Fast Parameter Estimation Of Wideband LFM Signals Based On MP Decomposition And Array Error Correction

The study on the following three aspects has important significance in signal analysis and signal processing: (1) the method of signal denotation and the decomposition (2) fast algorithm of signal decomposition. (3) signal denotation in the applications of signal processing. The signal sparse decomposition method is a recent and concise denotation and decomposition method. It has great research value and broad application prospects. The applications of signal sparse decomposition is widely regarded, however the computational complexity of that is very high, which become a key factor in impeding its development. To promote the research and application of signal sparse denotation and signal sparse decomposition, it is necessary to study the fast algorithm. Otherwise, signal sparse denotation and decomposition can not be practical, it can only remain in the searching phases. This thesis applies genetic algorithm and particle swarm optimization algorithm to parameter estimation of wideband LFM signal, and proposes two differently fast algorithm to the estimated frequency, one is using the hybrid PSO (DS-PSO) algorithm which is based on direct search method and particle swarm optimization by, and the other is using the hybrid optimization (GA-PSO) algorithm on the basis of genetic algorithm and particle swarm optimization. Meanwhile, it will greatly improve the speed of the parameter estimation of LFM signal when the two fast algorithms of the estimated frequency are combined with sectors estimation algorithm for estimating the signal DOA.Since the array manifold has unavoidable errors in practice, the parameter estimation performance can be affected. This thesis proposes a novel algorithm based on the PSO algorithm for the estimation of uniform linear array gain and phase uncertainties. The main work and contributions of the dissertation are in the several aspects as follows:1. Detailly described a recent signal denotation theory in recent years——signalsparse denotation and decomposition, and mainly the most commonly used methods of the sparse decomposition——matching pursuit (MP) algorithm. 2. On the analysis of the time and frequency domain characteristic of wideband linear frequency modulation, proposed the method for frequency and DOA estimation which is based on matching pursuit (MP) decomposition. And introduced the structure of over-complete dictionary and MP decomposition process in detail, then present the steps of the algorithm. Analysis of the performance is given through the numerical simulations.3. Introduced the basic theory of genetic algorithm and particle swarm optimization algorithm in detail, and discussed the hybrid optimization algorithm emphatically. Contraposing the specific problems of the calculation in frequency estimation and DOA estimation of wideband LFM signal based on MP decomposition, two differently fast algorithm of the estimated frequency are proposed, one is using the hybrid PSO (DS-PSO) algorithm based on direct search method and particle swarm optimization, and the other is using the hybrid optimization (GA-PSO) algorithm on the basis of genetic algorithm and particle swarm optimization. Meanwhile introduced the theory of the fast algorithm in detail, and presented the steps of the algorithm. When the two fast algorithm of the estimated frequency are combined with a fast algorithm based sectors estimation, it will greatly accelerated the speed of the parameter estimation of LFM signal. Finally, the validity of these fast algorithm are proved by vast experimental results.4. Studied the different array error and current array error correction methods, and known the parameter estimation performance can be affected by the array errors. In this thesis, a novel algorithm based on the PSO algorithm is proposed for the estimation of uniform linear array gain and phase uncertainties. This method can exactly estimate the array gain and phase uncertainties, and can gain the exact DOA estimation by correcting the guide vector. The effectiveness and validity of this algorithm are proved by simulation results.