Kernel DOA estimation in nonuniform arrays

An algorithm to estimate the direction of arrival (DOA) in non uniform linear antenna arrays based on the use of Mercer's kernel nonlinear interpolation is introduced. Interpolation is followed by the synthesis of a virtual uniform array. The methodology uses kernel functions, Support Vector Regression (SVR) for the interpolation, and standard MUSIC and root- MUSIC for the DOA estimation from the interpolated data. Multiple experiments are conducted using this algorithm. The results compare the interpolated DOA detection with a standard detection algorithm using non uniform array and shows a significant improvement in performance of the proposed detector.