Parameter Estimation For Frequency-hopping Signals Based On Compressed Sensing

Frequency hopping(FH)signal has been widely used in military communications because of its excellent characteristics such as low interception and anti-jamming,and its investigative technology has always been a hot spot of research.In recent years,the high sampling rate and massive data processing problem due to high band broadband bring great challenges to the detection technology of FH signal.The application of compressed sensing(CS)to FH signal parameter estimation can effectively solve the above problems.In this paper,the parameter estimation algorithm for FH signal based on CS is studied.The hopping time,the hopping period and the hopping frequency of the transmitted FH signal are estimated according to the received compressed sampling information.The work of this thesis can be summarized as following aspects:(1)Most existing parameter estimation algorithms do not consider the structural characteristics of FH signal,and have the disadvantages of high computational complexity or low estimation accuracy in low signal-to-noise ratio circumstance.To solve this problem,a parameter estimation algorithm for FH signal in compressed domain based on sliding window and atomic dictionary is proposed.Hopping time and hopping frequency are estimated by establishing a sliding window model based on the known hopping speed.But the estimation precision of hopping time is low,and the time sequence of hopping frequency is difficult to be determined only use this method.Therefore,the Fourier orthogonal basis of block diagonalization is constructed to determine the time sequence of hopping frequency.Based on the estimated frequency,a redundant dictionary that matches the local time-frequency characteristics of FH signals is constructed,and the hopping time is precisely determined by matching pursuit algorithm.Simulation results show that this algorithm can significantly reduce the sampling data and computational complexity,while maintaining the high accuracy estimation.(2)The segment length is difficult to select in the traditional parameter estimation algorithm for FH signal based on subsection.If the segment length is too long,the frequency resolution will increase,but the time resolution will decrease,and vice versa.The algorithm proposed in this paper adds the atom matching process to the segmentation idea.Frequency resolution is considered firstly when selecting the segment length.An atomic dictionary that can represent the local time-frequency characteristics of the FH signal is constructed based on the estimated frequency.The hopping time is estimated accurately according to the correlation between compressed sampling signal and atoms,and the problem of the contradiction between time and frequency resolution is solved.The algorithm takes advantage of the sparse structure of FH signal in the frequency domain.The blind parameter estimation of the FH signal is realized directly according to the compression measurement value,and the parameter accuracy estimation is improved on the premise of significantly reducing the complexity.(3)The existing parameter estimation algorithms for multiple FH signals based on CS often take the number of source signals as a priori information,which is difficult to obtain in the actual detection process.A parameter estimation algorithm for multiple FH signals based on compression reconfiguration is proposed in order to solve the problem above.The received signals from the multiple receiving antennas are reconstructed.Comparing the location of the peak value in the reconstructed vector,the peaks corresponding to the positions in all reconstruction vectors are generated by signals,and the number of peaks is the sparsity of multiple FH signals,that is,the number of source signals.The OMP algorithm is used to estimate the time frequency matrix under the condition of the known sparsity.Experimental results show that the parameters of multiple FH signals in compressed domain can be estimated accurately throngh this algorithm.