Research Of Blind Parameters Estimation Algorithm For Frequency Hopping Signals Under Complicated Environment

Frequency hopping(FH)communication has been widely used in both military and civilian communication field,because of its strong anti-interference ability,anti-interception ability,the capability against multi-path fading.At the same time,communication reconnaissance is more and more difficult as while as frequency hopping communication technology makes great progress,blind detection techniques of frequency hopping signals under complicated environment is not perfect.In response to this problem,blind parameters estimation algorithm for frequency hopping signals based on time-frequency analysis methods is systematically studied.The blind parameters estimation methods are proposed respectively from time-frequency matrix denoising preprocessing and parameters estimation and separation.Aiming at the problem of time-frequency matrix denoising preprocessing,two ways to denoise for time-frequency matrix are proposed in this dissertation: probability denoising and histogram denoising.According to the distribution of noise in frequency domain,the distribution parameters are obtained using probability distribution fitting for the sample data in probability denoising algorithm,and then denoising threshold for the signals under white gaussian noise can be calculated by setting the probability parameter of denoising;In the histogram denoising algorithm,by summarizing the regularities of distribution in timefrequency matrix and calculating the histogram of the sample data,the denoising threshold between signal points and noise points can be obtained without priori-knowledge.Finally,the data points are distinguished using denoising threshold through the method of detection noise.Simulation results show that,probability denoising can be applied to lower SNR environment than histogram denoising,and the range of SNR they fit is related to FH bandwidth.Meanwhile the former is limited to the white gaussian noise environment because of its design;histogram denoising can be used under other types of noises.Aiming at the problem of parameters estimation and separation,a parameters estimation and separation algorithm is proposed,including a parameter estimation method based on connected component labeling and parameter separation methods based on time and energy characteristics of signals.Firstly,the primary parameter list is obtained from the denoised time-frequency matrix by connected component labeling,including time parameters,frequency parameters and energy parameters.Secondly,the secondary parameter list can be obtained through duration and bandwidth parameters and joining between internal and external signals.Then,a time continuity labeling algorithm is proposed to obtain all the parameter sets meeting time continuity from the secondary parameter list.At last,according to non-overlapping time and energy-consistent,the sets are classified to obtain the target set,which is the one with the shortest immersion time.Simulation results show that,the algorithm this dissertation proposed can realize blind parameters estimation under multiple disturbance environment,such as impulse interference,burst interference and fixed-frequency interference,and satisfactory results are obtained when few overlap areas exist between FH signals and interfere signals in time-frequency matrix.