On the Common AOA Error in CKM-Based Integrated Sensing and Communications

In integrated sensing and communications (ISAC), the channel information can be acquired via the channel knowledge map (CKM). However, due to the lack of the precise knowledge of the user equipment's heading, angle of arrival (AOA) estimations of all paths share an identical error, which is defined as the common AOA error (CAE). In this letter, we prove that the CAE imposes rotational effects on the unitary matrices of the singular value decomposition (SVD) of the channel matrix, leading to non-orthogonality among data layers for the SVD-based precoding. Then we develop two algorithms utilizing the rotational effect to estimate and eliminate the CAE. The two algorithms are suitable for scenarios with low and high Rician K-factors, respectively. By applying the proposed algorithms, accurate wideband channel information can be attained based on narrowband reference signals. Simulation results show that the proposed algorithms obtain higher precoding performance than the baseline method.