Paremeter Estimation Of Frequency Hopping Signal Based On Optimal Kernel Function In Noise And Disturbance Environment

Frequency-hopping communication is one branch of the spectrum communication,which is used to spread the spectrum of the transmitted information far beyond the bandwidth of the signal itself.Due to its outstanding advantages of robust stability and information integrity,the application range of frequency-hopping technology has been extended from the initial military communications to a wide range of civil and commercial fields.Therefore,whether in military electronic warfare,radar signal processing,bluetooth transmission,wireless LAN technology and other technical directions,frequency-hopping technology has shown indispensable application potential.The instantaneous frequency of FH signal is not constant and its spectrum function changes with time,so it is a non-stationary signal.The traditional Fourier transform algorithm has limitations in processing non-stationary signals,which can only represent the parameter information which from time-domain or frequency-domain.The time-frequency analysis algorithm has developed rapidly since it was proposed.It combines the time-domain and frequency-domain,then constructs the joint function about time and frequency,and describes the non-stationary signal accurately.In this thesis,the time-frequency analysis algorithm is used to process the FH signal and estimates its parameter information.At the same time,the effectiveness of the algorithm is verified in a variety of electromagnetic environments.The main work and arrangements are as follows:This thesis firstly introduces the basic theory of FH communication and the commonly used signal analysis algorithms.Specifically,it includes the research status of time-frequency analysis algorithm,the working principle of FH communication system,simulation analysis of various time-frequency conversion algorithms.Finally,it comprehensively evaluates the advantages and disadvantages of each algorithm from the aspects of time-frequency clustering,cross interference and noise suppression,so as to lay a theoretical foundation for the improved time-frequency analysis algorithm proposed in this thesis.Secondly,a bilinear time-frequency analysis algorithm based on optimized kernel function is proposed.According to the distribution characteristics of FH signal in the fuzzy domain,the kernel function is designed pertinently.At the same time,the parameters of the kernel function are optimized by entropy measure theory to ensure that the frequency spectrum of FH signal with excellent distribution performance is obtained.Finally,the parameter estimation process of FH signal is designed by extracting the time-frequency ridge.Finally,the time-frequency performance of FH signal in the case of noise and interference is studied.According to the influence of different noises and interference signals on the frequency hopping signal in the fuzzy domain,based on the principle of preserving the self term of frequency hopping signal in the fuzzy domain and filtering out the cross interference term,the parameters of the kernel function are adjusted pertinently,then,the estimation of hop time,hop period and hop frequency is completed in turn,and the estimation accuracy is described by using the relative mean square error curve.Finally,it is compared with the SPWVD algorithm to verify the comprehensive performance of the time-frequency analysis algorithm proposed in this thesis.