Constructing Angular-Domain CKM Via Subregion-based Interpolation and Sequential Sampling Optimization

The angular-domain channel knowledge map (A-CKM) stores channel path angles distributed in a target area, which is particularly useful for multiple-input multiple-output (MIMO) transmission. However, the spatial non-stationary nature of angular features poses a challenge to A-CKM construction, especially in complex propagation environments. For this problem, we propose a framework for A-CKM construction, which consists of key steps including subregion partition, spatial interpolation, and sampling optimization. Specifically, the target area is firstly partitioned into multiple disjoint subregions based on cluster-level angular features, ensuring local environmental similarity in every subregion. Subsequently, a subregion-based inverse distance weighting (IDW) interpolation method is proposed to construct the A-CKM from sparsely sampled measurements. In addition, a sequential sampling optimization scheme is proposed to improve the sampling efficiency, while enabling continuous map refinement over time. Simulations validate the effectiveness of our proposed A-CKM construction scheme over conventional approaches. Moreover, through a case study of MIMO beamforming, we further demonstrate that the proposed CKM construction method effectively enhances communication performance.