The Research On A Low-Complexity Parametric Localization Algorithm For Incoherently Distributed Sources

Array signal processing is widely used in signal processing,and developed rapidly increasing in recent years.It is extensively adopted in many fields such as radar monitoring,geological prospecting,seismic monitoring,communications engineering and biomedicine.The Direction of Arrival(DOA)estimation is one of the core tasks in array signal processing.The estimation methods are mainly divided into the spatial spectrum estimation and beamforming algorithm.Spatial spectrum estimation is an important issue in array signal processing,which determines the exact direction of a signal by receiving the energy distribution of the signal in space.The results of the spatial signal to DOA estimation algorithm are related to the geometry of the receiving array.Different array structures are suitable for different scenarios and the results are also different.In this thesis,the Manifold Separation Technique(MST)is used to decompose the array steering vector,and the signal covariance matrix is further expressed as a closed expression.In this way,we get an estimator suitable for any array structure,which is computationally efficient and has excellent estimation performance.This article mainly has the following two contributions:(1)The closed-form expression of the covariance matrix of the uncorrelated source signal is derived using the Manifold Separation Technique.The expression can be applied to any array structure and large diffusion angle,it almost relieves all common restrictions.(2)For the cases where the source diffusion angle is uniformly distributed(Gaussian distribution),the 2-D spatial spectrum is calculated efficiently using an unweighted(Gaussian weighted)moving average.When calculating a series of continuous spatial spectral values,the complexity of each point spectrum is simplified by using the characteristics of sliding windows.A large number of simulation results show that the proposed algorithm is superior to several classical algorithms in terms of computational complexity under the premise of good performance.