| Summary: | Direction-of-arrival (DOA) estimation in the presence of gain-phase errors for uniform-circular-array (UCA) is addressed in this paper. Thanks to the uniformity and cycle property of UCA, the long data vector constructed with the proper entries of the Hadamard product of autocorrelation matrix and its conjugate, can be regarded as the received data corresponding to virtual two-dimensional sources impinging on virtual multiple UCAs. This finding indicates that azimuths and elevations of virtual sources related to DOAs of original sources can be decoupled and estimated via multiple-UCAs-estimating signal parameter via rotational invariance techniques proposed in this paper. Furthermore, two novel methods named as spatial-filtering-method and parameter-estimation-method are proposed to cope with one-component, which affects the DOA estimation of adjacent sources. Compared with the one-component elimination method proposed by Cao and Ye, previously mentioned two methods have closed-form solutions and they do not require iteration. The proposed method requires neither the calibration sources nor multidimensional parameter search. Furthermore, it can cope with more sources and work for various apertures of UCA. Simulations verify the effectiveness of the proposed method.
|
|---|