Underdetermined Super-resolution Direction Of Arrival Estimation With Uniform Circular Array

In the community of array signal processing, high-resolution direction-of-arrival(DOA) estimation has received considerable attention in past two decades. The arrayantenna can be of various structures, such as uniform linear array (ULA), uniformcircular array (UCA), non-uniform linear array (NLA) and so on. Among these arraystructures, the UCA has been widely used in radar and sonar due to its good property.However, since the number of elements in the UCA must be larger than the number ofsource signal impinging upon the UCA. Thus, the existing DOA estimationalgorithms cannot handle the underdetermined situation. To cure this problem, thisdissertation addresses an underdetermined DOA estimation scheme forquasi-stationary signals impinging upon the UCA.This dissertation first introduces several classical DOA estimation methods andthen addresses three schemes for underestimate DOA estimation in UCA application.These UCA DOA estimation schemes convert the steering matrix to Vandermondematrix which is convenient to be calculated. Among these three methods, modeexcitation scheme and interpolated scheme requires smaller complexity burden butlarger DOA estimation declination than the manifold separation technique. Throughvectorizing the covariance matrix, the UCA gains additional virtual sensors andtherefore enlarges its aperture. Based on the manifold separation technique, the DOAestimation problem is converted to a polynomial rooting problem. This schemereduces its complexity burden. Simulation results are included to demonstrate theeffectiveness of the proposed scheme.