CN-122001723-A - Blind recognition method and system for signal modulation patterns under uncooperative communication

CN122001723ACN 122001723 ACN122001723 ACN 122001723ACN-122001723-A

Abstract

The invention discloses a signal modulation pattern blind identification method and a system under uncooperative communication, wherein the method comprises the steps of receiving signals through an antenna, processing and outputting broadband intermediate frequency signals through a front end of a receiver, detecting or selecting target signals on broadband frequency spectrums, carrying out narrowband preprocessing according to the bandwidth and center frequency of the target signals to obtain narrowband IQ data, calculating a high-order spectrum for the narrowband IQ data, measuring signal carrier frequency parameters based on a high-order spectrum searching method or a gravity center method, calculating delay spectrum or square spectrum for the narrowband IQ data, measuring signal code rate parameters based on the delay spectrum or the square spectrum, extracting signal characteristic parameters for distinguishing multiple modulation patterns, designing a multi-characteristic combined modulation classifier based on the signal characteristic parameters, and identifying various signal modulation patterns through the modulation classifier. The invention has comprehensive modulation pattern recognition, low calculation complexity and easy engineering realization, and can meet the real-time application requirement of engineering.

Inventors

HUANG WEIYING
MI WENLONG
LU CHAO
ZHU ZIHAN
FENG JIA
LV WEI
XU JING

Assignees

中国电子科技集团公司第十研究所

Dates

Publication Date: 20260508
Application Date: 20260205

Claims (10)

1. A blind recognition method for a signal modulation pattern under non-cooperative communication is characterized by comprising the following steps: The narrow-band preprocessing, namely after receiving signals through an antenna, outputting a broadband intermediate frequency signal through the front end processing of a receiver, detecting or selecting a target signal on a broadband frequency spectrum, and carrying out narrow-band preprocessing according to the bandwidth and the center frequency of the target signal to obtain narrow-band IQ data; Signal parameter measurement, namely calculating a high-order spectrum for the narrow-band IQ data, measuring signal carrier frequency parameters based on a high-order spectrum line searching method or a gravity center method, calculating a delay spectrum or a square spectrum for the narrow-band IQ data, and measuring signal code rate parameters based on the delay spectrum or the square spectrum; extracting characteristic parameters, namely extracting signal characteristic parameters for distinguishing multiple modulation patterns based on the signal carrier frequency parameters, the signal code rate parameters and the narrowband IQ data; and identifying the modulation patterns, namely designing a multi-feature combined modulation classifier based on the signal feature parameters, and identifying various signal modulation patterns through the modulation classifier.
2. The method for blind recognition of signal modulation patterns in uncooperative communication according to claim 1, wherein in the signal parameter measurement, a higher order spectrum is calculated for the narrowband IQ data, and signal carrier frequency parameters are measured based on a higher order spectrum spectral line search method or a gravity center method, comprising: measuring signal carrier frequency parameters in a mode of searching carrier frequency spectral lines by a square spectrum; if the carrier frequency spectrum line is not obvious on the fourth power spectrum, measuring the carrier frequency parameter of the signal by adopting a mode of searching the carrier frequency spectrum line by using the eighth power spectrum; if the two methods are failed to measure, the center of gravity method is adopted to measure the carrier frequency parameters of the signals.
3. The method for blind recognition of signal modulation patterns in uncooperative communication according to claim 1, wherein in the signal parameter measurement, a delay spectrum or a square spectrum is calculated for the narrowband IQ data, and a signal code rate parameter is measured based on the delay spectrum or the square spectrum, comprising: Measuring signal code rate parameters by searching code rate spectral lines through a delay spectrum; if the code rate spectral lines on the delay spectrum are not obvious, a mode of searching the code rate spectral lines on two sides of the carrier frequency by using a square spectrum is adopted, so that signal code rate parameters are measured.
4. The method according to claim 1, wherein in the feature parameter extraction, the extracted signal feature parameters include a power spectrum feature parameter, a square-average ratio parameter, a zero-center normalized instantaneous amplitude spectrum density parameter, a square spectrum carrier frequency spectrum line feature, a square spectrum code rate spectrum line feature, a fourth power spectrum carrier frequency spectrum line feature, a fourth power spectrum code rate spectrum line feature, an eighth power spectrum carrier frequency spectrum line feature, a frequency cluster point feature, a constellation point cluster point feature, and a constellation circle number cluster feature; In the modulation pattern recognition, a multi-feature combined modulation classifier is provided with a plurality of feature judgment thresholds, including a power spectrum feature parameter judgment threshold, a square-average ratio parameter judgment threshold, a zero-center normalized instantaneous amplitude spectrum density parameter judgment threshold, a square spectrum line number judgment threshold, a fourth power spectrum line number judgment threshold and an eighth power spectrum line number judgment threshold.
5. The method for blind recognition of signal modulation patterns in uncooperative communication according to claim 1, wherein in the recognition of the modulation patterns, the recognition of each type of signal modulation patterns by the modulation classifier comprises: distinguishing a first type of modulation signal from a non-first type of modulation signal through a power spectrum characteristic parameter judgment threshold, wherein the first type of modulation signal comprises AM, FM and 2ASK modulation signals; and combining the square-average ratio parameter judgment threshold and the zero center normalized instantaneous amplitude spectrum density parameter judgment threshold to further distinguish AM, FM and 2ASK modulation signals.
6. The method for blind recognition of signal modulation patterns in uncooperative communication according to claim 1, wherein in the recognition of the modulation patterns, the recognition of each type of signal modulation patterns by the modulation classifier comprises: And combining a square spectrum spectral line search judgment threshold and a fourth-order spectrum spectral line search judgment threshold to obtain carrier frequency spectral line characteristics and square spectrum code rate spectral line characteristics of square spectrum and fourth-order spectrum so as to distinguish second-class modulation signals, wherein the second-class modulation signals comprise BPSK, QPSK, 8PSK, MSK, OQPSK and MQAM modulation signals.
7. The method for blind recognition of signal modulation patterns in uncooperative communication according to claim 1, wherein in the recognition of the modulation patterns, the recognition of each type of signal modulation patterns by the modulation classifier comprises: Distinguishing a third type of modulation signal from an MFSK modulation signal through a zero-center normalized instantaneous amplitude spectral density parameter decision threshold, wherein the third type of modulation signal comprises PI/4DQPSK and 8PSK modulation signals; And combining the fourth-order spectral line search judgment threshold and the eighth-order spectral line search judgment threshold to obtain the code rate spectral line characteristic of the fourth-order spectrum and the eighth-order spectrum carrier frequency spectral line characteristic, and further distinguishing PI/4DQPSK and 8PSK modulation signals.
8. The method for identifying the signal modulation pattern in the uncooperative communication according to claim 1 is characterized in that in the modulation pattern identification, the method for identifying the modulation pattern of the MQAM modulation signal comprises the steps of carrying out matched filtering and carrier estimation on the preprocessed baseband IQ data, completing carrier synchronization based on a carrier estimation result, completing bit synchronization based on a maximum energy output principle, outputting an optimal sampling point, carrying out subtractive clustering on a sampled constellation, judging the modulation order according to the number of clustering points, carrying out subtractive clustering on the amplitude value of the baseband signal after bit synchronization when the carrier estimation or constellation point clustering judgment fails, and judging the modulation order according to the number of turns of a theoretical constellation and the minimum circle radius value.
9. The method for identifying the signal modulation pattern in the uncooperative communication according to claim 1 is characterized in that in the method for identifying the modulation pattern, the method for identifying the modulation pattern of the MFSK modulation signal comprises the steps of calculating the phase of baseband IQ data, obtaining the instantaneous frequency through phase difference, smoothing the instantaneous frequency to remove burrs generated by phase jump, carrying out histogram statistics on the smoothed instantaneous frequency, carrying out subtractive clustering, and judging the modulation order according to the number of clustered frequency points.
10. A signal modulation pattern blind identification system in uncooperative communication, comprising: the narrow-band preprocessing module is configured to output a broadband intermediate frequency signal through the front end processing of the receiver after receiving the signal through the antenna, detect or select a target signal on a broadband frequency spectrum, and perform narrow-band preprocessing according to the bandwidth and the center frequency of the target signal to obtain narrow-band IQ data; The signal parameter measurement module is configured to calculate a high-order spectrum for the narrow-band IQ data, measure signal carrier frequency parameters based on a high-order spectrum line search method or a gravity center method, calculate a delay spectrum or a square spectrum for the narrow-band IQ data, and measure signal code rate parameters based on the delay spectrum or the square spectrum; A characteristic parameter extraction module configured to extract signal characteristic parameters for distinguishing a plurality of modulation styles based on the signal carrier frequency parameter, the signal code rate parameter and the narrowband IQ data; and the modulation pattern recognition module is configured to design a multi-feature joint modulation classifier based on the signal feature parameters and recognize various signal modulation patterns through the modulation classifier.

Description

Blind recognition method and system for signal modulation patterns under uncooperative communication Technical Field The invention relates to the technical field of signal modulation pattern recognition, in particular to a method and a system for blind recognition of a signal modulation pattern under non-cooperative communication. Background The signal modulation pattern identification is a key technology for radio spectrum monitoring and spectrum resource management, has an important role in non-cooperative communication, and is a basis for further signal demodulation and interpretation processing. With the increasing complexity of communication systems, the variety of modulation patterns used in communication is becoming rich, and different modulation patterns are often required to be selected according to different application scene requirements, so as to realize the maximum utilization of spectrum resources. Signal modulation pattern recognition is widely used in communication signal monitoring, analysis and communication countermeasure, and is a necessary function of radio spectrum monitoring equipment. The existing modulation pattern recognition algorithm has a modulation pattern recognition algorithm based on maximum likelihood, calculates a maximum likelihood function, and judges the modulation pattern used according to the maximum likelihood probability value. But this method requires more a priori information, which is often difficult to obtain in uncooperative communication, and requires a larger computational effort. In recent years, deep learning is widely applied, and a modulation pattern recognition algorithm based on the deep learning is used for carrying out modulation pattern recognition through a convolutional neural network or a cyclic neural network. However, the recognition accuracy of the modulation recognition algorithm has a large correlation with the number of training samples, and the computational power requirement on a hardware platform is high, so that the modulation recognition algorithm is easy to be limited in engineering application. The modulation recognition algorithm based on the feature extraction is a type of modulation recognition algorithm which is widely focused at present, and the type of algorithm recognizes modulation patterns by selecting proper feature parameters and designing a classifier. The stability of the characteristic parameters selected by the algorithm under different signal to noise ratios has great influence on the identification accuracy, and the complexity of the classifier design has a decisive role on the execution efficiency of the algorithm. Most of the existing modulation recognition algorithms based on feature extraction mainly recognize partial modulation patterns, have limited application scenes, and have important significance in researching multi-feature joint modulation pattern recognition algorithms with more complete coverage modulation types. Disclosure of Invention In order to solve the problems, the invention provides a blind recognition method and a blind recognition system for signal modulation patterns under non-cooperative communication, which have the advantages of comprehensive coverage modulation recognition types, low calculation complexity, high recognition speed, easy engineering realization, and high engineering actual measurement recognition speed which can be up to 8 signals per second, and can meet the real-time application requirement of engineering. The technical scheme adopted by the invention is as follows: a blind recognition method of signal modulation patterns under non-cooperative communication comprises the following steps: The narrow-band preprocessing, namely after receiving signals through an antenna, outputting a broadband intermediate frequency signal through the front end processing of a receiver, detecting or selecting a target signal on a broadband frequency spectrum, and carrying out narrow-band preprocessing according to the bandwidth and the center frequency of the target signal to obtain narrow-band IQ data; Signal parameter measurement, namely calculating a high-order spectrum for the narrow-band IQ data, measuring signal carrier frequency parameters based on a high-order spectrum line searching method or a gravity center method, calculating a delay spectrum or a square spectrum for the narrow-band IQ data, and measuring signal code rate parameters based on the delay spectrum or the square spectrum; extracting characteristic parameters, namely extracting signal characteristic parameters for distinguishing multiple modulation patterns based on the signal carrier frequency parameters, the signal code rate parameters and the narrowband IQ data; and identifying the modulation patterns, namely designing a multi-feature combined modulation classifier based on the signal feature parameters, and identifying various signal modulation patterns through the modulation classifier. Further, in the signal param