Frequency Hopping Signal Detection And Parameter Estimation Under Stable Alpha Distribution

With the continuous deepening of research,frequency hopping technology has become one of the representatives of today's communication anti-jamming technology.It has a very good ability to resist interception and fading interference,and has been widely used in military battlefields and civilian communications.The most typical feature of a frequency hopping signal is that the frequency changes continuously with time,so it is far from enough to analyze and process it only by Fourier transform.A method that can combine time and frequency is needed to perform it.In view of the problem that the noise model of most of the current algorithms is Gaussian noise,and the effect of the algorithm under the non-Gaussian model is not good,this thesis mainly studies how to detect the frequency hopping signal more accurately and carry out parameter estimation with less error when the communication environment is a stable distribution of alpha.First of all,this thesis studies the existing time-frequency analysis methods,mainly introducing linear and bilinear time-frequency analysis methods.On the basis of the two,the basic principle of combined time-frequency distribution is expounded,and each time-frequency analysis algorithm is evaluated through the calculation amount and information entropy of each algorithm.Subsequently,the alpha stable distribution is introduced,and the impulse noise is modeled and simulated.Finally,in order to resist the influence of the noise,the concept of fractional low-order operation is studied,and the time-frequency analysis algorithm based on fractional low-order is introduced.Secondly,since most of the current literature on frequency hopping signal detection is based on Gaussian model,and the actual environment noise is mostly non-Gaussian model,so this thesis uses the stable distribution of alpha as the environment,and proposes the combination of fractional low-order operations and combining time-frequency analysis methods.The algorithm is subsequently combined with morphological processing to extract the time-frequency detection quantity of the frequency hopping signal.The simulation results show that the detection probability of the algorithm can reach 90% when the generalized signal-to-noise ratio is greater than-1dB,and it has higher detection ability and better detection effect.Finally,when the generalized signal-to-noise ratio is low,in view of the problem that the traditional time-frequency matrix denoising algorithm will filter out a large number of signals,this thesis analyzes and compares the traditional mean threshold,median threshold and adaptive threshold,and proposes a weighted threshold algorithm based on the traditional threshold to be used in the combined time-frequency algorithm to process frequency hopping signals.Subsequently,combined with the morphological processing to obtain the time-frequency ridge of the signal,based on the analysis of the time-frequency ridge,the formula for the parameter estimation of the frequency hopping signal is obtained,and the method of evaluating the parameter estimation value is explained.The simulation results show that when the generalized signal-to-noise ratio of the weighted threshold proposed in this paper is lower than 2dB,the error obtained by parameter estimation is obviously lower than that obtained by the traditional threshold,and is closer to the actual parameter value of the frequency hopping signal.