital baseband signal to transform the digital baseband signal into a frequency-domain signal, and determine a decision threshold according to the estimated background noise power. The components greater than this threshold are judged as interference signals;

[0008] S2. Based on the sequential nulling algorithm or sequential pruning algorithm, cut off the frequency points of the interference signal components in each frequency domain under AWGN conditions;

[0009] S3. Adopt windowed FFT operation to extend the operation under AWGN conditions in S2 to the operation under narrowband interference conditions;

[0010] S4. Use IFFT operation to restore the frequency-domain signal after removing the frequency points of the interference signal components to a time-domain signal.

[0011] Further, in S2, based on the sequential zeroing algorithm, arrange the amplitude values in order and remove the frequency points with the maximum amplitude value.

[0012] Further, map a fixed number of frequency-domain samples to zero in descending order of amplitude values, and keep other samples unchanged. The expression is:

[0013] S(k) = FFT(s(n)) (1)

[0014] a k = sort(abs(S k )), where a k = 0,

[0015] Where:

[0016] s(n) is the time-domain signal, and S(k) is the frequency-domain amplitude signal;

[0017] a is the sorted frequency-domain amplitude, k is the amplitude number, and M is the number of removed points;

[0018] The signal-to-noise ratio and signal-to-noise ratio loss after processing are respectively expressed as:

[0019]

[0020] Where:

[0021] is the energy coefficient of each frequency point;

[0022] E{·} represents the statistical average value of the signal;

[0023] Var{·} represents the variance of the signal;

[0024] Re[·] represents taking the real part of the complex signal;

[0025] represents the power of the signal;

[0026] represents the power of the noise;

[0027] It can be seen that in this sequential zeroing algorithm, L is only related to M / N, and the L value increases monotonically with the increase of M / N, and is independent of the fluctuations of other parameters, that is, for the change of received signal power or received noise power, L is a constant, that is, the greater the proportion of removal, the greater the signal-to-noise ratio loss caused by the processing. Conversely, in theoretical conditions, under AWGN small signal conditions, if the link allowable loss is known, the maximum value of M / N when using this sequential zeroing algorithm can be calculated.

[0028] Furthermore, in S2, based on a sequential pruning algorithm, the amplitude values ​​are arranged in sequence, and the frequency point with the maximum amplitude value is cut off.

[0029] Furthermore, for the frequency point to be removed, the phase information of the frequency point is saved, and the amplitude of the frequency point is corrected to a constant value r. At this time, the decision statistic after processing is:

[0030]

[0031] Where: W is the noise at the frequency point after FFT operation transformation;

[0032] make:

[0033]

[0034] Expand A in formula (5) to obtain:

[0035]

[0036] Using small signal processing method, we get:

[0037]

[0038] in:

[0039]

[0040] Combining equations (7)-(10) above, we can get the expression of the signal-to-noise ratio after pruning:

[0041]

[0042] Based on equations (11)-(12), taking the derivative of r and setting it to zero, we can solve for the optimal trimming level value r under the given frequency ratio and input noise power. oc :

[0043]

[0044] The optimal trimming level value r is obtained oc is a function of the cutoff ratio and the noise variance.

[0045] Furthermore, conditional median filtering is adopted to avoid the background noise calculation in the sequential pruning algorithm, forming a sequential median filtering algorithm.

[0046] Furthermore, based on the sequential median filtering algorithm, conditional median filtering is only performed on the frequency points that need to be pruned, that is, local smoothing is performed on the interfering frequency points to replace the calculation of the pruning level value r. The corresponding expression is:

[0047]

[0048] Performance analysis is carried out through simulation.

[0049] Furthermore, in step S3, in the frequency domain, the energy of the interference is divided into two regions: the main lobe and the side lobe. A window function is selected to suppress the side lobe energy of the narrowband interference, and under the following two approximate conditions, the excision loss when there is interference is calculated:

[0050] The first approximate condition: As long as the excision ratio is greater than the ratio of the frequency points containing the main lobe energy, all the frequency points within the main lobe will be completely excised or pruned.

[0051] The second approximate condition: The side lobe is lower than the noise power, so its order statistic characteristics are approximately the order statistic of an AWGN.

[0052] Furthermore, it is implemented by using a narrowband interference cancellation system based on spread spectrum communication. The narrowband interference cancellation system based on spread spectrum communication includes: a superposition window module, an FFT module, a background noise estimation module, a frequency domain cancellation module, an IFFT module, a de-overlap module, an amplitude detection module, and a threshold determination module;

[0053] The FFT module is connected to the frequency domain cancellation module through the superposition window module. The background noise estimation module is connected to the frequency domain cancellation module through the threshold determination module. The frequency domain cancellation module, the IFFT module, and the de-overlap module are connected in sequence. The amplitude detection module is connected to the output end of the de-overlap module to detect the amplitude value on the output frequency points and compare it with a preset threshold value.

[0054] Furthermore, the interference components below the preset threshold value are kept unchanged, and the interference components above the preset threshold value are subjected to conditional median filtering using the sequential median filtering algorithm.

[0055] The beneficial effects of the present invention are embodied in:

[0056] 1. Statistical-based interference threshold estimation technology: For a wideband direct-sequence spread-spectrum system, the bandwidth of narrowband interference occupies a relatively small proportion of the useful signal bandwidth. In the frequency bands without interference, the frequency components are mainly Gaussian white noise, and the amplitude follows a Rayleigh distribution. The mean value of the entire frequency component can be derived from the spectral components without interference according to the statistical characteristics of the Rayleigh distribution, and the background noise power can be estimated. Then, the interference decision threshold can be determined for interference signal detection.

[0057] 2. Low-complexity interference cancellation strategy: Using the median filtering theory in image processing, conditional median filtering is performed on the frequency points of the interference signal to be eliminated, and the interference signal energy is smoothed and filtered out. There is no need to calculate the elimination target value of the interference frequency points, which simplifies the cancellation algorithm and reduces the computational complexity.

[0058] 3. Low-damage signal reconstruction technology: The Blackman window is used to preprocess the time-domain signal before interference cancellation. When there is interference, the number of interference cancellation frequency points is reduced, and the loss of useful signals is reduced. Data overlapping windowing and de-overlapping processing are used to reduce the influence of the window function on the signal envelope, and the time-domain signal is reconstructed. Description of the Drawings

[0059] The drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application.

[0060] Figure 1 It is a schematic diagram of the interference threat faced by communication signals in a satellite communication system.

[0061] Figure 2 It is a schematic diagram of the overall process of the narrowband interference cancellation method according to an embodiment of the present invention.

[0062] Figure 3 It is a simulation diagram of the signal-to-noise ratio loss curve under the sequential zeroing algorithm according to an embodiment of the present invention.

[0063] Figure 4 It is a simulation diagram of the average excision ratio factor curve according to an embodiment of the present invention.

[0064] Figure 5 It is a simulation diagram of the standard deviation curve of the excision ratio factor according to an embodiment of the present invention.

[0065] Figure 6 It is a schematic diagram of the overall structure of the narrowband interference cancellation system according to an embodiment of the present invention.

[0066] Figure 7 It is a block diagram of the structure of a computer device according to an embodiment of the present invention. Detailed Embodiments

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0068] It should be noted that the meaning of "and / or" appearing throughout the text includes three parallel solutions. Taking "A and / or B" as an example, it includes solution A, or solution B, or the solution where A and B are satisfied simultaneously. In addition, "a plurality of" means more than two. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0069] In recent years, high-throughput satellites have developed rapidly and are more and more widely used. High-throughput satellites generally use transparent transponders and do not have anti-jamming capabilities. In some application fields, high reliability of satellite communication is required, and the ability to adapt to complex electromagnetic environments is required.

[0070] In terms of interference paths, in a satellite communication system, the main station's service signal reception mainly faces two types of interference, namely uplink interference and downlink interference, as Figure 1 shown. The main station uses a large-aperture parabolic antenna with a narrow antenna beam. It is difficult for downlink interference to enter the receiver through the antenna, while the receiving antenna of the satellite covers the national territory and surrounding areas. Uplink interference is easy to enter the satellite's receiver. The interference faced by the main station is mainly the uplink interference forwarded by the satellite, and the direction of arrival of the incoming wave is the same as that of the useful signal. It is impossible to use the nulling antenna technology to achieve anti-jamming. Therefore, researching signal-domain anti-jamming technology to counter the interference against high-throughput satellites and improve the availability and throughput rate of satellite communication has important application value.

[0071] The explanations of important proprietary terms involved in this application are as follows:

[0072] AWGN: Additive White Gaussian Noise, additive white Gaussian noise

[0073] FFT: Fast Fourier Transform, fast Fourier transform

[0074] IFFT: Inverse Fast Fourier Transform, inverse fast Fourier transform

[0075] DS: Direct Sequence

[0076] See Figure 2 , embodiments of the present invention provide a narrowband interference cancellation method based on spread spectrum communication, including the following steps:

[0077] S1. Perform FFT operation on the digital baseband signal to transform the digital baseband signal into a frequency-domain signal, and determine a decision threshold according to the estimated background noise power. Components greater than this threshold are judged as interference signals;

[0078] S2. Based on the sequential nulling algorithm or the sequential pruning algorithm, remove the frequency points with interference signal components in each frequency domain under AWGN conditions;

[0079] S3. Adopt windowed FFT operation to extend the operation under AWGN conditions in S2 to the operation under narrowband interference conditions;

[0080] S4. Perform IFFT operation to restore the frequency-domain signal after removing the frequency points of the interference signal components to a time-domain signal.

[0081] In this embodiment, in S2, based on the sequential nulling algorithm, arrange the amplitude values in order and remove the frequency point with the maximum amplitude value.

[0082] In DS spread spectrum communication, at the receiver input, the useful signal power is much lower than the noise power, and the usual intentional interference component is also much larger than the noise component. Therefore, at the frequency points with interference, or at the frequency points within the main lobe of the interference, its amplitude value must be a large component. Conversely, assuming there is interference, there must be interference at the frequency points with large amplitudes. Based on such a simple principle, a sorting method can be used to remove the frequency points with the maximum amplitude value to eliminate interference.

[0083] In this embodiment, map a fixed number of frequency-domain samples to zero in descending order of amplitude values, and keep other samples unchanged. The expression is:

[0084] S(k) = FFT(s(n)) (1)

[0085] a k = sort(abs(S k ))), where a k = 0,

[0086] Where:

[0087] s(n) is the time-domain signal, and S(k) is the frequency-domain amplitude signal;

[0088] a is the sorted frequency-domain amplitude, k is the amplitude number, and M is the number of points removed;

[0089] It can be seen that the two elements of the sequential nulling algorithm are: one is to use the sorting method instead of the interference detection algorithm, and the other is that the notch algorithm is the nulling algorithm;

[0090] The signal-to-noise ratio and the signal-to-noise ratio loss after processing are respectively expressed as:

[0091]

[0092]

[0093] where:

[0094] is the energy coefficient of each frequency point;

[0095] E{·} represents the statistical average value of the signal;

[0096] Var{·} represents the variance of the signal;

[0097] Re[·] represents taking the real part of the complex signal;

[0098] represents the power of the signal;

[0099] represents the power of the noise.

[0100] The relationship curve between the excision ratio factor ρ = M / N (percentage) and the signal-to-noise ratio loss L (dB value) is as Figure 3 shown.

[0101] It can be known that in this sequential nulling algorithm, L is only related to the excision factor ρ, and the value of L increases monotonically with the increase of the value of ρ, and is independent of the fluctuations of the other parameters. That is, for the changes in the received signal power or the received noise power, L is a constant. That is, the larger the excision ratio, the greater the signal-to-noise ratio loss brought by the processing. On the contrary, in the theoretical situation, under the condition of small AWGN signals, if the allowable loss of the link is known, the maximum value of the excision factor ρ when using this sequential nulling algorithm can be calculated. For example, when L is 3 dB, ρ = 38.4%, and when L is 10 dB, ρ = 59.7%.

[0102] Next, on the basis of analyzing the sequential nulling algorithm, a similar method is used to obtain the performance of the traditional threshold nulling algorithm, and the performances of the two algorithms are compared.

[0103] The threshold zeroing algorithm is different from the sequential zeroing algorithm. This algorithm needs to detect interference first, determine the threshold of the interference level, and then set all frequency points that exceed the threshold value to zero. In the case of AWGN determining the frequency domain sample value, the probability that the kth sample point does not exceed the threshold value T is:

[0104]

[0105] The probability that M points out of N points exceed T is a probability of a binomial distribution:

[0106]

[0107] Therefore, the mean of the decision statistic of the threshold zeroing algorithm can be expressed as conditional probability:

[0108]

[0109] in:

[0110]

[0111] The above equations (6) and (7) give The expression of the mean of can be obtained by the same method. Using the total probability formula, we can derive the loss of SNR and signal-to-noise ratio under the threshold zeroing algorithm as follows:

[0112]

[0113] Combining the above equations (8) and (9), it can be seen that the denominator in L2 is actually a weighted average. In order to analyze the impact of this average, the mean and standard deviation of the resection scale factor are calculated.

[0114] According to the properties of binomial distribution, the mean and variance of M / N can be obtained as follows:

[0115]

[0116] var{M / N}=p(1-p) / N (11)

[0117] The curves of the mean and standard deviation of the cut-off scale factor are as follows: Figure 4 and Figure 5 As shown, in Figure 4 It is assumed that a fixed threshold is used to determine which frequencies need to be cut off. Figure 5Curves for different values of N are given, reflecting the influence of different FFT lengths. When N = 256, the maximum value of the standard deviation is 3.125%. That is to say, when N > 256, the standard deviation of the scaling factor will be quite small. Therefore, when the noise power is known and the sequence length is large, the performance difference between the sequential cancellation algorithm and the threshold cancellation algorithm is very small.

[0118] Although under AWGN conditions, when the interference characteristics are known, the performance of the sequential cancellation algorithm and the threshold cancellation algorithm is quite close, but under channel conditions with changing interference, their performances are quite different. Compared with the sequential cancellation algorithm, the threshold cancellation algorithm has a problem of channel estimation and needs to calculate the power of the received signal. In the estimation of the noise power, there is a trade-off between the estimation accuracy and the response time, that is, the more accurate the calculation of the noise variance, the longer the time required when the channel conditions change; vice versa. When the noise power suddenly increases, before the threshold value converges to the new value, the excision ratio will be very large, and the resulting SNR loss will be very large.

[0119] In this embodiment, in S2, based on the sequential pruning algorithm, each amplitude value is arranged in order, and the frequency point with the maximum amplitude value is excised.

[0120] In this embodiment, the difference between the sequential pruning algorithm and the sequential cancellation algorithm is that for the frequency points to be excised, the phase information of the frequency points is saved, and the amplitude value of the frequency points is corrected to a constant value r. At this time, the processed decision statistic is:

[0121]

[0122] Where: W is the noise at the frequency points after the FFT operation transformation;

[0123] Let:

[0124]

[0125] Expand A in Equation (13) to get:

[0126]

[0127] Using the small-signal processing method to get:

[0128]

[0129] Where:

[0130]

[0131] Combining the above equations (15)-(18) to obtain the signal-to-noise ratio expression after pruning:

[0132]

[0133] Derive the derivative of r based on equations (11)-(12) and set it to zero, then the optimal pruning level value r can be solved under the given cut-off frequency point ratio and input noise power. oc :

[0134]

[0135] Obtain the optimal pruning level value r oc is a function related to the cut-off ratio and noise variance.

[0136] The sequential pruning algorithm is actually a class of algorithms. Different methods for selecting the value of r can result in different specific algorithms. If r oc is used as the pruning level value r, the optimal pruning OC (Optimum Clip) algorithm is obtained; if the threshold value is used as the pruning level value r, the FCT (Fraction Clip to Threshold) algorithm is obtained; if the noise mean square deviation is used as the pruning level value r, the NC (Noise Clip) algorithm is obtained. These algorithms are all well implemented in the threshold cut-off algorithm, but in the sequential cut-off algorithm, since the noise variance is not calculated, the value of r cannot actually be determined. Therefore, in the sequential pruning algorithm, other methods must be found to give the value of r.

[0137] This application proposes a sequential pruning algorithm using conditional median filtering, which avoids the problem of accurately estimating the background noise in the sequential pruning method. For simplicity, it is called the "sequential median filtering algorithm", and the specific description is as follows:

[0138] Use conditional median filtering to avoid the calculation of the background noise in the sequential pruning algorithm, forming the sequential median filtering algorithm;

[0139] The application of median filtering in image processing is quite extensive. Its idea is: for the input sequence {x}, a sliding window with a length of L is used, and the output y at each sample point i is the median value of all input sample values within the window. The digital expression is:

[0140] y i = median(x j |j = i - k,..., i + k) (22)

[0141] where L = 2k + 1 is the length of the sliding window.

[0142] Two basic characteristics of median filtering:

[0143] Any pulse with a width narrower than k in the input sequence that can well maintain a smoothly varying signal will be removed in the output regardless of its polarity, amplitude, and position.

[0144] That is to say, if a smooth signal is interfered by random pulses, median filtering is an ideal choice to remove the pulses. However, if the signal cannot be considered smooth, such filtering will introduce obvious signal distortion.

[0145] There are often oscillating components modulated on the envelope of the sampling function in the spectrum of DSSS signals, so it is not suitable to use median filtering (MF). However, when there is narrowband interference, the frequency components of narrow pulses are significantly larger than other signal components. Conditional median filtering (CMF) can be used to remove those large pulse components without affecting the useful signal frequency components at low levels.

[0146] Median filtering processes each input sample point, while conditional median filtering CMF is different. Conditional median filtering CMF will selectively filter out large pulse components, which depends on the relative width of the pulse (relative to the window size) and its relative amplitude (relative to the relevant signal value). The mathematical description of conditional median filtering CMF is as follows:

[0147]

[0148] Among them, C is the threshold parameter.

[0149] The above formula (23) is the expression in the threshold algorithm.

[0150] In the sequential excision algorithm, combined with the above analysis, it can be seen that the performance of using the pruning algorithm is better than that of the zeroing algorithm. However, since there is no threshold value as a reference, the value of r needs to be determined. By modifying the above conditional median filtering, a new notch algorithm can be obtained, that is, based on the sequential median filtering algorithm, conditional median filtering is only performed on the frequency points that need to be pruned, that is, local smoothing is performed on the interfered frequency points to replace the calculation of the pruning level value r, and the corresponding expression is:

[0151]

[0152] It is very difficult to derive a closed-form expression for the performance of the conditional median filtering algorithm, but its performance can be analyzed through simulation.

[0153] In this embodiment, in step S3, in the frequency domain, the energy of the interference is divided into two regions: the main lobe and the side lobe. A window function is selected to suppress the side lobe energy of the narrowband interference, and under the following two approximate conditions, the excision loss when there is interference is calculated:

[0154] The first approximation condition: As long as the excision ratio is greater than the ratio of the frequency points containing the main lobe energy, all the frequency points within the main lobe will be completely excised or trimmed.

[0155] The second approximation condition: The sidelobes are lower than the noise power, so their order statistics characteristics are approximately the order statistics of an AWGN.

[0156] In the presence of narrowband interference at a fixed position, assuming that the narrowband interference occupies K frequency points in the frequency domain and M frequency points are excised, where M = K + L. The performance at this time is equivalent to the performance of an algorithm that cuts L frequency points into r under the AWGN condition with a length of N - K points. The following equations (25)-(26) can be obtained:

[0157]

[0158] In this way, the sequential excision algorithm under the AWGN condition is extended to the narrowband interference condition. When different notch algorithms are adopted, equation (25) above gives the E b / N0 value of the output after processing. When r = 0, it is the result of the sequential zeroing algorithm. When the sequential trimming algorithm is adopted, equation (26) above gives the optimal trimming value r oc in the presence of interference. Strictly speaking, equation (26) above does not give the result when conditional median filtering is adopted, because the trimming value of each point in conditional median filtering is not a given value, but only a local random value. However, if the median point obtained is mainly determined by the noise component, then conditional median filtering is actually equivalent to trimming the interference points to the noise level.

[0159] See Figure 6 , an embodiment of the present invention also provides a narrowband interference cancellation system based on spread spectrum communication, applicable to the narrowband interference cancellation method based on spread spectrum communication, including: a superposition window module, an FFT module, a background noise estimation module, a frequency domain cancellation module, an IFFT module, a de-overlap module, an amplitude detection module, and a threshold determination module;

[0160] The FFT module is connected to the frequency domain cancellation module through the superposition window module, the background noise estimation module is connected to the frequency domain cancellation module through the threshold determination module, the frequency domain cancellation module, the IFFT module, and the de-overlap module are connected in sequence, and the amplitude detection module is connected to the output end of the de-overlap module for detecting the amplitude value of the output frequency points and comparing it with a preset threshold value.

[0161] In this embodiment, the interference components below the preset threshold value are kept unchanged, and the interference components above the preset threshold value are subjected to conditional median filtering using the sequential median filtering algorithm.

[0162] This application adopts an adaptive frequency-domain interference cancellation method. The basic idea is to perform an FFT operation on the digital baseband signal, transform the signal into the frequency domain, determine a decision threshold according to the estimated background noise power, determine the components greater than the threshold as interference signals, then correct the interference signal components in the transform domain to eliminate the interference signal energy, and then restore it to the time-domain signal through IFFT.

[0163] A specific frequency cancellation algorithm is determined by two factors: one is interference detection, that is, determining which frequency points have interference; the other is notch filtering, that is, on the basis of interference detection, mapping the amplitude at the interference frequency points into new values to remove the energy of the interference. Different interference detection algorithms and notch filtering strategies determine different frequency cancellation algorithms.

[0164] Performing interference detection in real time and accurately is a relatively difficult task. This application adopts a method of detecting interference based on background noise power estimation. Generally speaking, for a broadband direct-sequence spread-spectrum (DS) system, the bandwidth of narrowband interference accounts for a relatively small proportion of the useful signal bandwidth. Therefore, in the frequency band without interference, the frequency components are mainly Gaussian white noise, and the amplitude follows a Rayleigh distribution. The mean value of the entire frequency component can be deduced from the spectral components without interference according to the statistical characteristics of the Rayleigh distribution, and the magnitude of the background noise power can be estimated, so as to determine a decision threshold for interference signal detection.

[0165] Different notch filtering strategies cause different damages to the received signal. Ideally, since the DS signal has a broadband flat spectral characteristic, while the interference has a narrowband sharp spectrum, if the power spectral density of the input signal is known, there is an optimal notch filtering algorithm, that is, the frequency response of the notch filter is the reciprocal of the power spectrum, and at this time the damage to the useful signal is the smallest. However, the actual problem is how to accurately estimate the power spectral distribution. Usually, a single DFT observation is used as the estimate of the power spectrum to reduce the response time after the interference changes. However, under the AWGN condition, this algorithm is not stable. A commonly used simplified algorithm is the threshold zeroing method (TZ). Its idea is to detect interference in advance, determine the threshold value of the interference, compare the amplitude value of each frequency point with the threshold value, and if the amplitude value of this frequency point is greater than the threshold value, set it to zero, otherwise keep it unchanged. The advantage of this algorithm is simplicity and fast response to interference, and the disadvantage is that it causes great damage to the useful signal. This application adopts conditional median filtering to eliminate interference signals. First, a threshold is determined according to the interference detection method described above. The components below the threshold are kept unchanged, and the interference components above the threshold are filtered by median filtering. Compared with the threshold zeroing method, this notch filtering method causes less damage to the useful signal and is also convenient for engineering implementation.

[0166] Before the FFT operation of this application, windowing processing is performed using an appropriate window function, which can solve the problem of spectral leakage. When performing the DFT operation, the first side lobe of the signal is only 13 dB lower than the main lobe, that is, the side lobe suppression ratio is only -13 dB. For interference that is dozens of dB larger than the useful signal, its side lobe is also much larger than the signal. From the frequency domain perspective, the entire signal frequency domain is contaminated by interference. If a third-order Blackman window is added, its side lobe suppression ratio is -57 dB, and the side lobe can basically be ignored. If a fourth-order Blackman window is used, the side lobe suppression ratio can reach above -100 dB, but the price paid is a reduction in frequency resolution.

[0167] An embodiment of the present invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor is caused to execute the steps of the narrowband interference cancellation method based on spread spectrum communication as described above.

[0168] See Figure 7 , an embodiment of the present invention also provides a computer device including a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor is caused to execute the steps of the narrowband interference cancellation method based on spread spectrum communication as described above.

[0169] An embodiment of the present invention also provides a computer program product containing instructions. When it runs on a computer, the computer is caused to execute the steps of the narrowband interference cancellation method based on spread spectrum communication as described above.

[0170] It can be understood that the system, device, and storage medium provided by the embodiments of the present invention correspond to the method provided by the embodiments of the present invention. The explanations, examples, and beneficial effects of the relevant content can refer to the corresponding parts in the narrowband interference cancellation method based on spread spectrum communication as described above.

[0171] It should be noted that those of ordinary skill in the art can understand that all or part of the steps implemented in the embodiments of the present invention can be realized in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using hardware, it can be realized in whole or in part in the form of purchasing standard parts or modified parts. When implemented using software, it can be realized in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

[0172] In summary, the present invention discloses a narrowband interference cancellation method based on spread spectrum communication for a high-throughput satellite communication system. The interference cancellation method based on the transform domain is adopted. The filtering process that is very complex in the time domain can be realized by simple multiplication in the frequency domain, and the transfer function of an ideal filter that cannot be realized in the time domain, such as a rectangular filter, can also be conveniently realized in the frequency domain. In the field of spread spectrum communication, the transform domain processing technology can effectively suppress interference and improve system performance, and has the advantage that the convergence rate is independent of the number of interferences compared with time domain filters, so it is suitable for the case of multiple interferences.

[0173] It should be understood that the examples and embodiments described herein are for illustrative purposes only and are not intended to limit the present invention. Those skilled in the art can make various modifications or changes according to it. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A narrowband interference elimination method based on spread spectrum communication, characterized in that: The following steps are involved: S1. Perform FFT operation on the digital baseband signal to transform the digital baseband signal into a frequency domain signal, determine a decision threshold based on the estimated background noise power, and judge the component greater than the threshold as an interference signal; S2, based on the sequential zeroing algorithm or the sequential pruning algorithm, the frequency points with interference signal components in each frequency domain under the AWGN condition are cut off; S3, using windowed FFT operation to extend the application under AWGN conditions in S2 to the application under narrowband interference conditions; S4. Using IFFT operation, the frequency domain signal after removing the interference signal component frequency point is restored to the time domain signal.

2. The narrowband interference elimination method based on spread spectrum communication according to claim 1, characterized in that: In S2, based on the sequential zeroing algorithm, the amplitude values ​​are arranged in sequence, and the frequency point with the maximum amplitude value is cut off.

3. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 2, characterized in that: Map a fixed number of frequency domain samples to zero in descending order, leaving the other samples unchanged. The expression is: S(k)=FFT(s(n)) (1) a k =sort(abs(S k )), where a k =0, in: s(n) is the time domain signal, S(k) is the frequency domain amplitude signal; a is the frequency domain amplitude after sorting, k is the amplitude number, and M is the number of points removed; The signal-to-noise ratio and signal-to-noise ratio loss after processing are expressed as: in: ， is the energy coefficient of each frequency point; E{·} represents the statistical mean of the signal; Var{·} represents the variance of the signal; Re[·] means taking the real part of the complex signal; Indicates the power of the signal; represents the power of the noise; It can be seen that in this sequential zeroing algorithm, L is only related to M / N, and the L value increases monotonically with the increase of M / N, and is independent of the fluctuations of other parameters, that is, for the change of received signal power or received noise power, L is a constant, that is, the greater the proportion of removal, the greater the signal-to-noise ratio loss caused by the processing. Conversely, in theoretical conditions, under AWGN small signal conditions, if the link allowable loss is known, the maximum value of M / N when using this sequential zeroing algorithm can be calculated.

4. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 1, characterized in that: In S2, based on a sequential pruning algorithm, the amplitude values ​​are arranged in sequence, and the frequency point with the maximum amplitude value is cut off.

5. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 4, characterized in that: For the frequency point to be removed, save the phase information of the frequency point, and correct the amplitude of the frequency point to a constant value r. At this time, the decision statistic after processing is: Where: W is the noise at the frequency point after FFT operation transformation; make: Expand A in formula (5) to obtain: Using small signal processing method, we get: in: Combining equations (7)-(10) above, we can get the expression of the signal-to-noise ratio after pruning: Based on equations (11)-(12), taking the derivative of r and setting it to zero, we can solve for the optimal trimming level value r under the given frequency ratio and input noise power. oc : The optimal trimming level value r is obtained oc is a function of the cutoff ratio and the noise variance.

6. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 5, characterized in that: Conditional median filtering is used to avoid background noise calculation in the sequential pruning algorithm, thereby forming a sequential median filtering algorithm.

7. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 6, characterized in that: Based on the sequential median filtering algorithm, conditional median filtering is performed only on the frequency points that need to be trimmed, that is, local smoothing is performed on the frequency points with interference to replace the calculation of the trimming level value r. The corresponding expression is: Performance analysis via simulation.

8. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 1, characterized in that: In S3, in the frequency domain, the energy of the interference is divided into two regions: the main lobe and the side lobe, a window function is selected to suppress the side lobe energy of the narrowband interference, and the removal loss in the presence of interference is calculated under the following two approximate conditions: First approximate condition: As long as the cut-off ratio is greater than the frequency ratio containing the main lobe energy, all the frequencies in the main lobe will be cut off or trimmed; The second approximation condition is that the sidelobe power is lower than the noise power, so its order statistics are similar to those of an AWGN.

9. The narrowband interference elimination method based on spread spectrum communication as claimed in claim 1, characterized in that: The invention adopts a narrowband interference elimination system based on spread spectrum communication, and the narrowband interference elimination system based on spread spectrum communication comprises: an overlapping windowing module, an FFT module, a background noise estimation module, a frequency domain elimination module, an IFFT module, a de-overlapping module, an amplitude detection module and a threshold determination module; The FFT module is connected to the frequency domain elimination module via the overlapping windowing module, the background noise estimation module is connected to the frequency domain elimination module via the threshold determination module, the frequency domain elimination module, the IFFT module and the de-overlapping module are connected in sequence, and the amplitude detection module is connected to the output end of the de-overlapping module for detecting the amplitude value at the output frequency point and comparing it with a preset threshold value.

10. The narrowband interference elimination method based on spread spectrum communication according to claim 9, characterized in that: Interference components below a preset threshold value remain unchanged, and interference components above a preset threshold value are subjected to conditional median filtering using a sequential median filtering algorithm.

Images