CN-121980320-A - Verilog-based self-adaptive time-frequency aliasing interference identification method and system

CN121980320ACN 121980320 ACN121980320 ACN 121980320ACN-121980320-A

Abstract

The invention discloses a method and a system for identifying self-adaptive time-frequency aliasing interference based on Verilog, wherein the method comprises the steps of collecting time-frequency aliasing signals; according to the prior carrier frequency and sampling frequency of the target signal, the actual fuzzy frequency is dynamically calculated by utilizing the proposed HOBs-DEC algorithm, the power spectrum of the signal is obtained, the number of effective spectral lines is identified based on the actual fuzzy frequency and a preset threshold, the signal is classified for the first time, if the number of the effective spectral lines is not smaller than the first threshold, the signal is judged to comprise an AM or single carrier signal, otherwise, the signal is judged to comprise a comb spectrum, an MSK or a Gaussian white noise signal, the square spectrum characteristic of the signal is extracted, the signal is identified to be the AM signal or the single carrier signal, the single spectrum and/or the square spectrum characteristic of the signal is extracted, and the signal is identified to be the comb spectrum, the MSK signal or the Gaussian white noise signal. The method of the invention can improve the processing speed of the identification process and enhance the applicability and stability in the actual scene.

Inventors

ZENG CAO
ZHOU RENJUN
YANG ZHAO
LIU JINHUI
LI SHIDONG
QI RUIXUE
WANG ZIXIAO

Assignees

西安电子科技大学

Dates

Publication Date: 20260505
Application Date: 20260401

Claims (10)

1. A Verilog-based adaptive time-frequency aliasing interference identification method is characterized by comprising the following steps: S1, acquiring a time-frequency aliasing signal to be identified according to a preset sampling frequency, wherein the time-frequency aliasing signal comprises a target signal and an interference signal; S2, dynamically calculating the actual fuzzy frequency corresponding to the prior carrier frequency by using the proposed HOBs-DEC algorithm according to the prior carrier frequency and the sampling frequency of the target signal; S3, obtaining a power spectrum of the time-frequency aliasing signal, identifying the number of effective spectral lines based on the actual fuzzy frequency and a preset first threshold, and classifying the time-frequency aliasing signal for the first time, wherein if the number of the effective spectral lines is not smaller than the first threshold, the time-frequency aliasing signal to be identified is judged to comprise an AM signal or a single carrier signal, and then the step S4 is carried out, otherwise, the time-frequency aliasing signal to be identified is judged to comprise a comb spectrum, an MSK signal or a Gaussian white noise signal, and then the step S5 is carried out; s4, extracting square spectrum characteristics of the time-frequency aliasing signals, and judging whether the time-frequency aliasing signals to be identified are AM signals or single carrier signals; s5, extracting single spectrum and/or square spectrum characteristics of the time-frequency aliasing signals, and judging whether the time-frequency aliasing signals to be identified are comb spectrum, MSK signals or Gaussian white noise signals; and S6, outputting a recognition result of the time-frequency aliasing interference communication interference.
2. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 1, wherein the S2 comprises: S2.1 assume a priori carrier frequencies 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 Initial value is a priori carrier frequency Intercepting N+1st to mth bits as hobs values, and dividing the hobs values into M-N intervals; s2.2 judgment Whether or not the value of (2) is less than the sampling frequency If yes, jumping to S2.4, and if not, jumping to S2.3; S2.3, judging the section number where hobs value is located At the same time update with the first formula And returning to S2.2, the expression of the first formula is: , s2.4 judgment Whether or not the value of (2) is greater than the sampling frequency If yes, update with the second formula If not, directly jump to S2.5, the expression of the second formula is: , s2.5 outputting the current Is the a priori carrier frequency Corresponding to the actual blurring frequency.
3. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 2, wherein the S3 comprises: s3.1, after the time-frequency aliasing signal is filtered, invoking an FFT IP core of Vivado to carry out 1024-point FFT, and generating a power spectrum of the time-frequency aliasing signal; S3.2, setting a register variable max1 with an initial value of 0, comparing spectral line power values of the power spectrum by frequency points by adopting a pipeline comparison method in the power spectrum output process, iteratively updating max1 into a current maximum power value, setting a register variable target1 with the initial value of 0 for frequency points in a preset interval near the prior frequency, and updating target1 into the maximum power value in the current preset interval by comparing iteration; S3.3, judging whether the value of target1 is larger than the value of max1 after shifting 5 bits right after the power spectrum output is finished, if so, judging that the interference signal does not exist, jumping to the step S6, if not, judging that the interference signal exists, jumping to the step S3.4; S3.4, taking the value of max1 to the right of 2 bits as a value of a first threshold Thre1, counting the number of spectral lines exceeding the first threshold Thre1 in all single spectrums of a power spectrum, if the number of the spectral lines exceeding the first threshold Thre1 is more than or equal to 4, judging that the time-frequency aliasing signal comprises an AM signal or a single carrier signal, jumping to the step S4, and if the number of the spectral lines exceeding the first threshold Thre1 is less than 4, judging that the time-frequency aliasing signal comprises a comb spectrum, an MSK signal or a Gaussian white noise signal, jumping to the step S5.
4. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 3, wherein the S4 comprises: S4.1, collecting a time-frequency aliasing signal to be identified under an observation window meeting the frequency resolution of an AM signal, and carrying out downsampling treatment on the time-frequency aliasing signal; S4.2, square operation is carried out on the down-sampled time-frequency aliasing signals, FFT IP core is called to calculate square spectrum, direct current components in the square spectrum are set to zero, and then the maximum spectral line of the square spectrum after the direct current components are set to zero is searched to obtain a maximum power value P max1 corresponding to the square spectrum; S4.3, searching the frequency of the modulation signal in a low-frequency interval of 0-50 kHz of the square spectrum according to the maximum power value P max1 corresponding to the square spectrum; S4.4, taking the maximum spectral line appearing in the first 10 frequency points of the square spectrum, traversing the maximum spectral line, and taking the spectral line power value with the largest occurrence number in statistics as a noise power estimation value P noise ; And S4.5, setting a sideband detection threshold as P noise multiplied by 64, and verifying whether sideband spectral lines exist according to the sideband detection threshold so as to judge whether the time-frequency aliasing signal is an AM signal or a single carrier signal.
5. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 4, wherein S4.3 comprises: setting a temporary threshold as P max multiplied by 0.5, traversing frequency points in a low-frequency interval of 0-50 kHz of the square spectrum when the amplitude modulation depth of an AM signal is greater than or equal to 0.15, searching whether spectral lines greater than the temporary threshold exist, and recording the frequency corresponding to the current spectral line as the frequency of a modulation signal if the spectral lines exist.
6. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 5, wherein S4.5 comprises: Setting the sideband detection threshold as P noise multiplied by 64, searching a frequency point corresponding to the frequency of the modulated signal obtained in the step S4.3 and two adjacent frequency points on the left and right, checking whether the power values of the current three frequency points exceed the sideband detection threshold, if so, judging that the time-frequency aliasing signal is an AM signal, otherwise, judging that the time-frequency aliasing signal is a single carrier signal.
7. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 6, wherein S5 comprises: S5.1, carrying out time domain square operation on the acquired time-frequency aliasing signals, calling FFT (fast Fourier transform) kernel to calculate a square spectrum, and setting a direct current component of the square spectrum to zero to obtain a preprocessed square spectrum; S5.2, obtaining a maximum power value P max2 in the preprocessed square spectrum, setting a second threshold Thre2=P max ×0.5, counting the number of spectral lines exceeding the second threshold Thr2 in the preprocessed square spectrum, preliminarily judging as an MSK signal if the number of spectral lines exceeding the second threshold Thr2=2, preliminarily judging as Gaussian white noise if the number of spectral lines exceeding the second threshold Thr2 is greater than 10, and performing a discreteness check if the number of spectral lines exceeding the second threshold Thr2 is less than or equal to 5 to further judge as the MSK signal or the Gaussian white noise signal; S5.3, carrying out FFT on the time-frequency aliasing signal to obtain a single spectrum, finding out a maximum power value P max_single of the single spectrum, and setting a third threshold as Thre3=P max_single multiplied by 0.5; s5.4, counting the number of the single spectrums exceeding the third threshold and calculating the distance between the adjacent single spectrums; And S5.5, judging whether the time-frequency aliasing signal is a comb spectrum or not according to the occurrence times of the distance between the adjacent single spectrums.
8. The Verilog-based adaptive time-frequency aliasing interference identification method of claim 7, wherein if the number of spectral lines of the second threshold Thre2 is less than or equal to 5, performing a discretization check to further determine whether the time-frequency aliasing signal is the MSK signal or the gaussian white noise, comprises: and taking eight spectral lines on two sides of the maximum spectral line of the square spectrum, calculating the average amplitude of all the spectral lines, comparing whether the maximum spectral line amplitude is more than 3 times of the average amplitude, if so, judging that the time-frequency aliasing signal is an MSK signal, and otherwise, judging that the time-frequency aliasing signal is a Gaussian white noise signal.
9. The Verilog-based adaptive time-frequency aliasing interference identification method according to claim 8, wherein S5.5 comprises: Setting 5 registers for storing distance values between the first 5 adjacent single spectrums, comparing the stored distance values with all distance values in a distance array in a parallel running mode, counting the occurrence times of each distance value, and taking out the frequency C max of the distance value with the largest occurrence times; Calculating the difference between the numbers of spectral lines Q and C max of the single spectrum exceeding the third threshold If delta is less than or equal to 3, the time-frequency aliasing signal is judged to be comb spectrum, and if delta is more than 3, the comb spectrum is eliminated.
10. A self-adaptive time-frequency aliasing interference identification system based on Verilog is characterized by comprising an ADC sampling module, a fuzzy frequency acquisition module, a guiding module, an AM and single carrier signal identification module, an MSK comb spectrum noise signal identification module and a timing control and judgment output module, wherein, The ADC sampling module is used for collecting time-frequency aliasing signals received by the receiving antenna according to a preset sampling frequency; the fuzzy frequency acquisition module is used for dynamically calculating the actual fuzzy frequency corresponding to the prior carrier frequency by utilizing the proposed HOBs-DEC algorithm according to the prior carrier frequency and the sampling frequency of the target signal; The guiding module is used for obtaining the power spectrum of the time-frequency aliasing signal, identifying the number of effective spectral lines based on the actual fuzzy frequency and a preset first preset threshold, and classifying the time-frequency aliasing signal for the first time, wherein if the number of the effective spectral lines is not smaller than the first threshold, the time-frequency aliasing signal to be identified is judged to comprise an AM or single carrier signal and the judgment result is transmitted to the AM and single carrier signal identification module, otherwise, the time-frequency aliasing signal to be identified is judged to comprise a comb spectrum, an MSK signal or a Gaussian white noise signal, and the judgment result is transmitted to the MSK comb spectrum noise signal identification module; The AM and single carrier signal identification module is used for extracting square spectrum characteristics of the time-frequency aliasing signals and judging whether the time-frequency aliasing signals to be identified are AM signals or single carrier signals; The MSK comb spectrum noise signal identification module is used for extracting single spectrum and/or square spectrum characteristics of the time-frequency aliasing signal and judging whether the time-frequency aliasing signal to be identified is a comb spectrum, an MSK signal or a Gaussian white noise signal; The time sequence control and judgment output module is used for generating an enabling signal under the control of a time sequence signal to control the execution of the ADC sampling module, the fuzzy frequency acquisition module, the guiding module, the AM and single carrier signal identification module and the MSK comb spectrum noise signal identification module, and outputting the identification result of time-frequency aliasing interference communication interference.

Description

Verilog-based self-adaptive time-frequency aliasing interference identification method and system Technical Field The invention belongs to the technical field of signal identification, and particularly relates to a Verilog-based self-adaptive time-frequency aliasing interference identification method and system. Background Along with the rapid jump of the modern radio communication technology, the application scenes such as the fifth generation and sixth generation mobile communication technologies, the large-scale internet of things, electronic information countermeasure and the like are continuously expanded, the number of wireless communication devices presents an exponential growth situation, data received by a signal receiver are signals with high aliasing in a time-frequency domain, and the traditional method based on single signal feature extraction is difficult to apply to the identification of time-frequency aliasing signals. The current methods for processing the time-frequency aliasing signals can be divided into two categories, namely multi-channel and single-channel, according to the number of the utilized receiving channels. The multichannel signal processing method represented by the array signal processing technology utilizes the difference of time-frequency aliasing signals in a space domain to separate, and converts the time-frequency aliasing problem into the identification of a single signal, thereby completing the detection and identification of the signal by utilizing the traditional single signal identification algorithm. However, the performance of the algorithm mainly depends on whether the estimation of the number and the direction information of the blind signals in the space domain is accurate, taking a Capon algorithm as an example, the algorithm generally comprises operations such as channel amplitude and phase error correction, matrix inversion, regularization when the matrix inversion is performed, and the like, a large number of floating point number calculation results in complex algorithm flow, and great difficulty is caused to the realization and floor debugging of the algorithm, and the algorithm is generally realized by a DSP (DIGITAL SIGNAL Processor). The single-channel signal processing method represented by cyclic spectrum estimation and high-order accumulation is characterized in that the data information of a single channel is utilized to extract the high-order information characteristics of signals, the number of identifiable time-frequency aliasing signals and the accuracy of identification under low signal-to-noise ratio are excellent, but related operations such as large amount of high-order multiplication, cyclic convolution and the like bring great challenges to engineering realization, the algorithm time delay is high, and the real-time requirement of practical application is difficult to meet. In addition, intelligent algorithms such as machine learning, deep learning and the like popular in recent years have not been colloquially represented in the recognition of partially analog and digitally modulated aliased signals by training a model by using a large amount of data, but besides being extremely high in algorithm complexity, algorithms are generally developed on the FPGA (Field-Programmable GATE ARRAY, field Programmable gate array) end by using the PS (Processing System ) end, a large amount of convolution operations consume resources, the recognition accuracy is seriously dependent on a trained data set, and the performance of the algorithms is also influenced by the number of samples. As described above, the existing detection and identification methods of time-frequency aliasing signals mainly include a multi-channel signal processing technology and a single-channel signal processing technology. Among them, the multi-channel signal processing separates aliased signals by utilizing the difference of the signals in the spatial domain, but it involves a large amount of matrix operations, especially inversion operations, not only is difficult to develop by using pure Verilog, but also the consumption of hardware resources is large. A single-channel processing method represented by cyclic spectrum estimation and high-order cumulant feature extraction relates to a large number of related operations such as high-order multiplication, cyclic convolution and the like, brings great challenges to engineering realization, and is difficult to meet the real-time requirements of engineering. The above solutions are all unfavorable for efficient deployment on FPGA platforms using pure Verilog language. Disclosure of Invention In order to solve the problems in the prior art, the invention provides a Verilog-based self-adaptive time-frequency aliasing interference identification method and a Verilog-based self-adaptive time-frequency aliasing interference identification system. The technical problems to be solved by the invention are realized by the following techni