Research On The Parameter Estimation Algorithmbased On Conformal Array

The direction of arrival(DOA)estimation and joint estimation of multiple parameters of incident signal have been widely used in radar,sonar and wireless communication systems.The array constructed by the antennas,detectors and various sensors has become an important part of these systems.So far most researches have focused on uniform linear array(ULA)and two dimensions array,and the corresponding algorithms are very mature.However,the researches about the conformal array with complex geometrical structure are still at the preliminary stage,the corresponding literatures are very rare.Based on the Euler rotation transformation and the special snapshot data model of the conformal array(the polarization characteristic is considered.),the DOA and multiple parameters joint estimation for conformal array is deeply studied in this dissertation.The aim is to improve the accuracy of parameter estimation,the resolution of angular and the capacity of real-time processing.The detailed works are described as follows:In order to improve the capacity of real-time processing and estimation accuracy of the spatial spectrum estimation algorithms for conformal array,two algorithms are proposed based on propagator method(PM)and parallel factor(PARAFAC)analysis theory.First,based on the symmetry of the cylindrical conformal array,four sub-arrays that have rotation invariant relationship are constructed.According to well design of the array,the polarization information coupled in the received data can be eliminated;the DOA estimation of incident signal can be achieved by PM finally.Second,the PARAFAC analysis theory is introduced in DOA estimation for conformal array,the DOA estimation of the incident signals can be achieved by constructing the rotation invariant relationship matrices.The PARAFAC model is constructed in cumulant domain,the trilinear alternating least squares regression algorithm is used to achieve the DOA estimation finally.The proposed algorithm has better estimation accuracy than the ESPRIT(Estimation of Signal Parameters via Rotational Invariance Techniques)algorithm;the parameter pairing is no longer needed.The computer simulation verifies the effectiveness of these two proposed algorithms.For the multiple parameter joint estimation problems for conformal array,i.e.,frequency and DOA joint estimation in this dissertation,two parameter estimation methods are proposed.First,based on the state space matrix and PM,the frequency and DOA joint estimation algorithm is proposed.The frequencies are obtained based on state space matrix.With well design of the cylindrical conformal array antennas,the coupling between the polarization information and angle information can be decoupled.The DOA estimation can be achieved based on PM.Then a novel pairing method between frequencies and DOAs is proposed based on array interpolation,the Cramer-Rao bound(CRB)is derived finally.Second,the delayed correlation function combined with PARAFAC analysis theory,another frequency and DOA joint estimation algorithm is proposed.The delayed correlation function can suppress the effect of noise.The sampling both in time and space are used to construct the space-time matrix.The spectrum speaks searching and parameter pairings are no longer needed.According to PARAFAC analysis theory,the frequency and DOA joint estimation is achieved finally.The computer simulation verifies the effectiveness of these two proposed algorithms.Due to the effect of mutual coupling among the antennas,the performance of DOA estimation and the frequency and DOA joint estimation will degrade seriously.Based on ULA and conformal array,the corresponding DOA estimation algorithms are proposed to eliminate the effect of mutual coupling.For the conformal array,the corresponding frequency and DOA joint estimation algorithm is proposed as well.First,for ULA,the mutual coupling matrix(MCM)with specific structure is constructed based on auxiliary antennas,the effect of mutual coupling is eliminated.The bearing space is discretized into lots of grids exploiting the spatial sparsity characteristic.The aperture loss is compensated by taking the vectorization operator of the new covariance matrix;the compressive sensing theory is utilized to estimate the DOA of the incident signals.Then,the construction of the MCM with specific structure is extended to conformal array,the PARAFAC analysis theory is used to estimate DOA and frequency-DOA of the incident signals finally.The computer simulation verifies the effectiveness of these two proposed algorithms.