An Analytical Covariance Matrix Change Detector

Next-generation sixth generation (6G) technologies require massive antenna deployments, increasing overhead for channel estimation and network reconfiguration. Covariance matrix (CM)-based methods help reduce this overhead but face two challenges: the need for many channel samples and the matrix’s slow but non-negligible variation. This paper proposes a two-step change point detection algorithm to address both issues. The first step detects angle of arrival (AoA) shifts to trigger channel updates only when necessary. The second evaluates spectral efficiency (SE) loss to decide when network reconfiguration is required. The proposed method strikes a balance between overhead, performance, and computational complexity, making it well suited for next-generation large-scale dynamic wireless systems.