Research On Parameter Estimation Of Frequency Hopping Signal Based On Gridless Compressived Sensing

Frequency-Hopping?FH?communication technology has the advantages of anti-interference and anti-fading,which is widely used in both military and civil communications.Therefore,the detection,reception and parameter estimation of FH signals have important practical significance.Under the background that the FH communication is developing towards the direction of super bandwidth and high hopping speed,the Compressed Sensing?CS?technology solves the problems of high sampling rate and high computation of FH signals.However,because of the Basis Mismatch?BM?problem in CS technology,the estimation accuracy of FH signals parameters based on CS is not very high.Gridless CS technology overcomes the shortcoming of BM problem and can accurately estimate sparse signals composed of linear superposition of multiple frequencies.The parameter estimation of frequency hopping signal using Gridless CS technique is an effective way to reduce the sampling rate and improve the estimation accuracy.In this paper,the parameter estimation of frequency hopping signal based on Gridless CS technology is studied.Aiming at the problem of high computational complexity and poor real-time performance of ANM algorithm,a fast atomic norm minimization algorithm is proposed.Furthermore,a high-precision algorithm for estimating hopping time based on Gridless CS technology and a joint algorithm for estimating frequency and hopping time are proposed.The simulation results show that the mean square error of the proposed algorithm decreases from 10^-1to 10^-3 for the hopping time estimation when the compression ratio is 0.5 and the SNR is 10dB.In addition,the precision of the proposed joint estimation algorithm for frequency estimation is at least twice as high as that of the contrast algorithm.The innovations of this paper are as follows:1?A Fast Atomic Norm Minimization?FANM?algorithm for frequency estimation is proposed.The proposed algorithm analyzes the reconstructed signal obtained in the early stage of ANM algorithm to roughly estimate the model order to reduce the number of parameters estimated in the Vandermonde decomposition of the Topriz matrix,improving the computing speed of the existing ANM algorithm,and enhancing the real-time performance of the line spectrum estimation.Compared with the ANM algorithm,the FANM algorithm has the advantages of low computational complexity and high estimation effectiveness.2?In the linear spectrum estimation based on gridless compressived sensing theory,the proposed algorithm has the advantages of low computational complexity and high efficiency compared with the ANM algorithm.A hopping time estimation algorithm based on numerical characteristics in compressed domain and atomic norm is proposed.Firstly,the numerical characteristics of the adjacent compressed sampling data compression through compressed sampling values,and the rough estimation of hopping time is realized.Then the atomic dictionary with full coverage of hop timimg is designed to accurately estimate of hopping time..3?A joint estimation algorithm of frequency hopping signal parameters based on mixed norm is proposed.Firstly,the compressed sampling model of multi period frequency hopping signals is constructed,and the sparsity of the coefficient matrix of two-dimensional frequency hopping signals is depicted byl_2,1-mixed norm.The problem of parameter estimation for frequency hopping signals is transformed to solve the problem of mixed norm minimization,the equivalence between mixed norm minimization and SDP problem is proved,and the accurate estimation of hopping frequency is realized by Vandermonde decomposition.Then the relationship between the amplitude of frequency and the energy is analysed.By calculating the energy ratio of the adjacent frequency in the fixed time period,The accurate estimation of hopping time is achieved by calculating the energy ratio of the adjacent hop frequencies in fixed time.