Research On Type Recognition And Parameter Estimation Of Electromagnetic Signals In Complex Scenes

In this era of information explosion,business demands are also increasing,which has brought about continuous improvement in signal transmission quality requirements.While enjoying the dividends brought about by the rapid development of the information technology industry,we must also deal with its negative effects,especially a series of new challenges in the field of electromagnetic signal parameter estimation and type identification.It is worth noting that whether the type of electromagnetic signal can be accurately identified and whether the parameters can be accurately estimated is an indispensable foundation for the research and application of various fields.This paper focuses on the blind equalization,type recognition and parameter estimation of electromagnetic signals in complex scenarios,and realizes the blind equalization of electromagnetic signals in multipath channels,the feature extraction and type recognition of electromagnetic signals,and the intelligent analysis of different signals and different parameters.First,a blind equalization algorithm of feedforward neural network based on the optimization of the intrusion weed algorithm(IWO)is designed.In this paper,based on the feedforward neural network blind equalization algorithm,combined with the intrusion weed optimization algorithm,the two algorithms cooperate to optimize the initialization process of the neural network weights,achieve the purpose of balancing,and improve the convergence speed.Secondly,an electromagnetic signal type recognition system based on signal feature extraction and machine learning algorithms is proposed.This paper proposes an electromagnetic signal recognition system.The key of the system is to extract features that can reflect the characteristics of different types of electromagnetic signals.First,extract multi-dimensional features including frequency domain features,instantaneous features,time domain features,transform domain features,etc.from the received electromagnetic signals such as communications and radar,and then use a variety of machine learning algorithms to complete the training of the classification model,verifying The effectiveness of the electromagnetic signal identification system.Finally,on the basis of signal type recognition,this paper further studies the electromagnetic signal parameter estimation in the case of insufficient prior information of the electromagnetic signal.This article combines the Metropolis-Hastings(MH)sampling in the MCMC algorithm with the electromagnetic signal parameter estimation.And through simulation experiments and Universal Software Radio Peripheral(USRP)hardware measurement experiments,the effectiveness of the parameter estimation algorithm is verified.