Enhancement of Spatial-Wideband Channel Estimation Based on Beam Split Pattern Detection

Channel state information (CSI) is crucial to achieve the promising gain in millimeter-wave (mmWave) wideband system. Several existing wideband channel estimation schemes only address the frequency-wideband effect, assuming common sparsity supports across subcarriers yet ignoring the spatial-wideband effect, where beams are split into different physical directions at different frequencies. Considering spatial-wideband effect, the beam split pattern (BSP), a set which contains different spatial directions for different subcarriers, can be exploited to perform BSP detection (BSPD) based channel estimation scheme. However, the straightforward BSPD scheme shows poor accuracy of channel estimation due to the limited angle-domain resolution. In this paper, we propose two refined channel estimation schemes for more accurate CSI acquisition. Firstly, direct refinement of BSPD (DR-BSPD) based channel estimation is developed through explicit angle-domain resolution enhancement. Secondly, multiple-subcarrier aided BSPD (MSA-BSPD) based channel estimation is proposed to implicitly refine the physical channel direction through exploiting the property that multiple subcarriers with different array gains are matched to one angle-domain sample. Numerical results demonstrate the superior performance of the two proposed schemes against the existing BSPD based channel estimation scheme.