Research On Optimization Method For Recognition Of Radiation Source Modulation Signal

Radiation source modulation recognition is one of the essential links in electronic reconnaissance systems,directly affecting the performance of electronic countermeasure equipment,and is related to subsequent operational decision-making.It is an important basis for judging enemy weapon threats and plays a vital role in electronic warfare.With the continuous development of military technology,the emergence of new modulation signals also puts forward new requirements for radiation source modulation recognition.In view of the recognition problems of known and unknown modulation signals in complex environments,research on known and unknown radiation source modulation signal identification,known radiation source modulation signal recognition,and unknown modulation signal incremental recognition technologies is conducted.Firstly,the basic principles of radiation source signal modulation recognition are studied,the generation mechanism of modulation signals is expounded,the mathematical model of radiation source signals is established,and the characterization forms of radiation source modulation signals are analyzed in conjunction with time-frequency transformation methods.Secondly,research on known radiation source modulation signal recognition is conducted.Based on the idea of spectral filtering and joint coding branch structure,a spectral mapping network is designed.By integrating differential evolution and Cauchy mutation mechanisms,swarm intelligence algorithms are improved,classifier combination control parameters are optimized,recognition models are optimized,and a known radiation source modulation recognition algorithm based on spectral mapping network and hybrid butterfly optimization is proposed.Next,the identification of unknown radiation source modulation signals is studied.By fusing spatial channel attention and Nash equilibrium principle,a discrimination algorithm for unknown modulation signals based on spatial spectrum attention generative adversarial network is proposed.Self-organizing iterative theory is introduced,a multi-dimensional comprehensive fitness function is designed,hybrid butterfly technology is combined,and an iterative algorithm is optimized,resulting in a clustering algorithm for unknown radiation source modulation signals based on improved iterative self-organizing data analysis.Finally,research on incremental recognition of unknown radiation source modulation signals is conducted.Based on the category incremental mechanism,dynamic incremental network structure is optimized,and a loss function is combined to provide an incremental modulation recognition scheme based on spatial spectrum coding embedding.Innovatively,the nearest ideal solution is introduced,the kernel mapping idea is combined,the ternary loss is improved,and the incremental scheme is further optimized,resulting in a radiation source incremental modulation recognition algorithm based on the ideal ternary loss.On this basis,the radiation source modulation recognition system is completed.Simulation results show that under the condition of a signal-to-noise ratio of 0d B,the designed system’s correct identification rate of known and unknown radiation source signal modulation types reaches 94.88%,the known radiation source modulation recognition rate reaches 90.48%,and the unknown radiation source clustering and incremental recognition accuracy rates reach 94.04% and 88.14% respectively.This enhances the reliability of radiation source modulation signal recognition algorithms and provides new ideas for battlefield radiation source reconnaissance.