Verilog-based adaptive time-frequency aliasing interference identification method and system

Abstract

The application discloses a kind of based on Verilog's adaptive time-frequency aliasing interference identification method and system, method includes: collection time-frequency aliasing signal;According to the priori carrier frequency of target signal and sampling frequency, actual blur frequency is dynamically calculated using the proposed HOBs-DEC algorithm;Obtain the power spectrum of signal, based on actual blur frequency and preset threshold to identify the number of effective spectral lines, signal is classified for the first time: if the number of effective spectral lines is not less than the first threshold, then determine that signal includes AM or single carrier signal;Otherwise, determine that signal includes comb spectrum, MSK or Gaussian white noise signal;Extract the square spectral feature of signal, identify the signal as AM signal or single carrier signal;Extract the single spectrum and / or square spectral feature of signal, identify the signal as comb spectrum, MSK signal or Gaussian white noise signal.The method of the application can improve the processing speed of the identification process, enhance the applicability and stability in actual scene.

Description

Technical Field

[0001] This invention belongs to the field of signal recognition technology, specifically relating to an adaptive time-frequency aliasing interference recognition method and system based on Verilog. Background Technology

[0002] With the rapid advancement of modern radio communication technology, the application scenarios of fifth-generation and sixth-generation mobile communication technologies, large-scale Internet of Things, and electronic information warfare are constantly expanding, and the number of wireless communication devices is growing exponentially. The data received by the signal receiver is often a signal with high time-frequency aliasing. Traditional methods based on single signal feature extraction are difficult to apply to the identification of time-frequency aliased signals.

[0003] Currently, methods for processing time-frequency aliasing signals can be broadly categorized into multi-channel and single-channel methods based on the number of receiving channels utilized. Multi-channel signal processing methods, exemplified by array signal processing, separate time-frequency aliasing signals by exploiting spatial differences, transforming the aliasing problem into the identification of a single signal. This allows for signal detection and identification using traditional single-signal recognition algorithms. However, the performance of these algorithms heavily relies on the accuracy of estimating the number and direction of blind signals in the spatial domain. For example, the Capon algorithm typically involves channel amplitude and phase error correction, matrix inversion, and regularization for non-positive definite matrices. The extensive floating-point calculations complicate the algorithm's flow, significantly hindering its implementation and debugging. Furthermore, it is generally implemented using a DSP (Digital Signal Processor). Single-channel signal processing methods, such as cyclic spectrum estimation and higher-order cumulants, utilize data from a single channel to extract higher-order information features of the signal. They exhibit excellent performance in terms of the number of identifiable time-frequency aliased signals and the accuracy of recognition under low signal-to-noise ratio conditions. However, the numerous higher-order multiplications and circular convolutions involved pose significant challenges to engineering implementation, resulting in high algorithm latency and difficulty in meeting the real-time requirements of practical applications. Furthermore, popular intelligent algorithms in recent years, such as machine learning and deep learning, have achieved remarkable performance in recognizing some analog and digital modulated aliased signals by training models with large amounts of data. However, in addition to their inherently high algorithm complexity, these algorithms are typically developed on the PS (Processing System) side of FPGAs (Field-Programmable Gate Arrays). The numerous convolution operations are resource-intensive, and the recognition accuracy is heavily dependent on the training dataset; the algorithm's performance is also affected by the number of samples.

[0004] As mentioned above, existing methods for detecting and identifying time-frequency aliasing signals mainly include multi-channel signal processing and single-channel signal processing techniques. Multi-channel signal processing separates aliasing signals by utilizing differences in the spatial domain, but it involves numerous matrix operations, especially inversion operations, which are difficult to develop using pure Verilog and consume significant hardware resources. Single-channel processing methods, such as cyclic spectrum estimation and higher-order cumulant feature extraction, involve numerous higher-order multiplications and circular convolutions, posing significant challenges to engineering implementation and making it difficult to meet real-time requirements. None of the above solutions are conducive to efficient deployment on FPGA platforms using pure Verilog. Summary of the Invention

[0005] To address the aforementioned problems in existing technologies, this invention provides an adaptive time-frequency aliasing interference identification method and system based on Verilog. This method has low complexity, low resource consumption, and can improve the processing speed of the identification process, enhancing its applicability and stability in practical scenarios. The technical problem to be solved by this invention is achieved through the following technical solution:

[0006] One aspect of the present invention provides an adaptive time-frequency aliasing interference identification method based on Verilog, comprising:

[0007] S1: Acquire the time-frequency aliasing signal to be identified according to the preset sampling frequency. The time-frequency aliasing signal includes the target signal and the interference signal.

[0008] S2: Based on the prior carrier frequency and sampling frequency of the target signal, the actual fuzzy frequency corresponding to the prior carrier frequency is dynamically calculated using the proposed HOBs-DEC algorithm;

[0009] S3: Obtain the power spectrum of the time-frequency aliasing signal, and identify the number of effective spectral lines based on the actual ambiguity frequency and a preset first threshold. Perform the first classification of the time-frequency aliasing signal: if the number of effective spectral lines is not less than the first threshold, determine that the time-frequency aliasing signal to be identified includes an AM signal or a single-carrier signal, and proceed to step S4; otherwise, determine that the time-frequency aliasing signal to be identified includes a comb spectrum, an MSK signal, or a Gaussian white noise signal, and proceed to step S5.

[0010] S4: Extract the square spectrum features of the time-frequency aliasing signal to determine whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal;

[0011] S5: Extract the single-order spectrum and / or square spectrum features of the time-frequency aliasing signal, and determine whether the time-frequency aliasing signal to be identified is a comb spectrum, MSK signal or Gaussian white noise signal;

[0012] S6: Output the identification results of time-frequency aliasing interference and communication interference.

[0013] Another aspect of the present invention provides an adaptive time-frequency aliasing interference detection and identification system based on Verilog, including an ADC sampling module, an ambiguous frequency acquisition module, a guidance module, an AM and single-carrier signal identification module, and an MSK comb spectrum noise signal identification module, wherein,

[0014] The ADC sampling module is used to acquire the time-frequency aliasing signal received by the receiving antenna at a preset sampling frequency;

[0015] The fuzzy frequency acquisition module is used to dynamically calculate the actual fuzzy frequency corresponding to the prior carrier frequency based on the prior carrier frequency and sampling frequency of the target signal using the proposed HOBs-DEC algorithm.

[0016] The guidance module is used to obtain the power spectrum of the time-frequency aliasing signal, and to identify the number of effective spectral lines based on the actual ambiguity frequency and a preset first threshold. The module performs a first classification of the time-frequency aliasing signal: if the number of effective spectral lines is not less than the first threshold, the time-frequency aliasing signal to be identified is determined to include an AM or single-carrier signal, and the determination result is transmitted to the AM and single-carrier signal identification module; otherwise, the time-frequency aliasing signal to be identified is determined to include a comb spectrum, MSK signal, or Gaussian white noise signal, and the determination result is transmitted to the MSK comb spectrum noise signal identification module.

[0017] The AM and single-carrier signal identification module is used to extract the square spectrum features of the time-frequency aliasing signal and determine whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal.

[0018] The MSK comb-spectrum noise signal identification module is used to extract the single-frequency spectrum and / or square spectrum features of the time-frequency aliasing signal and determine whether the time-frequency aliasing signal to be identified is a comb spectrum, an MSK signal, or a Gaussian white noise signal.

[0019] The timing control and decision output module is used to generate an enable signal under the control of the timing signal to control the execution of the ADC sampling module, the fuzzy frequency acquisition module, the guidance module, the AM and single-carrier signal identification module and the MSK comb spectrum noise signal identification module, and output the identification result of time-frequency aliasing interference communication interference.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0021] This invention provides an adaptive time-frequency aliasing communication interference identification method based on Verilog. This method utilizes Verilog to detect and identify time-frequency aliasing communication interference while meeting the requirements of low algorithm complexity and low resource consumption. First, the proposed HOBs-DEC algorithm is used to calculate the actual ambiguity frequency corresponding to the prior carrier frequency without multiplication or division operations. Then, the difference in frequency resolution required for different signal identifications is used to quickly complete the first classification of the signal. Subsequently, the identification algorithms for AM and MSK signals are improved, enabling the identification of AM signals and single-carrier signals, and MSK signals and Gaussian white noise signals, respectively. The feature extraction thresholds set during the identification process can all be obtained in Verilog using simple shifts and additions. This provides a simple, efficient, and easily implemented scheme for developing algorithms for detecting and identifying interference in related time-frequency aliasing signals using Verilog.

[0022] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an adaptive time-frequency aliasing communication interference identification method based on Verilog provided in an embodiment of the present invention.

[0024] Figure 2 This is a detailed algorithm flowchart of signal classification and identification in the Verilog-based adaptive time-frequency aliasing communication interference identification method provided in the embodiments of the present invention;

[0025] Figure 3 This is a hardware implementation logic architecture diagram of an adaptive time-frequency aliasing communication interference identification system based on Verilog provided in an embodiment of the present invention;

[0026] Figure 4 This is a diagram showing the recognition accuracy of different types of interference signals in the interference-to-signal ratio range of 15-35dB when the target signal is an MSK signal, as provided in the embodiments of the present invention.

[0027] Figure 5 This is a diagram showing the recognition accuracy of different types of interference signals when the target signal is an AM signal, provided by an embodiment of the present invention, in the range of 15-35dB between interference and signal-to-interference ratio. Detailed Implementation

[0028] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following describes in detail the Verilog-based adaptive time-frequency aliasing interference identification method proposed according to the present invention, in conjunction with the accompanying drawings and specific embodiments.

[0029] The foregoing and other technical contents, features, and effects of the present invention will be clearly presented in the following detailed description of specific embodiments in conjunction with the accompanying drawings. Through the description of the specific embodiments, a more in-depth and concrete understanding can be gained of the technical means and effects adopted by the present invention to achieve its intended purpose. However, the accompanying drawings are for reference and illustration only and are not intended to limit the technical solutions of the present invention.

[0030] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that an article or apparatus comprising a list of elements includes not only those elements but also other elements not expressly listed. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or apparatus that includes said element.

[0031] This invention provides an adaptive time-frequency aliasing communication interference identification method based on Verilog, such as... Figure 1 and Figure 2 As shown, the method includes the following steps:

[0032] S1: Collect the time-frequency aliasing signal to be identified according to the preset sampling frequency. The time-frequency aliasing signal includes the target signal and the interference signal.

[0033] Specifically, the signal generator transmits a signal according to preset transmission parameters and collects the time-frequency aliasing signal received by the receiving antenna according to a preset sampling frequency. The time-frequency aliasing signal processed in this invention includes a target signal and an interference signal. The target signal is an AM (amplitude modulation) or MSK (minimum shift keying) communication signal, and the interference signal is a comb spectrum, Gaussian white noise, a single-carrier signal, or a signal with the same modulation type as the transmitted signal. The communication frequency band (i.e., the signal frequency) is 300-600MHz, the maximum sampling frequency of the ADC (analog-to-digital converter) is 80MHz (less than the actual signal frequency, indicating undersampling), the modulation frequency of the AM signal is less than 50kHz, and the FPGA development board used has a chip model of xc7vx690tffg1761-2. The signal-to-interference ratio is above 15dB, and the carrier frequency and modulation pattern of the target signal are known.

[0034] In signal processing, according to the Nyquist-Shannon sampling theorem, to avoid frequency aliasing, the sampling frequency should be at least twice the maximum frequency of the signal. When the ADC's sampling frequency... When the frequency is lower than the actual frequency of the signal, it is undersampled. Under undersampled conditions, the actual high-frequency signal is "folded" into the low-frequency region after sampling, resulting in frequency ambiguity.

[0035] As shown above, in this embodiment, since the sampling frequency of the ADC is lower than the signal frequency, the frequency of the acquired signal is a "fuzzy" representation of the actual frequency, so dynamic correction is required.

[0036] S2: Based on the prior carrier frequency and sampling frequency of the target signal, the actual ambiguity frequency corresponding to the prior carrier frequency is dynamically calculated using the proposed HOBs-DEC algorithm.

[0037] This step utilizes the HOBs-DEC (Higher-Order Bits Difference Extraction and Comparison) algorithm to dynamically calculate the actual ambiguity frequency corresponding to the prior carrier frequency, thus resolving the carrier frequency ambiguity problem caused by undersampling. As shown above, when the ADC sampling frequency... Below the prior carrier frequency At this time, the obtained frequency is the frequency after blurring the actual frequency with respect to the sampling frequency. Therefore, the prior carrier frequency cannot be used directly, and the actual communication frequency is changing. Therefore, it is necessary to dynamically calculate the actual blurred frequency corresponding to the prior carrier frequency. In FPGA, the magnitude of the frequency is represented by a fixed-point number. S2 of this embodiment includes the following steps:

[0038] S2.1: Assume prior carrier frequency The most significant bit of the corresponding binary number is M sampling frequency The most significant bit of the corresponding binary number is N ,set up The initial value is the prior carrier frequency. , cut The N +1 position to the M Position as hobs value, and will hobs Values ​​are divided into M - N There are intervals, where the th interval is... i The intervals are , No. M - N The intervals are .

[0039] S2.2: Judgment Is the value less than the sampling frequency? If yes, then proceed to S2.4; otherwise, proceed to S2.3.

[0040] S2.3: Judgmenthobs The interval number where the value is located Simultaneously update using the first formula The value is returned to S2.2, and the expression of the first formula is:

[0041] (1)

[0042] S2.4: Judgment Is the value greater than the sampling frequency? If it is half, then update using the second formula. If the value is not specified, proceed directly to S2.5. The expression for the second formula is:

[0043] (2)

[0044] S2.5: Output the current... The value is the prior carrier frequency. The corresponding actual ambiguity frequency.

[0045] Please refer to Tables 1 and 2, which respectively represent the FPGA resource consumption for calculating the fuzzy frequency value of the prior carrier frequency relative to the sampling frequency using the traditional method and the HOBs-DEC algorithm proposed in this invention.

[0046] Table 1. FPGA resource consumption for calculating fuzzy frequency values ​​using traditional methods

[0047]

[0048] Table 2. FPGA resource consumption for calculating fuzzy frequency values ​​using this invention

[0049]

[0050] The traditional method obtains the integer part of the remainder by calling the divider IP. Then, the actual fuzzy frequency is obtained using formula (3):

[0051] (3)

[0052] in, Indicates the prior carrier frequency, This represents the actual fuzzy frequency obtained using traditional methods.

[0053] Specifically, a sampling frequency of 96MHz was used, and the sampled data was decimated by a factor of 16. The actual ambiguity frequency corresponding to the prior carrier frequency was calculated using both the traditional method and the method proposed in this invention, and the resource consumption of both methods after compilation on the Vivado platform was statistically analyzed. As shown in Tables 1 and 2, it can be seen that the method of this invention greatly optimizes the use of LUT, LUTRAM, FF, and DSP resources.

[0054] S3: Obtain the power spectrum of the time-frequency aliasing signal, identify the number of effective spectral lines based on the actual ambiguity frequency and the preset first threshold, and perform the first classification of the time-frequency aliasing signal: if the number of effective spectral lines is not less than the first threshold, it is determined that the time-frequency aliasing signal to be identified includes AM or single-carrier signal, and proceed to step S4; otherwise, it is determined that the time-frequency aliasing signal to be identified includes comb spectrum, MSK or Gaussian white noise signal, and proceed to step S5.

[0055] This step is implemented through a guidance module, which performs a Fast Fourier Transform (FFT) on the time-frequency aliasing signal to obtain the power spectrum. Based on the actual ambiguity frequency and a preset first threshold, the number of effective spectral lines is identified, and the time-frequency aliasing signal is classified for the first time.

[0056] Specifically, when the time-frequency aliasing signal to be identified contains an AM signal modulated by a smaller modulation frequency, accurately extracting the frequency features of the modulating signal is necessary to effectively distinguish the AM signal from other signals. For single-carrier signals, comb spectra, Gaussian white noise signals, and MSK signals, signal identification can be completed at a frequency resolution of 10-100kHz. However, the AM signal in this invention has an amplitude modulation frequency of no more than 50kHz, therefore, feature extraction and identification need to be completed at a frequency resolution of 1kHz. This situation leads to a large difference in the required frequency resolution for different signals during detection and identification, corresponding to different numbers of data points to be processed. Traditional algorithms do not distinguish between them and directly identify them according to the same frequency resolution, which will seriously reduce the efficiency of the algorithm. Therefore, this invention proposes to use a guidance module to perform the first classification of the time-frequency aliasing signal.

[0057] For the detection and recognition scenarios mentioned above, step S3 of this embodiment includes the following steps:

[0058] S3.1: After filtering the time-frequency aliasing signal, send it to the boot module, call Vivado's FFT IP core to perform a 1024-point FFT, and generate the power spectrum of the time-frequency aliasing signal.

[0059] Specifically, the time-frequency aliased signal sampled by the ADC and filtered by the filter is sent to the boot module. In the boot module, Vivado's FFT IP core is called to perform a 1024-point FFT to generate the power spectrum of the time-frequency aliased signal, so as to highlight the energy distribution characteristics of the signal.

[0060] S3.2: Set the register variable to an initial value of 0 max 1. During the power spectrum output process, a pipelined comparison method is used to compare the spectral power values ​​of the power spectrum at each frequency point and iteratively update the output. max 1 represents the current maximum power value. For frequency points located within a preset interval near the prior frequency, register variables are set to an initial value of 0. target 1. Update through comparison and iteration. target 1 represents the maximum power value within the current preset range.

[0061] Specifically, when the output value of the power spectrum is valid, a register variable is set to an initial value of 0. max 1. During the power spectrum output process, a pipelined comparison method is used to compare the spectral power values ​​of the power spectrum at each frequency point and iteratively update the output. max 1 represents the current maximum power value. During the comparison process, it is necessary to avoid the spectral power value corresponding to the actual ambiguous frequency of the prior carrier frequency to prevent the target signal from interfering with the classification. Therefore, for frequency points located within a preset interval near the prior frequency, register variables are set to an initial value of 0. target 1. Update through comparison and iteration. target 1 represents the maximum power value within the current preset range. The significance of avoiding prior frequencies lies in the fact that the prior frequencies are known to be the carrier frequencies of the target signal (AM signal or MSK signal), and their spectral energy is high. If they are included in the comparison, they will mask the weak spectral lines of the interference signal, leading to classification errors.

[0062] Specifically, when the power spectrum output index is near the index corresponding to the prior frequency, this embodiment is designed to take the first two points and the last two points, for a total of five frequency points. That is, when comparing the frequency point corresponding to the prior frequency and the two frequency points before and after, a register variable with an initial value of 0 is set. target 1. Through comparison and iteration, using target 1. Save the maximum power value within the current preset interval as the maximum value of the prior frequency interval. This aims to eliminate the influence of the target signal's main spectrum and focus on interference feature extraction.

[0063] S3.3: After the power spectrum output is completed, determine... target Is the value of 1 greater than the value after shifting right by 5 bits? max If the value is 1, it means there are no significant interference spectral lines near the prior frequency, and the interference signal is determined to be non-existent, so proceed to step S6; otherwise, the interference signal is determined to be present, so proceed to step S3.4.

[0064] S3.4: Will max The value of 1 is shifted right by 2 bits to obtain the first threshold Thre1 value, and the power spectrum is calculated for all single-order power spectral values ​​exceeding [a certain threshold]. ThreThe number of spectral lines is 1. Here, a spectral line is defined as a maximum point (peak), that is, a point where the power value is greater than the two adjacent frequency points, and the power value must be greater than the peak value. Thre 1. If the first threshold is exceeded Thre If the number of spectral lines is ≥4, the time-frequency aliasing signal to be identified is determined to be an AM signal or a single-carrier signal, and the process proceeds to step S4; if the first threshold is exceeded... Thre If the number of spectral lines is less than 4, the time-frequency aliasing signal to be identified is determined to be a comb spectrum, MSK signal or Gaussian white noise signal, and the process jumps to step S5.

[0065] This step achieves rapid initial diversion based on the different characteristics of time-frequency aliasing signals. By utilizing the differences in frequency resolution requirements of different signals, interference signals can be efficiently classified through a small number of sampling points (1024-point FFT) and intelligent threshold settings, avoiding resource waste in subsequent processing and improving the overall algorithm efficiency.

[0066] S4: Extract the square spectrum features of the time-frequency aliasing signal to determine whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal.

[0067] If, according to step S3, the time-frequency aliasing signal to be identified is determined to be an AM signal or a single-carrier signal, this step further identifies the AM signal or single-carrier signal. The modulation signal of the AM signal is a sine wave signal, and the time-domain expression of the AM signal is:

[0068] (4)

[0069] in, This is the externally applied DC component. This is the modulating signal for the AM signal, and its mean is usually 0. To modulate the angular frequency of the signal, For time; For carrier wave, The carrier angular frequency, and the carrier frequency The relationship between them is .

[0070] After squaring formula (4), the time-domain squared expression of the AM signal is:

[0071] (5)

[0072] Furthermore, the time-domain expression of a single-carrier signal is:

[0073] (6)

[0074] After squaring equation (6), the time-domain squared expression for the modulated signal of a single carrier is:

[0075] (7)

[0076] When the target signal is an AM signal, the spectral energy is mainly concentrated at the carrier spectrum, and setting it to zero directly can eliminate its influence. When the target signal is an MSK signal, directly setting the spectral lines to zero based on the prior carrier frequency will still cause the MSK signal's spectral lines to mask the AM signal's sideband spectral lines when the interference-to-signal ratio in a single spectrum is not less than 15dB, affecting the identification of the AM signal. However, performing a square calculation in the time domain followed by an FFT operation can separate the modulation signal from the carrier signal. The ratio of the amplitude of the separated modulation signal to the amplitude of the carrier signal is greater than the ratio of the amplitude of the sideband signal's spectral lines to the carrier signal before the square calculation. At the same time, the spectral energy of the MSK signal will mainly concentrate at two spectral lines, greatly reducing its influence on the identification of the AM signal's spectral lines.

[0077] Based on this, step S4 of this embodiment specifically includes the following steps:

[0078] S4.1: Acquire the time-frequency aliasing signal to be identified within the observation window that meets the frequency resolution required for AM signal identification, and perform downsampling processing on the acquired time-frequency aliasing signal.

[0079] In this embodiment, the original time-frequency aliasing signal is downsampled, for example, the number of data points is reduced from 80,000 to 16,384. The downsampling parameters need to ensure that the Nyquist frequency still covers the modulation frequency range of the AM signal. Downsampling can reduce the data size, improve the processing speed, and at the same time retain key frequency components, such as modulation signals below 50 kHz.

[0080] S4.2: Perform a squaring operation on the downsampled time-frequency aliased signal, call the FFT IP core to calculate the square spectrum, set the DC component in the square spectrum to zero, and then search for the maximum value spectral line of the square spectrum to obtain the maximum power value corresponding to the square spectrum.

[0081] Specifically, the downsampled time-frequency aliased signal is squared (i.e., each sample point is multiplied by itself), and the FFTIP core is called to calculate the squared spectrum. This squared spectrum data includes the carrier, modulation signal, and sideband components. This square operation enhances the modulation characteristics of the AM signal (such as sideband spectral lines), making it easier to distinguish single-carrier signals. Subsequently, the mean (DC component) of the squared spectrum is calculated, and a subtractor is used to subtract this mean from all sample points to set the DC component to zero. In an FPGA, the mean calculation can be implemented using accumulators and shifters, avoiding floating-point operations.

[0082] Subsequently, the maximum spectral line of the square spectrum after the DC component is set to zero is searched using the same method as in step S3: a register variable with an initial value of 0 is set. max 2. The pipeline compares the power values ​​at each frequency point of the current square spectrum and iteratively updates.max 2 represents the current maximum power value. P max1 .

[0083] S4.3: Based on the maximum power value corresponding to the square spectrum P max1 The frequency of the modulation signal is searched in the low-frequency range of 0-50 kHz of the square spectrum.

[0084] Specifically, set a temporary threshold as P max1 ×0.5 (achieved by right shifting by 1 bit), when the amplitude modulation depth of the AM signal is greater than or equal to 0.15, traverse the frequency points in the low-frequency range of 0-50 kHz of the square spectrum to find whether there is a spectral line greater than the temporary threshold. If it exists, record the frequency corresponding to the current spectral line as the modulation signal frequency.

[0085] S4.4: Take the top 10 spectral lines with the highest values ​​among the top 100 frequency points of the square spectrum, traverse the spectral lines with the highest values, and take the power value of the spectral line that appears most frequently in the statistics as the noise power estimate. P noise .

[0086] Specifically, the first 10 maxima in the squared spectrum are taken as spectral lines. Then, the frequency of each of these 10 spectral lines is counted among the first 100 frequency points of the squared spectrum. When the amplitude difference between two spectral lines is less than twice, they are considered the same spectral line and counted together to avoid duplicate counting of noise spectral lines. Finally, the power value of the spectral line with the highest frequency in the statistics is taken as the noise power estimate. P noise .

[0087] S4.5: Set the sideband detection threshold and verify the existence of sideband spectral lines based on the sideband detection threshold to determine whether the time-frequency aliasing signal is an AM signal or a single-carrier signal.

[0088] Specifically, the sideband detection threshold is set as follows: P noise ×64, in FPGA, is achieved by left shifting by 6 bits, without the need for a multiplier. The frequency point corresponding to the modulation signal frequency obtained in step S4.3 and its two adjacent frequency points are searched. It is checked whether the power values ​​of the current three frequency points exceed the sideband detection threshold. If there are frequency points exceeding the sideband detection threshold, then a sideband spectral line exists, and the time-frequency aliasing signal is determined to be an AM signal; otherwise, it is a single-carrier signal.

[0089] In the FPGA implementation, after obtaining the square spectrum of the signal using conventional methods, the spectral line search is performed according to the above steps. In addition, the weights in the algorithm can be obtained by shifting the amplitude left by 1 bit and left by 6 bits, which can completely avoid floating-point operations. The algorithm is efficient and has low complexity.

[0090] After completing the identification of AM signal and single-carrier signal in step S4, proceed to step S6.

[0091] S5: Extract the single-frequency spectrum and / or square spectrum features of the time-frequency aliasing signal to determine whether the time-frequency aliasing signal to be identified is a comb spectrum, MSK signal or Gaussian white noise signal.

[0092] The purpose of this step is to distinguish between comb-shaped spectra, MSK signals, and Gaussian white noise signals by extracting the single-order spectrum (original spectrum) and square spectrum features of the time-frequency aliasing signal. On the FPGA, only basic logic units (such as adders and comparators) are required, avoiding floating-point operations in traditional methods and significantly improving efficiency.

[0093] MSK signal is a special type of 2FSK signal, and its time-domain expression is as follows:

[0094] (8)

[0095] in, , , The carrier frequency of the MSK signal. For the first k The initial phase of each symbol, For output symbols, Let be the symbol rate. It can be seen that after squaring the MSK signal, two discrete frequency components will appear, and the distance between each frequency component and twice the carrier frequency is half the symbol rate.

[0096] Step S5 in this embodiment specifically includes:

[0097] S5.1: Perform time-domain squaring on the acquired time-frequency aliased signal, call the FFT IP core to calculate the squared spectrum, and set the DC component of the squared spectrum to zero to avoid interference with spectral line analysis, thus obtaining the preprocessed squared spectrum. In the FPGA, setting the DC component to zero can be achieved using a simple subtractor.

[0098] S5.2: Obtain the maximum power value in the preprocessed square spectrum. P max2 And set a second threshold. Thre 2= P max2 ×0.5, the square spectrum after statistical preprocessing exceeds the second threshold. Thre 2. Number of spectral lines. During statistics, only the maximum value spectral line (the peak value whose power is greater than the peak values ​​of its left and right adjacent points) is considered. If it exceeds the second threshold... Thre If the number of spectral lines is equal to 2, it is initially determined to be an MSK signal. If it exceeds the second threshold... ThreIf the number of spectral lines in 2 is much greater than 5, for example, greater than 10, then it is initially determined to be Gaussian white noise; if the second threshold... Thre If the number of spectral lines in signal 2 does not exceed 5, a discreteness check is performed to further determine whether it is an MSK signal or a Gaussian white noise signal.

[0099] In this embodiment, the purpose of the discreteness check is to distinguish between MSK signals and Gaussian white noise signals based on the energy concentration characteristic when the number of spectral lines is small (≤5). Specifically, 16 spectral lines around the maximum spectral line (index range is ±8 of the index corresponding to the maximum spectral line) are selected, and the average amplitude of these spectral lines (excluding the maximum spectral line itself) is calculated. In the FPGA, the average is calculated by summing the values ​​using an adder and then right-shifting by 4 bits (dividing by 16). Subsequently, it is compared whether the amplitude of the maximum spectral line is greater than 3 times the average amplitude. If it is, it is determined to be an MSK signal (energy concentration); otherwise, it is a Gaussian white noise signal (energy dispersion).

[0100] In the FPGA implementation, the amplitude of the maximum spectral line and its corresponding position can be obtained by iterative comparison using registers when calculating the square spectrum. An adder is set up according to the position of the maximum spectral line to calculate the sum of the spectral lines around the maximum spectral line. Apart from the necessary FFT calculation, the entire feature extraction process only requires one adder.

[0101] S5.3: Perform FFT on the time-frequency aliased signal to obtain the single spectrum and find the maximum power value of the single spectrum. P max_single Set the third threshold to Thre3= P max_single ×0.5.

[0102] S5.4: Count the number of spectral lines in a single spectrum that exceeds the third threshold and calculate the distance between adjacent single spectra.

[0103] Specifically, count the number of spectral lines in a single spectrum that exceeds the third threshold. Q Record the frequency indices of these spectral lines. Calculate the distances (i.e., index differences) between adjacent spectral lines that exceed the third threshold, and output the distance array.

[0104] S5.5: Determine whether the time-frequency aliasing signal is a comb spectrum based on the number of occurrences of the distance between adjacent single spectra.

[0105] Specifically, five registers are set up to store the distance values ​​between the first five adjacent single spectra (for parallel processing). The stored distance values ​​are compared with all distance values ​​in the distance array in a parallel pipeline, the frequency of each distance value is counted, and the frequency of the distance value with the most occurrences is taken. C max ;

[0106] Calculate the number of spectral lines in a single spectrum that exceeds the third threshold. Q Frequency of the most frequently occurring distance value C max The difference between If Δ≤3, it is determined to be a comb pattern; if Δ>3, it is excluded from the comb pattern.

[0107] After completing the identification of MSK signal, comb spectrum and Gaussian white noise signal in step S5, proceed to step S6.

[0108] S6: Output the detection and identification results of time-frequency aliasing interference and communication interference.

[0109] Specifically, the identification results include whether the time-frequency aliasing signal includes an interference signal and the type of interference signal. If there is no interference signal, the output number is 0; if there is an interference signal, the output number corresponding to the interference signal is 0.

[0110] Please see Figure 4 and Figure 5 , Figure 4 This is a diagram showing the recognition accuracy of different types of interference signals in the interference-to-signal ratio range of 15-35dB when the target signal is an MSK signal, as provided in the embodiments of the present invention. Figure 5 This is a graph showing the recognition accuracy of different types of interference signals when the target signal is an AM signal, provided in an embodiment of the present invention, within an interference-to-signal ratio (ISR) of 15-35 dB. Figure 4 and Figure 5 As can be seen, under the set interference-to-signal ratio, the identification method of the present invention achieves an effective identification rate of over 90% for all five types of signals under different target signals.

[0111] This embodiment provides an adaptive time-frequency aliasing communication interference identification method based on Verilog. It utilizes Verilog to detect and identify time-frequency aliasing communication interference while meeting the requirements of low algorithm complexity and low resource consumption. The method first uses the proposed HOBs-DEC algorithm to calculate the actual ambiguity frequency corresponding to the prior carrier frequency without multiplication or division operations. Then, it quickly completes the first classification of signals by utilizing the difference in frequency resolution required for different signal identifications. Subsequently, it improves the identification algorithms for AM and MSK signals, respectively identifying AM signals and single-carrier signals, and MSK signals and Gaussian white noise signals. The feature extraction thresholds set during the identification process can all be obtained in Verilog using simple shifts and additions. This provides a simple, efficient, and easily implemented scheme for developing algorithms for detecting and identifying interference in related time-frequency aliasing signals using Verilog.

[0112] Example 2

[0113] Based on Embodiment 1, this embodiment provides an adaptive time-frequency aliasing interference identification system based on Verilog, such as... Figure 3 As shown, the system includes an ADC sampling module, an ambiguity frequency acquisition module, a guidance module, an AM and single-carrier signal identification module, an MSK comb-spectrum noise signal identification module, and a timing control and decision output module. The ADC sampling module is used to acquire the time-frequency aliasing signal received by the receiving antenna at a preset sampling frequency. The ambiguity frequency acquisition module is used to dynamically calculate the actual ambiguity frequency corresponding to the prior carrier frequency using the proposed HOBs-DEC algorithm based on the prior carrier frequency and sampling frequency of the target signal. The guidance module is used to obtain the power spectrum of the time-frequency aliasing signal, identify the number of effective spectral lines based on the actual ambiguity frequency and a preset first threshold, and perform a first classification of the time-frequency aliasing signal: if the number of effective spectral lines is not less than the first threshold, it determines that the time-frequency aliasing signal to be identified includes an AM signal or a single-carrier signal and transmits the determination result to the AM and single-carrier signal identification module. The module determines whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal; otherwise, it determines that the time-frequency aliasing signal to be identified includes a comb spectrum, MSK signal, or Gaussian white noise signal, and transmits the determination result to the MSK comb spectrum noise signal identification module; the AM and single-carrier signal identification module is used to extract the square spectrum features of the time-frequency aliasing signal and determine whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal; the MSK comb spectrum noise signal identification module is used to extract the single spectrum and / or square spectrum features of the time-frequency aliasing signal and determine whether the time-frequency aliasing signal to be identified is a comb spectrum, MSK signal, or Gaussian white noise signal; the timing control and decision output module is used to generate an enable signal under the control of the timing signal to control the execution of the ADC sampling module, the fuzzy frequency acquisition module, the guidance module, the AM and single-carrier signal identification module, and the MSK comb spectrum noise signal identification module, and outputs the identification result of the time-frequency aliasing interference communication interference.

[0114] This invention addresses the scenario where the target signal is an AM or MSK signal under undersampling conditions, and the mixed interference signals include comb spectrum, Gaussian white noise, single-carrier signal, and similar modulation signals. Utilizing the carrier frequency and modulation pattern information of the target signal, as well as the different spectral line characteristics of different signals in the single and square spectra, a threshold that can be constructed using shift and addition operations is designed. Simultaneously, based on the different frequency resolution requirements of various signals to be identified, a guidance module is designed using a few samples to dynamically adjust the signal observation window according to different signals. Furthermore, leveraging the characteristics of Verilog, the HOBs-DEC algorithm is designed to quickly solve for the actual frequency after prior carrier frequency ambiguity, effectively reducing algorithm complexity, minimizing hardware resource consumption, optimizing the threshold calculation scheme, and improving the method's implementation efficiency.

[0115] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. An adaptive time-frequency aliasing interference identification method based on Verilog, characterized in that, include: S1: Acquire the time-frequency aliasing signal to be identified according to the preset sampling frequency. The time-frequency aliasing signal includes the target signal and the interference signal. S2: Based on the prior carrier frequency and sampling frequency of the target signal, the actual fuzzy frequency corresponding to the prior carrier frequency is dynamically calculated using the proposed HOBs-DEC algorithm; S3: Obtain the power spectrum of the time-frequency aliasing signal, and identify the number of effective spectral lines based on the actual ambiguity frequency and a preset first threshold. Perform the first classification of the time-frequency aliasing signal: if the number of effective spectral lines is not less than the first threshold, determine that the time-frequency aliasing signal to be identified includes an AM signal or a single-carrier signal, and proceed to step S4; otherwise, determine that the time-frequency aliasing signal to be identified includes a comb spectrum, an MSK signal, or a Gaussian white noise signal, and proceed to step S5. S4: Extract the square spectrum features of the time-frequency aliasing signal to determine whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal; S5: Extract the single-order spectrum and / or square spectrum features of the time-frequency aliasing signal, and determine whether the time-frequency aliasing signal to be identified is a comb spectrum, MSK signal or Gaussian white noise signal; S6: Output the identification results of time-frequency aliasing interference and communication interference; S2 includes: S2.1: Assume prior carrier frequency The most significant bit of the corresponding binary number is M sampling frequency The most significant bit of the corresponding binary number is N ,set up The initial value is the prior carrier frequency. , cut The N +1 position to the M Position as hobs The value, and the hobs Values ​​are divided into M - N A range; S2.2: Judgment Is the value less than the sampling frequency? If yes, then proceed to S2.4; otherwise, proceed to S2.

3. S2.3: Judgment hobs The interval number where the value is located Simultaneously update using the first formula The value of is returned to S2.2, and the expression of the first formula is: ， S2.4: Judgment Is the value greater than the sampling frequency? If it is half, then update using the second formula. If the value is not specified, proceed directly to S2.

5. The expression for the second formula is: ， S2.5: Output the current... The value is the prior carrier frequency. The corresponding actual ambiguity frequency; S4 includes: S4.1: Acquire the time-frequency aliasing signal to be identified within the observation window that meets the frequency resolution required for AM signal identification, and perform downsampling processing on the time-frequency aliasing signal; S4.2: Perform a squaring operation on the downsampled time-frequency aliased signal, call the FFT IP core to calculate the squared spectrum, and set the DC component in the squared spectrum to zero. Then, search for the maximum spectral line of the squared spectrum after setting the DC component to zero to obtain the maximum power value corresponding to the squared spectrum. P max1 ; S4.3: Based on the maximum power value corresponding to the square spectrum. P max1 The modulation signal frequency is searched within the 0-50 kHz low-frequency range of the square spectrum; S4.4: Take the top 10 maximum spectral lines from the top 100 frequency points of the square spectrum, traverse the maximum spectral lines, and take the power value of the spectral line that appears most frequently in the statistics as the noise power estimate. P noise ; S4.5: Set the sideband detection threshold to P noise ×64, and verify the existence of sideband spectral lines according to the sideband detection threshold to determine whether the time-frequency aliasing signal is an AM signal or a single-carrier signal; S5 includes: S5.1: Perform time-domain squaring on the acquired time-frequency aliasing signal, call the FFT kernel to calculate the square spectrum, set the DC component of the square spectrum to zero, and obtain the preprocessed square spectrum; S5.2: Obtain the maximum power value in the preprocessed square spectrum. P max2 And set a second threshold. Thre 2= P max2 ×0.5, statistically analyze the preprocessed squared spectrum that exceeds the second threshold. Thre The number of spectral lines in 2; if it exceeds the second threshold Thre If the number of spectral lines is 2, it is initially determined to be an MSK signal. If it exceeds the second threshold... Thre If the number of spectral lines in 2 is greater than 10, then the time-frequency aliasing signal is preliminarily determined to be Gaussian white noise; if the second threshold Thre If the number of spectral lines in 2 is ≤ 5, then a discreteness check is performed to further determine whether the time-frequency aliasing signal is the MSK signal or the Gaussian white noise signal; S5.3: Perform FFT on the time-frequency aliased signal to obtain the single spectrum, and find the maximum power value of the single spectrum. P max_single Set the third threshold to Thre3= P max_single ×0.5; S5.4: Count the number of single spectra that exceed the third threshold and calculate the distance between adjacent single spectra; S5.5: Determine whether the time-frequency aliasing signal is a comb spectrum based on the number of occurrences of the distance between adjacent single spectra.

2. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 1, characterized in that, S3 includes: S3.1: After filtering the time-frequency aliased signal, call Vivado's FFT IP core to perform a 1024-point FFT to generate the power spectrum of the time-frequency aliased signal; S3.2: Set the register variable to an initial value of 0 max 1. During the power spectrum output process, a pipelined comparison method is used to compare the spectral power values ​​of the power spectrum at each frequency point and iteratively update the output. max 1 represents the current maximum power value. For frequency points located within a preset interval near the prior frequency, register variables are set to an initial value of 0. target 1. Update through comparison and iteration. target 1 represents the maximum power value within the current preset range; S3.3: After the power spectrum output is completed, determine... target Is the value of 1 greater than the value after shifting right by 5 bits? max If the value is 1, it is determined that the interference signal does not exist and the process jumps to step S6; otherwise, it is determined that the interference signal exists and the process jumps to step S3.

4. S3.4: The above max The value of 1 is shifted right by 2 bits to form the first threshold. Thre The value of 1 indicates that the power spectrum exceeds the first threshold in all single-order spectra. Thre If the number of spectral lines of 1 exceeds the first threshold Thre If the number of spectral lines is ≥4, then the time-frequency aliasing signal is determined to include an AM signal or a single-carrier signal, and the process proceeds to step S4; if the first threshold is exceeded... Thre If the number of spectral lines is less than 4, then the time-frequency aliasing signal is determined to include a comb spectrum, an MSK signal, or a Gaussian white noise signal, and the process proceeds to step S5.

3. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 2, characterized in that, S4.3 includes: Set temporary threshold as P max1 ×0.5, when the amplitude modulation depth of the AM signal is greater than or equal to 0.15, traverse the frequency points in the 0-50 kHz low-frequency range of the square spectrum to find whether there is a spectral line greater than the temporary threshold. If there is, record the frequency corresponding to the current spectral line as the modulation signal frequency.

4. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 3, characterized in that, S4.5 includes: Set the sideband detection threshold to P noise ×64, search for the frequency point corresponding to the modulation signal frequency obtained in S4.3 and the two adjacent frequency points on the left and right, check whether the power value of the current three frequency points exceeds the sideband detection threshold. If there is a frequency point that exceeds the sideband detection threshold, then there is a sideband spectral line, and the time-frequency aliasing signal is determined to be an AM signal; otherwise, it is a single-carrier signal.

5. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 4, characterized in that, If the second threshold Thre If the number of spectral lines in signal 2 is ≤ 5, a discreteness check is performed to further determine whether the time-frequency aliasing signal is the MSK signal or the Gaussian white noise, including: Take eight spectral lines on each side of the maximum spectral line of the square spectrum, calculate the average amplitude of all spectral lines, and compare whether the amplitude of the maximum spectral line is greater than three times the average amplitude. If so, the time-frequency aliasing signal is determined to be an MSK signal; otherwise, it is a Gaussian white noise signal.

6. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 5, characterized in that, S5.5 includes: Five registers are set up to store the distance values ​​between the first five adjacent single spectra. The stored distance values ​​are compared with all distance values ​​in the distance array in parallel pipelined manner. The frequency of each distance value is counted, and the frequency of the distance value with the most occurrences is taken. C max ; Calculate the number of spectral lines in a single spectrum that exceeds the third threshold. Q and C max The difference between them Δ= Q C max If Δ≤3, the time-frequency aliasing signal is determined to be a comb spectrum; if Δ>3, the comb spectrum is excluded.

7. An adaptive time-frequency aliasing interference identification system based on Verilog, characterized in that, The system for executing the Verilog-based adaptive time-frequency aliasing interference identification method according to any one of claims 1 to 6, the system comprising an ADC sampling module, an ambiguous frequency acquisition module, a guidance module, an AM and single-carrier signal identification module, an MSK comb spectrum noise signal identification module, and a timing control and decision output module, wherein... The ADC sampling module is used to acquire the time-frequency aliasing signal received by the receiving antenna at a preset sampling frequency; The fuzzy frequency acquisition module is used to dynamically calculate the actual fuzzy frequency corresponding to the prior carrier frequency based on the prior carrier frequency and sampling frequency of the target signal using the proposed HOBs-DEC algorithm. The guidance module is used to obtain the power spectrum of the time-frequency aliasing signal, and to identify the number of effective spectral lines based on the actual ambiguity frequency and a preset first threshold. The module performs a first classification of the time-frequency aliasing signal: if the number of effective spectral lines is not less than the first threshold, the time-frequency aliasing signal to be identified is determined to include an AM or single-carrier signal, and the determination result is transmitted to the AM and single-carrier signal identification module; otherwise, the time-frequency aliasing signal to be identified is determined to include a comb spectrum, MSK signal, or Gaussian white noise signal, and the determination result is transmitted to the MSK comb spectrum noise signal identification module. The AM and single-carrier signal identification module is used to extract the square spectrum features of the time-frequency aliasing signal and determine whether the time-frequency aliasing signal to be identified is an AM signal or a single-carrier signal. The MSK comb-spectrum noise signal identification module is used to extract the single-frequency spectrum and / or square spectrum features of the time-frequency aliasing signal and determine whether the time-frequency aliasing signal to be identified is a comb spectrum, an MSK signal, or a Gaussian white noise signal. The timing control and decision output module is used to generate an enable signal under the control of the timing signal to control the execution of the ADC sampling module, the fuzzy frequency acquisition module, the guidance module, the AM and single-carrier signal identification module and the MSK comb spectrum noise signal identification module, and output the identification result of time-frequency aliasing interference communication interference.

Images