Research On Blind Source Separation Algorithm For Frequency Hopping Signals Under Time-Varying Channels

Frequency hopping signals are widely used in various communication fields due to their low probability of interception,strong anti-interference and high confidentiality,especially in the field of communication reconnaissance.However,when conducting communication reconnaissance,it is necessary to extract the required signal from the received mixed signal,in order to analyze,locate,and estimate the signal,so how to separate the desired signal from the mixed signal is the key premise of technology in communication reconnaissance.Blind source separation technology can effectively solve the problems in communication reconnaissance.In the actual application scenario,due to the influence of the signal receiver movement and the communication environment,the hybrid matrix in the blind source separation model is time-varying.Therefore,the blind separation algorithm for frequency hopping signals in time-varying channels is proposed.The research is carried out and the main research work is as follows:(1)Based on the blind separation of time-varying channel hopping signals under over-determined conditions,this paper improves on the basis of EASI algorithm.Crosstalk error is often used to evaluate the separation performance of the algorithm.When the energy of each component of the source signal has a large difference,the crosstalk error can not well characterize the separation performance of the algorithm.For this reason,this thesis constructs the separation index of the signal according to the convergence condition of EASI algorithm,and further describes the algorithm to the source signal.The separation performance is adjusted by the separation indicator and the crosstalk error to adjust the step size of the algorithm.In order to improve the separation speed of the signal,the algorithm is performed with a larger step value in the early iterative process.After the algorithm converges to the equilibrium point,the algorithm captures the smaller data and improves the stability of the algorithm.The algorithm continues the iteration with a smaller step size.In order to avoid the algorithm falling into local optimum,the momentum term is introduced.Experimental simulations show that the proposed algorithm improves the convergence speed and improves the steady-state error performance of the algorithm compared with the traditional algorithm.(2)For the blind separation of time-varying channel hopping signals under underdetermined conditions,the blind separation of time-varying channel hopping signals under underdetermined conditions is studied by SCA method.Firstly,the hybrid matrix is estimated,and the source signal separation on this basis.In order to satisfy the sparseness requirements of this method,the thesis transforms the signal into the time-frequency domain for analysis and processing,and uses the single-source time-frequency point to estimate the hybrid matrix.To improve the filtering precision of single-source points,the frequency-hopping signal is transformed by Gabor transform to the time frequency domain.For the noise point in the signal,the Hough transform is used to detect the time-frequency pattern of the frequency hopping signal,and the noise point is filtered out.The time-frequency points of the reservations are filtered by the single-source point detection standard,and the time-frequency ratio matrix of the observed signals is constructed after screening.For the time-varying hybrid matrix,the time of change is located,and the time-frequency ratio matrix is divided,and each time is estimated.In the different time period,The mixed matrix column vectors does not match the source signal,which are rearranged by the estimated incident angle of the source signal to obtain the final estimated mixing matrix.The experimental results show that the proposed method has good results in the accuracy of the estimated number of source signals and the accuracy of the hybrid matrix estimation.(3)For the recovery of time-varying channel hopping signals under underdetermined conditions,the sub-space projection algorithm is used to separate the source signal.The noise is introduced for the number of source signal preset by the algorithm,and a threshold is set for filtering.Discrete noise in the frequency domain;after the signal is separated in the time-frequency domain,the source signal is converted from the time-frequency domain back to the time domain by the short-time Fourier inverse transform,and the time domain hopping source signal is recovered.Experimental simulation results show The algorithm has better separation effect on the source signal.