Research On Parameter Estimation And Separation Technology Of Multi-component Linear Frequency Modulation Signals

As transmission channels and electromagnetic environments become increasingly complex and diverse,Linear Frequency Modulation(LFM)signals play an increasingly important role in signal processing fields such as communications,radar and biomedicine.Therefore,how to detect and estimate the target signal more accurately and efficiently,and then extract the relevant attribute features has gradually become a research problem of great significance and application value.Starting from the center frequency and frequency modulation slope of LFM signals,this article investigates the unique advantages of Fractional Fourier Transform(FRFT)and sparse representation theory in LFM signal processing.The problem is transformed into the search of the optimal rotation order and the weighted reconstruction of the atomic parameters of the LFM signal,and two methods for parameter estimation and separation of multi-component LFM signals are proposed.It provides a new signal processing approach and solution for the detection,time-frequency analysis,parameter estimation and separation of multiple different component LFM signals in the real environment.The main work content of this paper includes the following two aspects:(1)Aiming at the problems of low accuracy and large amount of computation in estimating center frequency and frequency modulation slope of LFM signals,a multi-component LFM signals parameter estimation and separation method based on the fourth-order Origin Moment of Fractional Spectrum(OMFr S)is investigated.The key innovation idea of this method is to use the fourth-order OMFr S to develop a reasonable search strategy of the optimal rotation order to achieve parameter estimation,and then put forward the signal step-by-step filtering technology in the FRFT domain to achieve the signal separation we are interested in.By virtue of its excellent impulse characteristics and anti noise performance,it overcomes the problems of low estimation accuracy and high computational complexity in traditional two-dimensional search and time-frequency analysis methods.Different from the existing solutions,this method is efficient and stable,and can avoid the interference problem between multiple signals with close time-frequency distances,overlapping and cross components.In this paper,the error analysis of LFM signals is carried out and the Cramer-Rao Lower Bound(CRLB)for parameter unbiased estimation is derived.The proposed method has been analyzed and compared with traditional and existing methods in multiple aspects,and simulation experimental results have proven the accuracy and effectiveness of this method in noisy environments.(2)Aiming at the limitation of LFM signal parameter estimation accuracy,a multi-component LFM signal parameter estimation and separation method based on Stage-wise Orthogonal Matching Pursuit(St OMP)and dictionary dynamic update is studied.In this method,the signal to be represented is modeled as a weighted combination of dictionary elements,and the parameter estimation of each component LFM signal whose energy is distributed in the whole observation interval of the dictionary matrix is realized by constrainting its weight sparsity.The limitation of signal length and sampling frequency on the accuracy of parameter estimation is overcome.Roughly estimate the FRFT parameters of the LFM signal to obtain prior data for constructing the dictionary center atom for each signal component.The dynamic update mode of dictionary matrix is designed according to the characteristics of parameter structure of LFM signal,and the parameter space is continuously redivided according to the matching results and search step size,which improves the accuracy of parameter estimation and reduces the calculation amount of algorithm.This method improves the reconstruction accuracy problem of the existing St OMP algorithm,achieving higher accuracy and fewer iterations of signal reconstruction by optimizing and screening candidate atoms,and improving the reliability of the separation results of multi-component LFM signals.In this paper,the advantages and disadvantages of various methods are analyzed and compared,and the accuracy and efficiency of the method are verified according to the experimental results from different angles.