Array Parameter Estimation Direction Finding Algorithm, With The Short Data

Array signal processing, an important part of signal processing, which is widely used in many fields such as radar, communications, sonar, seismic exploration, astronomy, has been developed over the recent three decades. This dissertation offers an in-depth study of joint multi-parameter estimation and direction-of-arrival(DOA)estimation with a small sample size.Firstly, the algorithms for estimating the number of sources, joint estimation of frequency and 2-D DOA are studied. We simply introduce the traditional methods and then propose a new method based on pseudocovariance matrix. The proposed method provides superior estimation performance to the traditional ones by taking advantage of sufficient temporal correlations to extend the array. The reconstructed pseudocovariance matrix smoothes the noise by averaging the data which makes the method can work well in the presence of color noise. The simulation results demonstrate that the proposed method performs better than the traditional ones. We study the method of joint estimation of frequency and 2-D DOA. After analyzing some algorithms of joint estimation of frequency and angles, a new method of joint multi-parameter estimation by extending the pseudocovariance matrix method to the L-shaped array is proposed. The proposed method provides very attractive 3-D estimation performance, high DOA resolution by making use of the temporal and spatial information of the array. The numerical simulations demonstrate the validity of the proposed method.Then, the estimation of DOA with a small sample size is studied. The performance of the traditional DOA estimators degrades rapidly in the case of small sample size, which does not satisfy the requirement of DOA estimation. The small sample size, caused by the non-alignment of the antenna, the widely used low probability of intercep(tLPI)technology, brings a new challenge to the array direction-finding system. To analyze the problem, the effect of the sample size on the performance of DOA estimator is studied by random matrix theory. Two DOA estimators based on adjusting the subspace are discussed and parametric iterative adaptive approach for amplitude and phase estimation(P-IAA-APES), which suits small sample size, is proposed. The proposed method has very attractive estimation performance, high DOA resolution in the case of small sample size for avoiding the eigenvalue decomposition(EVD)and calculating the sample correlation matrix iteratively to approximate the truth.