Research Of Blind Reconnaissance Algorithm For Hop Signals Based On Time-frequency Analysis

Frequency-hopping communication has excellent anti-interference ability,and is widely used in military communication.Nowadays,it is more and more difficult to carry out electronic reconnaissance under complex electromagnetic environment.With time-frequency analysis,this paper proposes complete blind reconnaissance algorithm,based on the study of the blind detection algorithm,blind source separation algorithm and blind parameter estimation algorithm for frequency-hopping signal.Firstly,this paper studies the basic principle and non-stationary characteristics of frequency-hopping signal with time-frequency analysis.According to the characteristics of frequency-hopping signal in time-frequency domain,this paper proposes a kind of blind detection algorithm based on clustering,which can extract frequency-hopping signal under complex electromagnetic environment.And the simulation results show that the algorithm has good detection performance.In the study of the underdetermined blind source separation,this paper studies the linear delay mixed signal model by means of sparse component analysis algorithm,which estimates the mixed matrix firstly and then separates source signal.For the estimation of mixed matrix,the time-frequency ratio algorithm,based on single source detection,is adopted;for the separation of source signal,the subspace projection separation algorithm is adopted,which improves algorithm flow and debase the requirement of sparse hypothesis.Compared with other algorithms,simulation results show that the blind source separation algorithm of this paper has better performance.As for the study of blind parameter estimation,this paper mainly estimates the frequency and hopping cycle of frequency-hopping signal.Based on the results of blind detection and blind source separation,this paper proposes two kinds of parameter estimation algorithm and analyzes their characteristics theoretically.At last,this paper applies blind estimation algorithm for engineering implementation on the FPGA platform.On the platform,an embedded processor is built and the algorithm is implemented by means of C language.The test results show that the system can realize parameter estimation function and meet the predetermined performance index.